To our families

Mathematical Principles of
Theoretical Physics

(理论物理的数学原理)

Tian Ma Shouhong Wang (马天 汪守宏)

Responsible Editors: Jingke Li and Yanchao Zhao

Copyright© 2015 by Science Press

Published by Science Press
16 Donghuangchenggen North Street
Beijing 100717, P. R. China

Printed in Beijing

All rights reserved. No part of this publication may be

reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, photocopying,
recording or otherwise, without the prior written permission of
the copyright owner.

ISBN 978-7-03-045289-4
Chapter 1
General Introduction

1.1 Challenges of Physics and Guiding Principle

Challenges of theoretical physics

Physics is an important part of science of Nature, and is one of the oldest science disci-
plines. It intersects with many other disciplines of science such as mathematics and chem-
istry.
Great progresses have been made in physics since the second half of the 19th century.
The Maxwell equation, the Einstein special and general relativity and quantum mechanics
have become cornerstones of modern physics. Nowadays physics faces new challenges. A
partial list of most important and challenging ones is given as follows.

1. What is dark matter?

2. What is dark energy?

Dark matter and dark energy are two great mysteries in physics. Their gravitational
effects are observed and are not accounted for in the Einstein gravitational field equations,
and in the Newtonian gravitational laws.

3. Is there a Big-Bang? What is the origin of our Universe? Is our Universe

static? What is the geometric shape of our Universe?

These are certainly most fundamental questions about our Universe. The current dom-
inant thinking is that the Universe was originated from the Big-Bang. However, there are
many unsolved mysteries associated with the Big-Bang theory, such as the horizon problem,
the cosmic microwave radiation problem, and the flatness problem.

4. What is the main characteristic of a black hole?

Black holes are fascinating objects in our Universe. However, there are a lot of confu-
sions about black holes, even its very definition.

5. Quark Confinement: Why has there never been observed free quarks?
2 Chapter 1 General Introduction

There are 12 fundamental subatomic particles, including six leptons and six quarks. This
is a mystery for not being able to observe free quarks and gluons.

6. Baryon asymmetry: Where are there more particles than anti-particles?

Each particle has its own antiparticle. It is clear that there are far more particles in this
Universe than anti-particles. What is the reason? This is another mystery, which is also
related to the formation and origin of our Universe.

7. Are there weak and strong interaction/force formulas?

We know that the Newton and the Coulomb formulas are basic force formulas for grav-
itational force and for electromagnetic force. One longstanding problem is to derive similar
force formulas for the weak and the strong interactions, which are responsible for holding
subatomic particles together and for various decays.

8. What is the strong interaction potential of nucleus? Can we derive the

Yukawa potential from first principles?

9. Why do leptons not participate in the strong interaction?

10. What is the mechanism of subatomic decays and scattering?

11. Can the four fundamental interactions be unified, as Einstein hoped?

Objectives and guiding principles

The objectives of this book are

1) to derive experimentally verifiable laws of Nature based on a few fundamental math-

ematical principles, and

2) to provide new insights and solutions to some outstanding challenging problems of

theoretical physics, including those mentioned above.

The main focus of this book is on the symbiotic interplay between theoretical physics
and advanced mathematics. Throughout the entire history of science, the searching for
mathematical representations of the laws of Nature is built upon the believe that the Nature
speaks the language of Mathematics. The Newton’s universal law of gravitation and laws of
mechanics are clearly among the most important discoveries of the mankind based on the
interplay between mathematics and natural sciences. This viewpoint is vividly revealed in
Newton’s introduction to the third and final volumes of his great Principia Mathematica: “I
now demonstrate the frame of the system of the world.”
1.2 Law of Gravity, Dark Matter and Dark Energy 3

It was, however, to the credit of Albert Einstein who envisioned that the laws of Nature
are dictated by a few fundamental mathematical principles. Inspired by the Albert Einstein’s
vision, our general view of Nature is synthesized in two guiding principles, Principles 2.1
& 2.2, which can be recapitulated as follows:

Nature speaks the language of Mathematics: the laws of Nature 1) are repre-
sented by mathematical equations, 2) are dictated by a few fundamental prin-
ciples, and 3) always take the simplest and aesthetic forms.

1.2 Law of Gravity, Dark Matter and Dark Energy

Gravity is one of the four fundamental interactions/forces of Nature, and is certainly the first
interaction/force that people studied over centuries, dating back to Aristotle (4th century
BC), to Galileo (late 16th century and early 17th century), to Johannes Kepler (mid 17th
century), to Isaac Newton (late 17th century), and to Albert Einstein (1915).

Newtonian gravity

Newton’s universal law of gravity states that the gravitational force between two massive
objects with m and M is given by
GmM
F =− 2 , (1.2.1)
r
which is an empirical law.
Einstein’s General theory of relativity
One of the greatest discovery in the history of science is Albert Einstein’s general the-
ory of relativity (Einstein, 1915, 1916). He derives the law of gravity, his gravitational
field equations by postulating two revolutionary fundamental principles: the principle of
equivalence (PE) and the principle of general relativity (PGR):

1) PE says that the space-time is a 4-dimensional Riemannian manifold

{M , g µν } with the Riemannian metric {g µν } representing the gravita-
tional potential;
2) PGR says that the law of gravity is covariant under general coordinate
transformations of both the inertial and non-inertial reference frames;
3) PGR, together with simplicity principle of law of Nature, uniquely dictates
the Lagrangian action, also called the Einstein-Hilbert functional:
! ! "
8π G √
LEH ({g µν }) = R+ S −gdx; (1.2.2)
M c4
4) The Einstein gravitational field equations are then derived using the least
action principle, also called the principle of Lagrangian dynamics (PLD):
4 Chapter 1 General Introduction

1 8π G
Rµν − gµν R = − 4 Tµν . (1.2.3)
2 c
This is the most profound theory of science. The PGR is a symmetry principle, and
the law of gravity, represented as a set of differential equations (1.2.3), is dictated by this
profound and simple looking symmetry principle. The connection to the Newtonian gravi-
tational law (1.2.1) is achieved through the following Schwarzschild solution of the Einstein
field equations in the exterior of a ball of spherically symmetric matter field with mass M:
! " ! "
2 2MG 2 2 2MG −1 2
ds = − 1 − 2 c dt + 1 − 2 dr + r2 d θ 2 + r2 sin2 θ d ϕ 2 . (1.2.4)
c r c r
Here the temporal component of the metric and the gravitational force of the ball exerted on
an object of mass mass m are
! " ! "
2MG 2 GMm
g00 = − 1 − 2 = − 1+ 2ψ , F = −m∇ψ = − 2 . (1.2.5)
c r c r

New law of gravity (Ma and Wang, 2014e)

Gravity is the dominant interaction governing the motion and structure of the large scale
astronomical objects and the Universe. The Einstein law of gravity has been a tremendous
success when it received many experimental and observational supports, mainly in a scale
of the solar system.
Dark matter and dark energy are two great mysteries in the scale of galaxies and beyond
(Riess et al., 1989; Perlmutter et al., 1999; Zwicky, 1937; Rubin and Ford, 1970). The
gravitational effects are observed and are not accounted for in the Einstein gravitational
field equations. Consequently, seeking for solutions of these two great mysteries requires
a more fundamental level of examination for the law of gravity, and has been the main
inspiration for numerous attempts to alter the Einstein gravitational field equations. Most
of these attempts, if not all, focus on altering the Einstein-Hilbert action with fine tunings,
and therefore are phenomenological. These attempts can be summarized into two groups:
(a) f (R) theories, and (b) scalar field theories, which are all based on artificially modifying
the Einstein-Hilbert action.
Our key observation is that due to the presence of dark matter and dark energy, the
energy-momentum tensor Tµν of visible baryonic matter in the Universe may not conserved:
∇µ Tµν ̸= 0. which is a contradiction to the Einstein field equations (1.2.3), since the left-
# $
µ 1
hand side of the (1.2.3) is conserved: ∇ Rµν − gµν R = 0. The direct consequence of
2
this observation is to take the variation of the Einstein-Hilbert action under energy momen-
tum conservation constraint:
d %%
(δ LEH (g µν ), X) = % LEH (g µν + λ Xµν ) = 0 ∀X = {Xµν } with ∇µ Xµν = 0. (1.2.6)
d λ λ =0
1.2 Law of Gravity, Dark Matter and Dark Energy 5

The div-free condition, ∇ µ Xµν = 0, imposed on the variation element represents energy-
momentum conservation. We call this variational principle, the principle of interaction
dynamics (PID), which will be discussed hereafter again.
Using PID (1.2.6), 1 we then derive the new gravitational field equations (Ma and Wang,
2014e):

1 8π G
Rµν − gµν R = − 4 Tµν − ∇µ ∇ν φ , (1.2.7)
# 2 c $
8 π G
∇µ Tµν + ∇µ ∇ν φ = 0. (1.2.8)
c4

In summary, we have derived new law of gravity based solely on first principles, the PE, the
PGR, and the constraint variational principle (1.2.6):

Law of gravity (Ma and Wang, 2014e)

1) (Einstein’s PE). The space-time is a 4D Riemannian manifold {M , g µν },

with the metric {g µν } being the gravitational potential;
2) The Einstein PGR dictates the Einstein-Hilbert action (1.2.2);
3) The gravitational field equations (1.2.7) and (1.2.8) are derived using
PID, and determine gravitational potential {g µν } and its dual vector field
Φµ = ∇µ φ ;
4) Gravity can display both attractive and repulsive effect, caused by the
duality between the attracting gravitational field {g µν } and the repulsive
dual vector field {Φµ }, together with their nonlinear interactions gov-
erned by the field equations (1.2.7) and (1.2.8).

We remark that it is the duality and both attractive and repulsive behavior of gravity that
maintain the stability of the large scale structure of our Universe.

Modified Newtonian formula from first principles (Ma and Wang, 2014e)

Consider a central matter field with total mass M and with spherical symmetry. We
derive an approximate gravitational force formula:
# $
1 k0
F = mMG − 2 − + k1r , k0 = 4 × 10−18km−1 , k1 = 10−57 km−3 . (1.2.9)
r r
1 The new field equations (1.2.7) can also be equivalently derived using the orthogonal decomposition theo-

rem, reminiscent to the Helmholtz decomposition, so that

c4
Tµν = T!µν − ∇ µ Φν .
8π G
See Chapter 4 and (Ma and Wang, 2014e) for details.
6 Chapter 1 General Introduction

Here the first term represents the Newton gravitation, the attracting second term stands for
dark matter and the repelling third term is the dark energy. We note that the dark matter
property of the gravity and the approximate gravitational interaction formula are consistent
with the MOND theory proposed by (Milgrom, 1983); see also (Milgrom, 2014) and the
references therein. In particular, our modified new formula is derived from first principles.
Dark matter and dark energy: a property of gravity (Ma and Wang, 2014e)
We have shown that it is the duality between the attracting gravitational field {g µν } and
the repulsive dual field {Φ µ = ∇µ φ } in (1.2.7), and their nonlinear interaction that gives
rise to gravity, and in particular the gravitational effect of dark energy and dark matter.
Also, we show in (Hernandez, Ma and Wang, 2015) that consider the gravitational field
outside of a ball of centrally symmetric matter field. There exist precisely two physical
parameters dictating the two-dimensional stable manifold of asymptotically flat space-time
geometry, such that, as the distance to the center of the ball of the matter field increases,
gravity behaves as Newtonian gravity, then additional attraction due to the curvature of
space (dark matter effect), and repulsive (dark energy effect). This also clearly demonstrates
that both dark matter and dark energy are just a property of gravity.
Of course, one can consider dark matter and dark energy as the energy carried by the
gravitons and the dual gravitons, to addressed in the unified field theory later in this book.

1.3 First Principles of Four Fundamental Interactions

The four fundamental interactions of Nature are the gravitational interaction, the electro-
magnetism, the weak and the strong interactions. Seeking laws of the four fundamental
interactions is the most important human endeavor. In this section, we demonstrate that
laws for the four fundamental interactions are determined by the following principles:

1) the principle of general relativity, the principle of gauge invariance, and

the principle of Lorentz invariance, together with the simplicity principle
of laws of Nature, dictates the Lagrangian actions of the four interactions,
and
2) the principle of interaction dynamics and the principle of representation
invariance determines the field equations.

Symmetry principles
We have shown that for gravity, the basic symmetry principle is the Einstein PGR, which
dictates the Einstein-Hilbert action, and induces the gravitational field equations (1.2.7)
using PID.
Quantum mechanics provides a mathematical description about the Nature in the molecule,
the atomic and subatomic levels. Modern theory and experimental evidence have suggested
1.3 First Principles of Four Fundamental Interactions 7

that the electromagnetic, the weak and the strong interactions obey the gauge symmetry.
In fact, these symmetries and the Lorentz symmetry, together with the simplicity of laws
of Nature dictate the Lagrangian actions for the electromagnetic, the weak and the strong
interactions:

Symmetry Dictates Actions of fundamental interactions:

(a) The principle of general relativity dictates the action for gravity, the Einstein-
Hilbert action.
(b) The principle of Lorentz invariance and the principle of gauge invari-
ance, together with the simplicity principle of laws of Nature, dictate the
Lagrangian actions for the electromagnetic, the weak and the strong in-
teractions.

This represents clearly the intrinsic beauty of Nature.

The abelian U(1) gauge theory describes quantum electrodynamics (QED). The non-
abelian SU(N) gauge theory was originated from the early work of (Weyl, 1919; Klein,
1938; Yang and Mills, 1954). Physically, gauge invariance refers to the conservation of
certain quantum property of the underlying interaction. Such quantum property of the N
particles with wave functions cannot be distinguished for the interaction, and consequently,
the energy contribution of these N particles associated with the interaction is invariant under
the general SU(N) phase (gauge) transformations:
! "
(Ψ,
& G &aµ τa ) = ΩΨ, Gaµ Ωτa Ω−1 + i (∂µ Ω)Ω−1 , ∀Ω = eiθ k (x)τk ∈ SU(N). (1.3.1)
g

Here the N wave functions: Ψ = (ψ1 , · · · , ψN )T represent the N particle fields, the N 2 − 1
gauge fields Gaµ represent the interacting potentials between these N particles, and {τa | a =
1, · · · , N 2 − 1} is the set of representation generators of SU(N). The gauge symmetry is then
stated as follows:

Principle of Gauge Invariance. The electromagnetic, the weak, and the strong
interactions obey gauge invariance. Namely, the Dirac equations involved in
the three interactions are gauge covariant and the actions of the interaction
fields are gauge invariant under the gauge transformations (1.3.1).

The field equations involving the gauge fields Gaµ are determined by the corresponding
Yang-Mills action, is uniquely determined by both the gauge invariance and the Lorentz
invariance, together with simplicity of laws of nature:
# $
' ( ! 1 ) mc *
LY M Ψ, {Gaµ } = − Gab gµα gνβ Gaµν Gbαβ + Ψ iγ µ Dµ − Ψ dx, (1.3.2)
M 4 h̄
8 Chapter 1 General Introduction

which is invariant under both the Lorentz and gauge transformations (1.2.3). Here Ψ =
1
Ψ† γ 0 , Ψ† = (Ψ∗ )T is the transpose conjugate of Ψ, Gab = tr(τa τb† ), λab c are the structure
2
constants of {τa | a = 1, · · · , N 2 − 1}, Dµ is the covariant derivative and Gaµν stands for the
curvature tensor associated with D µ :

Dµ = ∂µ + igGkµ τk ,
(1.3.3)
Gaµν = ∂µ Gaν − ∂ν Gaµ + gλbc
a b c
G µ Gν .

Principle of Interaction Dynamics (PID) (Ma and Wang, 2014e, 2015a)

With the Lagrangian action at our disposal, the physical law of the underlying system
is then represented as the Euler-Lagrangian equations of the action, using the principle of
Lagrangian dynamics (PLD). For example, all laws of classical mechanics can be derived
using PLD. However, in Section 1.1, we have demonstrated that the law of gravity obeys
the principle of interaction dynamics (PID), which takes the variation of the Einstein-Hilbert
action under energy-momentum conservation constraint (1.2.6).
We now state the general form of PID, and then illustrate the validity of PID for all
fundamental interactions.
PID was discovered in (Ma and Wang, 2014e, 2015a), and requires that for the four
fundamental interactions, the variation be taken under the energy-momentum conservation
constraints:

PID (Ma and Wang, 2014e, 2015a)

1) For the four fundamental interactions, the Lagrangian actions are given
by

!
√
L(g, A, ψ ) = L (g µν , A, ψ ) −gdx, (1.3.4)
M

where g = {g µν } is the Riemannian metric representing the gravitational

potential, A is a set of vector fields representing the gauge potentials, and
ψ are the wave functions of particles.
2) The actions (1.3.4) satisfy the invariance of general relativity, Lorentz
invariance, gauge invariance and the gauge representation invariance.
3) The states (g, A, ψ ) are the extremum points of (1.3.4) with the div A -free
constraint.

We now illustrate that PID is also valid for both the weak and strong interactions, re-
placing the classical PLD.
First, PID applied to the Yang-Mills action (1.3.2) takes the following form:
1.3 First Principles of Four Fundamental Interactions 9

+ ,
Gab ∂ ν Gbν µ − gλcd
b αβ c
g Gα µ Gdβ − gΨγµ τa Ψ = (∂µ + αb Gbµ )φa , (1.3.5)
(iγ µ Dµ − m)Ψ = 0, (1.3.6)

where the Dirac equations (1.3.6) are the gauge covariant, and are the dynamic equations
for fermions participating the interaction. The right hand side of (1.3.5) is due to PID, with
the operator
∇A = ∂µ + αb Gbµ
taken in such a way that αb Gbµ is PRI-covariant under the representation transformations
(1.3.8) below.
One important consequence of the PID SU(N) theory is the natural introduction of the
scalar fields φa , reminiscent of the Higgs field in the standard model in particle physics.
Second, for the Yang-Mills action for an SU(N) gauge theory, the resulting Yang-Mills
equations are the Euler-Lagrange equations of the Yang-Mills action:
+ ,
Gab ∂ ν Gbν µ − gλcd
b αβ c
g Gα µ Gdβ − gΨγµ τa Ψ = 0, (1.3.7)

supplemented with the Dirac equations (1.3.6).

Historically, the most important difficulty encountered by the gauge theory and the field
equations (1.3.7) is that the gauge vectorial field particles {Gaµ } for the weak interaction are
massless, in disagreement with experimental observations.
A great deal of efforts have been made toward to the modification of the gauge theory
by introducing proper mass generation mechanism. A historically important breakthrough
is the discovery of the mechanism of spontaneous symmetry breaking in subatomic physics.
Although the phenomenon was discovered in superconductivity by Ginzburg-Landau in
1951, the mechanism of spontaneous symmetry breaking in particle physics was first pro-
posed by Y. Nambu in 1960; see (Nambu, 1960; Nambu and Jona-Lasinio, 1961a,b). The
Higgs mechanism, introduced in (Higgs, 1964; Englert and Brour, 1964; Guralnik, Hagen
and Kibble, 1964), is an ad hoc method based on the Nambu-Jona-Lasinio spontaneous
symmetry breaking, leading to the mass generation of the vector bosons for the weak inter-
action.
In all these efforts associated with Higgs fields, the modification of the Yang-Mills ac-
tion is artificial as in the case for modifying the Einstein-Hilbert action for gravity.
As we indicated before, the Yang-Mills action is uniquely determined by the gauge sym-
metry, together with the simplicity principle of laws of Nature. All artificial modification of
the action will only lead to certain approximations of the underlying physical laws.
Third, the PID SU(N) gauge field equations (1.3.5) and (1.3.6) provide a first principle
based mechanism for the mass generation and spontaneous gauge symmetry-breaking: The
αb Gbµ on the right-hand side of (1.3.5) breaks the SU(N) gauge symmetry, and the mass
generation follows the Nambu-Jona Lasinio idea.
10 Chapter 1 General Introduction

Fourth, one of the most challenging problems for strong interaction is the quark confine
ment-no free quarks have been observed. One hopes to solve this mystery with the quantum
chromodynamics (QCD) based on the classical SU(3) gauge theory. Unfortunately, as we
shall see later that the classical SU(3) Yang-Mills equations produces only repulsive force,
and it is the dual fields in the PID gauge field equations (1.3.5) that give rise to the needed
attraction for the binding quarks together forming hadrons.
Hence experimental evidence of quark confinement, as well as many other properties
derived from the PID strong interaction model, clearly demonstrates the validity of PID for
strong interactions.

Fifth, from the mathematical point of view, the Einstein field equations are in general
non well-posed, as illustrated by a simple example in Section 4.2.2. In addition, for the
classical Yang-Mills equations, the gauge-fixing problem will also pose issues on the well-
posedness of the Yang-Mills field equations; see Section 4.3.5. The issue is caused by the
fact that there are more equations than the number of unknowns in the system. PID induced
model brings in additional unknowns to the equations, and resolves this problem.

Principle of representation invariance (PRI)

Recently we have observed that there is a freedom to choose the set of generators for
representing elements in SU(N). In other words, basic logic dictates that the SU(N) gauge
theory should be invariant under the following representation transformations of the gener-
ator bases:
τ&a = xba τb , (1.3.8)

where X = (xba ) are non-degenerate (N 2 − 1)-th order matrices. Then we can define natu-
rally SU(N) tensors under the transformations (1.3.8). It is clear then that θ a , Gaµ , and the
structure constants λabc are all SU(N)-tensors. In addition, G = 1 Tr(τ τ † ) is a symmetric
ab a b
2
positive definite 2nd-order covariant SU(N)-tensor, which can be regarded as a Riemannian
metric on SU(N). Consequently we have arrived at the following principle of representation
invariance, first discovered by the authors (Ma and Wang, 2014h):

PRI (Ma and Wang, 2014h) For the SU(N) gauge theory, under the repre-
sentation transformations (1.3.8),

1) the Yang-Mills action (1.3.2) of the gauge fields is invariant, and

2) the gauge field equations (1.3.5) and (1.3.6) are covariant.

It is clear that PRI is a basic logic requirement for an SU(N) gauge theory, and has
profound physical implications.
1.3 First Principles of Four Fundamental Interactions 11

First, as indicated in (Ma and Wang, 2013a, 2014h) and in Chapter 4, the field model
based on PID appears to be the only model which obeys PRI. In fact, based on PRI, for
the gauge interacting fields A µ and {Wµa }3a=1, corresponding to two different gauge groups
U(1) for electromagnetism and SU(2) for the weak interaction, the following combination

α Aµ + β Wµ3 (1.3.9)

is prohibited. The reason is that A µ is an U(1)-tensor with tensor, and Wµ3 is simply the
third component of an SU(2)-tensor. The above combination violates PRI. This point of
view clearly shows that the classical electroweak theory violates PRI, so does the standard
model. The difficulty comes from the artificial way of introducing the Higgs field. The PID
based approach for introducing Higgs fields by the authors appears to be the only model
obeying PRI.
Another important consequence of PRI is that for the term αb Gbµ in the right-hand side
of the PID gauge field equations (1.3.5), both {αb | b = 1, · · · , N 2 − 1} and {Gbµ | b =
1, · · · , N 2 − 1} are SU(N) tensors under the representation transformations (1.3.8).
The coefficients αb represent the portions distributed to the gauge potentials by the
charge, represented by the coupling constant g. Then it is clear that
In the field equations (1.3.5) and (1.3.6) of the SU(N) gauge theory for an
fundamental interaction,

(a) the coupling constant g represents the interaction charge, playing the
same role as the electric charge e in the U(1) abelian gauge theory for
quantum electrodynamics (QED);
(b) the potential
def
Gµ = αb Gbµ (1.3.10)

represents the total interacting potential, where the SU(N) covector αb

represents the portions of each interacting potential Gbµ contributed to
the total interacting potential; and
(c) the temporal component G0 and the spatial components G ⃗ = (G1 , G2 , G3 )
represent, respectively the interaction potential and interaction magnetic
potential. The force and the magnetic force generated by the interaction
are given by: g
F = −g∇G0 , Fm = ⃗v × curl G,⃗
c
where ∇ and curl are the spatial gradient and curl operators.

Geometric interaction mechanism

A simple yet the most challenging problem throughout the history of physics is the
mechanism or nature of a force.
12 Chapter 1 General Introduction

One great vision of Albert Einstein is his principle of equivalence, which, in the math-
ematical terms, says that the space-time is a 4-dimensional (4D) Riemannian manifold
{M , g µν } with the metric g µν representing the gravitational potential. In other words,
gravity is manifested as the curved effect of the space-time manifold {M , g µν }. In essence,
gravity is manifested by the gravitational fields {g µν , ∇µ φ }, determined by the gravitational
field equations (1.2.7) and (1.2.8) together with the matter distribution {Tµν }.
The gauge theory provides a field theory for describing the electromagnetic, the weak
and the strong interactions. The geometry of the SU(N) gauge theory is determined by
1) the complex bundle 1 M ⊗ p (C4 )N for the wave functions Ψ = (ψ1 , · · · , ψN )T , repre-
senting the N particles,

2) the gauge interacting fields {Gaµ | a = 1, · · · , N 2 − 1}, and their dual fields {φ a | a =
1, · · · , N 2 − 1}, and

3) the gauge field equations (1.3.5) coupled with the Dirac equations (1.3.6).
In other words, the geometry of the complex bundle M ⊗ p (C4 )N , dictated by the gauge
field equations (1.3.5) together with the matter equations (1.3.6), manifests the underlying
interaction.
Consequently, it is natural for us to postulate the Geometric Interaction Mechanism 4.1
for all four fundamental interactions:
Geometric Interaction Mechanism(Ma and Wang, 2014d)

1) (Einstein, 1915) The gravitational force is the curved effect of the time-
space; and
2) the electromagnetic, weak, strong interactions are the twisted effects of
the underlying complex vector bundles M ⊗ p Cn .

We note that Yukawa’s viewpoint, entirely different from Einstein’s, is that the other
three fundamental forces—the electromagnetism, the weak and the strong interactions–take
place through exchanging intermediate bosons such as photons for the electromagnetic in-
teraction, the W± and Z intermediate vector bosons for the weak interaction, and the gluons
for the strong interaction.
It is worth mentioning that the Yukawa Mechanism is oriented toward to computing the
transition probability for particle decays and scattering, and the above Geometric Interac-
tion Mechanism is oriented toward to establishing fundamental laws, such as interaction
potentials, of the four interactions.
1 Throughout this book, we use the notation ⊗ to denote ”gluing a vector space to each point of a manifold”
p
to form a vector bundle. For example, !
M ⊗ p Cn = {p} × Cn
p∈M
is a vector bundle with base manifold M and fiber complex vector space C n .
1.4 Symmetry and Symmetry-Breaking 13

1.4 Symmetry and Symmetry-Breaking

As we have discussed so far, symmetry plays a fundamental role in understanding Nature.
In mathematical terms, each symmetry, associated with particular physical laws, consists
of three main ingredients: 1) the underlying space, 2) the symmetry group, and 3) tensors,
describing the objects which possess the symmetry.
For example, the Lorentz symmetry is made up of 1) the 4D Minkowski space-time M 4
with the Minkowski metric, 2) the Lorentz group LG, and the Lorentz tensors. For example,
the electromagnetic potential A µ is a Lorentz tensor, and the Maxwell equations are Lorentz
invariant.
One important point to make is that different physical systems enjoy different symme-
try. For example, gravitational interaction enjoys the symmetry of general relativity, which,
amazingly, dictates the Lagrangian action for the law of gravity. Also, the other three fun-
damental interactions obey the gauge and the Lorentz symmetries.
In searching for laws of Nature, one inevitably encounters a system consisting of a num-
ber of subsystems, each of which enjoys its own symmetry principle with its own symmetry
group. To derive the basic law of the system, one approach is to seek for a large symmetry
group, which contains all the symmetry groups of the subsystems. Then one uses the large
symmetry group to derive the ultimate law for the system.
However, often times, the basic logic would dictate that the approach of seeking large
symmetry group is not allowed. For example, as demonstrated earlier, PRI specifically disal-
low the mixing the U(1) and SU(2) gauge interacting potentials in the classical electroweak
theory.
In fact, this demonstrates an inevitably needed departure from the Einstein vision of
unification of the four interactions using large gauge groups.
Our view is that the unification of the four fundamental interactions, as well as the mod-
eling of multi-level physical systems, is achieved through a symmetry-breaking mechanism,
together with PID and PRI. Namely, we postulated in (Ma and Wang, 2014a) and in Sec-
tion 2.1.7 the following Principle of Symmetry-Breaking 2.14:

1) The three sets of symmetries — the general relativistic invariance, the

Lorentz and gauge invariances, and the Galileo invariance — are mutu-
ally independent and dictate in part the physical laws in different levels
of Nature; and
2) for a system coupling different levels of physical laws, part of these sym-
metries must be broken.

Here we mention three examples.

First, for the unification of the four fundamental interaction, the PRI demonstrates that
the unification through seeking large symmetry is not feasible, and the gauge symmetry-
14 Chapter 1 General Introduction

breaking is inevitably needed for the unification. The PID-induced gauge symmetry-breaking,
by the authors (Ma and Wang, 2015a, 2014h,d), offers a symmetry-breaking mechanism
based only on the first principle; see also Chapter4 for details.

Second, for a multi-particle and multi-level system, its action is dictated by a set of
SU(N1 ), · · · , SU(Nm ) gauge symmetries, and the governing field equations will break some
of these gauge symmetries; see Chapter 6 .

Third, in astrophysical fluid dynamics, one difficulty we encounter is that the Newtonian
Second Law for fluid motion and the diffusion law for heat conduction are not compatible
with the principle of general relativity. Also, there are no basic principles and rules for com-
bining relativistic systems and the Galilean systems together to form a consistent system.
The distinction between relativistic and Galilean systems gives rise to an obstacle for estab-
lishing a consistent model of astrophysical dynamics. This difficulty can be circumvented
by using the above mentioned symmetry-breaking principle, where we have to chose the
coordinate system
xµ = (x0 , x), x0 = ct and x = (x1 , x2 , x3 ),

such that the metric is in the form:

! "
2
ds2 = − 1 + 2 ψ (x,t) c2 dt 2 + gi j (x,t)dxi dx j .
c

Here gi j (1 ! i, j ! 3) are the spatial metric, and ψ represents the gravitational potential.
The resulting system breaks the symmetry of general coordinate transformations, and we
call such symmetry-breaking as relativistic-symmetry breaking.

1.5 Unified Field Theory Based on PID and PRI

One of the greatest problems in physics is to unify all four fundamental interactions. Albert
Einstein was the first person who made serious attempts to this problem.
Most attempts so far have focused on unification through large symmetry, following
Einstein’s vision. However, as indicated above, one of the most profound implication of
PRI is that such a unification with a large symmetry group would violate PRI, which is
a basic logic requirement. In fact, the unification should be based on coupling different
interactions using the principle of symmetry-breaking (PSB) instead of seeking for a large
symmetry group.
The basic principles for the four fundamental interactions addressed in the previous
sections have demonstrated that the three first principles, PID, PRI and PSB, offer an entirely
different route for the unification, which is one of the main aims of this book:

1) the general relativity and the gauge symmetries dictate the Lagrangian;
1.5 Unified Field Theory Based on PID and PRI 15

2) the coupling of the four interactions is achieved through PID, PRI and
PSB in the unified field equations, which obey the PGR and PRI, but break
spontaneously the gauge symmetry; and
3) the unified field model can be easily decoupled to study individual inter-
action, when the other interactions are negligible.

Hereafter we address briefly the main ingredients of the unified field theory.
Lagrangian action
Following the simplicity principle of laws of Nature as stated in Principle 2.2, the three
basic symmetries—the Einstein general relativity, the Lorentz invariance and the gauge
invariance—uniquely determine the interaction fields and their Lagrangian actions for the
four interactions:

• The Lagrangian action for gravity is the Einstein-Hilbert functional given by (1.2.2);

• The field describing electromagnetic interaction is the U(1) gauge field {A µ }, repre-
senting the electromagnetic potential, and the Lagrangian action density is
1
LEM = − Aµν Aµν + ψ e (iγ µ Dµ − me )ψ e , (1.5.1)
4
in which the first term stands for the scalar curvature of the vector bundle M ⊗ p C4 .
The covariant derivative and the field strength are given by

Dµ ψ e = (∂µ + ieAµ )ψ e , Aµν = ∂µ Aν − ∂ν Aµ .

• For the weak interaction, the SU(2) gauge fields {Wµa |a = 1, 2, 3} are the interacting
fields, and the SU(2) Lagrangian action density LW for the weak interaction is the
standard Yang-Mills action density as given by (1.3.2).

• The SU(3) gauge action density LS for the strong interaction is also in the standard
Yang-Mills form given by (1.3.2), and the strong interaction fields are the SU(3)
gauge fields {Skµ | 1 ! k ! 8}, representing the 8 gluon fields.

It is clear that the action coupling the four fundamental interactions is the natural com-
bination of the Einstein-Hilbert functional, the standard U(1), SU(2), SU(3) gauge actions
for the electromagnetic, weak and strong interactions:
) * ! √
L {g µν }, Aµ , {Wµa }, {Skµ } = [LEH + LEM + LW + LS ] −gdx, (1.5.2)
M

which obeys all the symmetric principles, including principle of general relativity, the
Lorentz invariance, the U(1) × SU(2) × SU(3) gauge invariance and PRI.
16 Chapter 1 General Introduction

PID unified field equations

With PID, the PRI covariant unified field equations are then given by:

1 8π G + ,
Rµν − gµν R + 4 Tµν = ∇µ + α 0 Aµ + αb1Wµb + αk2 Skµ φνG , (1.5.3)
2 c + ,
µ
∂ (∂µ Aν − ∂ν Aµ ) − eJν = ∇ν + β 0 Aν + βb1Wνb + βk2Sνk φ e , (1.5.4)
+ ,
w
Gab ∂ µ Wµνb
− gw λcdb αβ c
g Wαν Wβd − gw Jν a (1.5.5)
# $
1 2
= ∇ν + γ Aν + γb Wν + γk Sν − mw xν φaw ,
0 1 b 2 k
4
+ ,
j j αβ c
Gksj ∂ µ Sµν − gs Λcd g Sαν Sβd − gs Qν k (1.5.6)
# $
1
= ∇ν + δ 0 Aν + δb1Wνb + δk2 Sνk − m2s xν φks ,
4
(iγ µ Dµ − m)Ψ = 0, (1.5.7)

where Ψ = (ψ e , ψ w , ψ s )T stands for the wave functions for all fermions, participating re-
spectively the electromagnetic, the weak and the strong interactions, and the current densi-
ties are defined by

Jν = ψ e γν ψ e , Jν a = ψ w γν σa ψ w , Qν k = ψ s γν τk ψ s . (1.5.8)

Equations (1.5.7) are the gauge covariant Dirac equations for all fermions participating
the four fundamental interactions. The left-hand sides of the field equations (1.5.3)-(1.5.6)
are the same as the classical Einstein equations and the standard U(1) × SU(2) × SU(3)
gauge field equations. The right-hand sides of are due to PID, and couple the four funda-
mental interactions.
Duality
In the field equations (1.5.3)-(1.5.6), there exists a natural duality between the inter-
action fields (g µν , Aµ ,Wµa , Skµ ) and their corresponding dual fields (φ µG , φ e , φaw , φks ). This
duality can be viewed as the duality between mediators and the duality between interacting
forces, summarized as follows:

(a) Duality of mediators. Each interaction mediator possesses a dual field

particle, called the dual mediator, and if the mediator has spin-k, then its
dual mediator has spin-(k − 1).
(b) Duality of interacting forces. Each interaction generates both attracting
and repelling forces. Moreover, for each pair of dual fields, the even-
spin field generates an attracting force, and the odd-spin field generates
a repelling force.
1.6 Theory of Strong Interactions 17

A few remarks are now in order.

First, from the (phenomenological) weakton theory of elementary particles to be intro-
duced later, that each mediator and its corresponding dual mediator are the same type of
particles with the same constituents, but with different spins.
Second, from the duality of the mediators induced from the first principle PID, we con-
jecture that in addition to the neutral Higgs found by LHC in 2012, there must be two
charged massive spin-0 Higgs particles.
Third, the duality of interacting forces demonstrates that all four forces can be both
attractive and repulsive. For example, as the distance of two massive objects increases, the
gravitational force displays attraction as the Newtonian gravity, more attraction than the
Newtonian gravity in the galaxy scale, and repulsion in a much larger scale.
Fourth, as we shall see later, the 8 gluon fields produces repulsive force, and it is their
dual fields that produce the needed attraction for quark confinement. This gives a strong
observational evidence for the existence of the dual gluons and for the valid and necessity
of PID.
Decoupling
The unified field model can be easily decoupled to study each individual interaction
when other interactions are negligible. In other words, PID is certainly applicable to each
individual interaction. For gravity, for example, PID offers to a new gravitational field
model, leading to a unified model for dark energy and dark matter as described earlier in
this introduction.

1.6 Theory of Strong Interactions

The strong interaction is responsible for the nuclear force that binds protons and neutrons
(nucleons) together to form the nucleus of an atom, and is the force that holds quarks to-
gether to form protons, neutrons, and other hadron particles. The current theory of strong
interaction is the quantum chromodynamics (QCD), based on an SU(3) non-Abelian gauge
theory.
Two most important observed basic properties of the strong interaction are the asymp-
totic freedom and the quark confinement. The theoretical understanding of these properties
are still lacking. There have been many attempts such as the lattice QCD, which is devel-
oped in part for the purpose of understanding the quark confinement.
Modern theory of QCD leaves a number of key problems open for a long time, including

1) (Quark confinement) Why is there no observed quark?

2) Why is there asymptotic freedom?

18 Chapter 1 General Introduction

3) What is the strong interaction potential?

4) Can one derive the Yukawa potential from first principles?

5) Why is the strong force short-ranged?

These are longstanding problems, and the new field theory for strong interaction based
on PID and PRI completely solves these open problems. We now present some basic in-
gredients of this new development, and we refer the interested readers to (Ma and Wang,
2014c) and Section 4.5 for details.

PID SU(3) gauge field equations

The strong interaction field equations decoupled from the unified field model are
+ , # $
j j 1
Gksj ∂ µ Sµν − gs Λcd gαβ Sαν
c
Sβd − gs Qν k = ∂ν + δk2 Sνk − m2s xν φks , (1.6.1)
4
+ ) * ,
iγ µ ∂µ + igs Skµ τk − mq ψ = 0. (1.6.2)

1. Gluons and dual scalar gluons. For the strong interaction, the field equations induce
a duality between the eight spin-1 massless gluons, described by the SU(3) gauge fields
{Skµ | k = 1, · · · , 8}, and the eight spin-0 dual gluons, described by the dual fields {φ ks | a =
1, · · · , 8}. As we shall see from the explianation of quark confinements gluons and their
dual gluons are confined in hadrons.

2. Attracting and repulsive behavior of strong force. As indicated before and from the
strong interaction potentials derived hereafter, strong interaction can display both attracting
and repulsive behavior. The repulsive behavior is due mainly to the spin-1 gluon fields, and
the attraction is caused by the spin-0 dual gluon fields.

Layered strong interaction potentials

Different from gravity and electromagnetic force, strong interaction is of short-ranged

with different strengths in different levels. For example, in the quark level, strong interaction
confines quarks inside hadrons, in the nucleon level, strong interaction bounds nucleons
inside atoms, and in the atom and molecule level, strong interaction almost diminishes.
This layered phenomena can be well-explained using the unified field theory based on PID
and PRI.
One key ingredient is total interaction as the consequence of PRI addressed above. For
the strong interaction, in view of (1.3.10), the total strong interaction potential is defined by
δk2 Sνk on the right hand side of (1.6.1):

def def
Sµ = δk2 Skµ = (S0 , S1 , S2 , S3 ), Φ = S0 . (1.6.3)
1.6 Theory of Strong Interactions 19

The temporal component S0 , denoted by Φ, gives rise to the strong force between two
elementary particles carrying strong charges:

F = −gs ∇Φ. (1.6.4)

From the field equations (1.6.1) and (1.6.2), we derive in Section 4.5.2 that for a particle
with N strong charges and radius ρ , its strong interaction potential are given by (4.5.39) and
recast here convenience:
# $
1 A
Φ = gs (ρ ) − (1 + kr)e −kr
,
r ρ
! "3 (1.6.5)
ρw
gs (ρ ) = N gs ,
ρ
where gs is the strong charge of w∗ -weakton, the ρw is the radius of the w∗ -weakton, A is a
dimensionless constant depending on the particle, and 1/k is the radius of strong attraction
of particles. Phenomenologically, we can take
⎧ −18
⎪
⎪ 10 cm for w∗ − weaktons,
⎪
⎪
1 ⎨ 10
−16
cm for quarks,
= (1.6.6)
k ⎪ 10
⎪ −13
cm for neutreons,
⎪
⎪
⎩ −7
10 cm for atom/molecule,
and the resulting layered formulas of strong interaction potentials are called the w ∗ -weakton
potential Φ0 , the quark potential Φq , the nucleon/hadron potential Φn and the atom/molecule
potential Φa .
Quark confinement
Quark confinement is a challenging problem in physics. Quark model was confirmed by
many experiments. However, no any single quark is found ever. This fact suggests that the
quarks were permanently bound inside a hadron, which is called the quark confinement. Up
to now, no other theories can successfully describe the quark confinement phenomena. The
direct reason is that all current theories for interactions fail to provide a successful strong
interaction potential to explain the various level strong interactions.
With the strong interaction potential (1.6.5), we can calculate the strong interaction
bound energy E for two same type of particles:
# $
1 A
E = g2s (ρ ) − (1 + kr)e−kr . (1.6.7)
r ρ
The quark confinement can be well explained from the viewpoint of the strong quark
bound energy Eq and the nucleon bound energy En . In fact, by (1.6.7) we can show that

Eq = 1020 En ∼ 1018 GeV. (1.6.8)

20 Chapter 1 General Introduction

This clearly shows that the quarks is confined in hadrons, and no free quarks can be found.

Asymptotic freedom and short-range nature of the strong interaction

Asymptotic freedom was discovered and described by (Gross and Wilczek, 1973; Politzer,
1973). Using the strong interaction potential (1.6.5), we can clearly demonstrate the asymp-
totic freedom property and the short-ranged nature of the strong interaction.

1.7 Theory of Weak Interactions

The new weak interaction theory based on PID and PRI was first discovered by (Ma and
Wang, 2013a). As addressed earlier, the weak interaction obeys the SU(2) gauge symmetry,
which dictates the standard SU(2) Yang-Mills action. By PID and PRI, the field equations
of the weak interaction are given by:
+ , + 1 ,
w
Gab ∂ µ Wµν
b b αβ c
− gwλcd g Wαν Wβd − gw ψ w γ ν σa ψ w = ∂µ + γb1Wµb − m2w xµ φaw , (1.7.1)
4
(iγ µ Dµ − ml )ψ w = 0. (1.7.2)

1. Higgs fields from first principles. The right-hand side of (1.7.1) is due to PID, leading
naturally to the introduction of three scalar dual fields. The left-hand side of (1.7.1) repre-
sents the intermediate vector bosons W ± and Z, and the dual fields represent two charged
Higgs H ± (to be discovered) and the neutral Higgs H 0 , with the later being discovered by
LHC in 2012.
It is worth mentioning that the right-hand side of (1.7.1), involving the Higgs fields, can
not be generated by directly adding certain terms in the Lagrangian action, as in the case for
the new gravitational field equations derived in (Ma and Wang, 2014e).

2. Duality. We establish a natural duality between weak gauge fields {Wµa }, representing
the W ± and Z intermediate vector bosons, and three bosonic scalar fields φaw , representing
both two charged and one neutral Higgs particles H ± , H 0 .

3. Spontaneous gauge symmetry-breaking. PID induces naturally spontaneous symme-

try breaking mechanism. By construction, it is clear that the Lagrangian action LW obeys
the SU(2) gauge symmetry, the PRI and the Lorentz invariance. Both the Lorentz invariance
and PRI are universal principles, and, consequently, the field equations (1.7.1) and (1.7.2)
are covariant under these symmetries. The gauge symmetry is spontaneously breaking in
the field equations (1.7.1), due to the presence of the terms in the right-hand side, derived
by PID. This term generates the mass for the vector bosons.

4. Weak charge and weak potential. By PRI, the SU(2) gauge coupling constant g w
plays the role of weak charge, responsible for the weak interaction.
1.8 New Theory of Black Holes 21

Also, PRI induces an important SU(2) constant vector {γb1 }. The components of this
vector represent the portions distributed to the gauge potentials Wµa by the weak charge gw .
Hence the (total) weak interaction potential is given by the following PRI representation
invariant
Wµ = γa1Wµa = (W0 ,W1 ,W2 ,W3 ), (1.7.3)
and the weak charge potential and weak force are as

Φw = W0 the time component of Wµ ,

(1.7.4)
Fw = −gw (ρ )∇Φw ,

where gw (ρ ) is the weak charge of a particle with radius ρ .

5. Layered formulas for the weak interaction potential. The weak interaction is also
layered, and we derive from the field equations the following
# $
1 B
Φw = gw (ρ )e−kr − (1 + 2kr)e−kr ,
r ρ
! "3 (1.7.5)
ρw
gw (ρ ) = N gw ,
ρ
where Φw is the weak force potential of a particle with radius ρ and carrying N weak charges
gw , taken as the unit of weak charge gs for each weakton (Ma and Wang, 2015b), ρw is the
weakton radius, B is a parameter depending on the particles, and 1/k = 10 −16 cm represents
the force-range of weak interactions.
6. The layered weak interaction potential formula (1.7.5) shows clearly that the weak
interaction is short-ranged. Also, it is clear that the weak interaction is repulsive, asymptot-
ically free, and attractive when the distance of two particles increases.

1.8 New Theory of Black Holes

The concept of black holes is essentially developed following the Karl Schwarzschild’s
derivation of the Schwarzschild solution for the Einstein gravitational field equations. In
the exterior of spherically symmetric ball of mass, the solution is given by
! " ! "
Rs 2 2 Rs −1 2
ds2 = − 1 − c dt + 1 − dr + r2 d θ 2 + r2 sin2 θ d ϕ 2 , (1.8.1)
r r
where
2MG
Rs = (1.8.2)
c2
is called the Schwarzschild radius, at which the metric displays a singularity. There have
been many confusions about black holes throughout the history of general relativity and
black holes.
22 Chapter 1 General Introduction

In this section, we present the black hole theorem proved in Section 7.3, and clarify
some of the confusions in the literature. This black hole theorem leads to important insights
to many problems in astrophysics and cosmology, which will be addressed in details in
Chapter 7.

Blackhole Theorem (Ma and Wang, 2014a) Assume the validity of the Einstein
theory of general relativity, then the following assertions hold true:

1) black holes are closed: matters can neither enter nor leave their interiors,
2) black holes are innate: they are neither born to explosion of cosmic ob-
jects, nor born to gravitational collapsing, and
3) black holes are filled and incompressible, and if the matter field is non-
homogeneously distributed in a black hole, then there must be sub-blackholes
in the interior of the black hole.

This theorem leads to drastically different views on the structure and formation of our
Universe, as well as the mechanism of supernovae explosion and the active galactic nucleus
(AGN) jets. Here we only make a few remarks, and refer interested readers to Section 7.3
for the detailed proof.

1. An intuitive observation. One important part of the theorem is that all black holes are
closed: matters can neither enter nor leave their interiors. Classical view was that nothing
can get out of blackholes, but matters can fall into blackholes. We show that nothing can
get inside the blackhole either.
To understand this result better, let’s consider the implication of the classical theory
that matters can fall inside a blackhole. Take as an example the supermassive black hole
at the center of our galaxy, the Milky Way. By the classical theory, this blackhole would
continuously gobble matters nearby, such as the cosmic microwave background (CMB). As
the Schwarzschild radius of the black hole r = Rs is proportional to the mass, then the radius
Rs would increase in cubic rate, as the mass M is proportional to the volume. Then it would
be easy to see that the black hole will consume the entire Milky Way, and eventually the
entire Universe. However, observational evidence demonstrates otherwise, and supports our
result in the blackhole theorem.

2. Singularity at the Schwarzschild radius is physical. One important ingredient is that

the singularity of the space-time metric at the Schwarzschild radius R s is essential, and
cannot be removed by any differentiable coordinate transformations. Classical transforma-
tions such as those by Eddington and Kruskal are non-differentiable, and are not valid
for removing the singularity at the Schwarzschild radius. In other words, the singularity
displayed in both the Schwarzschild metric and the Tolman-Oppenheimer-Volkoff (TOV)
1.9 The Universe 23

metric at r = Rs
! "−1
2 u 2 2 r2
ds = −e c dt + 1 − 2 dr2 + r2 d θ 2 + r2 sin2 θ d ϕ 2 , (1.8.3)
Rs
is a true physical singularity, and defines the black hole boundary.
3. Geometric realization of a black hole. As described in Section 4.1 in (Ma and Wang,
2014a) and in Section 7.3.1, the geometrical realization of a black hole, dictated by the
Schwarzschild and TOV metrics, clearly manifests that the real world in the black hole is a
hemisphere with radius Rs embedded in R4 , and at the singularity r = Rs , the tangent space
of the black hole is perpendicular to the coordinate space R3 .
This geometric realization clearly demonstrates that the disk in the realization space R 3
is equivalent to the real world in the black hole. If the outside observer observes that nothing
gets inside the black hole, then nothing will get inside the black hole in the reality as well.

1.9 The Universe

Geometry and origin of our Universe
Modern cosmology adopts the view that our Universe is formed through the Big-Bang
or Big-Bounce; see among others (Harrison, 2000; Kutner, 2003; Popławski, 2012; Roos,
2003). Based on our new insights on black holes, we have reached very different conclu-
sions on the structure and formation of our Universe.
It is clear that the large scale structure of our Universe is essentially dictated by the law
of gravity, which is based on Einstein’s two principles: the principle of general relativity
and the principle of equivalence, as addressed in Section 1.1. Also, strong cosmological ob-
servational evidence suggests that the large scale Universe obey the cosmological principle
that the Universe is homogeneous and isotropic.

Basic assumptions
(a) the Einstein theory of general relativity, and
(b) the cosmological principle.

With Assumption (a) above, we have our blackhole theorem, Theorem 7.15, at our dis-
posal. Then we can draw a number of important conclusions on the structure of our Uni-
verse.
1. Let E and M be the total energy and mass of the Universe:

E = kinetic + electromagnetic + thermal+ Ψ, M = E/c2 , (1.9.1)

where Ψ is the energy of all interaction fields. The total mass M dictates the Schwarzschild
radius Rs .
24 Chapter 1 General Introduction

If our Universe were born to the Big-Bang, assuming at the initial stage, all energy is
concentrated in a ball with radius R0 < Rs , by the theory of black holes, then the energy
contained in BR0 must generate a black hole in R3 with fixed radius Rs as defined by (1.8.2).
If we assume that at certain stage, the Universe were contained in ball of a radius R with
R0 < R < Rs , then we can prove that the Universe must contain a sub-black hole with radius
r given by
1
R
r= R.
Rs
Based on this property, the expansion of the Universe, with increasing R to R s , will give rise
to an infinite sequence of black holes with one embedded to another. Apparently, this sce-
nario is clearly against the observations of our Universe, and demonstrates that our Universe
cannot be originated from a Big-Bang.

2. By the cosmological principle, given the energy density ρ 0 > 0 of the Universe, based
on the Schwarzschild radius, the Universe will always be bounded in black hole, which is
an open ball of radius:
2
3c2
Rs = .
8π Gρ0
This immediately shows that there is no unbounded universe. Consequently, since a black
hole is unable to expand and shrink, we arrive immediately from the above analysis that our
Universe must be static, and not expanding.
Notice that the isotropy requirement in the cosmological principle excludes the globular
open universe scenario. Consequently, we have shown that our Universe must be a closed
3D sphere S3 .
In summary, we have proved two theorems, Theorems 7.27-7.28, on the geometry and
structure of our Universe, which have been discovered and proved in (Ma and Wang, 2014a,
Theorems 6.2 & 6.3):

Theorem on Structure of our Universe Assume the Einstein theory of gen-

eral relativity, and the principle of cosmological principle, then the following
assertions hold true:

1) our Universe is not originated from a Big-Bang, and is static;

2) the topological structure of our Universe is the 3D sphere S 3 such that to
each observer, the corresponding equator with the observer at the center
of the hemisphere can be viewed as the black hole horizon;
3) the total mass Mtotal in the Universe includes both the cosmic observable
mass M and the non-observable mass, regarded as dark matter, due to the
space curvature energy; and
1.9 The Universe 25

4) a negative pressure is present in our Universe to balance the gravitational

attracting force, and is due to the gravitational repelling force, also called
dark energy.

It is clear that this theorem changes drastically our view on the geometry and the origin
of the Universe. Inevitably, a number of important questions need to be addressed for this
scenario of our Universe. Hereafter we examine a few most important problems.
Redshift problem
The natural and important question that one has to answer is the consistency with astro-
nomical observations, including the cosmic edge, the flatness, the horizon, the redshift, and
the cosmic microwave background (CMB) problems. These problems can now be easily
understood based on the structure of the Universe and the blackhole theorem we derived.
Hereafter we focus only on the redshift and the CMB problems.
The most fundamental problem is the redshift problem. Observations clearly show that
light coming from a remote galaxy is redshifted, and the farther away the galaxy is, the
larger the redshift. In modern astronomy and cosmology, it is customary to characterize the
redshift by a dimensionless quantity z in the formula
λobserv
1+z = , (1.9.2)
λemit
where λobserv and λemit represent the observed and emitting wavelenths.
There are three sources of redshifts: the Doppler effect, the cosmological redshift, and
the gravitational redshift. If the Universe is not considered as a black hole, then the gravi-
tational redshift and the cosmological redshift are both too small to be significant. Hence,
modern astronomers have to think that the large port of the redshift is due to the Doppler
effect.
However, due to black hole properties of our Universe, the black hole and cosmological
redshifts cannot be ignored. Due to the horizon of the sphere, for an arbitrary point in the
spherical Universe, its opposite hemisphere relative to the point is regarded as a black hole.
Hence, g00 can be approximatively taken as the Schwarzschild solution for distant objects
as follows
! "
Rs
−g00 (r) = α (r) 1 − , α (0) = 2, α (Rs ) = 1, α ′ (r) < 0,
r
&
where &r = 2Rs − r for 0 ! r < Rs is the distance from the light source to the opposite radial
point, and r is the distance from the light source to the point. Then we derive the following
redshift formula, which is consistent with the observed redshifts:
√
1 2Rs − r
1+z = 2 ! " = 3 for 0 < r < Rs . (1.9.3)
Rs α (r)(Rs − r)
α (r) 1 −
r
&
26 Chapter 1 General Introduction

CMB problem

In 1965, two physicists A. Penzias and R. Wilson discovered the low-temperature cos-
mic microwave background (CMB) radiation, which fills the Universe, and it has been re-
garded as the smoking gun for the Big-Bang theory. However, based on the unique scenario
of our Universe we derived, it is the most natural thing that there exists a CMB, because
the Universe has always been there as a black-body, and CMB is a result of blackbody
equilibrium radiation.

Dark matter and dark energy

Conclusion 4) in the above results for our Universe shows that the observable cosmic
mass M, and the total mass Mtotal which includes both M and the non-observable mass
caused by space curvature energy, enjoy the following relation:

Mtotal = 3π M/2 (1.9.4)

The difference Mtotal − M can be regarded as the dark matter. Astronomical observations
have shown that the measurable mass M is about one fifth of total mass Mtotal .
Also, the static Universe has to possess a negative pressure to balance the gravitational
attracting force. The negative pressure is actually the effect of the gravitational repelling
force, attributed to dark energy.
Equivalently, the above interpretation of dark matter and dark energy is consistent with
the theory based on the new PID gravitational field equations discovered by (Ma and Wang,
2014e) and addressed in detail in Chapter 4. It is clear now that gravity can display both at-
tractive and repulsive effect, caused by the duality between the attracting gravitational field
{g µν } and the repulsive dual vector field {Φ µ }, together with their nonlinear interactions
governed by the field equations. Consequently, dark energy and dark matter phenomena are
simply a property of gravity. The detailed account of this relation is addressed in Section
7.6, based in part on (Hernandez, Ma and Wang, 2015).

PID-cosmological model

One of the main motivations for the introduction of the Big-Bang theory and for the
expanding universe is that the Friedamann solution of the Einstein gravitational field equa-
tions demonstrated that the Einstein theory must produce a variable size universe; see Con-
clusions of Friedmann Cosmology 7.23. Also, it is classical that bringing the cosmological
constant Λ into the Einstein field equations will lead to an unstable static Universe.

We have demonstrated that the right cosmological model should be derived from the new
gravitational field equations (1.2.7), taking into consideration the presence of dark matter
and dark energy. In this case, based on the cosmological principle, the the metric of a
1.10 Supernovae Explosion and AGN Jets 27

homogeneous spherical universe is of the form

# $
dr2
ds2 = −c2 dt 2 + R2 + r 2
(d θ 2
+ sin2
θ d ϕ 2
) , (1.9.5)
1 − r2

where R = R(t) is the cosmic radius. We deduce then from (1.2.7) the following PID-
cosmological model, with ϕ = φ ′′ :
! "
4π G 3p ϕ
R′′ = − ρ+ 2 + R,
3 c 8π G
1 (1.9.6)
(R′ )2 = (8π Gρ + ϕ )R2 − c2 ,
3
3R′ 24π G R′
ϕ′ + ϕ =− 2 p,
R c R
supplemented with the equation of state:

p = f (ρ , ϕ ). (1.9.7)

Note that only two equations in (1.9.6) are independent.

Also, the model describing the static Universe is the equation of state (1.9.7) together
with the stationary equations of (1.9.6), which are equivalent to the form
! "
3p
ϕ = −8π G ρ + 2 ,
c
(1.9.8)
4
c
p=− .
8π GR2

1.10 Supernovae Explosion and AGN Jets

Supernovae explosion and the active galactic nucleus (AGN) jets are among the most im-
portant astronomical phenomena, which are lack of reasonable explanation.
Relativistic, magnetic and thermal effects are main ingredients in astrophysical fluid
dynamics, and are responsible for many astronomical phenomena. The thermal effect is
described by the Rayleigh number Re:

mGr0 r1 β T0 − T1
Re = , (1.10.1)
κν r1 − r0
where T0 and T1 are the temperatures at the bottom and top of an annular shell region r0 <
r < r1 ; see e.g. (7.1.74) for other notations used here.
Based on our theory of black holes, including in particular the incompressibility and
closedness of black holes, the relativistic effect is described by the δ -factor:
2mG
δ= . (1.10.2)
c2 r0
28 Chapter 1 General Introduction

where m is the mass of the core 0 < r < r0 .

Consider e.g. an active galactic nucleus (AGN), which occupies a spherical annular
shell region Rs < r < R1 , where Rs is the Schwarzschild radius of the black hole core of the
galaxy. Then r0 = Rs and the δ -factor is δ = 1. The relativistic effect is then reflected in the
radial force ! " ! "
ν ∂ 1 dα Rs −1
Fr = Pr , α = 1− ,
2α ∂ r α dr r
which gives rise to a huge explosive force near r = Rs :
ν R2s
Pr . (1.10.3)
1 − Rs/r r4
The relativistic effect is also reflected in the electromagnetic energy:
! "
ν0 ∂ 1 d α
Hr ,
2α ∂ r α dr
which consists of a huge explosive electromagnetic energy near r = R s :
ν0 R2s
Hr . (1.10.4)
1 − Rs/r r4
The basic mechanism of the formation of AGN jets is that the radial temperature gradient
causes vertical Bénard convection cells. Each Bénard convection cell has a vertical exit,
where the circulating gas is pushed by the radial force, and then erupts leading to a jet.
Each Bénard convection cell is also an entrance, where the external gas is attracted into the
nucleus, is cycloaccelerated by the radial force as well, goes down to the interior boundary
r = Rs , and then is pushed toward to the exit. Thus the circulation cells form jets in their
exits and accretions in their entrances.
This mechanism can also be applied to supernovae explosion. When a very massive
red giant completely consumes its central supply of nuclear fuels, its core collapses. Its
radius r0 begins to decrease, and consequently the δ -factor increases. The huge mass m and
the rapidly reduced radius r0 make the δ -factor approaching one. The thermal convection
gives rise to an outward radial circulation momentum flux Pr . Then the radial force as in
(1.10.3) will lead to the supernova explosion. Also, Pr = 0 at r = r0 , where r0 is the radius
of blackhole core of supernovae. Consequently, the supernova’s huge explosion preserves a
smaller ball, yielding a neutron star.

1.11 Multi-Particle Systems and Unification

The field theory for multi-particle system was discovered by (Ma and Wang, 2014d). Clas-
sical quantum dynamic equations describe single particle systems. The existing model for a
multi-particle system is non-relativistic and is based on prescribing the interaction between
particles using such potentials as the Coulomb potential.
1.11 Multi-Particle Systems and Unification 29

As far as we know, there is still no good model for a multi-particle system, which takes
also into consideration the dynamic interactions between particles. The main obstacle for
establishing a field theory for an interacting multi-particle system is the lack of basic prin-
ciples to describe the dynamic interactions of the particles.
Basic postulates for interacting multi-particle systems
To seek the needed principles, we proceed with three observations.
1. One natural outcome of the field theory of four interactions developed recently by
the authors and addressed in the previous sections is that the coupling constants for the
U(1) × SU(2) × SU(3) gauge theory play the role of the three charges e, g w and gs for
electromagnetism, the weak and the strong interaction. These charges generate interacting
fields among the interacting particles.
Now we consider an N-particle system with each particle carrying an interaction charge
g. Let this be a fermionic system, and the Dirac spinors be given by

Ψ = (ψ1 , · · · , ψN )T ,

which obeys the Dirac equations:

iγ µ Dµ Ψ + MΨ = 0, (1.11.1)

where M is the mass matrix, and

Dµ Ψ = ∂µ Ψ + igGΨ, (1.11.2)

where G = (Giµj ) is an Hermitian matrix, representing the interacting potentials between the
N-particles generated by the interaction charge g.
Now let
{τ0 , τ1 , · · · , τK | K = N 2 − 1}
be a basis of the set of all Hermitian matrices, where τ0 = I is the identity, and τa (1 ! a !
ij
N 2 − 1) are the traceless Hermitian matrices. Then the Hermitian matrix G = (G µ ) and the
differential operator D µ in (1.11.1) can be expressed as
G = G0µ I + Gaµ τa ,
Dµ = ∂µ + igG0µ + igGaµ τa .

Consequently the Dirac equations (1.11.1) are rewritten as

iγ µ [∂µ + igG0µ + igGaµ τa ]Ψ + MΨ = 0. (1.11.3)

2. The energy contributions of the N particles are indistinguishable, which implies the
SU(N) gauge invariance. Hence (1.11.3) are exactly the Dirac equations in the form of
SU(N) gauge fields {Gaµ | 1 ! a ! N 2 − 1} with a given external interaction field G0µ .
30 Chapter 1 General Introduction

3. The SU(N) gauge theory of N particles must obey PRI. Consequently there exists a
constant SU(N) tensor
αaN = (α1N , · · · , αNN ),

such that the contraction field using PRI

Gµ = αaN Ga (1.11.4)

is independent of the SU(N) representation τa , and is the interaction field which can be
experimentally observed.

With these three observations, it is natural for us to introduce the following postulate,
which is also presented as Postulates 6.25-6.27 in Chapter 6:

Postulates for interacting multi-particle systems

1) the Lagrangian action for an N-particle system satisfy the SU(N) gauge
invariance;
2) gGaµ represent the interaction potentials between the particles; and
3) for an N-particle system, only the interaction field G µ in (1.11.4) can be
measured, and is the interaction field under which this system interacts
with other external systems.

Field Equations for Multi-Particle System

Multi-particle systems are layered, and with the above postulates and basic symmetry
principles, we are able to determine in a unique fashion field equations for different multi-
particle systems.
For example, given an N-particle system consisting of N fermions with given charge g,
the SU(N) gauge symmetry dictates uniquely the Lagrangian density, given in two parts:
1) the sector of SU(N) gauge fields LG , and 2) the Dirac sector of particle fields LD ,
as described earlier in the SU(N) gauge theory. The combined action is 1) SU(N) gauge
invariant, 2) representation invariant (PRI), and 3) Lorentz invariant. The field equations are
then derived by using PID:

+ , # $
ν b b αβ c d k2 0
Gab ∂ Gν µ − gλcd g Gα µ Gβ − gΨγµ τa Ψ = ∂µ − xµ + gGµ + gGµ φa , (1.11.5)
4
4 5
iγ µ ∂µ + igG0µ + igGaµ τa Ψ − MΨ = 0, (1.11.6)

for 1 ! a ! N 2 − 1, where G0µ is the interaction field of external systems, and G µ = αa Gaµ .
It is important to note that coupling to the external fields is achieved by the terms involving
G0µ .
1.12 Weakton Model of Elementary Particles 31

Unification and geometrization of matter fields

Also, we have established a unified field model coupling matter fields, which matches
the vision of Einstein and Nambu, as stated in Nambu’s Nobel lecture (Nambu, 2008):
Einstein used to express dissatisfaction with his famous equation of gravity
Gµν = 8π Tµν

His point was that, from an aesthetic point of view, the left hand side of the
equation which describes the gravitational field is based on a beautiful geo-
metrical principle, whereas the right hand side, which describes everything
else,... looks arbitrary and ugly.
... [today] Since gauge fields are based on a beautiful geometrical principle,
one may shift them to the left hand side of Einstein s equation. What is left
on the right are the matter fields which act as the source for the gauge fields ...
Can one geometrize the matter fields and shift everything to the left?

Basically, one needs to geometrize the energy-momentum tensor Tµν appearing in the
Einstein field equations. For example, for multi-particle system under gravity and elec-
tromagnetism, using the basic postulates as outlined above, a unified field model can be
naturally derived so that the energy-momentum tensor Tµν is derived from first principles
and is geometrized.
The above vision of Einstein and Nambu is achieved by using the postulate for interact-
ing multi-particle systems, the PID, the PRI, as well as the principle of symmetry-breaking
(PSB):
1) fundamental symmetries of Nature dictate the actions for both the subsys-
tems as well as the global system,
2) the subsystems are coupled through PID, PRI and PSB.

We believe that this is the essence of physical modeling and a unique route for unifica-
tion.

1.12 Weakton Model of Elementary Particles

The weakton model of elementary particles was first introduced by (Ma and Wang, 2015b).
Hereafter we address the basic ingredients of this theory.
Motivation and requirements of weaktons
The matter in the Universe is made up of a number of fundamental constituents, called
elementary particles. Based on the current knowledge of particle physics, all forms of mat-
ter are made up of 6 leptons and 6 quarks, and their antiparticles, which are treated as
elementary particles.
32 Chapter 1 General Introduction

Great achievements and insights have been made for last 100 years on the understand-
ing of the structure of subatomic particles and on their interactions; see among many others
(Halzen and Martin, 1984; Griffiths, 2008; Kane, 1987; Quigg, 2013). However, there are
still many longstanding open questions and challenges. Here are a few fundamental ques-
tions which are certainly related to the deepest secrets of our Universe. One such problem
is that why leptons do not participate in strong interactions.
The starting point of the study is the puzzling decay and reaction behavior of subatomic
particles. For example, the electron radiations and the electron-positron annihilation into
photons or quark-antiquark pair clearly shows that there must be interior structure of elec-
trons, and the constituents of an electron contribute to the making of photon or the quark in
the hadrons formed in the process. In fact, all sub-atomic decays and reactions show clearly
the following conclusion:

There must be interior structure of charged leptons, quarks and mediators.

This conclusion motivates us to propose a model for sub-lepton, sub-quark, and sub-
mediators. It is clear that any such model should obey four basic requirements:

1) Mass generation mechanism.

2) Consistency of quantum numbers for both elementary and composite particles.

3) Exclusion of nonrealistic compositions of the elementary particles.

4) Weakton confinement.

The model should lead to consistency of masses for both elementary particles, which
we call weaktons, and composite particles (the quarks, leptons and mediators). Since the
mediators, the photon γ and the eight gluons are all massless, a natural requirement is that

the proposed elementary particles—weaktons—are massless. Namely, these

proposed elementary particles must have zero rest mass.

Weaktons and their quantum numbers

Careful examinations of the above requirements and subatomic decays/reactions lead us
to propose six elementary particles, which we call weaktons, and their anti-particles:

w∗ , w1 , w2 , νe , νµ , ντ ,
(1.12.1)
w , w1 , w2 , ν e , ν µ , ν τ ,
∗

where νe , νµ , ντ are the three generation neutrinos, and w∗ , w1 , w2 are three new particles,
which we call w-weaktons. These weaktons in (1.12.1) are endowed with the quantum
1.12 Weakton Model of Elementary Particles 33

numbers: electric charge Qe , weak charge Qw , strong charge Qs , weak color charge Qc ,
baron number B, lepton numbers Le , Lµ , Lτ , spin J, and mass m. The quantum numbers of
weaktons are listed in Table 1.1.

Table 1.1 Weakton quantum numbers

Weakton Qe Qw Qs Qc Le Lµ Lτ B J m
w∗ 2/3 1 1 0 0 0 0 1/3 ±1/2 0
w1 −1/3 1 0 1 0 0 0 0 ±1/2 0
w2 −2/3 1 0 −1 0 0 0 0 ±1/2 0
νe 0 1 0 0 1 0 0 0 −1/2 0
νµ 0 1 0 0 0 1 0 0 −1/2 0
ντ 0 1 0 0 0 0 1 0 −1/2 0

Weakton constituents and duality of mediators

The weakton constituents of charged leptons and quarks are given by

e = νe w1 w2 , µ = νµ w1 w2 , τ = ντ w1 w2 ,
∗
u = w w1 w1 , c = w∗ w2 w2 , t = w∗ w2 w2 , (1.12.2)
∗ ∗ ∗
d = w w1 w2 , s = w w1 w2 , b = w w1 w2 ,

where c,t and d, s, b are distinguished by the spin arrangements; see (5.3.20).
The weakton constituents of the mediators and their dual mediators are given by

γ = cos θw w1 w1 − sin θw w2 w2 (", #), γ0 = cos θw w1 w1 − sin θw w2 w2 (↑↓, ↓↑),

Z = sin θw w1 w1 + cos θw w2 w2 (", #), H 0 = sin θw w1 w1 + cos θw w2 w2 (↑↓, ↓↑),
W − = w1 w2 (", #), H − = w1 w2 (↑↓, ↓↑),
W + = w1 w2 (", #), H + = w1 w2 (↑↓, ↓↑),
gk = w∗ w∗ (", #), gk0 = w∗ w∗ (↑↓, ↓↑).
(1.12.3)
Remarkably, both the spin-1 mediators and the spin-0 dual mediators in the unified field
theory have the same weakton constituents, but with different spins. The spin arrangements
clearly demonstrate that there must be spin-0 particles with the same weakton constituents
as the mediators. Consequently, there must be dual mediators with spin-0. This observation
clearly supports the unified field model presented earlier. Conversely, the existence of the
dual mediators makes the weakton constituents perfectly fit.
Also, a careful examination of weakton constituents predicts the existence of an addi-
tional mediator, which we call the ν -mediator:
φν0 = ∑ αl νl ν l (↓↑), ∑ αl2 = 1, (1.12.4)
l l
34 Chapter 1 General Introduction

taking into consideration of neutrino oscillations. When examining decays and reactions of
sub-atomic particles, it is apparent for us to predict the existence of this mediator.
With the above weakton constituents of charged leptons, quarks and mediators, we can
then verify the four basic requirements as mentioned earlier, based in part on the weak
interaction potentials.
1. Mass generation. One important conclusion of the aforementioned weakton model is
that all particles—both matter particles and mediators—are made up of massless weaktons.
A fundamental question is how the mass of a massive composite particle is generated. In
fact, based on the Einstein formulas:
2 2
d⃗ v2 ⃗ v2 E
P = 1 − 2 F, m = 1− 2 2, (1.12.5)
dt c c c
we observe that a particle with an intrinsic energy E has zero mass m = 0 if it moves in the
speed of light v = c, and possess nonzero mass if it moves with a velocity v < c. Hence by
this mass generation mechanism, for a composite particle, the constituent massless weak-
tons can decelerate by the weak force, yielding a massive particle.
In principle, when calculating the mass of the composite particle, one should also con-
sider the bounding and repelling energies of the weaktons, each of which can be very large.
Fortunately, the constituent weaktons are moving in the “asymptotically-free” shell region
of weak interactions as indicated by the weak interaction potential/force formulas, so that
the bounding and repelling contributions to the mass are mostly canceled out. Namely,
the mass of a composite particle is due mainly to the dynamic behavior of the constituent
weaktons.
2. Consistency and removal of unrealistic compositions. The consistency can be easily
checked. Also a few simple quantum rules can be devised so that unrealistic combinations
of weaktons are easily excluded.
3. The weakton confinement is simply the direct consequence of the weak interaction
potential/force formulas.
Mechanism of decays
Remarkably, the weakton model offers a perfect explanation for all sub-atomic decays.
In particular, all decays are achieved by 1) exchanging weaktons and consequently exchang-
ing newly formed quarks, producing new composite particles, and 2) separating the new
composite particles by weak and/or strong forces.
One aspect of this decay mechanism is that we know now the precise constituents of
particles involved in all decays/reactions both before and after the reaction. It is therefore
believed that the new decay mechanism provides clear new insights for both experimental
and theoretical studies.
Chapter 2
Fundamental Principles of Physics

The purposes of this chapter are 1) to provide an intuitive introduction to fundamental prin-
ciples of Nature, and 2) to explain how the laws of Nature are derived based on these princi-
ples. The main focus is on the symbiotic interplay between advanced mathematics and the
laws of Nature. For this purpose, we start with a brief overview on the perspective and the
physical significance of a few fundamental principles.
Section 2.1 provides a basic intuitive introduction to fundamental principles, symme-
tries, the geometric interaction mechanism and the symmetry-breaking principle. We start
with two guiding principles of theoretical physics, which can be synthesized as: the laws
of Nature are represented by mathematical equations, are dictated by a few fundamental
principles, and always take the simplest and aesthetic forms.
The geometric interaction mechanism was originally motivated by Albert Einstein’s vi-
sion revealed in his principle of equivalence, and was first postulated in (Ma and Wang,
2014d).
Symmetry plays a fundamental role in physics. Many, if not all, physical systems obey
certain symmetry. For this reason, many fundamental principles in physics address the
underlying symmetries of physical systems.
However, a crucial component of the unification of four fundamental interactions as
well as the modeling of multi-level physical systems is the symmetry-breaking mechanism.
Consequently, we postulated in (Ma and Wang, 2014a) and in Section 2.1.7 a principle
of symmetry-breaking 2.14, which states that for a system coupling different levels of
physical laws, part of these symmetries of the subsystems must be broken.
Section 2.2 addresses essentials of the classical Lorentz invariance, and its applications
to the derivation of Schrödinger equation, the Klein-Gordon equation, the Weyl equations
and the Dirac equations in relativistic quantum mechanics.
Section 2.3 presents a brief introduction to the Einstein theory of general relativity,
focusing on Einstein’s two basic principles, the principle of equivalence and the principle
of general relativity, and on the derivation of the Einstein gravitational field equations.
A brief introduction to both the U(1) abelian and the SU(N) non-abelian gauge theories
are presented in Section 2.4. In particular, the U(1) abelian gauge theory is defined on the
36 Chapter 2 Fundamental Principles of Physics

complex spinor bundle M 4 ⊗ p C4 . The SU(N) non-abelian gauge theory, also called the
Yang-Mills theory, is the connection on the complex spinor bundle M 4 ⊗ p (C4 )N . Note that
the presentation here leads clearly to PRI, to be introduced in Chapter 4.

Section 2.5 introduces the principle of Lagrangian dynamics (PLD) and its many appli-
cations such as in classical mechanics and electrodynamics. The Noether Theorem, Theo-
rem 2.38, is also proved.

Section 2.6 is on Hamiltonian dynamics and its connections to the Lagrangian dynamics.

2.1 Essence of Physics

2.1.1 General guiding principles
Physics studies fundamental interactions, motion and formation of matter in our Universe.
The heart of fundamental physics is to seek experimentally verifiable, fundamental laws and
principles of Nature. Namely, physical concepts and theories are transformed into mathe-
matical models, and the predictions derived from these models can be verified experimen-
tally and conform to the reality.
Modeling is a crucial step to understand a physical phenomena. A good model should
be derived based on a few fundamental principles, and often presented in the form of dif-
ferential equations. The first principles are often connected with symmetries, which also
dictate specific forms of mathematical models, represented by differential equations.
In a nutshell, we have the following first principle of physics.

First Principle of Physics 2.1 Physical systems obey laws and principles of Nature:

1) These laws and principles can be expressed using mathematical models:

Model = Mathematical representation of physical laws and principles;

2) These laws and principles are universal, and the universality is reflected by symme-
tries in physics; and

3) Physical symmetries dictate the precise forms of the mathematical models, which are
often presented in the forms of ordinary or partial differential equations.

In addition, the following principle provides the essence of fundamental physics, and is
supported by known experimental facts.

Essence of Physics 2.2 The essence of physics can be drawn in the following two
important ingredients:

1) Theoretical physics is built upon a few fundamental principles of Nature; and

2.1 Essence of Physics 37

2) The laws of Nature always take the simplest and aesthetic forms.

First Principle 2.1 and Essence of Physics 2.2 serve as guiding principles to understand
and explore the physical world of our Universe and the laws of Nature.

2.1.2 Phenomenological methods

Theories of physics fall into two categories: first principle theories and phenomenological
theories. The first principle theories are derived based on a few fundamental laws and
principles of Nature, and phenomenological theories are conjectured and synthesized from
observational data.
For example, the Newton’s gravitational law is given by
m1 m2 G
F =− , (2.1.1)
r2
which is essentially deduced by using the phenomenological technique based on a large
number of astronomical data. First, one readily conjectures that the gravitational force is
proportional to the masses m and M of the two bodies, and is a function of distance r:

F = −mMΦ(r),

where Φ(r) is an undetermined function.

Then one considers two planets with masses m1 and m2 , rotating around the Sun with
velocities v1 and v2 . By the balance between the gravitational and centrifugal forces, we
have
v2i
= MΦ(ri ) for i = 1, 2,
ri
where M is the mass of the Sun. Consequently,

Φ(r1 ) r2 v21
= . (2.1.2)
Φ(r2 ) r1 v22

By the Kepler’s third law, the periods T1 and T2 of the two planets satisfy

T12 r3
2
= 13 , Ti vi = 2π ri for i = 1, 2,
T2 r2

which implies that

v21 r2
2
= . (2.1.3)
v2 r1
Then, it follows from (2.1.2) and (2.1.3) that
G
Φ(r) = ,
r2
38 Chapter 2 Fundamental Principles of Physics

where G is a constant, called the gravitational constant. Thus, by the Kepler’s third law, one
can postulate naturally the Newton’s gravitational law (2.1.1). In other words, the Newton’s
gravitational law can be regarded as a phenomenological theory.
Albert Einstein was the first who tried to deduce basic physical laws from the first prin-
ciples. For example, the Einstein theory of general relativity and the Einstein gravitational
field equations:
1 8π G
Rµν − gµν R = − 4 Tµν , (2.1.4)
2 c
are derived based on the following three basic principles:

1) the principle of general relativity,

2) the principle of equivalence, and

3) the principle of Lagrangian dynamics.

The Schwarzschild solution of the Einstein gravitational field equations (2.1.4) offers a nat-
ural link between the field equations and the Newtonian gravitational law (2.1.1).

2.1.3 Fundamental principles in physics

Fundamental first principles refer to the laws of Nature that cannot be derived from other
more basic laws.
Based on Essence of Physics 2.2, if we would like to better understand theoretical
physics, it is crucial to know and find out all fundamental laws. In this subsection, we
shall list the known and important principles in various physical fields, most of which will
be introduced in later chapters.
We start with the introduction of Principle of Lagrangian Dynamics (PLD), which is of
special importance.
In classical mechanics, we know the least action principle. For a mechanical system
with the position and velocity variables

x = (x1 , · · · , xn ), ẋ = (ẋ1 , · · · , ẋn ), (2.1.5)

let the kinetic energy T and the potential energy V be functions of the position and velocity
variables in (2.1.5). Then the states of the system are the extremum points of the functional
! t1
L(x) = L (x, ẋ)dt, (2.1.6)
t0

where the integrand L

L (x, ẋ) = T − V (2.1.7)
2.1 Essence of Physics 39

is called the Lagrange density, and the functional (2.1.6) is called the Lagrange action. The
extremum points x satisfy the variational equation, called the Euler-Lagrange equations of
(2.1.6):
δ L(x) = 0, (2.1.8)
where δ L is the variational derivative operator of L. The equation (2.1.8) can be equivalently
expressed as
d ∂L ∂L
− = 0.
dt ∂ ẋ ∂x
The most important point is that the least action principle can be generalized to all
physical systems describing motions, as well as to fundamental interactions of Nature. The
generalization in a motion system is called the PLD, and in an interaction system is called
the Principle of Interaction Dynamics (PID). PID will be introduced in detail In Chapter 4
of this book, and PLD is stated as follows.

Principle of Lagrangian Dynamics 2.3 For a physical motion system, there are
functions
u = (u1 , · · · , un ), (2.1.9)
which describe the states of this system, and there exists a functional of u, given by
!
L(u) = L (u, Du, · · · , Dm u)dx, (2.1.10)
Ω

where Ω is the domain of u, and Dk u is the k-th derivative of u for any 0 ! k ! m. Then the
state functions of this system are the extremum points of (2.1.10). Namely the state functions
satisfy the variational equation of (2.1.10):

δ L(u) = 0. (2.1.11)

The functional L is called the Lagrange action, and L is called the Lagrange density.

By PLD, to derive dynamical equations for a physical system, it suffices to find the
corresponding Lagrange action. In the next subsection, we demonstrate that symmetries in
physics dictate the precise forms of the Lagrange actions.
A list of known important physical principles and laws of Nature in various subfields of
physics is given as follows.

1) Universal principles in all fields:

• Principle of Lagrangian Dynamics (PLD),

• Principle of Hamiltonian Dynamics (PHD),
• Principle of Interaction Dynamics (PID),
• Lorentz Invariance,
40 Chapter 2 Fundamental Principles of Physics

• Principle of General Relativity,

• Principle of Gauge Invariance,

• Principle of Representation Invariance (PRI),

• Principle of Symmetry-Breaking

2) Classical mechanics:

• Newton’s Second Law,

• Principle of Least Action,

• Principle of Galilean Invariance,

• Fick and Fourier Diffusion Laws.

3) Quantum physics:

• Basic postulates of quantum physics,

• Pauli Exclusion Principle.

4) Statistical physics:

• Basic Laws of Thermodynamics,

• Le Châterlier Principle.

5) Nonlinear sciences:

• Principle of Phase Transition Dynamics.

A couple of remarks are now in order.

Remark 2.4 Because various conservation laws can be deduced by PHD, PLD and
symmetries based on the Noether theorem, these laws are not listed in the fundamental
principles.

Remark 2.5 Mathematical models form the skeleton of theoretical physics, and most,
if not all, mathematical models (differential equations) in theoretic physics can be derived
based on the principles listed above. One of the ultimate goals of this book is to derive
physical models based on a first principles.
2.1 Essence of Physics 41

2.1.4 Symmetry
Symmetry plays an important role in physics. We start with intuitively examining how
symmetry works, keeping in mind the relation between equations and laws of physics in
First Principle 2.1, which can be simply recast as

Laws of Physics = Differential Equations. (2.1.12)

The laws of Nature on the left hand side of (2.1.12) are often beyond words, and are
best expressed by differential equations. It is this characteristic, together with the Noether
theorem, that illustrates the importance of symmetry in physics, as illustrated below:
Invariance Covariance Symmetry
⇓ ⇓ ⇓
Form of Equation Space Structure Conservation Law

Namely, symmetry dictates and determines

1) the explicit form of differential equations governing the underlying physical system,

2) the space-time structure of the Universe, and the mechanism of fundamental interac-
tions of Nature, and

3) physical conservation laws.

We now give a simple example to demonstrate how symmetry determines the explicit
form of equations. We begin with two basic implications of a symmetry:

a) Fundamental laws of Nature are universal, and their validity is independent of the
space, time, and directions of experiments and observations, and

b) By (2.1.12), the universality of laws of Nature implies that the differential equations
representing them are covariant. Equivalently the Lagrange actions are invariant un-
der the corresponding coordinate transformations.

As an example, consider a physical system obeying PLD 2.3. For simplicity, we assume
that the Lagrange action of this fictitious system is finite-dimensional. Namely, let

F : Rn → R 1

be a finite dimensional function regarded as the action of the physical system, and we need
to determine the explicit form of this function

F = F(x) for x = (x1 , · · · , xn ) ∈ Rn . (2.1.13)

Consider the case where the system satisfies the following invariance principle, called the
principle of rotational invariance.
42 Chapter 2 Fundamental Principles of Physics

Principle 2.6 (Rotational Invariance) The laws that the system (2.1.13) obeys are the
same under all orthogonal coordinate systems.

Based on the implication b) of symmetry above, Principle of Rotational Invariance 2.6

says that the action (2.1.13) is invariant under orthogonal coordinate transformations:

x& = Ax where A is an arbitrary orthogonal matrix. (2.1.14)

We know that the two forms as

|x|2 = x21 + · · · + x2n , x · y = x 1 y1 + · · · + x n yn

are invariant, where y ∈ Rn is given, and the invariant function F(x) must be of the form:

F(x) = F(|x|, x · y)

In addition, by the simplicity principle in Essence of Physics 2.2, the exponent of F in x is

two. Thus the action (2.1.13) is defined in the form as

F(x) = α |x|2 + β x · y, (2.1.15)

where α , β are two constants, and y is a given vector.

Thus we deduce an explicit expression (2.1.15) for the function F, which are invariant
under rotational symmetry, Principle 2.6. The graph of (2.1.15) is a sphere, and the invari-
ance of (2.1.15) means that the sphere looks the same from different directions. This is the
reason why an invariance is called a symmetric principle.

2.1.5 Invariance and tensors

Symmetries are characterized by three ingredients:

spaces, transformation groups, and tensors,

which are applicable to different physical fields. Here are the five currently known important
symmetries and their corresponding ingredients:

1) Galileo Invariance:

• Space: Euclidean space R3 ,

• Transformation group: Galileo group and SO(3),
• Tensor types: Cartesian tensors,
• Fields: classical mechanics, fluid dynamics, and astrophysics.
2.1 Essence of Physics 43

2) Lorentz Invariance:

• Space: Minkowski space,

• Transformation group: Lorentz groups,
• Tensor types: 4-dimensional tensors and spinors,
• Fields: Quantum physics and Interactions.

3) Einstein General Relativity:

• Space: 4-dimensional Riemann manifolds,

• Transformation group: GL(4),
• Tensor types: General tensors and Riemann metric,
• Fields: Gravitation and Astrophysics.

4) Gauge Invariance:

• Space: complex vector bundle M ⊗ p Cn ,

• Transformation groups: U(1) and SU(n),
• Tensor types: Wave functions and gauge fields,
• Fields: Quantum physics and Interactions.

5) Representation Invariance:

• Space: Tangent space of SU(n),

• Transformation: GL(Cn ),
• Tensor types: SU(n) tensors,
• Fields: Quantum physics and Interactions.

We now use examples in classical mechanics to illustrate the main characteristics of

these symmetries and to clarify how tensors describe invariances.

Principle 2.7 (Galilean Invariance) Mechanical laws are invariant under the following
transformations:

1) Galilean transformation

t = t,
& x& = x + vt for x ∈ R3 , (2.1.16)

where v is a constant velocity, and

44 Chapter 2 Fundamental Principles of Physics

2) translational and rotational transformations

t = t + t0 ,
& x& = Ax + b for x ∈ R3 , (2.1.17)

where A ∈ SO(3) is an orthogonal matrix, b ∈ R3 and t0 ∈ R1 are constant.

By (2.1.12) we know that the invariance of physical laws is equivalent to the covariance
of the corresponding differential equations. We now demonstrate that the Newton’s Second
Law obeys the Galileo Invariance Principle 2.7.
The Newton’s Second Law reads as

ma = F, (2.1.18)

and the corresponding differential equation describing the motion of a particle is given by

d 2x
m = F. (2.1.19)
dt 2
When we investigate the motion in other reference frame (&
x, &
t ) with the transformation

t = t + t0 ,
& x& = Ax + vt + b, (2.1.20)

by (2.1.20) we derive
d 2 x& d 2x
= A . (2.1.21)
d&t2 dt 2
On the other hand, experiments show that the expressions F& in (&
x, &
t ) and F in (x,t) of force
satisfy the relation:
F& = AF. (2.1.22)
Hence, by (2.1.19) we deduce from (2.1.21) and (2.1.22) that

d 2 x& &
m = F. (2.1.23)
d&t2
It is clear that both forms of (2.1.19) and (2.1.23) are the same under the transformation
(2.1.20), i.e. under the transformations (2.1.16) and (2.1.17).
Mathematically, the transformations corresponding to (2.1.16) and (2.1.17) are the Galileo
group and SO(3), and the functions satisfying (2.1.21) and (2.1.22) under the transformation
(2.1.20) are called first-order Cartesian tensors, which are also the usual vectors in R 3 .
Also, this example demonstrates that invariance principles must be characterized by
corresponding tensors. In the later sections, we shall give precise definitions of various
types of tensors.
Remark 2.8 In classical mechanics, the laws were first discovered phenomenolog-
ically, and then were known to obey the Galileo Principle of Invariance. It is Einstein’s
vision that laws of physics can be derived from symmetries.
2.1 Essence of Physics 45

Remark 2.9 The transformation (2.1.17) consists of three subclasses of transforma-

tions:
the rotation transformation: x& = Ax,
the space translation: x& = x + b, (2.1.24)
the time translation: t = t + t0 .
&
Unlike the Galileo invariance of (2.1.16), which is only valid in the classical mechanics,
the invariance for the transformations (2.1.24) are universally valid in all physical fields.
In particular, based on the Noether theorem, Theorem 2.38, we can deduce the following
corresponding conservation laws:
Time translation invariance ⇒ Energy conservation,
Space translation invariance ⇒ Momentum conservation,
Rotational invariance ⇒ Angular momentum conservation.
2.1.6 Geometric interaction mechanism
Albert Einstein was the first physicist who postulated that the gravitational force is caused by
the time-space curvature. However, Yukawa’s viewpoint, entirely different from Einstein’s,
is that the other three fundamental forces take place through exchanging intermediate bosons
such as photons for the electromagnetic interaction, W± and Z intermediate vector bosons
for the weak interaction, and gluons for the strong interaction.
Based on our recent studies on field theory of the four interactions, in the same spirit as
the Einstein’s mechanism of gravitational force, it is natural for us to postulate a mechanism
for all four interactions different from that of Yukawa.
To proceed, we recall that in geometry, the square ds2 of an infinitesimal arc-length in a
flat space can be written as
ds2 = dx21 + · · · + dx2n ,
which is the well known Pythagorean theorem, and in a curved space ds 2 is given by

ds2 = gi j (x)dxi dx j with gi j ̸= δi j , (2.1.25)

and the Pythagorean theorem is in general not true in a curved space. Mathematically, a
space M being flat indicates that one can choose properly a coordinate system so that the
metric
g i j = δi j ,
and otherwise, M will be curved.
Regarding to the laws of Nature on our Universe M , physical states are described by
functions u = (u1 , · · · , un ) defined on M :

u : M → M ⊗ p Rn for non-quantum system, (2.1.26)

u : M → M ⊗ p Cn for quantum system, (2.1.27)

46 Chapter 2 Fundamental Principles of Physics

which are solutions of differential equations associated with the laws of the underlying
physical system:
δ L(Du) = 0, (2.1.28)
where D is a derivative operator, and L is the Lagrange action. Consider two transformations
for the two physical systems (2.1.26) and (2.1.27):
x& = T x for (2.1.26), (2.1.29)

u& = eiθ τ u for (2.1.27), (2.1.30)

where x is a coordinate system in M , T : Rn → Rn is a linear transformation, eiθ τ : Cn → Cn

is an SU(n) transformation, θ is a function of x, and τ is a traceless Hermitian matrix; see
Section 2.4 on gauge theory.
We now state two important symmetric principles: the Einstein principle of general
relativity and the gauge invariance.

Principle 2.10 (General Relativity) Laws of Physics are the same under all coordinate
systems. Namely, equations (2.1.28) are covariant and equivalently the action L is invariant
under all coordinate transformations (2.1.29).

Principle 2.11 (Gauge Invariance) A quantum system with electromagnetic, weak,

and strong interactions is invariant under the corresponding SU(n) gauge transformations
(2.1.30).

One important consequence of the invariance of (2.1.28) under the transformations

(2.1.29) and (2.1.30) is that the derivatives D in (2.1.28) must take the following form (see
Section 2.4 for detailed derivations):
D = ∇+Γ for (2.1.26), (2.1.31)

D = ∇ + igAτ for (2.1.27), (2.1.32)

where Γ depends on the metrics gi j , A is a gauge field, representing the interaction potential,
and g is the coupling constant, representing the interaction charge.
The derivatives defined in (2.1.31) and (2.1.32) are called connections respectively on
M and on the complex vector bundle M ⊗ p Cn . For the connections, we have the following
theorem, providing a mathematical basis for our interaction mechanism.

Theorem 2.12

1) The space M is curved if and only if Γ ̸= 0 in all coordinates, or equivalently g i j ̸= δi j

under all coordinate systems;

2) The complex bundle M ⊗ p Cn is geometrically nontrivial or twisted if and only if

A ̸= 0.
2.1 Essence of Physics 47

Consequently, by Principle 2.10 the presence of the gravitational field implies that
the space-time manifold is curved, and, by Principle 2.11, the presence of the electromag-
netic, the weak and strong interactions indicates that the complex vector bundle M ⊗ p Cn
is twisted:
Principle 2.10 ⇒ gi j ̸= δi j ⇒ M is curved,
(2.1.33)
Principle 2.11 ⇒ A ̸= 0 ⇒ M ⊗ p Cn is twisted.
This analogy, together with Einstein’s vision on gravity as the curved effect of space-
time manifold, it is natural for us to postulate the following mechanism for all four interac-
tions.
Geometric Interaction Mechanism 2.13 The gravitational force is the curved effect
of the time-space, and the electromagnetic, weak, strong interactions are the twisted effects
of the underlying complex vector bundles M ⊗ p Cn .
As mentioned earlier, traditionally one adopts Yukawa’s viewpoint that forces of the
interactions of Nature are caused by exchanging the field mediators. Namely, the gravitation
is due to exchanging gravitons, the electromagnetic force is due to exchanging photons, the
weak force is due to exchanging the intermediate vector bosons, and the strong force is due
to exchanging gluons. The point of view we are taking is the Interaction Mechanism 2.13,
following Einstein’s version.

2.1.7 Principle of symmetry-breaking

Different physical systems obey different physical principles. The four fundamental inter-
actions of Nature, the quantum systems, the fluid dynamics and heat conduction obey the
following symmetry principles:
the general relativity for gravity,
the Lorentz and gauge invariances for the other three interactions,
(2.1.34)
the Lorentz invariance for quantum systems, and
the Galilean invariance for fluid and heat conductions.
The corresponding fields and systems in (2.1.34) are governed by the following physical
laws and first principles:
PID and PRI for four fundamental interactions,
PLD for quantum systems,
(2.1.35)
the Newton Second Law for fluids,
diffusion laws for heat conductivity,
Here PID stands for the principle of interaction dynamics, and PRI stands for the principle
of representation invariance, both discovered recently by the authors. PLD stands for the
principle of Lagrangian dynamics.
48 Chapter 2 Fundamental Principles of Physics

Astrophysics is the only field that involving all the fields in (2.1.34) and (2.1.35). Con-
sequently, one needs to couple the basic laws in (2.1.35) to model astrophysical dynamics.
One difficulty we encounter now is that the Newtonian Second Law for fluid motion
and the diffusion law for heat conduction are not compatible with the principle of general
relativity. Also, there are no basic principles and rules for combining relativistic systems and
the Galilean systems together to form a consistent system. The reason is that in a Galilean
system, time and space are independent, and physical fields are 3-dimensional; while in a
relativistic system, time and space are related, and physical fields are 4-dimensional.
The distinction between relativistic and Galilean systems gives rise to an obstacle for
establishing a consistent model of astrophysical dynamics, coupling all the physical systems
in (2.1.34) and (2.1.35).
In the unified field theory based on PID, the key ingredient for coupling gravity with
the other three fundamental interactions is achieved through spontaneous gauge symme-
try breaking. Here we propose that the coupling between the relativistic and the Galilean
systems through relativistic-symmetry breaking.
In fact, the model given by (7.1.75)-(7.1.76) follows from this symmetry-breaking prin-
ciple, where we have to chose the coordinate system

xµ = (x0 , x), x0 = ct and x = (x1 , x2 , x3 ),

such that the metric is in the form:

! "
2
ds = − 1 + 2 φ (x,t) c2 dt 2 + gi j (x,t)dxi dx j .
2
(2.1.36)
c
Here gi j (1 ! i, j ! 3) are the spatial metric, and

φ = the gravitational potential. (2.1.37)

With this metric (2.1.36)-(2.1.37), we can establish the fluid and heat equations as in (7.1.78).
It is then clear that by fixing the coordinate system to ensure that the metric is in the form
(2.1.36)-(2.1.37), the system breaks the symmetry of general coordinate transformations,
and we call such symmetry-breaking as relativistic-symmetry breaking.
We believe the symmetry-breaking is a general phenomena when we deal with a physical
system coupling different subsystems in different levels. The unified field theory for the four
fundamental interactions is a special case, which couples the general relativity, the Lorentz
and the gauge symmetries. Namely, the symmetry of general relativity needs to be linked to
both the Lorentz invariance and the gauge invariance in two aspects as follows:

1) In the Dirac equations for the fermions:

c
iγ µ Dµ ψ − mψ = 0,
h̄
γ µ have to obey two different transformations.
2.2 Lorentz Invariance 49

2) The gauge-symmetry breaks in the gravitational field equations coupling the other
interactions:
1 8π G
Rµν − gµν R = − 4 Tµν + Dµ Φν , (2.1.38)
2 c
where
k1 k2 k3
Dµ = ∇µ + eAµ + gwWµ + gs Sµ , (2.1.39)
h̄c h̄c h̄c
∇µ is the covariant derivative, ki (1 ! i ! 3) are parameters, A µ ,Wµ , S µ are the total
electromagnetic, weak and strong interaction potentials. It is the terms A µ ,Wµ , S µ in
(2.1.39) that break the gauge symmetry of (2.1.38).

In summary, we are ready to postulate a general symmetry-breaking principle.

Principle of Symmetry-Breaking 2.14
1) The three sets of symmetries,
the general relativistic invariance,
the Lorentz and gauge invariances, and (2.1.40)
the Galileo invariance,
are mutually independent and dictate in part the physical laws in different levels of
Nature; and

2) for a system coupling different levels of physical laws, part of these symmetries must
be broken.

This principle of symmetry-breaking holds the key component for us to establish the
PID unified field theory for four fundamental interactions in Chapter 4, the field equations
for multi-particle systems in Chapter 6, and the astrophysical fluid dynamical equations in
Chapter 7, resolving a number of important physical problems.

2.2 Lorentz Invariance

2.2.1 Lorentz transformation
In 1903, H. A. Lorentz discovered the Lorentz coordinate transformation, under which the
laws of electromagnetism are invariant. Consequently, physicists discovered two different
invariances:

the Galilean invariance in classical mechanics, (2.2.1)

the Lorentz invariance in electromagnetism. (2.2.2)

In 1905, Albert Einstein introduced the special theory of relativity, based on the follow-
ing two first principles: the principle of special relativity and the principle of invariance of
the speed of light.
50 Chapter 2 Fundamental Principles of Physics

Principle 2.15 (Special Relativity) Physical laws are covariant in all inertial systems.

Principle 2.16 (Invariance of speed of light) The vacuum speed of light is a universal
constant.

A few remarks are now in order. First, a system of reference is always needed to describe
the nature, and an inertial system is a system on which a freely moving object moves with
constant velocity (Landau and Lifshitz, 1975). Second, it is clear that under the Galilean in-
variance, the speed of light changes in different inertial systems. Consequently, the Galilean
invariance and the Lorentz invariance are incompatible. Third, the invariance of the vacuum
speed of light was verified by the Michelson and Morley experiment. Einstein discovered
his special theory of relativity by postulating, based on Principles 2.15-2.16, the following
principle of Lorentz invariance.

Principle 2.17 (Lorentz Invariance) Physical laws are invariant under Lorentz trans-
formations.

With this principle, the special theory of relativity was developed in two directions:
a) the introduction of relativistic mechanics, replacing the classical mechanics, and b) the
development of relativistic quantum mechanics.

Figure 2.1

We now introduce typical Lorentz transformations 1 . Let (x,t) and (& x, &
t ) be two inertial
systems which are in relativistic motion with a constant velocity v, as shown in Figure 2.1,
where v is in the x1 -axis direction. Then the Lorentz transformation for the two inertial
systems is given by
6 7
x 1 − vt t − x1 v/c2
x1 , x&2 , x&3 ) = 3
(& , x2 , x3 , t=3
& , (2.2.3)
1 − v2/c2 1 − v2/c2

where c is the speed of light.

1 General Lorentz transformations are linear transformations that preserves the Minkowski metric (2.2.6),

and the transformation (2.2.3) is called a boost.

2.2 Lorentz Invariance 51

The vacuum speed of light is invariant under the above Lorentz transformation (2.2.3).
x1 /d&
To see this, notice that the speed of light is c& = d& t measured in the x&-coordinate system,
1
and is c = dx /dt measured in the x-coordinate system. Consequently,

x1
d& dx1 − vdt 8 dt − dx1 v/c2 dx1 /dt − v c−v
c& = =3 3 = 1 2
= = c.
dt
& 2
1 − v /c 2 2
1 − v /c 2 1 − (dx /dt)(v/c ) 1 − v/c

2.2.2 Minkowski space and Lorentz tensors

We recall that each symmetry is characterized by three ingredients: space/manifold, trans-
formation group, and tensors. For the Lorentz invariance, the transformation group is the
Lorentz group, consisting of Lorentz transformations such as those given by (2.2.3), and the
corresponding space is the Minkowski space introduced below.
Minkowski Space
We know that the Newtonian mechanics is defined in the Euclidean space R 3 , and the
time t is regarded as a parameter. From the Lorentz transformation (2.2.3), we see that
in relativistic physics, there is no absolute space and time, and the time t cannot be sepa-
rated from the space R3 . Namely the Minkowski space M 4 is a 4-dimensional space-time
manifold defined by

M 4 = {(x0 , x1 , x2 , x3 ) | x0 = ct, (x1 , x2 , x3 ) ∈ R3 }, (2.2.4)

where c is the speed of light, and M 4 is endowed with the Riemannian metric:
⎛ ⎞
−1 0 0 0
⎜ 0 1 0 0⎟
(g µν ) = ⎜
⎝ 0 0 1 0⎠ .
⎟ (2.2.5)
0 0 0 1

The doublet {M 4 , g µν } given by (2.2.4) and (2.2.5) is called the Minkowski space, and the
metric (2.2.5) can be equivalently expressed in the form

ds2 = g µν dxµ dxν = −c2 dt 2 + dx2 = −c2 dt 2 + (dx1 )2 + (dx2 )2 + (dx3 )2 . (2.2.6)

Lorentz transformations are linear transformations of the Minkowski space M 4 that

preserve the metric (2.2.6). The following coordinate transformation of the Minkowski
space is a Lorentz transformation, called boost, corresponding to Figure 2.1:
⎛ ⎞
1 β
⎛ 0 ⎞ 3 −3 0 0 ⎛ 0 ⎞
x& ⎜ 1−β2 1−β2 ⎟ x
⎜ ⎟
⎜ x&1 ⎟ ⎜
⎜ 2 ⎟ = ⎜− 3 β 1 ⎟ ⎜ x1 ⎟
⎟⎜ 2 ⎟,
⎝ x& ⎠ ⎜ 3 0 0 ⎟⎝ x ⎠ (2.2.7)
⎜ 1−β2 1−β2 ⎟
3 ⎝ 3
x& 0 0 1 0⎠ x
0 0 0 1
52 Chapter 2 Fundamental Principles of Physics

where β = ν /c. Both transformations (2.2.3) and (2.2.7) are the same in form, but the math-
ematical implication is changed. Here, (2.2.7) represents the coordinate transformation for
the Minkowski space M 4 . With Einstein’s summation convention, (2.2.7) is often denoted
by
µ
x&µ = Lν xν , (2.2.8)
where
⎛ ⎞
1 β
3 −3 0 0
⎜ 1 −β2 1−β2 ⎟
⎜ ⎟
µ ⎜ β 1 ⎟
(Lν ) = ⎜
⎜− 31 − β 2 3 0 0⎟
⎟ (2.2.9)
⎜ 1−β2 ⎟
⎝ 0 0 1 0⎠
0 0 0 1
µ µ
is the Lorentz matrix, and its inverse (lν ) = (Lν )−1 is given by
⎛ ⎞
1 β
3 3 0 0
⎜ 1−β2 1−β2 ⎟
⎜ ⎟
µ ⎜ β 1 ⎟
(lν ) = ⎜
⎜ 31 − β 2 3 0 0⎟
⎟. (2.2.10)
⎜ 1−β2 ⎟
⎝ 0 0 1 0⎠
0 0 0 1

An important property of the Minkowski space is that its Riemannian metric is invariant
under the Lorentz transformation (2.2.7).

Theorem 2.18 The Minkowski metric (2.2.5) or (2.2.6) is invariant under the co-
ordinate transformation (2.2.7). Namely the metric (& xµ } is the same as that in
g µν ) in {&
µ
{x }:
g µν ) = (g µν ).
(& (2.2.11)
xµ } is also expressed as
In other words, ds2 in {&

ds2 = −c2 d&

t 2 + d&
x2 , x1 , x&2 , x&3 ).
x& = (&

The proof of Theorem 2.18 needs to use the properties of tensors. In fact, the Minkowski
metric (2.2.5) is a second-order covariant Lorentz tensor, i.e. under the transformation
(2.2.8), g µν transforms as
g µν ) = (lβα )(g µν )(lβα )T ,
(& (2.2.12)
µ µ
where (lµ ) is the inverse of (Lν ) given by (2.2.10). Then, a direct computation we can get
(2.2.11) from (2.2.12).
2.2 Lorentz Invariance 53

Lorentz Transformation Group

µ
Each transformation (2.2.8) corresponds to a Lorentz matrix (Lν ) given by (2.2.9).
These matrices constitute a group in the multiplication as
⎛ ⎞⎛ ⎞
1 β 1 γ
3 −3 0 0 3 −3 0 0
⎜ 1−β2 1−β2 ⎟⎜ 1 − γ2 1 − γ2 ⎟
⎜ ⎟⎜ ⎟
⎜ β 1 ⎟⎜ γ 1 ⎟
⎜
⎜ −3 3 0 0 ⎟
⎟⎜
⎜ −3 3 0 0 ⎟
⎟
⎜ 1−β2 1−β2 ⎟⎜ 1 − γ2 1 − γ2 ⎟
⎝ 0 0 1 0 ⎠⎝ 0 0 1 0 ⎠
0 0 0 1 0 0 0 1
⎛ 1 α ⎞
√ −√ 0 0
⎜
⎜ 1 − α2 1 − α2 ⎟
⎟
⎜ α 1
=⎜ −√ √ 0 0 ⎟⎟,
⎜ 1 − α2 1 − α2 ⎟
⎝ 0 0 1 0 ⎠
0 0 0 1

where
v u w u+v
β= , γ= , α= , w= ,
c c c 1 + uv/c2
where w is the velocity composed by u and v by the theory of special relativity.
Thanks to Theorem 2.18, we can define Lorentz transformation group as all linear trans-
formations of the Minkowski space M 4 , that preserve the Minkowski metric; see (2.2.12):
µ µ
LG = {(Lν ) : M 4 → M 4 | (g µν ) = (Lαβ )(g µν )(Lαβ )T , det(Lν ) = 1}. (2.2.13)

Elements of LG are also called Lorentz matrices, and relativistic physics is referred to the
invariance of action under the Lorentz group. The invariance of the Minkowski metric
implies the Lorentz invariance of physical laws.
µ
In this definition, we require det(Lν ) = 1, and such transformations are often called
proper Lorentz transformation in physics literatures. The parity transformation and time
reversal are two special linear transformations of the Minkowski space, which are not ele-
ments in LG as defined in (2.2.13) and are often dealt with separately:
⎛ ⎞ ⎛ ⎞
1 0 0 0 −1 0 0 0
⎜0 −1 0 0⎟ ⎜ 0 1 0 0⎟
P=⎜ ⎝0 0 −1 0 ⎠ ,
⎟ T =⎜⎝ 0 0 1 0⎠ .
⎟ (2.2.14)
0 0 0 −1 0 0 0 1

Lorentz Tensors

We now define Lorentz tensors, also called 4-dimensional (4-D) tensors, corresponding
to the Lorentz invariance.
54 Chapter 2 Fundamental Principles of Physics

Definition 2.19 (Lorentz Tensors) The following quantities are called Lorentz tensors,
or 4-dimensional tensors:

1) A function T with 4k components

T = {Tµ1 ···µk }, µ1 , · · · , µk = 0, 1, 2, 3,

is called a k-th order covariant tensor, if under the Lorentz transformation (2.2.13),
the components of T transform as
ν
T&µ1 ···µk = lµν11 · · · lµkk Tν1 ···νk ,
µ
where (lµν ) = (Lν )−1 is the inverse transformation of the Lorentz transformation
µ
(Lν ) ∈ LG.

2) A tensor
T = {T µ1 ··· µk }, µ1 , · · · , µk = 0, 1, 2, 3,
µ
is called a k-th order contra-variant tensor, if under the Lorentz transformation (L ν ) ∈
LG, the components of T change as
µ µ
T& µ1 ···µk = Lν11 · · · Lνkk T ν1 ···νk .

3) A function
µ ··· µ
T = {Tν11···νs r }, r + s = k,
µ
is called a k-th order (r, s)-type tensor, if under the Lorentz transformation (L ν ) ∈ LG,
µ ··· µ µ µ β ···β
T&ν11···νs r = lνα11 · · · lναss Lβ 1 · · · Lβrr Tα11···αsr .
1

In physics, the most important tensors are first- and second-order tensors. The following
is a list of commonly encountered 4-D tensors, and we always use (g µν ) = (g µν )−1 with
(g µν ) being the Minkowski metric given by (2.2.5):

1) Position vectors:

xµ = (x0 , x1 , x2 , x3 ), x0 = ct,
xµ = (x0 , x1 , x2 , x3 ) = g µν xν = (−x0 , x1 , x2 , x3 ). (2.2.15)

2) The 4-D electromagnetic potential:

Aµ = (A0 , A1 , A2 , A3 ),
Aµ = (A0 , A1 , A2 , A3 ) = g µν Aν , (2.2.16)

where A0 = −A0 is the electric potential, (A1 , A2 , A3 ) = (A1 , A2 , A3 ) is the magnetic

vector potential.
2.2 Lorentz Invariance 55

3) The 4-D current density:

Jµ = (J0 , J1 , J2 , J3 ),
J µ = (J 0 , J 1 , J 2 , J 3 ) = g µν Jν , (2.2.17)

where J 0 = −J0 = cρ , ρ is the charge density, and J⃗ = (J1 , J2 , J3 ) is the current

density field.

4) The 4-D energy-momentum vector:

Eµ = (E, cP1 , cP2 , cP3 ),
E µ = g µν Eν = (−E, cP1 , cP2 , cP3 ), (2.2.18)

where E is the energy, and P = (P1 , P2 , P3 ) is the momentum vector,

5) The 4-D gradient operators:

! " ! "
∂ ∂ ∂ ∂ 1∂
∂µ = , , , = ,∇ ,
∂ x0 ∂ x1 ∂ x2 ∂ x3 c ∂t
! " ! " (2.2.19)
µ µν ∂ ∂ ∂ ∂ 1∂
∂ = g ∇µ = − 0 , 1 , 2 , 3 = − ,∇ ,
∂x ∂x ∂x ∂x c ∂t
transform as the first-order Lorentz tensors.
2.2.3 Relativistic invariants
All Lagrange actions in relativistic physics are Lorentz invariants. The most common
Lorentz invariants are contractions of 4-D tensors. For example, let

Aµ = (A0 , A1 , A2 , A3 ), Bµ = (B0 , B1 , B2 , B3 )

be the covariant and contra-variant vectors. Then the contraction

A µ B µ = A0 B0 + A1 B1 + A2 B2 + A3 B3 (2.2.20)

is a Lorentz invariant. In fact, under the Lorentz transformation (2.2.8), A µ and Bµ satisfy
&µ = Lνµ Aν , µ
A B&µ = lν Bν .

It follows that
&µ B&µ = Lαµ l µ Aα Bβ .
A (2.2.21)
β
Since (lµν ) = (Lνµ )−1 is the inverse matrix of (Lνµ ), then
µ
Lαµ lβ = δβα .

Thus, (2.2.21) becomes

&µ B&µ = Aµ Bµ ,
A
which shows that the contraction (2.2.20) is a Lorentz invariant. In fact, the following
theorem can be easily verified in the same fashion.
56 Chapter 2 Fundamental Principles of Physics

Theorem 2.20 (Lorentz Invariants)

1) The contractions
Aµ1 ···µk Bµ1 ···µk , (2.2.22)
is a Lorentz invariant, and so are the following contractions:

gαβ Aα Bβ , gαβ Aα Bβ . (2.2.23)

µ
2) Let Aν be a (1,1) Lorentz tensor, then the contraction
µ
Aµ = A00 + A11 + A22 + A33 (2.2.24)

is a Lorentz invariant.

3) The following differential operators

! "
∂ ∂ ∂ ∂
A µ ∂ µ = A0 , A1 , A2 , A3 ,
∂ x0 ∂ x1 ∂ x2 ∂ x3
! "
∂ ∂ ∂ ∂
A µ ∂ µ = A0 0 , A1 1 , A2 2 , A3 3 , (2.2.25)
∂x ∂x ∂x ∂x
1 ∂2 ∂2 ∂2 ∂2
∂ µ ∂µ = − + + +
c2 ∂ t 2 ∂ x21 ∂ x22 ∂ x23

are Lorentz invariant operators.

Theorem 2.20 provides a number of typical Lorentz invariants through contractions, and
in fact, all Lagrange actions in relativistic physics are combinations of the Lorentz invariants
in (2.2.22)-(2.2.25).

2.2.4 Relativistic mechanics

First, the 4-D velocity is defined by
! "
dx0 dx1 dx2 dx3
uµ = (u0 , u1 , u2 , u3 ) = , , , ,
ds ds ds ds

where ds is the arc-length element in (2.2.6), and is given by

?
dxk
ds = c 1 − v2/c2 dt, v = (v1 , v2 , v3 ), vk = .
dt
It is clear that the 4-D velocity u µ can be expressed as

u µ = (u0 , u1 , u2 , u3 ),
1 vk /c (2.2.26)
u0 = 3 , uk = 3 for 1 ! k ! 3.
1 − v2/c2 1 − v2/c2
2.2 Lorentz Invariance 57

Second, the 4-D acceleration is defined by

du µ
aµ = = (a0 , a1 , a2 , a3 ),
ds 6 7
0 1 d 1
a = 3 3 , (2.2.27)
c 1 − v2/c2 dt 1 − v2/c2
6 7
k 1 d vk
a = 3 3 for 1 ! k ! 3.
c2 1 − v2/c2 dt 1 − v2/c2

Third, the 4-D energy-momentum vector is

E µ = (E, cP1 , cP2 , cP3 ),

mc2
E=3 ,
1 − v2/c2 (2.2.28)
mvk dxk
Pk = 3 , vk = for 1 ! k ! 3.
1 − v2/c2 dt

Fourth, by (2.2.28), we derive the most important formula in relativistic, called the Ein-
stein energy-momentum relation:

E 2 = c 2 P 2 + m2 c4 . (2.2.29)

Remark 2.21 In 1905, Albert Einstein conjectured that a static object with mass m
have energy E, and satisfy the relation, called the Einstein formula:

E = mc2 . (2.2.30)

Thus, in a static coordinate system the energy-momentum is in the form

(E, cP) = (mc2 , 0) (2.2.31)

Then, it follows from (2.2.31) that the energy-momentum (E, cP) of a moving object with
velocity v is taken in the form of (2.2.28), which implies that the relation (2.2.29) holds true.
Hence, the energy-momentum relation (2.2.29) is based on postulating (2.2.30), which was
verified by numerous experiments.

Fifth, in classical mechanics, the Newtonian second law takes the form
d
P = F, P = mv is the momentum.
dt
In the relativistic mechanics, the 4-D force is
! "
dE dP
Fµ = ,c ,
ds ds
58 Chapter 2 Fundamental Principles of Physics

and the relativistic force is F = (F 1 , F 2 , F 3 ):

1 d k
Fk = 3 P for1 ! k ! 3,
1 − v2/c2 dt

where P = (P1 , P2 , P3 ) is as in (2.2.28). Thus, it follows that the relativistic motion law is
given by
?
d k mvk
P = 1 − v2/c2 F k , Pk = 3 for k = 1, 2, 3. (2.2.32)
dt 1 − v2/c2
2.2.5 Lorentz invariance of electromagnetism
The Maxwell equations for electromagnetic fields are
1 ∂H
curl E = − , (2.2.33)
c ∂t
div H = 0, (2.2.34)
1 ∂ E 4π
curl H = + J, (2.2.35)
c ∂t c
div E = 4πρ , (2.2.36)

where E, H are the electric and magnetic fields, and J is the current density, and ρ the charge
density.
To show the Lorentz invariance of the Maxwell equations (2.2.33)-(2.2.36), we need to
express them in the form of the 4-D electromagnetic potential and current density.
The electromagnetic potential and current density are briefly introduced in (2.2.16) and
(2.2.17):
Aµ = (A0 , A1 , A2 , A3 ),
(2.2.37)
Jµ = (J0 , J1 , J2 , J3 ), J0 = −cρ .
Using the Lorentz tensor operator
! "
∂ ∂ ∂ ∂ ∂
∂µ = = , 1, 2, 3 ,
∂ xµ ∂x ∂x ∂x ∂x
0

we can construct two second-order Lorentz tensors:

∂ Aν ∂ A µ
Fµν = − ν,
∂ xµ ∂x (2.2.38)
1
Gµν = εµναβ gακ gβ λ Fκλ ,
2
where gαβ is the Minkowski metric, and
⎧
⎪ 1 (µ , ν , α , β ) is an even permutation of (0123),
⎨
εµναβ = − 1 (µ , ν , α , β ) is an odd permutation of (0123),
⎪
⎩
0 otherwise,
2.2 Lorentz Invariance 59

is a 4-th order Lorentz tensor. Thanks to the relations:

1 ∂ ⃗A
H = curl ⃗A, E = ∇A0 − , ⃗A = (A1 , A2 , A3 ),
c ∂t
or equivalently,
∂ A3 ∂ A2 ∂ A1 ∂ A3 ∂ A2 ∂ A1
H1 = − 3, H2 = 3 − 1 , H3 = 1 − 2 ,
∂x 2 ∂x ∂x ∂x ∂x ∂x
∂ A0 ∂ A1 ∂ A0 ∂ A2 ∂ A0 ∂ A3
E1 = 1 − 0 , E2 = 2 − 0 , E3 = 3 − 0 ,
∂x ∂x ∂x ∂x ∂x ∂x
the first pair of the Maxwell equations (2.2.33) and (2.2.34) are in the form:
∂ Gµν
=0 for µ = 0, 1, 2, 3. (2.2.39)
∂ xν
The second pair of the Maxwell equations (2.2.35) and (2.2.36) are
∂ Fµν 4π
= Jµ for µ = 0, 1, 2, 3. (2.2.40)
∂ xν c
It is clear that the Maxwell equations (2.2.39) and (2.2.40) are covariant under the Lorentz
transformations. In fact, (2.2.39) and (2.2.40) can be equivalently written as
4π
∂ ν Gµν = 0, ∂ ν Fµν = Jµ .
c
Also, two electromagnetic dynamic equations and the charge conservation law are writ-
ten as
∂ρ
+ div J = 0. (2.2.41)
∂t
The motion equation in an electromagnetic field is given by
dv e
m = eE + v × H. (2.2.42)
dt c
It is clear that (2.2.41) can be written as

∂µ J µ = 0, (2.2.43)

which is Lorentz invariant.

The Lorentz covariance of the motion equation in an electromagnetic field follows from
the following Lorentz covariant formulation of (2.2.42):
du µ e
m = 2 Fµν uν for µ = 0, 1, 2, 3, (2.2.44)
ds c
where u µ is the 4-D velocity given by (2.2.26), and Fµν is as in (2.2.38).
In summary, all electromagnetic equations can be written in the Lorentz covariant forms
given by (2.2.39)-(2.2.40), (2.2.43) and (2.2.44), and the Lorentz invariance of these equa-
tions follows.
60 Chapter 2 Fundamental Principles of Physics

2.2.6 Relativistic quantum mechanics

Quantum physics is based on several fundamental principles, also called basic postulates of
quantum mechanics, which will be introduced systematically in Chapter 6.
For our purpose, we introduce hereafter two of these basic postulates.
Basic Postulate 2.22 An observable physical quantity can be represented by a Her-
mitian linear operator. In particular, the energy E, momentum ⃗P, scalar-valued momentum
P are represented by the operators given by
∂ ⃗P = −ih̄∇,
E = ih̄ , (2.2.45)
∂t
P0 = ih̄(⃗σ · ∇) for massless fermions,
(2.2.46)
P1 = −ih̄(⃗α · ∇) for massive fermions,
where h̄ is the Plank constant, ⃗σ = (σ1 , σ2 , σ3 ), ⃗α = (α1 , α2 , α3 ), σk (1 ! k ! 3) are the
Pauli matrices defined by
! " ! " ! "
0 1 0 −i 1 0
σ1 = , σ2 = , σ3 = , (2.2.47)
1 0 i 0 0 −1
and αk (1 ! k ! 3) are 4-th order matrices given by
! " ! " ! "
0 σ1 0 σ2 0 σ3
α1 = , α2 = , α3 = . (2.2.48)
σ1 0 σ2 0 σ3 0
Basic Postulate 2.23 For a quantum system with observable Hermitian operators
L1 , · · · , Lm , if the physical quantities lk corresponding to Lk (1 ! k ! m) satisfy a relation

H(l1 , · · · , lm ) = 0, (2.2.49)

then the following equation induced by (2.2.49)

H(L1 , · · · , Lm )ψ = 0 (2.2.50)

may give a model describing this system provided the operator H(L 1 , · · · , Lm ) in (2.2.49) is
Hermitian.
Remark 2.24 If the operator H(L1 , · · · , Lm ) in (2.2.49) is irreducible, then (2.2.49)
must describe the system.
Based on Postulates 2.22 and 2.23, we derive three basic equations of relativistic quan-
tum mechanics: the Klein-Gordon equation describing the bosons, the Weyl equations de-
scribing massless free fermions, and the Dirac equations describing massive fermions.
1. Klein-Gordon equation. By the Einstein energy m omentum relation for energy E,
momentum P and the rest mass m:

E 2 − c2⃗P2 − m2c4 = 0,
2.2 Lorentz Invariance 61

and by (2.2.45), we derive from Postulate 2.23 an relativistic equation, called the Klein-
Gordon (KG) equation: #
1 ∂2 ) mc *2 $
2
− ∇ + ψ = 0. (2.2.51)
c2 ∂ t 2 h̄
The KG equation describes a free boson with spin J = 0.
When an electromagnetic field presents, the operators in (2.2.45) and (2.2.46) are rewrit-
ten as1
∂ e
E = ih̄ − eA0 , ⃗P = −ih̄∇ − ⃗A,
∂t c (2.2.52)
P0 = ih̄(⃗σ · ⃗D), P1 = −ih̄(⃗α · ⃗D),
where A µ = (A0 , ⃗A) is the electromagnetic potential, and

⃗D = ∇ + ie ⃗A.
h̄c
Thus, as electromagnetic field presents, the Klein-Gordon equation (2.2.51) is expressed as
! ) mc *2 "
µ
Dµ D − ψ = 0, (2.2.53)
h̄

where e
Dµ = ∂µ + i
Aµ , (2.2.54)
h̄c
is a 4-dimensional vector operator. The expression (2.2.53) is clearly Lorentz invariant.
2. Weyl equations. Based on the de Broglie relation

E = h̄ω , P = h̄/λ , c = ωλ ,

we obtain
E = cP. (2.2.55)
The relation (2.2.55) is valid to a massless free fermion, e.g. as neutrinos. Inserting E and
P0 in (2.2.45) and (2.2.46) into (2.2.55), we obtain that

∂ψ
= c(⃗σ · ∇)ψ , (2.2.56)
∂t
which are called the Weyl equations describing the free massless neutrinos. Here ψ =
(ψ1 , ψ2 )T is a two component Weyl spinor.
3. Dirac equations. For a massive fermion, the de Broglie matter-wave duality relation
can be generalized in the form

E = h̄ω ± mc2 , P = h̄/λ , c = ωλ .

1 Here e is the electric charge of an electron, and is negative.
62 Chapter 2 Fundamental Principles of Physics

Then we have
E = cP ± mc2 . (2.2.57)

Inserting the operators E and P1 in (2.2.45) and (2.2.46) into (2.2.57), and taking the mass
operator as ! " ! "
I 0 1 0
±mc2 = mc2 α0 , α0 = , I= ,
0 −I 0 1
we derive the following Dirac equations:

∂ψ
ih̄ = −ih̄c(⃗α · ∇)ψ + mc2α0 ψ , (2.2.58)
∂t
where ψ = (ψ1 , ψ2 , ψ3 , ψ4 )T is a four-component Dirac spinor, and ⃗α = (α1 , α2 , α3 ) is as
defined by (2.2.48).
Usually, we multiply both sides of (2.2.58) by the matrix α0 , and denote

γ µ = (γ 0 , γ 1 , γ 2 , γ 3 ), (2.2.59)

where
! " ! "
I 0 0 σk
γ 0 = α0 = , γ k = α0 αk = for 1 ! k ! 3.
0 −I − σk 0

The matrices γ µ are the Dirac matrices. Then the Dirac equations (2.2.58) are in the form
) mc *
iγ µ ∂µ − ψ = 0. (2.2.60)
h̄
When we consider the case under an electromagnetic field, we need to replace ∂ µ in
(2.2.60) by D µ given by (2.2.54).

Remark 2.25 The reason why the scalar-valued momentum operators P are taken in
the form of (2.2.46) is due to the following Einstein energy-momentum formulas:

E 2 = c2 P02 = c2⃗P2 for mass m = 0,

2 2 2 2⃗ 2 2 4
E = c (P1 + mcα0 ) = c P + m c for m ̸= 0,

where P0 , P1 are as in (2.2.46), and ⃗P is as in (2.2.45).

2.2.7 Dirac spinors

The Dirac matrices (2.2.59) are not invariant under the Lorentz transformations, i.e. γ µ =
(γ 0 , γ 1 , γ 2 , γ 3 ) is not a 4-D vector operator. Hence, the covariance of the Dirac equations
(2.2.60) requires that under the transformation
µ
x&µ = Lxν , L = (Lν ) as in (2.2.9), (2.2.61)
2.2 Lorentz Invariance 63

the left-hand side of (2.2.60) should be

) mc * ) mc *
iγ µ ∂&µ − & = R iγ µ ∂µ −
ψ ψ, (2.2.62)
h̄ h̄

where ∂&µ = ∂ /∂ x&µ , and

ψ
& = Rψ . (2.2.63)

Here R is a 4 × 4 matrix depending on the Lorentz matrix L in (2.2.61). To determine the

matrix R, we note that
∂ ∂
µ
= Lνµ ν , (Lνµ ) = L. (2.2.64)
∂ x& ∂x
Inserting (2.2.63) and (2.2.64) into (2.2.62), we deduce that
! " ! "
∂ mc ∂ mc
iR−1 γ µ Lνµ R ν + ψ = iγ ν ν + ψ.
∂x h̄ ∂x h̄

It follows that
R−1 γ µ Lνµ R = γ ν ,

which is equivalent to
µ
R−1 γ µ R = Lν γ ν for µ = 0, 1, 2, 3. (2.2.65)

Hence the covariance of (2.2.62) is equivalent to the transformation matrix R in (2.2.63)

obeying equations (2.2.65).
To derive an explicit form of R in (2.2.65), we need to write the Lorentz matrix L in the
form ⎛ ⎞
cosh θ − sinh θ 0 0
µ ⎜− sinh θ cosh θ 0 0⎟
Lν = ⎜
⎝ 0
⎟,
0 1 0⎠
0 0 0 1
where cosh θ and sinh θ are the hyperbolic functions and θ satisfies

1 v/c
cosh θ = 3 , sinh θ = 3 .
1 − v2/c2 1 − v2/c2

In this form, the equations (2.2.65) can be written as

R−1 γ 0 R = cosh θ γ 0 − sinh θ γ 1 ,

R−1 γ 1 R = − sinh θ γ 0 + cosh θ γ 1 ,
(2.2.66)
R−1 γ 2 R = γ 2 ,
R−1 γ 3 R = γ 3 .
64 Chapter 2 Fundamental Principles of Physics

By the expressions of γ µ , we infer from (2.2.66) that

⎛ θ θ⎞
cosh 0 0 − sinh
⎜ 2 2⎟
⎜ θ θ ⎟
θ θ 0 1 ⎜⎜ 0 cosh
2
− sinh
2
0 ⎟
⎟
R = cosh I − sinh γ γ = ⎜ θ θ ⎟. (2.2.67)
2 2 ⎜ 0 − sinh cosh 0 ⎟
⎜ ⎟
⎝ 2 2 ⎠
θ θ
− sinh 0 0 cosh
2 2
The four-component function ψ satisfying (2.2.63) under the Lorentz transformation
(2.2.61) is called the Dirac spinor, which ensures the Lorentz covariance for the Dirac equa-
tions.
Noting that
1 1
cosh θ = (eiθ + e−iθ ), sinh θ = (eiθ − e−iθ ),
2 2
and cosh2 θ − sinh2 θ = 1, we infer from (2.2.67) that
R−1 = R† . (2.2.68)

2.3 Einstein’s Theory of General Relativity

2.3.1 Principle of general relativity
First Principle of Physics 2.1 amounts to saying that

Laws of Physics = Differential Equations

(2.3.1)
Universality of Laws = Covariance of Equations.

In retrospect, Albert Einstein must have followed the spirit of this principle for his dis-
covery of the general theory of relativity. As mentioned earlier, coordinate systems, also
called reference systems, are just an indispensable tool to express the laws of physics in the
form of differential equations. Consequently the validity of laws of physics is independent
of coordinate systems. Hence Einstein proposed the following principle of general relativity.

Principle 2.26(General Relativity) Laws of physics are the same under all coordinate
systems, both inertial and non-inertial. In other words, the models describing the laws of
physics are invariant under general coordinate transformations.

In Section 2.1.5, we mentioned that each symmetry in physics is characterized by three

ingredients:

space, transformation group, tensors.

The special theory of relativity or the Lorentz invariance is dictated by

2.3 Einstein’s Theory of General Relativity 65

Minkowski space, Lorentz group, and Lorentz tensors.

The three ingredients of the Theory of General Relativity are

• the space-time Riemannian space,

• general coordinate transformations, and

• general tensors,

under which the theory of general relativity is developed.

First, the Minkowski space is now replaced by the Riemannian space. In Theorem 2.18,
we see that the Minkowski metric

ds2 = −c2 dt 2 + (dx1 )2 + (dx2 )2 + (dx3 )2 (2.3.2)

is invariant under the Lorentz transformations. However, when we consider a non-inertial

reference system
t , x&1 , x&2 , x&3 ),
(c&
rotating with constant angular velocity Ω around the x3 -axis of an inertial system (ct, x1 , x2 , x3 ),
the coordinate transformation is given by

x1 = x&1 cos Ω&

t − x&2 sin Ω&
t,
x2 = x&1 sin Ω&
t + x&2 cos Ω&
t,
x3 = x&3 ,
t =&
t.

Under this transformation, the metric (2.3.2) becomes

ds2 = − [c2 − (&

x1 )2 Ω2 − (&
x2 )2 Ω2 ]d&
t 2 − 2Ω&
x2d&
x1 d&
t (2.3.3)
2 1 2 2 2 3 2
+ 2Ω&
x1d&
x d&
t + (d&
x ) − (d&
x ) + (d&
x ) .

Hence in a general non-inertial system, the metric ds2 takes the form

ds2 = g µν dxµ dxν , (2.3.4)

where {g µν } is a Riemannian metric different from the Minkowski metric given by (2.2.5).
In mathematics, the space M endowed with the metric (2.3.4), denoted by {M , g µν },
is called a Riemannian space, or a Riemannian manifold, and (2.3.4) or {g µν } is the Rie-
mannian metric.
In Theorem 2.12, we know that the Minkowski space M 4 is flat, and the Riemannian
space {M , g µν } is curved provided that the metric g µν is not the same as the Minkowski
metric in any coordinate systems.
66 Chapter 2 Fundamental Principles of Physics

2.3.2 Principle of equivalence

In the last subsection we see that the underlying space for the general theory of relativity is
the 4-dimensional Riemannian space {M , g µν }, instead of the Minkowski space. Now, a
crucial step is that we have to make sure the physical significance of the Riemannian metric
{g µν }.
Consider a free particle moving in a Riemannian space {M , g µν }, which satisfies the
motion equations ! k"
dx
D =0 for k = 1, 2, 3, (2.3.5)
ds
where D is the covariant derivative, and
3
ds = g µν dxµ dxν .

According to the theory of Riemannian geometry, the motion equations (2.3.5) in the Minkowski
space are in the form
d 2 xk
=0 for k = 1, 2, 3,
ds2
which are the motion equations of special relativity, and (2.3.5) in the Riemannian space
become
d 2 xk dxµ dxν
2
+ Γkµν =0 for k = 1, 2, 3. (2.3.6)
ds ds ds
where Γαµν is the Levi-Civita connection, given by
! "
1 ∂ g µβ ∂ gνβ ∂ g µν
Γαµν = gαβ + − , (2.3.7)
2 ∂ xν ∂ xµ ∂ xα

and (gαβ ) is the inverse of (gαβ ).

The equations (2.3.6) are the generalized Newtonian Second Law, the first term in the
left-hand side of (2.3.6) is the acceleration, and the second terms represent the force acting
on the particle. In view of (2.3.7) we see that the force is caused by the curvature of space-
time, i.e. by the non-flat metric
∂µ gαβ ̸= 0.
In addition, from this fact we can also think that the gravitation results in a non-flat
Riemann metric. However, in (2.3.3) we obviously see that the inertial force also lead to
the non-flat metrics. Thus, we encounter a difficulty that the curved Riemann space can be
caused by both gravitational and inertial forces.
Einstein proposed the principle of equivalence overcoming this difficulty.

Principle 2.27(Principle of Equivalence) One cannot distinguish the gravitational and

the inertial forces at any space-time point by experiments. In other words, any non-inertial
system can be equivalently regarded as an inertial system located in a gravitational field.
2.3 Einstein’s Theory of General Relativity 67

With the principle of equivalence, the Riemann metric is regarded as the effects caused
only by the gravitation. Furthermore, by the classical gravitational theory, the acting force
! "
∂ϕ ∂ϕ ∂ϕ
F = −∇ϕ = − , , , (2.3.8)
∂ x1 ∂ x2 ∂ x3
where ϕ is the gravitational potential. In comparing the force (2.3.8) with (2.3.7) which
contains first-order derivative terms ∂α gµν , the connections (2.3.7) also play a role of acting
forces in (2.3.6). Hence we need to regard the Riemann metric {g µν } as the gravitational
potential.
Thus, the principle of general relativity requires that the underlying space-time manifold
be a 4-dimensional Riemann space {M , g µν }, and the principle of equivalence defines the
Riemann metric g µν as the gravitational potential.
In conclusion, by Principles 2.26 and 2.27, we derive the following crucial physical
conclusion of the general theory of relativity, and the second crucial physical conclusion is
the Einstein field equations for the gravitational potential:
Physical Conclusion 2.28 (General Theory of Relativity) The physical space of our
Universe is a 4-dimensional Riemannian space {M , g µν }, and the Riemann metric g µν
represents the gravitational potential. The physical laws associated with gravitation are
covariant under the general coordinate transformations in {M , g µν }.

The principle of equivalence have received many supports by experiments. In fact, the
principle is based on the fact that the inertial mass is the same as the gravitational mass. By
the Newton second law,
F = am,
where m is called the inertial mass, denoted by mI , and the gravitational law provides
Gm1 m2
F= ,
r2
where m1 and m2 are called the gravitational masses, and denoted by mg .
Theoretically both masses mI and mg are different physical quantities, and we cannot a
prior claim that they are the same. It must be verified by experiments. Newton was the first
man to check them, resulting
mg
= 1 + o(10−3),
mI
i.e., mg = mI in the error of 10−3 . In 1890, Eötvös obtained the precision to 10−8, and in
1964, Dicke reached at 10−11 .

2.3.3 General tensors and covariant derivatives

In order to obtain the second crucial conclusion of general theory of relativity, i.e., the
Einstein gravitational field equations for the gravitational potential g µν , we introduce, in
this subsection, the concept of general tensors, general invariants, and covariant derivatives.
68 Chapter 2 Fundamental Principles of Physics

1. General tensors. Let {M , g µν } be an n-dimensional Riemannian space, and x =

(x1 , · · · , xn ) be a local coordinate system of M . We call the following transformation

x&k = ϕ k (x) for 1 ! k ! n, (2.3.9)

a general coordinate transformation if the functions ϕ k (x) (1 ! k ! n) in (2.3.9) satisfy that

the Jacobian ! i"
' i( ∂ϕ
aj = for x ∈ M (2.3.10)
∂xj
is continuous and non-degenerate. The inverse of (2.3.10) is denoted by

(bij ) = (aij )−1 . (2.3.11)

We remark that if the matrices in (2.3.10) and (2.3.11) are not continuous, then the
transformation (2.3.8) are not permitted.

Definition 2.29(General Tensors) Let T be a set of quantities defined on the Riemann

space {M , gi j }, and T have nr+s components in each coordinate system x ∈ M :

T = {T ji11···
···ir
js (x)} for x ∈ M .

If under the coordinate transformation (2.3.9), the components of T transform as

T&ji11···
···ir i1 ir l1 ls k1 ···kr
js = ak1 · · · akr b j1 · · · b js Tl1 ···ls , (2.3.12)

then T is called a (r, s) type of kth-order general tensor with k = r + s, where a ij and bm
l are
as in (2.3.10) and (2.3.11). The (r, 0) type tensors are called contra-variant tensors, and the
(0, s) type tensors are called covariant tensors.

On a Riemannian space {M , gi j }, the metric {gi j } and its inverse {gi j } are second-order
covariant and contra-variant symmetric tensors, i.e. they satisfy

gi j = g ji , gi j = g ji ,
(2.3.13)
g&i j = bli bkj glk , g&i j = ail akj glk ,

under the transformation (2.3.9). For first-order tensors, we have

⎛ ⎞ ⎛ ⎞ ⎛ 1 ⎞ ⎛ 1 ⎞
A&1 A1 &
A A
⎜ .. ⎟ i T ⎜ .. ⎟ ⎜ .. ⎟ i ⎜ .. ⎟
⎝ . ⎠ = (b j ) ⎝ . ⎠ , ⎝ . ⎠ = (a j ) ⎝ . ⎠ (2.3.14)
A&n An &n
A An

and for second-order tensors, we have

(T&i j ) = (bkl )T (Ti j )(bkl ), (T&i j ) = (akl )(T i j )(akl )T . (2.3.15)

2.3 Einstein’s Theory of General Relativity 69

2. General Invariants. General Invariants are derived from contractions of general

tensors. We infer from (2.3.14) that
⎛ ⎞ ⎛ 1 ⎞
&1
A A
& &k & & ⎜ .. ⎟ i i ⎜ .. ⎟
Ak A = (A1 , · · · , An ) ⎝ . ⎠ = (A1 , · · · , An )(b j )(a j ) ⎝ . ⎠
&n
A An
⎛ 1 ⎞
A
⎜ .. ⎟
= (A1 , · · · , An ) ⎝ . ⎠ = Ak Ak , by (bij ) = (aij )−1 ,
An

and by (2.3.15) we have

+ ,
T&i j T&i j = tr (T&i j )(T&i j )T
+ ,
= tr (bkl )T (Ti j )(bkl )(akl )(T i j )T (akl )T
+ ,
= tr (bkl )T (Ti j )(T i j )T (akl )T
= tr[(Ti j )(T i j )T ]
= Ti j T i j .

Here we have used the following property for matrices:

tr[ABA−1] = tr B, (2.3.16)

which can be shown by the fact that the eigenvalues λ 1 , · · · , λn of B are the same as those of
ABA−1 . Indeed we have

tr B = λ1 + · · · + λn = tr[ABA−1].

Moreover, all general invariants are in contraction form:

Alk11···l s k1 ···kr
···kr Bl1 ···ls , (2.3.17)

where the invariance holds true as well if the indices are exchanged. For example, A i j B ji is
also invariant.
By the invariant form (2.3.17), the metric form ds2 of a Riemannian space is a general
invariant, i.e.
ds2 = gi j dxi dx j = g&i j d&
xi d&
x j.

3. Covariant Derivatives. The covariance of differential equations requires that the

derivative operator ∇ be also covariant, i.e. ∇ is a general tensor operator.
70 Chapter 2 Fundamental Principles of Physics

Let us consider the usual derivatives

∂k = ∂ /∂ xk for 1 ! k ! n. (2.3.18)

For a vector field A = (A1 , · · · , Ak ), in the transformation (2.3.9) we have

&k = ak Al .
A (2.3.19)
l

Differentiating both sides of (2.3.19), we have

∂ A&k ∂ Al ∂ xi ∂ akl ∂ xi l ∂ Al ∂ a k
=akl i + i A = akl bij i + li bij Al , (2.3.20)
∂ x&j ∂ x ∂ x&j ∂ x ∂ x&j ∂x ∂x
where bij = ∂ xi /∂ x&j are as in (2.3.11). In view of Definition 2.29, we infer from (2.3.20)
@ iA
∂A
that the usual derivative operators (2.3.18) are not tensor operators. Namely, is not
∂xj
∂ akl i l
a tensor, and the transformation formula (2.3.20) contains an extra term: bA.
∂ xi j
To solve this problem, as shown in (2.1.31), we need to add a term Γ to the derivative
operator ∂k , resulting a new derivative operator ∇:

∂ Ak
∇ j Ak = + Γkij Ai ,
∂xj
such that ∇ = {∇ j } is a tensor operator. Namely, {∇ j Ak } is a (1,1) type tensor, and trans-
forms as ! l "
k ∂ A&k &k &i k i ∂A
&
∇ jA =
& + Γi j A = al b j + Γri A = akl bij ∇i Al .
l r
(2.3.21)
∂ x&j ∂ xi
By (2.3.20), it follows from (2.3.21) that
k
Γ &i = ak bij Γlri Ar − bij ∂ al Al .
&kij A (2.3.22)
l
∂ xi
By (2.3.22) we deduce the transformation rule for Γ as
k
&kij = ak bri bsj Γlrs − bri bsj ∂ ar .
Γ (2.3.23)
l
∂ xs
Fortunately, for a Riemannian space {M , gi j }, there exists a set of functions

Γ = {Γkij }, (2.3.24)

called the Levi-Civita connection, which satisfies the transformation given by (2.3.23), and
are given by ! "
1 ∂ gil ∂ g jl ∂ gi j
Γkij = gkl + − , (2.3.25)
2 ∂xj ∂ xi ∂ xl
2.3 Einstein’s Theory of General Relativity 71

which we have seen in (2.3.7).

Based on the connection (2.3.24)–(2.3.25), we now define the covariant derivatives in
the Riemannian space {M , gi j } as follows:

∂u
∇k u = for a scalar field u,
∂ xk
∂uj j
∇k u j = k + Γkl ul for a vector field {u j },
∂x
∂uj
∇k u j = k − Γlk j ul for a covector field {u j },
∂x

and for a (r, s) type general tensor field {uij11···i r

··· js },

∂ uij11···i r
··· js
∇k uij11···i r
··· js = − Γlk j1 uil1j···i r
2 ··· js
− · · · − Γlk js uij11···i r
··· js−1 l (2.3.26)
∂ xk
i ···i l
+ Γikl1 ulij12······ijsr + · · · + Γiklr u j11 ··· s−1
js .

Consequently, physical laws obeying Principle 2.26 of Einstein general relativity must
be in a form of covariant partial differential equations:

L(u, ∇u, · · · , ∇m u) = 0.

Now, what remains is to establish the field equations governing the Riemann metric
{gi j } (the gravitational potential), which will be introduced in the next two subsections.

2.3.4 Einstein-Hilbert action

In view of Principle 2.3 (PLD), to derive the gravitational field equations it suffices to derive
the Lagrange action for the gravitational potential {g µν }.
To this end, we first introduce the Ricci curvature tensor R µν and the scalar curvature R
in a Riemannian manifold.
1. Ricci tensor. It is natural to conjecture that the Lagrange density L for the field
equations depends on the terms ∂ g µν , i.e.

L = L (g µν , · · · , ∂ m gµν ). (2.3.27)

However, it is known that all covariant derivatives of the Riemann metric are zero:

∇g µν = 0, ∇g µν = 0. (2.3.28)

Hence, we are not able to directly use the terms ∇g µν , · · · , ∇m g to construct the density
(2.3.27), and have to look for invariants depending on ∂ g µν , · · · , ∂ m gµν (m $ 1) in a differ-
ent fashion.
72 Chapter 2 Fundamental Principles of Physics

The Ricci curvature tensor provides a natural way for us to find the Lagrange action.
For a covector field A = {Ak } we have

∂ β β
∇µ ∇ν Aα = (∇ν Aα ) − Γµν (∇β Aα ) − Γµα (∇ν Aβ ).
∂ xµ
Note that
∂ Aα γ
∇ν Aα = − Γνα Aγ .
∂ xν
Then we can deduce that
β
[∇µ ∇ν − ∇ν ∇µ ]Aα = Rα µν Aβ , (2.3.29)

where
β
β ∂ Γα µ ∂ Γβαν γ β γ β
Rα µν = − + Γα µ Γγν − Γαγ Γγ µ . (2.3.30)
∂ xν ∂ xµ
The tensor in (2.3.30) is called the curvature tensor, and its self-contraction given by

Rµν = Rαµαν (2.3.31)

is the Ricci tensor.

2. Scalar curvature. Again by contraction of the Ricci tensor with the metric tensor,
we derive an invariant, called scalar curvature:

R = g µν Rµν . (2.3.32)

3. Lagrangian action. In the Riemannian space {M , g µν }, the scalar curvature (2.3.32)

is a unique invariant which contains up to second-order derivatives of g µν . Hence it is
natural to choose the scalar curvature R as the main part of the Lagrange density. Namely,
L should be in the form
L = R + S, (2.3.33)
where S is the energy-momentum term of baryonic matter in the Universe. Physically, the
energy-momentum density term S is taken as
8π G αβ
S= g Sαβ , (2.3.34)
c4
where G is the gravitational constant, and Sαβ is the energy-momentum stress tensor. There-
fore we obtain the Lagrange action of gravitational fields:
! # $
8π G √
LEH = R + 4 gαβ Sαβ −gdx, (2.3.35)
M c
√
where g = det(gαβ ), and −gdx is the volume element.
2.3 Einstein’s Theory of General Relativity 73

The functional (2.3.35) is called the Einstein-Hilbert action or Einstein-Hilbert func-

tional. Historically, the functional was first introduced by David Hilbert in 1915 after he
listened to the lecture on the general theory of relativity by Einstein. In fact, it is not easy to
determine the expression of the variational derivative operator δ LR for the functional
!
√
LR = R −gdx. (2.3.36)
M

2.3.5 Einstein gravitational field equations

The gravitational field equations based on PLD are the variational equations of the Einstein-
Hilbert action LEH in (2.3.35):
δ LEH = 0. (2.3.37)
By (2.3.35), the equation (2.3.37) can be explicitly expressed as
1 8π G
Rµν − gµν R = − 4 Tµν , (2.3.38)
2 c
where R µν is the Ricci tensor, R the scalar curvature, and Tµν is the energy-momentum
tensor. The equations (2.3.38) are the well-known Einstein gravitational field equations.

Remark 2.30 The terms in (2.3.38) are

1
Rµν − gµν R = δ LR , LR as in (2.3.36),
2 !
√
Tµν = δ Ls , Ls = gµν Sµν −gdx.
M

In Section 3.3, we shall give the derivation of δ LR .

Einstein derived the field equations (2.3.38) in 1915 using his great physical intuition.
According to the classical gravity theory, the Newton potential ϕ satisfies the Laplace equa-
tion
∆ϕ = 4π Gρ , (2.3.39)
where ρ is the mass density. Einstein thinks the gravitational field equations of general
relativity should be in the form

Gµν = β Tµν , β is a constant, (2.3.40)

where Tµν represents the energy-momentum tensor, which are provided by physical obser-
vations, and G µν represents gravitational potential which are the generalization of ∆ϕ in
(2.3.39). Hence G µν contains the derivatives of g µν up to the second order. In Riemannian
geometry, only the curvature tensors satisfy the needed properties. Thus G µν must be in the
form
Gµν = Rµν + λ1gµν R + λ2gµν .
74 Chapter 2 Fundamental Principles of Physics

where λ1 , λ2 are constants.

By the conservation of energy-momentum:

∇µ Tµν = 0,

the second-order tensor G µν should be divergence-free, i.e.

∇µ Gµν = 0. (2.3.41)

By the Bianchi identity, G µν satisfying (2.3.41) are uniquely determined in the form up to
a constant λ ,
1
Gµν = Rµν − gµν R + λ g µν , (2.3.42)
2
where λ g µν are divergence-free due to (2.3.28). Furthermore, Einstein determined the con-
stant β in (2.3.40) by comparing with (2.3.39), and β is given by

8π G
β =− .
c4
Thus, by (2.3.40) and (2.4.42), Einstein deduced the field equations in the general form as
follows
1 8π G
Rµν − gµν R + λ g µν = − 4 Tµν , (2.3.43)
2 c
and the constant λ is usually called the cosmological constant. Today, a large number of
physical experiments manifest that λ = 0. However, the recent discovery of the acceleration
of our universe leads to some physicists to think that λ ̸= 0.
In Chapter 7, based on the unified field theory developed by the authors, we present a
theory of dark matter and dark energy, which clearly explains the phenomena of dark matter
and dark energy, and shows that the constant λ should be zero, i.e. λ = 0.

2.4 Gauge Invariance

2.4.1 U(1) gauge invariance of electromagnetism

Gauge symmetry is one of fundamental invariance principles of physics, and determines the
Lagrange actions of the electromagnetic, the weak and the strong interactions. In order to
show the origin of gauge theory, we first introduce the U(1) gauge invariance of electro-
magnetic fields.
Fermions under an electromagnetic field with potential A µ obey the Dirac equations:
mc
iγ µ Dµ ψ − ψ = 0, (2.4.1)
h̄
Dµ = (∂µ + ieAµ ), (2.4.2)
2.4 Gauge Invariance 75

where the electromagnetic potential A µ satisfies the Maxwell equations (2.2.40):

4π µ
∂ν F µν = J , (2.4.3)
c
! "
µα νβ ∂ Aα ∂ Aβ
F µν =g g − α . (2.4.4)
∂ xβ ∂x

It is easy to see that the system of equations (2.4.1)-(2.4.4) is invariant under the follow-
ing transformation:
& = eiθ ψ ,
ψ A&µ = Aµ − 1 ∂µ θ . (2.4.5)
e
Since eiθ ∈ U(1), (2.4.5) is called a U(1)-gauge transformation.
Now, we consider the gauge invariance from another point of view. Let ψ be a Dirac
spinor describing a fermion:
ψ : M 4 → M 4 ⊗ p C4 .

Experimentally, we cannot observe the phase angles of ψ . Namely, under a phase rotation
transformation
& = eiθ ψ , θ = θ (xµ ),
ψ (2.4.6)

we cannot distinguish the two states ψ

& and ψ experimentally. Mathematically speaking, the
phenomenon amounts to saying that the Dirac equations are covariant under the transfor-
mation (2.4.6). This covariance requires that the derivative be covariant:

D & = eiθ Dµ ψ .
&µ ψ (2.4.7)

However, we note that

& = ∂µ (eiθ ψ ) = eiθ (∂µ + i∂µ θ )ψ ,

∂µ ψ (2.4.8)

indicating that ∂ µ is not covariant as defined by (2.4.7). In view of (2.4.8), to obtain a

covariant derivative, we have to add a term G µ to ∂µ . Namely, we need to define D µ by

Dµ = ∂µ + igGµ with g being a coupling constant, (2.4.9)

where Gµ is a 4-dimensional vector field and for (2.4.6) G µ transforms as

&µ = G µ − 1 ∂ µ θ .
G (2.4.10)
g

Thus, it is readily to check that

&µ ψ
D & µ )(eiθ ψ ) = eiθ (∂µ + igGµ )ψ = eiθ Dµ ψ .
& =(∂µ + igG
76 Chapter 2 Fundamental Principles of Physics

Hence, the derivative operator defined by (2.4.9) is covariant under the gauge transformation
(2.4.6) and (2.4.10):
& = eiθ ψ ,
ψ G& = Gµ − 1 ∂µ θ . (2.4.11)
g
In view of (2.4.11) and (2.4.5), the field G µ and the coupling constant g in (2.4.9) are
the electromagnetic potential A µ and electric charge e:
Gµ = Aµ the electromagnetic potential,
g=e the electric charge.
The above process illustrates that the electromagnetic potential A µ and the electric
charge e are the outcomes from the U(1) gauge invariance. In other words, from the phase
angle symmetry of particles (U(1) gauge symmetry) we deduce the following conclusions:
U(1) gauge symmetry ⇒ U(1) gauge field A µ ,
U(1) gauge field A µ = electromagnetic potential, (2.4.12)
U(1) gauge coupling constant g = electric charge.
Recall that in the Einstein general theory of relativity, the principle of general relativity
not only leads to the equivalence of gravitational potential to the Riemannian metric g µν ,
but also determines the field action—the Einstein-Hilbert action (2.3.35). Now, by the U(1)
gauge invariance, we are able to deduce both the electromagnetic potential A µ as (2.4.12)
and the U(1) gauge action as follows.
In fact, in the same fashion as used in (2.3.29) for gravity, by the covariance of (2.4.7)
we derive the covariant field Fµν as
4 5
Dµ Dν − Dν Dµ ψ = ieFµν ψ , (2.4.13)

where Fµν is given by (2.4.4). The left-hand side of (2.4.13) is covariant under the gauge
transformation (2.4.5), and so does Fµν . Consequently the contraction

F = Fµν F µν , (2.4.14)

is invariant under both the gauge and the Lorentz transformations.

Thus, the invariant F in (2.4.14) is the main part of the action of the U(1) gauge field
Aµ , which is given by
! # $
1 4π
LEM = − F+ Aµ J µ dxdt. (2.4.15)
M4 4 c
This is the action of the Maxwell field.
The form of (2.4.15) is similar to that of the Einstein-Hilbert action (2.3.35). In math-
ematics, the gauge field A µ in (2.4.2) is the connection of the complex vector bundle
M 4 ⊗ p C4 , with which the Dirac spinor ψ is defined:

ψ : M 4 → M 4 ⊗ p C4 .
2.4 Gauge Invariance 77

The tensor Fµν in (2.4.13) is the curvature tensor of the bundle M 4 ⊗ p C4 , and F = Fµν F µν
in (2.4.15) is the scalar curvature. In addition, the current energy A µ J µ in (2.4.15) corre-
sponds to the energy-momentum g µν Sµν in (2.3.35).
In the same fashion as the electromagnetism, we can derive the gauge theories for both
the weak and the strong interactions from the SU(2) and SU(3) gauge invariances. In the
next subsections we shall introduce the mathematical theory of SU(N) gauge fields and the
principle of gauge invariance in physics.

2.4.2 Generator representations of SU(N)

Both the weak and strong interactions are described by the SU(N) gauge theory, which is a
generalization of the U(1) gauge theory introduced in the last subsection.
In an SU(N) gauge theory, there are N wave functions, representing N fermions:

Ψ = (ψ 1 , · · · , ψ N )T ,

where each ψ k (1 ! k ! N) is a 4-component Dirac spinor. We seek for a set of gauge fields
Gkµ such that the Dirac equations
+ cm ,
iγ µ Dµ − Ψ=0 (2.4.16)
h̄
are invariant under the SU(N) gauge transformation
k (x)τ
Ψ
& = ΩΨ, ∀Ω = eiθ k ∈ SU(N), (2.4.17)

where m is the mass matrix, and θ k are real parameters,

Dµ = ∂µ + igGkµ τk , (2.4.18)

g is a coupling constant, τk (1 ! k ! N 2 − 1) are the generators of SU(N).

We recall that SU(N) is the group consisting of N × N unitary matrices with unit deter-
minant:

SU(N) = {Ω| Ω the N × N matrix, Ω† = Ω−1 , det Ω = 1}. (2.4.19)

In the SU(N) gauge theory for (2.4.16)-(2.4.19), we encounter a mathematical concept,

generators τk (1 ! k ! N 2 − 1) of SU(N). We now introduce the representation of SU(N).
By (2.4.19) we see that each matrix Ω ∈ SU(N) satisfies

Ω† = Ω−1 , det Ω = 1, Ω ∈ SU(N), (2.4.20)

where Ω† = (ΩT )∗ is the complex conjugate of the transport of Ω. Note that an exponent
eiA of a matrix iA is an N × N complex matrix satisfying
†
(eiA )† = e−iA .
78 Chapter 2 Fundamental Principles of Physics

It follows that
† †
(eiA )† (eiA ) = e−iA eiA = ei(A−A ) .
Hence we see that

(eiA )† = (eiA )−1 if and only if A = A† is Hermitian. (2.4.21)

In addition, we have
det eA = etrA , (2.4.22)
which will be proved at the end of this subsection.
Based on (2.4.20)-(2.4.22), it is not difficult to understand the representation of SU(N)
in a neighborhood of the unit matrix, stated in the following theorem.
Theorem 2.31 (SU(N) Representation) The matrices Ω in a neighborhood U ⊂
SU(N) of the unit matrix can be expressed as

Ω = eiA , A† = A, trA = 0. (2.4.23)

Furthermore, the tangent space Te SU(N) of SU(N) at the unit matrix e = I is a K-dimensional
linear space (K = N 2 − 1), generated by K linear independent traceless Hermitian matrices
τk (1 ! k ! K):

Te SU(N) = span{τ1 , · · · , τK } with τk† = τk , tr τk = 0. (2.4.24)

In particular, the matrices Ω in (2.4.23) can be written as

kτ
Ω = eiθ k , τk (1 ! k ! K) as in (2.4.24), (2.4.25)

where θ k (1 ! k ! K) are real numbers, and

[τk , τl ] = τk τl − τl τk = iλklj τ j , (2.4.26)

where λklj are called the structure constants of SU(N).

We remark here that the basis {τk | 1 ! k ! K} of the tangent space (2.4.24) is usually
called the generators of SU(N), and (2.4.25) is called the generator representation of SU(N).
We are now in position to give a proof of (2.4.22). By the classical Jordan theorem, for
a given matrix A there is a non-degenerate matrix B such that
⎛ ⎞
λ1 ∗
BAB−1 = ⎝
⎜ .. ⎟
⎠, (2.4.27)
.
0 λN

is an upper triangular matrix, where λ1 , · · · , λN are the eigenvalues of A. Recall that eA is

defined by
∞
Ak
eA = ∑ .
k=0 k!
2.4 Gauge Invariance 79

By (2.4.27) we have
⎛ ⎞k ⎛ ⎞
∞ ∞
λ1 ∗ eλ 1 ∗
1 1⎜ .. ⎟ ⎜ .. ⎟
BeA B−1 = ∑ k! (BAB−1)k = ∑ k! ⎝ . ⎠ =⎝ . ⎠.
k=0 k=0
0 λN 0 eλN

Hence we get
det eA = det (BeA B−1 ) = eλ1 +···+λN . (2.4.28)
It is known that
tr A = λ1 + · · · + λN .
Thus, the formula (2.4.22) follows from (2.4.28).

2.4.3 Yang-Mills action of SU(N) gauge fields

An SU(N) gauge theory mainly deals with the invariance of the Dirac equations for N
spinors under SU(N) gauge transformations. The physical meaning of SU(N) gauge invari-
ance is that in a system of N fermions we cannot distinguish one particle from others by the
weak or strong interaction.
Based on the SU(N) representation Theorem 2.31, we now introduce the SU(N) gauge
theory.
Consider N Dirac spinors ψ k (1 ! k ! N) and K = (N 2 − 1) Lorentz vector fields
Gaµ (1 ! a ! K), called the gauge fields, given by

Ψ = (ψ 1 , · · · , ψ N )T , Gµ = (G1µ , · · · , GKµ ). (2.4.29)

The Dirac equations for the fermions are:

+ mc ,
iγ µ Dµ − Ψ = 0, (2.4.30)
h̄
where
⎛ ⎞
m1 0
⎜
m=⎝ .. ⎟
is the the mass matrix, (2.4.31)
. ⎠
0 mN
a
Dµ = ∂µ + igGµ τa , (2.4.32)

and τa (1 ! a ! K) are a set of SU(N) generators as in Theorem 2.31.

For the N spinors Ψ in (2.4.29), consider the following SU(N) gauge transformations
aτ
Ψ
& = ΩΦ ∀Ω = eiθ a
∈ SU(N), (2.4.33)

where θ a (1 ! a ! K) are functions of x µ ∈ M 4 .

80 Chapter 2 Fundamental Principles of Physics

To ensure that (2.4.30)-(2.4.32) are covariant under the transformation (2.4.33), we need
to determine
(1) the transformation form of the gauge fields G µ , and
(2) the action for the gauge fields G µ .
First, consider (1). The covariance of Dirac equations (2.4.30) is equivalent to the co-
variance of the derivative D µ Ψ as given by
&µ Ψ
D & = ΩDµ Ψ ∀Ω as in (2.4.33). (2.4.34)

The left-hand side of (2.4.34) can be directly computed as

&aµ τa )ΩΨ = Ω∂µ Ψ + (∂µ Ω)Ψ + igG
& =(∂µ + igG
&µ Ψ
D &aµ τa ΩΨ,

and the right-hand side is

ΩDµ Ψ =Ω(∂µ + igGaµ τa )Ψ = Ω∂µ Ψ + igGaµ Ωτa Ψ.

Hence, it follows from (2.4.34) that

&a τa = Gaµ Ωτa Ω−1 + i (∂µ Ω)Ω−1 ,

G (2.4.35)
g
which is called the SU(N) gauge transformation of the gauge fields G aµ .
In addition, the mass matrix (2.4.31) satisfies that

& = ΩmΩ−1 .
m (2.4.36)

In summary, the transformations (2.4.33), (2.4.35) and (2.4.36) constitute the SU(N)
gauge transformations defined as
aτ
Ψ
& = ΩΨ ∀Ω = eiθ a
∈ SU(N),
&aµ τa = Gaµ Ωτa Ω−1 + i (∂µ Ω)Ω−1 ,
G (2.4.37)
g
& = ΩmΩ−1 .
m
Usually, (2.4.37) is taken as an infinitesimal transformation as follows

Ω = I + iθ aτa , θ a are infinitesimal.

In this case, (2.4.37) can be written as

& = Ψ + iθ aτa Ψ,
Ψ
&aµ = Gaµ − λ a θ b Gcµ − 1 ∂µ θ a ,
G (2.4.38)
bc
g
a a
& = m + iθ (τa m − mτa),
m
2.4 Gauge Invariance 81

a are the structure constants as defined in (2.4.26).

where λbc
We have seen that the Dirac equations (2.4.30)-(2.4.32) are covariant under the SU(N)
gauge transformations (2.4.37) or (2.4.38).
What remains to do is to find out the action of the gauge fields Gaµ , which is invariant
under both the Lorentz transformation and the SU(N) gauge transformation (2.4.37).
Recall that in (2.4.13), the U(1) gauge field action (2.4.15) is derived by using the
commutator of the covariant derivative operator:
4 5
D µ , Dν = D µ Dν − D ν D µ .

We now derive in the same fashion the SU(N) action, called the Yang-Mills action.
By (2.4.34),
& ν Ψ)
& µ (D
D & =D& µ (ΩDν Ψ) = Ω(Dµ Dν Ψ).

It follows that + , 4 5
D &ν Ψ
&µ , D & = Ω Dµ , Dν Ψ. (2.4.39)

On the other hand, by (2.4.32), we have

i4 5 i i
Dµ , Dν = (∂µ + igGaµ τa )(∂ν + igGaν τa ) − (∂ν + igGaν τa )(∂µ + igGaµ τa ). (2.4.40)
g g g

Notice that
∂µ ∂ν = ∂ν ∂µ , ∂ν (Gaµ τa ) = ∂ν Gaβ τa + Gaβ τa ∂ν .

Then (2.4.40) becomes

i4 5 + ,
Dµ , Dν =(∂µ Gaν − ∂ν Gaµ )τa − ig Gaµ τa , Gbν τb (2.4.41)
g
=(∂µ Gaν − ∂ν Gaµ + gλbc
a b c
Gµ Gν )τa (by (2.4.26))

Therefore we define
a
Fµν = ∂µ Gaν − ∂ν Gaµ + gλbc
a b c
G µ Gν . (2.4.42)

Then, thanks to (2.4.41), we deduce that

i + & & , & &a
Dµ , Dν Ψ =Fµν τa ΩΨ
g
i 4 5
= Ω Dµ , Dν Ψ (by (2.4.39))
g
a
=Fµν Ωτa Ψ,

which implies that

a a
F&µν τa = Fµν Ωτa Ω−1 . (2.4.43)
82 Chapter 2 Fundamental Principles of Physics

As µ , ν are the indices of 4-D tensors, a Lorentz invariant can be constructed using the
following contraction
a
F&µν τa F&µν b τb† = F&µν
a &µν b
F τa τb† , (2.4.44)
where F µν b = g µα gνβ Fαβ . By (2.4.43) we have
a
F&µν τa F&µν b τb† =Fµν
a
(Ωτa Ω−1 )F µν b (Ωτb Ω−1 )† (2.4.45)
a
=Fµν F µν b Ωτa τb† Ω−1 †
(by Ω = Ω ).
−1

Therefore we deduce from (2.4.44) and (2.4.45) that

a &µν b
F&µν F τa τb† = Fµν
a
F µν b Ωτa τb† Ω−1 . (2.4.46)

In (2.3.16) we have verified that

tr A = tr (BAB−1 ), ∀ matrices A and B.

Hence it follows from (2.4.46) that

a &µν b
Gab F&µν F a
= Gab Fµν F µν b , (2.4.47)

1
where Gab = tr(τa τb† ).
2
The equality (2.4.47) shows that
a
F = Gab Fµν F µν b = Gab g µα gνβ Fµν
a b
Fαβ (2.4.48)

is invariant under both the Lorentz transformations and the SU(N) gauge transformations,
a
where Fµν are given by (2.4.42).
The function (2.4.48) is a unique form which is both Lorentz and the SU(N) gauge
invariant, and contains up to first-order derivatives of the gauge fields G aµ . Hence, F in
(2.4.48) is a unique candidate to be the Lagrange density.
In Section 3.5, we have shown that {Gab} in (2.4.48) is a Riemannian metric of SU(N),
and
1 1 c d
Gab = tr(τa τb† ) = λ λ , (2.4.49)
2 4N ad cb
where λab c are the structure constants of SU(N).

In the classical SU(N) gauge theory, the SU(N) generator τk (1 ! k ! K) are taken to
be Hermitian and traceless, and satisfy
1
tr (τa τb† ) = δab .
2
In this case, the Lagrange density F in (2.4.48) becomes
a
F = Fµν F µν a .
2.4 Gauge Invariance 83

The Lagrange action of an SU(N) gauge theory is usually taken in the following form,
called the Yang-Mills action:
! # ) $
1 a µν a µ cm *
LY M = − Fµν F + Ψ iγ Dµ − Ψ dx, (2.4.50)
M4 4 h̄

where Ψ = Ψ† γ 0 ,
a
Fµν = ∂µ Gaν − ∂ν Gaµ + gλbc
a b c
G µ Gν ,
(2.4.51)
Dµ = ∂µ + igGaµ τa .
The second term on the right-hand side of (2.4.50) is the action for the Dirac equations
(2.4.30)-(2.4.32).

2.4.4 Principle of gauge invariance

In Sections 2.4.2-2.4.3, we introduced the mathematical framework of SU(N) gauge fields,
leading to the following principle of gauge invariance.
Principle 2.32 (Gauge Invariance) The electromagnetic, the weak, and the strong in-
teractions obey gauge invariance. Namely, the motion equations involved in the three inter-
actions are gauge covariant and the actions of the interaction fields are gauge invariant.

A few remarks are now in order.

Remark 2.33 The Standard Model in particle physics is currently a prevailing theory
describing all, except the gravity, fundamental interactions. It consists of the Glashow-
Weinberg-Salam (GWS) electroweak theory, the transition theory of weak interaction decay,
the quark model, and the Quantum Chromodynamics (QCD). Based on the Standard Model,
the strong interaction is described by an SU(3) gauge theory, and the electromagnetic and
weak interactions are unified in an action of U(1) × SU(2) gauge fields, combing with the
Higgs mechanism and the Yukawa coupling.
Remark 2.34 All up-to-date experiments illustrate that the electromagnetic, weak,
strong (EWS) interactions obey Principle of Gauge Invariance 2.32.
Remark 2.35 All current theories about interactions, concluding the Standard Model
and the String Theory, have a remarkable drawback: they cannot provide a set of acceptable
field equations for the weak and strong interactions from which we can deduce the weak and
strong interaction potentials. It is known that, the variational equations of the Yang-Mills
action (2.4.50)-(2.4.51) are in the form

∂ µ Fµν
a a αβ b
− gλbc g Fαν Gcβ − gJνa = 0,
) cm * (2.4.52)
iγ µ Dµ − Ψ = 0,
h̄
where
Jνa = Ψγν τa Ψ, γν = gνβ γ β .
84 Chapter 2 Fundamental Principles of Physics

However, it appears that we are not able to derive from (2.4.52) any weak and strong in-
teraction potentials in agreement with experiments. An important reason is that the SU(N)
gauge theory has N 2 − 1 fields:

G1µ , · · · , Gkµ for K = N 2 − 1, (2.4.53)

and we don’t know which or what combination of the potentials in (2.4.53) gives rise to an
interaction force formula.

Remark 2.36 The authors have developed a new unified field theory for the four fun-
damental interactions recently , based only on the following fundamental principles: the
Einstein Principle of General Relativity, the Principle of Lorentz Invariance, the Principle
of Gauge Invariance, and the two newly developed principles by the authors: the Principle
of Interaction Dynamics (PID) and the Principle of Representation Invariance (PRI). The
unified field theory will be presented in detail in Chapter 4, and we shall see that the theory
provides sound explanations and resolutions to the following problems:

1) dark matter and the dark energy,

2) spontaneous symmetry breaking based on first principles,

3) quark confinement,

4) asymptotically freedom,

5) strong interaction potential of nucleons, and

6) layered formula for the weak and strong interaction potentials.

2.5 Principle of Lagrangian Dynamics (PLD)

2.5.1 Introduction
PLD has been briefly introduced in Subsection 2.1.3, and we now address PLD in detail.
In classical mechanics, a physical motion system can be described by three dynamical
principles: the Newtonian Dynamics, the Lagrangian Dynamics, and the Hamiltonian Dy-
namics. Both the Lagrangian dynamics and the Hamiltonian dynamics remain valid in other
physical fields such as the electrodynamics and quantum physics.
The three dynamical principles are equivalent in describing the motion of an N-body
system. Consider an N-body system of planets, with masses m1 , · · · , mN and coordinates
xk = (x1k , x2k , x3k ).

1. Newtonian dynamics. The motion equations governing the N planets are the Newto-
nian second law:
d 2 xk
mk 2 = Fk for 1 ! k ! N, (2.5.1)
dt
2.5 Principle of Lagrangian Dynamics (PLD) 85

where Fk is the gravitational force acting on the k-th planet by the other planets, which can
be expressed as
mk m j G
Fk = − ∑ 3
(x j − xk ), (2.5.2)
j̸=k j − xk |
|x
where G is the gravitational constant.
2. Lagrangian dynamics. Based on the least action principle, the Lagrange density L
of the N-body system is
L = T − V,
where T is the total kinetic energy, and V is the potential energy:
! "
1 N dxk 2 mi m j G
T = ∑ mk , V =− ∑ , (2.5.3)
2 k=1 dt i, j=1,i̸= j 2|x i − x j|

Hence, the Lagrange action is

! t1 ! t1
B C
N
1 mi m j G
L=
t0
L dt =
2 t0
∑ mk ẋ2k + ∑ |x − x j | dt. (2.5.4)
k=1 i̸= j i

The variational derivative operator δ L of L is

mk m j G
δ L = −mk ẍk − ∑ (x − xk ),
3 j
(2.5.5)
j̸=k j − xk |
|x
and the motion equations of the Lagrangian dynamics are given by
δ L = 0. (2.5.6)
It is clear that the motion equations (2.5.6) derived from the Lagrangian dynamics is the
same as the equations (2.5.1)-(2.5.2) of the Newtonian dynamics.
3. Hamiltonian dynamics. In the next section we shall introduce the principle of
Hamiltonian dynamics (PHD), and we derive here the Hamiltonian system for the N-body
motion.
The total energy H of the N-body system is given by
N
1
H(x, y) = ∑ 2mk y2k + V (x), (2.5.7)
k=1
where yk is the momentum of the k-th planet, and V (x) is the potential energy as in (2.5.3).
Then the motion equations derived from the PHD are as follows
c∂ x k ∂H
= ,
∂t ∂ yk
(2.5.8)
∂ yk ∂H
=− .
∂t ∂ xk
where H(x, y) is defined by (2.5.7). It is easy to check that the equations (2.5.8) are equiva-
lent to those in (2.5.1)-(2.5.2).
86 Chapter 2 Fundamental Principles of Physics

2.5.2 Elastic waves

In an elastic continuous medium, the wave vibration is described by the PLD.

Let Ω ⊂ R3 be a domain of elastic continuous medium, and the function u(x,t) represent
the displacement of the medium at time t and x ∈ Ω. Then the total kinetic energy is
! % %
1 %% ∂ u %%2
T= ρ % % dx, (2.5.9)
Ω2 ∂t

where ρ is the density of the medium.

For an elastic material, the deformation potential energy V is taken in the general form
! # $
1 2
V= k|∇u| + F(x, u) dx, (2.5.10)
Ω 2

where k is a constant.
The Lagrange action for the elastic wave is written as
! t1 ! t1 ! # $
1 2 1 2
L= (T − V )dt = ρ u̇ − k|∇u| − F(x, u) dxdt. (2.5.11)
t0 t0 Ω 2 2

By the PLD, the wave equation is derived by

δ L = 0,

and it follows from (2.5.11) that

∂ 2u
ρ − k∆u = f (x, u) for x ∈ Ω. (2.5.12)
∂ t2
where
∂ F(x, z)
f (x, z) = .
∂z
The equation (2.5.12) is the usual wave equation describing elastic vibration in a continuous
medium.

2.5.3 Classical electrodynamics

Electrodynamics consists of two parts: the Maxwell field equations and the motion of
charged particles, each described by the related Lagrange actions.

Lagrange density of electromagnetic fields

In classical Maxwell theory, we usually take the electric field E, the magnetic field H,
the current density J⃗ and the electric charge density ρ as state functions, because these
2.5 Principle of Lagrangian Dynamics (PLD) 87

physical quantities are observable. However, the fields E and H are not fundamental physi-
cal quantities, and the basic fields describing electromagnetism are the 4-D electromagnetic
potential A µ and the current density Jµ :

Aµ = (A0 , A1 , A2 , A3 ),
(2.5.13)
Jµ = (J0 , J1 , J2 , J3 ).

Hence the Lagrange density for the Maxwell theory should be constructed with the fields in
(2.5.13).
Based on the Lorentz Invariance and the U(1) Gauge Invariance, the action of (2.5.13)
is uniquely determined and is given in the form (2.4.15):
! T!
L= L (Aµ , Jµ )dxdt, (2.5.14)
0 Ω

where
1 1
L = Fµν F µν + Aµ J µ Fµν as in (2.4.4). (2.5.15)
16π c
We now derive the variational equation, also called the Euler-Lagrange equation, of
(2.5.14)-(2.5.15). Since Jµ is an applied external field, we only take variation with respect
to the field A µ .
It is known that δ L is a 4-dimensional field

δ L = (δ L0 , δ L1 , δ L2 , δ L3 ),

&µ with A
and satisfies that for any A &µ |∂ Q = 0, we have
T

!
d %%
(δ L)µ A
&µ dxdt = % L(Aµ + λ A
&µ ), (2.5.16)
QT d λ λ =0

where QT = Ω × (0, T ) and λ is a real parameter. By (2.5.14) we see that

L = L1 + L2 ,
!
1 µ a νβ
L1 = g g Fµν Fαβ dxdt, Fµν = ∂ν Aµ − ∂µ Aν ,
QT 16 π
!
1
L2 = Aµ J µ dxdt.
QT c

It is clear that
! !
d %% &µ ) = 1 d &µ )J µ dxdt = 1
% L2 (Aµ + λ A (Aµ + λ A Jµ A
&µ dxdt.
d λ λ =0 c QT d λ c QT

We infer from (2.5.16) that

1
δ L2µ = J µ . (2.5.17)
c
88 Chapter 2 Fundamental Principles of Physics

Noting that g µν = gν µ , we have

!
d %% 1 d %%
% L1 (Aµ + λ Aµ ) =
& gµα gνβ Fαβ % (Fµν + λ F&µν )dxdt
d λ λ =0 8π QT d λ λ =0
!
6 7
1 µα νβ ∂ A&µ ∂ A
&ν
= g g Fαβ − µ dxdt
8π QT ∂ xν ∂x

By the Gauss formula,

! !
∂A&µ ∂ Fαβ
g µα gνβ Fαβ ν
dxdt = − gµα gνβ &µ dxdt,
A
QT ∂x QT ∂ xν
! !
∂ A&ν ∂ Fαβ
gµα gνβ Fαβ µ dxdt = − gµα gνβ &ν dxdt
A
QT ∂x QT ∂ xµ
=(by the permutation of µ and ν )
!
∂ Fαβ
=− gνα gµβ &µ dxdt
A
QT ∂ xν
!
∂ Fαβ
= g µα gνβ &µ dxdt,
A
QT ∂ xν
where a permutation on α and β is performed, and Fβ α = −Fαβ .
Thus, we obtain that
!
d %% &µ ) = − 1
∂ Fαβ
% L1 (Aµ + λ A gµα gνβ &µ dxdt.
A
d λ λ =0 4π QT ∂ xν
We infer then from (2.5.16) that

1 µα νβ ∂ Fαβ 1 ∂ F µν
δ L1 = − g g = − . (2.5.18)
4π ∂ xν 4 π ∂ xν
Hence it follows from (2.5.17) and (2.5.18) that
1 ∂ F µν 1 µ
δL = − + J ,
4 π ∂ xν c
which implies that the equation δ L = 0 takes the form:
∂ F µν 4π µ
ν
= J . (2.5.19)
∂x c
This is the second pair of the Maxwell equations (2.2.35) and (2.2.36). Since the first pair
of the Maxwell equations (2.2.33) and (2.2.34) are direct consequence of (2.4.4) and

1 ∂ ⃗A
H = curl ⃗A, E = ∇A0 − , ⃗A = (A1 , A2 , A3 ), (2.5.20)
c ∂t
the action (2.5.14)-(2.5.15) completely determines the Maxwell equations (2.2.33)-(2.2.36).
2.5 Principle of Lagrangian Dynamics (PLD) 89

Lagrangian for dynamics of charged particles

A particle with an electric charge e moving in an external electromagnetic field (E, H)

is governed by
dp e
= eE + v × H, (2.5.21)
dt c
e
where p is the momentum of the particle, v is the velocity, and f = v × H is the Lorentz
c
force.
The action for the motion equation (2.5.21) is taken as
! T
e
L= −mcds + Aµ dxµ . (2.5.22)
0 c

It is clear that (2.5.22) is Lorentz invariant. By

?
ds = 1 − v2/c2 cdt, dxµ = (1, v1 , v2 , v3 )cdt,

the action (2.5.22) can be written as

! T
L= L (v, Aµ )dt,
0 (2.5.23)
?
e
L = −mc 1 − v2/c2 + Ak vk + eA0 .
2
c
All physical properties of electromagnetic kinematics can be derived from the action
(2.5.23).
We shall deduce (2.5.21) from (2.5.23). The Euler-Lagrange equation of (2.5.23) is
given by
! "
d ∂L ∂L
= ∇L , ∇L = for x ∈ R3 . (2.5.24)
dt ∂ v ∂x
By (2.5.23), we have
∂L mv e
=3 + ⃗A.
∂v 2
1 − v /c 2 c
Then (2.5.24) becomes

dP e d⃗A
+ = ∇L ,
dt c dt (2.5.25)
mv
P= 3 is the momentum.
1 − v2/c2

By (2.5.23), we have
e
∇L = ∇⃗A · v + e∇A0.
c
90 Chapter 2 Fundamental Principles of Physics

Thanks to
∇⃗A · v =∇(⃗A · v) (by ∂ v = 0)
=(⃗A · ∇)v + (v · ∇)⃗A + v × curl ⃗A + ⃗Acurl v
=(v · ∇)⃗A + v × curl ⃗A,
the equation (2.5.25) reads
dP e d⃗A e e
+ = (v · ∇)⃗A + v × curl ⃗A + e∇A0 . (2.5.26)
dt c dt c c
It is known that
d⃗A ∂ ⃗A ∂ ⃗A dxk ∂ ⃗A
= + k = + (v · ∇)⃗A.
dt ∂ t ∂ x dt ∂t
Hence (2.5.26) is rewritten as
dP e ∂ ⃗A e
=− + e∇A0 + v × curl ⃗A. (2.5.27)
dt c ∂t c
Then, by (2.5.20), the equation (2.5.27) takes the form
dP e
= eE + v × curl ⃗A.
dt c
Thus, we have deduced (2.5.21) from the action (2.5.23).
Now, we deduce the Einstein energy-momentum formula for the 4-dimensional energy-
momentum under an electromagnetic field. Corresponding to the least action principle of
classical mechanics, the momentum P and energy E of a charged particle are given by
∂L ∂L
P= , E = v· −L. (2.5.28)
∂v ∂v
By (2.5.23), we refer from (2.5.28) that
mv e
P= 3 + ⃗A, ⃗A the magnetic potential,
2
1 − v /c 2 c
(2.5.29)
mc2
E=3 + eA0 , A0 the electric potential.
1 − v2/c2
It follows from (2.5.29) that
) e *2
(E − eA0)2 = c2 ⃗P − ⃗A + m2 c4 , (2.5.30)
c
which is the Einstein energy-momentum relation under with an electromagnetic field.
It is the formula (2.5.30) that makes us to take the energy and momentum operators in
quantum mechanics in the following form:
∂ c
ih̄− eA0, −ih̄∇ − ⃗A, (2.5.31)
∂t e
which were also given in (2.2.52). In particular, the particular form of (2.5.31) leads to the
origin of gauge theory.
2.5 Principle of Lagrangian Dynamics (PLD) 91

2.5.4 Lagrangian actions in quantum mechanics

PLD is also valid in quantum physics. In this subsection we shall introduce the actions
for basic equations of quantum mechanics: the Schrödinger equation, the Klein-Gordon
equation, and the Dirac equations.
Action for the Schrödinger equation
A particle moving at lower velocity can be approximatively described by the Schrödinger
equation, given by
∂ψ h̄2
ih̄ = − ∆ψ + V (x)ψ , (2.5.32)
∂t 2m
where m is the mass of the particle, V (x) is the potential energy, and ψ = ψ 1 + iψ2 is a
complex valued wave function. By the Basic Postulates 2.22-2.23, the equation (2.5.32) is
derived using the following the non-relativistic energy momentum relation:
1 ⃗2
E= P + V.
2m
Hence the Schrödinger equation (2.5.32) is a basic equation in non-relativistic quantum
mechanics.
The Lagrange action for the Schrödinger equation is
! T!
L= L (ψ , ψ ∗ )dxdt,
0 R3
# $ (2.5.33)
∂ ψ 1 h̄2
L = ih̄ψ ∗ − |∇ψ |2 + V |ψ |2
∂t 2 2m
We now compute δ L to show that (2.5.33) is indeed the action of (2.5.32). Take the
variation for ψ ∗ for (2.5.33):
!
d %
& ∗ dxdt = %% L(ψ , ψ ∗ + λ ψ
(δ L)ψ & ∗)
QT d λ λ =0
!
d %%
= % L (ψ , ψ ∗ + λ ψ
& ∗ )dxdt
QT d λ λ =0
! # $
∂ ψ ∗ h̄2
= ih̄ ψ
& − ∇ψ ∇ψ & +Vψψ
∗ ∗
& dxdt,
QT ∂t 2m

where QT = R3 × (0, T ), and ψ

& ∗ satisfies

ψ
& ∗ (0, x) = ψ
& ∗ (T, x) = 0 ∀x ∈ R3 ,
ψ
& →0
∗ as |x| → ∞.

Then by the Gauss formula, we have

! !
∇ψ · ∇ψ
& ∗ dxdt = − ∆ψ ψ
& ∗ dxdt,
QT QT
92 Chapter 2 Fundamental Principles of Physics

which implies that

! ! # $
∂ψ h̄2
δ Lψ
& ∗ dxdt = ih̄ + ∆ψ − V ψ ψ& ∗ dxdt.
QT QT ∂t 2m

Since ψ
& ∗ is arbitrary, we derive that

∂ψ h̄2
δ L = ih̄ + ∆ψ − V ψ .
∂t 2m
Hence we have

δL ∂ψ h̄2
=0 ⇔ ih̄ = − ∆ψ + V (x)ψ . (2.5.34)
δ ψ∗ ∂t 2m

We can derive in the same fashion that

δL ∂ ψ∗ h̄2
=0 ⇔ −ih̄ = − ∆ψ ∗ + V (x)ψ ∗ . (2.5.35)
δψ ∂t 2m

It follows from (2.5.34) and (2.5.35) that

! "∗
δL δL
= .
δ ψ∗ δψ

In other words, (2.5.34) and (2.5.35) are equivalent, and are exactly the Schrödinger equa-
tion.

Action for Klein-Gordon fields

The field equations governing the spin-0 bosons are the Klein-Gordon equations:

1 ∂ 2ψ ) mc *2
− ∆ψ + ψ = 0. (2.5.36)
c ∂t
2 2 h̄
It is easy to introduce the Lagrange action for (2.5.36):
! # ) mc *2 $
√
L= ∇µ ψ ∇µ ψ ∗ + |ψ |2 −gdx, (2.5.37)
M4 h̄

where ∇µ and ∇µ are 4-dimensional gradient operators as defined by (2.2.19), M 4 is the

Minkowski space, and g = det(g µν ).
It is readily to see that
! "∗ ) mc *2
δL δL 1 ∂ 2ψ
= =0 ⇔ − ∆ψ + ψ = 0.
δ ψ∗ δψ c ∂t
2 2 h̄

Hence, the Klein-Gordon equation is a variational equation of the action (2.5.37).

2.5 Principle of Lagrangian Dynamics (PLD) 93

Action for Dirac spinor fields

To introduce an action for the Dirac equations introduced in Subsections 2.2.5 and 2.2.6,
we first recall the Dirac equations:
∂ ψ mc
iγ µ µ − ψ = 0, (2.5.38)
∂x h̄
where ψ = (ψ1 , ψ2 , ψ3 , ψ4 )T is the Dirac spinor, and γ µ = (γ 0 , γ 1 , γ 2 , γ 3 ) is the Dirac ma-
trices, defined by
! " ! "
I 0 0 σk
γ0 = , γk = for k = 1, 2, 3, (2.5.39)
0 −I − σk 0
where the Pauli matrices are given by
! " ! " ! "
0 1 0 −i 1 0
σ1 = , σ2 = , σ3 = .
1 0 i 0 0 −1
The Lagrange action of (2.5.38) is
!
mc √
L= ψ (iγ µ ∂µ − )ψ −gdx, (2.5.40)
M4 h̄
where ψ = ψ † γ 0 .
The Lagrange density of (2.5.40) is
) mc *
L = ψ iγ µ ∂µ − ψ
h̄
is Lorentz invariant. To see this, recall the spinor transformation defined by (2.2.63):
µ
x&µ = Lν xν ⇒ ψ
& = Rψ ,

which satisfies that (2.2.62):

mc mc
(iγ µ ∂&µ − & = R(iγ µ ∂µ −
)ψ )ψ , (2.5.41)
h̄ h̄
where ∂&µ = ∂ /∂ x&µ , R is as in (2.2.67). By (2.2.68) and
R† γ 0 = γ 0 R† , (γ 0 )† = γ 0 . (2.5.42)

It follows from (2.5.41) and (2.4.2) that

mc ) mc *
ψ& (iγ µ ∂&µ − & =ψ † R† γ 0 R iγ µ ∂µ −
)ψ ψ
h̄ h̄
=(by ψ † γ 0 = ψ )
mc
=ψ (iγ µ ∂µ − )ψ .
h̄
It implies that the Lagrange action (2.5.40) is Lorentz invariant.
It is easy to see that
δL mc
=0 ⇔ (iγ µ ∂µ − )ψ = 0.
δ ψ∗ h̄
94 Chapter 2 Fundamental Principles of Physics

2.5.5 Symmetries and conservation laws

The importance of PLD lies in the following three points:

1) Physics are established based on a few universal principles, which provides a solid
foundation for physics;

2) Based on PLD, many physical problems become simpler. In particular, by means of

invariance it is easier to find the Lagrange actions than to seek for the differential
equations; and

3) Lagrange actions contain more physical information than the differential equations.
In fact, a conservation law of a physical system can be derived from the invariance
of the Lagrange action under the associated symmetric transformation.

The correspondence between symmetries and conservation laws are revealed by the
Noether theorem, to be introduced below. For this purpose, we need to introduce some
related concepts on group action and symmetry.
We begin with a simple example. A circle with radius r is described by

x2 + y 2 = r 2 . (2.5.43)

Figure 2.2

We can clearly see the symmetry of the circle shown in Figure 2.2—the graph is the same
from whatever the direction we look at it. This phenomenon is expressed in mathematics as
the invariance of the equation (2.5.43) under the following coordinate transformation
! ′" ! "! "
x cos θ sin θ x
= . (2.5.44)
y′ − sin θ cos θ y
2.5 Principle of Lagrangian Dynamics (PLD) 95

Namely, in the coordinate system (x′ , y′ ), the equation describing the circle is invariant and
still takes form:
x′2 + y′2 = r2 . (2.5.45)

Now, we discuss the symmetry from the viewpoint of action functional and group action.
The key feature is then characterized by the Noether theorem. Let F : R 2 → R1 be a function
(a finite dimensional functional) defined by

F(u) = x2 + y2 for u = (x, y) ∈ R2 . (2.5.46)

All the transformation matrices in (2.5.44) constitute a group, denoted by SO(2), called the
orthogonal group:
@ % ! " A
% cos θ sin θ
SO(2) = Aθ %% Aθ = , θ ∈ R1 . (2.5.47)
− sin θ cos θ

The symmetry here is the invariance of the function (2.5.46) under the SO(2) group action:

F(Aθ u) = F(u) ∀Aθ ∈ SO(2).

The generalization to the Lagrangian dynamics is what we shall introduce in this subsection.

Definition 2.37 Let G be a group, X be a Banach space, and F : X → R 1 a continuous

functional. Let G be a group acting on X:

Au ∈ X ∀u ∈ X and A ∈ G,
(2.5.48)
A(Bu) = (AB)u ∀u ∈ X, A, B ∈ G.

The functional F is called invariant under the action (2.5.48) of G, if

F(Au) = F(u) ∀A ∈ G.

We consider a Lagrange action defined on a function space X:

! T
L(u, u̇) = L (u, u̇)dt for u ∈ X, (2.5.49)
0

where the Lagrange density L is defined by

⎧
⎨ L (x, ẋ) for an N-body system,
L (u, u̇) = !
⎩ g(u, u̇, Du, · · · , Dm u)dx otherwise.
Ω

The following theorem is the well-known Noether theorem, which provides a correspon-
dence between symmetries and conservation laws in the Lagrangian system (2.5.49).
96 Chapter 2 Fundamental Principles of Physics

Theorem 2.38 (Noether Theorem) Let G = {Aλ | λ ∈ R1 } be a parameterized group,

and L(u, u̇) be the Lagrange action given by (2.5.49). If L is invariant under the group action
of G:
L (u, u̇) = L (Aλ u, Aλ u̇), ∀Aλ ∈ G, (2.5.50)
then the system has a conserved quantity induced by G, expressed as
D δ L dA (u) % E
%
I(u, u̇) = , λ % for Aλ ∈ G. (2.5.51)
δ u̇ d λ λ =0
In other words,
d
I(u, u̇) = 0 for any solutions u of δ L = 0. (2.5.52)
dt
Remark 2.39 The correspondence between symmetries and conservation laws in
the Noether Theorem holds true as well for discrete transformation groups, such as the
reflections of time and space.

Proof of Theorem 2.38 Let (u, ut ) be a solution of the variational equation of L given
by (2.5.49): ! "
d δ L (u, u̇) δ
= L (u, u̇). (2.5.53)
dt δ u̇ δu
By the invariance (2.5.50), (Aλ u, Aλ u̇) are also solutions of (2.5.53). Then it follows from
(2.5.50) that
∂ L (Aλ u, Aλ u̇) D δ L dΦ E D δ L d Φ̇ E
0= = , + , , (2.5.54)
∂λ ∂ Φ dλ δ Φ̇ d λ
where Φ = Aλ u. Since (Aλ u, Aλ u̇) satisfies (2.5.53), we have
! "
δ d δ L (Aλ u, Aλ u̇)
L (Aλ u, Aλ u̇) = . (2.5.55)
δΦ dt δ Φ̇
Inserting (2.5.55) in (2.5.54) we deduce that for any Φ = A λ u and any λ ∈ R1 ,
D d ! δ L " dΦ E D ∂ L d ! dΦ " E d D δ L dΦ E
0= , + , = , ,
dt δ Φ̇ dλ ∂ Φ̇ dt d λ dt δ Φ̇ d λ
which implies that
d d D δ L dΦ E%%
I(u, u̇) = , % = 0,
dt dt δ Φ̇ d λ λ =0
where Aλ u = u if λ = 0. The proof is complete.
We now give two examples to show how to apply the Noether theorem to a specific
physical problem.
Example 2.40 Consider an N-body motion, such as a system of N planets. Let m k
and xk be the mass and the coordinates of the k-th body. The Lagrange density is

1 N
L = ∑ mk |ẋk |2 − ∑ V (|xi − x j |).
2 k=1
(2.5.56)
i̸= j
2.5 Principle of Lagrangian Dynamics (PLD) 97

Note that x is the u in Theorem 2.38. Let G be the translation group:

G = {Aλ | Aλ : R3 → R3 , Aλ x = x + λ⃗r},

where ⃗r is a given vector. It clear that (2.5.56) is invariant under the transformation of G.
By
! "
δL ∂L ∂L
= ,··· , = (m1 ẋ1 , · · · , mN ẋN ),
δ ẋ ∂ ẋ1 ∂ ẋN
d %% d %%
% Aλ x = % (x + λ⃗r) = (⃗r, · · · ,⃗r ).
d λ λ =0 d λ λ =0 F GH I
N

Hence we derive from the Noether Theorem that

DδL d % E N
%
I= , % Aλ x = ∑ mk ⟨ẋk ,⃗r⟩
δ ẋ d λ λ =0 k=1

is conserved, and is the total momentum in the direction ⃗r. Thus we have shown that the
translation invariance corresponds to momentum conservation.
Example 2.41 Consider the Schrödinger equation, the action is given by (2.5.33), and
! # ! "$
1 h̄2
L (ψ , ψ̇ ) = −ih̄ψ ∗ ψ̇ + |∇ψ |2 + V |ψ |2 dx. (2.5.57)
R3 2 2m
Let G be the phase rotation group
G = {Aλ = eiλ | λ ∈ R1 }.

It is clear that (2.5.57) is invariant under the phase rotation:

ψ → Aλ ψ = eiλ ψ .
We see that
δL d
= −ih̄ψ ∗ , A ψ| = iψ .
δ ψ̇ d λ λ λ =0
Thus, the quantity (2.5.51) reads as
DδL d E !
I(ψ , ψ̇ ) = , Aλ ψ = h̄|ψ |2 dx.
δ ψ̇ d λ R3

Hence, the modulus of ψ !

|ψ |2 dx is conserved. (2.5.58)
R3
The property (2.5.58) is just what we need because in quantum mechanics, we have
!
|ψ |2 dx = 1. (2.5.59)
R3

Here, the conservation (2.5.59) corresponds to the invariance of phase rotation of the wave
function ψ .
98 Chapter 2 Fundamental Principles of Physics

Remark 2.42 In classical mechanics, the energy conservation can be deduced from
the time translation invariance

t →t +λ for λ ∈ R1 . (2.5.60)

However, instead of the formula (2.5.51), it is derived using following relation:

∂H ∂H ∂L
dH = dp+ dq − dt, (2.5.61)
∂p ∂q ∂t

where H = H(q, p,t) is the total energy, L is the Lagrangian, and q, p satisfy the Hamilton
equations
dq ∂ H dp ∂H
= , =− . (2.5.62)
dt ∂p dt ∂q
If L = L (q, q̇,t) is invariant under the translation (2.5.60), then L does not explicitly
contain time t, i.e.
∂L
= 0.
∂t
Then it follows from (2.5.61) and (2.5.62) that

dH ∂L
=− = 0.
dt ∂t
Hence the energy conservation is deduced using the time translation invariance.

We end this section with some relations of symmetries and conservation laws:

energy ⇔ time translation,

momentum ⇔ space translation,
angular Momentum ⇔ space rotation,
particle number ⇔ phase rotation,
parity ⇔ space reflection.

2.6 Principle of Hamiltonian Dynamics (PHD)

2.6.1 Hamiltonian systems in classical mechanics

In classical mechanics, the principle of Hamiltonian dynamics (PHD) consists of the fol-
lowing three main ingredients:

1) for an isolated (conserved) mechanical system, its states are described by a set of state
variables given by
q1 , · · · , qN , p1 , · · · , pN ; (2.6.1)
2.6 Principle of Hamiltonian Dynamics (PHD) 99

2) its total energy H is a function of (2.6.1), i.e.

H = H(q, p), (2.6.2)

3) the state variables q and p satisfy

dqk ∂H
= ,
dt ∂ pk
for 1 ! k ! N. (2.6.3)
d pk ∂H
=− ,
dt ∂ qk

The state variables qk (1 ! k ! N) represent positions, and pk (1 ! k ! N) represent mo-

mentums. The system (2.6.1)-(2.6.3) is called Hamiltonian system.
In physics, PLD and PHD are two independent fundamental principles. However, the
two sets of equations derived from PLD and PHD are usually equivalent.
In classical mechanics, Hamiltonian systems and Lagrange systems can be transformed
to each other by the Legendre transformation. We start with a simple example. Consider a
particle with mass m in a force field F. The Lagrange action for this system is given by
! T
L= L (q, q̇)dt,
0 (2.6.4)
1
L = T − V = mq̇2 − Fq,
2
where q stands for position, and the Euler-Lagrange equation of (2.6.4) is the Newtonian
Second Law, written as
d δL δL
= ⇒ mq̈ = F. (2.6.5)
dt δ q̇ δq
Based on (2.6.1)-(2.6.3), the state variables q, p of PHD are

q as in (2.6.4) and p = mq̇. (2.6.6)

The total energy is

1 2
H= p + Fq, (2.6.7)
2m
and the Hamilton equations of (2.6.7) are given by
dp ∂H
= = F,
dt ∂q
(2.6.8)
dq ∂H 1
=− = p.
dt ∂p m
By (2.6.6), it is clear that

Hamilton Eqs (2.6.8) = Lagrange Eq (2.6.5) = Newton 2nd Law. (2.6.9)

100 Chapter 2 Fundamental Principles of Physics

The equivalences in (2.6.9) only manifest that the three principles have an intrinsic relation.
In particular, by (2.6.4)-(2.6.7), the variables q, q̇ of PLD and the variables q, p of PHD have
the relation:
δL δL ∂L
p= (by = = mq̇), (2.6.10)
δ q̇ δ q̇ ∂ q̇
and the Lagrange density L (q, q̇) and the Hamilton energy H(q, p) are related by

pq̇ − L (q, q̇) = H(q, p). (2.6.11)

In this example, the two relations (2.6.10) and (2.6.11) are obtained directly by their
physical meanings, and by which we can deduce one by another.
In fact, the relations (2.6.10) and (2.6.11) are also valid in [Link] discuss this prob-
lem as follows.
Consider a mechanical system. For PLD, the state variables are positions q k and veloci-
ties q̇k :
q1 , · · · , qN and q̇1 , · · · , q̇N , (2.6.12)

The Lagrange action is given by

! T
L= L (q, q̇)dt. (2.6.13)
0

For PHD, the state variables are

q1 , · · · , qN and p 1 , · · · , pN , (2.6.14)

The Hamilton energy is

H = H(q, p). (2.6.15)

The two systems (2.6.12)-(2.6.13) and (2.6.14)-(2.6.15) of PLD and PHD satisfy the
following relations, which implies the equivalence of PLD and PHD in classical mechanics.

Dynamical Relation 2.43 (PLD and PHD) For the two systems of PLD and PHD in
classical mechanics, the following conclusions hold true:

1) The two sets (2.6.12) and (2.6.14) of variables satisfy the following relation:

∂ L (q, q̇)
pk = for 1 ! k ! N. (2.6.16)
∂ q̇k

2) The two functions (2.6.13) and (2.6.15) satisfy the following relation, which is usually
called the Legendre transformation:

pk q̇k − L (q, q̇) = H(q, p). (2.6.17)

2.6 Principle of Hamiltonian Dynamics (PHD) 101

By (2.6.16)-(2.6.17), if we have obtained the PLD system (2.6.12)-(2.6.13), then by the

implicit function theorem we can solve from (2.6.16) the functions

q̇k = fk (q, p). (2.6.18)

Then, inserting (2.6.18) in the left-hand side of (2.6.17) we deduce the expression of the
Hamilton energy:
H(q, p) = pk fk (q, p) − L (q, f (q, p)), (2.6.19)
which gives rise to the Hamiltonian system (2.6.3). In other words, we can derive the
Hamiltonian dynamics from the Lagrangian dynamics by the relations (2.6.16) and (2.6.17).
Conversely, if we know the PHD system (2.6.14) and (2.6.15), then it follows from
(2.6.16) and (2.6.17) that the Lagrange density L satisfies
! "
∂L ∂L
H x, − yk + L (x, y) = 0. (2.6.20)
∂y ∂ yk
Theoretically we can solve the differential equation (2.6.20), and obtain the solution L =
L (x, y). Let x = q and y = q̇, then we deduce the expression for the Lagrange density from
the Hamilton function H(q, p).
Remark 2.44 We remark that both PLD and PHD are independent to each other.
However, the PLD dynamical system (2.1.5)-(2.1.8) and the PHD dynamical system (2.6.1)-
(2.6.3) are equivalent in mechanics by (2.6.16) and (2.6.17). But we shall see later that the
relations (2.6.16) and (2.6.17) are not valid in electromagnetism where PLD and PHD still
hold true.
In addition, the differential dH is given by
∂H ∂H ∂H
dH = dqk + d pk + dt. (2.6.21)
∂ qk ∂ pk ∂t
By (2.6.17),
∂L ∂L
dH = q̇k d pk − dqk − dt. (2.6.22)
∂ qk ∂t
It follows from (2.6.21) and (2.6.22) that
∂H
q̇k = , (2.6.23)
∂ pk
∂H ∂L
=− , (2.6.24)
∂ qk ∂ qk
∂H ∂L
=− . (2.6.25)
∂t ∂t
From (2.6.21) and (2.6.25) we deduce (2.5.61). Then by (2.6.16) and the Lagrange equation
(2.1.8), we have ! "
d pk d δL ∂L
= = .
dt dt δ q̇k ∂ qk
102 Chapter 2 Fundamental Principles of Physics

Hence (2.6.24) becomes

∂H
ṗk = − . (2.6.26)
∂ qk
Thus, by the relations (2.6.16) and (2.6.17) we deduce the equivalence of PLD dynamics
and PHD dynamics in another fashion.

2.6.2 Dynamics of conservative systems

Energy conservation is a universal law in physics. It implies that the PHD is a universal
principle to describe conservative physical systems. In other words, the Hamilton principle
introduced in the last subsection can be generalized to all physical fields. The most remark-
able characteristic of PHD is that the total energy H of the physical system is conserved:

d
H(q(t), p(t)) = 0, (2.6.27)
dt
where (q, p) are the solutions of the Hamiltonian system.
Now, we introduce the PHD.

Principle 2.45 (Hamiltonian Dynamics) For any conservative physical system, there
are two sets of state functions

u = (u1 , · · · , uN ) and v = (v1 , · · · , vN ), (2.6.28)

such that the energy density H is a function of (2.6.28):

H = H (u, v, · · · , Dm u, Dm v), m $ 0. (2.6.29)

The total energy of the system is

!
H= H (u, v, · · · , Dm u, Dm v)dx, Ω ⊂ R3 , (2.6.30)
Ω

provided that the system is described by continuous fields. Moreover the state functions u
and v satisfy the equations
∂u δH
=α ,
∂t δv (2.6.31)
∂v δH
= −α ,
∂t δu
where α is a constant.

In general, the energy density (2.6.29) for a continuous field system depends on u, v only
up to the first-order derivatives Du and Dv:

H = H (u, v, Du, Dv). (2.6.32)

2.6 Principle of Hamiltonian Dynamics (PHD) 103

Then the Hamilton equations (2.6.31) can be expressed in the following form:
# ! " $
∂ uk ∂H ∂H
= α −∂ j + ,
∂t ∂ ζ jk ∂ vk
# ! " $ (2.6.33)
∂ vk ∂H ∂H
= α ∂j − ,
∂t ∂ ξ jk ∂ uk
where ξ jk , ζ jk are variables corresponding to ∂ j uk and ∂ j vk .
Hereafter, we always assume that H is in the form (2.6.32). The following theorem
shows that the total energy H of the Hamiltonian system (2.6.31) (or (2.6.33)) is conserved.
Theorem 2.46 (Energy Conservation) Let (u, v) be the solutions of (2.6.33), and H
does not explicitly contain time t. Then the energy H(u, v) is conserved, i.e. H(u, v) satisfies
(2.6.27) with q = u and p = v, if Ω = R3 . In addition, if Ω ̸= R3 , then we have
! # $
dH ∂H ∂uj ∂H ∂vj
= + ni ds, (2.6.34)
dt ∂ Ω ∂ ξi j ∂ t ∂ ζi j ∂ t
where n = (n1 , n2 , n3 ) is the unit outward normal at ∂ Ω.
Proof For (2.6.32), we have
! # ! " ! "$
dH ∂ H ∂ uk ∂ H ∂uj ∂ H ∂ vk ∂ H ∂vj
= + ∂i + + ∂i dx. (2.6.35)
dt Ω ∂ uk ∂ t ∂ ξi j ∂t ∂ vk ∂ t ∂ ζi j ∂t
By the Gauss formula,
! ! " ! ! ! "
∂H ∂uj ∂H ∂uj ∂H ∂uj
∂i dx = ni ds − ∂i dx,
Ω ∂ ξi j ∂t ∂ Ω ∂ ξi j ∂ t Ω ∂ ξi j ∂t
! " ! ! ! "
∂H ∂vj ∂H ∂vj ∂H ∂vj
∂i dx = ni ds − ∂i dx.
∂ ζi j ∂t ∂ Ω ∂ ζi j ∂ t Ω ∂ ζi j ∂ t
Hence (2.6.35) is rewritten as
! #! ! "" ! ! "" $
d ∂H ∂H ∂ uk ∂H ∂H ∂ vk
H= − ∂i + − ∂i dx (2.6.36)
dt Ω ∂ uk ∂ ξik ∂t ∂ vk ∂ ζik ∂t
! # $
∂H ∂uj ∂H ∂vj
+ + ni ds.
∂ Ω ∂ ξi j ∂ t ∂ ζi j ∂ t
Since (u, v) is a solution of (2.6.33), then (2.6.36) becomes
! # $
dH ∂H ∂uj ∂H ∂vj
= + ni ds,
dt ∂ Ω ∂ ξi j ∂ t ∂ ζi j ∂ t
and (2.6.34) follows.
If Ω = R3 (i.e. ∂ Ω = ∅) or u = v = 0 on ∂ Ω, then
d
H(u, v) = 0 for (u, v) satisfy (2.6.33).
dt
The proof is complete.
104 Chapter 2 Fundamental Principles of Physics

Remark 2.47 The integral functions in (2.6.34):

P = (P1 , P2 , P3 ),
# $
∂H ∂uj ∂H ∂vj (2.6.37)
Pk = − + ,
∂ ξk j ∂ t ∂ ζk j ∂ t
are the energy fluxes. Hence (2.6.34) can be expressed as
!
d
H(u, v) = − P · nds,
dt ∂Ω
which means that the rate of energy change in Ω equals to the difference of the input and
output of energy flow crossing the boundary of Ω per unit time. Consequently (2.6.34) is
equivalent to energy conservation.
In the Lagrangian dynamics, Noether Theorem 2.38 gives a correspondence between
symmetries and conservations, and provides a way to seek the conservation laws. Likewise,
the Hamiltonian dynamics provides another criterion to find conservation laws.
Let S(u, v) be a functional given by
!
S(u, v) = S(u, v, Du, Dv)dx. (2.6.38)
Ω

The following theorem provides a condition for S(u, v) to be a conserved quantity.

Theorem 2.48 (Conservation Laws of Hamiltonian System) Let S(u, v) be a functional
as given by (2.6.38). If S and the Hamilton energy H satisfy the following relation
! # ! "$ # ! "$
∂S ∂S ∂H ∂H
− ∂i − ∂i dx (2.6.39)
Ω ∂ uk ∂ ξik ∂ vk ∂ ζik
! # ! "$ # ! "$
∂S ∂S ∂H ∂H
= − ∂i − ∂i dx,
Ω ∂ vk ∂ ζik ∂ uk ∂ ξik
for solutions (u, v) of (2.6.33), then S is a conserved quantity of the Hamiltonian system.
Namely, S satisfies that !
dS
=− Ps · nds, (2.6.40)
dt ∂Ω
where Ps = (Ps1 , Ps2 , Ps3 ) is the flux given by
# $
k ∂S ∂uj ∂S ∂vj
Ps = − + .
∂ ξk j ∂ t ∂ ζk j ∂ t
Proof The proof is similar to that of Theorem 2.46. By
! # ! " ! "$
dS ∂S ∂uj ∂S ∂uj ∂S ∂vj ∂S ∂vj
= + ∂i + + ∂i dx
dt Ω ∂uj ∂t ∂ ξi j ∂t ∂vj ∂t ∂ ζi j ∂t
! #! ! "" ! ! "" $
∂S ∂S ∂uj ∂S ∂S ∂vj
= − ∂i + − ∂i dx
Ω ∂uj ∂ ξi j ∂t ∂vj ∂ ζi j ∂t
!
+ Psk · nk dS.
∂Ω
2.6 Principle of Hamiltonian Dynamics (PHD) 105

By (2.6.33) and (2.6.39) we deduce (2.6.40). The proof is complete.

In fact, Theorem 2.46 is a special case of Theorem 2.48 for S = H. Theorem 2.48 is
useful for Hamiltonian systems.

2.6.3 PHD for Maxwell electromagnetic fields

PHD is also valid in the Maxwell electrodynamics. To establish the Hamiltonian dynamics
for electromagnetism, we first determine the conjugate field functions as follows

u = E = (E1 , E2 , E3 ),
(2.6.41)
v = A = (A1 , A2 , A3 ),

where E is the electric field, and A is the magnetic potential. The Hamilton energy H is
given by !
H= H (E, A)dx,
Ω
(2.6.42)
1 1 1
H = (E 2 + |curl A|2 ) + ∇ϕ · E − J · A,
8π 4π c
where ϕ is the electric potential, and J is the current. Using
DδH E d %%
, E& = % H(E + λ E,
& A),
δE d λ λ =0
DδH E %
,A & = d %% H(E, A + λ A),
&
δA d λ λ =0
we can compute the derivatives of (2.6.42) as follows:

δH 1 δH 1 1
= (E + ∇ϕ ), = curl2 A − J.
δE 4π δA 4π c
Thus, the Hamilton equations (2.6.31) are in the form:

1 ∂E δH 4π
= 4π = curl2 A − J,
c ∂t δA c (2.6.43)
1 ∂A δH
= −4π = −E − ∇ϕ ,
c ∂t δE
which are the classical Maxwell equations (2.2.33) and (2.2.35).

Remark 2.49 We note that the Lagrange action of electromagnetism defined by

(2.4.15) can be expressed as
! # $
1 1 4π
LEM = − |curl A|2 + E 2 + Aµ J µ dxdt, (2.6.44)
M4 2 2 c

where
1 ∂A
E =− − ∇A0 , Aµ = (A0 , A1 , A2 , A3 ).
c ∂t
106 Chapter 2 Fundamental Principles of Physics

The fields of PLD are given by

A0 , A1 , A2 , A3 and Ȧ0 , Ȧ1 , Ȧ2 , Ȧ3 (A0 = ϕ ),

and the fields of PHD are

A1 , A2 , A3 and E 1 , E2 , E3 .

It is clear that
1 1 1
E(−Ȧ) − LEM = H (A, E) − ϕ J 0 ,
4π 4π c
and J 0 = ρ . Hence, the relation (2.6.17) does not hold true in general.

Now, we consider the energy conservation. When H contains ∇ϕ and J, which depend
on t, the Hamilton H = H (A, E,t) contains explicitly t. It implies that the Maxwell fields
have energy exchange with other charged particles. Then H is not conserved.
As there is no charged particles in Ω, then

(ρ , J) = 0 in Ω.

In this case, ! ! !
∇ϕ · Edx = − ϕ div Edx = − 4πϕρ dx = 0.
Ω Ω Ω

Hence, the Hamilton energy (2.6.42) becomes

!
1
H= (E 2 + |curl A|2 )dx. (2.6.45)
8π Ω

By (2.6.37), the energy flux P defined in (2.6.45) reads

1 ∂A
P= curl A ×
4π ∂t
Note that
∂A
= −cE, curl A = B (B the magnetic field),
∂t
Then, by (2.6.34) we have that
! ! ) c *
dH
=− P · nds = − E × B · nds.
dt ∂Ω ∂ Ω 4π

The field
c
P=
E ×B
4π
is the Poynting vector, represents the energy flux density of an electromagnetic field.
2.6 Principle of Hamiltonian Dynamics (PHD) 107

2.6.4 Quantum Hamiltonian systems

In quantum physics, the state function is a set of complex valued wave functions:

ψ = (ψ1 , · · · , ψN )T , N $ 1,

and ψk are as
ψk = ψk1 + iψk2 for 1 ! k ! N. (2.6.46)
In view of PHD for a quantum system, the conjugate fields are taken as real and imagi-
nary parts of the wave functions in (2.6.46):

ψ11 , · · · , ψN1 and ψ12 , · · · , ψN2 . (2.6.47)

Let H = H(ψ ) be the Hamilton energy. Then the Hamilton equations for the quantum
system are as follows:
∂ ψk1 δ
α = H,
∂t δ ψk2
for 1 ! k ! N, (2.6.48)
∂ ψ2 δ
α k = − 1 H,
∂t δ ψk
where α is a constant.
We now introduce quantum Hamiltonian systems for the Schrödinger equation, the Weyl
equation, the Dirac equations, the Klein-Gordon equation, and the BEC equation.
1. Schrödinger equation:
∂ψ h̄2
ih̄ = − ∆ψ + V (x)ψ , (2.6.49)
∂t 2m
where ψ = ψ 1 + iψ 2 . The Hamilton energy of (2.6.49) is given by
! # 2 $
1 h̄
H(ψ ) = |∇ψ |2 + V (x)|ψ |2 dx.
2 Ω 2m
It is easy to derive that
δ h̄2
H = − ∆ψ 1 + V ψ 1 ,
δ ψ1 2m
δ h̄2
H = − ∆ψ 2 + V ψ 2 .
δψ 2 2m
Hence, the Hamiltonian system is
∂ ψ1 h̄2
h̄ = − ∆ψ 2 + V ψ 2 ,
∂t 2m (2.6.50)
∂ ψ2 h̄2
h̄ = ∆ψ 2 − V ψ 2 .
∂t 2m
108 Chapter 2 Fundamental Principles of Physics

It is clear that (2.6.50) and (2.6.49) are equivalent.

2. Weyl equations:
1 ∂ψ
= (⃗σ · ∇)ψ , (2.6.51)
c ∂t
where ψ = (ψ1 , ψ2 )T , and ψk = ψk1 + iψk2 (1 ! k ! 2). The Hamilton energy H of (2.6.51)
is in the form
!
H= iψ † (⃗σ · ∇)ψ dx (2.6.52)
R3
! # $
∂ ψ12 1 ∂ ψ22 1 ∂ ψ12 2 ∂ ψ11 1 ∂ ψ12 1 ∂ ψ21 1 ∂ ψ21 2
= ψ
1 2
+ ψ + ψ + ψ + ψ + ψ + ψ dx.
R3 ∂ x ∂ x1 1 ∂ x2 2 ∂ x2 2 ∂ x3 1 ∂ x3 1 ∂ x3 2

The Hamilton equations are

1 ∂ ψk1 δ
= H,
c ∂t δ ψk2
for k = 1, 2,
1 ∂ ψk2 δ
=− H,
c ∂t δ ψk1

which, in view of for (2.6.52), are in the form:

! "
1 ∂ ψ11 ∂ ψ21 ∂ ψ22 ∂ ψ11
= + 2 + 3 ,
c ∂t ∂ x1 ∂x ∂x
2 ! "
1 ∂ ψ1 ∂ ψ2 ∂ ψ21
2 ∂ ψ12
= − 2 + 3 ,
c ∂t ∂ x1 ∂x ∂x
1 ! " (2.6.53)
1 ∂ ψ2 ∂ ψ1 ∂ ψ12
1 ∂ ψ21
= − 2 − 3 ,
c ∂t ∂ x1 ∂x ∂x
! "
1 ∂ ψ22 ∂ ψ12 ∂ ψ11 ∂ ψ22
= + 2 − 3 .
c ∂t ∂ x1 ∂x ∂x
It is readily to check that (2.6.53) and (2.6.51) are equivalent.
3. Dirac equations:

∂ψ
ih̄ = −ih̄c(⃗α · ∇)ψ + mc2α0 ψ , (2.6.54)
∂t
where ⃗α , α0 are as in (2.2.58), ψ = (ψ1 , ψ2 , ψ3 , ψ4 )T , and

ψk = ψk1 + iψk2 for 1 ! k ! 4.

The Hamilton energy of (2.6.54) is defined by

! 4 5
H= −ih̄cψ † (⃗α · ∇)ψ + mc2ψ † α0 ψ dx. (2.6.55)
R3
2.6 Principle of Hamiltonian Dynamics (PHD) 109

It is clear that the expansions of (2.6.54) in terms of real and imaginary parts are in the form

∂ ψk1 δ
h̄ =− H,
∂t δ ψk2
for 1 ! k ! 4,
∂ ψk2 δ
h̄ = H,
∂t δ ψk1
where H is given by (2.6.55).
4. Klein-Gordon equation:

1 ∂ 2ψ ) mc *2
− ∆ ψ + ψ = 0. (2.6.56)
c2 ∂ t 2 h̄
The conjugate fields are
∂ψ
ψ1 = ψ, ψ2 = .
∂t
The Hamilton energy is
! # ) mc *2 $
1
H= |ψ 2 |2 + |∇ψ 1 |2 + |ψ 1 |2 dx. (2.6.57)
2 Ω h̄
Then (2.6.56) can be equivalently written as

1 ∂ ψ1 δ
= H,
c ∂t δ ψ2
1 ∂ ψ2 δ
=− H.
c ∂t δ ψ1

5. Bose-Einstein condensation (BEC) equation:

∂ψ h̄2
ih̄ = − ∆ψ + V (|ψ |2 )ψ , (2.6.58)
∂t 2m
where V (|ψ |2 ) is a function of |ψ |2 , and ψ = ψ 1 + iψ 2 . The Hamilton energy of (2.6.58) is
given by
! # 2 $
h̄ 2 2
H= |∇ψ | + G(|ψ | ) dx,
Ω 4m
! (2.6.59)
1 z
G(z) = V (s)ds.
2 0
For (2.6.59), the equation (2.6.58) can be rewritten as

∂ ψ1 δ
h̄ = H,
∂t δ ψ2
∂ ψ2 δ
h̄ =− H.
∂t δ ψ1
110 Chapter 2 Fundamental Principles of Physics

The examples given above show that PHD holds true in general in quantum physics.
In particular, we now show that the relations (2.6.16) and (2.6.17) also valid for quantum
Hamiltonian systems.
In fact, the Lagrange action of a quantum system with the Hamilton energy H(ψ ) is
given by
! T!
L= L (ψ , ψ̇ )dxdt,
0 R3 (2.6.60)
L = ih̄ψ ψ̇ − H (ψ , Dψ ),
∗

where H reads !
H= H (ψ , Dψ )dx.
R3
When taking the variation, the functional
!
&
L= iψ ∗ ψ̇ dxdt
QT
!
= i(ψ 1 − iψ 2 )(ψ̇ 1 + iψ̇ 2 )dxdt
QT
! # $
i ∂ ∂
= |ψ |2 − (ψ 1 ψ 2 ) + 2ψ 2ψ̇1 dxdt
QT 2 ∂ t ∂t
! T!
= 2ψ 2 ψ̇ 1 dxdt,
0 R3

where QT = R3 × (0, T ). Hence, the density L in (2.6.60) is equivalent to

L = 2h̄ψ 2 ψ̇ 1 − H . (2.6.61)

It follows from (2.6.61) that

1
ψ2 = ∂ L /∂ ψ̇ 1 ,
2h̄ (2.6.62)
1 1
ψ 2 ψ̇ 1 − L = H .
2h̄ 2h̄
The relations (2.6.62) are what we expected.
Chapter 3
Mathematical Foundations

The aim of this chapter is to provide mathematical foundations for the remaining part of the
book on the unified field theory, elementary particles, quantum physics, astrophysics and
cosmology.
As addressed in Chapter 2, Nature speaks the language of mathematics. Thanks to
Einstein’s principle of equivalence and the geometric interaction mechanism, the space-time
is a 4D Riemannian manifold M , and physical fields are regarded as functions on vector
bundles over the base manifold M . For example, the Riemannian metric g µν , representing
the gravitational potential, is a function on the second-order cotangent bundle, g µν : M →
T ∗ M ⊗ T ∗ M ; the electromagnetic field A µ is a function on the cotangent bundle A µ : M →
T ∗ M ; the SU(2) gauge fields for the weak interaction is a function {Wµa } : M → (T ∗ M )3 ;
and the SU(3) gauge fields for the strong interaction is a function {S kµ } : M → (T ∗ M )8 .
Also, the Dirac spinor field is defined a complex bundle: Ψ : M → M ⊗ p C4 .
The three most fundamental symmetries of Nature are the principle of Lorentz invari-
ance, the principle of general relativity, and the principle of gauge invariance. They lead
to transformations in either the base space-time manifold M , or the corresponding vector
bundles.
In summary, we have the following
1) The space-time is a 4D Riemannian manifold, and all physical fields are
functions on vector bundles over the space-time manifold M ⊗ p E N ; and
2) Fundamental symmetries are invariances of physical field equations un-
der the underlying transformations on the corresponding vector bundles.
The basic concepts in Section 3.1 include Riemannian manifolds such as the space-time
manifold, vector bundles, tensor fields, connections and linear transformations on vector
bundles. It is particularly important that we identify all physical fields as functions on
proper vector bundles defined on the space-time manifold, and symmetries correspond then
to linear transformations on the corresponding vector bundles.
Section 3.2 is on basic functional analysis and partial differential equations on mani-
folds, which are needed for rigorous proofs of theorems and concepts used in later chapters
of the book. The readers may consult other references such as (Ma, 2011).
112 Chapter 3 Mathematical Foundations

Section 3.3 proves the orthogonal decomposition theorem for tensors on Riemannian
manifolds, which forms the mathematical foundation for the principle of interaction dy-
namics (PID), postulated by (Ma and Wang, 2014e, 2015a). Basically, PID takes the vari-
ation of the Lagrangian action L(u) under the generalized energy-momentum conservation
constraints, which we call divA -free constraints with A being the gauge potentials:

d !! def
⟨δ L(u0 ), X⟩ = ! L(u0 + λ X) = 0, ∀ divA X = divX − X · A = 0. (3.0.1)
d λ λ =0
Here divA X = 0 represents a generalized energy-momentum conservation. The study of the
constraint variation (3.0.1) requires the decomposition of all tensor fields into the space of
divA -free fields and its orthogonal complements. Such a decomposition is reminiscent to
the classical Helmholtz decomposition of a vector into the sum of an irrotational (curl-free)
vector field and a solenoidal (divergence-free) vector field. In particular, we show in this
chapter that (see Theorem 3.17):

L2 (Trk M ) = G(Trk M ) ⊕ L2D (Trk M ),

" #
G(Trk M ) = v ∈ L2 (Trk M )| v = Dϕ + A ⊗ ϕ , ϕ ∈ H 1 (Tr−1
k
M) , (3.0.2)

L2D (Trk M ) = {v ∈ L2 (Trk M )| divA v = 0},

where D is the connection on the space-time manifold.

The first part of Section 3.4 deals with classical variations of the Lagrangian actions,
and gives detailed calculations for the variations of the Einstein-Hilbert functional and the
Yang-Mills action. The second part studies variations under div A -free constraints, where
A stands for gauge potentials. These constraints represent generalized energy-momentum
conservation and provide basic mathematical theorems for deriving the unified field model
for the fundamental interactions. This section is based entirely on (Ma and Wang, 2014e,
2015a).
Section 3.5 explores the inner/hidden symmetry behind the SU(N) representation for
the non-abelian gauge theory. Basically, we have realized in (Ma and Wang, 2014h) that
the set of generators SU(N) plays exactly the role of a coordinate system, leading to a new
invariance, which we call the principle of representation invariance (PRI), first discovered
in (Ma and Wang, 2014h).
Section 3.6 addresses the spectral analysis of the Dirac and Weyl operators, which will
play an important role in studying the energy-levels of subatomic particles in Section 6.4,
which is based entirely on (Ma and Wang, 2014g).
3.1 Basic Concepts 113

3.1 Basic Concepts

3.1.1 Riemannian manifolds
The n-sphere Sn and the Euclidean space Rn are two typical examples of n-dimensi-
onal manifolds. The most common manifolds, which we can visually see, are one-dimensional
curves and two-dimensional surfaces. It is, however, difficult for us to tell whether the
three-dimensional space we live in is curved or flat by our common sense. The Riemannian
geometry provides a theory with which the intelligent beings living in an n-dimensional
manifold are able to determine the curvature of this space from its metric.
A plane in Figure 3.1 (a) is expressed by

⃗r(x1 , x2 ) = (x1 , x2 , x3 (x1 , x2 )), (3.1.1)

where x3 = α1 x1 + α2 x2 + α3 , α j (1 ! j ! 3) are constants, and the metric of the plane

(3.1.1) are given by

ds2 = d⃗r · d⃗r =(dx1 )2 + (dx2 )2 + (α1 dx1 + α2 dx2 )2 (3.1.2)

=(1 + α12)(dx1 )2 + 2α1 α2 dx1 dx2 + (1 + α22)(dx2 )2 .

A sphere in Figure 3.1 (b) is expressed by

⃗r(x1 , x2 ) = (x1 , x2 , x3 (x1 , x2 )), (3.1.3)

Figure 3.1 (a) a plane, and (b) a sphere with radius R

$
where x3 = R2 − (x1 )2 − (x2 )2 , and R is the radius. The metric of this sphere (3.1.3) is
given by
% &2
2 1 2 ∂ x3 1 ∂ x3 2
2 2
ds =(dx ) + (dx ) + dx + 2 dx (3.1.4)
∂ x1 ∂x
=(1 + ϕ )(dx1 )2 + 2ϕ dx1 dx2 + (1 + ϕ )(dx2)2 ,
114 Chapter 3 Mathematical Foundations

where ϕ = 1/(R2 − (x1 )2 − (x2 )2 ).

It is clear that the metric
ds2 = gi j dxi dx j
defined on a surface dictates its curvature. We see that the metric (3.1.2) of the plane (3.1.1),
i.e. % & % &
g11 g12 1 + α12 α1 α2
= (3.1.5)
g21 g22 α1 α2 1 + α22
is a constant metric, and the metric (3.1.4) of the sphere (3.1.3), i.e.
% & % &
g11 g12 1 + ϕ (x1, x2 ) ϕ (x1 , x2 )
= (3.1.6)
g21 g22 ϕ (x1 , x2 ) 1 + ϕ (x1 , x2 )

is not constant. In fact, for the plane shown by Figure 3.1 (a), we can find a coordinate
transformation % 1& % 1 & % 1&
x' a1 a12 x
= ,
x'2 a21 a22 x2
x1 , x'2 ), the metric (3.1.2) is expressed in the diagonal
such that in the coordinate system ('
form:
ds2 = (d' x1 )2 + (d'
x2 )2 . (3.1.7)
In other words, the metric is gi j = δi j . However, it is impossible to achieve this for the
metric (3.1.4) of the sphere.
The conclusion in this example holds true as well for all Riemannian manifolds. Namely,
for a Riemannian manifold {M , gi j }, M is flat if and only if there is a coordinate system x,
such that the metric {gi j } can be expressed as gi j = δi j under the x-coordinate system.
The following are a few general properties of an n-dimensional Riemannian manifold
{M , gi j }.
1. The Riemannian metric
ds2 = gi j dxi dx j (3.1.8)
is invariant. In fact, under a coordinate transformation

x' = ϕ (x), x = ϕ −1 ('

x), (3.1.9)

the second-order covariant tensor field {gi j } satisfies that

⎛ 1⎞ ⎛ 1⎞
d'
x dx
⎜ .. ⎟ i ⎜ .. ⎟
gi j ) = (bki )T (gkl )(blj ),
(' ⎝ . ⎠ = (a j ⎝ . ⎠,
)
xn
d' dxn
% i
&
∂ϕ
(aij ) = , (bij ) = (aij )−1 .
∂xj
3.1 Basic Concepts 115

Hence we have
⎛ ⎞
x1
d'
⎜ .⎟
xi d'
g'i j d' x j =(d'
x1 , · · · , d'
xn )(' ⎝ ..⎠
gi j ) ⎜ ⎟ (3.1.10)
xn
d'
⎛ ⎞
dx1
=(dx1 , · · · , dxn )(aki )T (bki )T (gkl )(blj )(alj ) ⎝ ... ⎠
⎜ ⎟

dxn
⎛ 1
⎞
dx
1 n ⎜ .. ⎟
=(dx , · · · , dx )(gi j ) ⎝ . ⎠
dxn
=gi j dxi dx j .

It follows that the metric (3.1.8) is invariant.

2. The length of a curve on M is determined by the metric {g i j }. Let γ (t) be a curve
connecting two points p, q ∈ M, and γ (t) be expressed by

x(t) = (x1 (t), · · · , xn (t)), 0 ! t ! T, x(0) = p, x(T ) = q. (3.1.11)

By (3.1.8) the infinitesimal arc-length is given by

. .
ds = gi j dxi dx j = gi j (x(t))ẋi ẋ j dt.

Hence the length L of γ (t) is given by

/
! q ! T
dxi dx j
L= ds = gi j (x(t)) dt. (3.1.12)
p 0 dt dt

The invariance of ds as shown in (3.1.10) implies that the length L in (3.1.12) is independent
of the coordinate systems.
3. The volume of a bounded domain U ⊂ M is invariant. The reason why {g i j } is called
a metric is that such quantities as the length, area, volume and angle are all determined by
the metric {gi j }.
Given a Riemannian manifold {M , gi j }, let U ⊂ M be a bounded domain. Then the
volume of U is written as !
V = Ω(x)dx. (3.1.13)
U
where the volume element Ωdx reads
√
Ω(x)dx = −gdx, g = det(gi j ). (3.1.14)
116 Chapter 3 Mathematical Foundations

To derive (3.1.14), let M ⊂ Rn+1 be an embedding, and

⃗r(x) = {r1 (x), · · · , rn+1 (x)},

be the embedding function. Consider the vector product of n vectors in R n :

! !
! ⃗e1 · · · ⃗en+1 !
0 1 !!∂ r · · · ∂ r !!
∂⃗r ∂⃗r ! 1 1 1 n+1 !
,··· , n = ! . .. ! ,
∂x 1 ∂x ! . . . !!
!
!∂n r1 · · · ∂n rn+1 !

where {⃗e1 , · · · ,⃗en+1 } is an orthogonal basis of Rn+1 . Then the volume element Ωdx is
!0 1!
! ∂⃗r ∂⃗r !!
Ωdx = !! , · · · , dx.
∂ x1 ∂ xn !
∂⃗r ∂⃗r
By gi j = · , the norm of the vector [∂1⃗r, · · · , ∂n⃗r] is
∂ xi ∂ x j
√
|[∂1⃗r, · · · , ∂n⃗r]| = −g, g = det (gi j ).

Thus (3.1.14) follows.

We now verify the invariance of the volume element. Under the transformation (3.1.9),

d' x1 ∧ · · · ∧ d'
x = d' xn (3.1.15)
% 1 & % n &
∂ϕ 1 ∂ ϕ1 n ∂ϕ 1 ∂ ϕn n
= dx + · · · + dx ∧ · · · ∧ dx + · · · + dx
∂ x1 ∂ xn ∂ x1 ∂ xn
% i&
∂ϕ
= det dx1 ∧ · · · ∧ dxn .
∂xj
On the other hand,
% & % &T
∂ ψi ∂ ψi
gi j ) =
(' (gi j ) , ψ = ϕ −1 . (3.1.16)
∂yj ∂yj
Hence % i &−1 .
.
∂ϕ
det('
gi j ) = det det(gi j ). (3.1.17)
∂xj
We deduce from (3.1.15) and (3.1.17) that
. .
det('
gi j )d'
x = det(gi j )dx.

Namely, both the volume and the volume element in (3.1.13)-(3.1.14) are invariant.
4. The metric {gi j } gives rise to an inner product structure on the tangent space of a
Riemann manifold M , and defines the angle between two tangent vectors.
3.1 Basic Concepts 117

Let p ∈ M be a given point, and Tp M be the tangent space at p ∈ M. For two vectors
X,Y ∈ Tp M ,
X = {X 1, · · · , X n }, Y = {Y 1 , · · · ,Y n },

the inner product of X and Y is defined by

⟨X,Y ⟩ = gi j (p)X iY j . (3.1.18)

It is clear that the inner product (3.1.18) is an invariant. The angle between X and Y is
defined as .
⟨X,Y ⟩
cos θ = , |Z| = gi j Z i Z j , for Z = X,Y.
|X||Y |

5. The inverse of (gi j ), denoted by

(gi j ) = (gi j )−1 ,

is a second order contra-variant tensor. In fact, by (3.1.16) we have that

% &−1 2% &T 3−1
ij ∂ϕ ∂ϕ
I = ('
gi j )('
g )= (gi j ) gi j ),
('
∂x ∂x

where I is unit matrix. It follows that

% &T % &
ij ∂ϕ ij ∂ϕ
g )=
(' (g ) .
∂x ∂x

Hence {gi j } is a second-order contra-variant tensor.

3.1.2 Physical fields and vector bundles

Let M be a manifold. A vector bundle on M is obtained by gluing an N-dimensional linear
space E N at each point p ∈ M , denoted by

def "
M ⊗p EN = {p} × E pN , (3.1.19)
p∈M

where E N = RN , or CN , or the Minkowski space.

In a vector bundle (3.1.19), the geometric position of E pN is related with p ∈ M . For
example, the set of all tangent spaces is a vector bundle on M , called the tangent bundle of
M , denoted by
T M = M ⊗ p Tp M .
118 Chapter 3 Mathematical Foundations

For (3.1.19), if the positions of bundle spaces E pN (p ∈ M ) are independent of p, then it

is called a geometrically trivial vector bundle. Vector bundles on a flat manifold are always
geometrically trivial.
The physical background of vector bundles are very clear. Each physical field must be a
mapping from the base space M to a vector bundle M ⊗ p E N . Actually, all physical events
occur in the space-time universe M , and the physical fields describing these events are
defined on vector bundles. This point of view is well illustrated in the following examples.

Example 3.1 The motion of the air or seawater can be ideally considered as a fluid
motion on a two-dimensional sphere S 2 , and the velocity field u is defined on the tangent
planes T S2 :

u : S2 → T S2 with u(p) ∈ Tp S2 , ∀p ∈ S2 .

Example 3.2 The electromagnetic interaction takes place on the 4D space-time mani-
fold M , and the electromagnetic field A µ is defined on the tangent space:

Aµ : M → T M with Aµ (p) ∈ Tp M , ∀p ∈ M ,

where at each point p ∈ M , Tp M is the Minkowski space.

Example 3.3 An electron moves in the space-time manifold M , and the field de-
scribing the electron state is the Dirac spinor ψ , which is defined on a 4D complex space
C4 :

ψ : M → M ⊗ p C4p with ψ (p) ∈ C4p , ∀p ∈ M .

Examples 3.1-3.3 clearly illustrate that all physical fields are defined on a vector bundle
on M . Namely, a physical field F is a mapping:

F : M → M ⊗ p E pN with F(p) ∈ E pN , ∀p ∈ M . (3.1.20)

The physical field F takes its values in the N-dimensional linear space E N , and can be
written as N components,
F = (F1 , · · · , FN )T . (3.1.21)
Considering invariance under certain symmetry, the transformation group always acts on
the bundle space E N , and induces a corresponding transformation for the field F in (3.1.21).
Also, in order to ensure the covariance of the field equations for F, a differential operator
D acting F must also be covariant, leading to the introduction of connections on the vector
bundle M ⊗ p E N . This process is shown as follows:
1. Symmetric group G act on E N :

G p : E pN → E pN . (3.1.22)
3.1 Basic Concepts 119

2. The fields F induced to be transformed:

F → Tp F. (3.1.23)

3. The vector bundle M ⊗ p E N is endowed with connections Γ µ :

Dµ = ∂µ + Γµ . (3.1.24)

Finally, we say that the vector bundle M ⊗ p E N is geometrically trivial if and only if the
connections Γ µ in (3.1.24) are zero, i.e. Γ µ = 0 on M . It implies that the transformations
(3.1.22)-(3.1.23) determines whether M ⊗ p E pN is geometrically trivial, and if G p , Tp in
(3.1.22) and (3.1.23) are independent of p ∈ M , then M ⊗ p E N is geometrically trivial,
and otherwise it’s not. This viewpoint is important for the unified field theory introduced
in Chapter 4, because it implies that the geometry of M ⊗ p E N is determined by symmetry
principles.
The other reason to adopt vector bundles as the mathematical framework to describe
physical fields F is that the types of F can be directly reflected by the bundle space E N .
Hereafter we list a few useful physical fields:

1) For a real (complex) scalar field φ , the associated vector bundle is M ⊗ p R1 (M ⊗ p

C):
φ : M → M ⊗ p R1 (M ⊗ p C).

2) Let Aµ be a 4-dimensional vector field. Then

Aµ : M → T M ,

and for any p ∈ M , Tp M is the Minkowski space.

3) For a 4-dimensional covector field A µ , we have

Aµ : M → T ∗ M ,

and for any p ∈ M , Tp∗ M is the dual space of Tp M .

4) A (k, r)-tensor field T on M :

µ ··· µ
T = {Tν11···νr k },
is expressed by
T : M → Trk M ,
where Trk M is the (k, r)-tensor bundle on M , denoted by

Trk M = T M ⊗ · · · ⊗ T M ⊗ T ∗ M ⊗ · · · ⊗ T ∗ M , (3.1.25)
4 56 7 4 56 7
k r

and ⊗ represents the tensor product.

120 Chapter 3 Mathematical Foundations

5) The Dirac spinor field Ψ is defined by

Ψ : M → M ⊗ p C4 .

In particular, for N Dirac spinor fields Ψ = (ψ 1 , · · · , ψ N )T ,

Ψ : M → M ⊗ p (C4 )N .

6) The Riemann metric g µν defined on a 4D space-time manifold M , representing the

gravitational potential, is a mapping:

g µν : M → T20 M ,

where T20 M = T ∗ M ⊗ T ∗ M is a (0,2)-tensor bundle as defined by (3.1.25).

The fields given by 1)-6) above include all types of physical fields, and the associated
vector bundles are physically significant.
In classical theories of interactions, the physical fields and the associated bundle spaces
are given as follows:

1) Gravity:
g µν : M → T ∗ M ⊗ T ∗ M . (3.1.26)

2) Electromagnetism: U(1) gauge and fermion fields,

Aµ : M → T ∗ M , Ψ : M → M ⊗ p C4 . (3.1.27)

3) Weak interaction: SU(2) gauge fields

Wµa : M → (T ∗ M )3 , (Ψ1 , Ψ2 )T : M → M ⊗ p (C4 )2 . (3.1.28)

4) Strong interaction: SU(3) gauge fields and fermion fields,

Skµ : M → (T ∗ M )8 , (Ψ1 , Ψ2 , Ψ3 )T : M → M ⊗ p (C4 )3 . (3.1.29)

In the unified field theory to be introduced in Chapter 4, each of the four interactions
(3.1.26)-(3.1.29) possesses corresponding dual fields as follows:

φµG ↔ g µν , φ E ↔ Aµ , φaw ↔ Wµa , φks ↔ Skµ .

The corresponding vector bundles for the dual fields are as follows
φµG : M → T ∗ M ,
φ E : M → M ⊗ p R1 ,
(3.1.30)
φaw : M → M ⊗ p R3 ,
φks : M → M ⊗ p R8 .
3.1 Basic Concepts 121

The fields (3.1.26)-(3.1.30) are all physical fields in the unified field theory. The dual
fields (3.1.30) are introduced only in the unified field equations using PID, but they do not
appear in the actions of the unified field theory.

3.1.3 Linear transformations on vector bundles

In the last subsection, a field on a manifold M can be regarded as a mapping from the base
manifold M to some vector bundle M ⊗ p E N :

F : M → M ⊗p EN . (3.1.31)

Let G be a transformation group acting on the fiber space E N of the bundle. This group
action induces naturally a transformation on the bundle M ⊗ p E N as follows:

X → gX, ∀X ∈ E pN , g ∈ G, p ∈ M. (3.1.32)

For the field (3.1.31), we know that

F(p) ∈ E pN , ∀p ∈ M .

Hence the transformation (3.1.32) induces a natural transformation on the field F:

F → gF, ∀g ∈ G. (3.1.33)

As discussed in Subsection 2.1.5, each symmetry possesses three ingredients: space

(manifold), transformation group, and tensors. Consequently, the spaces M are different for
different symmetries. However, in the fashion provided by (3.1.31)-(3.1.33), the base space
M is fixed, and all symmetric transformations as groups act on the bundle spaces E N as in
(3.1.32), and the fields of (3.1.31) are automatically transformed as in (3.1.33). Thus, the
physical invariances are referred to a few group actings on bundle spaces, i.e. the linear
transformation on E N .
To illustrate this idea, we start with general tensors on an n-dimensional Riemannian
manifold M . Let two coordinate systems x = (x1 , · · · , xn ) and x' = ('
x1 , · · · , x'n ) are trans-
forming under the following coordinate transformation:

x'k = ϕ k (x), 1 ! k ! n. (3.1.34)

Then (k, r)-tensors transform under

i ···i s ···s
T'j11··· jkr = blj11 · · · bljrr ais11 · · · aiskk Tl11···lr k , (3.1.35)

where aij = ∂ x'i /∂ x j , bij = ∂ xi /∂ x'j .

Now, we consider the general tensors from the viewpoint of vector bundles. A (k, r)
tensor field is a mapping:
T : M → Trk M = M ⊗ p E p , (3.1.36)
122 Chapter 3 Mathematical Foundations

and the bundle space E p is given by

E p = Tp M ⊗ · · · ⊗ Tp M ⊗ Tp∗ M ⊗ · · · ⊗ Tp∗ M . (3.1.37)

4 56 7 4 56 7
k r

In the x-coordinate system, the bases of Tp M and Tp∗ M are:

Tp M : ei = ∂ /∂ xi for 1 ! i ! n,
i i
(3.1.38)
Tp∗ M : e = dx for 1 ! i ! n.

These bases induce a basis of the linear space E p in (3.1.37):

j ··· j
ei11···ikr = ei1 ⊗ · · · ⊗ eik ⊗ e j1 ⊗ · · · ⊗ e jr ,

and the field (3.1.36) can be expressed as

i ···i
T = T j11··· jkr eij11···i
··· jr
k
. (3.1.39)

When we take linear transformations on Tp M and Tp∗ M :

A : Tp M → Tp M , A = (aij ),
(3.1.40)
B : Tp∗ M → Tp∗ M , B = (bij ) = (AT )−1 ,

then the field T of (3.1.36) (i.e. (3.1.39)) will transform in the same fashion as (3.1.35).
The above two ways to define general tensors are equivalent. However, in the second
fashion we replace the coordinate transformation (3.1.39) by (3.1.40). This approach is very
convenient to uniformly treat the transformations in the unified field theory.
Hereafter we list a few typical linear transformations of fiber spaces for various physical
fields.
1. Lorentz transformations. A (k, r) type Lorentz field is a mapping

F : M → M ⊗p EN , (3.1.41)

where the fiber space E pN is

E pN = Tp M ⊗ · · · ⊗ Tp M ⊗ Tp∗ M ⊗ · · · ⊗ Tp∗ M .
4 56 7 4 56 7
k r

When Tp M undergoes a Lorentz transformation

L : Tp M → Tp M , L is a Lorentz matrix, (3.1.42)

the dual space Tp∗ M transforms as

L : Tp∗ M → Tp∗ M ,
' L = (LT )−1 ,
' (3.1.43)
3.1 Basic Concepts 123

This leads to a natural linear transformation for the fiber space E pN , which induces a trans-
formation for the field F in (3.1.41).
2. SU(N) gauge transformation. A set of N Dirac spinor fields Ψ are referred to the
mapping:
Ψ : M → M ⊗ p (C4 )N . (3.1.44)
When we take the bundle space transformation

Ω : (C4p )N → (C4p )N for Ω ∈ SU(N),

then the mapping Ψ of (3.1.44) transforms as

Ψ → ΩΨ for Ω ∈ SU(N).

This is the SU(N) gauge transformation for the Dirac spinors.

3. Spinor transformation. The SU(N) gauge fields are the set of functions

Gaµ (1 ! a ! N 2 − 1) and Ψ = (ψ 1 , · · · , ψ N )T ,

and for each a, Gaµ is a 4-dimensional vector field

Gaµ : M → T M .

Hence, the SU(N) gauge fields (Gaµ , Ψ) are the mapping

8 9
(Gaµ , Ψ) : M → M ⊗ p (Tp M )K × (C4 )N for K = N 2 − 1. (3.1.45)

By the definition of spinors in Subsection 2.2.6, under the Lorentz transformation of

(3.1.42), we have
Gaµ → 'LGaµ for 1 ! a ! N 2 − 1,
(3.1.46)
ψ i → Rψ i for 1 ! i ! N,
L = (LT )−1 , and R is the spinor representation matrix determined by the Lorentz matrix
and '
L, as given by (2.2.67).
Hence, for SU(N) gauge fields (3.1.45), if we take the linear transformations for the
bundle spaces of (3.1.45) as
' : Tp M → Tp M ,
L
R : C4p → C4p ,

then the gauge fields (3.1.45) transform as (3.1.46).

Remark 3.4 The Lorentz transformations (3.1.42) and (3.1.43) are independent of
p ∈ M . Hence, the associated vector bundles in (3.1.41) are trivial. In other words, the
vector bundles corresponding only to the Lorentz transformations are trivial. But, other
vector bundles given above are in general nontrivial.
124 Chapter 3 Mathematical Foundations

3.1.4 Connections and covariant derivatives

In the last subsection, we see that a field

F : M → M ⊗p EN (3.1.47)

undergoes a transformation:
F' = Tp F, p ∈ M, (3.1.48)
if the bundle space E pN undergoes a linear transformation

Tp : E pN → E pN . (3.1.49)

By PLD, for a physical field F as defined by (3.1.47), there is a Lagrangian action

!
L= L (F, DF, · · · , Dm F)dx. (3.1.50)

Physical symmetry principles amount to saying that the action (3.1.50) is invariant under
the transformation (3.1.48). This requires that the derivative DF in (3.1.50) be covariant.
Namely, for (3.1.48) we have
' F' = NDF,
D N is a matrix depends on Tp . (3.1.51)

If the transformation Tp in (3.1.48) is independent of p ∈ M , then

Dµ = ∂µ (∂µ = ∂ /∂ xµ ).

However, if Tp depends on p ∈ M, then we have

∂'µ (Tp F) = Tp ∂'µ F + ∂'µ Tp F, (3.1.52)

which violates the covariance of (3.1.51), because it has a superfluous term ∂'µ Tp F in the
right-hand side of (3.1.52). Hence, the derivative operator D µ must be in the form

Dµ = ∂µ + Γµ , (3.1.53)

such that (3.1.51) holds true. The field Γ µ is the connection of the vector bundle M ⊗ p E N
under the transformation (3.1.49).
To make explicit of Γ' µ , we assume that ∂'µ and ∂µ have the following relation

∂'µ = A∂µ , (3.1.54)

and A is a matrix in the following form

⎧
⎪ identity for SU(N) gauge fields,
⎨
A = Lorentz matrix for Lorentz tensors, (3.1.55)
⎪
⎩
affine matrix for general tensors.
3.1 Basic Concepts 125

By (3.1.48) and (3.1.53)-(3.1.54), we have

' µ F' = (∂'µ + Γ

D ' µ )(Tp F) = ATp ∂µ F + A∂µ Tp F + Γ
' µ Tp F.

Thanks to (3.1.51),

N(∂µ + Γµ )F = ATp ∂µ F + A∂µ Tp F + Γ

' µ Tp F.

It follows that

N = ATp , ' µ = ATp Γµ Tp−1 − A∂µ Tp Tp−1 .

Γ

In other words, under the linear transformation (3.1.49), the system transforms as follows
' F' = (ATp )DF,
D
(3.1.56)
' = ATp ΓTp−1 − A(∂ Tp )Tp−1 .
Γ

The following summarizes the connections of all symmetry transformations:

1. Connection for Lorentz group. As the Lorentz transformation Tp is independent of
p ∈ M , the connections Γ µ are zero:

Γµ = 0 for the Lorentz action.

2. Connection for SU(N) group. For the SU(N) group action, (3.1.49) is

Ω p : (C4 )N → (C4 )N , Ω p ∈ SU(N), p ∈ M.

By (3.1.55), A = I. The connection of SU(N) group is the gauge fields G aµ :

" 2
#
Γµ = igGaµ τa | {τa }Na=1−1 is a set of generators of SU(N) .

Hence, relations (3.1.56) for the SU(N) gauge fields Gaµ are written as

'Ψ
D ' = ΩΨ, Ψ : M → M ⊗ p (C4 )N ,
'aµ τa = ΩGaµ τa Ω−1 − 1 ∂ ΩΩ−1 ,
G Ω ∈ SU(N).
ig

3. Connections for general linear group GL(n). For (k, r)-tensors:

F : M → Trk M , (3.1.57)

the GL(n) group action

Ap : T M → T M , A p ∈ GL(n), (3.1.58)
induces a linear transformation:
126 Chapter 3 Mathematical Foundations

k k
Tp = A p ⊗ · · · ⊗ A p ⊗ A−1 ⊗ · · · ⊗ A−1
p : Tr M → Tr M , (3.1.59)
4 56 7 4p 56 7
k r
which can be equivalently expressed a K × K matrix with K = n k+r , and ⊗ is the tensor
product of matrices defined by (3.1.67); see Remark 3.5. In this case, the matrix A of
(3.1.54)-(3.1.55) is precisely the A p in (3.1.58). Hence, by (3.1.56) we have
' F' = (A ⊗ T )DF,
D (3.1.60)

where F is as in (3.1.57), T is as in (3.1.59), and A is as in (3.1.58).

The covariant derivative operator D depends on the indices k and r of bundle spaces
Trk M, and are derived by induction.
4. Derivative on T M . For F = (F 1 , · · · , F n ),

D j F i = ∂ j F i + Γijl F l , (3.1.61)

where Γijl are connections defined on T M . As M is a Riemannian manifold, {Γijk } are the
Levi-Civita connection as given by (2.3.25). It follows from (3.1.56) that the connection of
(3.1.61) transforms as
' = A ⊗ A ⊗ (AT)−1 Γ − A ⊗ ∂ A ⊗ A−1,
Γ (3.1.62)

which is the the same as those of (2.3.23).

5. Derivative on T ∗ M . Consider F = (F1 , · · · , Fn ). The derivative operators satisfy that
D(A · B) = DA · B + A · DB,
(3.1.63)
D(A ⊗ B) = DA ⊗ B + A ⊗ DB.
It is known that

Fk Gk = a scalar field, and D(Fk Gk ) = ∂ (Fk Gk ). (3.1.64)

We infer then from (3.1.63) and (3.1.64) that

Dk Fi Gi + Fi Dk Gi = ∂k Fi Gi + Fi∂k Gi . (3.1.65)

Inserting (3.1.61) in (3.1.65) we deduce that

j
Dk Fi = ∂k Fi − Γki Fj .

6. Derivative on Trk M . For two vector fields Ak and Bk , their tensor product A ⊗ B =
{Ai B j } is a second order tensor. By (3.1.63) we have

Dk (A ⊗ B) =Dk (Ai B j ) = Dk Ai B j + Ai Dk B j
j
=(∂k Ai + Γikl Al )B j + Ai (∂k B j + Γkl Bl )
=∂k (Ai B j ) + Γikl Al B j + Γklj Ai Bl .
3.2 Analysis on Riemannian Manifolds 127

Replacing A ⊗ B by T i j , we obtain

Dk T i j = ∂k T i j + Γikl T l j + Γklj T il .

In the same fashion, for general (k, r)-tensors

i ···i
T = T j11··· jkr : M → Trk M ,

its covariant derivative can be expressed as

i ···i i ···i li ···i
Dk T j11··· jkr =∂k T j11··· jkl + Γikl1 T j1 2··· jrk + · · · (3.1.66)
i i ···i l i ···i i ···i
+ Γklk T j11··· jk−1
r
− Γlk j1 T j21··· jkr − · · · − Γlk jr T j11··· jkk−1 l .

The derivative (3.1.66) were given in (2.3.26).

Remark 3.5 We have encountered tensor products in (3.1.59) and (3.1.62). A further
explanation of this considered is now in order. Let A and B be two matrices with orders n
and m respectively. Then A ⊗ B is a matrix of order N = nm, defined by
⎛ ⎞ ⎛ ⎞
a11 B · · · a1n B ai j b11 · · · ai j b1m
A ⊗ B = ⎝ ... .. ⎟ , ai j B = ⎝ ... .. ⎟ .
⎜ ⎜
. ⎠ . ⎠ (3.1.67)
an1 B ··· ann B ai j bm1 ··· ai j bmm
i ···i
In (3.1.60) and (3.1.62), the components of DF = {Dk Fj11··· jkr } and Γ = {Γkij } are arranged
to be in two vectorial forms.

3.2 Analysis on Riemannian Manifolds

3.2.1 Sobolev spaces of tensor fields
The gravitational field equations are defined on a Riemannian manifold. To study these
equations, it is necessary to introduce various types of function spaces on manifolds, which
possess different differentiability. In particular, we need to introduce the concept of weak
derivatives.
Let M be an n-dimensional manifold with metric {gi j }. The following functions are
defined on M .
1. L p spaces. For a real number p (1 ! p < ∞), we denote
> !! ?
! √
L p (M ⊗ p E N ) = u : M → M ⊗ p E N ! |u| p −gdx < ∞ , (3.2.1)
M
√
where −gdx is the volume element, and |u| is the modulus of u. For example, if u : M →
T M is a vector field, then
|u| = |gi j ui u j |1/2 .
128 Chapter 3 Mathematical Foundations

The space (3.2.1) is endowed with L p -norm as

0! 11/p
p√
||u||L p = |u| −gdx .
M

By Functional Analysis, the spaces L p (M ⊗ p E N ) are Banach spaces, and for 1 < p < ∞,
L p (M ⊗ p E N ) are reflective and separable. The dual spaces of L p (M ⊗ p E N ) (1 < p < ∞)
are
L p (M ⊗ p E N )∗ = Lq (M ⊗ p (E N )∗ ), 1 1
+ = 1,
q N ∗ p
L (M ⊗ p E ) = L (M ⊗ p (E ) ), N ∗ p q

where (E N )∗ is the dual space of E N .

For p = ∞, we define that

L∞ (M ⊗ p E N ) = {u : M → M ⊗ p E N | u is bounded almost everywhere}.

The norm of L∞ (M ⊗ p E N ) is defined by

||u||L∞ = sup |u|.

M

The space L∞ (M ⊗ p E N ) is a Banach space, but not reflective and separable. The space
L∞ (M ⊗ p E N ) is the dual space of L1 (M ⊗ p E N ), i.e.

L1 (M ⊗ p E N )∗ = L∞ (M ⊗ p (E N )∗ ).

2. Weakly differentiable functions. A field u ∈ L p (M ⊗ p E N ) is called k-th order

weakly differentiable, if each component u j is k-th order weakly differentiable, i.e. for each
u j there exists uniquely a function ϕ such that for all v ∈ C0∞ (M ) we have
! !
√
ϕ v −gdx = (−1)k u j ∂ k vdx. (3.2.2)
M M

In this case, ϕ is called the k-th weak derivative of u j , denoted by

ϕ = ∂ ku j. (3.2.3)

The space C0∞ (M ) consists of infinitely differentiable functions, and

@ ∞
∞
C (M ) if M is compact and ∂ M = ∅,
C0 (M ) = ∞
(3.2.4)
{u ∈ C (M )| u ̸= 0 in a compact set of M }.

The definition (3.2.2)-(3.2.3) for weak derivatives is very abstract. In the following, we
discuss the distinction between continuity and weak differentiability in an intuitive fashion.
3.2 Analysis on Riemannian Manifolds 129

Let u be a function defined on Rn . It is known that if u is differentiable at x = 0, then u

can be Taylor expanded as
u(x) = ax + o(|x|), (3.2.5)
where a = ∇u(0) is the first order derivative of u at x = 0, i.e. the gradient of u at x = 0.
However, if u is weakly but not continuously differentiable at x = 0, then in a neighbor-
hood of x = 0, u must contain at least a term as |xi |α (α ! 1). Without loss of generality, we
express u in the form

u = a|x|α + continuously differentiable terms, (3.2.6)

where a ̸= 0 is a constant, and 1 − n < α ! 1. The index α in (3.2.6) determines the

regularity of u as follows:
u = Lipschitz, if α = 1,
u = Hölder, if 0 < α < 1, (3.2.7)
u = singularity, if α < 1.
Expressions (3.2.5) and (3.2.6) exhibit the essential difference between continuity and
weak differentiability.
Remark 3.6 For a weakly differentiable function as (3.2.6), its index α has to satisfy
1 − n < α ! 1, which is crucial in the Sobolev embedding theorems in the next subsection.

3. W k,p spaces (Sobolev spaces). Let u : M → M ⊗ p E N . Then, u = {u j | 1 ! j ! N},

and each component u j of u is a function on M. We denote
∂ ku j
∂αuj = for α = (α1 , · · · , αn ),
(∂ x1 )α1 · · · (∂ xn )αn
n
and k = |α | = ∑ αi (αi " 0). Then, we define W k,p spaces as
i=1

W k,p (M ⊗ p E N ) = {u ∈ L p (M ⊗ p E N )| ∂ β u j ∈ L p (M ), 1 ! j ! n, 0 ! β ! k}, (3.2.8)

and ∂ β u j in (3.2.8) are weak derivatives of u j .

The spaces W k,p (M ⊗ p E N ) are called Sobolev spaces, and the norms are defined by
k 0! √
11/p
||u||W k,p = ∑ |Dβ u| p −gdx . (3.2.9)
β =0 M

It is clear that

u ∈ W k,p (M ⊗ p E N ) ⇒ u is k-th weakly differentiable.

In addition, we introduce the spaces W0k,p (M ⊗ p E N ) as

130 Chapter 3 Mathematical Foundations

W0k,p (M ⊗ p E N ) = Closure of C0∞ (M ⊗ p E N ) under W k,p norm (3.2.9).

Here C0∞ (M ⊗ p E N ) is as defined in (3.2.4).
If ∂ M = ∅, then W0k,p (M ⊗ p E N ) = W k,p (M ⊗ p E N ), and if ∂ M ̸= ∅ then for u ∈
W0k,p (M ⊗ p E N ) we have

u|∂ M = 0, · · · , ∂ β u|∂ M = 0, ∀|β | ! k − 1.

4. H k spaces. As M is a Riemannian manifold, M ⊗ p E N and its dual bundle M ⊗ p

(E N )∗ are isomorphic. In this case, the spaces W k,2 (M ⊗ p E N ) are Hilbert spaces, denoted
by
H k (M ⊗ p E N ) = W k,2 (M ⊗ p E N ),
(3.2.10)
H0k (M ⊗ p E N ) = W0k,2 (M ⊗ p E N ).
The inner products of (3.2.10) are defined by
! k √
⟨u, v⟩H k = ∑
M |β |=0
Dβ u · Dβ v∗ −gdx

where v∗ ∈ H k (M ⊗ p (E N )∗ ) is the dual field of v ∈ H k (M ⊗ p E N ).

5. Lipschitz spaces. Let k " 0 be integrals. The Lipschitz space C k,1 (M ⊗ p E N ) consists
of all k-th order continuously differentiable functions u with Dk u being Lipschitz continu-
ous:
Ck,1 (M ⊗ p E N ) = {u ∈ Ck (M ⊗ p E N )| [∂ k u]Lip < ∞},
where [∂ k u]Lip is the Lipschitz modulus, defined by

|v(x) − v(y)|
[v]Lip = sup .
x,y∈M , x̸=y |x − y|

A Lipschitz continuous function u ∈ C 0,1 (M ⊗ p E N ) is as shown in (3.2.6)-(3.2.7) with

α = 1.
6. Hölder spaces. The Hölder space C k,α (M ⊗ p E N ) (0 < α < 1) consists of all k-th
order continuously differentiable functions u with Dk u being Hölder continuous:

Ck,α (M ⊗ p E N ) = {u ∈ Ck (M ⊗ p E N )| [Dk u]α < ∞},

and [v]α is the Hölder modulus:

|v(x) − v(y)|
[v]α = sup , 0 < α < 1.
x,y∈M , x̸=y |x − y|α

A Hölder continuous function u ∈ C 0,α (M ⊗ p E N ) is as shown in (3.2.6)–(3.2.7) with 0 <

α < 1.
3.2 Analysis on Riemannian Manifolds 131

The norm of Ck,α (M ⊗ p E N )(0 < α ! 1) is given by

||u||Ck,α = ||u||Ck + [Dk u]α (0 < α ! 1),

where || · ||Ck is the norm of Ck (M ⊗ p E N ):

k
||u||Ck = ∑ sup |∂ β u|.
|β |=0 M

3.2.2 Sobolev embedding theorem

In (3.2.6)–(3.2.7) we see that for Ω ⊂ Rn , a function u ∈ W 1,p (Ω) does’t imply that u ∈
Lq (Ω) for any q > p. For example, for the function
1
u = |x|−α , 0<α < , x ∈ Ω ⊂ R3 , Ω bounded,
2
it is known that
α +4
∇u = |x|− 2 x ∈ L2 (Ω).
Obviously we have
n
u ∈ W 1,2 (Ω), u ̸∈ Lq (Ω), . ∀q >
α
The following embedding problem of Sobolev spaces provides a solution for this prob-
lem.
Theorem 3.7 (Sobolev Embedding Theorem) Let M be an n-dimensional compact
manifold. Then we have the embeddings:
⎧ np
⎪
⎪
⎪ Lq (M ⊗ p E N ) for 1 ! q ! , if n > p,
⎨ n− p
W 1,p (M ⊗ p E N ) "→ Lq (M ⊗ E N ) for 1 ! q < ∞, if n = p, (3.2.11)
⎪ p
⎪
⎪
⎩ 0,α
C (M ⊗ p E N ) for α = 1 − n/p, if n < p.

Here C0,α (M ⊗ p E N ) are the Hölder spaces. Moreover we have the following inequalities
for the norms:
np
||u||Lq ! C||u||W 1,p for 1 ! q ! , if n > p,
n− p
||u|| q ! C||u|| 1,p for 1 ! q < ∞, if n = p, (3.2.12)
L W
||u||C0,α ! C||u||W 1,p for α = 1 − n/p, if n < p,
where C > 0 are constants depending on n, p and M.
Remark 3.8 When M is non-compact, the conclusions (3.2.11) and (3.2.12) are also
valid only for q satisfying
np
p!q! if n > p,
n− p
p!q<∞ if n = p.
132 Chapter 3 Mathematical Foundations

Remark 3.9 By the recurrence relations

W k,p (M ⊗ p E N ) "→ W k−1,q (M ⊗ p E N ),

we readily deduce from (3.2.11) that

⎧ q np
⎪
⎪ L (M ⊗ p E N ) for 1 ! q ! , if n > kp,
⎨ n − kp
W k,p (M ⊗ p E N ) "→ Lq (M ⊗ p E N ) for 1 ! q < ∞, if n = kp,
⎪
⎪
⎩ M ,α
C (M ⊗ p E N ) for m + α = k − n/p, if n < kp.

Based on weakly differentiability properties (3.2.6) and (3.2.7), essence of Theorem 3.7
can be seen from embeddings (3.2.11) using the following function:
|x|α
u(x) = , x ∈ Rn , (3.2.13)
| ln |x||β

where β > 1 is given, and the exponent α < 1 is to reflect the critical embedding index q ∗
in (3.2.11).
The derivatives of u given by (3.2.13) are as follows

∇u = (∂1 u, · · · , ∂n u),
% &
α β (3.2.14)
∂i u = − |x|α −2 xi .
| ln |x||β | ln |x||β +1

Let BR = {x ∈ Rn | |x| < R}, 0 < R < 1. Assume that

u ∈ W 1,p (BR ) for some 1 ! p < ∞.

Then, by (3.2.14) we have

! !
|x|(α −1)p
|∇u| p dx ! C dx. (3.2.15)
BR BR | ln |x||β p
In the spherical coordinate system,

dx = rn−1 drds,

and ds is the area element of the unit sphere, (3.2.15) becomes

! ! R! ! R
rn+(α −1)p−1 rk
|∇u| p dx !C drds ! C dr, (3.2.16)
BR 0 Sn−1 | ln r|β p 0 | ln |r||β p

for 0 < R < 1 and β p > 1. It follows from (3.2.16) that

!
|∇u| p dx < ∞ ⇔ k = n + (α − 1)p − 1 " −1.
BR
3.2 Analysis on Riemannian Manifolds 133

Hence we obtain that

|x|α n
u= ∈ W 1,p (BR ) ⇔ α " 1 − . (3.2.17)
| ln |x||β p

On the other hand, in the same fashion we see that

|x|α n
u= β
∈ Lq (BR ) ⇔ α " − . (3.2.18)
| ln |x|| q

Hence, by (3.2.17) we can see that as p > n,

n
u ∈ C0,α (BR ), α = 1− ,
p

and as p ! n, then from (3.2.17) and (3.2.18), at the critical embedding exponent q ∗ we have
n n
α = 1− , α =− .
p q∗

It follows that ⎧
⎨ = np for n > p,
q ∗ n− p
⎩
<∞ for n = p.
Thus, we deduce that
⎧ np
⎪
⎪
⎪ u ∈ Lq (BR ) for 1 ! q ! , if n > p,
⎪
⎨ n− p
u ∈ W 1,p (BR ) ⇒ u ∈ Lq (BR ) for 1 ! q < ∞, if n = p, (3.2.19)
⎪
⎪
⎪ n
⎩ u ∈ C0,α (BR )
⎪ for α = 1 − , if p > n.
p

The relations (3.2.19) are the embeddings given by (3.2.11).

3.2.3 Differential operators

A differential operator defined on a manifold M is a mapping given by

G : W k,p (M ⊗ p E1N1 ) → L p (M ⊗ p E2N2 ) for some k " 1,

and is called a k-th order differential operator.

The most important operators in physics are:

1) the gradient operator: ∇,

2) the divergent operator: div,

3) the Laplace operator: Dk Dk = div · ∇,

134 Chapter 3 Mathematical Foundations

4) the Laplace-Beltrami operator: ∆ = d δ + δ d,

5) the wave operator: # = D µ Dµ , with Dµ being the 4-D gradient operator.

We now give a detailed account on the above operators.

1. Gradient operator. The gradient operator ∇ is a mapping of the following types:

∇k : W 1,p (Trk M ) → L p (Trk+1 M ),

(3.2.20)
∇k : W 1,p (Trk M ) → L p (Tr+1
k
M ),

and ∇k and ∇k have the relation

∇k = gkl ∇l , ∇k = gkl ∇l ,

and {gkl } the Riemann metric of M . ∇ is expressed as

∇k = (D1 , · · · , Dn ),
(3.2.21)
D j the covariant derivative operators.

2. Divergence operator. The divergence operator div is a mapping of the following

types:
div : W 1,p (Trk+1 M ) → L p (Trk M ),
(3.2.22)
div : W 1,p (Tr+1
k
M ) → L p (Trk M ).
i ···i i ···i
For T = {T j11··· jk+1
r
} ∈ W 1,p (Trk+1 M ) and T = {T j11··· jkr+1 } ∈ W 1,p (Tr+1
k M),

i ···i ···ik+1
div T = Dil T j11··· jlr , and
i ···i
(3.2.23)
jl
div T = D T j11··· jkl ··· jk+1

As u ∈ W 1,p (T M ) and u ∈ W 1,p (T ∗ M ), we can give the expressions of div u in the

following.
Let u ∈ W 1,p (T M ), u = (u1 , · · · , un ). Then by (3.2.23),

∂ uk
div u = Dk uk = + Γkk j u j .
∂ xk
By the Levi-Civita connections (2.3.25), the contraction
√
1 ∂ gkl 1 ∂g 1 ∂ −g
Γkk j = gkl = = √ ,
2 ∂xj 2g ∂ x j −g ∂ x j

here g = det (gi j ). Thus we have

√ √
∂ uk 1 ∂ −g k 1 ∂ ( −guk )
div u = + √ u = √ . (3.2.24)
∂ xk −g ∂ xk −g ∂ xk
3.2 Analysis on Riemannian Manifolds 135

Let u ∈ W 1,p (T ∗ M ), u = (u1 , · · · , un ). Then,

div u =Dk uk = gkl Dl (uk ) = Dl (glk uk )

1 ∂ √
=√ ( −gglk uk ) (by (3.2.24)). (3.2.25)
−g ∂ xl

The formula (3.2.24) and (3.2.25) give expression of div u as:

⎧
⎪ 1 ∂ √
⎪
⎨ √ ( −guk ) foru ∈ W 1,p (T M ),
−g ∂ xk
div u = 1 ∂ √ (3.2.26)
⎪
⎪
⎩ √ ( −ggkl ul ) foru ∈ W 1,p (T ∗ M ).
−g ∂ xk

3. Laplace operators. The Laplace operator Dk Dk in Rn is in the familiar form

∂ ∂
Dk Dk = .
∂ xk ∂ xk

However, the Laplace operators Dk Dk defined on a Riemannian manifold are usually very
complex.
By (3.2.20) and (3.2.22), we have

∇ div
W 2,p (Trk M ) →k W 1,p (Tr+1
k
M ) → L p (Trk M ),
∇k div
W 2,p (Trk M ) → W 1,p (Trk+1 M ) → L p (Trk M ).

Hence, the Laplace operator div · ∇ is the mapping:

div · ∇ : W 2,p (Trk M ) → L p (Trk M ). (3.2.27)

By (3.2.21) and (3.2.23), div · ∇ is written as

div · ∇ = Dk Dk = gkl Dk Dl .

4. Expression of div · ∇ on M ⊗ p R1 . A scalar field u on M can be regarded as u : M →

M ⊗ p R1 . In this case, ∇u is written as
% &
∂u ∂u
∇u = ,··· , n .
∂x 1 ∂x

By (3.2.26), we get that

% &
1 ∂ √ ∂u
div · ∇u = √ −ggkl l . (3.2.28)
−g ∂ xk ∂x
136 Chapter 3 Mathematical Foundations

5. Expression of div · ∇ on T M . A vector field u : M → T M can be written as

u = (u1 , · · · , un ),

and its gradient (i.e. covariant derivatives) read as

∇u = {Dk ui },
∂ ui (3.2.29)
Dk u i = + Γikl ul ,
∂ xk
and the divergence of ∇u is
0 1
i k i kl i kl ∂ i i j j i
div · ∇u =D Dk u = g Dl (Dk u ) = g (Dk u ) + Γl j Dk u − Γkl D j u .
∂ xl
Then, by (3.2.29) we obtain that
0 % i & % j &
∂ ∂u ∂u j s
div · ∇u = gkl + Γ i
kj u j
+ Γ i
lj + Γ ks u
∂ xl ∂ xk ∂ xk
% i &1
j ∂u
−Γkl + Γijs us . (3.2.30)
∂xj

6. Laplace-Beltrami operators. The Laplace-Beltrami operators are defined by

∆ = d δ + δ d,

where d is the differential operator, and δ the Hodge operator, which are defined on the
spaces of all differential forms. For a scalar field, ∆ is the same as div · ∇, i.e.

∆u = div · ∇u, as u ∈ W 2,p (M ⊗ p R1 ).

For vector fields and covector fields, we have

∆ui = −div(∇ui ) − gi j R jk uk ,
(3.2.31)
∆ui = −div(∇u j ) − gk j Ri j uk ,
where Ri j is the Ricci curvature tensor defined by (2.3.31), and in (3.2.30) we give the
formula of div (∇uk ). The expression of div (∇ui ) is written as
> 0 1 0 1 0 1?
∂ ∂ ui j j ∂ ui j ∂uj
div (∇ui ) =gkl − Γ ik u j − Γ lk ∂ x j − Γ r
ij u r − Γ li ∂ xk − Γ r
jk u r . (3.2.32)
∂ xl ∂ xk
Physically, it suffices to introduce the Laplace-Beltrami operators ∆ only for vector and
covector fields, i.e. the formula in (3.2.31).
Remark 3.10 The Navier-Stokes equations governing a fluid flow on a sphere (e.g.
the surface of a planet) are expressed in the spherical coordinate system, and the viscosity
differential operators are the Laplace-Beltrami operators given by (3.2.31); see Chapter 7
for details.
3.2 Analysis on Riemannian Manifolds 137

3.2.4 Gauss formula

If M = Rn , it is known that the Gauss formula is

! !
u · ∇ f dx = − f div udx, ∀u ∈ H 1 (Rn × Rn ), f ∈ H 1 (Rn ),
Rn Rn

which can be generalized to tensor fields on Riemann manifolds.

Let M be a Riemannian manifold with ∂ M = ∅. Then, there is an inner product
defined on T M defined by

(X,Y ) = gi j X i X j , ∀X,Y ∈ T M .

Also, there is an inner product on T ∗ M:

(X ∗ ,Y ∗ ) = gi j Xi∗Y j∗ , ∀X ∗ ,Y ∗ ∈ T ∗ M .

The tensor bundle Trk M is

Trk M = T · · ⊗ T M7 ⊗ T
4 M ⊗ ·56
∗
· · ⊗ T ∗ M7 .
4 M ⊗ ·56
k r

The inner products on T M and T ∗ M induce a natural inner product on Trk M :

i ···i
(u, v) = gi1 j1 · · · gik jk gl1 s1 · · · glr sr ul11 ···lrk vsj11···s
··· jr
r
, ∀u, v ∈ Trk M . (3.2.33)

Hence we can define the inner product ⟨·, ·⟩ on L2 (Trk M ) by

!
√
⟨u, v⟩L2 = (u, v) −gdx, ∀u, v ∈ L2 (Trk M ), (3.2.34)
M

where (u, v) is as in (3.2.33).

Furthermore, we can also define an inner product on the spaces H m (Trk M ) as follows
!
√
⟨u, v⟩H m = [(−1)m (∆m u, v) + (u, v)] −gdx, ∀u, v ∈ H m (Trk M ), (3.2.35)
M

where ∆ = div · ∇.
In the following, we give the Gauss formula on Trk M , which are crucial for the orthog-
onal decomposition theory in the next section.

Theorem 3.11 (Gauss Formula) For any u ∈ H 1 (Trk M ) and v ∈ H 1 (Trk+1 M ) (or
v ∈ H 1 (Tr+1
k M )),
! !
√ √
(∇u, v) −gdx = − (u, divv) −gdx. (3.2.36)
M M
138 Chapter 3 Mathematical Foundations

Formula (3.2.36) is a corollary of the classical Gauss formula

! !
√
div w −gdx = w · nds. (3.2.37)
M ∂M

In fact, let wk = (u, vk ), then by (2.3.28) we have

div w = Dk (u, vk ) = (∇u, v) + (u, div v).

If ∂ M = ∅, then we derive from (3.2.37) that

! !
√ √
div w −gdx = [(∇u, v) + (u, div v)] −gdx = 0,
M M

which is the formula (3.2.36).

The Gauss formula (3.2.36) can be generalized to more general gradient and divergent
operators, denoted by DA and divA .
Let A be a vector field or a covector field, and u ∈ L2 (Trk M ). We define the operators
DA and divA by
DA u = Du + u ⊗ A,
(3.2.38)
divA u = divu − u · A.
Based on (3.2.36), it is readily to verify that the following formula holds true for the opera-
tors (3.2.38). ! !
√ √
(DA u, v) −gdx = − (u, divA v) −gdx. (3.2.39)
M M

Remark 3.12 The motivation to generalize the Gauss formulas (3.2.36) to the opera-
tors DA and divA is to develop a new unified field theory for the fundamental interactions.
The vector fields A in (3.2.38) and (3.2.39) represent gauge fields in the interaction field
equations, and lead to a mass generation mechanism based on a first principle, called PID,
different from the famous Higgs mechanism.

3.2.5 Partial differential equations on Riemannian manifolds

To develop an orthogonal decomposition theory for general (k, r)-tensor fields, we need
to introduce the existence theorems for linear elliptic and hyperbolic equations on closed
Riemann and Minkowski manifolds. The existence results are well-known. In the following,
we give the definition of weak solutions and the basic existence theorems for PDEs without
proofs.

Linear elliptic equations

Consider the following PDEs defined on a Riemannian manifold {M , g i j } with ∂ M =

∅:
gi j Di D j u = div f + g, (3.2.40)
3.2 Analysis on Riemannian Manifolds 139

where Di are the covariant derivative operators, the unknown function u : M → Trk M is a
(k, r) tensor field, g : M → Trk M and f : M → Trk+1 M (or f : M → Tr+1k M ) are given.

We need to introduce the concept of weak solutions for (3.2.40).

Definition 3.13 Let f ∈ L2 (Trk+1 M )(or f ∈ L2 (Tr+1 k M )) and g ∈ L2 (T k M ). A field

r
u ∈ H (Tr M ) is called a weak solution of (3.2.40), if for all v ∈ H 1 (Trk M ) the following
1 k

equality holds true,

! !
√ √
(∇u, ∇v) −gdx = [( f , ∇v) − (g, v)] −gdx,
M M

where (·, ·) is the inner product as defined in (3.2.33).

The following existence theorem is a classical result, which is a corollary of the well-
known Fredholm Alternative Theorem.

Theorem 3.14 Let the metric gi j be W 2,∞ , and f , g be L2 . If

!
√
[( f , ∇φ ) − (g, φ )] −gdx = 0 (3.2.41)
M

holds for all φ satisfying

gi j Di D j φ = 0, (3.2.42)
then the equation (3.2.40) possesses a weak solution u ∈ H 1 (Trk M ). In particular, if gi j , f , g
are C∞ , then u ∈ C∞ (Trk M ).

Two remarks are now in order. First, the solutions φ ̸= 0 of (3.2.42) are the eigenfunc-
tions of the Laplace operator div · ∇ = Di Di corresponding to the zero eigenvalue λ = 0, and
the condition (3.2.41) represents that div f + g is orthogonal to all eigenfunctions for λ = 0
of div · ∇.
Second, by the L p -estimate theorem, if gi j ∈ W m+2,p , f ∈ W m+1,p , and g ∈ W m,p , then
u ∈ W m,p (Trk M ). Therefore, by the Sobolev Embedding Theorem 3.7, if g i j , f , g are C∞ ,
then the solution u is also C ∞ .
Linear hyperbolic equations
Let {M , g µν } be a Minkowski manifold, i.e., its metric g µν can be written in some
coordinate system as % &
−1 0
(g µν ) = , G = (gi j ) (3.2.43)
0 G
where G is an (n − 1) × (n − 1) positive definite symmetric matrix. Then, the Laplace
operator
∂2
Dµ Dµ = g µν Dµ Dν = − 2 + gi j Di D j
∂t
is a hyperbolic operator.
140 Chapter 3 Mathematical Foundations

Let M = S1 × M A with the metric (3.2.43), and M A is a Riemannian manifold with

metric {gi j } as in (3.2.43), and ∂ MA= ∅. Then, for the Minkowski manifold S 1 × M A, the
equation
gµν Dµ Dν u = div f , (3.2.44)
is a hyperbolic equation, written as

∂ 2u
− + gi j Di D j u = div f , (3.2.45)
∂ t2
with the periodic condition

u(t + T ) = u(t), ∀t ∈ R1 . (3.2.46)

The following existence theorem is classical.

Theorem 3.15 Let M = S1 × M Ais a Minkowski manifold with metric (3.2.43), and
∂ M = ∅. Assume that g µν and f are C∞ , then the problem (3.2.45)-(3.2.46) has a C ∞
A
solution u.

Remark 3.16 If M = R1 × M Awith metric (3.2.43) and ∂ M ' = ∅, then the problem
(3.2.45)-(3.2.46) is replaced by the following initial value problem:

∂ 2u
− + gi j Di D j u = div f ,
∂ t2
(3.2.47)
u(0) = ϕ ,
ut (0) = ψ .

The same existence theorem holds true as well for problem (3.2.47).

3.3 Orthogonal Decomposition for Tensor Fields

3.3.1 Introduction
Let M be a Riemannian manifold (or a Minkowski manifold), ∂ M = ∅, and u is a tensor
field on M :
u : M → Trk M . (3.3.1)
The orthogonal decomposition of tensor fields means that the field u given by (3.3.1) can be
decomposed as
u = ∇φ + v and div v = 0, (3.3.2)
for some φ : M → Trk−1 M (or φ : M → Tr−1
k M ). Moreover ∇φ and v are orthogonal in

the following sense: !

√
⟨∇φ , v⟩ = (∇φ , v) −gdx = 0. (3.3.3)
M
3.3 Orthogonal Decomposition for Tensor Fields 141

In this section, we shall show that all tensor fields as given by (3.3.1) can be decomposed
into the form (3.3.2) satisfying (3.3.3).
In order to understand the problem well, we first introduce some classical results: the
Helmholtz decomposition and the Leray decomposition on M = R n .

1. Helmholtz decomposition. Let u ∈ L2 (T R3 ) be a 3-dimensional vector field, i.e.

u(x) = (u1 (x), u2 (x), u3 (x)) for x ∈ R3 ,

then there exist a function φ ∈ H 1 (R3 ) and a vector field A ∈ H 1 (T R3 ), such that u can be
decomposed as
u = ∇φ + curl A,
!
(3.3.4)
∇φ · curl Adx = 0.
R3

Note that div(curl A) = 0. Hence, the Helmholtz decomposition is an important initial result
on orthogonal decompositions.

2. Leray decomposition. Let Ω ∈ Rn be a domain, and u ∈ L2 (T Ω) be an n-dimensional

vector field. Then u can be decomposed as

u = ∇φ + v,
v · n|∂ Ω = 0, div v = 0, φ ∈ H 1 (Ω), (3.3.5)
!
∇φ · vdx = 0.
Ω

The Leray decomposition (3.3.5) is crucial in fluid dynamics.

The decompositions (3.3.4) and (3.3.5) can be generalized to more general tensor fields
as shown in (3.3.1)-(3.3.3). Now we discuss the simplest case to illustrate the main idea.
Let u : Rn → T Rn be a given vector field. Then div u is a known function. It is known
that the Poisson equation
∆φ = div u for x ∈ Rn (3.3.6)

has a weak solution φ ∈ H 1 (Rn ), enjoying

!
[∇φ − u] · ∇ϕ dx = 0 ∀ϕ ∈ H 1 (Rn ). (3.3.7)
Rn

Let v = u − ∇φ . Then, by (3.3.7) we have

!
v · ∇φ dx = 0,
R

which means that div v = 0. Thus we obtain the orthogonal decomposition u = ∇φ + v with
div v = 0.
142 Chapter 3 Mathematical Foundations

3.3.2 Orthogonal decomposition theorems

The aim of this subsection is to derive an orthogonal decomposition for (k, r)-tensor fields,
with k + r " 1, into divergence-free and gradient parts. This decomposition plays a crucial
role for the unified field theory coupling four fundamental interactions to be introduced in
Chapter 4 of this book.
Let M be a closed Riemannian manifold or M = S 1 × M A be a closed Minkowski
2 k
manifold with metric (3.2.43), and v ∈ L (Tr M ) (k + r " 1). We say that v is divA −free,
denoted by divA v = 0, if
!
√
(∇A ψ , v) −gdx = 0, ∀∇A ψ ∈ L2 (Trk M ). (3.3.8)
M

Here ψ ∈ H 1 (Trk−1 M ) or H 1 (Tr−1

k M ), ∇ and div are as in (3.2.38).
A A
We remark that if v ∈ H 1 (Trk M ) satisfies (3.3.8), then v is weakly differentiable, and
div v = 0 in L2 -sense. If v ∈ L2 (Trk M ) is not differentiable, then (3.3.8) means that v is
divA -free in the distribution sense.
Theorem 3.17 (Orthogonal Decomposition Theorem) Let A be a given vector field or
covector field, and u ∈ L2 (Trk M ). Then the following assertions hold true:

1) The tensor field u can be orthogonally decomposed into

u = ∇A ϕ + v with divA v = 0, (3.3.9)

where ϕ ∈ H 1 (Trk−1 M ) or ϕ ∈ H 1 (Tr−1

k M ).

2) If M is a compact Riemannian manifold, then u can be orthogonally decomposed

into
u = ∇A ϕ + v + h, (3.3.10)
where ϕ and v are as in (3.3.9), and h is a harmonic field, i.e.

divA h = 0, ∇A h = 0.

In particular, the subspace of all harmonic tensor fields in L2 (Trk M ) is of finite di-
mensional:
H(Trk M ) = {h ∈ L2 (Trk M )| ∇A h = 0, divA h = 0}, and
(3.3.11)
dim H(Trk M ) < ∞.

Remark 3.18 The above orthogonal decomposition theorem implies that L 2 (Trk M )
can be decomposed into

L2 (Trk M ) = G(Trk M ) ⊕ L2D (Trk M ) for general case,

2
L (Trk M ) = G(Trk M ) ⊕ H(Trk M ) ⊕ L2N (Trk M ) for M compact Riemainnian.
(3.3.12)
3.3 Orthogonal Decomposition for Tensor Fields 143

Here H is as in (3.3.11), and

G(Trk M ) = {v ∈ L2 (Trk M )| v = ∇A ϕ , ϕ ∈ H 1 (Tr−1

k
M )},
L2D (Trk M ) = {v ∈ L2 (Trk M )| divA v = 0},
L2N (Trk M ) = {v ∈ L2D (Trk M )| ∇A v ̸= 0}.

They are orthogonal to each other:

L2D (Trk M )⊥G(Trk M ), L2N (Trk M )⊥H(Trk M ), G(Trk M )⊥H(Trk M ).

Remark 3.19 The orthogonal decomposition (3.3.12) of L2 (Trk M ) implies that if a

tensor field u ∈ L2 (Trk M ) satisfies that
!
√
⟨u, v⟩L2 = (u, v) −gdx = 0, ∀divA v = 0,
M

then u must be a gradient field, i.e.

u = ∇A ϕ for some ϕ ∈ H 1 (Trk−1 M ) or H 1 (Tr−1

k
M ).

Likewise, if u ∈ L2 (Trk M ) satisfies that

⟨u, v⟩L2 = 0, ∀v ∈ G(Trk M ),

then u ∈ L2D (Trk M ). It is the reason why we define a divA -free field by (3.3.8).

Proof of Theorem 3.17 We proceed in several steps as follows.

S TEP 1. P ROOF OF A SSERTION (1). Let u ∈ L2 (E), E = Trk M (k + r " 1). Consider
the equation
∆ϕ = divA u in M , (3.3.13)
where ∆ is the Laplace operator defined by

∆ = divA · ∇A . (3.3.14)

Without loss of generality, we only consider the case where divA u ∈ E' = Trk−1 M . It is
clear that if (3.3.13) has a solution ϕ ∈ H 1 (E),
' then by (3.3.14), the following vector field
must be divA -free
v = u − ∇Aϕ ∈ L2 (E). (3.3.15)
Moreover, by (3.3.8), we have
!
√
⟨v, ∇A ψ ⟩L2 = (v, ∇A ψ ) −gdx = 0, ∀∇A ψ ∈ L2 (Trk M ). (3.3.16)
M

Namely v and ∇A ϕ are orthogonal. Therefore, the orthogonal decomposition u = v + ∇ A ϕ

follows from (3.3.15) and (3.3.16).
144 Chapter 3 Mathematical Foundations

It suffices then to prove that (3.3.13) has a weak solution ϕ ∈ H 1 (E):

'

⟨∇A ϕ − u, ∇A ψ ⟩L2 = 0, ∀ψ ∈ H 1 (E).

' (3.3.17)

Obviously, if φ satisfies
∆φ = 0, (3.3.18)
where ∆ is as in (3.3.14), then, by (3.2.39),
! !
√ √
(∆φ , φ ) −gdx = − (∇A φ , ∇A φ ) −gdx = 0.
M M

Hence (3.3.18) is equivalent to

∇A φ = 0. (3.3.19)
Therefore, for all φ satisfying (3.3.18) we have
!
√
(u, ∇A φ ) −gdx = 0.
M

By Theorem 3.14, we derive that the equation (3.3.13) has a unique weak solution ϕ ∈
H 1 (E).
'
For Minkowski manifolds, by Theorem 3.15, the equation (3.3.13) also has a solution.
Thus Assertion (1) is proved.
S TEP 2. P ROOF OF A SSERTION (2). Based on Assertion (1), we have

H k (E) = HDk ⊕ Gk , L2 (E) = L2D ⊕ G,

where

HDk = {u ∈ H k (E)| divA u = 0},

Gk = {u ∈ H k (E)| u = ∇A ψ }.

' : H 2 (E) → L2 (E) by

Define an operator ∆ D D

∆u
' = P∆u, (3.3.20)

where P : L2 (E) → L2D (E) is the canonical orthogonal projection.

We known that the Laplace operator ∆ can be expressed as

∂2
∆ = divA · ∇A = gkl + B, (3.3.21)
∂ xk ∂ xl
where B is a lower-order differential operator. Since M is compact, the Sobolev embeddings

H 2 (E) "→ H 1 (E) "→ L2 (E)

3.3 Orthogonal Decomposition for Tensor Fields 145

are compact. Hence the lower-order differential operator

B : H 2 (M , RN ) → L2 (M , RN )

is a linear compact operator. Therefore the operator in (3.3.21) is a linear completely con-
tinuous field
∆ : H 2 (E) → L2 (E),
which implies that the operator of (3.3.20) is also a linear completely continuous field

' = P∆ : HD2 (E) → L2D (E).

∆

By the spectrum theorem of completely continuous fields (Ma and Wang, 2005), the space

' = {u ∈ HD2 (E)| ∆u

H ' = 0}

is finite dimensional, and is the eigenspace of the eigenvalue λ = 0. By (3.2.39), for u ∈ H

'
we have
! !
' u)√−gdx =
(∆u,
√
(∆u, u) −gdx (by divA u = 0)
M M
!
√
=− (∇A u, ∇A u) −gdx
M
=0 (by ∆u
' = 0).

It follows that
' ⇔ ∇A u = 0,
u∈H
' is the same as the harmonic space H of (3.3.11), i.e. H
which implies that H ' = H. Thus we
have

L2D (E) = H ⊕ L2N (E),

L2N (E) = {u ∈ L2D (E)| ∇A u ̸= 0}.

The proof of Theorem 3.17 is complete.

3.3.3 Uniqueness of orthogonal decompositions

In this subsection we only consider the case where M is a closed manifold with zero first
Betti number.
In Theorem 3.17, a tensor field u ∈ L2 (Trk M ) with k + r " 1 can be orthogonally de-
composed into

u = ∇ϕ + v for general closed manifolds,

(3.3.22)
u = ∇ϕ + v + h for compact Riemannian manifolds.
146 Chapter 3 Mathematical Foundations

Now we address the uniqueness problem of the decomposition (3.3.22). In fact, if u is a

vector field or a covector field:

u ∈ L2 (T M ) or u ∈ L2 (T ∗ M ),

then the decomposition (3.3.22) is unique.

We see that if u ∈ L2 (Trk M ) with k + r " 2, then there are different types of decompo-
sitions of (3.3.22). For example, for u ∈ L2 (T20 M ), in a local coordinate system, u is given
by
u = {ui j (x)}.
In this case, u admits two types of decompositions

u i j = Di ϕ j + v i j , Di vi j = 0, (3.3.23)
j
ui j = D j ψi + wi j , D wi j = 0. (3.3.24)

It is easy to see that if ui j ̸= u ji , then (3.3.23) and (3.3.24) can be two different decomposi-
tion of ui j . Namely

{vi j } ̸= {wi j }, (ϕ1 , · · · , ϕn ) ̸= (ψ1 , · · · , ψn ).

The reason is that the two differential equations generating the two decompositions (3.3.23)
and (3.3.24) as
Di Di ϕ j = Di ui j and Di Di ψ j = Di u ji (3.3.25)
are different because Di ui j ̸= Di u ji .
However for a symmetric tensor field ui j = u ji , as

Di ui j = Di u ji ,

the two equations in (3.3.25) are the same. By the uniqueness of solutions of (3.3.25), the
two solutions ϕ j and ψ j are the same:

ϕi = ψi for 1 ! i ! n.

Thus (3.3.23) and (3.3.24) can be expressed as

u i j = Di ϕ j + v i j , Di vi j = 0, (3.3.26)
j
ui j = D j ϕi + wi j , D wi j = 0. (3.3.27)

From (3.3.26) and (3.3.27) we can deduce the following theorem.

Theorem 3.20 Let u ∈ L2 (T20 M ) be symmetric, i.e. ui j = u ji , and the first Betti number
β1 (M ) = 0 for M . Then the following assertions hold true:
3.3 Orthogonal Decomposition for Tensor Fields 147

1) u has a unique orthogonal decomposition if and only if there is a scalar function

ϕ ∈ H 2 (M ) such that u can be expressed as

u i j = v i j + Di D j ϕ ,
(3.3.28)
vi j = v i j , Di vi j = 0.

2) If vi j in (3.3.26) is symmetric: vi j = v ji , then u can be expressed by (3.3.28).

3) u can be orthogonally decomposed in the form (3.3.28) if and only if the follow-
ing differential equations have a solution ϕ ∈ H 2 (M ), and ϕ is the scalar field in
(3.3.28):
∂ ∂ϕ
∆ϕ + Rki k = −D j u ji for 1 ! i ! n, (3.3.29)
∂ xi ∂x
where Rki = gk j Ri j and Ri j are the Ricci curvature tensors, and ∆ is the Laplace
operator for scalar fields as defined by (3.2.28).

Proof We only need to prove Assertions (2) and (3).

We first prove Assertion (2). Since vi j in (3.3.26) is symmetric, then we have

Di ϕ j = D j ϕi . (3.3.30)

Note that
∂ϕj
Di ϕ j = − Γkij ϕk ,
∂ xi
and Γkij = Γkji . We infer then from (3.3.30) that

∂ϕj ∂ ϕi
= . (3.3.31)
∂x i ∂xj
By assumption, the 1-dimensional homology of M is zero,

H1 (M ) = 0,

and by the de Rham theorem (Ma, 2010), it follows that all closed 1-forms are complete
differentials, i.e. for any
ω = ψi dxi , d ω = 0,
there is a scalar function ψ such that

ψi = ∂ ψ / ∂ x i for 1 ! i ! n.

In view of (3.3.31), it implies that the 1-form

ω = ϕk dxk

is closed. Hence it follows that there is a ϕ ∈ H 1 (M ) such that

148 Chapter 3 Mathematical Foundations

∂ϕ
ϕk = for 1 ! k ! n.
∂ xk
Assertion (2) is proved.

Now we prove Assertion (3). Taking the divergence on both sides of (3.3.26), we obtain
that
Di Di ϕ j = D i u i j . (3.3.32)

By the Laplace-Beltrami operator in (3.2.31),

' ϕ j = Di Di ϕ j + Rk ϕk ,
−∆ (3.3.33)
j

' = δ d + d δ . By the Hodge theory, for ω = ϕi dxi we have

where ∆

dω = 0 ⇔ ϕi = ∂ ϕ / ∂ x i ,
(3.3.34)
δ ω = ∆ϕ ⇔ ϕi = ∂ ϕ / ∂ x i .

Here ∇ is the gradient operator, and ∆ the Laplace operator as in (3.2.28). It follows from
(3.3.34) that
'ω = (δ d + d δ )ω = d δ ω ⇔ ϕi = ∂ ϕ /∂ xi ,
∆

and
∂
dδ ω = (∆ϕ )dxi .
∂ xi
Hence we deduce from (3.3.33) that

∂ ∂ϕ
Di D j ϕ j = − (∆ϕ ) − Rkj k ⇔ ϕi = ∂ ϕ /∂ xi . (3.3.35)
∂x j ∂x
Inserting (3.3.32) in (3.3.35) we obtain that the equations

∂ ∂ϕ
(∆ϕ ) + Rkj k = −Di ui j
∂xj ∂x

have a solution ϕ if and only if ϕ j in (3.3.26) is a gradient field of ϕ , i.e. ϕ j = ∂ ϕ /∂ x j .

Assertion (3) is proven, and the proof of the theorem is complete.

Remark 3.21 The conclusions of Theorem 3.20 are also valid for second-order contra-
variant symmetric tensors u = {ui j }, and the decomposition is given as follows:

ui j = vi j + gik g jl Dk Dl ϕ
Di vi j = 0, vi j = v ji , ϕ ∈ H 2 (M ).
3.3 Orthogonal Decomposition for Tensor Fields 149

3.3.4 Orthogonal decomposition on manifolds with boundary

In the above subsections, we mainly consider the orthogonal decomposition of tensor fields
on the closed Riemannian and Minkowski manifolds. In this section we discuss the problem
on manifolds with boundary.
1. Orthogonal decomposition on Riemannian manifolds with boundaries. The Leray
decomposition (3.3.5) is for the vector fields on a domain Ω ⊂ R n with ∂ Ω ̸= ∅. This result
can be also generalized to general (k, r)-tensor fields defined on manifolds with boundaries.
Theorem 3.22 Let M be a Riemannian manifold with boundary ∂ M ̸= ∅, and

u : M → Trk M (3.3.36)

be a (k, r)-tensor field. Then we have the following orthogonal decomposition:

u = ∇A ϕ + v, (3.3.37)
!
√
divA v = 0, v · n|∂ M = 0, (∇A ϕ , v) −gdx = 0,
M

where ∂ v/∂ n = ∇A v · n is the derivative of v in the direction of outward normal vector n on

∂ Ω.

Proof For the tensor field u in (3.3.36), consider

divA · ∇A ϕ = divA u, ∀x ∈ M ,
∂ϕ (3.3.38)
= u · n, ∀x ∈ ∂ M .
∂n
This Neumann boundary problem possesses a solution provided the following condition
holds true: ! !
∂ϕ
ds = u · nds, (3.3.39)
∂M ∂n ∂M
which is ensured by the boundary condition in (3.3.38). Hence by (3.3.38) the field

v = u − ∇A ϕ (3.3.40)

is divA -free, and satisfies the boundary condition

v · n|∂ M = 0. (3.3.41)

Then it follows from (3.3.40) and (3.3.41) that the tensor field u in (3.3.36) can be orthogo-
nally decomposed into the form of (3.3.37). The proof is complete.

2. Orthogonal decomposition on Minkowski manifolds. Let M be a Minkowski mani-

fold in the form
M =M A× (0, T ), (3.3.42)
150 Chapter 3 Mathematical Foundations

with the metric % &

−1 0
(g µν ) = . (3.3.43)
0 G

Here M Ais a closed Riemannian manifold, and G = (gi j ) is the Riemann metric of M A.
In view of the Minkowski metric (3.3.43), we see that the operator divA · ∇A is a hyper-
bolic differential operator expressed as
% &2
∂
divA · ∇A = − + A0 + gi j DAi DA j . (3.3.44)
∂t

Now a tensor field u ∈ L2 (Trk M ) has an orthogonal composition if the following hyper-
bolic equation
divA · ∇A ϕ = divA u, in M (3.3.45)
has a weak solution ϕ ∈ H 1 (Trk−1 M ) in the following sense:
! !
√ √
(DA ϕ , DA ψ ) −gdx = (u, DA ψ ) −gdx, ∀ψ ∈ H 1 (Trk−1 M ). (3.3.46)
M M

Theorem 3.23 Let M be a Minkowski manifold as defined by (3.3.42)-(3.3.43), and

u ∈ L2 (Trl M ) (k + r " 1) be an (k, r)-tensor field. Then u can be orthogonally decomposed
into the following form
u = ∇A ϕ + v, divA v = 0,
! (3.3.47)
√
(∇A ϕ , v) −gdx = 0,
M

if and only if equation (3.3.45) has a weak solution ϕ ∈ H 1 (T k−1 M ) in the sense of (3.3.46).

3.4 Variations with divA -Free Constraints

3.4.1 Classical variational principle
Variational approach originates from the minimization problem of a functional. Let X be a
Banach space, and F be a functional on X:

F : X → R1 . (3.4.1)

The minimization problem of F is to find a point u ∈ X, which is a minimal point of F.

Namely, there is a neighborhood U ⊂ X of u, such that F is minimal at u in U:

F(u) = min F(v). (3.4.2)

v∈U

In the classical variational principle we know that the minimal point u of F in (3.4.2) is
a solution of the variational equation of F:

δ F(u) = 0, (3.4.3)
3.4 Variations with divA -Free Constraints 151

where δ F is the derivative operator of F.

Given a variational problem, it is important to compute δ F for a given functional F.
Hereafter we give a brief introduction of general methods to compute the derivative opera-
tors from F.
Let F be the functional given by (3.4.1), and X ∗ be the dual space of X. The derivative
operator δ F(u) of F at u ∈ X is a linear functional on X, i.e. δ F(u) ∈ X ∗ for each u ∈ X. In
other words, δ F is a mapping from X to X ∗ :

δ F : X → X ∗. (3.4.4)

Denote ⟨·, ·⟩ the product between X and X ∗ , i.e.

⟨·, ·⟩ : X × X ∗ → R.

Then the derivative operator δ F in (3.4.4) satisfies the relation:

d !!
⟨δ F(u), v⟩ = ! F(u + λ v) for u, v ∈ X, (3.4.5)
d λ λ =0
where λ ∈ R1 is real number.
Based on (3.4.5), it is easy to see that the minimal point u satisfying (3.4.2) is a solution
of the variational equation (3.4.3). In fact, by (3.4.2) for any given v ∈ X the function
f (λ ) = F(u + λ v) is minimal at λ = 0:

d f (0) d !!
=0⇒ ! F(u + λ v) = 0, ∀v ∈ X.
dλ d λ λ =0
It follows then by (3.4.5) that

⟨δ F(u), v⟩ = 0, ∀v ∈ X,

which means that u satisfies (3.4.3).

In the following, we give a simple example to show how to compute δ F using formula
(3.4.5).
Let X = H 1 (Rn ), and F : H 1 (Rn ) → R1 be given by
! 0 1
1 2
F(u) = |∇u| + f (u) dx for u ∈ H 1 (Rn ). (3.4.6)
Rn 2

We see that
! 0 1
d d 1 2
F(u + λ v) = |∇u + λ ∇v| + f (u + λ v) dx
dλ dλ Rn 2
! 8 9
= (∇u + λ ∇v) · ∇v + f ′(u + λ v)v dx.
Rn
152 Chapter 3 Mathematical Foundations

Hence we have
!
d !! 8 9
! F(u + λ v) = ∇u · ∇v + f ′(u)v dx
d λ λ =0 n
!R 8 9
= −div (∇u) + f ′ (u) vdx. (by(3.2.36))
Rn

On the other hand, by (3.4.5) and

!
⟨δ F(u), v⟩ = δ F(u)vdx,
Rn

we deduce that
! !
δ F(u)vdx = (−∆u + f ′(u))vdx, ∀v ∈ H 1 (Rn ).
Rn Rn

Hence we obtain the derivative operator δ F(u) of (3.4.6) as

δ F(u) = −∆u + f ′(u).

3.4.2 Derivative operators of the Yang-Mills functionals

Let Gaµ (1 ! a ! N 2 − 1) be the SU(N) gauge fields. The Yang-Mills functional for Gaµ is
defined by
! 0 1
1 a µν a
F= − Fµν F dx, (3.4.7)
M 4
where M is the 4-dimensional Minkowski space, and
a
Fµν = ∂µ Gaν − ∂ν Gaµ + gλ abcGbµ Gcν . (3.4.8)

In Section 2.4.3, we have seen that the functional (3.4.7) is the scalar curvature part in
the Yang-Mills action (2.4.50).
Referring to derivative operators of functionals for the electromagnetic potential de-
duced in Subsection 2.5.3, we now derive the derivative operator for the Yang-Mills func-
tional (3.4.7):
!
d !! ' =− 1 a d !
!
! F(G + λ G) gµα gνβ Fαβ ! F a (G + λ G)dx
'
d λ λ =0 2 M d λ λ =0 µν
!
B C
'a ∂ G 'aµ
1 µν a ∂ Gν
=− F − ν dx
2 M ∂ xµ ∂x
!
1
− F µν a gλ abc (Gbµ G
'cν + G
'bµ Gcν )dx (by(3.4.8))
2 M
B a C
!
µα νβ
∂ Fαβ c cba b
= g g − gFαβ λ Gµ G 'aν dx.
M ∂ xµ
3.4 Variations with divA -Free Constraints 153

By (3.4.5) we deduce the derivative operator δ F of (3.4.7) as

δ F = ∂ α Fαβ
a
− ggα µ λcb
a c
Fαβ Gbµ , β = 0, 1, 2, 3, (3.4.9)
a = λ cba .
where λcb
For the general form of Yang-Mills functional given by
! 0 1
1 a µν b
F= − Gab Fµν F dx, (3.4.10)
M 4

where (Gab ) is the Riemann metric on SU(N) given by (2.4.49). The derivative operator of
F in (3.4.10) is as follows

δ F = Gab ∂ α Fαβ
b
− ggα µ Gbc λda
c b
Fαβ Gdµ . (3.4.11)

3.4.3 Derivative operator of the Einstein-Hilbert functional

The Einstein-Hilbert functional is in the form
!
√
F= R −gdx, (3.4.12)
M

where M is an n-dimensional Riemannian manifold with metric g i j , and R = gi j Ri j is the

scalar curvature of M , and Ri j is the Ricci curvature tensor:
% 2 &
1 ∂ gkl ∂ 2 gi j ∂ 2 gik ∂ 2 g jl
Ri j = gkl + − − + gkl grs (Γrkl Γsij − Γrik Γsjl ), (3.4.13)
2 ∂ xi ∂ x j ∂ xk ∂ xl ∂ x j ∂ xl ∂ xi ∂ xk
and the Levi-Civita connections Γrkl are written as
% &
r 1 rs ∂ gks ∂ gls ∂ gkl
Γkl = g + k − s . (3.4.14)
2 ∂ xl ∂x ∂x
First we verify the following derivative operator δ F of the Einstein-Hilbert functional
(3.4.12)-(3.4.14):
1
δ F = Ri j − gi j R. (3.4.15)
2
Note that gi j and gi j have the relations
1
gi j = × ||gi j ||, ||gi j || the cofactor of gi j , (3.4.16)
g
1
gi j = × ||gi j ||, ||gi j || the cofactor of gi j . (3.4.17)
g
Hence by (3.4.16), we have
d !!
gi j × ||gi j || = g'i j gi j g.
! det (gi j + λ g'i j ) =' (3.4.18)
d λ λ =0
154 Chapter 3 Mathematical Foundations

j
In addition, by gik gk j = δi , we obtain
d !!
! (gik + λ g'ik )(gk j + λ g'k j ) = 0.
d λ λ =0
It follows that
g'i j = −gik g jl g'kl . (3.4.19)
Thus, (3.4.18) is rewritten as
d !!
! det (gi j + λ g'i j ) = −ggi j g'i j . (3.4.20)
d λ λ =0
For the Einstein-Hilbert functional (3.4.12), we have
d !!
! F(gi j + λ g'i j )
d λ 0λ =0
!
√ 1 d
= Ri j g'i j −g − √ R det (gi j + λ g'i j )
M 2 −g d λ
1 !
d √ !
+ gi j Ri j (gi j + t g'i j ) −g dx!
dλ λ =0
!
1 √
= (Ri j − gi j R)' gi j −gdx
M 2
!
d !! √
+ gi j ! Ri j (gi j + λ g'i j ) −gdx (by (3.4.20)).
M d λ λ =0

In view of (3.4.5) and

!
√
⟨δ F, g'i j ⟩ = δ F g'i j −gdx,
M
d !! ∂ Ri j
! Ri j (gi j + λ g'i j ) = g'kl ,
d λ λ =0 ∂ gkl
we arrive at
! 0 % &1 !
1 i j√ ∂ Ri j √
δ F − Ri j − gi j R g' −gdx = gi j g'kl −gdx.
M 2 M ∂ gkl
To verify (3.4.15), it suffices to prove that
!
√
gi j δ Ri j −gdx = 0, (3.4.21)
M

where δ Ri j is the variational element

δ Ri j = Ri j (gkl + δ gkl ) − Ri j (gkl ),

which are equivalent to the following directional derivative:

d !!
! Ri j (gi j + λ g'i j ).
d λ λ =0
3.4 Variations with divA -Free Constraints 155

To get (3.4.21), we take Ri j in the form

∂ Γkki ∂ Γkij
Ri j = − k + Γlik Γklj − Γli j Γklk . (3.4.22)
∂xj ∂x
By the Riemannian Geometry, for each point x0 ∈ M there exists a coordinate system
under which
Γkij (x0 ) = 0, ∀1 ! i, j, k ! n. (3.4.23)
It is known that the covariant derivatives of gi j and gi j are zero, i.e. Dgi j = 0 and Dgi j = 0.
Hence we infer from (3.4.33) that

∂ gi j (x0 ) ∂ gi j (x0 )
= 0, = 0, ∀1 ! i, j, k ! n.
∂ xk ∂ xk
By (3.4.22), at x0 , we have
% &
ij ij ∂ k ∂ k ∂ D ik l ij k
E
g δ Ri j =g δ Γ ik − δ Γ i j = g δ Γ il − g δ Γ i j . (3.4.24)
∂xj ∂ xk ∂ xk

Although Γkij are not tensor fields, the variations

δ Γkij (x) = Γkij (x + δ x) − Γkij (x)

are (1,2)-tensor fields. Therefore at x0 , (3.4.24) can be rewritten as

∂ uk
g i j δ Ri j = = div u, at x0 ∈ M. (3.4.25)
∂ xk
where
uk = gik δ Γlil − gi j δ Γkij
is a vector field. Since (3.4.25) is independent of the coordinate systems, in a general
coordinate system the relation (3.4.25) becomes
1 ∂ √
gi j δ Ri j = div u = √ ( −guk ), at x0 ∈ M . (3.4.26)
−g ∂ xk
As x0 ∈ M is arbitrary, the formula (3.4.26) holds true on M . Hence we have
! !
√ √
gi j δ Ri j −gdx = div u −gdx.
M M

Since M is closed, i.e. ∂ M = ∅, it follows from (3.2.37) that

!
√
gi j δ Ri j −gdx = 0.
M

Thus we derive (3.4.21), and the derivative operator of the Einstein-Hilbert functional (3.4.12)
is as given by (3.4.15).
156 Chapter 3 Mathematical Foundations

3.4.4 Variational principle with divA -free constraint

Let M be a closed manifold. A Riemannian metric G on M is a mapping

G : M → T20 M = T ∗ M ⊗ T ∗ M ,

which is symmetric and nondegenerate. Namely, in a local coordinate system, G can be

expressed as
G = {gi j } with gi j = g ji , (3.4.27)

and the matrix (gi j ) is invertible on M :

G−1 = (gi j ) = (gi j )−1 : M → T02 M = T M ⊗ T M .

If we regard a Riemannian metric G = {gi j } as a tensor field on the manifold M , then

the set of all metrics G = {gi j } on M constitute a topological space, called the space of
Riemannian metrics on M . The space of Riemannian metrics on M is defined by

W m,2 (M , g) ={G| G ∈ W m,2 (T20 M ), G−1 ∈ W m,2 (T02 M ),

G is the Riemannian metric on M as in (3.4.27)}.

The space W m,2 (M , g) is a metric space, but not a Banach space. However, it is a subspace
of the direct sum of two Sobolev spaces W m,2 (T20 M ) and W m,2 (T02 M ):

W m,2 (M , g) ⊂ W m,2 (T20 M ) ⊕ W m,2 (T02 M ).

A functional defined on W m,2 (M , g):

F : W m,2 (M , g) → R (3.4.28)

is called the functional of Riemannian metric. In general, the functional (3.4.28) can be
expressed as
!
√
F(gi j ) = f (gi j , · · · , ∂ m gi j ) −gdx. (3.4.29)
M

Since (gi j ) is the inverse of (gi j ), we have

1
gi j = × the cofactor of gi j . (3.4.30)
g

Therefore, F(gi j ) in (3.4.29) also depends on gi j , i.e. putting (3.4.30) in (3.4.29) we get
!
√
F(gi j ) = f'(gi j , · · · , ∂ m gi j ) −gdx. (3.4.31)
M
3.4 Variations with divA -Free Constraints 157

We note that although W m,2 (M , g) is not a linear space, but for a given element gi j ∈
W m,2 (M , g) and any symmetric tensor fields Xi j , X i j , there is a number λ0 > 0 such that

gi j + λ Xi j ∈ W m,2 (M , g), ∀0 ! |λ | < λ0 ,

(3.4.32)
gi j + λ X i j ∈ W m,2 (M , g), ∀0 ! |λ | < λ0 .

With (3.4.32), we can define the following derivative operators of the functional F:

δ∗ F : W m,2 (M , g) → W −m,2 (T02 M ),

δ ∗ F : W m,2 (M , g) → W −m,2 (T20 M ),

where W −m,2 (E) is the dual space of W m,2 (E), and δ∗ F, δ ∗ F are defined by

d !!
⟨δ∗ F(gi j ), X⟩ = ! F(gi j + λ Xi j ), (3.4.33)
d λ λ =0
d !!
⟨δ ∗ F(gi j ), X⟩ = ! F(gi j + λ X i j ). (3.4.34)
d λ λ =0

For any give metric gi j ∈ W m,2 (M , g), the value of δ∗ F and δ ∗ F at gi j are second-order
contra-variant and covariant tensor fields:

δ∗ F(gi j ) : M → T M × T M ,
(3.4.35)
δ ∗ F(gi j ) : M → T ∗ M × T ∗ M .

Theorem 3.24 Let F be the functionals defined by (3.4.29) and (3.4.31). Then the
following assertions hold true:

1) For any gi j ∈ W m,2 (M , g), δ∗ F(gi j ) and δ ∗ F(gi j ) are symmetric tensor fields.

2) If {gi j } ∈ W m,2 (M , g) is an extremum point of F, i.e. δ F(gi j ) = 0, then {gi j } is also

an extremum point of F.

3) δ∗ f and δ ∗ F have the following relation

(δ ∗ F(gi j ))kl = −gkr gls (δ∗ F(gi j ))rs ,

where (δ ∗ F)kl and (δ∗ F)kl are the components of δ ∗ F and δ∗ F.

j
Proof We only need to verify Assertion (3). In view of g ik gk j = δi , we have the
variational relation
δ (gik gk j ) = gik δ gk j + gk j δ gik = 0.

It implies that
δ gkl = −gki gl j δ gi j . (3.4.36)
158 Chapter 3 Mathematical Foundations

In addition, in (3.4.33) and (3.4.34),

λ Xi j = δ gi j , λ X i j = δ gi j , λ ̸= 0 small.

Therefore, by (3.4.36) we get

⟨(δ∗ F)kl , δ gkl ⟩ = −⟨(δ∗ F)kl , gki gl j δ gi j ⟩ = −⟨gki gl j (δ∗ F)kl , δ gi j ⟩ = ⟨(δ ∗ F)i j , δ gi j ⟩.

Hence we have
(δ ∗ F)i j = −gki gl j (δ∗ F)kl .
Thus Assertion (3) follows and the proof is complete.

We are now in position to consider the variation with divA -free constraints. We know
that an extremum point gi j of a metric functional is a solution of the equation

δ F(gi j ) = 0, (3.4.37)

in the sense that for any Xkl = Xlk ∈ L2 (T20 M ),

!
d !! √
⟨δ F(gi j ), X⟩ = ! F(gi j + λ Xi j )|λ =0 = (δ ∗ F(gi j ))kl Xkl −gdx = 0. (3.4.38)
d λ λ =0 M

Note that the solution gi j of (3.4.37) in the usual sense should satisfy

⟨δ F(gi j ), X⟩ = 0, ∀X ∈ L2 (T20 M ). (3.4.39)

Notice that (3.4.38) has a symmetric constraint on the variational elements Xi j : Xi j = X ji .

Therefore, comparing (3.4.38) with (3.4.39), we may wonder if a solution g i j satisfying
(3.4.38) is also a solution of (3.4.39). Fortunately, note that L2 (T20 M ) can be decomposed
into a direct sum of symmetric and anti-symmetric spaces as follows

L2 (T20 M ) = L2s (T20 M ) ⊕ L2c (T20 M ),

L2s (T20 M ) = {u ∈ L2 (T20 M )| ui j = u ji },
L2c (T20 M ) = {u ∈ L2 (T20 M )| ui j = −u ji },

and L2s (T20 M ) and L2c (T20 M ) are orthogonal:

! !
√ √
gik g jl ui j vkl −gdx = − gik g jl ui j vlk −gdx
M M
= 0, ∀u ∈ L2s (T20 M ), v ∈ L2c (T20 M ).

Thus, due to the symmetry of δ F(gi j ), the solution gi j of (3.4.37) satisfying (3.4.38) must
also satisfy (3.4.39). Hence the solutions of (3.4.37) in the sense of (3.4.38) are the solutions
in the usual sense.
However, if we consider the variations of F under the divA -free constraint, then the
extremum points of F are not solutions of (3.4.37) in the usual sense. Motived by physical
considerations, we now introduce variations with divA -free constraints.
3.4 Variations with divA -Free Constraints 159

Definition 3.25 Let F(u) be a functional of tensor fields u. We say that u 0 is an

extremum point of F(u) under the devA -free constraint, if

d !!
⟨δ F(u0 ), X⟩ = ! F(u0 + λ X) = 0, ∀divA X = 0, (3.4.40)
d λ λ =0
where divA is as defined in (3.2.38).
In particular, if F is a functional of Riemannian metrics, and the solution u 0 = gi j is a
Riemannian metric, then the differential operator DA in divA X in (3.4.40) is given by

DA = D + A, D = ∂ + Γ, (3.4.41)

and the connection Γ is taken at the extremum point u 0 = gi j .

We have the following theorems for divA -free constraint variations.

Theorem 3.26 Let F : W m,2 (M , g) → R1 be a functional of Riemannian metrics.

Then there is a vector field Φ ∈ H 1 (T M ) such that the extremum points {gi j } of F with the
divA -free constraint satisfy the equation

δ F(gi j ) = DΦ + A ⊗ Φ, (3.4.42)

where D is the covariant derivative operator as in (3.4.41).

Theorem 3.27 Let F : H m (T M ) → R1 be a functional of vector fields. Then there is

a scalar function ϕ ∈ H 1 (M ) such that for a given vector field A, the extremum points u of
F with the divA -free constraint satisfy the equation

δ F(u) = (∂ + A)ϕ . (3.4.43)

Proof of Theorems 3.26 and 3.27 First we prove Theorem 3.26. By (3.4.40), the
extremum points {gi j } of F with the divA -free constraint satisfy
!
√
δ F(gi j ) · X −gdx = 0, ∀X ∈ L2 (T02 M ) with divA X = 0.
M

It implies that
δ F(gi j )⊥L2D (T20 M ) = {v ∈ L2 (T20 M )| divA v = 0}. (3.4.44)
By Theorem 3.17, L2 (T20 M ) can be orthogonally decomposed into

L2 (T20 M ) = L2D (T20 M ) ⊕ G2 (T20 M ),

G2 (T20 M ) = {DA Φ| Φ ∈ H 1 (T10 M )}.

Hence it follows from (3.4.44) that

δ F(gi j ) ∈ G2 (T20 M ),
160 Chapter 3 Mathematical Foundations

which means that the equality (3.4.42) holds true.

To prove Theorem 3.27, for an extremum vector field u of F with the divA -free con-
straint, we derive in the same fashion that u satisfies the following equation
!
√
δ F(u) · X −gdx = 0, ∀X ∈ L2 (T M ) with divA x = 0. (3.4.45)
M

In addition, Theorem 3.17 means that

L2 (T M ) = L2D (T M ) ⊕ G2 (T M ),
L2D (T M ) = {v ∈ L2 (T M )| divA v = 0},
G2 (T M ) = {DA ϕ | ϕ ∈ H 1 (M )}.

Then we infer from (3.4.45) that

δ F(u) ∈ G2 (T M ).

Thus we deduce the equality (3.4.43).

The proofs of Theorems 3.26 and 3.27 are complete.

3.4.5 Scalar potential theorem

In Theorem 3.26, if the vector field A in DA is zero, and the first Betti number β1 (M ) = 0
for M , then we have the following scale potential theorem. This result is also important for
the gravitational field equations and the theory of dark matter and dark energy introduced in
Chapter 7.

Theorem 3.28 (Scalar Potential Theorem) Assume that the first Betti number of M is
zero, i.e. β1 (M ) = 0. Let F be a functional of Riemannian metrics. Then there is a scalar
field ϕ ∈ H 2 (M ) such that the extremum points {gi j } of F with divergence-free constraint
satisfy the equation
(δ F(gi j ))kl = Dk Dl ϕ . (3.4.46)

Proof Let {gi j } be an extremum point of F under the divergence-free constraint:

!
√
(δ F(gi j ))kl X kl −gdx = 0, ∀X = {Xkl } with Dk X kl = 0.
M

By Theorem 3.26 and A = 0 in (3.4.42), δ F(gi j ) is in the form

(δ F(gi j ))kl = Dk Φl , (3.4.47)

for some {Φl } ∈ H 1 (T ∗ M ). By Theorem 3.24, δ F(gi j ) is symmetric. Hence we have

Dk Φl = Dl Φk . (3.4.48)
3.4 Variations with divA -Free Constraints 161

In addition, by
∂ Φl
Dk Φl = − Γklj Φ j ,
∂ xk
and Γklj = Γlkj , it follows from (3.4.48) that

∂ Φl ∂ Φk
= . (3.4.49)
∂x k ∂ xl
By assumption, the first Betti number of M is zero, i.e. the first homology of M is zero:
H1 (M ) = 0. It follows from the de Rham theorem that if
% &
k ∂ Φk ∂ Φl
d(Φk dx ) = − k dxl ∧ dxk = 0,
∂ xl ∂x
then there exists a scalar function ϕ such that
∂ϕ k
dϕ = dx = Φk dxk .
∂ xk
Thus, we infer from (3.4.49) that
∂ϕ
Φk = for some ϕ ∈ H 2 (M ).
∂ xk
Therefore, we derive (3.4.46) from (3.4.47), and the proof is complete.

If the first Betti number β1 (M ) ̸= 0, then we have the following theorem.

Theorem 3.29 Let the first Betti number of M is β1 (M ) = N with N ̸= 0. Then there
are a scalar field ϕ ∈ H 2 (M ) and N vector fields ψ j (1 ! j ! N) in H 1 (T ∗ M ) such that
the extremum point {gi j } of F with divergence-free constraint satisfies the equation
N
(δ F(gi j ))kl = Dk Dl ϕ + ∑ α j Dk ψl ,
j
(3.4.50)
j=1
j j
Dk Dk ψl = −Rkl ψk for 1 ! j ! N, 1 ! l ! n, (3.4.51)

where α j (1 ! j ! N) are constants, Rkl = gk j R jl and R jl are the Ricci curvature tensors.

Proof Since the first Betti number β1 (M ) = N(̸= 0), by the de Rham theorem, there
are N closed 1-forms

ω j = ψkj dxk ∈ Hd1 (M ) for 1 ! j ! N, (3.4.52)

they constitute a basis for the 1-dimensional de Rham homology Hd1 (M ). Hence ω j (1 !
j ! N) are not exact, and satisfy that
B C
∂ ψkj ∂ ψlj
dω j = − k dxl ∧ dxk = 0,
∂ xl ∂x
162 Chapter 3 Mathematical Foundations

which imply that

∂ ψkj ∂ ψlj
= for 1 ! j ! N.
∂ xl ∂ xk
or equivalently
j j
Dl ψk = D k ψl for 1 ! j ! N.
Namely, ∇ψ j ∈ L2 (T ∗ M ⊗ T ∗ M ) are symmetric second-order covariant tensor fields.
Hence any covector field Φ ∈ L2 (T ∗ M ) satisfying

Dl Φk = Dk Φl (3.4.53)

must be in the following form

N
Φk = Dk ϕ + ∑ α j ψkj , (3.4.54)
j=1

where α j (1 ! j ! N) are constants, ϕ is some scalar field. Hence, by A = 0 in D A , the

equation (3.4.42) in Theorem 3.26 can be expressed as

(δ F(gi j ))lk = Dl Φk ,

where Φk satisfy (3.4.53) and (3.4.54), which are the equations given by (3.4.50).
On the other hand, by the Hodge decomposition theorem, the 1-forms in (3.4.52) are
harmonic, i.e.
d ω j = 0, δ ω j = 0 for 1 ! j ! N.
It follows that the covector fields ψ j (1 ! j ! N) in (3.4.52) satisfy

∆ψ j = 0 for 1 ! j ! N, (3.4.55)

where ∆ = d δ + δ d is the Laplace-Beltrami operator as defined in (3.2.31). Hence, the

equations in (3.4.55) can be equivalently rewritten in the form

Dk Dk ψlj = −Rkl ψkj for 1 ! j ! N, 1 ! l ! n,

which are exactly the equations in (3.4.51). The proof is complete.

3.5 SU(N) Representation Invariance

3.5.1 SU(N) gauge representation
We briefly recapitulate the SU(N) gauge theory. In the general case, a set of SU(N) gauge
fields consists of K = N 2 − 1 vector fields Aaµ and N spinor fields ψ j :

A1µ , · · · , AKµ , Ψ = (ψ 1 , · · · , ψ N )T . (3.5.1)

3.5 SU(N) Representation Invariance 163

For the fields (3.5.1), a gauge invariant functional is the Yang-Mills action:
!
LY M = [LG + LD ] dx, (3.5.2)

where LG and LD are the gauge field section and Dirac spinor section, and are written as
1 a µν a
LG = − Fµν F ,
4 (3.5.3)
a
Fµν = ∂µ Aaν − ∂ν Aaµ + gλbc
a b c
Aµ Aµ ,

and
LD = Ψ(iγ µ Dµ − m)Ψ,
(3.5.4)
Dµ = ∂µ + igAaµ τa ,
where τa (1 ! a ! K = N 2 − 1) are the generators of SU(N).
The actions (3.5.2)-(3.5.4) are invariant under both the Lorentz transformation and the
SU(N) gauge transformation as follows
Ψ
' = ΩΨ,
(3.5.5)
'aµ τa = ΩAaµ τa Ω−1 + i (∂µ Ω)Ω−1 ,
A
g
and Ω ⊂ SU(N) can be expressed as
aτ
Ω = eiθ a
, τa is in (3.5.4).
For the gauge theory, a vary basic and important problem is that in the gauge transfor-
mation (3.5.5) the generators of SU(N) given by
{τa | 1 ! a ! K}, (3.5.6)
have infinite numbers of families, and each family of (3.5.6) corresponds to a set of gauge
fields:
{τa | 1 ! a ! K} ↔ {Aaµ | 1 ! a ! K}. (3.5.7)
Now, we assume that the generators of (3.5.6) undergo a linear transformation as follows
τ'b = xab τa , (3.5.8)
where (xab ) is a K-th order complex matrix. Then the corresponding gauge field A aµ in (3.5.7)
has to change. Namely, under the transformation (3.5.8) A aµ will transform as
'aµ = ya Abµ ,
A (3.5.9)
b

and (yab ) is a K-order matrix depending on (xab ).

Intuitively, any gauge theory should be independent of the choice of {τ a }, otherwise the
basically logical rationality will be broken. In other words, the Yang-Mills density (3.5.3)
should be invariant under the transformation (3.5.9). However, in view of (3.5.3), if

Aaµ → yab Abµ ⇒ Fµν

a
→ yab Fµb ,
164 Chapter 3 Mathematical Foundations

then LG will be changed as

a
LG = Fµν F µν a → yab yac Fµν
b
F µν c .

Hence the Yang-Mills action (3.5.2) violates the invariance.

To solve this problem, the authors have developed in (Ma and Wang, 2014h) a mathe-
matical theory of SU(N) representation invariance, where the SU(N) tensors and the Rie-
mannian metric on SU(N) are defined. Furthermore the Yang-Mills action is revised. In this
subsection, we shall introduce this theory.

3.5.2 Manifold structure of SU(N)

To establish the representation invariance theory for SU(N) gauge fields, we need to intro-
duce the SU(N) tensors and the Riemannian metric defined on SU(N). The main objective
of this subsection is to introduce some basic concepts on SU(N), including manifold struc-
ture, tangent space, coordinate systems and coordinate transformations.
1. Manifold structure on SU(N). In mathematics, a space M is an n-dimensional
manifold means that for each point p ∈ M there is a neighborhood U ⊂ M of p, such that
U is homeomorphic to Rn , i.e. there exists an one to one mapping

ψ : U → Rn

and ψ has a continuous inverse ψ −1 : Rn → U.

The group SU(N) consists of all N-th order unitary matrices with unit determinant:

SU(N) = {A | A is an N-th order matrix , A† A = I, det A = 1}.

Each matrix A ⊂ SU(N) can be written as

⎛ ⎞
a11 ··· a1N
⎜ .. .. ⎟ ,
A=⎝ . . ⎠
aN1 ··· aNN

and akl = xkl + iykl ∈ C (1 ! k, l ! N). Thus the matrix A can be regarded as a point p A in
2
R2N :
2
pA = (x11 , y11 , · · · , x1N , y1N , · · · , xN1 , yN1 , · · · , xNN , yNN ) ∈ R2N . (3.5.10)
2
Therefore, we have that SU(N) can be regarded as a subspace of R 2N .
By A† A = I and det A = 1, the entries akl (1 ! k, l ! N) of A satisfy

ai j a∗jk = δik for 1 ! i, k ! N,

(3.5.11)
det (ai j ) = 1,
3.5 SU(N) Representation Invariance 165

which are N 2 + 1 equations, as constraints for the point pA in (3.5.10). Hence SU(N) can
2
be regarded as subspace of R2N has dimension N 2 − 1.
Mathematically SU(N) is a manifold. In fact, at any point p A of (3.5.10), each equation
2
of (3.5.11) represents a hypersurface near pA in R2N :

Σik : ai j a∗jk = δik for 1 ! i, k ! N,

Σ1 : det (ai j ) = 1,
2
and the N 2 + 1 hypersurface Σik and Σ1 transversally interact in R2N to constitute an N 2 − 1
2
dimensional surface Γ(pA ) near each pA ∈ R2N , and the sum of all Σ(pA ) is the SU(N)
space:
"
SU(N) = Γ(pA ).
pA

Hence, SU(N) possesses the manifold structure.

2. Tangent space TA SU(N). Since SU(N) is an N 2 − 1 dimensional manifold, at each

point A ∈ SU(N) there is a tangent space, denoted by TA SU(N), which is an N 2 − 1 dimen-
sional linear space, as shown in Figure 3.2.

Figure 3.2 Tangent space TA SU(N) at A ∈ SU(N)

Now we derive some properties of tangent vectors τ on TA SU(N). To this end, let
γ (t) ⊂ SU(N) be a curve passing through the point A ∈ SU(N) with τA ∈ TA SU(N) as its
tangent vector at A, as shown in Figure 3.2. Let γ (0) = A. Then the curve γ (t) satisfies the
following equation

d γ (t)
= τt , τt ∈ T SU(N) with τA = τ |t=0 ,
dt (3.5.12)
γ (0) = A.
166 Chapter 3 Mathematical Foundations

For infinitesimal t, the solution of (3.5.12) is

γ (t) = A + t τA ⊂ SU(N).

It follows that
(A + t τA)† (A + t τA) = I. (3.5.13)
As A† A = I and t is infinitesimal, we deduce from (3.5.13) that

A† τA + τA† A = 0. (3.5.14)

Hence, (3.5.14) is the condition for an N-th order complex matrix τ ∈ TA SU(N): A† τ is
anti-Hermitian. Namely,

TA SU(N) = {τ | τ satisfies (3.5.14)}. (3.5.15)

Note that
†τ
A + t τA = A(I + tA†τA ) = AetA A , (3.5.16)
for infinitesimal t. If we replace τ by iτ , then (3.5.15) can be expressed as

TA SU(N) = {iτ | A† τ = τ † A, Tr(A† τ ) = 0}. (3.5.17)

By (3.5.17), ∀A ∈ SU(N) there is a neighborhood UA ⊂ SU(N) such that for any Ω ∈ UA , Ω

can be written as
†
Ω = AeiA τ for some iτ ∈ TA SU(N). (3.5.18)
Here the traceless condition in (3.5.17)

Tr(A† τ ) = 0

is derived by (2.2.51) and det Ω = 1 for Ω as in (3.5.18).

In particular, at the unit matrix e = I, we have

Te SU(N) = {iτ | τ † = τ , Trτ = 0},

(3.5.19)
Ω = eiτ , τ ∈ Te SU(N),

where Ω ⊂ SU(N) is in a neighborhood of e = I.

3. Coordinate systems on T SU(N). Since the representation (3.5.17)-(3.5.18) of SU(N)
is essentially the same for all A ∈ SU(N), it suffices to only consider the representation
(3.5.19) of SU(N) at the unit matrix e = I.
It is known that Te SU(N) is an N 2 − 1 dimensional linear space. Therefore we can take
a coordinate basis, called the generator basis of SU(N), denoted by

τ1 , · · · , τK , K = N 2 − 1, (3.5.20)
3.5 SU(N) Representation Invariance 167

such that for any τ ∈ Te SU(N), we have

τ = θ a τa , (3.5.21)
and θ a (1 ! a ! K) are complex numbers, called the coordinate system on Te SU(N). In this
case, Ω in (3.5.19) can be expressed as
aτ
Ω = eiθ a
, Ω ⊂ SU(N).
3.5.3 SU(N) tensors
Let (τ1 , · · · , τK ) ⊂ Te SU(N) be a generator basis of SU(N). If the basis undergoes a linear
transformation:
τ'a = xba τb ,
(3.5.22)
X = (xba ) is a complex K-th order matrix,
then the coordinate system (θ 1 , · · · , θ K ) of Te SU(N) also undergoes a corresponding trans-
formation as follows
θ'a = yab θ b ,
(3.5.23)
Y = (yab ) is a complex K-th order matrix.

Since the expression (3.5.21) is independent of the choice of the generator bases of SU(N),
we have ⎛ ⎞
τ1
θ'a τ'a = (θ 1 , · · · , θ K )Y T X ⎝ ... ⎠ = θ a τa ,
⎜ ⎟

τK
which requires that
Y = (X −1 )T . (3.5.24)
Thus, we see that (θ 1 , · · · , θ K ) is a first order contra-variant tensor defined on Te SU(N).
We are now ready to define more general SU(N) tensors.
Definition 3.30 (SU(N) Tensors) Let T be given as

T = {Tba11···b
···ai
j
| 1 ! ak , bl ! K = N 2 − 1}.

We say that T is a (i, j)-type of SU(N) tensor, under the generator basis transformation as
(3.5.22), we have
d
T'ba11···b
···ai
j
= yac11 · · · yacii xdb11 · · · xb jj Tdc11···d
···ci
j
,
where (xab ) and (yab ) are as in (3.5.22) and (3.5.23).

Based on Definition 3.30, it is easy to see that the SU(N) gauge fields (A 1µ , · · · , Aaµ ) is a
contra-variant SU(N) tensor. In other words, under (3.5.22), (A 1µ , · · · , Aaµ ) transforms as

'aµ = ya Abµ ,
A Y = (yab ) as in (3.5.24). (3.5.25)
b
168 Chapter 3 Mathematical Foundations

This can be seen from the fact that the operator Aaµ τa in the differential operator D µ in
(3.5.4) is independent of generator bases τa of SU(N).
a of SU(N) constitute a (1,2)-type of SU(N)
We now verify that the structure constants λbc
a,
tensor. By the definition of λbc

[τb , τc ] = τb τc† − τc τb† = iλbc

a
τa .

By (3.5.22),
f
[τ'b , τ'c ] = xab xdc [τa , τd ] = ixab xdc λad τf ,
and by definition
a' a d
[τ'b , τ'c ] = i'
λbc τa = i'
λbc xa τd .
Then it follows that
a f
'
λbc = xb xgc yad λ fdg , (3.5.26)
which means that a}
{λbc is a (1,2)-type SU(N) tensor.
Next, we introduce two second-order covariant SU(N) tensors G ab and gab , and later we
shall prove that they are equivalent.
Let A ∈ SU(N). Then the tangent space TA SU(N) is given by (3.5.17). Let

ω1 , · · · , ωK ∈ TA SU(N), (3.5.27)

be a generator basis of SU(N) at A.

1) SU(N) tensor Gab (A). By the basis (3.5.27) we can get a 2-covariant SU(N) tensor
defined by
1
Gab (A) = tr(ωa ωb† ), A ∈ SU(N), (3.5.28)
2
where ωa (1 ! a ! K) are as in (3.5.27).

2) SU(N) tensor gab . The structure constants λbc a generated by the generators ω in
a
(3.5.27) satisfy
a
[ωb , ωc ] = iλbc ωa A † , ∀A ∈ SU(N) (3.5.29)
then we can define another 2-covariant tensor gab on TA SU(N) by the structure con-
a as follows
stants λbc
1 c d
gab = λ λ . (3.5.30)
4N ad cb
The following theorem shows that both Gab and gab are symmetric second-order SU(N)
tensors.
Theorem 3.31 The fields Gab (A) and gab given by (3.5.28) and (3.5.30) are 2-order
symmetric SU(N) tensors, and for any A ∈ SU(N), the generator basis (3.5.27) satisfies
a are the structure constants of SU(N) independent of A.
(3.5.29), where λbc
3.5 SU(N) Representation Invariance 169

Proof We first consider gab . By the anti-symmetry of the structure constants:

c c
λab = −λba ,

we deduce the symmetry of gab :

c d c d
gab = λad λcb = λda λbc = gba .

Now we verify the symmetry of Gab . By (3.5.17) the basis (ω1 , · · · , ωK ) of TA SU(N)
satisfy that
A† ωa = (A† ωa )† , tr(A† ωa ) = 0, A† = A−1 , (3.5.31)

Namely A† ωa are Hermitian. Let A† ωa = τa , then

τa = τa† , trτa = 0. (3.5.32)

Thus we have
1 1 1
Gab = tr(ωa ωb† ) = tr(Aτa τb† A† ) = tr(τa τb† ). (3.5.33)
2 2 2
Thanks to (3.5.32), (τ1 , · · · , τK ) ∈ Te SU(N). Hence we have

τa τb† = τb τa† + iλab

c
τc , trτc = 0. (3.5.34)

It follows from (3.5.33) and (3.5.34) that

1 1
Gab = tr(τa τb† ) = tr(τb τa† ) = Gba .
2 2
Hence Gab is symmetry.
Finally we prove that for any A ∈ SU(N), each generator basis (ω1 , · · · , ωK ) of TA SU(N)
satisfy (3.5.29). In fact, by (3.5.31), τa = A† ωa (1 ! a ! K) constitute a basis of Te SU(N).
Therefore τa satisfies (3.5.34), i.e.

A† [ωa , ωb ]A = iλab
c †
A ωc .

Hence we obtain (3.5.29). The proof is complete.

3.5.4 Intrinsic Riemannian metric on SU(N)

By Theorem 3.31, Gab (A) and gab are symmetric. We now show that both Gab and gab are
positive definite, and consequently Gab (A) is an intrinsic Riemannian metric on SU(N).

Theorem 3.32 For the 2-order SU(N) tensors gab and Gab , the following assertions
hold true:

1) The tensor {gab } given by (3.5.30) is positive definite; and

170 Chapter 3 Mathematical Foundations

2) For each point A ∈ SU(N), there is a coordinate system in where G ab (A) = gab . There-
fore Gab (A) is a Riemannian metric on SU(N).

Proof First, we prove Assertion 1). As a 2-order covariant tensor, g ab transforms as

gab ) = X(gab)X T ,
(' X = (xab ) as in (3.5.22). (3.5.35)

1 c d
Hence, if for a given basis {τa | 1 ! a ! K} of Te SU(N) we can verify that gab = λ λ
4N ad cb
is positive definition, then by (3.5.35) we derive Assertion 1). In the following, we proceed
first for SU(2) and SU(3), then for the general SU(N).
For SU(2), we take the Pauli matrices
% & % & % &
0 1 0 −i 1 0
σ1 = , σ2 = , σ3 = (3.5.36)
1 0 i 0 0 −1
c of (3.5.36) are as follows
as the generator basis of SU(2). The structure constants λ ab
⎧
⎪ 1 if (abc) is even,
⎨
c
λab = 2εabc , εabc = − 1 if (abc) is odd, (3.5.37)
⎪
⎩
0 if otherwise.

Based on (3.5.37), direct calculation shows that

1 c d
gab = λad λcb = δab .
8
Namely (gab ) = I is identity.
For SU(3), we can take the following Gell-Mann matrices as the generator basis of
SU(3):
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 1 0 0 −i 0 1 0 0
λ1 = ⎝1 0 0 ⎠ , λ2 = ⎝ i 0 0 ⎠ , λ3 = ⎝0 −1 0⎠ ,
0 0 0 0 0 0 0 0 0
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0 0 1 0 0 −i 0 0 0
λ4 = ⎝ 0 0 0⎠ , λ5 = ⎝0 0 0 ⎠, λ6 = ⎝0 0 1⎠ , (3.5.38)
−1 0 0 i 0 0 0 1 0
⎛ ⎞ ⎛ ⎞
0 0 0 1 0 0
1 ⎝
λ7 = ⎝0 0 −i⎠ , λ8 = √ 0 1 0 ⎠.
0 i 0 3 0 0 −2

The structure constants are

c
λab = 2 fabc , 1 ! a, b, c ! 8,
3.5 SU(N) Representation Invariance 171

and fabc are anti-symmetric, given by

1
f123 = 1, f147 = f246 = f257 = f345 = ,
2
1 √
f156 = f367 = − , f458 = f678 = 3/2, (3.5.39)
2
fabc = 0, for others.

By (3.5.39) we can deduce that

@
c d
0, if a ̸= b,
λad λcb =
12, if a = b.
Hence we get
1 c d
gab = λ λ = δab .
12 ad cb
Namely, under the Gell-Mann representation (3.5.38), (g ab ) = I.
We are now in position to consider general SU(N). In fact, for all N " 2, there exists
1
a generator basis {τa | 1 ! a ! N 2 − 1} of SU(N) such that (gab ) = (λ c λ d ) = I. The
4N ad cb
generators τa are given by the following traceless Hermitian matrices:
172 Chapter 3 Mathematical Foundations

where σk (1 ! k ! 3) are as in (3.5.36) and λ j (3 ! j ! 8) as in (3.5.38). Corresponding to

(3.5.40), gab are as follows
1 c d
gab = λ λ = δab . (3.5.41)
4N ad cb
Hence the 2-order SU(N) tensor {gab } is positive definition.

Now, we prove Assertion 2). For each A ∈ SU(N) we take the matrices

ωa = Aτa for 1 ! a ! K, (3.5.42)

where τa (1 ! a ! K) form a basis of Te SU(N). It is clear that ωa satisfy the properties

(3.5.31). Hence the matrices {ωa | 1 ! a ! K} constitute a basis of TA SU(N). On the other
hand, we see that
1
Gab (A) = tr(ωa ωb† )
2
1
= tr(Aτa τb† A† ) (by (3.5.41))
2
1
= tr(τa τb† ) (by (2.3.16)). (3.5.43)
2
If we take (3.5.40) as the basis τa (1 ! a ! K), then we have

1
tr(τa τb† ) = δab .
2
It follows from (3.5.43) that with the basis (3.5.40) of TA SU(N),

Gab (A) = Gab (I) = δab = gab (by (3.5.42)).

Assertion 2) and the theorem are proved.

3.5.5 Representation invariance of gauge theory

In this subsection, we consider the representation invariance for the SU(N) gauge theory.
In Subsection 3.5.1 we see that the classical Yang-Mills action (3.5.2)-(3.5.4) will change
under the transformation of generator bases of SU(N). The modified version of the Yang-
Mills action obeying the representation invariance is given by

LY M = LG + LD ,
1
LG = − Gab Fµνa
F µν b , (3.5.44)
4
8 9
LD = Ψ iγ µ (∂µ + igAaµ τa ) − m Ψ,

where Gab is as in (3.5.28).

3.6 Spectral Theory of Differential Operators 173

It is clear that the action density (3.5.44) is invariant under the transformation of TA SU(N):

τ'a = xba τb . (3.5.45)

In fact, the following are three terms in (3.5.44), which involve contractions of SU(N)
tensors:
a
Gab Fµν F µν b , Aaµ τa , λbc
a b c
A µ Aν .

Obviously, these terms are also Lorentz invariant.

Remark 3.33 The purely mathematical logic requires the introduction of the modified
Yang-Mills action (3.5.44) and the SU(N) tensors. There is a profound physical significance.
This invariance dictates that mixing different gauge potentials from different gauge groups
will often lead to the violation this simple principle. As we shall see, the new invariance
theory is very important and crucial in the unified field model presented in the next chapter,
where this principle is called the Principle of Representation Invariance (PRI).

3.6 Spectral Theory of Differential Operators

3.6.1 Physical background
Based on the Bohr atomic model, an atom consists of a proton and its orbital electron,
bounded by electromagnetic energy. Due to the quantum effect, the orbital electron is in
proper discrete energy levels:
0 < E1 < · · · < EN , (3.6.1)

which can be expressed as

En = E0 + λ n (λn < 0), (3.6.2)

where λn (1 ! n ! N) are the negative eigenvalues of a symmetric elliptic operator. Here

E0 stands for the intrinsic energy, and λn stands for the bound energy of the atom, holding
the orbital electrons, due to the electromagnetism. Hence there are only N energy levels E n
for the atom, which are certainly discrete.
To see this, let Z be the atomic number of an atom. Then the potential energy for
electrons is given by
Ze2
V (r) = − .
r
With this potential, the wave function ψ of an orbital electron satisfies the Schrödinger
equation
∂ψ h̄2 2 Ze2
ih̄ + ∇ ψ+ ψ = 0. (3.6.3)
∂t 2m0 r
Let ψ take the form
ψ = e−iλ t/h̄ ϕ (x),
174 Chapter 3 Mathematical Foundations

where λ is the bound energy. Putting ψ into (3.6.3) leads to

h̄2 2 Ze2
− ∇ ϕ− ϕ = λ ϕ.
2m0 r
Since the orbital electrons are bound in the interior of the atom, the following condition
holds true:
ϕ =0 for |x| > r0 ,
where r0 is the radius of an atom. Thus, if ignoring the electromagnetic interactions be-
tween orbital electrons, then the bound energy of an electron is a negative eigenvalue of the
following elliptic boundary problem

h̄2 2 Ze2
− ∇ ϕ− ϕ = λϕ for x ∈ Br0 ,
2m0 r (3.6.4)
ϕ =0 for x ∈ ∂ Br0 ,

where Br0 is a ball with the atom radius r0 .

According to the spectral theory for elliptic operators, the number of negative eigenval-
ues of (3.6.4) is finite. Hence, it is natural that the energy levels in (3.6.1) and (3.6.2) are
finite and discrete.

3.6.2 Classical spectral theory

Consider the eigenvalue problem of linear elliptic operators as follows

− D2 ψ + A ψ = λ ψ for x ∈ Ω,
(3.6.5)
ψ =0 for x ∈ ∂ Ω,

where Ω ⊂ Rn is a bounded domain, ψ = (ψ1 , · · · , ψm )T : Ω → Cm is a complex-valued

function with m components,

D = ∇ + i⃗B, ⃗B = (B1 , · · · , Bn ), (3.6.6)

and A, Bk (1 ! k ! n) are m-th order Hermitian matrices:

A = (Ai j (x)), Bk = (Bkij (x)). (3.6.7)

Let λ0 be an eigenvalue of (3.6.5). The corresponding eigenspace at λ 0 is

Eλ0 = {ψ ∈ L2 (Ω, Cm )| ψ satisfy (3.6.5) with λ = λ0 }

is finite dimensional, and its dimension

N = dim Eλ0
3.6 Spectral Theory of Differential Operators 175

is called the multiplicity of λ0 . Physically, N is also called the degeneracy provided N > 1.
Usually, we count the multiplicity N of λ0 as N eigenvalues, i.e., we denote

λ1 = · · · = λ N = λ0 .

Based on the physical background, we mainly concern the negative eigenvalues. How-
ever, for our purpose the following classical spectral theorem is very important.

Theorem 3.34 (Spectral Theorem of Elliptic Operators) Let the matrices in (3.6.7)
are Hermitian, and the functions Ai j , Bkij ∈ L∞ (Ω). The the following assertions hold true:

1) All eigenvalues of (3.6.5) are real with finite multiplicities, and form an infinite con-
sequence as follows:

−∞ < λ1 ! λ2 ! · · · ! λk ! · · · , λk → ∞ as k → ∞.

where λk is counting the multiplicity.

2) The eigenfunctions ψk corresponding to λk are orthogonal to each other, i.e.

!
ψk† ψ j dx = 0, ∀k ̸= j.
Ω

In particular, {ψk } is an orthogonal basis of L2 (Ω, Cm ).

3) There are only finite number of negative eigenvalues in {λ k },

− ∞ < λ1 ! · · · ! λN < 0, (3.6.8)

and the number N of negative eigenvalues depends on the matrices A, B j in (3.6.7)

and the domain Ω.

Remark 3.35 For the energy levels of subatomic particles introduced in Chapter 5,
we are mainly interested in the negative eigenvalues of (3.6.5) and in the estimates of the
number N in (3.6.8).

Theorem 3.34 is a corollary of the classical Lagrange multiplier theorem. We recall the
variational principle with constraint. Let H be a linear normed space, and F and G are two
functionals on H:
F, G : H → R.
Let Γ ⊂ H be the set
Γ = {u ∈ H| G(u) = 1}.
If u0 ∈ Γ is a minimum point of F with constraint on Γ:

F(u0 ) = min F(u),

u∈Γ
176 Chapter 3 Mathematical Foundations

then u0 satisfies the equation

δ F(u0 ) = λ δ G(u0 ), (3.6.9)

where λ is a real number.

For the eigenvalue equation of (3.6.5), the corresponding functional is
! 8 9
F(ψ ) = |Dψ |2 + ψ † Aψ dx, (3.6.10)
Ω

and the constraint functional G is given by

!
G(ψ ) = |ψ |2 dx. (3.6.11)
Ω

It is easy to see that the equation of (3.6.5) is of the form:

δ F(ψ ) = λ δ G(ψ ),

and F, G are as in (3.6.10) and (3.6.11), which is as the variational equation (3.6.9) with the
constraint on Γ.
Hence, the eigenvalues λk (k = 1, 2, · · · ) of (3.6.5) can be expressed in the following
forms
λ1 = min F(ψ ),
ψ ∈Γ
(3.6.12)
λk = min F(ψ ),
ψ ∈Γ,ψ ∈Hk−1
⊥

where F is as in (3.6.10), Γ and Hk−1

⊥ are the sets:

Γ = {ψ ∈ H01 (Ω, Cm )| ||ψ ||L2 = 1},

> !! ? (3.6.13)
!
⊥
Hk−1 = ψ ∈ H01 (Ω, Cm ) ! ψ † ψ j dx = 0, 1 ! j ! k − 1 ,
Ω

and ψ j (1 ! j ! k − 1) are the eigenfunctions corresponding to the first (k − 1) eigenvalues

λ1 , · · · , λk−1 . Namely Hk−1
⊥ is the orthogonal complement of H
k−1 = span{ψ1 , · · · , ψk−1 }
in H01 (Ω, Cm ).
Based on (3.6.12)-(3.6.13), we readily deduce the spectral theorem, Theorem 3.34, for
the elliptic eigenvalue problem (3.6.5).

3.6.3 Negative eigenvalues of elliptic operators

The following theorem provides a necessary and sufficient condition for the existence of
negative eigenvalues of (3.6.5), and a criterion to estimate the number of negative eigenval-
ues.

Theorem 3.36 For the eigenvalue problem (3.6.5), the following assertions hold true:
3.6 Spectral Theory of Differential Operators 177

1) Equations (3.6.5) have negative eigenvalues if and only if there is a function ψ ∈

H01 (Ω, Cm ), such that
!
[(Dψ )† (Dψ ) + ψ † Aψ ]dx < 0, (3.6.14)
Ω

where D is as in (3.6.6).

2) If there are K linear independent functions ψ1 , · · · , ψK ∈ H01 (Ω, Cm ), such that

ψ satisfies (3.6.14) for any ψ ∈ E K = span {ψ1 , · · · , ψK }, (3.6.15)

then the number N of negative eigenvalues is larger than K, i.e., N " K.

Proof Assertion 1) follows directly from the following classical formula for the first
eigenvalue λ1 of (3.6.5):
! 8
1 9
λ1 = min (Dψ )† (Dψ ) + ψ † Aψ dx.
ψ ∈H0 (Ω,C ) ||ψ ||L2
1 m Ω

We now prove Assertion 2) by contradiction. Assume that it is not true, then K > N. By
Theorem 3.34, the K functions ψ j in (3.6.15) can be expended as
N ∞
ψ j = ∑ α ji ei + ∑ β jl ϕl for 1 ! j ! K, (3.6.16)
i=1 l=1

where ei (1 ! i ! N) and ϕl are eigenfunctions corresponding to negative and nonnegative

eigenvalues. Since K > N, there exists a K-th order matrix P such that
% &
0 ··· 0
Pα = , (3.6.17)
∗

where ⎛ ⎞
α11 ··· α1N
⎜ .. .. ⎟
α =⎝ . . ⎠ with αi j as in (3.6.16).
αK1 ··· αKN
Thus, under the transformation P,
B C
ψ
ψ
'=P ∈ EK, ψ = (ψ1 , · · · , ψN )T , (3.6.18)
0

where E K is as in (3.6.15).
However, by (3.6.14) and (3.6.15), the first term ψ
'1 in (3.6.18) can be expressed in the
form
∞
ψ
'1 = ∑ θl ϕl ∈ E K . (3.6.19)
l=1
178 Chapter 3 Mathematical Foundations

Note that ϕl are the eigenfunctions corresponding to the nonnegative eigenvalues of (3.6.5).
Hence we have
! ! ∞

Ω
'1 )† (Dψ
[(Dψ '1† Aψ
'1 ) + ψ '1 ]dx =
Ω
'1† (−D2 ψ
ψ '1 + Aψ
'1 )dx = ∑ |θl |2 λl > 0. (3.6.20)
l=1

Here λl " 0 are the nonnegative eigenvalues of (3.6.5). Hence we derive, from (3.6.19) and
(3.6.20), a contradiction with the assumption in Assertion (2). The proof of the theorem is
complete.

3.6.4 Estimates for number of negative eigenvalues

For simplicity, it is physically sufficient for us to consider the eigenvalue problem of the
Laplace operators, given by

− ∇2 ψ + V (x)ψ = λ ψ for x ∈ Br ,
(3.6.21)
ψ =0 for x ∈ ∂ Br ,

where Br ⊂ Rn is a ball with radius r.

In physics, V represents a potential function and takes negative value in a bound state,
ensuring by Theorem 3.36 that (3.6.21) possesses negative eigenvalues.
Here, for the potential function V (x), we assume that

V (ρ x) ≃ ρ α V0 (x) (α > −2), (3.6.22)

where V0 (x) is defined in the unit ball B1 , and

Ω = {x ∈ B1 | V0 (x) < 0} ̸= ∅. (3.6.23)

Let θ > 0 be defined by

!
1
θ= inf |V (x)| |ψ |2 dx. (3.6.24)
ψ ∈L (Ω,C ) ||ψ ||L2
2 m Ω

The main result in this section is the following theorem, which provides a relation be-
tween N, θ and r, where N is the number of negative eigenvalues of (3.6.21). Let λ 1 be the
first eigenvalue of the equation

− ∆e = λ e for x ∈ Ω,
(3.6.25)
e=0 for x ∈ ∂ Ω,

where Ω ⊂ B1 is as defined by (3.6.23).

To state the main theorem, we need to introduce a lemma, leading to the Weyl asymptotic
relation (Weyl, 1912).
3.6 Spectral Theory of Differential Operators 179

Lemma 3.37(H. Weyl) Let λN be the N-th eigenvalue of the m-th order elliptic oper-
ator
(−1)m ∆m e = λ e for x ∈ Ω ⊂ Rn ,
(3.6.26)
Dk e|∂ Ω = 0 for 0 ! k ! m − 1,
then λN has the asymptotical relation

λN ∼ λ1 N 2m/n , (3.6.27)

where λ1 is the first eigenvalue of (3.6.26).

We are now ready for the main theorem.

Theorem 3.38 Under the assumptions of (3.6.22) and (3.6.23), the number N of the
negative eigenvalues of (3.6.21) satisfies the following approximative relation
% &n/2
θ r2+α
N≃ , (3.6.28)
λ1

provided that θ r2+α /λ1 ≫ 1 is sufficiently large, where r and θ are as in (3.6.21) and
(3.6.24), and λ1 is the first eigenvalue of (3.6.25).

Proof The ball Br can be written as

Br = {y = rx | x ∈ B1 }.

Note that ∂ /∂ y = r−1 ∂ /∂ x, (3.6.21) can be equivalently expressed as

− ∆ϕ + r2V (rx)ϕ = β ϕ for x ∈ B1 ,

(3.6.29)
ϕ =0 for x ∈ ∂ B1 ,

and the eigenvalue λ of (3.6.21) is

1
λ= β, where β is the eigenvalue of (3.6.29).
r2
Hence the number of negative eigenvalues of (3.6.21) is the same as that of (3.6.29), and we
only need to prove (3.6.28) for (3.6.29).
By (3.6.22), the equation (3.6.29) is approximatively in the form

− ∆ϕ + r2+α V0 (x)ϕ = β ϕ for x ∈ B1 ,

(3.6.30)
ϕ =0 for x ∈ ∂ B1 .

Based on Assertion (2) in Theorem 3.36, we need to find N linear independent functions
ϕn ∈ H01 (B1 ) (1 ! n ! N) satisfying
! 8 9
|∇ϕ |2 + r2+α V0 (x)ϕ 2 dx < 0, (3.6.31)
B1
180 Chapter 3 Mathematical Foundations

for any ϕ ∈ span {ϕ1 , · · · , ϕN } with ||ϕ ||L2 = 1.

To this end, we take the eigenvalues {λn } and eigenfunctions {en } of (3.6.25) such that

0 < λ1 ! · · · ! λN < λN+1 ,

and
λN < θ r2+α ! λN+1 . (3.6.32)

For the eigenfunctions en , we make the extension

@
en for x ∈ Ω,
ϕn =
0 for x ∈ B1 /Ω.

It is known that ϕn is weakly differentiable, and ϕn ∈ H01 (Ω). These functions ϕn (1 ! n !

N) are what we need. Let
N
ϕ= ∑ αn ϕn, ||ϕ ||L2 = 1.
n=1

By Assertion (2) in Theorem 3.36, ϕn (1 ! n ! N) are orthonormal:

! !
ϕi ϕ j dx = ei e j dx = δi j .
B1 Ω

Therefore we have
N
||ϕ ||L2 = ∑ αn2 = 1. (3.6.33)
n=1

Thus the integral in (3.6.31) is

!
[|∇ϕ |2 + r2+α V0 (x)ϕ 2 ]dx
B1
!
B CB C !
N N
= − ∑ αn en ∑ n n
α ∆e dx + r 2+α
V0 (x)ϕ 2 dx
Ω n=1 n=1 Ω
N !
= ∑ αn2 λn + r2+α V0 (x)ϕ 2 dx
n=1 Ω
N
! ∑ αn2 λn − θ r2+α (by (3.6.24))
n=1
<0 (by (3.6.32) and (3.6.33)).

It follows from Theorem 3.36 that there are at least N negative eigenvalues for (3.6.30).
When θ r2+α ≫ 1 is sufficiently large, the relation (3.6.32) implies that

λN ≃ θ r2+α . (3.6.34)
3.6 Spectral Theory of Differential Operators 181

On the other hand, by (3.6.27) in Lemma 3.37,

λN ∼ λ1 N 2/n (m = 1). (3.6.35)

Hence the relation (3.6.28) follows from (3.6.34) and (3.6.35). The proof is complete.

Remark 3.39 In Section 6.4.6, we shall see that for particles with mass m, the param-
eters in (3.6.28) are

α = 0, r = 1, n = 3, θ = 4mρ12 Ag2 /h̄2 ρ ,

where g = gw or gs is the weak or strong interaction charge, ρ is the particle radius, ρ 1 is

the weak or strong attracting radius, and A is the weak or strong interaction constant. Hence
the number of energy levels of massive particles is given by
0 13/2
4 ρ12 A mc g2
N= .
λ1 ρ h̄ h̄c
where λ1 is the first eigenvalue of −∆ in the unit ball B1 .
Example 3.40 (Number of Atomic Energy Levels) As an application of Theorem
3.38, we consider here the estimates for the number of atomic energy levels. Let the atom
number be Z. If ignoring interactions between orbital electrons, then the spectral equation
is as follows
− ∆ψ + V (x)ψ = λ ψ for x ∈ Br0 ,
ψ =0 for x ∈ ∂ Br0 ,
and r0 is the atom radius, V is the potential energy, given by

2Zme2 1
V (x) = − , m the mass of electron.
h̄2 r
The parameters in (3.6.28) for this system are

α = −1, n = 3, r = r0 = 10−8cm. (3.6.36)

In addition, V0 and Ω in (3.6.23) and (3.6.24) are as

2Zme2
V0 = − , Ω = B1 .
h̄2
Therefore, the parameter θ is given by

2Zme2
θ= . (3.6.37)
h̄2
According to physical parameters, it is known that

mc 1 e2 1
r0 × = × 103, = . (3.6.38)
h̄ 4 h̄c 137
182 Chapter 3 Mathematical Foundations

Hence, by (3.6.36)-(3.6.38) the formulas (3.6.28) becomes

% &3/2 % &3/2
2Z e2 mcr0 103 Z
N= = , (3.6.39)
λ1 h̄c h̄ 274λ1

where λ1 is the first eigenvalue of −δ on B1 .

The formulas is derived in the ideal situation ignoring interactions between orbital elec-
trons, and only holds for a bigger atom number Z.

3.6.5 Spectrum of Weyl operators

In Section 6.4.2, we shall deduce from Basic Postulates of Quantum Mechanics that the
spectral equations for the massless subatomic particles are in the following form
ig
− h̄c(⃗σ · ⃗D)2 ϕ + {(⃗σ · ⃗D), A0 }ϕ = iλ (⃗σ · ⃗D)ϕ ,
2 (3.6.40)
ϕ |∂ Ω = 0,

where ϕ = (ϕ1 , ϕ2 )T : Ω → C2 is a complex-valued function with two components, called

the Weyl spinor, ⃗σ = (σ1 , σ2 , σ2 ) is the Pauli matrix operator as given by (3.5.36), ⃗D =
(D1 , D2 , D3 ) is the derivative operator given by

Dk = ∂k + igAk , for k = 1, 2, 3, (3.6.41)

and {(⃗σ · ⃗D), A0 } is the anti-commutator defined by

{(⃗σ · ⃗D), A0 } = (⃗σ · ⃗D)A0 + A0 (⃗σ · ⃗D).

Remark 3.41 The equation (3.6.40) is essentially an eigenvalue problem of the first
order differential operator:
ih̄c(⃗σ · ⃗D) + gA0,
which is called the Weyl operator. In addition, the operator (⃗σ · ⃗D)2 in (3.6.40) is elliptic
and can be written as
g
(⃗σ · ⃗D)2 = D2 − ⃗σ · curl⃗A, (3.6.42)
h̄c
and ⃗A = (A1 , A2 , A3 ) as in (3.6.41). The ellipticity of (3.6.42) g in (3.6.41)-(3.6.42) repre-
sents the weak or strong interaction charge, and A µ = (A0 , A1 , A2 , A3 ) the weak or strong
interaction potential.
Note also that ⃗A = (A1 , A2 , A3 ) stands for the magnetic component of the weak or strong
interaction. Hence, in (3.6.42), the term

g⃗σ · curl⃗A

represents magnetic energy generated by the weak or strong interactions, which is an im-
portant byproduct of the unified field theory based on PID and PRI.
3.6 Spectral Theory of Differential Operators 183

Since (3.6.40) is essentially an eigenvalue problem of first-order differential equations,

its negative and positive eigenvalues are infinite. However, if we only consider the phys-
ically meaningful eigenstates, then the number of negative eigenvalues of (3.6.40) is fi-
nite. We now give introduce these physical meaningful eigenvalues and eigenfunctions for
(3.6.40).

Definition 3.42 A real number λ and a two-component wave function ϕ ∈ H01 (Ω, C2 )
are called the eigenvalue and eigenfunction of (3.6.40), if (λ , ϕ ) satisfies (3.6.40) and
! F G
ϕ † i(⃗σ · ⃗D)ϕ dx > 0. (3.6.43)
Ω

The physical significance of (3.6.43) is that the kinetic energy E of the eigenstate ϕ is
positive: E > 0.
The following theorem ensures the mathematical rationality of the eigenvalue problem
of the Weyl operators.

Theorem 3.43 (Spectral Theorem of Weyl Operators) For the eigenvalue problem
(3.6.40), the following assertions hold true:

1) The eigenvalues of (3.6.40) are real and discrete, with finite multiplicities, and satisfy

−∞ < λ1 ! · · · ! λk ! · · · , λk → ∞ as k → ∞.

2) The eigenfunctions are orthogonal in the sense that

! F G
ψk† i(⃗σ · ⃗D)ψ j dx = 0, ∀k ̸= j. (3.6.44)
Ω

3) The number of negative eigenvalues is finite

−∞ < λ1 ! · · · ! λN < 0.

4) Equations (3.6.40) have negative eigenvalues if and only if there exists a function
ϕ ∈ H01 (Ω, C2 ) satisfying (3.6.43) such that
! 0 1
ig
h̄c|(⃗σ · ⃗D)ϕ |2 + ϕ † {(⃗σ · ⃗D), A0 }ϕ < 0.
Ω 2

Proof It is easy to see that the operator

L = i(⃗σ · ⃗D) : H01 (Ω, C2 ) → L2 (Ω, C2 )

is a Hermitian operator. Consider a functional F : H 1 (Ω, C) → R:

! F G
g
F(ψ ) = h̄c|Lψ |2 + ψ † {L, A0 }ψ dx.
Ω 2
184 Chapter 3 Mathematical Foundations

By (3.6.42), the operator L2 = −(⃗σ · ⃗D)2 is elliptic. Hence F has the following lower bound
on S: > !! ?
!
S = ψ ∈ H01 (Ω, C2 )!! ψ † Lψ dx = 1 ,
Ω

namely
min F(ψ ) > −∞.
ψ ∈S

Based on the Lagrange multiplier theorem of constraint minimization, the first eigenvalue
λ1 and the first eigenfunction ϕ1 ∈ S satisfy

λ1 = F(ψ1 ) = min F(ψ ). (3.6.45)

ψ ∈S

In addition, if
λ1 ! · · · ! λ m
are the first m eigenvalues with eigenfunctions ψk , 1 ! k ! m, then the (m + 1)-th eigenvalue
λm+1 and eigenfunction ψm+1 satisfy

λm+1 = F(ψm+1 ) = min F(ψ ), (3.6.46)

ψ ∈S,ψ ∈Hm⊥

where Hm = span{ψ1 , · · · , ψm }, and Hm⊥ is the orthogonal complement of Hm in the sense

of (3.6.44).
It is clear that Assertions (1)-(4) of the theorem follow from (3.4.45) and (3.4.26). The
proof is complete.

In the following, we consider the estimates of the number of negative eigenvalues for
the Weyl operators. If the interaction potential A µ takes approximatively the following

Aµ = (K, 0, 0, 0) with K > 0 being a constant.

Then (3.6.40) becomes

− ∆ϕ = i(λ + K)(⃗σ ·⃗∂ )ϕ for x ∈ Ω ⊂ Rn ,

(3.6.47)
ϕ =0 for x ∈ ∂ Ω.

Obviously, the number N of negative eigenvalues of (3.6.47) satisfies

βN < K < βN+1 , (3.6.48)

where βk is the k-th eigenvalue of the equation

− ∆ϕk = iβk (⃗σ ·⃗∂ )ϕk ,

(3.6.49)
ϕk |∂ Ω = 0.
3.6 Spectral Theory of Differential Operators 185

For (3.6.49) we have the Weyl asymptotical relation

βN ∼ β1 N 1/n . (3.6.50)

Here the exponent is 1/n, since (3.6.47) is an 2m-th order elliptic equation with m = 1/2.
Hence we deduce, from (3.6.48) and (3.6.50), the estimates of the number N of negative
eigenvalues of (3.6.47) as % &n
K
N≃ , (3.6.51)
β1
where β1 is the first eigenvalue of (3.6.49).

Remark 3.44 For the mediators such as the photon and gluons, the number of energy
levels is given by N in (3.6.51), which can be estimated as
% &3
A ρ1 g2w
N= ,
β1 ρ h̄c

where ρ1 , ρ , A, gw are as in Remark 3.39.

Chapter 4
Unified Field Theory of Four Fundamental
Interactions

Once again, the goal of this book is to derive experimentally verifiable laws of Nature based
on a few fundamental mathematical principles. The aims of this chapter are are as follows:

• to address the basic principles for the unified field theory coupling the four funda-
mental interactions,

• to derive the unified model based on these principles, and

• to study the mechanism and nature of individual interactions.

This chapter is based entirely on the recent work of the authors (Ma and Wang, 2015a,
2014h,c, 2013a, 2014e). The key ingredients of the unified field theory include the follow-
ing.
First, we have established two new principles, the principle of interaction dynamics
(PID) and the principle of representation invariance (PRI). PID was first discovered by (Ma
and Wang, 2014e, 2015a). It requires that for the four fundamental interactions, the variation
be taken under the energy-momentum conservation constraints. The validity of PID for the
four fundamental interactions of Nature has been demonstrated through strong experimental
and observation supports. For gravity, PID is induced by the presence of the dark matter and
dark energy phenomena. PID is also required by the Higgs field and the quark confinement,
as we explained in Chapter 1.
PRI, originally discovered by (Ma and Wang, 2014h), states that the SU(N) gauge theory
should be invariant, under the representation transformations of the generator bases. PRI is
simply a logic requirement for the SU(N) theory.
Second, with PRI, the unification through a large symmetry group appears to be not
feasible. Then we have demonstrated that the two first principles, PID and PRI, together
with the principle of symmetry-breaking, offer an entirely different route for the unification:

1) the general relativity and the gauge symmetries dictate the Lagrangian;
and
4.1 Principles of Unified Field Theory 187

2) the coupling of the four interactions is achieved through PID and PRI
in the unified field equations, which obey the PGR and PRI, but break
spontaneously the gauge symmetry.

In this chapter, there are four sections. Section 4.1 presents the general view of the
unified field theory, the geometry of unified fields, gauge-symmetry breaking, PID, and
PRI. Section 4.2 presents the experimental and observational physical supports for PID.
The mathematical reason from the well-posedness point of view is also given. Section 4.3
introduces the unified field equations coupling the four fundamental interactions based on
PID. Section 4.4 presents the natural duality between the gauge fields and their dual fields,
as well as the decoupling of the unified field equations to the field equations for individual
interaction when the other interactions are negligible.

4.1 Principles of Unified Field Theory

4.1.1 Four interactions and their interaction mechanism
The four fundamental interactions/forces of Nature include

1) the gravitational force, generated by the mass charge M, which is responsible to all
macroscopic motions;

2) the electromagnetic force, generated by the electric charge e, which holds the atoms
and molecules together;

3) the strong force, generated by the strong charge g s , which mainly acts at three levels:
quarks and gluons, hadrons, and nucleons;

4) the weak force, generated by the weak charge g w , which provides the binding energy
to hold the mediators, the leptons and quarks together.

The most crucial ingredient of each interaction is its corresponding interaction potential
Φ and charge g. The relation between the corresponding force F, its associated potential Φ
and its charge g is as follows:
F = −g∇Φ, (4.1.1)
where ∇ is the three-dimensional spatial gradient operator. The charge g for each interaction
is given as follows:

m the mass charge for gravity,

e the electric charge for electromagnetism,
(4.1.2)
gs the strong charge for strong interaction, and
gw the weak charge for the weak interaction.
188 Chapter 4 Unified Field Theory of Four Fundamental Interactions

We know now that the four interactions are dictated respectively by the following sym-
metry principles:

gravity: principle of general relativity,

electromagnetism: U(1) gauge invariance,
(4.1.3)
weak interaction: SU(2) gauge invariance,
strong interaction: SU(3) gauge invariance.

The last three interactions also obey the Lorentz invariance. As a natural consequence, the
three charges e, gw , gs in (4.1.2) are the coupling constants of the U(1), SU(2), SU(3) gauge
fields.
Following the simplicity principle of laws of Nature as stated in Principle 2.2, the three
basic symmetries—the Einstein general relativity, the Lorentz invariance and the gauge
invariance—uniquely determine the interaction fields and their Lagrangian actions for the
four interactions:
1. Gravity. The gravitational fields are the Riemannian metric defined on the space-time
manifold M :
ds2 = g µν dxµ dxν , (4.1.4)
and then second-order tensor {g µν } stands for the gravitational potential. The Lagrangian
action for the metric (4.1.4) is the Einstein-Hilbert functional
8π G
LEH = R + S, (4.1.5)
c4
where R stands for the scalar curvature of the tangent bundle T M of M .
2. Electromagnetism. The field describing electromagnetic interaction is the U(1) gauge
field
Aµ = (A0 , A1 , A2 , A3 ),
representing the electromagnetic potential, and the Lagrangian action is
1
LEM = − Aµν Aµν , (4.1.6)
4
which stands for the scalar curvature of the vector bundle M ⊗ p C4 , with

Aµν = ∂µ Aν − ∂ν Aµ .

being the curvature tensor.

3. Weak interaction. The weak fields are the SU(2) gauge fields

Wµa = (W0a ,W1a ,W2a ,W3a ) for 1 ! a ! 3,

4.1 Principles of Unified Field Theory 189

and their action is

1 w a µν b
LW = − Gab Wµν W , (4.1.7)
4
which also stands for the scalar curvature of the spinor bundle: M ⊗ p (C4 )2 . Here
a
Wµν = ∂µ Wνa − ∂ν Wµa + gw λbc
a
WµbWνc for 1 ! a ! 3.

4. Strong interaction. The strong fields are the SU(3) gauge fields

Skµ = (S0k , S1k , S2k , S3k ) for 1 ! k ! 8,

and the action is

1
LS = − Gkls Skµν S µν l , (4.1.8)
4
which corresponds to the scalar curvature of M ⊗ p (C4 )3 . Here

Skµν = ∂µ Sνk − ∂ν Skµ + gs Λkrl Srµ Sνl for 1 ! k ! 8.

The Yukawa Interaction Mechanism, briefly mentioned in Section 2.1.6 and restated
below, is the main reason why the weak interaction is described by an SU(2) gauge theory
and the strong interaction is described by an SU(3) gauge theory.
One great vision of Albert Einstein is his principle of equivalence, which says that grav-
ity is manifested as the curved effect of the space-time manifold {M , g µν }. Based on the
recent work by the authors (Ma and Wang, 2015a, 2014h,d), the Einstein vision leads us
to postulate the Geometric Interaction Mechanism 2.13, which is restated here for conve-
nience:

Geometric Interaction Mechanism 4.1 The gravitational force is the curved effect of
the time-space, and the electromagnetic, weak, strong interactions are the twisted effects of
the underlying complex vector bundles M ⊗ p Cn .

Yukawa’s viewpoint, entirely different from Einstein’s, is that the other three funda-
mental forces—the electromagnetism, the weak and the strong interactions——take place
through exchanging intermediate bosons:

Yukawa Interaction Mechanism 4.2 The four fundamental interactions of Nature are
mediated by exchanging interaction field particles, called the mediators. The gravitational
force is mediated by the graviton, the electromagnetic force is mediated by the photon, the
strong interaction is mediated by the gluons, and the weak interaction is mediated by the
intermediate vector bosons W ± and Z.
190 Chapter 4 Unified Field Theory of Four Fundamental Interactions

It is the Yukawa mechanism that leads to the SU(2) and SU(3) gauge theories respec-
tively for the weak and the strong interactions. In fact, the three mediators W ± and Z for
the weak interaction are regarded as the SU(2) gauge fields Wµa (1 ! a ! 3), and the eight
gluons for the strong interaction are considered as the SU(3) gauge fields S kµ (1 ! k ! 8).
Of course, the three color quantum numbers for the quarks are an important evidence for
choosing the SU(3) gauge theory to describe the strong interaction.
The two interaction mechanisms lead to two entirely different directions to develop the
unified field theory. The need for quantization for all current theories for the four interac-
tions is based on the Yukawa Interaction Mechanism. The new unified field theory in this
article is based on the Geometric Interaction Mechanism, which focuses directly on the four
interaction forces as in (4.1.1), and does not involve a quantization process.
A radical difference for these two mechanisms is that the Yukawa Mechanism is ori-
ented toward to computing the transition probability for the particle decays and scatterings,
and the Geometric Interaction Mechanism is oriented toward to fundamental laws, such as
interaction potentials, of the four interactions.

4.1.2 General introduction to unified field theory

Einstein’s unification

The aim of a unified field theory is to establish a set of field equations coupling the four
fundamental interactions. Albert Einstein was the first person who attempted to establish a
unified field theory. The basic philosophy of his unification is that all fundamental forces
of Nature should be dictated by one large symmetry group, which can degenerate into a
sub-symmetry for each interaction:

Unification through an action under a large symmetry (4.1.9)

In essence, with the Einstein unification, under the large symmetry, the four fundamental
forces can be regarded as one fundamental force.
Recall that there are four fundamental interactions of Nature: gravitational, electromag-
netic, strong, and weak, whose fields and actions:

1) fields g µν , Aµ , {Wµa | 1 ! a ! 3} and {Skµ | 1 ! k ! 8}, and

2) their actions LEH , LEM , LW and LS ,

are dictated by the symmetries in (4.1.3), as described in (4.1.5)-(4.1.8).

Basically, the Einstein unification is to search for a large symmetry, which dictates an N
component field G:
G = (G1 , · · · , GN ), (4.1.10)
4.1 Principles of Unified Field Theory 191

and an action
L = L (G1 , · · · , GN ). (4.1.11)
The basic requirements for such a unification are as follows: Under certain conditions,

1) the large symmetry degenerates into sub-symmetries: the general invariance, the
Lorentz invariance, and the U(1), SU(2), SU(3) invariance;

2) the field G of (4.1.10) is then decomposed into the fields of the four fundamental
interactions:
degenerate
(G1 , · · · , GN ) −→ g µν , Aµ ,Wµa , Skµ , (4.1.12)

3) the action (4.1.11) also becomes the simple sum of these actions given in (4.1.5)-
(4.1.8):
L (G1 , · · · , GN ) → LEH + LEM + LW + LS . (4.1.13)

For almost a century, a great deal of effort has been made to find the unified field model
based on the above mentioned approach. However, all efforts in this aspect are not suc-
cessful. In fact, among other reasons, this route of unification violates the principle of
representation invariance (PRI), discovered in (Ma and Wang, 2014h); see also remaining
part of this chapter for details.
Moreover, the basic principles and the field equations from all attempted unified field
theories based on (4.1.9) are often too complex, and violate the simplicity principle of
physics. Most importantly, despite of many attempts, the current theories following (4.1.9)
fail to provide solutions to the following longstanding problems and challenges:

1) interaction force formulas

2) quark confinement,

3) asymptotical freedom,

4) strong interaction potentials of nucleus,

5) dark matter and dark energy phenomena,

6) decoupling to the individual interactions as required by (4.1.13),

7) spontaneous symmetry breaking from first principles,

8) mechanism of subatomic particle decay and scattering,

9) violation of PRI, and

10) reasons why leptons do not participate in strong interactions.

192 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Unified Field Theory based on PID, PRI and PSB

The unified field theory introduced in this chapter is based on the three new principles,
PID, PRI and PSB, postulated recently by the authors. In this theory, the Lagrangian is a
simple sum of the known four actions in in (4.1.5)-(4.1.8):

L = LEH + LEM + LW + LS , (4.1.14)

and with which the unification is achieved as follows:

Unification through the Field Equations Based on PID, PRI and PSB (4.1.15)

A few remarks are now in order.

First, the unified field model of (4.1.14) and (4.1.15) not only solves the most basic
problems 1)-10) mentioned above, but also is the simplest with respect to the underlying
physical principles and to the explicit form of the field equations coupling four forces.
Second, the new unified field theory based on PID, PRI and PSB addressed in this chap-
ter offers an answer to dark energy and dark matter problem; see also Section 7.6.3.
Third, thanks to PRI, we have shown that the classical SU(3) Yang-Mills theory will
only provide a repulsive force. The attractive bounding force between quarks are due to the
dual fields in the PID-induced SU(3) gauge theory. In other words, the quark confinement
problem is solved in (Ma and Wang, 2014c), and will be addressed in detail in Section 4.5.3.
4.1 Principles of Unified Field Theory 193

Also, the route of unification (4.1.14) and (4.1.15) is readily applied to multi-particle
interacting systems, and gives rise to a first dynamic interacting model for multi-particle
systems; see Chapter 6 for details.
Finally, we present a diagram to illustrate the framework of the unified field theory,
based on (4.1.14) and (4.1.15). We note that quantization is used mainly for deriving tran-
sition probability from the field equations for each interaction.

4.1.3 Geometry of unified fields

Hereafter we always assume that the space-time manifold M of our Universe is a 4-dimensional
Riemannian manifold. We adopt the view that symmetry principles determine the geometric
structure of M , and the geometries of M associated with the fundamental interactions of
Nature dictate all motion laws defined on M . The process that symmetries determine the
geometries of M and the associated vector bundles is achieved in the following three steps:

1) The symmetric principles, such as the Einstein general relativity, the Lorentz invari-
ance, and the gauge invariance, determine that the fields reflecting geometries of M
are the Riemannian metric {g µν } and the gauge fields {Gaµ }. In addition, the sym-
metric principles also determine the Lagrangian actions of g µν and Gaµ ;

2) PID determines the field equations governing g µν and Gaµ ; and

3) The solutions g µν and Gaµ of the field equations determine the geometries of M and
the vector bundles.

The geometry of the unified fields refers to the geometries of M , determined by the
following known physical symmetry principles:

principle of general relativity,

principle of Lorentz invariance,
(4.1.16)
U(1) × SU(2) × SU(3) gauge invariance,
principle of representation invariance (PRI).

We shall introduce the two principles PID and PRI in Section 4.1.5.
The fields determined by the symmetries in (4.1.16) are given by

• general relativity: g µν : M → T20 M , the Riemannian metric,

• Lorentz invariance: ψ : M → M ⊗ p (C4 )N , the Dirac spinor fields,

• U(1) gauge invariance: A µ : M → T ∗ M , the U(1) gauge field,

• SU(2) gauge invariance: Wµa : M → (T ∗ M )3 , the SU(2) gauge fields,

194 Chapter 4 Unified Field Theory of Four Fundamental Interactions

• SU(3) gauge invariance: S kµ : M → (T ∗ M )8 , the SU(3) gauge fields.

The Lagrange action for the geometry of the unified fields is given by
!
√
L= [LEH + LEM + LW + LS + LD ] −gdx (4.1.17)
M

where LEH , LEM , LW and LS are the Lagrangian actions for the four interactions defined
by (4.1.5)–(4.1.8), and the action for the Dirac spinor fields is given by

LD = Ψ(iγ µ Dµ − m)Ψ. (4.1.18)

Here

Ψ = (ψ e , ψ w , ψ s ),
m = (me , mw , ms ),

and
ψ e : M → M ⊗ p C4 1-component Dirac spinor,
w 4 2
ψ : M → M ⊗ p (C ) 2-component Dirac spinors, (4.1.19)
ψ s : M → M ⊗ p (C4 )3 3-component Dirac spinors.

The derivative operators D µ are given by

Dµ ψ e = (∂µ + ieAµ )ψ e ,
Dµ ψ w = (∂µ + igwWµa σa )ψ w , (4.1.20)
s
Dµ ψ = (∂µ + igsSkµ τk )ψ s .

The geometry of unified fields consists of 1) the field functions and 2) the Lagrangian
action (4.1.17), which are invariant under the following seven transformations:

µ T µ
1) the general linear transformation Q p = (aν ) : Tp M → Tp M with Q −1
p = (bν ) , for
any p ∈ M :
g µν ) = Q p (g µν )Q Tp ,
('
'µ = aνµ Aν ,
A
W'µa = aνµ Wνa for1 ! a ! 3, (4.1.21)
S'kµ = aνµ Sνk for1 ! k ! 8,
µ
γ'µ = bν γ ν , ∂'µ = aνµ ∂ν ,

with no change on other fields, where γ µ are the Dirac matrices;

4.1 Principles of Unified Field Theory 195

2) the Lorentz transformation on Tp M :

L = (lµν ) : Tp M → Tp M , L is independent of p ∈ M,
g µν ) = L(g µν )LT ,
(' 'µ = lµν Aν ,
A
W'µa = lµν Wνa for 1 ! a ! 3,
(4.1.22)
S'kµ = lµν Sνk for 1 ! k ! 8,
Ψ' = RL Ψ, RL is the spinor transformation matrix,
∂'µ = l ν ∂ν ,
µ

with no change on other fields;

3) the U(1) gauge transformation on M ⊗ p C4 :

Ω = eiθ : C4 → C4 ∈ U(1),
D E % 1
&
(4.1.23)
e iθ e
ψ' , Aµ = e ψ , Aµ − ∂µ θ ,
'
e

4) SU(2) gauge transformation on M ⊗ p (C4 )2 :

a
Ω = eiθ σa : (C4 )2 → (C4 )2 ∈ SU(2),
D E % i
&
(4.1.24)
ψ'w,W'µa σa , m
'w = Ωψ w ,Wµa Ωσa Ω−1 + ∂µ ΩΩ−1 , Ωmw Ω−1 .
gw

5) SU(3) gauge transformation on M ⊗ p (C4 )3 :

k
Ω = eiθ τk : (C4 )3 → (C4 )3 ∈ SU(3),
D E % i
& (4.1.25)
ψ' s , S'kµ τk , m
' s = Ωψ s , Skµ Ωτk Ω−1 + ∂µ ΩΩ−1 , Ωms Ω−1 .
gs

6) SU(2) representation transformation on Te SU(2):

X = (xba ) : Te SU(2) → Te SU(2), (yab )T = X −1 ,

σ's = xba σb , 'w ) = X(Gw )X T ,
(G ab ab (4.1.26)
'µa
W = yabWµb .

7) SU(3) representation transformation on Te SU(3):

X = (xlk ) : Te SU(3) → Te SU(3), (ykl )T = X −1 ,

τ'k = xlk τl , 's ) = X(Gs )X T ,
(G kl kl (4.1.27)
S'kµ = yk Slµ .
l
196 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Remark 4.3 Here we adopt the linear transformations of the bundle spaces instead
of the coordinate transformations in the base manifold M . In this case, the two transfor-
mations (4.1.21) and (4.1.22) are compatible. Otherwise, we have to introduce the Vierbein
tensors to overcome the incompatibility between the Lorentz transformation and the general
coordinate transformation.

4.1.4 Gauge symmetry-breaking

In physics, symmetries are displayed in two levels in the laws of Nature:

the invariance of Lagrangian actions L, (4.1.28)

the covariance of variation equations of L. (4.1.29)

The following three symmetries:

the Einstein principle of general relativity (PGR),

the Lorentz invariance, (4.1.30)
the principle of representation invariance (PRI),

represent the universality of physical laws— the validity of laws of Nature is independent of
the coordinate systems expressing them. Consequently, the symmetries in (4.1.30) cannot
be broken at both levels of (4.1.28) and (4.1.29).
The physical implication of the gauge symmetry, however, is different at the two levels:

(1) the gauge invariance of the Lagrangian action, (4.1.28), says that the energy contri-
butions of particles in a physical system are indistinguishable; and

(2) the gauge invariance of the variational equations, (4.1.29), means that the particles
involved in the interaction are indistinguishable.

It is clear that the first aspect (1) above is universally true, while the second aspect (2)
is not universally true. In other words, the Lagrange actions obey the gauge invariance,
but the corresponding variational equations break the gauge symmetry. This suggests us to
postulate the following principle of gauge symmetry breaking for interactions described by
the gauge theory.

Principle 4.4(Gauge Symmetry Breaking)

1) The gauge symmetry holds true only for the Lagrangian actions for the electromag-
netic, week and strong interactions; and

2) the field equations of these interactions spontaneously break the gauge symmetry.
4.1 Principles of Unified Field Theory 197

The principle of gauge symmetry breaking can be regarded as part of the spontaneous
symmetry breaking, which is a phenomenon appearing in various physical fields. In 2008,
the Nobel Prize in Physics was awarded to Y. Nambu for the discovery of the mechanism
of spontaneous symmetry breaking in subatomic physics. In 2013, F. Englert and P. Higgs
were awarded the Nobel Prize for the theoretical discovery of a mechanism that contributes
to our understanding of the origin of mass of subatomic particles.
This phenomenon was discovered in superconductivity by Ginzburg-Landau in 1951,
and the mechanism of spontaneous symmetry breaking in particle physics was first pro-
posed by Y. Nambu in 1960; see (Nambu, 1960; Nambu and Jona-Lasinio, 1961a,b). The
Higgs mechanism, introduced in (Higgs, 1964; Englert and Brout, 1964; Guralnik, Hagen
and Kibble, 1964), is an artificial method based on the Nambo-Jona-Lasinio spontaneous
symmetry breaking, leading to the mass generation for the vector bosons of the weak inter-
action.
PID discovered by the authors, to be stated in detail in the next section, provides a new
mechanism for gauge symmetry breaking and mass generation. The difference between the
PID and the Higgs mechanisms is that the first one is a natural sequence of the first principle,
and the second is to add artificially a Higgs field in the Lagrangian action. Also, the PID
mechanism obeys PRI, and the Higgs mechanism violates PRI. symmetry-breaking!PID-
induced

4.1.5 PID and PRI

The main objective in this subsection is to postulate two fundamental principles of physics,
the principle of interaction dynamics (PID) and the principle of representation invariance
(PRI), which are based on rigorous mathematical foundations established in Sections 3.3-
3.5.
Let {M , g µν } be the 4-dimensional space-time Riemannian manifold with {g µν } the
Minkowski type Riemannian metric. For an (r, s)-tensor u we define the A-gradient and
A-divergence operators ∇A and divA as

∇A u = ∇u + u ⊗ A,
divA u = div u − A · u,

where A is a vector field and here stands for a gauge field, ∇ and div are the usual gradient
and divergent covariant differential operators. Let F = F(u) be a functional of a tensor field
u. A tensor u0 is called an extremum point of F with the divA -free constraint, if u0 satisfies
the equation
!
d !! √
! F(u0 + λ X) = δ F(u0 ) · X −gdx = 0, ∀X with divA X = 0. (4.1.31)
d λ λ =0 M

Principle 4.5(Principle of Interaction Dynamics)

198 Chapter 4 Unified Field Theory of Four Fundamental Interactions

1) For all physical interactions there are Lagrangian actions

!
√
L(g, A, ψ ) = L (g µν , A, ψ ) −gdx, (4.1.32)
M

where g = {g µν } is the Riemannian metric representing the gravitational potential,

A is a set of vector fields representing the gauge potentials, and ψ are the wave
functions of particles;

2) The action (4.1.32) satisfy the invariance of general relativity, Lorentz invariance,
gauge invariance and the gauge representation invariance;

3) The states (g, A, ψ ) are the extremum points of (4.1.32) with the divA -free constraint
(4.1.31).

Based on PID and Theorems 3.26 and 3.27, the field equations with respect to the action
(4.1.32) are given in the form
δ
L(g, A, ψ ) = (∇µ + αb Abµ )Φν , (4.1.33)
δ g µν
δ
L(g, A, ψ ) = (∇µ + βba Abµ )ϕ a , (4.1.34)
δ Aaµ

δ
L(g, A, ψ ) = 0 (4.1.35)
δψ
where Aaµ = (Aa0 , Aa1 , Aa2 , Aa3 ) are the gauge vector fields for the electromagnetic, the weak
and strong interactions, Φν = (Φ0 , Φ1 , Φ2 , Φ3 ) in (4.1.33) is a vector field induced by grav-
itational interaction, ϕ a is the scalar fields generated from the gauge fields Aaµ , and αb , βba
are coupling parameters.
Consider the action (4.1.32) as the natural combination of the actions for all four inter-
actions, as given in (4.1.17)-(4.1.20):

L = LEH + LEM + LW + LS + LD .

Then (4.1.33)-(4.1.35) provide the unified field equations coupling all interactions. More-
over, we see from (4.1.33)-(4.1.35) that there are too many coupling parameters which need
to be determined. Fortunately, this problem can be satisfactorily resolved, leading also to
the discovery of PRI (Ma and Wang, 2014h). Meanwhile, we remark that it is the gauge
fields Aaµ appearing on the right-hand sides of (4.1.33) and (4.1.34) that break the gauge
symmetry, leading to the mass generation of the vector bosons for the weak interaction
sector.
We are now in position to introduce the principle of representation invariance (PRI). We
end this section by recalling the principle of representation invariance (PRI) first postulated
4.1 Principles of Unified Field Theory 199

in (Ma and Wang, 2014h). We proceed with the SU(N) representation. In a neighborhood
U ⊂ SU(N) of the unit matrix, a matrix Ω ∈ U can be written as
aτ
Ω = eiθ a
,

where
τa = {τ1 , · · · , τK } ⊂ Te SU(N), K = N 2 − 1, (4.1.36)
is a basis of generators of the tangent space Te SU(N); see Section 3.5 for the mathematical
theory. An SU(N) representation transformation is a linear transformation of the basis in
(4.1.36) as
τ'a = xba τb , (4.1.37)
where X = (xba ) is a nondegenerate complex matrix.
Mathematical logic dictates that a physically sound gauge theory should be invariant
under the SU(N) representation transformation (4.1.37). Consequently, the following prin-
ciple of representation invariance (PRI) must be universally valid and was first postulated in
(Ma and Wang, 2014h).

Principle 4.6(Principle of Representation Invariance) All SU(N) gauge theories are

invariant under the transformation (4.1.37). Namely, the actions of the gauge fields are
invariant and the corresponding gauge field equations as given by (4.1.33)-(4.1.35) are
covariant under the transformation (4.1.37).

Both PID and PRI are very important. As far as we know, it appears that the only unified
field model, which obeys not only PRI but also the principle of gauge symmetry breaking,
Principle 4.4, is the unified field model based on PID introduced in this chapter. From this
model, we can derive not only the same physical conclusions as those from the standard
model, but also many new results and predictions, leading to the solution of a number of
longstanding open questions in physics, including the 10 problems mentioned in Section
4.1.2.

A few further remarks on PID and PRI are now in order.

First, there are strong theoretical, experimental and observational evidence for PID; see
the next section for details.
Second, PID is based on variations with divA -free constraint defined by (4.1.31). Phys-
ically, the divA -free condition: divA X = 0 stands for the energy-momentum conservation
constraints.
Third, PRI means that the gauge theory is universally valid, and therefore should be
independent of the choice of the generators τa of SU(N). In other words, PRI is basic a
logic requirement.
200 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Fourth, the electroweak interactions is a U(1) × SU(2) gauge theory coupled with the
Higgs mechanism. An unavoidable feature for the Higgs mechanism is that the gauge fields
with different symmetry groups are combined linearly into terms in the corresponding gauge
field equations. For example, in the Weinberg-Salam electroweak gauge equations with
U(1) × SU(2) symmetry breaking, there are such linearly combinations as

Zµ = cos θW Wµ3 + sin θW Bµ ,

Aµ = − sin θW Wµ3 + cos θW Bµ , (4.1.38)
1
Wµ± = √ (Wµ1 ± iWµ2 ),
2
where Wµa (1 ! a ! 3) are the SU(2) gauge fields, and B µ is the U(1) gauge field. It is
clear that these terms in (4.1.38) are not covariant under the general SU(2) representation
transformations. Hence, the Higgs mechanism violates the PRI. Since the standard model
is based on the Higgs mechanism, it violates PRI as well.

4.2 Physical Supports to PID

The original motivation for PID was to explain the dark matter and dark energy phenomena.
We have demonstrated that the presence of dark matter and dark energy leads directly to
PID for gravity.
There are strong theoretical, experimental and observational evidence for PID. The need
of spontaneous symmetry-breaking for generating mass of the vector bosons for the weak
interaction is a physical evidence for PID for the weak interaction. The quark confinement
requires the introduction of the dual gluon fields demonstrates the necessity of PID; see
Section 4.5. In this section, we address the physical evidence from the following viewpoints:

1) the discovery of dark matter and dark energy,

2) the non-existence of solutions for the classical Einstein gravitational field equations
in general situations,

3) the principle of spontaneous gauge-symmetry breaking,

4) the Ginzburg-Landau superconductivity theory, and

5) the gauge-fixing problem.

4.2.1 Dark matter and dark energy

The presence of dark matter and dark energy provides a strong support for PID. We recall
the Einstein gravitational equations, which are expressed as
1 4π G
Rµν − gµν R = − 4 Tµν , (4.2.1)
2 c
4.2 Physical Supports to PID 201

where Tµν is the usual energy-momentum tensor of visible matter. By the Bianchi identity,
the left-hand side of (4.2.1) is divergence-free, i.e.

1
∇µ (Rµν − gµν R) = 0. (4.2.2)
2
It implies that the usual energy-momentum tensor satisfies

∇µ Tµν = 0. (4.2.3)

However, due to the presence of dark matter and dark energy, the energy-momentum
tensor of visible matter Tµν may no longer be conserved, i.e. (4.2.3) is not true. Hence we
have
∇µ Tµν ̸= 0,

which is a contradiction to (4.2.1) and (4.2.2).

On the other hand, by the Orthogonal Decomposition Theorem 3.17, Tµν can be orthog-
onally decomposed into
c4
Tµν = T'µν − ∇µ Φν , (4.2.4)
8π G
and T'µν is divergence-free:
∇µ T'µν = 0. (4.2.5)

Hence, by (4.2.2) and (4.2.5) the gravitational field equations (4.2.1) should be in the form

1 8π G
Rµν − gµν R = − 4 T'µν . (4.2.6)
2 c
By (4.2.4) we have
c4
T'µν = Tµν + ∇µ Φν ,
8π G
which stands for all energy and momentum including the visible and the invisible matter
and energy, and which, by (4.2.5), is conserved. Thus, the equations (4.2.6) are rewritten as

1 8π G
Rµν − gµν R = − 4 Tµν − ∇µ Φν , (4.2.7)
2 c
The equations (4.2.7) are just the variational equations of LEH with the div-free constraint
as (4.1.33). Namely, (4.2.7) are the gravitational field equations obeying PID.
We remark that the term ∇ µ Φν in (4.2.7) has no variational structure, and cannot be
derived by modifying the Einstein-Hilbert functional. Hence, (4.2.7) are just the variational
equations due to PID.
202 Chapter 4 Unified Field Theory of Four Fundamental Interactions

4.2.2 Non well-posedness of Einstein field equations

The second strong theoretical evidence to PID is that the classical Einstein gravitational
field equations are an over-determined system.
The Einstein field equations (4.2.1) possess 10 unknown functions g µν and 10 inde-
pendent equations due to symmetry for the indices µ and ν . However, since the general
coordinate system can be arbitrarily chosen, under proper coordinate transformation

x'µ = aνµ xν , 0 ! ν ! 3,

the 10 unknown functions become

% &
−1 0
, gi j = g ji for 1 ! i, j ! 3.
0 gi j
This observation implies that the number of independent unknown functions for the Einstein
field equations (4.2.1) is six. Namely,

NEQ = 10,
(4.2.8)
NUF = 6,
where NEQ is the number of independent equations in (4.2.1), and NUF is the number of
independent unknown functions.
Consequently, the Einstein field equations (4.2.1) have no solutions in the general case.
Some readers may think that the Bianchi identity
% &
1 8π G
∇µ Rµν − gµν R + 4 Tµν = 0, (4.2.9)
2 c
reduce the number NEQ to six: NEQ = 6. But we note that (4.2.9) generates also four new
equations
∇µ Tµν = 0 for 0 ! ν ! 3,
because there are unknown functions g µν in the covariant derivative operators ∇ µ ; see
(3.1.66) or (2.3.26). Hence, the Einstein field equations (4.2.1) should be in the form
1 8π G
Rµν − gµν R = − 4 Tµν ,
2 c (4.2.10)
∇µ Tµν = 0.
Thus, the fact (4.2.8) still holds for (4.2.10).
Now we note the gravitational field equations (4.2.7) derived from PID, where there
are four additional new unknown functions Φν (0 ! ν ! 3). In this case, the numbers of
independent unknown functions and equations are the same.
In the following, we give an example to show the non well-posedness of the classical
Einstein field equations.
4.2 Physical Supports to PID 203

It is known that the metric of central gravitational field takes the form

ds2 = c2 g00 dt 2 + g11dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ), (4.2.11)

where (ct, r, θ , ϕ ) is the spherical coordinate system. The metric g µν in (4.2.11) can be
expressed in the form
g00 = −eu (u = u(r)),
v
g11 = e (v = v(r)),
2
g22 = r , (4.2.12)
2 2
g33 = r sin θ ,
gµν = 0 for µ ̸= ν .

Consider the influence of cosmic microwave background (CMB) radiation, the energy-
momentum tensor can be approximatively written as
% &
−g00ρ 0
Tµν = , (4.2.13)
0 0

where ρ is the energy density, a constant.

For the metric (4.2.12), the nonzero components of the Ricci tensor are
0 1
u′′ u′ u′ ′
R00 = −eµ −ν + + (u − v′ ) ,
2 r 4
u′′ v′ u′ ′
R11 = − + (u − v′ ),
2 r 4 (4.2.14)
F r G
R22 = e−v 1 − ev + (u′ − v′ ) ,
2
R33 = sin2 θ R22 .

On the other hand, equations (4.2.10) can be equivalently written as

8π G 1
Rµν = − (Tµν − gµν T ),
c4 2 (4.2.15)
∇µ Tµν = 0,

and by (4.2.13),
T = g µν Tµν = g00 T00 = −ρ .

Thus, the Einstein field equations for the spherically symmetric gravitation fields (4.2.15)
204 Chapter 4 Unified Field Theory of Four Fundamental Interactions

are in the form

4π G
R00 = g00 ρ ,
c4
4π G
R11 = − g11 ρ ,
c4 (4.2.16)
4π G
R22 = − 4 g22 ρ ,
c
∇µ Tµν = 0.
Now, we deduce that the equations (4.2.16) have no solutions. By D µ Tµν = 0, we have
1
Γ010 T00 = u′ ρ = 0,
2
which implies that u′ = 0. Hence, by (4.2.14) we have

R00 = 0,

which is a contradiction to the first equation of (4.2.16). Therefore the equations (4.2.16)
have no solutions.
However, if we consider this example by using the field equations derived from PID,
then the problem must have a solutions; see the theory of dark matter and dark energy in
Chapter 7.

4.2.3 Higgs mechanism and mass generation

Principle 4.4 of gauge symmetry breaking is also a main motivation to postulate PID in
our program for a unified field theory. In fact, the Higgs mechanism is one way to achieve
the spontaneous gauge-symmetry breaking. In the Glashow-Weinberg-Salam (GWS) elec-
troweak theory, the three intermediate vector bosons W ± and Z for the weak interaction
retain their masses by the Higgs mechanism. We now show that the masses of the inter-
mediate vector bosons can be also obtained by PID. Furthermore, we shall show in Section
4.6 that all conclusions of the GWS electroweak theory confirmed by experiments can be
derived by the unified field theory based on PID.
For convenience, we first introduce some related basic knowledge of quantum physics.
In quantum field theory, a field ψ is called a fermion with mass m, if it satisfies the Dirac
equation
(iγ µ Dµ − m)ψ = 0, (4.2.17)
and the action of (4.2.17) is
!
LF = LF dx, LF = ψ (iγ µ Dµ − m)ψ , (4.2.18)

where ψ = ψ † γ 0 .
4.2 Physical Supports to PID 205

A field Φ is called a boson with mass m, if Φ satisfies a Klein-Gordon type of wave

equation:
D mc E2
#Φ + Φ = g(Φ), (4.2.19)
h̄
where g(Φ) is the terms of Φ other than kΦ (k a constant), and # is the wave operator given
by
1 ∂2
# = −∂ µ ∂µ = 2 2 − ∆.
c ∂t
The bosonic field Φ is massless if it satisfies

#Φ = g(Φ). (4.2.20)

The physical significances of the fermionic and bosonic fields ψ and Φ are as follows:

1) Macro-scale: ψ and Φ represent field energy density. In particular, if Φ is a gauge

field then it stands for the interaction potential corresponding to the gauge theory.

2) Micro-scale (Quantization): ψ represents a spin- 21 fermionic particle, and Φ repre-

sents a bosonic particle with an integer spin k if Φ is a k-th order tensor field.

In particular, in the classical Yang-Mills theory, the SU(N) gauge fields A aµ = (Aa0 , Aa1 ,
Aa2 , Aa3 )
(1 ! a ! N 2 − 1) satisfy the following field equations:

∂ µ Fµν
a
= o(Aµ ), a
Fµν = ∂µ Aaν − ∂ν Aaµ + gλbc
a b c
A µ Aν , (4.2.21)

which are the variational equations of the Yang-Mills action

!
1 a µν a
LY M = − Fµν F + LF dx, (4.2.22)
4
where LF is as in (4.2.18) with D µ = ∂µ + igAaµ τa , and

∂ µ Fµν
a
= −#Aaν − ∂ν (∂ µ Aaµ ) + o(A).

Thus, the gauge field equations (4.2.21) are reduced to the bosonic field equations (4.2.20).
In other words, the gauge fields Aaµ (1 ! a ! N 2 − 1) satisfying (4.2.21) represent N 2 − 1
massless bosons with spin-1 because each Aaµ is a vector field.
We are now in position to introduce the Higgs mechanism. Physical experiments show
that the weak interacting fields should be SU(2) gauge fields with masses, representing 3
massive bosonic particles. However, as mentioned in (4.2.21), the gauge fields satisfying
the SU(2) Yang-Mills theory are 3(= N 2 − 1) massless bosons. To overcome this difficulty,
(Higgs, 1964; Englert and Brout, 1964; Guralnik, Hagen and Kibble, 1964) suggested to
add a scalar field φ into the Yang-Mills action (4.2.22) to create masses. In fact, we cannot
add a massive term mAaµ Aµ a into the Yang-Mills action (4.2.22); otherwise this action will
206 Chapter 4 Unified Field Theory of Four Fundamental Interactions

violate the gauge symmetry. But the Higgs mechanism can ensure the gauge invariance for
the Yang-Mills action, and spontaneously break the gauge symmetry in field equations at a
ground state of the Higgs field φ .
For clearly revealing the essence of the Higgs mechanism, we only take one gauge field
(there are four gauge fields in the GWS theory). In this case, the Yang-Mills action density
is in the form
1
LY M = − gµα gνβ (∂µ Aν − ∂ν Aµ )(∂α Aβ − ∂β Aα ) + ψ (iγ µ Dµ − m)ψ , (4.2.23)
4
where g µν is the Minkowski metric, and

Dµ ψ = (∂µ + igAµ )ψ . (4.2.24)

It is clear that the action (4.2.23) is invariant under the following U(1) gauge transformation
1
ψ → eiθ ψ , Aµ → Aµ − ∂µ θ . (4.2.25)
g
The variation equations of (4.2.23) are
#Aµ + ∂µ (∂ ν Aν ) + gJµ = 0,
(iγ µ Dµ − m)ψ = 0, (4.2.26)
µ
Jµ = ψγ ψ ,
which are invariant under the gauge transformation (4.2.25). It is clear that the bosonic
particle A µ in (4.2.26) is massless.
To generate mass for A µ , we add a Higgs sector LH to the Yang-Mills action (4.2.23):
1 1
LH = − gµν (Dµ φ )† (Dµ φ ) + (φ † φ − ρ 2 )2 ,
2 4
Dµ φ = (∂µ + igAµ )φ , (4.2.27)

(Dµ φ )† = (∂µ − igAµ )φ † ,

where ρ ̸= 0 is a constant. Obviously, the following action
!
L= (LY M + LH )dx, (4.2.28)

and its variational equations

δL ig
µ
= ∂ ν (∂ν Aµ − ∂µ Aν ) − gJµ − (φ (Dµ φ )† − φ † Dµ φ ) = 0,
δA 2
δL
= (iγ µ Dµ − m)ψ = 0, (4.2.29)
δψ
δL
= Dµ Dµ φ + (φ 2 − ρ 2 )φ = 0,
δφ
4.2 Physical Supports to PID 207

are invariant under the gauge transformation

1
(ψ , φ ) → (eiθ ψ , eiθ φ ), Aµ → Aµ − ∂µ θ . (4.2.30)
g

Equations (4.2.29) are still massless. However, we note that (0, 0, ρ ) is a solution of
(4.2.29), which represents a ground state in physics, i.e. a vacuum state. Consider a trans-
lation for Φ = (A, ψ , φ ) at Φ0 = (0, 0, ρ ) as

Φ=Φ
' + Φ0 , Φ 'ψ
' = (A, ' , φ'),

then the equations (4.2.29) become

∂ ν (∂ν A'µ − ∂µ A'ν ) − g2ρ 2 A

'µ − gJ'µ + gJ'µ (φ') = 0,
(iγ µ Dµ − m)ψ
' = 0, (4.2.31)
µ
D Dµ (φ' + ρ ) − 2ρ 2φ' + (φ' + ρ )φ'2 = 0,

where
i
J'µ (φ ) = (φ'(Dµ φ')† − φ'† Dµ φ').
2
We see that A 'µ attains its mass m = gρ in (4.2.31), but equations (4.2.31) break the
invariance for the gauge transformation (4.2.30). The process that masses are created by
the spontaneous gauge-symmetry breaking is called the Higgs mechanism. Meanwhile, the
√
field φ' in (4.2.31), called the Higgs boson, is also obtain its mass m = 2ρ .
In the following, we show that PID provides a new mechanism for generating masses,
drastly different from the Higgs mechanism.
In view of (4.1.34) and (4.1.35), based on PID, the variational equations of the Yang-
Mills action (4.2.23) with the divA -free constraint are in the form
0 1
1 D mc E2
∂ ν (∂ν Aµ − ∂µ Aν ) − gJµ = ∂µ − xµ + λ Aµ φ ,
4 h̄ (4.2.32)
µ
(iγ Dµ − m f )ψ = 0,

1 D mc E2
where φ is a scalar field. The term − xµ is the mass potential of φ , and is also
4 h̄
regarded as the interacting length of φ . If φ has a nonzero ground state φ 0 = ρ , then for the
translation
φ = φ' + ρ , Aµ = A 'µ , ψ = ψ ',

the first equation of (4.2.32) becomes

D E2 0 D E2 1
'ν ) − m0 c A
∂ ν (∂ν A'µ − ∂µ A 'µ − gJ'µ = ∂µ − 1 mc xµ + λ A
'µ φ', (4.2.33)
h̄ 4 h̄
208 Chapter 4 Unified Field Theory of Four Fundamental Interactions

D m c E2 h̄ $
0
where = λ ρ . Thus the mass m0 = λ ρ is generated in (4.2.33) as the Yang-
h̄ c
Mills action takes the divA -free constraint variation. Moreover, when we take divergence
on both sides of (4.2.33), and by

∂ µ ∂ ν (∂ν A
'µ − ∂µ A
'ν ) = 0, ∂ µ J'µ = 0,

we derive the field equation of φ' as follows

D mc E2 1 D mc E2
∂ µ ∂µ φ' − φ' = −λ ∂ µ (A'µ φ') + xµ ∂ µ φ'. (4.2.34)
h̄ 4 h̄
This equation (4.2.34) is the field equation with mass m for the Higgs bosonic particle φ'.
Remark 4.7 In (4.2.28) we see that the essence of the Higgs mechanism is to add
artificially a Higgs sector LH into the Yang-Mills action. However, for the PID model, the
masses of A µ and the Higgs field φ are generated naturally for the first principle, PID, taking
the variation with energy-momentum conservation constraint.

4.2.4 Ginzburg-Landau superconductivity

Superconductivity studies the behavior of the Bose-Einstein condensation and electromag-
netic interactions. The Ginzburg-Landau theory provides a support for PID.
The Ginzburg-Landau free energy for superconductivity is
! 0 1
1 !!D es E !!2 2 b 4 1 2
G= ! ih̄∇ + A ψ ! + a|ψ | + |ψ | + | curl A| dx, (4.2.35)
Ω 2ms c 2 8π
where A is the electromagnetic potential, ψ is the wave function of superconducting elec-
trons, Ω is the superconductor, es and ms are charge and mass of a Cooper pair.
The superconducting current equations determined by the Ginzburg-Landau free energy
(4.2.35) are:
δG
= 0, (4.2.36)
δA
which implies that

c e2 h̄es
curl2 A = − s |ψ |2 A − i (ψ ∗ ∇ψ − ψ ∇ψ ∗ ). (4.2.37)
4π ms c ms
Let
c e2s h̄es ∗
J= curl2 A, Js = |ψ |2 A − i (ψ ∇ψ − ψ ∇ψ ∗).
4π ms c ms
Physically, J is the total current in Ω, and Js is the superconducting current. Since Ω is a
medium conductor, J contains two types of currents as

J = Js + σ E,
4.3 Unified Field Model Based on PID and PRI 209

where σ is dielectric constant, σ E is the current generated by the electric field E,

1 ∂A
E =− − ∇Φ = −∇Φ,
c ∂t
and Φ is the electric potential. Since At = 0, the superconducting current equations should
be taken as
1 σ e2 ih̄es ∗
curl2 A = − ∇Φ − s 2 |ψ |2 A − (ψ ∇ψ − ψ ∇ψ ∗ ). (4.2.38)
4π c ms c ms c
Since (4.2.37) is the expression of (4.2.36), then the equation (4.2.38) can be written in
the abstract form
δG σ
= − ∇Φ. (4.2.39)
δA c
In addition, for conductivity, the gauge fixing is given by

div A = 0, A · n|∂ Ω = 0,

which imply that !

∇Φ · Adx = 0.
Ω
σ
Hence, the term − ∇Φ in (4.2.39) can not be added into the Ginzburg-Landau free energy
c
(4.2.35).
However, the equation (4.2.39) are just the variational equation with divergence-free
constraint as follows
H I
δG d
,X = G(A + λ X)|λ =0 = 0, ∀ div X = 0.
δA dλ

Thus, we see that the Ginzburg-Landau superconductivity theory obeys PID.

4.3 Unified Field Model Based on PID and PRI

4.3.1 Unified field equations based on PID
The abstract unified field equations (4.1.33)-(4.1.34) are derived based on PID. We now
present the detailed form of this model, ensuring that these field equations satisfy both the
principle of gauge-symmetry breaking, Principle 4.4, and PRI.
In Section 4.1.3, we showed that the action functional obeys all the symmetric princi-
ples, including principle of general relativity, the Lorentz invariance, the U(1) × SU(2) ×
SU(3) gauge invariance and PRI, is the natural combination of the Einstein-Hilbert func-
tional, the U(1), SU(2), SU(3) Yang-Mills actions for the electromagnetic, weak and strong
interactions: !
√
L= [LEH + LEM + LW + LS ] −gdx. (4.3.1)
M
210 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Here
8π G
LEH = R + S,
c4
1
LEM = − Aµν Aµν + ψ e (iγ µ Dµ − m)ψ e ,
4
(4.3.2)
1 w a bµν
LW = − GabWµν W + ψ w (iγ µ Dµ − ml )ψ w ,
4
1
LS = − Gkls Skµν S µν l + ψ s (iγ µ Dµ − mq )ψ s ,
4
where R is the scalar curvature of the space-time Riemannian manifold (M , g µν ) with
Minkowski type metric, S is the energy-momentum density, G ab w and G s are the SU(2)
kl
and SU(3) metrics as defined by (3.5.28), ψ e , ψ w and ψ s are the Dirac spinors for fermions
participating in the electromagnetic, weak, strong interactions, and
Aµν = ∂µ Aν − ∂ν Aµ ,
a
Wµν = ∂µ Wνa − ∂ν Wµa + gwλbc
a
WµbWνc , (4.3.3)
Skµν = ∂µ Sνk − ∂ν Skµ + gsΛklr Slµ Sνr .
Here Aµ is the electromagnetic potential, Wµa (1 ! a ! 3) are the SU(2) gauge fields for the
weak interaction, Skµ (1 ! k ! 8) are the SU(3) gauge fields for the strong interaction, g w
and gs are the weak and strong charges, and
Dµ ψ e = (∂µ + ieAµ )ψ e ,
Dµ ψ w = (∂µ + igwWµa σa )ψ w , (4.3.4)
s
Dµ ψ = (∂µ + igsSkµ τk )ψ s ,
where σa (1 ! a ! 3) and τk (1 ! k ! 8) are the generators of SU(2) and SU(3).
Remark 4.8 For a vector field Xµ and an antisymmetric tensor field Fµν , we have
∇µ Xν − ∇ν Xµ = ∂µ Xν − ∂ν Xµ ,
∇µ Fµν = ∂ µ Fµν ,
where ∇µ is the Levi-Civita covariant derivative. Hence, the tensor fields in (4.3.3) and the
action (4.3.1) obey both the Einstein general relativity and the Lorentz invariance simulta-
neously under the transformations (4.1.21)-(4.1.22).
Remark 4.9 In the standard model, the wave functions ψ w and ψ s in LW and LS are
as follows % &
l
ψw = 1 the left-hand lepton pairs,
l2 L
⎛ ⎞ (4.3.5)
q1
s
ψ = ⎝q2 ⎠ qi the quark with ith color.
q3
4.3 Unified Field Model Based on PID and PRI 211

The reason why ψ w and ψ s are taken in the form (4.3.5) is that the standard model, in partic-
ular the GWS electroweak theory, is oriented toward to computing the transition probability
for the decay and scattering. In the PID field theory, it is unnecessary to take ψ w and ψ s
as in (4.3.5), because this model is oriented toward to interaction potentials and the basic
mechanism.

Remark 4.10 According to the standard model, the field particles corresponding to
electromagnetic, weak, and strong interactions are described by U(1), SU(2), SU(3) gauge
fields. Hence we take the U(1) × SU(2) × SU(3) Yang-Mills action together with L EH
as the action. However, if only consider the field theory for an N-particle system with N1
electric, N2 weak, N3 strong charges, then the action sectors in (4.3.1) should be taken as

LEH = R,
LEM = SU(N1 ) Yang-Mills action,
LW = SU(N2 ) Yang-Mills action, and
LS = SU(N3 ) Yang-Mills action.

In Section 6.5, we shall discuss the unified field theory for multi-particle systems.

We are now in position to establish unified field equations obeying PRI and PSB. By
PID, the unified field model (4.1.33)-(4.1.34) are derived as the variational equation of the
action (4.1.17) under the divA -constraint

⟨δ L, X⟩ = 0 for any X with divA X = 0.

Here it is required that the gradient operator ∇A corresponding to divA are PRI covariant.
The gradient operators in different sectors are given as follows:

Dgµ = ∇µ + α 0 Aµ + αb1Wµb + αk2 Skµ ,

Deµ = ∇µ + β 0Aµ + βb1Wµb + βk2 Skµ ,
1 (4.3.6)
Dwµ = ∇µ + γ 0 Aµ + γb1Wµb + γk2Skµ − m2w xµ ,
4
1
Dsµ = ∇µ + δ 0 Aµ + δb1Wµb + δk2 Skµ − m2s xµ ,
4

where
mw , ms , α 0 , β 0 , γ 0 , δ 0 are scalar parameters,
αa1 , βa1 , γa1 , δa1 are first-order SU(2) tensors, (4.3.7)
αk2 , βk2 , γk2 , δk2 are first-order SU(3) tensors.
212 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Thus, the PID equations (4.1.33)-(4.1.34) can be expressed as

δL
= Dgµ φνg ,
δ g µν
δL
= Deµ φ e ,
δ Aµ
(4.3.8)
δL
= Dwµ φaw ,
δ Wµa
δL
= Dsµ φks ,
δ Skµ
g
where φν is a vector field, and φ e , φ w , φ s are scalar fields.
With the PID equations (4.3.8), the PRI covariant unified field equations are then given
as follows:1
1 8π G
Rµν − gµν R = − 4 Tµν + Dgµ φνg , (4.3.9)
2 c
∂ µ (∂µ Aν − ∂ν Aµ ) − eJν = Deν φ e , (4.3.10)
F G
w
Gab ∂ µ Wµνb b αβ c
− gw λcd g Wαν Wβd − gwJν a = Dνw φaw , (4.3.11)
F G
j j αβ c
Gksj ∂ µ Sµν − gsΛcd g Sαν Sβd − gsQν k = Dνs φks , (4.3.12)

(iγ µ Dµ − m)ψ e = 0, (4.3.13)

(iγ µ Dµ − ml )ψ w = 0, (4.3.14)
(iγ µ Dµ − mq )ψ s = 0, (4.3.15)

where Dgµ , Deν , Dwν , Dsν are given by (4.3.6), and

Jν = ψ e γν ψ e ,
Jν a = ψ w γ ν σa ψ w ,
Qν k = ψ s γν τk ψ s ,
(4.3.16)
δS c4 αβ w a b
Tµν = + g (GabWα µ Wβ ν + Gkls Sαk µ Sβl ν + Aα µ Aβ ν )
δ g µν 16π G
c4
− gµν (LEM + LW + LS ).
16π G
Remark 4.11 It is clear that the action (4.3.1)-(4.3.4) for the unified field model is
invariant under the U(1) × SU(2) × SU(3) gauge transformation as follows

1 We ignore the Klein-Gordon fields.

4.3 Unified Field Model Based on PID and PRI 213

D E % 1
&
e iθ e
ψ' , Aµ = e ψ , Aµ − ∂µ θ ,
'
e
D E % i
&
a
w 'a w a
ψ , Wµ σa = U ψ ,Wµ U σaU + ∂µ UU
' −1 −1
, U = eiθ σa ,
gw
D E % k i
&
k
(4.3.17)
iϕ τk s k
' s 'k
ψ , S τk = e ψ , Sµ Ωτk Ω + ∂µ ΩΩ
−1 −1
, Ω = eiϕ τk ,
gs
aσ aσ
' l = eiθ
m a
ml e−iθ a
,
iϕ k τ −iϕ k τ
m
'q = e k mq e k .

However, the equations (4.3.9)-(4.3.15) are not invariant under the transformation (4.3.17)
g g
due to the terms D µ φν , Deν φ e , Dwν φaw , Dsν φks on the right-hand sides of (4.3.9)-(4.3.12) con-
taining the gauge fields A µ ,Wµa and Skµ .
Hence, the unified field model based on PID and PRI satisfies the spontaneous gauge-
symmetry breaking as stated in Principle 4.4 and PRI.

4.3.2 Coupling parameters and physical dimensions

There are a number of to-be-determined coupling parameters in the general form of the
unified field equations (4.3.9)-(4.3.15), and the SU(2) and SU(3) generators σ a and τk are
taken arbitrarily. With PRI we are able to substantially reduce the number of these to-be-
determined parameters in the unified model to two SU(2) and SU(3) tensors

{αaw } = (α1w , α2w , α3w ), {αks } = (α1s , · · · , α8s ),

containing 11 parameters, representing the portions distributed to the gauge potentials by

the weak and strong charges.
Also, if we take σa (1 ! a ! 3) as the Pauli matrices (3.5.36) and τk = λk (1 ! k ! 8)
w and G s are Euclidian:
as the Gell-Mann matrices (3.5.38), then the two metrics G ab kl

w
Gab = δab , Gkls = δkl .

Hence we usually take the Pauli matrices σa and the Gell-Mann matrices λk as the SU(2)
and SU(3) generators.
For convenience, we first introduce dimensions of related physical quantities. Let E
represent energy, L be the length and t be the time. Then we have
$ √
(Aµ ,Wµa , Skµ ) : E/L, (e, gw , gs ) : EL,
√
E
(Jµ , Jµ a , Qµ k ) : 1/L3 , e w s
(φ , φa , φk ) : √ ,
LL
(h̄, c) : (Et, L/t), mc/h̄ : 1/L.
214 Chapter 4 Unified Field Theory of Four Fundamental Interactions

In addition, for gravitational fields we have

gµν : dimensionless, R : 1/L2 , Tµν : E/L3 ,

(4.3.18)
φµg : 1/L, G : L5 /Et 4.

The dimensions of the parameters in (4.3.9)–(4.3.15) are as follows

√
(mw , ms ) : 1/L, (α 0 , β 0 , γ 0 , δ 0 ) : 1/ EL,
√ √ (4.3.19)
(αa1 , βa1 , γa1 , δa1 ) : 1/ EL, (αk2 , βk2 , γk2 , δk2 ) : 1/ EL.

Thus the parameters in (4.3.7) can be rewritten as

Dm c m cE
H π
(mw , ms ) = , ,
h̄ h̄
e
(α 0 , β 0 , γ 0 , δ 0 ) = (α e , β e , γ e , δ e ),
h̄c (4.3.20)
gw
(αa , βa , γa , δa ) = (αaw , βaw , γaw , δaw ),
1 1 1 1
h̄c
gs
(αk , βk , γk , δk ) = (αks , βks , γks , δks ),
2 2 2 2
h̄c

where mH and mπ represent the masses of φ w and φ s , and all the parameters (α , β , γ , δ )
on the right hand side of (4.3.20) with different super and sub indices are dimensionless
constants.

4.3.3 Standard form of unified field equations

With (4.3.20) at our disposal, the unified field equations (4.3.9)-(4.3.15) can be simplified
in the form
0 1
1 8π G eα e gw αaw a gs αks k g
Rµν − gµν R = − 4 Tµν + ∇µ + Aµ + Wµ + S φ , (4.3.21)
2 c h̄c h̄c h̄c µ ν
F e gw gs G
∂ ν Aν µ − eJν = ∂µ + β e Aµ + βawWµa + βks Skµ φ e , (4.3.22)
h̄c h̄c h̄c
gw a αβ b c
∂ ν Wνaµ − εbc g Wα µ Wβ − gwJµa (4.3.23)
h̄c
0 1
e gw gs 1 D mH c E 2
= ∂µ + γ e Aµ + γbwWµb + γks Skµ − xµ φwa ,
h̄c h̄c h̄c 4 h̄
ν k gs k αβ i j k
∂ Sν µ − f i j g Sα µ Sβ − gs Q µ (4.3.24)
h̄c
0 1
e e gw w b gs s l 1 D mπ c E 2
= ∂µ + δ Aµ + δb Wµ + δl Sµ − xµ φsk ,
h̄c h̄c h̄c 4 h̄
(iγ µ Dµ − m)Ψ = 0, (4.3.25)
4.3 Unified Field Model Based on PID and PRI 215

where Ψ = (ψ e , ψ w , ψ s ), and

Aµν = ∂µ Aν − ∂ν Aµ ,
a gw a b c
Wµν = ∂µ Wνa − ∂ν Wµa + ε W W , (4.3.26)
h̄c bc µ ν
gs
Skµν = ∂µ Sνk − ∂ν Skµ + fikj Siµ Sνj .
h̄c
Equations (4.3.21)-(4.3.25) need to be supplemented with coupled gauge equations to
compensate the new dual fields (φ e , φwa , φsk ). In different physical situations, the coupled
gauge equations may be different.
From the field theoretical point of view instead of the field particle point of view, the
coefficients in (4.3.21)-(4.3.24) should be

(α1w , α2w , α3w ) = α w (ω1 , ω2 , ω3 ),

(β1w , β2w , β3w ) = β w (ω1 , ω2 , ω3 ),
(4.3.27)
(γ1w , γ2w , γ3w ) = γ w (ω1 , ω2 , ω3 ),
(δ1w , δ2w , δ3w ) = δ w (ω1 , ω2 , ω3 ),

and
(α1s , · · · , α8s ) = α s (ρ1 , · · · , ρ8 ),
(β1s , · · · , β8s ) = β s (ρ1 , · · · , ρ8 ),
(4.3.28)
(γ1s , · · · , γ8s ) = γ s (ρ1 , · · · , ρ8 ),
(δ1s , · · · , δ8s ) = γ s (ρ1 , · · · , ρ8 ),
with the unit modules:
.
|ω | = ω12 + ω22 + ω32 = 1,
.
|ρ | = ρ12 + · · · + ρ82 = 1,

using the Pauli matrices σa and the Gell-Mann matrices λk as the generators for SU(2) and
SU(3) respectively.
The two SU(2) and SU(3) tensors in (4.3.27) and (4.3.28),

ωa = (ω1 , ω2 , ω3 ), ρk = (ρ1 , · · · , ρ8 ), (4.3.29)

are very important, by which we can obtain SU(2) and SU(3) representation invariant gauge
fields:
Wµ = ωaWµa , S µ = ρk Skµ . (4.3.30)

which represent respectively the weak and the strong interaction potentials.
216 Chapter 4 Unified Field Theory of Four Fundamental Interactions

In view of (4.3.27)-(4.3.30), the unified field equations for the four fundamental forces
are written as
0 1
1 8π G eα e gw α w gs α s
Rµν − gµν R + 4 Tµν = ∇µ + Aµ + Wµ + Sµ φνg , (4.3.31)
2 c h̄c h̄c h̄c
0 1
ν eβ e gw β w gs β s
∂ Aν µ − eJµ = ∂µ + Aµ + Wµ + Sµ φ e , (4.3.32)
h̄c h̄c h̄c
gw a αβ b c
∂ ν Wνaµ − ε g Wα µ Wβ − gw Jµa (4.3.33)
h̄c bc
0 1
1 eγ e gw γ w gs γ s
= ∂µ − kw2 xµ + Aµ + Wµ + Sµ φwa ,
4 h̄c h̄c h̄c
ν k gs k αβ i j k
∂ Sν µ − f i j g Sα µ Sβ − gs Q µ (4.3.34)
h̄c
0 1
1 2 eδ e gw δ w gs δ s
= ∂ µ − ks x µ + Aµ + Wµ + Sµ φsk ,
4 h̄c h̄c h̄c
(iγ µ Dµ − m)Ψ = 0. (4.3.35)

4.3.4 Potentials of the weak and strong forces

It is known that the U(1) gauge fields

Aµ = (A0 , A1 , A2 , A3 ) (4.3.36)

represent the electromagnetic potentials, with

A0 = the Coulomb potential,

(4.3.37)
⃗A = magnetic potential, ⃗A = (A1 , A2 , A3 ),

and the electric charge e is

e = the U(1) gauge coupling constant. (4.3.38)

The electromagnetic forces are given by

Fe = −e∇A0 the Coulomb force,

e (4.3.39)
Fm = ⃗v × curl ⃗A the Lorentz force.
c
Now, we consider the SU(2) and SU(3) gauge fields:

SU(2) gauge fields: Wµa = (W0a ,W1a ,W2a ,W3a ), 1 ! a ! 3,

(4.3.40)
SU(3) gauge fields: Skµ = (S0k , S1k , S2k , S3k ), 1 ! k ! 8.

They are SU(N) tensors with N = 2, 3, and have N 2 − 1 components. These components will
change under the transformation of SU(N) generators. Thanks to PRI, the N 2 − 1 (N = 2, 3)
4.3 Unified Field Model Based on PID and PRI 217

gauge fields in (4.3.40) can be combined into two vector fields as in (4.3.30):
Wµ = ωaWµa = (W0 ,W1 ,W2 ,W3 ),
(4.3.41)
S µ = ρk Skµ = (S0 , S1 , S2 , S3 ),
which have the same role as (4.3.36)-(4.3.39) for the electro-magnetic U(1) gauge fields.
In the same spirit as the electromagnetic fields, for the two fields given by (4.3.41), we
have
W0 = the weak force potential,
(4.3.42)
⃗ = the weak magnetic potential, W
W ⃗ = (W1 ,W2 ,W3 ),
and
S0 = the strong force potential,
(4.3.43)
⃗S = the strong magnetic potential, ⃗S = (S1 , S2 , S3 ).
In addition, the weak and strong charges gw and gs are
weak charge gw = SU(2) gauge coupling constant,
(4.3.44)
strong charge gs = SU(3) gauge coupling constant.
The weak and strong forces are given by
weak force: Fw = −gw ∇W0 ,
gw (4.3.45)
⃗,
weak magnetic force: Fwm = ⃗v × curl W
c
and
strong force : Fs = −gs ∇S0 ,
gs (4.3.46)
strong magnetic force : Fsm = ⃗v × curl ⃗S.
c
Remark 4.12 It is the PRI that provides a physical approach to combine the N 2 − 1
components of the SU(N) gauge fields into the forms (4.3.41)-(4.3.46) for the interacting
forces. With this physical interpretation, the electromagnetic, weak and strong interactions
can be regarded as a unified force, separated by three different interaction charges: the
electric charge e, weak charge gw , strong charge gs .

4.3.5 Gauge-fixing problem

For a gauge theory, one has to supplement gauge-fixing equations to ensure a unique phys-
ical solution. In the U(1) gauge theory the gauge-fixing problem is well-posed. However
for the SU(N) gauge theory, the problem is generally not well-posed under the Principle of
Lagrange Dynamics (PLD).
We first recall the classical U(1) gauge theory describing electromagnetism. The field
equations by PLD are given by

∂ ν (∂ν Aµ − ∂µ Aν ) = eψγµ ψ (γµ = g µν γ ν ), (4.3.47)

µ
iγ (∂µ + ieAµ )ψ − mψ = 0, (4.3.48)
218 Chapter 4 Unified Field Theory of Four Fundamental Interactions

which are invariant under the U(1) gauge transformation

' = eiθ ψ ,
ψ 'µ = Aµ − 1 ∂µ θ .
A
e
It implies that A'µ , ψ
' are also solutions. Hence, the equations (4.3.47)-(4.3.48) have in-
finitely many solutions. However, in these solutions only one is physical. To find the phys-
ical solution, one has to provide a supplementary equation, called gauge-fixing equation,

F(Aµ ) = 0, (4.3.49)

such that the system (4.3.47)-(4.3.49) has a unique physical solution. Observe that

∂ µ ∂ ν (∂ν Aµ − ∂µ Aν ) = 0,
∂ µ ψγµ ψ = 0 (by (4.3.48)).

Only three equations in (4.3.47) are independent. Therefore the number of independent
equations in (4.3.47)-(4.3.49) is NEQ = 8, the same as the number of unknowns. Hence,
(4.3.47)-(4.3.49) are well-posed. Usually physical gauge-fixing equation (4.3.47) takes one
of the forms:
∂ A1 ∂ A2 ∂ A3
Coulomb gauge : + + 3 = 0,
∂ x1 ∂ x2 ∂x
Lorentz gauge : ∂ µ Aµ = 0,
Axial gauge : A3 = 0,
Temporal gauge : A0 = 0.

However, for the SU(N) gauge theory by PLD, the gauge-fixing problem is in general
not well-posed. In fact, the SU(N) gauge field equations are given by

∂ ν Fνaµ = gψγµ τa ψ , for 1 ! a ! N 2 − 1, (4.3.50)

µ
iγ (∂µ + igGaµ τa )ψ − mψ = 0, (4.3.51)

where
Fνaµ = ∂ν Gaµ − ∂µ Gaν + gλbc
a b c
G µ Gν .
The equations (4.3.50)-(4.3.51) are invariant under the SU(N) gauge transformations

ψ
a
' = eiθ τa ψ , 'aµ τa = Gaµ eiθ b τb τa e−iθ b τb − 1 ∂µ θ a τa .
G (4.3.52)
g
Hence, if there is a solution for (4.3.50)-(4.3.51), then there are infinitely many solutions.
Thus, we have to supplement N 2 − 1 gauge-fixing equations in order to get a unique physical
solution:
Fa (Gµ ) = 0 for 1 ! a ! N 2 − 1. (4.3.53)
4.4 Duality and Decoupling of Interaction Fields 219

The reason why take N 2 − 1 equations in (4.3.53) are that there are N 2 − 1 free functions θ a
in (4.3.52).
Now, the gauge-fixing problem (4.3.50)-(4.3.51) with (4.3.53) is not well-posed either,
because the number of independent equations of (4.3.50) are 4(N 2 − 1) due to

∂ µ (ψγµ τa ψ ) ̸= 0.

Namely, the number of independent equations in the gauge-fixing problem (4.3.50)-(4.3.51)

with (4.3.53) is NEQ = 5(N 2 − 1) + 4N, larger than the number of unknowns NUF = 4(N 2 −
1) + 4N.
The non well-posedness of SU(N) gauge-fixing problem implies the the PLD is not
applicable for the SU(N) gauge field theory. However, based on PID, the SU(N) gauge-
fixing problem is well-posed.

4.4 Duality and Decoupling of Interaction Fields

The natural duality of four fundamental interactions to be addressed in this section is a
direct consequence of PID. It is with this duality, together with the PRI invariant potentials
Sµ and Wµ given by (4.5.1) and (4.6.1), that we establish a clear explanation for many
longstanding challenging problems in physics, including for example the dark matter and
dark energy phenomena, the formulas of the weak and strong forces, the quark confinement,
the asymptotic freedom, and the strong interacting potentials of nucleons. Also, this duality
lay a solid foundation for the weakton model of elementary particles and the energy level
theory of subatomic particles, and gives rise to a new mechanism for sub-atomic decay and
scattering.
The unified field model can be easily decoupled to study each individual interaction
when other interactions are negligible. In other words, PID is certainly applicable to each
individual interaction. For gravity, for example, PID offers to a new gravitational field
model, leading to a unified model for dark energy and dark matter (Ma and Wang, 2014e).

4.4.1 Duality
In the unified field equations (4.3.21)-(4.3.24), there exists a natural duality between the
interaction fields (g µν , Aµ ,Wµa , Skµ ) and their corresponding dual fields (φ µg , φ e , φaw , φks ) :

g µν ↔ φµg ,
Aµ ↔ φ e,
(4.4.1)
Wµa ↔ φwa for 1 ! a ! 3,
Skµ ↔ φsk for 1 ! k ! 8.

Thanks to PRI, the SU(2) gauge fields Wµa (1 ! a ! 3) and the SU(3) gauge fields S kµ (1 !
k ! 8) are symmetric in their indices a = 1, 2, 3 and k = 1, · · · , 8 respectively. Therefore, the
220 Chapter 4 Unified Field Theory of Four Fundamental Interactions

corresponding relation (4.4.1) can be also considered as the following dual relation
gµν ↔ φµg ,
Aµ ↔ φ e,
(4.4.2)
{Wµa } ↔ {φwa },
{Skµ } ↔ {φsk }.
The duality relation (4.4.1) can be regarded as the correspondence between field particles
for each interaction, and the relation (4.4.2) is the duality of interacting forces. We now
address these two different dualities.
1. Duality of field particles. In the duality relation (4.4.1), if the tensor fields on the
left-hand side are of k-th order, then their dual tensor fields on the right-hand side are of
(k − 1)-th order. Physically, this amounts to saying that if a mediator for an interaction has
spin-k, then the dual mediator for the dual field has spin-(k − 1). Hence, (4.4.1) leads to the
following important physical conclusion:

Duality of Interaction Mediators 4.13 Each interaction mediator possesses a dual

field particle, called the dual mediator, and if the mediator has spin-k, then its dual mediator
has spin-(k − 1).
The duality between interaction mediators is a direct consequence of PID used for de-
riving the unified field equations. Based on this duality, if there exist a graviton with spin
J = 2, then there must exist a dual graviton with spin J = 1. In fact, for all interaction
mediators, we have the following duality correspondence:
graviton (J = 2) ↔ dual vector graviton (J = 1),
photon (J = 1) ↔ dual scalar photon (J = 0),
W ± bosons (J = 1) ↔ charged Higgs H ± (J = 0), (4.4.3)
0
Z boson (J = 1) ↔ neutral Higgs H (J = 0),
k
gluons g (J = 1) ↔ dual scalar gluons φgk (J = 0).

The neutral Higgs H 0 (the adjoint particle of Z) had been discovered experimentally. We
remark that the duality (4.4.3) can also be derived using the weakton model (Ma and Wang,
2015b), which is also presented in the next chapter.
2. Duality of interacting forces. The correspondence (4.4.2) provides a dual relation
between the attracting and repelling forces. In fact, from the interaction potentials we find
that the even-spin fields yield attracting forces, and the odd-spin fields yield repelling forces.
Duality of Interaction Forces 4.14 Each interaction generates both attracting and
repelling forces. Moreover, for each pair of dual fields, the even-spin field generates an
attracting force, and the odd-spin field generates a repelling force.
4.4 Duality and Decoupling of Interaction Fields 221

This duality of interaction forces is illustrated as follows:

Gravitation force = attraction due to g µν + repelling due to φµg ,
Electromagnetism = attraction due to φ e + repelling due to A µ ,
(4.4.4)
Weak force = attraction due to φw + repelling due to Wµ ,
Strong force = attraction due to φs + repelling due to S µ .
Here, we remark that the electromagnetic force in (4.4.4) is between the charged particles
with the same sign, and the force generated by φ e is attractive.

4.4.2 Gravitational field equations derived by PID

If we only consider the gravitational interaction, then the gravitational field equations can
be decoupled from the unified field model (4.3.21)-(4.3.25), and are given by
1 8π G F e G
Rµν − gµν R = − 4 Tµν + ∇µ + Aµ Φν , (4.4.5)
2 c h̄c
e
where the term Aµ Φν represents the coupling between the gravitation and the cosmic
h̄c
microwave background (CMB) radiation.
By the Bianchi identity (4.2.2), taking divergence on both sides of (4.4.5) yields
e µ 8π G
∇µ ∇µ Φν + ∇ (Aµ Φν ) = 4 ∇µ Tµν . (4.4.6)
h̄c c
The duality of gravity is based on the field equations (4.4.5) and (4.4.6).
1. Gravitons and dual gravitons. It is known that as the equations describing field parti-
cles, (4.4.5) characterize the graviton as a massless, neutral bosonic particle with spin J = 2,
and (4.4.6) indicate that the dual vector graviton is a massless, neutral bosonic particle with
J = 1. Hence, the gravitational field equations induced by PID and PRI provide a pair of
field particles:
tensor graviton: J = 2, m = 0, Qe = 0,
(4.4.7)
vector graviton: J = 1, m = 0, Qe = 0,
where Qe is the electric charge.
It is the nonlinear interaction of these two field particles in (4.4.7) that lead to the dark
matter and dark energy phenomena.
2. Gravitational force. If we consider the gravitational force only from the Einstein field
equations
1 8π G
Rµν − gµν R = − 4 Tµν , (4.4.8)
2 c
then by the Schwarzschild solution of (4.4.8), we can derive the classical Newton’s gravita-
tional force as
mMG
F =− 2 , (4.4.9)
r
222 Chapter 4 Unified Field Theory of Four Fundamental Interactions

which is an attracting force generated by g µν .

However, with the field equations (4.4.5), we can deduce a revised formula to (4.4.9).
Actually, ignoring the microwave background radiation, the equations (4.4.5) become (Ma
and Wang, 2014e):
1 8π G
Rµν − gµν R = − 4 Tµν − ∇µ ∇ν ϕ , (4.4.10)
2 c
where Φν = −∇ν ϕ , and ϕ is a scalar field. In Chapter 7 (see also (Ma and Wang, 2014e)),
we are able to derive from (4.4.10) that the gravitational force should be in the form
0 % 2 & 1
1 c2 c 1 dϕ
F = mMG − 2 + Φr − + , (4.4.11)
r 2MG MG r dr
where ϕ is the dual scalar field, representing the scalar potential, and

Φ = g µν ∇µ ∇ν ϕ . (4.4.12)

The first term in the right-hand side of (4.4.11) is the Newton’s gravitational force, and the
second term (4.4.12) represents the repelling force generated by the dual field ϕ , and the
third term % 2 &
c 1 dϕ
− +
MG r dr
represents the force due to the nonlinear coupling of g µν and its dual ϕ . Formula (4.4.11)
can be approximatively written as
% &
1 k0
F = mMG − 2 − + k1 r ,
r r (4.4.13)
−18 −1 −57 −3
k0 = 4 × 10 km , k1 = 10 km .
The formula (4.4.13) shows that a central gravitational field with mass M has an attract-
ing force −k0 /r in addition to the Newtonian gravitational force. This explains the dark
matter phenomenon. Also there is a repelling force k1 r, which explains the dark energy
phenomenon; see (Ma and Wang, 2014e) for details.

4.4.3 Modified QED model

For the electromagnetic interaction only, the decoupled QED field equations from (4.3.22)
and (4.3.25) are given by
% &
ν βe
∂ (∂ν Aµ − ∂µ Aν ) − eJµ = ∂µ + Aµ φ e , (4.4.14)
h̄c
D e E mc
iγ µ ∂µ + i Aµ ψ − ψ = 0, (4.4.15)
h̄c h̄
where β is a dimensionless constant, and Jµ = ψγ µ ψ is the current density satisfying

∂ µ Jµ = 0. (4.4.16)
4.4 Duality and Decoupling of Interaction Fields 223

Equations (4.4.14) and (4.4.15) are the modified QED model. Taking divergence on
both sides of (4.4.14), by (4.4.16) and

∂ µ ∂ ν (∂ν Aµ − ∂µ Aν ) = 0,

the equations (4.4.14)-(4.4.15) can be equivalently written as

% &
ν βe
∂ (∂ν Aµ − ∂µ Aν ) − eJµ = ∂µ + Aµ φ e , (4.4.17)
h̄c
βe
∂ µ ∂µ φ e + ∂ µ (Aµ φ e ) = 0, (4.4.18)
h̄c
µ
J K mc
iγ ∂µ + ieAµ ψ − ψ = 0. (4.4.19)
h̄
If we take
1 ∂ ⃗A
H = curl ⃗A, E = − − ∇ϕ , (4.4.20)
c ∂t
where A µ = (ϕ , ⃗A), ⃗A = (A1 , A2 , A3 ), then (4.4.17)-(4.4.18) and (4.4.20) are a modified
version of the classical Maxwell equations, which are expressed as
1 ∂H
= − curl E,
c ∂t
H = curl ⃗A,
1 ∂E βe
= curl H + J⃗ + ∇φ e + Aµ φ e , (4.4.21)
c ∂t h̄c
1 ∂φe βe e
div E = ρ + + φ ϕ,
c ∂t h̄c
% 2 & % &
∂ e β e 1∂ e e
−∆ φ + (ϕφ ) − div (Aφ ) = 0,
⃗
∂ t2 h̄c c ∂ t
where J⃗ is the electric current density and ρ is the electric charge density.
The equations (4.4.14)-(4.4.15) or (4.4.17)-(4.4.19) need to be supplemented with a
coupled equation to compensate the gauge-symmetry breaking and the induced dual field
φ e:
F(Aµ , φ e , ψ ) = 0. (4.4.22)
Remark 4.15 Usually, the compensating equation (4.4.22) is called the gauge-fixing
equation. The compensating equation (4.4.22) should be determined based on first princi-
ples. However, we don’t know whether there are such physical laws. In general, according
to different situations physicists take the gauge-fixing equation in the following forms:
Lorentz gauge: ∂ µ Aµ = 0,
Coulomb gauge: div ⃗A = 0,
(4.4.23)
Axial gauge: A3 = 0,
Temporal gauge: A0 = 0 (A0 = ϕ ).
224 Chapter 4 Unified Field Theory of Four Fundamental Interactions

The following are the two perspectives of the duality for electromagnetism.
1. Photon and dual scalar photon. If we view the field equations (4.4.17) and (4.4.18)
as describing the field particles, then we have

Jµ = 0, β = 0, in (4.4.17)-(4.4.18).

Thus, the usual photon equation is given by

#Aµ + ∂µ (∂ ν Aν ) = 0, (4.4.24)

and the dual photon, also called the scalar photon, is described by the following equation

#φ e = 0. (4.4.25)

By (4.4.24) and (4.4.25) we deduce the following basic properties for photons and scalar
photons:
photon: J = 1, m = 0, Qe = 0,
(4.4.26)
scalar photon: J = 0, m = 0, Qe = 0.

2. Electromagnetic force. If we consider the electromagnetic force, then the constant

β ̸= 0 in (4.4.17). It is known that in the classical theory, the Coulomb potential ϕ satisfies
the equation
∆ϕ = −4πρ , (4.4.27)
which is the stationary equation of (4.4.17) with µ = 0 and Φe = 0, and with the Coulomb
gauge in (4.4.23). For the case where ρ = eδ (r), the solution of (4.4.27) is the well-known
Coulomb potential:
e
ϕ= . (4.4.28)
r
For the modified equations (4.4.17) and (4.4.18), the stationary time-component equa-
tions with the Coulomb gauge are given by
βe e
∆ϕ − φ ϕ = 4π eδ (r), (4.4.29)
h̄c
∆φ e = 0. (4.4.30)

Only the radially symmetric solutions of (4.4.29) and (4.4.30) are physical. The radial
solutions of (4.4.30) are constants

φ e = φ0 the constants.

Thus the equations (4.4.29) and (4.4.30), in the spherical coordinate system, are reduced as
follows
1 d 2d
(r )ϕ − kϕ = −4π eδ (r), (4.4.31)
r2 dr dr
4.4 Duality and Decoupling of Interaction Fields 225

e
where k = β φ0 . The solutions of (4.4.31) are expressed as
h̄c
e √
ϕ = e− kr . (4.4.32)
r
The parameter k is to be determined by experiments.
We discuss the solutions (4.4.32) in the following three cases.

1) Case k = 0. The solution (4.4.32) in this case is reduced to the Coulomb potential
(4.4.28);

2) Case k > 0. This solution is similar to the Yukawa potential for the strong interac-
tions of nucleons.

3) Case k < 0. In this case, the solution can be written as

e √ αe √
ϕ = cos −kr + sin −kr, (4.4.33)
r r
where α is an arbitrary constant. The function
1 √
ϕ0 = sin −kr
r
in (4.4.33) satisfies that
∆ϕ0 − kϕ0 = 0.

4.4.4 Strong interaction field equations

The decoupled strong interaction field model from (4.3.21)-(4.3.25) is given by
0 1
ν k gs k αβ i j k gs s l 1 D mπ c E 2
∂ Sν µ − fi j g Sα µ Sβ − gsQµ = ∂µ + δl Sµ − xµ φsk , (4.4.34)
h̄c h̄c 4 h̄
F gs G mc
iγ µ ∂µ + i Sbµ τb ψ − ψ = 0, (4.4.35)
h̄c h̄
for 1 ! k ! 8, where τk = τ k are the Gell-Mann matrices as in (3.5.38), and
gs
Skµν = ∂µ Sνk − ∂ν Skµ + Siµ Sνj ,
h̄c (4.4.36)
Qkµ = ψγ µ τ k ψ .
Taking divergence on both sides of (4.4.34) and by

∂ µ ∂ ν Skµν = 0 for 1 ! k ! 8,

we deduce the following dual field equations for the strong interaction:
0% & 1
gs k l 1 m2π c2 gs
µ k
∂ ∂µ φs + ∂ µ
δ S − xµ φsk = −gs ∂ µ Qkµ − fikj gαβ ∂ µ (Sαi µ Sβj ).
h̄c l µ 4 h̄2 h̄c
(4.4.37)
226 Chapter 4 Unified Field Theory of Four Fundamental Interactions

The equations (4.4.34)-(4.4.35) also need 8 additional gauge equations to compensate

the induced dual fields φsk :

Fsk (S µ , φs , ψ ) = 0, 1 ! k ! 8. (4.4.38)

We have the following duality for the strong interaction.

1. Gluons and dual scalar gluons. Based on quantum chromodynamics (QCD), the field
particles for the strong interaction are the eight massless gluons with spin J = 1, which are
described by the SU(3) gauge fields S kµ (1 ! k ! 8). By the duality (4.4.1), for the strong
interactions we have the field particle correspondence

Skµ ↔ φsk for 1 ! k ! 8.

It implies that corresponding to the 8 gluons S kµ (1 ! k ! 8) there should be 8 dual gluons

represented by φsk , called the scalar gluons due to φsk being scalar fields. Namely we have
the following gluon correspondence

gluons gk ↔ scalar gluons gk0 for 1 ! k ! 8.

Gluons and scalar gluons are described by equations (4.4.34) and (4.4.37) respectively,
which are nonlinear. In fact, gk and gk0 are confined in hadrons.
2. Strong force. The strong interaction forces are governed by the field equations
(4.3.31)-(4.3.35). The decoupled field equations are given by
0 1
ν k gs k αβ i j k 1 2 gs δ
∂ S ν µ − f i j g S α µ S β − g s Q µ = ∂ µ − ks x µ + Sµ φsk , (4.4.39)
h̄c 4 h̄c
1 gs δ µ
∂ µ ∂µ φsk − k2 φsk + ks2 xµ ∂ µ φsk + ∂ (S µ φsk ) (4.4.40)
4 h̄c
g s
= −gs ∂ µ Qkµ − fikj gαβ ∂ µ (Sαi µ Sβj ),
F G h̄c
gs mc
iγ µ ∂µ + i Slµ τl ψ − ψ = 0, (4.4.41)
h̄c h̄
for 1 ! k ! 8, where δ is a parameter, and S µ is as in (4.3.41).
In the next section, we shall deduce the layeredformulas of strong interaction potentials
from the equations (4.4.39)-(4.4.41) with gauge equations (4.4.38).
Remark 4.16 We need to explain the physical significance of the parameters k s and
δ . Usually, ks and δ are regarded as masses of the field particles. However, since (4.4.39)-
(4.4.41) are the field equations for the interaction forces, the parameters k s and δ are no
longer viewed as masses. In fact, k −1 represents the range of attracting force for the strong
% &−1
gs φs0
interaction, and δ is the range of the repelling force, where φs0 is a ground state
h̄c
of φs .
4.4 Duality and Decoupling of Interaction Fields 227

4.4.5 Weak interaction field equations

The unified field model (4.3.21)-(4.3.25) can be decoupled to study the weak interaction
only, leading to the following weak interaction field equations:
0 1
ν a gw a αβ b c a 1 D mH c E 2 gw w b a
∂ Wν µ − εbc g Wα µ Wβ − gwJµ = ∂µ − xµ + γb Wµ φw , (4.4.42)
h̄c 4 h̄ h̄c
F g G mc
w
iγ µ ∂µ + i Wµa σa ψ − ψ = 0, (4.4.43)
h̄c h̄
where mH represents the mass of the Higgs particle, σa = σ a (1 ! a ! 3) are the Pauli
matrices as in (3.5.36), and
a gw a b c
Wµν = ∂µ Wνa − ∂ν Wµa + ε W W ,
h̄c bc µ ν (4.4.44)
Jµa = ψγµ σ a ψ , γµ = g µν γ ν .
Taking divergence on both sides of (4.4.42) we get
D m c E2 gw 1 D mH c E 2
H
∂ µ ∂µ φwa − φwa + γbw ∂ µ (Wµb φwa ) − xµ ∂ µ φwa (4.4.45)
h̄ h̄c 4 h̄
gw a αβ µ b c
= − εbc g ∂ (Wα µ Wβ ) − gw ∂ µ Jµa .
h̄c
Also, we need to supplement (4.4.42)-(4.4.43) with three additional 3 gauge equations
to compensate the induced dual fields φwa :

Fwa (Wµ , φw , ψ ) = 0 for 1 ! a ! 3. (4.4.46)

1. Duality between W ± , Z Bosons and Higgs Bosons H ± , H 0 . The three massive vec-
tor bosons, denoted by W ± , Z 0 , has been discovered experimentally. The field equations
(4.4.42) give rise to a natural duality:

Z0 ↔ H 0, W ± ↔ H ±, (4.4.47)

where H 0 , H ± are three dual scalar bosons, called the Higgs particles. The neutral Higgs H 0 ,
discovered by LHC in 2012, and the charged Higgs H ± , to be discovered experimentally.
In Section 4.6, we shall introduce the dual bosons (4.4.47) by applying the field equa-
tions (4.4.42)-(4.4.45) with gauge equations (4.4.46).
2. Weak force. If we consider the weak interaction force, we have to use the equations
decoupled from (4.3.31)-(4.3.35):
0 1
ν a gw a αβ b c 1 2 gw
∂ Wν µ − εbc g Wα µ Wβ − gw Jµ = ∂µ − kw xµ + γ Wµ φwa ,
a
(4.4.48)
h̄c 4 h̄c
gw 1
∂ µ ∂µ φwa − k2 φw2 + γ∂ µ (Wµ φwa ) − k2 xµ ∂ µ φwa (4.4.49)
h̄c 4
228 Chapter 4 Unified Field Theory of Four Fundamental Interactions

gw a αβ µ b c
= −gw ∂ µ Jµa − ε g ∂ (Wα µ Wβ ),
h̄c bc
gw mc
iγ µ (∂µ + i Wµa σa )ψ − ψ = 0, (4.4.50)
h̄c h̄
where γ , kw are constants, Wµ is as in (4.3.41).
In Section 4.6, we shall deduce the layered formulas of weak interaction potentials by
applying the equations (4.4.48)-(4.4.50).
Remark 4.17 The duality of four fundamental interactions is very important, and is a
direct consequence of PID. With this duality, and with the PRI invariant potentials Wµ and
Sµ given by (4.3.41), we obtain explanations for a number of physical problems such as the
dark matter and dark energy phenomena, the weak and strong forces and potential formulas,
the quark confinement, the asymptotic freedom, the strong potentials of nucleons etc. Also
this study leads to the needed foundation for the weakton model of elementary particles and
the energy level theory of subatomic particles, and gives rise to a mechanism for subatomic
decays and scatterings.

4.5 Strong Interaction Potentials

4.5.1 Strong interaction potential of elementary particles
Thanks to PRI, the strong interaction potential takes the linear combination of the eight
SU(3) gauge potentials as follows
S µ = ρk Skµ , (4.5.1)
where ρk = (ρ1 , · · · , ρ8 ) is the SU(3) tensor as given in (4.3.29).
Let gs be the strong charge of an elementary particle, i.e. the w∗ weakton introduced in
Chapter 5, and
Φ0 = S0 the time-component of (4.5.1)
be the strong charge potential of this particle. Then the strong force between two elementary
particles carrying strong charges is

F = −gs ∇Φ0 .

However, the strong interactions are layered, i.e. the strong forces only act on the same
level of particles, such as the quark level, the hadron level. Hence, the strong interaction
potentials are also layered. In fact, in the next subsection we shall show that for a particle
with N strong charges gs of the elementary particles, its strong interaction potential is given
by 0 1
1 A
Φs = Ngs (ρ ) − (1 + kr)e −kr
,
r ρ
% &3 (4.5.2)
ρw
gs (ρ ) = gs ,
ρ
4.5 Strong Interaction Potentials 229

where ρw is the radius of the elementary particle (i.e. the w∗ weakton), ρ is the particle
radius, k > 0 is a constant with k−1 being the strong interaction attraction radius of this
particle, and A is the strong interaction constant, which depends on the type of particles.
Thus, the strong force between such two particles is

F = −Ngs (ρ )∇Φs , (4.5.3)

where gs (ρ ) and Φs are as in (4.5.2).

In particular for the w∗ -weakton, which possesses one strong charge g s , the formula
(4.5.2) becomes 0 1
1 A0 −k0 r
Φ0 = gs − (1 + k0r)e , (4.5.4)
r ρw
where 1/k0 is the attraction radius of the strong interaction for the elementary particles, i.e.
the w∗ -weakton. According to the physical observation, we take the quantitative order

k0 = 1018 cm−1 . (4.5.5)

In this subsection, we shall deduce formula (4.5.4) for the w∗ -weaktons from the strong
interaction field equations (4.4.39)-(4.4.41). Taking inner product of the field equation
(4.4.39) and (4.4.40) with ρk = (ρ1 , · · · , ρ8 ), we derive that
0 1
ν gs αβ i j 1 2 gs δ
∂ S ν µ − λ i j g S α µ S β − g s Q µ = ∂ µ − k0 x µ + Sµ φs , (4.5.6)
h̄c 4 h̄c
% &
1 ∂2 1 gs
− ∆ φs + k02 φs + k02 xµ ∂ µ φs = gs ∂ µ Qµ + ∂ µ (λi j gαβ Sαi µ Sβj − δ S µ φs ),
c ∂t
2 2 4 h̄c
(4.5.7)

where
gs
φs = ρk φsk , Qµ = ρk Qkµ , λi j = ρk λikj , S µν = ∂µ Sν − ∂ν Sµ + λi j Siµ Sνj .
h̄c
Based on the superposition property of the strong potential for strong charges, Φ s = S0
and φs obey a linear relationship. Namely, the time-component µ = 0 equation of (4.5.6)
and the equation (4.5.7) should be linear. In other words, we have to take eight gauge fixing
equations as in (4.4.38) such that they contain the following two equations:
F G
λi j ∂ ν (Sνi S0j ) − gαβ Sαi o Sβj + δ S0 φs = 0,
F G (4.5.8)
∂ µ λi j gαβ Sαβ i S j − δ S φ = 0,
β µ s

Also, in the eight supplement equations we take

xµ ∂ µ φs = 0, ∂ µ Sµ = 0, (4.5.9)
230 Chapter 4 Unified Field Theory of Four Fundamental Interactions

together with the following static assumption:

∂ S0 ∂ φs
= 0, = 0. (4.5.10)
∂t ∂t
With the gauge fixing equations (4.5.8)-(4.5.9) and the static assumption (4.5.10), we
derive from (4.5.6) and (4.5.7) that
1
−∆Φs = gs Q − k02 cτφs , (4.5.11)
4
−∆φs + k0 φs = gs ∂ µ Qµ ,
2
(4.5.12)

where cτ is the wave length of φs , Q = −Q0 .

In the following, we deduce the solution Φs and φs of (4.5.11)-(4.5.12) in a few steps.
Step 1. Solution of (4.5.12). By definition of Q µ , we have

∂ µ Qµ = ρk ∂µ ψγ µ τk ψ + ρk ψγ µ τk ∂µ ψ .

In view of the Dirac equation (4.4.41),

gs j mc
∂µ ψγ µ τk ψ = i Sµ ψγ µ τ j τk ψ + i ψτk ψ ,
h̄c h̄
µ gs j µ mc
ψγ τk ∂µ ψ = −i Sµ ψγ τk τ j ψ − i ψτk ψ .
h̄c h̄
Hence we arrive at
igs 2gs µ
∂ µ Qµ = ρk Sµj ψγ µ [τ j , τk ] ψ = − ρk Sµj λ jk
i
Qi (4.5.13)
h̄c h̄c
µ
Since Qi = Qiµ is the current density, we have
i µ µ
ρk λ jk Qi = θ j δ (r),
µ
where δ (r) is the Dirac delta function, θ j is a constant tensor, inversely proportional to the
volume of the particle. Hence
µ j µ
ρk Sµj λ jk
i
θi = S µ θ j δ (r),
j j
where S µ ∼ S µ (0) takes the following average value
!
j 1 j
Sµ = Sµ dv. (4.5.14)
|Bρw | Bρw

Here ρw is the radius of a w∗ -weakton. Later, we shall see that

1
Sµj ∼ as r → 0.
r
4.5 Strong Interaction Potentials 231

Hence we deduce from (4.5.14) that

j j j
Sµ = ξµ ρw−1 (ξµ is a constant tensor).

Thus, (4.5.13) becomes

κ
∂ µ Qµ = − δ (r) (ρw is the radius of a w∗ -weakton), (4.5.15)
ρw

where κ is a parameter given by

2gs j µ
κ= ξµ θ j , (4.5.16)
h̄c
and κ is inversely proportional to the volume of w∗ -weakton.
Therefore, equation (4.5.12) is rewritten as
gs κ
− ∆φs + k02 φs = − δ (r), (4.5.17)
ρw

whose solution is given by

gs κ 1 −k0 r
φs = − e . (4.5.18)
ρw r

Step 2. Solution of (4.5.11). The quantity gs Q = −gs Q0 is the strong charge density of
a w∗ -weakton, and without loss of generality, we assume that

Q = β δ (r), (4.5.19)

and β > 0 is a constant, inversely proportional to the volume of the w∗ -weakton. Hence
(4.5.11) can be rewritten as

gs A 1 −k0 r
− ∆Φs = gs β δ (r) + e , (4.5.20)
ρw r

where A is a constant given by

k02 cτκ 1
A= with physical dimension .
4 L
Assume Φs = Φs (r) is radially symmetric, then (4.5.20) becomes
% &
1 d 2 d gs A 1 −k0 r
− 2 r Φs = gs β δ (r) + e ,
r dr dr ρw r

whose solution takes the form

0 1
β A
Φs = gs − ϕ (r)e−k0 r , (4.5.21)
r ρw
232 Chapter 4 Unified Field Theory of Four Fundamental Interactions

where ϕ solves % & % &

1 2k0 1
ϕ ′′ + 2 − k0 ϕ ′ − − k02 ϕ = . (4.5.22)
r r r
Assume that the solution ϕ of (4.5.22) is given by
∞
ϕ= ∑ αk rk . (4.5.23)
k=0

Inserting ϕ in (4.5.22) and comparing the coefficients of r k , then we obtain the following
relations
1
α1 = k0 α0 + ,
2
1 2 1
α2 = k0 α0 + k0 ,
2 3
..
. (4.5.24)
2k0
αk = αk−1 − k02 αk−2 , ∀k " 2.
k+1
where α0 is a free parameter with dimension L. Hence
r
ϕ (r) = α0 (1 + k0r + + o(r)),
2α0
and often it is sufficient to take a first-order approximation.
Step 3. Strong interaction potential of w∗ -weakton. The formula (4.5.21) with (4.5.23)-
(4.5.24) provides an accurate strong interaction potential for the w∗ -weaktons:
0 1
1 A0 r
Φ0 = gs β − ϕ'(r)e −k 0 , (4.5.25)
r ρw
where A0 = Aα0 /β is a dimensionless parameter, and ϕ'(r) is
r
ϕ' = 1 + k0r + + o(r),
2α0
which is a dimensionless function, and

β is inversely proportional to the particle volume. (4.5.26)

Since gs β is to be determined for the elementary particle, we can take g s β as the strong
charge of the w∗ -weakton, still denoted by gs . In addition, it is sufficient to take approxima-
tively ϕ' = 1 + k0r. Then the formula (4.5.25) is rewritten as
0 1
1 A0 −k0 r
Φ0 = gs − (1 + k0r)e .
r ρw
It is the strong interaction potential given in (4.5.4).
4.5 Strong Interaction Potentials 233

4.5.2 Layered formulas of strong interaction potentials

Different from gravity and electromagnetic force, strong interaction is of short-ranged with
different strengths in different levels. For example, in the quark level, strong interaction
confines quarks inside hadrons, in the nucleon level, strong interaction bounds nucleons
inside atoms, and in the atom and molecule level, strong interaction almost diminishes.
This layered phenomena can be well-explained using the unified field theory based on PID
and PRI. We derive in this subsection strong interaction potentials in different levels.
Without loss of generation, we shall discuss strong interaction nucleon potential. For
strong interaction of nucleons, we still use the SU(3) gauge action
1 D mc E
L = − Skµν S µν k + h̄cq iγ µ Dµ − q, (4.5.27)
4 h̄
where Skµ are the strong interaction gauge fields,

q = (q1 , q2 , q3 ) (4.5.28)

where q1 , q2 , q3 are the wave functions of the three quarks constituting a nucleon, and
% &
igs
Dµ q = h̄c∂µ + Skµ τk q. (4.5.29)
h̄c
The action L defined by (4.5.27) is SU(3) gauge invariant. Physically this means that the
contribution of each quark to the strong interaction potential energy is indistinguishable.
With PID, the corresponding field equations for the action (4.5.27) are
% &
gs k2
∂ ν Sνk µ + λikj gαβ Sαi µ Sβj − gs Qkµ = ∂µ − 1 xµ φnk , (4.5.30)
h̄c 4
D gs E mc
iγ µ ∂µ + i Skµ τk q − q = 0, (4.5.31)
h̄c h̄
where the parameter k1 is defined by
1
r1 = = 10−13 cm, (4.5.32)
k1
which is the strong attraction radius of nucleons, and

Qkµ = qγµ τ k q (τ k = τk ).

Let the strong interaction nucleon potential Φkn and its dual potential φn be defined by

Φn = ρk S0k , φn = ρk φnk .

In the same spirit as for deriving the weakton potential equations (4.5.11) and (4.5.12), for
Φn and φn we deduce that
1
− ∆Φn = gs Qn − k12 cτ1 φn ,
4 (4.5.33)
− ∆φn + k12 φn = gs ∂ µ Qµ ,
234 Chapter 4 Unified Field Theory of Four Fundamental Interactions

where cτ1 is the wave length of φn , and

Qµ = ρk Qkµ = (Q0 , Q1 , Q2 , Q3 )

represents the quark current density inside a nucleon. Similar to (4.5.15) and (4.5.19), for
Qn and ∂ µ Qµ we have
κn
∂ µ Qµ = − δ (r), Qn = βn δ (r), (4.5.34)
ρn
where ρn is the radius of a nucleon, βn and κn are constants, inversely proportional to the
volume of nucleons. Hence, in the same fashion as in deducing (4.5.25), from (4.5.33) and
(4.5.34) we derive the following strong nucleon potential as
0 1
1 An −k1 r
Φn = βn gs − ϕ (r)e , (4.5.35)
r ρn
where ϕ (r) is as
r
ϕ (r) = 1 + k1r + + o(r), α0 as in (4.5.24), (4.5.36)
2α0
and An is a dimensionless parameter:

κn k12 cτ1
An = .
4βn
Note that βn is inversely proportional to the volume Vn of a nucleon. Hence we have
% &3
βn NV0 ρw
= =N (N = 3), (4.5.37)
β Vn ρn
where N is the number of strong charges in a nucleon, β is the parameter as in (4.5.25)
and (4.5.26), and V0 the volume of a w∗ -weakton. Since gs β in (4.5.25) is taken as gs , by
(4.5.37) the formula (4.5.35) is expressed as
0 1
1 An
Φn = gs (ρn ) − ϕ (r)e−kn r ,
r ρn
% &3 (4.5.38)
ρw
gs (ρn ) = 3 gs ,
ρn
and ϕ (r) is as in (4.5.36).
In summary, for a particle with N strong charges and radius ρ , its strong interaction
potential can be written as
0 1
1 A
Φ = gs (ρ ) − ϕ (r)e −kr
,
r ρ
% &3 (4.5.39)
ρw
gs (ρ ) = N gs ,
ρ
4.5 Strong Interaction Potentials 235

where ρw is the radius of w∗ -weakton, A is a dimensionless constant depending on the

particle, ϕ (r) = 1 + kr, and 1/k is the radius of strong attraction of particles. Phenomeno-
logically, we take
⎧ −18
⎪
⎪ 10 cm for w∗ -weaktons,
⎪
⎪ −16
1 ⎨ 10 cm for quarks,
= (4.5.40)
k ⎪ ⎪ 10 −13
cm for neutreons,
⎪
⎪
⎩ −7
10 cm for atom/molecule.
More specifically, we give the following layered formulas of strong interaction poten-
tials for various level of particles as the w∗ -weakton potential Φ0 , the quark potential Φq ,
the nucleon/hadron potential Φn and the atom/molecule potential Φa are given as follows
(Ma and Wang, 2014c):
0 1
1 A0
Φ0 = gs − (1 + k0r)e−k0 r ,
r ρw
% &3 0 1
ρw 1 Aq −k1 r
Φq = gs − (1 + k1r)e ,
ρq r ρq
% &3 0 1 (4.5.41)
ρw 1 An −kn r
Φn = 3 gs − (1 + knr)e ,
ρn r ρn
% &3 0 1
ρw 1 Aa −ka r
Φa = N gs − (1 + kar)e .
ρa r ρa
Here, k0 , k1 , kn , ka are taken as in (4.5.40):
1 1
= 10−18 cm, = 10−16 cm,
k0 k1
(4.5.42)
1 1
= 10−13 cm, = 10−10 ∼ 10−7 cm.
kn ka
Remark 4.18 For two particles with N1 , N2 strong charges and radii ρ1 , ρ2 , their strong
potential energy is 0 1
1 B
V = gs (ρ1 )gs (ρ2 ) − (1 + kr)e−kr , (4.5.43)
r ρ
where B and k depend on the two types of particles, and ρ depends on ρ 1 and ρ2 with ρ = ρ
if ρ = ρ1 = ρ2 . The strong force between the two particles is

F = −∇V, V as in (4.5.43). (4.5.44)

4.5.3 Quark confinement

Quark model was confirmed by lots of experiments. However, no any single quark is found
ever. This fact suggests that the quarks were permanently bound inside a hadron, which is
called the quark confinement. Up to now, no other theories can successfully describe the
236 Chapter 4 Unified Field Theory of Four Fundamental Interactions

quark confinement phenomena. The direct reason is that all current theories for interactions
fail to provide a successful strong interaction potential to explain the various level strong
interactions.
Now, we can derive the quark confinement by using the layered strong interaction po-
tentials (4.5.41).
The strong interaction bound energy E for two particles is given by (4.5.43). In particu-
lar, for two same particles we have
0 1
2 1 A
E = gs (ρ ) − (1 + kr)e −kr
,
r ρ
% &3 (4.5.45)
ρw
gs (ρ ) = gs .
ρ

The quark confinement can be well explained from the viewpoint of the strong quark
bound energy Eq and the nucleon bound energy En . In fact, by (4.5.45) we have
% &7
Eq Aq ρn
$ . (4.5.46)
En An ρq

According to the physical observation fact,

ρn $ 10−16 cm, ρq $ 10−19 cm. (4.5.47)

Then, by (4.5.46) we derive that

Eq Aq
$ 1021 .
En An
Physically Aq /An is no small, and we assume that Aq /An $ 10−1 . Then we have

Eq = 1020En . (4.5.48)

It is known that the bound energy of nucleons is about

En ∼ 10−2 GeV.

Thus, we obtain by (4.5.48) that

Eq ∼ 1018 GeV,

which is at the Planck level. This clearly shows that the quarks is confined in hadrons, and
no free quarks can be found.

Remark 4.19 The magnitude order (4.5.42) and (4.5.47) are not accurate, but they are
enough precisely to explain the important physical problems as the quark confinement and
asymptotic freedom etc.
4.5 Strong Interaction Potentials 237

4.5.4 Asymptotic freedom

The strong interaction potentials provide also a natural explanation for the asymptotic free-
dom phenomena. To this end, we need to introduce the asymptotic freedom in two per-
sptives: the deep inelastic scattering experiments, and the QCD theory for the coupling
constant of quark potentials.

1. Deep inelastic scattering experiments. In physical experiments, the interior of a

proton is probed by using the accelerating electrons to hit this proton. The collision is called
the elastic scattering if there is no momentum exchange as in the e-p elastic scattering:

e− + p → e− + p. (4.5.49)

The collision is an inelastic scattering if the particles are changed after the collision. For
example the usual e-p inelastic scatterings are as follows

e− + p → e− + π + + n,
(4.5.50)
e− + p → e− + π 0 + p.

In 1967, three physicists J. L. Friedman, H. W. Kendall and R. E. Taylor performed a

series of deep inelastic experiments, which not only provided sufficient evidence for the
existence of quarks, but also exhibited the asymptotic freedom phenomena. Due to their
pioneering investigations concerning deep inelastic scattering, the three physicists were
awarded the Nobel Prize in 1990.
In the e-p scattering experiments, if an electron at lower energy collides with a proton,
then the proton looks as a point particle, and the collision is an elastic scattering. However,
if the accelerating electron is at higher energy, this electron will hit deeply into the interior
of a proton, and collides with a quark in the proton, as shown in Figure 4.1. The experiments
show that the the collided quark acts as if it is a free particle.

Figure 4.1 Black dots represent three quarks in a proton

238 Chapter 4 Unified Field Theory of Four Fundamental Interactions

2. QCD theory for asymptotic freedom. The notion of asymptotic freedom was first
introduced by (Gross and Wilczek, 1973; Politzer, 1973), who were awarded the Nobel Prize
in 2004. By using the renormalization group, they derived the strong interaction coupling
constant of quarks as follows

4π
αs = % & , (4.5.51)
2
11 − n f ln(q /λ )
2 2
3

where λ is a parameter, n f = 6 is the flavor number of quarks, and q2 is the transfer mo-
mentum by the incidence electron.
In QCD, the coupling constant αs stands for the strength of the strong interactions be-
tween the quarks in a proton. It is clear that as the transfer momentum q 2 tends to infinitely
large, αs tends to zero:
αs → 0 as q2 → ∞. (4.5.52)

It means that the greater the incidence electron kinetics is, the more freedom the quarks
exhibit. The asymptotic freedom is based on the fact (4.5.51)-(4.5.52).

3. Strong interaction potential for asymptotic freedom. The strong interaction potentials
provide also a natural explanation for the asymptotic freedom phenomena. By (4.5.41) and
(4.5.44), the strong force between quarks is as
0 1
dΦq 1 Aq
F = −gs (ρq ) = g2s (ρq ) 2 − k12 re−k1 r ,
dr r ρq

which implies that there are two radii r 1 and r2 such that

ρq < r1 < r1 < r2 (r1 = 1/k1),

and the strong force F satisfies

>
>0 for 0 < r < r1 ,
F (4.5.53)
<0 for r1 < r < r 2 .

By (4.5.53), we see that

F ∼0 near r = r1 .

This indicates that there is a free shell region inside a proton with radius R ≃ r 1 , such that
the three quarks are free in this shell region.
When a lower energy electron collides with the proton, the electromagnetic force cause
the electron moving away, leading to the elastic scattering as given by (4.5.49). However,
when a high speed electron collides with a proton, it can run into the inside of this proton,
interacting with one of the quarks. Since the quark was in a free shell region with no force
4.5 Strong Interaction Potentials 239

acting upon it, this particular quark will behavior as a free quark. As it moves into the
attracting region of the proton, the quark confinement will hold this quark, which, at the
same time, will collide with gluons, exchanging quarks, leading to the inelastic scatterings
as given in (4.5.50). This explains the asymptotic freedom.
Remark 4.20 The two phenomena of the quark confinement and the asymptotic free-
dom are consistent. On the one hand, the quarks in a proton are held tightly together by the
strong interaction to cause the quark confinement, and on the other hand, quarks look as if
they are free. The consistency is the direct consequence of the quark potential formula in
(4.5.41).

4.5.5 Modified Yukawa potential

One of the mysteries of the strong interaction is the conflict characteristics exhibited in the
quark level and in the nucleon level. By the strong interaction potentials (4.5.41), we see
that
quark strong force is infinitely attractive at r = 10−16 cm; (4.5.54)
while in the nucleon level,

nucleon strong force is repulsive in 0 < r < 10−13 cm. (4.5.55)

The conflicting characteristics of the strong interactions demonstrated in (4.5.54) and (4.5.55)
can hardly be explained by any existing theory. However, the layered strong interaction po-
tentials in (4.5.41) derived based on PID and PRI lead to a natural explanation of these
characteristics, as well as explanation of the quark confinement and asymptotic freedom in
previous subsections. In this subsection, we shall show that the strong interaction nucleon
potential (4.5.38), which can be regarded as a modified Yukawa potential, fits experiments.
Meanwhile, we also determine some parameters for the strong interaction potentials, in-
cluding the strong charge gs .
1. Experimental results. Experiments showed that the strong nucleon force has the
following properties:

• Nucleon force is of short-ranged, with the radius of the force range: r ∼ 2 × 10 −13 cm.

• Nucleon force has a repelling center and an attracting region: it is repulsive for r <
1 1
× 10−13 cm, is attractive for × 10−13 < r < 2 × 10−13 cm, and diminishes for
2 2
r > 2 × 10−13. Namely, experimentally, the nucleon force behaves as
⎧
⎪ 1 −13
⎨ > 0 for 0 < r < r = 2 × 10 cm,
⎪
⎪
F exp < 0 for r < r < 2 × 10−13 cm, (4.5.56)
⎪
⎪
⎪
⎩
$ 0 for r > 2 × 10−13 cm.
240 Chapter 4 Unified Field Theory of Four Fundamental Interactions

More precisely, the experimental data for the strong interaction nucleon potential can
be schematically shown in Figure 4.2; see (Weisskopf, 1972).

2. Yukawa potential. Based on the classical strong interaction theory, the potential
holding nucleons to form an atomic nucleus is the Yukawa potential
g 1
ΦY = − e−kn r , r1 = = 10−13 cm, (4.5.57)
r kn
where g is the Yukawa strong charge, and
g2 = 1 ∼ 10 h̄c. (4.5.58)

Figure 4.2 Experimental curve of nucleon potential energy

By (4.5.57) we can deduce the classical nucleon force as

% &
dΦY 2 1 1
FY = −g = −g + e−kn r . (4.5.59)
dr r2 r1 r
It is clear that the nucleon force (4.5.59) is alway attractive, i.e.
FY < 0 for any r > 0,
FY → −∞ for r → 0, (4.5.60)
−13
FY /F1 ≃ 0 for r > 10 × 10 cm,
where F1 = 2g2 /er12 , and e is the base of the natural logarithm.
Comparing (4.5.60) with (4.5.56), we find that the Yukawa theory has a large error in
1
0 < r < × 10−13 cm. In particular, by (4.5.57) the Yukawa potential can be shown in
2
Figure 4.3.
3. Modified Yukawa potential. The nucleon potential Φn derived by the unified field
model based on PID and PRI is given by
% &3 0 1
ρw 1 An −kn r
Φn = 3 gs − (1 + knr)e ,
ρn r ρn
4.5 Strong Interaction Potentials 241

Figure 4.3 Theoretic curve of Yukawa potential energy

and the potential energy Vn of two nucleons is

% &3 % &6 0 1
ρw ρw 1 An
Vn = 3 gs Φn = 9 g2s − (1 + knr)e−kn r . (4.5.61)
ρn ρn r ρn
The nucleon force is given by
% &6 0 1
dVn ρw 2 1 An 2 −kn r
Fn = − =9 gs 2 − kn re , (4.5.62)
dr ρn r ρn
where
ρn = 10−16 cm, kn = 1013 cm−1 . (4.5.63)

4. Parameters An and g2s . First, we use the experimental data in (4.5.56) to determine
the parameter An in (4.5.62). By (4.5.56) we know that
1
F exp = 0, at r = × 10−13 cm.
2
Hence, let
1
Fn = 0, at r = × 10−13 cm.
2
Then, it follows from (4.5.62) and (4.5.63) that

An = ρn ekn r /kn2 r3 = 8e1/2 × 10−3. (4.5.64)

Next, we assume that

Fn = FY , at r1 = 10−13 cm. (4.5.65)

Then, by (4.5.59) and (4.5.62), we deduce from (4.5.65) that

% &6 0 1
ρw 2 1 An 2 −1 2
9 gs 2 − kn r1 e = −g2 2 e−1 ,
ρn r1 ρn r1
242 Chapter 4 Unified Field Theory of Four Fundamental Interactions

which leads to
% &6
2 ρn ρn
g2s = g2 .
9 (An kn r12 − eρn) ρw

In view of (4.5.63)-(4.5.64) and r1 = 1/kn , we derive

% &6
2 e−1/2 ρn
g2s = g2 , (4.5.66)
9 8 − e1/2 ρw

where e = 2.718 and g is as in (4.5.58).

Finally, from Fn = 0 we can also deduce that

Fn > 0 as r > r2 $ 9 × 10−13 cm.

4.5.6 Physical conclusions for nucleon force

The discussion above leads to the following physical conclusions:

1. The modified Yukawa potential based on PID and PRI is

2 % & 3
1 8e1/2 r
Φn = β g − 1+ e −r/r1
, (4.5.67)
r r1 r1
√ $ √
where β = 2/ 8 e − e, g is the Yukawa charge, r1 = 10−13 cm, and the modified
Yukawa force is 2 3
2 2 1 8e1/2 r −r/r1
Fn = β g − 2 e . (4.5.68)
r2 r1 r1

2. The nucleon strong interaction constant An is given by

An = 8e1/2 × 10−3, e = 2.718. (4.5.69)

3. By (4.5.66), the strong charge gs is given by

% &6
2 e−1/2 ρn
g2s = g2 , (4.5.70)
9 8 − e1/2 ρw

where ρw is the w∗ -weakton radius and ρn the nucleon radius.

4. The largest attraction force of Fn is achieved at r = 1.5r1 , and the force is given by
0 1
8e−1/2 g2 2 1
Fmax = − 3e − . (4.5.71)
8 − e1/2 r12 9
4.5 Strong Interaction Potentials 243

5. By (4.5.68), the attractive and repulsive regions of the strong nucleon force Fn are as
follows ⎧
⎪ > 0 for 0 < r < r1 /2,
⎨
Fn < 0 for r1 /2 < r < 9r1 , (4.5.72)
⎪
⎩
> 0 for 9r1 < r.

6. Based on (4.5.67), the theoretical curve of the strong interaction potential energy
for two nucleons is as shown in Figure 4.4, where we see that the theoretical result is in
agreement with the experimental curve shown in Figure 4.2.

Figure 4.4 Theoretical curve of nucleon potential energy

4.5.7 Short-range nature of strong interaction

By the layered potentials (4.5.41) for the strong interaction, we see that when the nucleons
form an atom, the nucleon potential is no longer valid, and the correct potential becomes
the strong interaction potential for atoms given by the fourth formula in (4.5.41). The cor-
responding force formula is given by
% &6 0 1
ρw 1 Aa
Fa = N 2 g2s 2 − ka2 re−ka r , (4.5.73)
ρa r ρa
where
for atom : ka = 10−9 ∼ 10−10 cm, ρa = 10−8 cm,
(4.5.74)
for molecule: ka = 10−7 cm, ρa = 10−7 cm.
It is clear that the attractive force in (4.5.73) is of short-ranged. The repulsive force in
(4.5.73) looks as if it is long-ranged. However by (4.5.74) the factors (ρ w /ρa )6 for the atoms
and molecules are very small. Hence, the strong repulsive forces for atoms and molecules
almost vanishes.
In fact, the bound force between atoms and molecules is the electromagnetic force with
strength given by
e2 1
= , e the electric charge. (4.5.75)
h̄c 137
244 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Hence at the atomic and molecular scale, the ratio between strong repulsive force and the
electromagnetic force is
% &6
Fa ρw
= Ns2 g2s /Ne2 e2 , (4.5.76)
Fe ρa
where Ns is the number of strong charge, and Ne is the number of the electric charge. Note
that each nucleon has three strong charges, and the protons are almost the same as neutrons.
Therefore, we assume that
Ns = 6Ne .
In view of (4.5.70), the ratio (4.5.76) becomes
% &6
Fa ρn 2
= 4β 2 g2 /e2 , β2 = √ $ 0.2.
Fe ρa 8 e−e

By (4.5.58) and (4.5.75), we have

g2 1 1
2
is in the range of ∼ .
e 137 13.7
Then, by ρn = 10−16 cm and (4.5.74), we derive that
>
Fa 10−50 at the atom level,
∼
Fe 10−56 at the molecular level.

This clearly demonstrates the short-range nature of the strong interaction.

4.6 Weak Interaction Theory

4.6.1 Dual equations of weak interaction potentials
According to the standard model, weak interaction is described by the SU(2) gauge theory.
In Section 4.3.4, we have demonstrated that the weak interaction potential is given by the
following PRI representation invariant

Wµ = ωaWµa = (W0 ,W1 ,W2 ,W3 ), (4.6.1)

where {ωa | 1 ! a ! 3} is the SU(2) tensor as in (4.3.29).

Also, the weak charge potential and weak force are as

Φw = W0 the time component of Wµ ,

(4.6.2)
Fw = −gw (ρ )∇Φw ,

where gw (ρ ) is the weak charge of a particle with radius ρ .

In this subsection, we establish the field equations for the dual potential Φw and φw of
the weak interaction from the field equations (4.4.48)-(4.4.50).
4.6 Weak Interaction Theory 245

Taking inner products of (4.4.48) and (4.4.49) with ωa respectively, we derive the field
equations of the two dual weak interaction potentials Φw and φw as follows
gw 1 gw
∂ ν Wν µ − κab gαβ Wαaµ Wβb − gwJµw = (∂µ − k2 xµ + γ Wµ )φw , (4.6.3)
h̄c 4 h̄c
0 1
1 ∂2
− ∆ φw + k02 φw − gw ∂ µ Jµw (4.6.4)
c2 ∂ t 2
gw F G 1
= ∂ µ κab gαβ Wαaµ Wβb − γ Wµ φw − k02 xµ ∂ µ φw ,
h̄c 4
c ω , W is as in (4.6.1), φ = ω φ a , and
where κab = εab c µ w a w
gw
Jµw = ωa ψγµ σa ψ , Wµν = ∂µ Wν − ∂ν Wµ + κabWµaWνb . (4.6.5)
h̄c
Experiments showed that the SU(2) gauge fields Wµa for weak interaction field particles
possess masses. In addition, the dual Higgs fields φwa of Wµa also have masses. In (4.6.3),
there is no massive term of Wµ . However, we see that on the right-hand side of (4.6.3) there
is a term
gw
γφwWµ , (4.6.6)
h̄c
which is spontaneously generated by PID, and breaks the gauge-symmetry. Namely, (4.6.6)
will vary under the following SU(2) gauge transformation
1
Wµa → Wµa − εbc
a b c
θ Wµ − ∂µ θ a
gw
We shall show that it is the spontaneous symmetry breaking term (4.6.6) that generates mass
from the ground state of φw .
It is clear that the following state

(Wµa , φwa , ψ ) = (0, φ0a , 0) with φ0a being constants,

is a solution of (4.6.3) and (4.6.4), which is a ground state of φwa . Let a0 = φ0a ωa , which is a
constant. Take the transformation

φw → φw + a0 , Wµa → Wµa , ψ → ψ.

Then the equations (4.6.3) and (4.6.4) are rewritten as

gw
∂ ν Wν µ − k12Wµ − κab gαβ Wαaµ Wβb − gwJµw (4.6.7)
0 h̄c 1
1 2 gw 1
= ∂µ − k0 xµ + γ Wµ φµ − a0 k02 xµ ,
4 h̄c 4
F1 ∂ G
− ∆ φw + k02 φw − gw∂ µ Jµw + k02 a0 (4.6.8)
c2 ∂ t 2
gw µ F G 1
= ∂ κab gαβ Wαaµ Wβb − γ Wµ (φµ + a0 ) − k02 xµ ∂ µ φw ,
h̄c 4
246 Chapter 4 Unified Field Theory of Four Fundamental Interactions

$
where k1 = gw γ a0 /h̄c represents mass.
Thus, (4.6.7) and (4.6.8) have masses as

k0 = mH c/h̄, k1 = mW c/h̄, (4.6.9)

where mH and mW are the masses of Higgs and W ± bosons. Physical experiments measured
the values of mH and mW as

mH $ 160 GeV/c2 , mW $ 80 GeV/c2 . (4.6.10)

By (4.4.46), equations (4.6.7) and (4.6.8) need to add three gauge fixing equations.
Based on the superposition property of the weak charge forces, the dual potentials W0 and
φw should satisfy linear equations, i.e. the time-component µ = 0 equation of (4.6.7) and
the equation (4.6.8) should be linear. Therefore, we have to take the three gauge fixing
equations in the following forms
F G h̄c a0 k02
κab gαβ Wαa0Wβb − ∂ µ (WµaW0b ) + γ W0 φw − x0 = 0,
gw 4
gw µ F G k2 (4.6.11)
∂ κab gαβ Wαaµ Wβb − γ Wµ φw − 0 xµ ∂ µ φw − k02 a0 = 0,
h̄c 4
∂ µ Wµ = 0,
and with the static conditions
∂ ∂
Φw = 0, φw = 0, (Φw = W0 ). (4.6.12)
∂t ∂t
With the equations (4.6.11) and the static conditions (4.6.12), the time-component µ = 0
equation of (4.6.7) and its dual equation (4.6.8) become
1
− ∆Φw + k12 Φw = gw Qw − k02 cτφw , (4.6.13)
4
− ∆φw + k02 φw = gw ∂ µ Jµw , (4.6.14)

where cτ is the wave length of φw , Qw = −J0w , and Jµw is as in (4.6.5). The two dual equations
(4.6.13) and (4.6.14) for the weak interaction potentials Φw and φw are coupled with the
Dirac equations (4.4.50), written as
D gw E mc
iγ µ ∂µ + i Wµa σa ψ − ψ = 0. (4.6.15)
h̄c h̄
In addition, by (4.6.9) and (4.6.10) we can determine values of the parameters k 0 and k1 as
follows
k0 = 2k1 , k1 = 1016 cm−1 . (4.6.16)
The parameters 1/k0 and 1/k1 represent the attracting and repulsive radii of weak interaction
forces.
In the next subsection, we shall apply the equations (4.6.13)-(4.6.15) to derive the lay-
ered formulas of weak interaction potentials.
4.6 Weak Interaction Theory 247

4.6.2 Layered formulas of weak forces

Now we deduce from (4.6.13)-(4.6.16) the following layered formulas for the weak interac-
tion potential: 0 1
−kr 1 B
Φw = gw (ρ )e − (1 + 2kr)e −kr
,
r ρ
% &3 (4.6.17)
ρw
gw (ρ ) = N gw ,
ρ
where Φw is the weak force potential of a particle with radius ρ and carrying N weak
charges gw (gw is the unit of weak charge for each weakton, an elementary particle), ρ w is
the weakton radius, B is a parameter depending on the particles, and
1
= 10−16 cm, (4.6.18)
k
represents the force-range of weak interaction.
To derive the layered formulas (4.6.17), first we shall deduce the following weak inter-
action potential for a weakton
0 1
0 −kr 1 B0
Φw = gs e − (1 + 2kr)e −kr
. (4.6.19)
r ρw
To derive the solution φw of (4.6.14), we need to compute the right-hand term of (4.6.14).
By (4.6.5) we have
∂ µ Jµw = ωa ∂µ ψγ µ σa + ωaψγ µ σa ∂µ ψ .
Due to the dirac equation (4.6.15),
gw b µ mc
∂µ ψγ µ σa ψ = −i W ψγ σb σa ψ + i ψσa ψ ,
h̄c µ h̄
µ gw b µ mc
ψγ σa ∂µ ψ = i Wµ ψγ σa σb ψ − i ψσa ψ .
h̄c h̄
Thus we obtain
gw gw c
∂ µ Jµw = i ωaWµb ψγ µ [σa , σb ] ψ = −2 εab ωaWµb Jcµ . (4.6.20)
h̄c h̄c
µ
c σ and J = ψγ µ σ ψ . Note that
Here we used [σa , σb ] = i2εab c c c
! !
! ⃗i ⃗j ⃗k !
c
! !
εab ωaWµb = !! ω1 ω2 ω3 !! = ω ⃗ µ,
⃗ ×W
!W 1 W 2 W 3 !
µ µ µ

where ω ⃗ µ = (Wµ1 ,Wµ2 ,Wµ3 ). Hence, (4.6.20) can be rewritten as

⃗ = (ω1 , ω2 , ω3 ), W

2gw
∂ µ Jµw = − ⃗ µ ) · J⃗µ ,
ω ×W
(⃗ (4.6.21)
h̄c
248 Chapter 4 Unified Field Theory of Four Fundamental Interactions

µ µ µ µ
and J⃗µ = (J1 , J2 , J3 ). The weak current density Ja is as

Jaµ = θaµ δ (r), θaµ the constant tensor,

⃗ µ in (4.6.21) is replaced by the average value
and W
!
1 ⃗ µ dx,
W (4.6.22)
|Bρw | Bρw

where Bρw ⊂ Rn is the ball with radius ρw . Similar to the case (4.5.14) for the strong
interaction, the average value (4.6.22) is
⃗ µ = ⃗ζµ /ρw ,
W ⃗ζµ = (ζ 1 , ζ 2 , ζ 3 ).
µ µ µ

Thus, (4.6.21) can be expressed as

∂ µ Jµw = −κδ (r)/ρw , (4.6.23)

and κ is a parameter, written as

2gw ⃗ µ
κ= ω × ⃗ζµ ).
θ · (⃗ (4.6.24)
h̄c
Putting (4.6.23) in (4.6.14) we deduce that
gw κ
−∆φw + k02 φw = − δ (r),
ρw
whose solution is given by
gw κ 1 −k0 r
φw = − e . (4.6.25)
ρw r
Therefore we obtain the solution of (4.6.14) in the form (4.6.25).
Inserting (4.6.25) into (4.6.13) we get
gw B 1 −k0 r
− ∆Φw + k12 Φw = gw Qw + e , (4.6.26)
ρw r
where B is a parameter with dimension 1/L, given by
1
B = κ k02 cτ , κ is as in (4.6.24).
4
Since gw Qw = −gw J0w is the weak charge density, we have

Qw = β δ (r),

and β is a scaling factor. We can take proper unit for gs such that β = 1. Thus, putting
Qw = δ (r) in (4.6.26) we derive that
gw B 1 −k0 r
− ∆Φw + k12 Φw = gw δ (r) + e . (4.6.27)
ρw r
4.6 Weak Interaction Theory 249

Let the solution of (4.6.27) be radially symmetric, then the equation (4.6.27) can be equiv-
alently written as
% &
1 d 2 d gw B 1 −k0 r
− 2 r Φw + k12 Φw = gw δ (r) + e . (4.6.28)
r dr dr ρw r
The solution of (4.6.28) can be expressed as
gw −k1 r gw B
Φw = e − ϕ (r)e−k0 r , (4.6.29)
r ρw
where ϕ (r) satisfies the equation
% & % &
1 2k0 2 2 1
ϕ +2
′′
− k0 ϕ −
′
+ k1 − k0 ϕ = . (4.6.30)
r r r
Let ϕ be in the form
∞
ϕ= ∑ βn r n (the dimension of ϕ is L).
n=0

Inserting ϕ into (4.6.30), comparing coefficients of r n , we get

1
β 1 = k0 β 0 + ,
2
1 2 1
β2 = (k0 − k12 )β0 + k0 ,
2 3
..
.
2k0
βn = βn−1 − (k02 − k12 )βn−2 , n " 2.
n+1
Note that the dimensions of B and β0 in ϕ (r) are 1/L and L. The parameter B = B β0
is dimensionless. Physically, we can only measure the value of B. Therefore we take ϕ in
its second-order approximation as follows

ϕ = β0 (1 + k0r).

In addition, by (4.6.16) we take

k1 = k, k0 = 2k, k = 1016 cm−1 .

Thus, the formula (4.6.29) is written as

0 1
1 B
Φs = gw e −kr
− (1 + 2kr)e −kr
.
r ρw
This is the weak interaction potential of a weakton, which is as given in (4.6.19).
In the same fashion as used in the layered formula (4.5.39) for the strong interaction,
for a particle with radius ρ and N weak charges gw , we can deduce the layered formula of
weak interaction potentials as in the form given by (4.6.17).
250 Chapter 4 Unified Field Theory of Four Fundamental Interactions

4.6.3 Physical conclusions for weak forces

As mentioned earlier, the layered weak interaction potential formula (4.6.17) plays the same
fundamental role as the Newtonian potential for gravity and the Coulomb potential for elec-
tromagnetism. Hereafter we explore a few direct physical consequences of the weak inter-
action potentials.
1. Short-range nature of weak interaction. By (4.6.17) it is easy to see that for all par-
ticles, their weak interaction force-range is as
1
r= = 10−16 cm,
k
which is consistent with experimental data.
2. Weak force formula. For two particles with radii ρ1 , ρ2 , and with N1 , N2 weak charges
gw . Their weak charges are given by
% &3
ρw
gw (ρ j ) = N j gw for j = 1, 2. (4.6.31)
ρj
The weak potential energy for the two particles is
0 1
1 B
V = gw (ρ1 )gw (ρ2 )e−kr − (1 + 2kr)e−kr , (4.6.32)
r ρ
where gw (ρ1 ) and gw (ρ2 ) are as in (4.6.31), ρ is a radius depending on ρ1 and ρ2 , and B is
a constant depending on the types of these two particles.
The weak force between the two particles is given by
0 1
d 1 k 4B
F = − V = gw (ρ1 )gw (ρ2 )e−kr 2 + − k2 re−kr . (4.6.33)
dr r r ρ

3. Repulsive condition. For the two particles as above, if their weak interaction constant
B satisfies the inequality
1 k 4B 2 −kr 1
+ " k re , ∀0 < r ! ,
r2 r ρ k
or equivalently B satisfies
e
B ! ρ k (e = 2.718, k = 1016 cm−1 ), (4.6.34)
2
then the weak force between these two particles is always repulsive.
It follows from this conclusion that for the neutrinos:

(ν1 , ν2 , ν3 ) = (νe , νµ , ντ ),
(ν 1 , ν 2 , ν 3 ) = (ν e , ν µ , ν τ ),
4.6 Weak Interaction Theory 251

the weak interaction constants

Bi j for νi and ν j ∀1 ! i, j ! 3,
Bi j for νi and anti-neutrinos ν j ∀i ̸= j,

satisfy the exclusion condition (4.6.34).

4. Value of weak charge gw . Based on the Standard Model, the coupling constant Gw of
the β -decay of nucleons and the Fermi constant G f have the following relation

8 D mW c E2
G2w = √ Gf , (4.6.35)
2 h̄
and G f is given by
D m c E2
p
G f = 10−5h̄c/, (4.6.36)
h̄
where mW and m p are masses of W ± bosons and protons. By the gauge theory, Gw is also the
coupling constant of SU(2) gauge fields. Therefore we can regard G w as the weak charge
of nucleons, i.e.
Gw = gw (ρn ), ρn the nucleon radius.
In addition, it is known that % &3
ρw
gw (ρn ) = 9 gs .
ρn
Hence, we deduce from (4.6.35) and (4.6.36) that
% &6 % &2
ρw 8 mw
81 g2w = √ × 10−5h̄c.
ρn 2 mp

Then we derive the relation

% &2 % &6
8 mw ρn
g2w = √ × 10 −5
× h̄c. (4.6.37)
81 2 mp ρw

In a comparison with (4.5.66) and (4.6.37), we find that the strong charge g s and the
weak charge gw have the same magnitude order. Direct computation shows that

g2w gw
≃ 0.35 equivalently ≃ 0.6.
g2s gs

4.6.4 PID mechanism of spontaneous symmetry breaking

In the derivation of equations (4.6.7) and (4.6.8) we have used the PID mechanism of spon-
taneous symmetry breaking. In this subsection we shall discuss the intermediate vector
bosons W ± , Z and their dual scalar bosons H ± , H 0 , called the Higgs particles by using the
PID mechanism of spontaneous symmetry breaking.
252 Chapter 4 Unified Field Theory of Four Fundamental Interactions

We know that the interaction field equations are oriented toward two directions: i) to
derive interaction forces, and ii) to describe the field particles and derive the particle transi-
tion probability. The PID-PRI weak interaction field equations describing field particles are
given by (4.4.42)-(4.4.45). Here, for convenience, we write them again as follows:
% &
gw a αβ b c 1 D mH c E 2 gw
∂ ν Wνaµ − εbc g Wα µ Wβ − gw Jµa = ∂µ − xµ + γbWµb φ a , (4.6.38)
h̄c 4 h̄ h̄c
D m c E2 g
H w
− ∂ µ ∂µ φ a + φ a − γb ∂ µ (Wµb φ a ) (4.6.39)
h̄ h̄c
gw a αβ µ b c 1 D mH c E2
= εbc g ∂ (Wα µ Wβ ) + gw∂ µ Jµa − xµ ∂ µ φ a ,
h̄c 4 h̄

where mH is the Higgs particle mass, and Wµa (1 ! a ! 3) describe the intermediate vector
bosons as follows
W ± : Wµ1 ± iWµ2 ,
(4.6.40)
Z: Wµ3 ,

and φ a describe the dual Higgs bosons as

H ± : φ 1 ± iφ 2 ,
(4.6.41)
H0 : φ 3.

By (4.4.46), equations (4.6.38)-(4.6.39) need to be supplemented with three gauge fixing

equations. According to physical requirement, we take these equations as

∂ µ Wµa = 0 for 1 ! a ! 3. (4.6.42)

To match (4.6.40) and (4.6.41), we take the SU(2) generator transformation as follows
⎛ ⎞ ⎛ ⎞⎛ ⎞
σ'1 1 i 0 σ1
⎝σ'2 ⎠ = √1 ⎝1 −i 0 ⎠ ⎝σ2 ⎠ .
2 0 0 √
σ'3 2 σ3

Under this transformation, Wµa and φ a are changed into

D E
(W' µ1 , W 'µ3 ) = (Wµ± , Zµ ) = √1 (Wµ1 ± iWµ2 ),Wµ3 ,
'µ2 , W
2
D 1 E
(φ' , φ' , φ' ) = (φ , φ ) = √ (φ 1 ± iφ 2 ), φ 3 .
1 2 3 ± 0
2

The equations (4.6.38) and (4.6.39) obey PRI, and under the above transformation they
4.6 Weak Interaction Theory 253

becomes
igw αβ ±
∂ ν Wν±µ ± g (W )α µ Zβ − Zα µ Wβ± ) − gwJµ± (4.6.43)
0 h̄c 1
2 1 D mH c E2
= ∂µ + kW Wµ± + kZ2 Zµ − xµ φ ± ,
4 h̄
ig w αβ
∂ ν Zν µ − g (Wα+µ Wβ− − Wα−µ Wβ+ ) − gwJµ0 (4.6.44)
0 h̄c 1
2 2 1 D mH c E2
= ∂µ + kW Wµ + kZ Zµ −
±
xµ φ 0 ,
4 h̄
Dm cE gw
H
#H ± + H ± − ∂ µ (γ'bW ' µb H ± ) (4.6.45)
h̄ h̄c
gw αβ µ ± ' b ' c 1 D mH c E 2
= g ∂ (' εbcWα µ Wβ ) + gw ∂ µ Jµ± − xµ ∂ µ H ± ,
h̄c 4 h̄
D m c E2 gw
H
#H 0 + H 0 − ∂ µ (γ'bW ' µb H 0 ) (4.6.46)
h̄ h̄c
gw αβ µ 0 ' b ' c 1 D mH c E 2
= g ∂ (' εbcWα µ Wβ ) + gw ∂ µ Jµ0 − xµ ∂ µ H 0 ,
h̄c 4 h̄
1
where H ± = √ (φ 1 ± iφ 2 ), H 0 = φ 3 in (4.6.45) and (4.6.46), and
2
1
Jµ± = √ (Jµ1 ± iJµ2 ) the charged current,
2
Jµ0 = Jµ3 the neutral current,
igw (4.6.47)
Wν±µ = ∂ν Wµ± − ∂µ Wν± ± (Zµ Wν± − Zν Wµ± ),
h̄c
igw + −
Zν µ = ∂ν Zµ − ∂µ Zν + (W W − Wν+Wµ− ),
h̄c µ ν
and
2 gw γ1 gw γ3
kW =√ , kZ2 = , (4.6.48)
2h̄c h̄c
which are derived by the following transformation
⎛ 2⎞ ⎛ ⎞⎛ ⎞
kW 1 i 0 γ1
g
2 ⎠ = √ w ⎝1 −i ⎠ ⎝γ2 ⎠ ,
⎝kW
√0
k2 2h̄c 0 0 2 γ3
Z

with γ2 = 0 in the Pauli matrix representation.

The equation (4.6.43)-(4.6.46) are the model to govern the behaviors of the weak inter-
action field particles (4.6.40) and (4.6.41). Two observations are now in order.

First, we note that these equations are nonlinear, and consequently, no free weak inter-
action field particles appear.
254 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Second, there are two important solutions of (4.6.43)-(4.6.44), dictating two different
weak interaction procedures.
The first solution sets
Wµ± = 0, φ 0 = 1. (4.6.49)

Then Z satisfies the equation

1 D mH c E 2
#Zµ + kz2 Zµ = −gw Jµ0 + xµ . (4.6.50)
4 h̄
The second solution takes
Zµ = 0, φ ± = 1. (4.6.51)

Then Wµ± satisfy the equations

1 D mH c E 2
#Wµ± + kw2 Wµ± = −gw Jµ± + xµ . (4.6.52)
4 h̄
We are now ready to obtain the following physical conclusions for the weak interaction
field particles.

1. Duality of field particles. The field equations (4.6.43)-(4.6.46) provide a natural

duality between the field particles:

W ± ↔ H ±, Z ↔ H 0.

2. Lack of freedom for field particles. Due to the nonlinearity of (4.6.43)-(4.6.46), the
weak interaction field particles W ± , Z, H ± , H 0 are not free bosons.

3. PID mechanism of spontaneous symmetry breaking. In the field equations (4.6.50)

and (4.6.52), it is natural that the masses mW and mZ of W ± and Z bosons appear at the
ground states (4.6.49) and (4.6.51), with

h̄ h̄
mW = kW , mZ = k Z ,
c c
and kW and kZ are as in (4.6.48).

4. Basic properties of field particles. From (4.6.43)-(4.6.46) we can obtain some basic
information for the bosons as follows:

W± : spin J = 1, electric charge = ±e, mass mW ,

Z: spin J = 1, electric charge = 0, mass mZ ,
±
H : spin J = 0, electric charge = ±e, mass m+
H,
H0 : spin J = 0, electric charge = 0, mass m0H .
4.6 Weak Interaction Theory 255

Remark 4.21 Under the translation

1 D mH c E 2 1
Zµ → Zµ + xµ ,
4 h̄ kZ2
1 mH c 2 1
Wµ± → Wµ± + ( ) 2 xµ ,
4 h̄ kW
the equations (4.6.50) and (4.6.52) become

#Zµ + kZ2 Zµ = −gw Jµ0 ,

2
(4.6.53)
#Wµ± + kW Wµ± = −gw Jµ± .

Similarly, if we take the gauge fixing equations

xµ ∂ µ φ a = 0 for 1 ! a ! 3,

instead of (4.6.42), then under the conditions

Wµa = 0 for 1 ! a ! 3,

the equations (4.6.46) and (4.6.45) are in the form

D m c E2
H
#H 0 + H 0 = gw ∂ µ Jµ0 ,
h̄ (4.6.54)
D m c E2
H
#H ± + H ± = gw ∂ µ Jµ± .
h̄
The equations (4.6.53) and (4.6.54) are the standard Klein-Gordon models describing the
W ± , Z, H ± , H 0 bosons, which are derived only by the PID-PRI unified field model.

4.6.5 Introduction to the classical electroweak theory

In comparison with the PID-PRI weak interaction theory, in this subsection we briefly intro-
duce the classical U(1) × SU(2) electroweak theory; see among many others (Kaku, 1993;
Griffiths, 2008; Quigg, 2013).
1. The fields in the electroweak theory are as follows:

SU(2) gauge fields: Wµ1 ,Wµ2 ,Wµ3 ,

U(1) gauge field: Bµ , % &
νL
Dirac spinor doublets: L= ,
lL
Dirac spinor singlet: R = l%R , &
φ+
Higgs scalar doublet: φ= ,
φ0

where νL and lL are the left-hand neutrino and lepton, lR is the right-hand lepton, (φ + , φ 0 )
are scalar fields with electric charge (1,0).
256 Chapter 4 Unified Field Theory of Four Fundamental Interactions

2. The Lagrangian action of the electroweak theory is given by

LW S = LG + LD + LH , (4.6.55)
where LG is the gauge sector, LD is the Dirac sector, and LH is the Higgs sector:
1 a µν a 1
LG = − Wµν W − Bµν Bµν ,
4 4
LD = iLγ µ Dµ L + iRγ µ Dµ R, (4.6.56)
1 λ
LH = (Dµ φ )† (Dµ φ ) + (φ † φ − a2)2 − Gl (Lφ R + Rφ † L).
2 4
Here λ , a, Gl are constants,
a
Wµν = ∂µ Wνa − ∂ν Wµa + g1 εbc
a
WµbWνc ,
Bµν = ∂µ Bν − ∂ν Bµ ,
Dµ R = (∂µ + ig2 Bµ )R, (4.6.57)
g2 g1
Dµ L = (∂µ + i Bµ − i Wµa σa )L,
2 2
g2 g1
Dµ φ = (∂µ − i Bµ − i Wµa σa )φ ,
2 2
g1 , g2 are coupling constants of SU(2) and U(1) gauge fields, and σa (1 ! a ! 3) are the
Pauli matrices.
3. The action (4.6.55)-(4.6.57) is invariant under the following SU(2) and U(1) gauge
transformations:
• SU(2) gauge transformation:
i aσ
L → e2θ a
L,
i a
φ →e 2 θ σa φ,
R → R,
2
Wµa → Wµa − ∂µ θ a + εbc
a b c
θ Wµ , i.e. as in (2.4.38),
g1
Bµ → Bµ .

• U(1) gauge transformation:

i
L → e 2 β L,
i
φ → e− 2 β φ ,
R → eiβ R,
2
Wµa → Wµa − ∂µ β ,
g2
2
Bµ → Bµ + ∂µ β .
g2
4.6 Weak Interaction Theory 257

4. The electroweak field equations are derived by

δ LW S δ LW S δ LW S δ LW S δL
= 0, = 0, = 0, = 0, = 0. (4.6.58)
δ Wµa δ Bµ δφ δL δR
We remark here that the Higgs field in this setting is included in the Lagrangian action,
drastically different from the mechanism based on PID developed earlier.
In addition, the particle φ + represents a massless boson with a positive electric charge.
However, in reality no such particles exist. Hence we have to take the Higgs scalar field as
% &
0
φ= . (4.6.59)
ϕ
In fact, in the WS theory, the Higgs field φ is essentially taken as the form (4.6.59). Under
the condition (4.6.59), the variational equations (4.6.58) of the SW action (4.6.55)-(4.6.57)
are expressed as follows:
Gauge field equations (massless):
g1 1 g21 2 1
∂ ν Wν1µ − g1 gαα (Wα2µ Wα3 − Wα3µ Wα2 ) + J − ϕ Wµ = 0,
2 µ 2
g1 g2
∂ ν Wν2µ − g1 gαα (Wα3µ Wα1 − Wα1µ Wα3 ) + Jµ2 − 1 ϕ 2Wµ2 = 0,
2 2 (4.6.60)
g1 g1
∂ ν Wν3µ − g1 gαα (Wα1µ Wα2 − Wα2µ Wα1 ) + Jµ3 − ϕ 2 (g1Wµ3 − g2Bµ ) = 0,
2 2
ν g2 L R g2 2 3
∂ Bν µ − Jµ − g2Jµ − ϕ (g2 Bµ − g1Wµ ) = 0.
2 2
Higgs field equations:
1
∂ µ ∂µ ϕ − ϕ (g21WµaW µ a + g22 Bµ Bµ − 2g1g2Wµ3 Bµ ) (4.6.61)
4
− λ ϕ (ϕ − a2 ) + Gl (l L R + RlL ) = 0.
2

Dirac equations:
iγ µ (∂µ + ig2Bµ )R − Gl ϕ lL = 0,
% & % &
g2 g1 ν 0 (4.6.62)
iγ µ (∂µ + i Bµ − i Wµa σa ) L − Gl R = 0.
2 2 lL ϕ
Here
Jµa = Lγµ σa L 1 ! a ! 3,
JµL = ν L γµ νL + l L γµ lL ,
JµR = Rγµ R, γµ = g µα γ α .

5. Masses are generated at the ground states. It is clear that the following state

ϕ = a, Wµa = 0, Bµ = 0, L = 0, R = 0,
258 Chapter 4 Unified Field Theory of Four Fundamental Interactions

is a solution of (4.6.60)-(4.6.62), called a ground state. We take the translation transforma-

tion
ϕ → ϕ + a, Wµa → Wµa , Bµ → Bµ , L → L, R → R,

then the massless equations (4.6.60) become massive, written as

⎛ 1 ⎞ ⎛ 1⎞
Wν µ Wµ
⎜W 2 ⎟ ⎜W 2 ⎟
⎜ ⎟ ⎜ ⎟
∂ ν ⎜ ν3µ ⎟ − M ⎜ µ3 ⎟
⎝Wν µ ⎠ ⎝Wµ ⎠
Bν µ Bµ
⎛ g1 1 ⎞
g1 gαβ εab1
Wαaµ Wβb − Jµ1 + g21 (ϕ 2 + 2aϕ )Wµ1
⎜ 2 2 ⎟
⎜ g1 2 1 2 2 ⎟
⎜ αβ 2 a b
g1 g εabWα µ Wβ − Jµ + g1 (ϕ + 2aϕ )Wµ 2 ⎟
⎜ 2 2 ⎟
⎜
=⎜ ⎟, (4.6.63)
⎜g1 gαβ ε 3 W a W b − g1 J 3 + 1 g1 (ϕ 2 + 2aϕ )(g1W 3 − g2Bµ )⎟
⎟
⎜ ab α µ β 2 µ
2 µ ⎟
⎝ ⎠
1 1
g2 JµL + g2 JµR + g2 (ϕ 2 + 2aϕ )(g2Bµ − g1Wµ3 )
2 2
where M is the mass matrix given by
⎛ ⎞
m21 0 0 0
2
c ⎜0 m21 0 0 ⎟
M= 2⎜ ⎟, (4.6.64)
h̄ ⎝ 0 0 m21 −m23 ⎠
0 0 −m23 m22

and √
m1 c g 1 a m2 c g 2 a m3 c g1 g2 a
=√ , =√ , = √ .
h̄ 2 h̄ 2 h̄ 2

6. The masses mW and mZ can be derived from (4.6.63) and (4.6.64) as follows. Ac-
cording to the IVB theory for particle transition, the W ± particles are characterized as

W ± : Wµ1 ± iWµ2 .

Hence we need the following transformation for Wµ1 and Wµ2 :

% +& % & % 1&
Wµ 1 1 i Wµ
=√ , (4.6.65)
Wµ− 2 1 −i Wµ2

which requires the following transformation of SU(2) generators from the Pauli matrices
σa : ⎛ ⎞ ⎛ ⎞⎛ ⎞
σ'1 1 i 0 σ1
⎝σ'2 ⎠ = √1 ⎝0 −i 0 ⎠ ⎝σ2 ⎠ .
2 0 √
σ'3 0 2 σ3
4.6 Weak Interaction Theory 259

In addition, we also need to obtain the mass mZ of Z boson by diagonalizing the massive
matrix (4.6.64). It is easy to see that
⎛ 2 ⎞
mW 0 0 0
c2 ⎜ 0 mW 2 0 0⎟
UMU † = 2 ⎜ ⎝ 2
⎟,
⎠
h̄ 0 0 m Z 0
0 0 0 0
⎛ 1 i ⎞
√ √ 0 0 (4.6.66)
⎜ 2 2 ⎟
⎜ ⎟
⎜ 1 i ⎟ g1 g2
U =⎜ ⎜ √2 − √2 0 0⎟⎟, α= , β= ,
⎜ ⎟ |g| |g|
⎝ 0 0 α β⎠
0 0 −β α
.
where |g| = g21 + g22 , and

c2 mW 2 a2 2 c 2 mZ a2 2
= g , = |g| . (4.6.67)
h̄2 2 1 h̄2 2

7. The field equations governing W ± and Z bosons are obtained from the equations
(4.6.63) under the following transformation
⎛ +⎞ ⎛ 1⎞
Wµ Wµ
⎜Wµ− ⎟ ⎜Wµ2 ⎟
⎜ ⎟ ⎜ ⎟
⎝ Zµ ⎠ = U ⎝W 3 ⎠ for U as in (4.6.66). (4.6.68)
µ
Aµ Bµ
In this case, the equations (4.6.63) become
⎛ D E2 ⎞
∂ ν W + − mW c W + ⎛g ⎞
⎜ νµ µ ⎟ 1 +
⎜ h̄ √
⎟ ⎜ 2 ⎟ J µ
⎜ D E ⎟
⎜∂ ν W − − mW c 2 W − ⎟ ⎜ ⎜√g1 − ⎟ ⎟
⎟ = ⎜ 2 Jµ ⎟
⎜ νµ µ ⎟
⎜ h̄ ⎜
⎟ + higher order terms. (4.6.69)
⎜ D m c E2 ⎟ ⎜
⎜ ν Z ⎟ ⎝ |g|J 0 ⎟
⎜ ∂ Zν µ − Zµ ⎟ µ ⎠
⎝ h̄ ⎠ em
−eJµ
∂ ν Aν µ
where e = g1 g2 /|g|, and
1
Jµ± = (Jµ1 ± iJµ2 ),
2
1
Jµ0 = (g2 J 3 − g22 JµL − 2g22JµR ),
2|g|2 1 µ
1
Jµem = (Jµ3 + JµL + 2JµR).
2
260 Chapter 4 Unified Field Theory of Four Fundamental Interactions

Under the transformation (4.6.68),

Zµ = cos θwWµ3 + sin θw Bµ ,

(4.6.70)
Aµ = − sin θwWµ3 + cos θw Bµ ,

where θw is called the Weinberg angle, defined by

g1 g2
cos θw = , sin θw = .
|g| |g|

8. The Higgs field equation governed H 0 boson is given by (4.6.61), and at the ground
state ϕ = a which can be written as
D m c E2 1
H
∂ µ ∂µ ϕ − ϕ = (ϕ + a)(g21WµaW µ a + g22Bµ Bµ − 2g1g2Wµ3 Bµ ) (4.6.71)
h̄ 4
+ λ ϕ (ϕ 2 + 2aϕ ) + Gl (l L R + RlL ),

and the Higgs boson mass is

mH c √
= 2λ a. (4.6.72)
h̄
4.6.6 Problems in WS theory
The classical electroweak theory provides a model with some experimental supports. How-
ever, this theory faces a number of problems, which are difficult, if not impossible, to re-
solve.
1. Lack of weak force formulas. This problem is that all weak interaction theories have
to face, and it is also that all existed theories cannot solve.
In fact, in the original field equations (4.6.60) there are four gauge field components:

Wµ1 , Wµ2 , Wµ3 , Bµ , (4.6.73)

and we don’t know which of these potentials plays the role of weak interaction potential. In
fact, with the mixed fields
Wµ± , Zµ , Aµ ,
it is even more difficult to determine the weak force.
If we combine the classical electroweak theory with PRI, and take

Wµ = ω1Wµ1 + ω2Wµ2 + ω3Wµ3

as the weak interaction potential, then from the field equations (4.6.60), we can only deduce
the weak force potential in the following form:
g1 −kr g1 −kr 1
Φw = e or Φw = − e , k = √ ag1 .
r r 2
4.6 Weak Interaction Theory 261

It implies that the weak force is only repulsive or attractive, which is not consistent with
experiments.

2. Violation of PRI. In the classical electroweak theory, a key ingredient is the linear
combinations of Wµ3 and Bµ as given by (4.6.70). By PRI,

Wµ3 is the third component of a SU(2) tensor {Wµa },

Bµ is the U(1) gauge field.

Hence, for the combinations of two different types of tensors:

Zµ = cos θwWµ3 + sin θw Bµ ,

Aµ = − sin θwWµ3 + cos θw Bµ ,

their field equations (4.6.69) must vary under general SU(2) generator transformations as
follows
σ'a = xba σb . (4.6.74)

In other words, such linear combinations violates PRI.

3. Decoupling obstacle. The classical electroweak theory has a difficulty for decoupling
the electromagnetic and the weak interactions. In reality, electromagnetism and weak inter-
action often are independent to each other. Hence, as a unified theory for both interactions,
one should be able to decouple the model to study individual interactions. However, the
classical electroweak theory manifests a radical decoupling obstacle.
In fact, it is natural to require that under the condition

Wµ± = 0, Zµ = 0, (4.6.75)

the WS field equations (4.6.69) should return to the U(1) gauge invariant Maxwell equa-
tions. But we see that
Aµ = cos θw Bµ − sin θwWµ3 ,

where B µ is a U(1) gauge field, and Wµ3 is a component of SU(2) gauge field. Therefore,
Aµ is not independent of SU(2) gauge transformation. In particular, the condition (4.6.75)
means
Wµ1 = 0, Wµ2 = 0, Wµ3 = − tgθw Bµ . (4.6.76)

Hence, as we take the transformation (4.6.74), Wµa becomes

⎛ ⎞ ⎛ ⎞
W'µ1 y1Wµ3
⎜ ' 2 ⎟ ⎝ 32 3 ⎠
⎝Wµ ⎠ = y3Wµ , (yba )T = (xba )−1 .
W'µ3 y33Wµ3
262 Chapter 4 Unified Field Theory of Four Fundamental Interactions

It implies that under a transformation (4.6.74), a nonzero weak interaction can be generated
from a zero weak interaction field of (4.6.75)-(4.6.76):

' µ± ̸= 0,
W Z'µ ̸= 0 as ya3 ̸= 0 (1 ! a ! 3),

and the nonzero electromagnetic field A µ ̸= 0 will become zero:

'µ = 0
A as y33 = cot θw .

Obviously, it is not reality.

4. Artificial Higgs mechanism. In the classical electroweak action (4.6.55)-(4.6.57), the
Higgs sector LH is not based on a first principle, and is artificially added into the action.
5. Presence of a massless and charged boson φ + . In the WS theory, the Higgs scalar
doublet φ = (φ + , φ 0 ) contains a massless boson φ + with positive electric charge. Obviously
there are no such particles in reality. In particular, the particle φ + is ”formally suppressed”
in the WS theory by transforming it to zero in the WS theory. However, from a field theo-
retical point of view, this particle field still represents a particle. This is one of major flaws
for the electroweak theory and for the standard model.
Chapter 5
Elementary Particles

The aims of this chapter are as follows:

1) to give a brief introduction to the particle physics,
2) to introduce the weakton model of elementary particles,
3) to address the mechanism of subatomic decays,
4) to introduce color algebra, as the mathematical basis for the color quantum
number, and
5) to derive the structure of mediator clouds around subatomic particles.
Searching for the main constituents of matter has a long history going back to an-
cient Greeks, to Robert Boyle (1600s), John Dalton (early 1800s), J.J. Thomson and Ernest
Rutherford (end of the 19th century), and Niels Bohr (1913).
In the current standard model of particle physics, all forms of matter are made up of 6
leptons and 6 quarks, and their antiparticles, which are treated as elementary particles. The
forces are mediated by the mediators, including the photon mediating the electromagnetism,
the vector bosons W ± , Z and the Higgs H 0 mediating the weak interaction, the eight gluons
mediating the strong interaction, and the hypothetical graviton mediating gravity.
However, there are many challenging problems related to subatomic particles. One such
problem is that why leptons do not participate in strong interactions. In fact, the most dif-
ficult challenge is associated with the puzzling decay and reaction behavior of subatomic
particles. For example, the electron radiations and the electron-positron annihilation into
photons or quark-antiquark pair clearly shows that there must be interior structure of elec-
trons, and the constituents of an electron contribute to the making of photon or the quark in
the hadrons formed in the process. In fact, all sub-atomic decays and reactions show clearly
the following conclusion:

There must be interior structure of charged leptons, quarks and mediators.

This conclusion motivates us to propose a weakton model for sub-lepton, sub-quark, and
sub-mediators (Ma and Wang, 2015b), based on the new field theory and the new insights
for the weak and strong interactions presented in the last chapter.
One important theoretical basic for the weakton model is the field theory developed in
the last chapter. In particular, the weak and strong charges are responsible for the weak
264 Chapter 5 Elementary Particles

and strong interactions. The confinement property demonstrated by the weak and strong
interaction potentials give rise to the needed confinement for the weakton constituents of
the composite particles.
Remarkably, the weakton model offers a perfect explanation for all sub-atomic decays.
In particular, all decays are achieved by 1) exchanging weaktons and consequently exchang-
ing newly formed quarks, producing new composite particles, and 2) separating the new
composite particles by weak and/or strong forces.
One aspect of this decay mechanism is that now we know the precise constituents of
particles involved in all decays/reactions both before and after the reaction. It is therefore
believed that the new decay mechanism provides clear new insights for both experimental
and theoretical studies.
This chapter is organized as follows. Sections 5.1 and 5.2 recall the basic knowledge of
particle physics and the quark model. Sections 5.3 and 5.4 are based entirely on (Ma and
Wang, 2015b), with the weakton model introduced in Section 5.3, and with the mechanism
of subatomic decays and electron radiations presented in Section 5.4. The last section,
Section 5.5 studies the color algebra associated with the color quantum number, leading
to detailed structure on mediator clouds around subatomic particles. Section 5.5 is based
entirely on (Ma and Wang, 2014b).

5.1 Basic Knowledge of Particle Physics

5.1.1 Classification of particles
In particle physics, subatomic particles are classified into two basic classes, bosons and
fermions:
bosons = integral spin particles,
fermions = fractional spin particles.
Also, based on their properties and levels, all subatomic particles are currently classified
into four types:
leptons, quarks, mediators, hadrons.
In the following, we recall basic definitions and the quantum characterizations of these
particles; see among many others (Kane, 1987; Griffiths, 2008; Halzen and Martin, 1984).
1. Leptons. Leptons are fermions which do not participate in the strong interaction, and
consist of three generations:
! " ! " ! "
e µ, τ
, , ,
νe νµ ντ

where e, µ , τ are the electron, the muon, the tau, and νe , νµ , ντ are the e-neutrino, the µ -
neutrino, the τ -neutrino. Together with their antiparticles, there are total 12 leptons:
5.1 Basic Knowledge of Particle Physics 265

particles: (e− , νe ), (µ − , νµ ), (τ − , ντ ),
antiparticles: (e+ , ν e ), (µ + , ν µ ), (τ + , ν τ ).

2. Quarks. Based on the standard model, there are three generations of quarks consisting
of 6 particles and 6 antiparticles, which participate all four fundamental interactions:
quarks: (u, d), (c, s), (t, b),
antiquarks: (u, d), (c, s), (t, b),

where u, d, c, s,t, b are the up quark, down quark, charm quark, strange quark, top quark,
and bottom quark.
The quark model asserts that three quarks are bounded together to form a baryon, and
a pair of quark and antiquark are bounded to form a meson. As mentioned in Section
4.5.3, quarks are confined in hadrons, and no free quarks have been found in Nature. This
phenomena is called the quark confinement, which has been very well explained by the
layered formulas of strong interactions in the last chapter.
3. Mediators. The standard model shows that each interaction is associated with a class
of field particles, called mediators. Hence, there are four classes of mediators:

Gravitation: graviton G,
Electromagnetism: photon γ ,
Weak interaction: vector and Higgs bosons W ± , Z, H ± , H 0 ,
Strong interaction: gluons gk (1 ! k ! 8).

By the unified field theory introduced in the last chapter, there exists a natural duality
between the interacting fields {g µν , Aµ ,Wµa , Skµ } and their dual fields {φ µg , φ e , φwa , φsk }:
gµν ↔ φµg ,
Aµ ↔ φ e ,
Wµa ↔ φwa , 1 ! a ! 3,
Skµ ↔ φsk , 1 ! k ! 8.

This duality leads to four classes of new dual particle fields, called the dual mediators:
tensor graviton G ↔ vector graviton g,
vector photon γ ↔ scalar photon γ0 ,
±
charged vector bosons W ↔ charged Higgs bosons H ± , (5.1.1)
0
neutral vector boson Z ↔ neutral Higgs boson H ,
k
vector gluons g ↔ scalar gluons gk0 , 1 ! k ! 8,

where the upper index k in gk and gk0 represents the color index of gluons.
266 Chapter 5 Elementary Particles

The mediator duality (5.1.1) is derived from the unified field theory. However, the dual-
ity (5.1.1) can also be naturally deduced from the weakton model postulated in Section 5.3.
Namely, the mediator duality (5.1.1) is derived from two very different perspectives in two
different theories. This can hardly be a coincident.
4. Hadrons. Hadrons are classified into two types: baryons and mesons. Baryons are
fermions and mesons are bosons, which are all made up of quarks:

Baryons = qi q j qk , mesons = qi q j ,

where qk = {u, d, c, s,t, b}. The quark constituents of main hadrons are listed as follows:
1
• Baryons with J = :
2
√
p (uud), n (udd), Λ (s(du − ud)/ 2),
√
Σ+ (uus), Σ− (dds), Σ0 (s(du + ud)/ 2),
Ξ0 (uss), Ξ− (dss).

3
• Baryons with J = :
2

∆++ (uuu), ∆+ (uud), ∆− (ddd), ∆0 (udd),

Σ∗+ (uus), Σ∗− (dds), Σ0 (uds),
Ξ∗0 (uss), Ξ∗− (dss), Ω− (sss).

• Mesons with J = 0:
√
π + (ud), π − (ud), π 0 ((uu − dd)/ 2),
K + (us), K − (us), K 0 (ds),
0 √
K (ds), η ((uu + dd − 2ss)/ 6).

• Mesons with J = 1:
√
ρ + (ud), ρ − (ud), ρ 0 ((uu − dd)/ 2),
K ∗+ (us), K ∗− (us), K ∗0 (ds),
∗0
K (ds), ψ (cc), γ (bb).

5.1.2 Quantum numbers

It is easy to identify a macro-body, but not easy to identify a sub-atomic particle. Usually
experimental physicists distinguish these sub-atomic particles by measuring their physical
5.1 Basic Knowledge of Particle Physics 267

parameters. There are numerous types of particles, which obey various conservation laws.
Each conservation law is characterized by one parameter. These physical parameters are
called quantum numbers.
The main quantum numbers of particles are:
mass m, electric charge Q, lifetime τ ,
spin J, lepton number L, baryon number B,
(5.1.2)
parity π , isospin (I, I3 ), strange number S,
G-parity, hypercharge Y.
In addition, there are some new quantum numbers, including the weak charge Q w , strong
charge Qs , and colour index k, which play an important role in in the weakton model intro-
duced in Section 5.3.
We now briefly introduce the quantum numbers listed in (5.1.2).
1. Mass. This is a quantity to characterize the inertia, and also plays the role of gravita-
tional charge. In the macro-world, masses are continuous in distribution, however in quan-
tum world masses are discrete. The same particles have the same masses, but the different
massive particles have a gap between their masses.
In non-quantum physics, the masses are additive, i.e. if a body A consists of two bodies
B and C, then the mass mA of A is the sum of masses mB and mC of B and C:

A = B + C ⇒ mA = mB + mC . (5.1.3)

But, the additive relation (5.1.3) is not valid in a quantum system.

Mass is a most important quantum number, which plays a role of an identity card for all
massive particles.
2. Electric charge Q. This is an important quantum number. It is the source of electro-
magnetic force. In the four interaction charges: m, e, g w , gs , the electric charge e is unique
one possessing positive and negative values. The electric charges are discrete, they appear
only at an integral multiples of the electron charge e:1

Q = ±ne (n = 0, 1, 2, · · · ).

Electric charge is an additive conservation quantity. For any particle reaction:

A1 + · · · + A n → B1 + · · · + B N ,

we have
n N
∑ QA k = ∑ QB k ,
k=1 k=1
1 In the weakton model of elementary particles, w∗ carries 2/3 electric charge, and w and w carry −1/3
1 2
and −2/3 electric charges respectively.
268 Chapter 5 Elementary Particles

where QAk and QBk are the electric charges of particles Ak and Bk
3. Lifetime of particles. Due to the decay property for most particles, the lifetime
becomes a quantum number. Except the long lifetime particles: the electron e, proton p,
neutrino v, photon γ , gluon gk , graviton G, all other particles undergo decays.
Usual particles with decay have very short lifetime, and in general their lifetimes τ do
not exceed 10−5s:
τ < 10−5s on an average.
The neutron n is special, and has a longer lifetime:

τ = 885.7s on an average.

4. Spin J. The spin is an intrinsic property of particles. Although it exhibits angu-

lar momentum characteristics, the spin does not represent particle rotation around its axis.
However, for a better understanding, we may imaginarily illustrate the concept of spin in
Figure 5.1. Also, the spin of a right-handed particle is denoted by ↑, and a left-handed
particle by ↓.

Figure 5.1 (a) right-hand spin(J >0), and (b) left-hand spin(J < 0).

In quantum mechanics, the spin operator ⃗S is a 3-dimensional vector operator, defined

as
⃗S = sh̄(τ1 , τ2 , τ3 ), s the spin value,
! "
σk 0 (5.1.4)
τk = for 1 ! k ! 3,
0 σk
and σk are the Pauli matrices. The spin value s of the spin operator (5.1.4) represents the
right-hand and left-hand spins. For example, if ±s (s > 0) are the eigenvalues of third
component S3 ,
S3 ψR = sψR , S3 ψL = −sψL ,
then the eigenfunctions ψR and ψL are the wave functions of the right-handed and left-
handed particles with spin J = s.
5. Lepton numbers L. There are 3 types of lepton numbers: Le , Lµ , Lτ , which are special
parameters to characterize the three generations of leptons (e, νe ), (µ , νµ ), (τ , νµ ). They
5.1 Basic Knowledge of Particle Physics 269

only take three values L = ±1, 0. For all non-lepton particles, their lepton number is zero,
and
⎧
⎪ 1 for e− , νe ,
⎨
Le = − 1 for e+ , ν e ,
⎪
⎩
0 others,
⎧
⎪ 1 for µ − , νµ ,
⎨
Lµ = − 1 for µ + , ν µ ,
⎪
⎩
0 others,
⎧
⎪ 1 for τ − , ντ ,
⎨
Lτ = − 1 for τ + , ν τ ,
⎪
⎩
0 others.

The lepton numbers are additive and conservative quantities.

6. Baryon number B. The hadrons are classified into baryons and mesons. Baryons are
fermions and mesons are bosons. Also, baryons have an additive and conservative quantum
number: the baryon number B, defined by
⎧
⎪ 1 for a baryon,
⎨
B = −1 for an antibaryon,
⎪
⎩
0 for all other particles.

7. Parity π . Due to the Noether Theorem 2.38, each symmetry is associated with a
conservation law. The parity is a conservative quantum number corresponding to the spatial
reflection symmetry:
x → −x (x ∈ R3 ).

Parity π only takes two values:

π = ±1.

The weak interaction violates parity; see (Lee and Yang, 1956; Wu, Ambler, Hayward,
Hoppes and Hudson, 1957). The parities π of most particles are obtained by two methods:
the experimental and the artificial means.
The parity is a multiplication quantum number. For example, for an N particles system:

A = A1 + · · · + AN ,

its tatol parity πA is given by

πA = (−1)l πA1 · · · πAN ,
270 Chapter 5 Elementary Particles

where l is the sum of orbital quantum numbers of all particles, and π Ak is the parity of Ak
particle.
The parity conservation is defined as follows. For a particle reaction:

A1 + · · · + AN −→ B1 + · · · + BK , (5.1.5)

the total parities of both sides of (5.1.5) are the same:

(−1)l1 πA1 · · · πAN = (−1)l2 πB1 · · · πBK .

In Subsection 5.1.4, we shall introduce the violation of parity conservation in the weak
interaction in detail.
8. Isospin (I, I3 ). The isospin was first presented in 1932 by Heisenberg, and was used to
describe that the strong interaction between protons and neutrons, independently of electric
charges. Later, along with the development of particle physics, it was discovered that the
isospin is a good quantum number for all hadrons.
Isospin has two components (I, I3 ), where I is called the isospin and I3 is the third
component of isospin. In a many-particle system, the isospin I obeys the vectorial additive
rule and I3 obeys the usual additive rule. We shall introduce the isospin again in Subsection
5.1.4.
9. Strange number S. In 1947, C. D. Rochester and C. C. Butler first discovered the
strange particle K 0 , a neutral meson, from cosmic rays, and later many strange particles
such as K ± , Λ, Σ± , Ξ− , Ξ0 , Ω− were found. They are named the ”strange” because these
particles possess a kind quantum number called the strange number.
The strange number S is an additive quantum number, which is conserved only in the
electromagnetic and strong interactions, and takes the integral values:

S = 0, 1, 2, 3, · · · .

10. Hypercharge Y . Hypercharge Y is another quantum number for hadrons, defined by

Y = S + B,

where S and B are the strange number and the baryon number.
11. G-parity. G parity is a conservative quantity only for strong interactions under the
G-transformation
Gψ = Ĉeiπ I2 ψ ,
where Ĉ is the transformation of particles and antiparticles:

ĈA = A, ĈA = A, A the antiparticle of A,

and I2 is the second axis of the isospin space.

5.1 Basic Knowledge of Particle Physics 271

5.1.3 Particle transitions

Particle transition is the main dynamic form of sub-atomic particles, which includes particle
decays, scatterings and various radiations. It is called ”transition” because in these processes
particles will undergo a transition from an energy level to other energy levels.
The particle transition is a crucial way for us to understand the particle structures and
properties. In particular, these transition processes can reveal the mysteries of the weak and
the strong interactions.
1. Decays. Particle decay means that a particle is spontaneously decomposed into sev-
eral other particles. The most remarkable example is the β -decay, i.e., a neutron n decays
to a proton p, an electron e− and an anti-neutrino ν e :

n → p + e− + ν e . (5.1.6)

Current list of discovered particle consists of hundreds of numbers. Except electrons

and protons, all massive particles will decay therefore they are finite lifetime. The massless
particles, such as the photon, the gluons and the neutrinos, do not decay, and have infinite
lifetime.
All decays obey the following laws and rules:

1) Decays are caused by the three fundamental forces: the electromagnetic, weak, strong
interactions, and, therefore, are mainly classified into two types: weak interaction
decays and strong interaction decays;

2) Decays always take place from higher masses toward to lower masses. For example
in the β -decay (5.1.6), the masses of particles on the right-hand side are smaller than
the mass of the neutron n.

3) Both strong and weak interaction decays have to obey some basic conservation laws,
as shown in the next subsection.
Remark 5.1 The following transition

p → n + e+ + νe (5.1.7)

is often called β -decay as well. In fact, in Nature the process (5.1.7) cannot spontaneously
take place, and it always occurs under certain energetic excitation. Hence, (5.1.7) is an
excited scattering.

2. Scattering. If the decay is a spontaneous behavior, then scattering is a forced behavior

of particles under certain force actions, such as collisions and energetic excitations. The
transition (5.1.7) is a scattering, and the precise decay mechanism is written as

p + γ → n + e+ + νe .
272 Chapter 5 Elementary Particles

In general, a scattering reaction is written as

A1 + · · · + A N → B1 + · · · + B K , (5.1.8)

where An (1 ! n ! N) are the initial particles, and Bk (1 ! k ! K) are the final particles.
There are two forms of scatterings: the elastic and non-elastic scatterings, both of which
are caused by the three interactions: the electromagnetic, the weak and the strong interac-
tions. Collision is the most important experimental method to detect new particles and new
phenomena. The scatterings (5.1.8) satisfy various conservation relations.
Here are a few important scatterings in the history of quantum physics:

Compton scattering: γ + e − → γ + e− ,
Pair annihilation: e + + e− → γ + γ ,
Pair creation: γ + γ → e + + e− ,
Moller scattering: e − + e− → e− + e− ,
Bhahba scattering: e − + e+ → e− + e+ .
Deep inelastic scattering : e − + p → e− + p + π 0 .

3. Radiations. Radiations include the electromagnetic radiation and the gluon radiation.
The first one is that electrons emit photons:

e− → e − + γ , (5.1.9)

and the second one is that quarks emit gluons

q → q + gk , (5.1.10)

Electromagnetic radiations are caused in two scenaries:

1) As an electron changes its velocity, it emits photons. This radiation is called the
bremsstrahlung. The radiation energy in unit time can be expressed by the formula:

2 e2 2
W= a with a being the average acceleration.
3 c3

2) As an electron at a higher energy level E1 undergoes a transition to a lower energy

level E0 < E1 , it emits photons. The energy ε of the emitting photons equals to the
difference of energy levels:
ε = E1 − E0 .

In summary, there are three types of particle transitions: decays, scatterings, and ra-
diations. They are the main dynamic behavior for micro-particles, and reveal the interior
5.1 Basic Knowledge of Particle Physics 273

structure of particles, and provide crucial information about the three interactions associated
with the electromagnetic, the weak, and the strong forces.
Here we list some principal decay forms:

• Lepton decays:
µ − → e− + ν e + ν µ ,
µ + → e+ + νe + ν µ ,
τ − → e− + ν e + ντ ,
τ − → µ − + ν µ + ντ , (5.1.11)
τ → π + ντ ,
− −

τ − → ρ − + ντ ,
τ − → K − + ντ .

• Quark decays:
d → u + e− + ν e ,
s → u + e− + ν e ,
(5.1.12)
s → d + γ,
c → d + s + u.

• Mediator decays:
W + → e+ + νe , µ + + νµ , τ + + ντ ,
W − → e− + ν e , µ− + νµ, τ− + ντ , (5.1.13)
Z → e+ + e− , µ + + µ −, τ + + τ − .

• Baryon decays:
n → p + e− + ν e ,
Λ → p + π −, n + π 0,
Σ+ → p + π 0, n + π +,
Σ− → n + π −,
(5.1.14)
Σ0 → Λ + γ ,
Ξ− → Λ + π − ,
∆++ → p + π +,
∆+ → p + π 0 ,
274 Chapter 5 Elementary Particles

∆− → n + π − ,
∆0 → n + π 0 ,
Σ∗± → Σ± + π 0,
Σ∗0 → Σ0 + π 0,
Ξ∗− → Ξ− + π 0,
Ξ∗0 → Ξ0 + π 0.

• Meson decays:

π + → µ + + νµ ,
π− → µ− + νµ ,
π 0 → 2γ ,
K + → µ + + νµ , π + + π 0, π + + π + + π −,
K− → µ − + ν µ , π − + π 0, π − + π − + π +,
K 0 → π + + e− + ν e , π + + π −, π + + π − + π 0,
(5.1.15)
η → 2γ , π + + π − + π 0,
ρ ± → π ± + π 0,
ρ 0 → π + + π −,
K ∗± → K ± + π 0,
K ∗0 → K 0 + π 0,
ω → π0 + γ, π + + π − + π 0.

5.1.4 Conservation laws

All particle transitions can be expressed in the form:

A1 + · · · + A N → B1 + · · · + B K . (5.1.16)

They are driven by the three fundamental forces: the electromagnetic, the weak, and the
strong interactions. Particle transitions have to obey certain basic conservation laws in
the sense that for certain conservative quantum number q, its total values on both sides
of (5.1.16) are the same:
qA = qB . (5.1.17)
In addition, some conservation laws may not be valid for all interactions. For example,
parity is valid only in the strong and the electromagnetic interactions, and is violated in the
weak interaction.
5.1 Basic Knowledge of Particle Physics 275

We now introduce the basic conservation laws.

1. Energy conservation. Energy is an additive quantum number. The energy of a particle
is given by the formula '
ε = c P 2 + m2 c2 ,
where P is the momentum, and m the mass. Energy conservation is a law that all transitions
must obey. Namely, for (5.1.16),
N K
∑ εAn = ∑ εBk ,
n=1 k=1

where εAn and εBk are the energies of particles An and Bk .

2. Momentum conservation. Momentum is vectorial additive. For two particles A 1 and
A2 , their total momentum P = (P1 , P2 , P3 ) equals to

Pk = PkA1 + PkA2 , 1 ! k ! 3,

where PA1 and PA2 are the momentums of A1 and A2 .

Momentum conservation is a universal conservation law, and for (5.1.16) the following
equality holds true:
N K
∑ PjAn = ∑ Pj k ,
B
1 ! j ! 3.
n=1 k=1

3. Angular momentum conservation. This quantum number is vectorial additive. The

angular momentum J⃗ of a particle consists of the orbital angular momentum L and spin ⃗S:

J⃗ = ⃗L + ⃗S, ⃗L =⃗r × P, P is the momentum.

Angular momentum conservation law is also a universal conservation law, and for (5.1.16),
N K
∑ J⃗An = ∑ J⃗Bk .
n=1 k=1

4. Other universal conservation laws. In particle physics, the following quantum num-
bers are additive and conservative in all interactions: the electric charge Q e , the lepton
numbers Le , Lµ , Lτ , and the baryon number B.
Table 5.1 lists the conservation or non-conservation properties of quantum numbers for
the three interactions.

Remark 5.2 CP combines charge conjugation C and parity P, and CPT combines CP
and time reversal T .
276 Chapter 5 Elementary Particles

Table 5.1 Conservation Laws

Conservative Quantities Strong Electromagnetic Weak
Energy E Yes Yes Yes
Momentum P Yes Yes Yes
Angular Momentum J Yes Yes Yes
Electric charge Qe Yes Yes Yes
Lepton Numbers Le ,Lµ ,Lτ Yes Yes Yes
Baryon Number B Yes Yes Yes
Strange Number S Yes Yes No
Parity π Yes Yes No
Isospin I Yes No No
I3 Yes Yes No
G Parity G Yes No No
Charge Conjugation C Yes Yes No
Time Reversal T Yes Yes No
CP Yes Yes No
CPT Yes Yes Yes

Combined CPT conservation was proved by G. Lüders and W. Pauli independently

in 1954, called CPT theorem. The CPT conservation is an important result in quantum
physics. Based on CPT theorem, we can deduce that particles and antiparticles have the
same masses and lifetimes, and their magnetic moments are reversal with the same magni-
tudes.

5.1.5 Basic data of particles

Here, we list the basic data for leptons, quarks, and hadrons in Tables 5.2-5.5. The units are:
mass in MeV/c2 , lifetime in seconds, and electric charge in the unit of proton charge.
Note that the quantum number Qw in Table 5.2 represents the weak charge number. Here
the values of Qw are based on the weakton model in Section 5.3.

! "
1
Table 5.2 Leptons spin J =
2
Leptons m Qe Qw Le Lµ Lτ τ
e− 0.51 −1 3 1 0 0 ∞
e+ 0.51 1 3 −1 0 0
µ− 105.7 −1 3 0 1 0 2.2 × 10−6
µ+ 105.7 1 3 0 −1 0
τ− 1777 −1 3 0 0 1 2.9 × 10−13
τ+ 1777 1 3 0 0 −1
ν+ 0 0 1 1 0 0 ∞
νe 0 0 1 −1 0 0 ∞
νµ 0 0 1 0 1 0 ∞
νµ 0 0 1 0 −1 0 ∞
ντ 0 0 1 0 0 1 ∞
ντ 0 0 1 0 0 −1 ∞
5.2 Quark Model 277

1
Also, all neutrinos possess left-hand spin with J = − , and antineutrinos possess right-
2
1
hand spin with J = . Namely, neutrinos move at the speed of light.
2
In addition, the data of W ± , Z, H 0 are given as follows.

W ± : m = 80.4 GeV/c2 , Qe = ±1, τ = 3.11 × 10−25,

Z : m = 91.2 GeV/c2 , Qe = 0, τ = 2.67 × 10−25,
H 0 : m = 126 GeV/c2 , Qe = 0, τ = 10−21.
! "
1
Table 5.3 Quarks spin J =
2
Quarks m Qe Qs Qw B I I3 S Y
d 7 −1/3 1 3 1/3 1/2 −1/2 0 1/3
u 3 2/3 1 3 1/3 1/2 1/2 0 1/3
s 120 −1/3 1 3 1/3 0 0 −1 −2/3
c 1200 2/3 1 3 1/3 1/2 1/2 0 1/3
b 4300 −1/3 1 3 1/3 1/2 −1/2 0 1/3
t 174000 2/3 1 3 1/3 0 0 1 4/3

5.2 Quark Model

5.2.1 Eightfold way
We see from Tables 5.4 and 5.5 that there are many hadrons, and in fact their number arrives
at hundreds. The abundance of hadrons implies that there must be some rules to classify

Table 5.4 Baryons

Baryons J m Qe Qw Qs B I I3 S Y τ
p 1/2 938.3 1 9 3 1 1/2 1/2 0 1 ∞
n 1/2 939.6 0 9 3 1 1/2 −1/2 0 1 885.7
Λ 1/2 1115.7 0 9 3 1 0 0 −1 0 2.7 × 10−10
Σ+ 1/2 1189.4 1 9 3 1 1 1 −1 0 8 × 10−11
Σ− 1/2 1197.5 −1 9 3 1 1 0 −1 0 1.5 × 10−10
Σ0 1/2 1192.6 0 9 3 1 1 −1 −1 0 7.4 × 10−20
Ξ− 1/2 1321.3 −1 9 3 1 1/2 1/2 −2 −1 1.6 × 10−10
Ξ0 1/2 1314.8 0 9 3 1 1/2 −1/2 −2 −1 2.9 × 10−10
∆++ 3/2 1230 2 9 3 1 3/2 3/2 0 1 5.6 × 10−24
∆+ 3/2 1231 1 9 3 1 3/2 1/2 0 1 5.6 × 10−24
∆− 3/2 1234 −1 9 3 1 3/2 −1/2 0 1 5.6 × 10−24
∆0 3/2 1232 0 9 3 1 3/2 −3/2 0 1 5.6 × 10−24
Σ∗+ 3/2 1383 1 9 3 1 1 1 −1 0 1.8 × 10−23
Σ∗− 3/2 1387 −1 9 3 1 1 0 −1 0 1.8 × 10−23
Σ∗0 3/2 1384 0 9 3 1 1 −1 −1 0 1.8 × 10−23
Ξ∗− 3/2 1535 −1 9 3 1 1/2 1/2 −2 −1 6.9 × 10−23
Ξ∗0 3/2 1532 0 9 3 1 1/2 −1/2 −2 −1 6.9 × 10−23
Ω− 3/2 1672 −1 9 3 1 0 0 −3 −2 8.2 × 10−11
278 Chapter 5 Elementary Particles

them into groups based on certain common properties. The Eightfold Way can be regarded
as a successful classification for hadrons. In fact, the Eightfold Way, together with the
irreducible representations of SU(3), has played a navigation role for the introduction of the
quark model.

Table 5.5 Mesons

Mesons J m Qe Qw Qs B I I3 S Y τ
π0 0 135 0 6 2 0 1 0 0 0 8.4 × 10−17
π+ 0 139.6 1 6 2 0 1 1 0 0 2.6 × 10−8
π− 0 139.6 −1 6 2 0 1 −1 0 0 2.6 × 10−8
K+ 0 493.7 1 6 2 0 1/2 1/2 1 1 1.2 × 10−8
K− 0 493.7 −1 6 2 0 1/2 −1/2 1 1 1.2 × 10−8
K0 0 497.7 0 6 2 0 1/2 −1/2 1 1 5.1 × 10−8
η 0 547.5 0 6 2 0 0 0 0 0 5.1 × 10−19
η′ 0 957.8 0 6 2 0 0 0 0 0 3.2 × 10−21
D0 0 1864.5 0 6 2 0 1/2 −1/2 0 0 4.1 × 10−13
D+ 0 1869.3 1 6 2 0 1/2 1/2 0 0 10−12
D− 0 1869.3 −1 6 2 0 1/2 −1/2 0 0 10−12
B+ 0 5279 1 6 2 0 1/2 1/2 0 0 1.6 × 10−12
B− 0 5279 −1 6 2 0 1/2 −1/2 0 0 1.6 × 10−12
B0 0 5279.4 0 6 2 0 1/2 −1/2 0 0 1.5 × 10−12
ρ+ 1 775.5 1 6 2 0 1 1 0 0 4 × 10−24
ρ− 1 775.5 −1 6 2 0 1 −1 0 0 4 × 10−24
ρ0 1 775.5 0 6 2 0 1 0 0 0 4 × 10−24
K ∗+ 1 894 1 6 2 0 1/2 1/2 1 1 10−23
K ∗− 1 894 −1 6 2 0 1/2 −1/2 1 1 10−23
K ∗0 1 894 0 6 2 0 1/2 −1/2 1 1 10−23
ω 1 782.6 0 6 2 0 0 0 0 0 8 × 10−23
ψ 1 3097 0 6 2 0 0 0 0 0 7 × 10−21
D∗+ 1 2008 1 6 2 0 1/2 1/2 0 0 3 × 10−21
D∗− 1 2008 −1 6 2 0 1/2 −1/2 0 0 3 × 10−21
D∗0 1 2008 0 6 2 0 1/2 −1/2 0 0 3 × 10−21

The Eightfold Way was introduced by Gell-Mann and Ne’eman independently in 1961.
This scheme arranges the baryons and mesons into certain geometric patterns, according
to their charge, strangeness, hypercharge and isospin I3 . These geometric patterns include
hexagons and triangles, and all particles put in each diagram are considered as a class.
The first group of hadrons consists of the eight lightest baryons:

n, p, Σ− , Σ0 , Σ+ , Λ, Ξ− , Ξ0 , (5.2.1)

which fit into a hexagonal array with two particles Σ0 and Λ at the center; see Figure 5.2.
The eight baryons (5.2.1) with J = 1/2 are known as the baryon octet. Note that particles
in Figure 5.2 range with charges lie along the downward-sloping diagonal line: Q = 1 for p
and Σ+ , Q = 0 for n, Λ, Σ0 and Ξ0 , and Q = −1 for Σ− and Ξ− . Horizontal lines associate
particles of the same strangeness S = 0 and hyper-charge Y = 1 for the neutron and proton,
S = −1 and Y = 0 for Σ− , Σ0 , Λ and Σ+ , S = −2 and Y = −1 for Ξ− and Ξ0 . The third
component I3 of isospin indicates particles in column.
5.2 Quark Model 279

Figure 5.2 Baryon octet

The second group of hadrons consists of the eight lightest mesons:

0
π +, π 0, π − , K + , K 0 , K − , K , η . (5.2.2)

In the same fashion as the baryon octet (5.2.1), this group (5.2.2) fits into a hexagonal array
with π 0 and η at the center; see Figure 5.3.

Figure 5.3
280 Chapter 5 Elementary Particles

The third group of hadrons consists of ten baryons:

∆++ , ∆+ , ∆− , ∆0 , Σ∗+ , Σ∗− , Σ∗0 , Ξ∗− , Ξ∗0 , Ω− , (5.2.3)

which are fited into a triangular array as in Figure 5.4.

Figure 5.4

The fourth group of hadrons is given by the eight messons:

∗0
ρ + , ρ − , ρ 0 , K ∗+ , K ∗− , K ∗0 , K , ω , (5.2.4)

which are arranged as in Figure 5.5.

Figure 5.5
5.2 Quark Model 281

Remark 5.3 The Eightfold Way based on the patterns given by Figures 5.2-5.5 pro-
vides a beautiful classification for hadrons. However, the most important point is that the
Eightfold way and the irreducible representation of SU(N) also provide key clues to dis-
cover the quark model for the hadron structure.

5.2.2 Irreducible representations of SU(N)

To better understand the process from the Eightfold way to the quark model, it is necessary
to know the irreducible representation of SU(N) and its connection with particle physics.
We proceed in a few steps as follows.
1. Irreducible representation of Lie groups. We begin with the definition of represen-
tation of abstract groups. Let G and M be two groups, and M consist of N-th order real or
complex matrices satisfying certain properties. Let H be a mapping from G to M:

H : G → M, (5.2.5)

preserving the multiplication:

H(g1 · g2 ) = H(g1 )H(g2 ), ∀g1 , g2 ∈ G. (5.2.6)

Then, the image H(G) ⊂ M of group G is called an N-dimensional representation of G.

Because M is a group of matrices, for any g ∈ G, H(g) ∈ M is an N-th order matrix,
written as ⎛ ⎞
a11 · · · a1N
H(g) = ⎝ ... .. ⎟ .
⎜
. ⎠
aN1 ··· aNN
If there exists a matrix A (not necessary in M) such that A−1 H(g)A ∈ M for all g ∈ G, and
⎛ ⎞
H1 (g) 0
AH(g)A−1 = ⎝
⎜ .. ⎟
⎠, ∀g ∈ G, (5.2.7)
.
0 Hn (g)
n
where Hk (g) (1 ! k ! n) are mk -th order matrices with ∑ mk = N, then the representation
k=1
H(G) is reducible, and each block matrix Hk (g) in (5.2.7) is a smaller representation of G.
In mathematics, (5.2.7) can be equivalently expressed as

H(G) = H1 (G) ⊕ · · · ⊕ Hn (G). (5.2.8)

If all block matrices Hk (G) (1 ! k ! n) in (5.2.8) cannot be split into smaller pieces any-
more, then the direct sum of all sub-representations Hk (G) as

H1 (G) ⊕ · · · ⊕ Hn (G)
282 Chapter 5 Elementary Particles

is called an irreducible representation of G, which can be simply written in the following

form
H(G) = m1 ⊕ · · · ⊕ mn , (5.2.9)
where mk is the order of Hk (G).
2. Fundamental representation SU(N). Let G be a linear transformation group made up
of all linear norm-preserving mappings of CN :

g : CN → CN . (5.2.10)

It is known that for each linear operator g ∈ G as defined in (5.2.10), there is a unique matrix
U ∈ SU(N) such that
g(ψ ) = U ψ , ∀ψ ∈ CN . (5.2.11)
Hence, relation (5.2.11) provides a correspondence

g -→ U for g ∈ G and U ∈ SU(N),

which is a one to one and onto mapping

H : G → SU(N), (5.2.12)

and satisfies the multiplication relation (5.2.6).

Usually, the representation given by (5.2.12):

SU(N) = H(G)

is called an N-dimensional fundamental representation of linear norm-preserving transfor-

mation group G, which for simplicity is denoted by SU(N).
3. Conjugate representation SU(N). The conjugate group SU(N) of SU(N) is called
the conjugate representation, expressed as

SU(N) = {U | U ∈ SU(N)}, (5.2.13)

where U is the complex conjugate of U.

If SU(N) and SU(N) are regarded as linear norm-preserving transformation groups of
N-dimensional complex space CN , then they represent such linear operators as follows. Let

{e1 , · · · , eN } ⊂ CN

constitute a complex orthogonal basis of CN , i.e.

. / 0
N /
C = ∑ ck ek //ck ∈ C .
N
k=1
5.2 Quark Model 283

N
Then the conjugate space C of CN can be written as
. / 0
N /
N
C = ∑ βk ek //βk ∈ C .
k=1

Thus, each matrix U ∈ SU(N) is a linear transformation:

U : CN → CN , (5.2.14)

and each U ∈ SU(N) gives

∗ N N
U : C →C . (5.2.15)
Remark 5.4 In particle physics, a complex orthogonal basis {e 1 , · · · , eN } of CN
stands
for N different particles, and its conjugate basis {e1 , · · · , eN } stands for the N antiparticles.
Hence we have
CN = the space of all states of particles e1 , · · · , eN ,
N (5.2.16)
C = the space of all states of antiparticles e1 , · · · , eN .
Thus, the mappings U ∈ SU(N) in (5.2.14) stand for state transformations of particles
e1 , · · · , eN , and U ∗ ∈ SU(N) for state transformations of antiparticles e1 , · · · , eN .

4. Tensor product of matrices. In quantum physics we often see tensor products of

matrices. Here we give their definition. Let A, B be two matrices given by
⎛ ⎞ ⎛ ⎞
a11 · · · a1n b11 · · · b1m
A = ⎝ ... .. ⎟ , B = ⎜ .. .. ⎟ .
⎜
. ⎠ ⎝ . . ⎠
an1 ··· ann bm1 ··· bmm
Then the tensor product A ⊗ B is defined by
⎛ ⎞
a11 B ··· a1n B
⎜ .. .. ⎟ ,
A⊗B = ⎝ . . ⎠ (5.2.17)
an1 B ··· ann B
where ai j B are the block matrices
⎛ ⎞
ai j b11 ··· ai j b1m
⎜
ai j B = ⎝ .. .. ⎟
. . ⎠.
ai j bm1 ··· ai j bmm
Hence A ⊗ B is an (n × m)-th order matrix.
5. Irreducible representation of SU(N). In the quark model, we shall meet the notations:
meson = 3 ⊗ 3,
(5.2.18)
baryon = 3 ⊗ 3 ⊗ 3,
284 Chapter 5 Elementary Particles

representing the tensor products:

meson = SU(3) ⊗ SU(3),

(5.2.19)
baryon = SU(3) ⊗ SU(3) ⊗ SU(3).
To understand the implications of (5.2.18) and (5.2.19), we have to know the irreducible
representation of the tensor products

SU(N) ⊗ · · · ⊗ SU(N) ⊗ SU(N) ⊗ · · · ⊗ SU(N), (5.2.20)

1 23 4 1 23 4
k1 k2

and its physical significance, which will be discussed in more detail in the next two subsec-
tions. Here we just give a brief introduction to the irreducible representation of (5.2.20).
A representation of (5.2.20) is a mapping defined as

H : SU(N) → SU(N) ⊗ · · · ⊗ SU(N) ⊗ SU(N) ⊗ · · · ⊗ SU(N),

1 23 4 1 23 4
k1 k2
(5.2.21)
H(U) = U · · ⊗ U4 ⊗U
1 ⊗ ·23 · · ⊗ U4
1 ⊗ ·23 for U ∈ SU(N).
k1 k2

We can show that mapping (5.2.21) satisfies (5.2.6), i.e.

H(U1 ·U2 ) = H(U1 )H(U2 ), ∀U1 ,U2 ∈ SU(N).

By Definition (5.2.17) for tensor products of matrices, each representation H(U) in

(5.2.21) is an N k -th order matrix, and is also a linear transformation of the complex space
as
N N
H(U) : X → X, X = CN ⊗ · · · ⊗ CN ⊗ C ⊗ · · · ⊗ C . (5.2.22)
1 23 4 1 23 4
k1 k2

It is (5.2.16) that bestows the physical implication of the representation (5.2.21)-(5.2.22) of

SU(N), which will be explained in the next subsection.
Based on the irreducible representation theory of SU(N), if k = k 1 + k2 " 2 and N " 2,
the representation (5.2.21) must be reducible, i.e. H(U) can be split into smaller block
diagonal form:

A−1 H(U)A = H1 (U) ⊕ · · · ⊕ HK (U), ∀U ∈ SU(N). (5.2.23)

Usually, (5.2.23) is simply denoted as

N · · ⊗ N4 ⊗ N
1 ⊗ ·23 · · ⊗ N4 = m1 ⊕ · · · ⊕ mK ,
1 ⊗ ·23 (5.2.24)
k1 k2

where m j is the order of H j (U). In Subsection 5.2.4 we shall give the computational method
of (5.2.23) (or (5.2.24)) by the Young tableau.
5.2 Quark Model 285

5.2.3 Physical explanation of irreducible representations

The physical implications of irreducible representations of SU(N) were revealed first by

Sakata in 1950’s. In the following we give the Sakata explanation in a few steps.

1. The dimension N of SU(N) represents N particles:

ψ1 , · · · , ψ N , (5.2.25)

and their complex linear combination constitute CN :

. / 0
N /
C = ∑ zk ψk / zk ∈ C, 1 ! k ! N ,
N /
j=1

which contains all physical states of the N particles (5.2.25).

In addition, the N fundamental particles of SU(N) are the antiparticles of (5.2.25) given
by
ψ 1, · · · , ψ N , (5.2.26)

where ψ k is the complex conjugate of ψk . The linear space

. / 0
N /
N
C = ∑ yk ψ k / yk ∈ C, 1 ! k ! N
/
j=1

contains all physical states of the N antiparticles (5.2.26).

2. Each matrix U ∈ SU(N) and each U ∈ SU(N) represent the transformations of phys-
ical states of particles (5.2.25) and antiparticles (5.2.26) as follows
N N
∑ zk ψk → ∑ 5zk ψk ,
j=1 j=1
(5.2.27)
N N
∑ yk ψ k → ∑ y5k ψ k ,
j=1 j=1

where ⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
z1
5 z1 y51 y1
⎜ .. ⎟ ⎜ .. ⎟ , ⎜ .. ⎟ ⎜ .. ⎟ .
⎝ . ⎠=U⎝ . ⎠ ⎝ . ⎠=U⎝ . ⎠ (5.2.28)
zN
5 zN y5N yN

3. The tensor product of fundamental particles (5.2.25) and (5.2.26) are denoted by

N ⊗ · · · ⊗ N ⊗ N ⊗ · · · ⊗ N = {ψi1 · · · ψik ψ j1 · · · ψ jk }, (5.2.29)

1 23 4 1 23 4 1 2
k1 k2
286 Chapter 5 Elementary Particles

which stands for a new particle system where each particle

ψi1 ···ik1 j1 ··· jk2 = ψi1 · · · ψik1 ψ j1 · · · ψ jk (5.2.30)

2

is a composite particle made up of ψi1 , · · · , ψik1 , ψ j1 , · · · , ψ jk .

2
For example for N = 3, the tensor product
⎛ ⎞
ψ1 ψ 1 ψ1 ψ 2 ψ1 ψ 3
3 ⊗ 3 = ⎝ ψ2 ψ 1 ψ2 ψ 2 ψ2 ψ 3 ⎠
ψ3 ψ 1 ψ3 ψ 2 ψ3 ψ 3

contains 9 new particles

ψi j = ψ i ψ j , 1 ! i, j ! 3,
composed of a fundamental particle ψi and an anti-particle ψ j .
4. The state space of composite particle system (5.2.29) is the tensor product of k 1
N
complex spaces CN and k2 complex conjugate spaces C , expressed as
N N
CN ⊗ · · · ⊗ C N ⊗ C ⊗ · · · ⊗ C (5.2.31)
1 23 4 1 23 4
k1 k2
6 N N N N
= ∑ ··· ∑ ∑ ··· ∑ zi1 ···ik j ··· jk2
1 1
ψi1 ···ik1 j1 ··· jk2 |
i1 =1 ik1 =1 j1 =1 jk2 =1
7
zi1 ···ik j ··· jk2 ∈ C, ψi1 ···ik j ··· jk2 are as in (5.2.30) .
1 1 1 1

5. Denote the space (5.2.31) as

k N k2 N N
CN 1 ⊗ C = CN ⊗ · · · ⊗ C N ⊗ C ⊗ · · · ⊗ C .
1 23 4 1 23 4
k1 k2

Then a representation H(U) of (5.2.21) for U ∈ SU(N) is a linear transformation of state

space of composite particle system:
k N k2 k N k2
H(U) : CN 1 ⊗ C → CN 1 ⊗ C , (5.2.32)

which represents state transformation of composite particles, similar to the state transfor-
mation (5.2.27)-(5.2.28) for a single particle system.
6. We recall the irreducible representation of SU(N) given by (5.2.23) and (5.2.24).
The irreducible representation implies that there is a decomposition in the N k composite
particles (5.2.29), which are classified into K groups as

G1 = {Ψ11 , · · · , Ψ1m1 }, · · · , GK = {ΨK1 , · · · , ΨKmK }, (5.2.33)

5.2 Quark Model 287

j j
where each group G j contain m j composite particles Ψ1 , · · · , Ψm j , as in (5.2.30), such that
the space (5.2.31) can be decomposed into the direct sum of the K subspaces of span G j as
k N k2
CN 1 ⊗ C = E1 ⊕ · · · ⊕ EK , (5.2.34)

where
E j = span G j = span{Ψ1j , · · · , Ψmj j }, 1 ! j ! K, (5.2.35)
and then, the linear transformations H(U) in (5.2.32) are also decomposed into the direct
sum for all U ∈ SU(N) as in (5.2.23). Namely, under the decomposition (5.2.33) of basis of
k N k2
CN 1 ⊗ C , the representations

H(U) ∈ SU(N) ⊗ · · · ⊗ SU(N) ⊗ SU(N) ⊗ · · · ⊗ SU(N)

1 23 4 1 23 4
k1 k2

can also be decomposed in the form

H(U) = H1 (U) ⊕ · · · ⊕ HK (U), ∀U ∈ SU(N), (5.2.36)

and H j (U) (1 ! j ! K) are as in (5.2.23), such that

H j (U) : E j → E j , dim E j = m j , 1 ! j ! K. (5.2.37)

In other words, the subspace E j of (5.2.35) is invariant for the linear transformation (5.2.36)-
(5.2.37).
7. Sakata’s explanation of irreducible representation of SU(N). Now, we can deduce
the following physical conclusions from the discussions in above steps 1-6.
Physical Explanation 5.5 Let (5.2.29) be a family of composite particles as given by
(5.2.30). The irreducible representation (5.2.36), which usually is expressed as

N ⊗ · · · ⊗ N ⊗ N ⊗ · · · ⊗ N = m1 ⊕ · · · ⊕ m K ,
1 23 4 1 23 4
k1 k2

means that
• the composite particle system (5.2.29) can be classified into K groups of particles:
j
G j = {Ψ1 , · · · , Ψmj j }, 1 ! j ! K;

• each group G j has m j particles, such that under the state transformation (5.2.27)-
(5.2.28) of fundamental particles:
N N
U : CN → CN , U : C →C (U ∈ SU(N)),

the particles in G j only transform between themselves as in (5.2.37).

288 Chapter 5 Elementary Particles

We now examine the Physical Explanation 5.5 from the mathematical viewpoint. The
k N k2
Nk (k = k1 + k2 ) elements of (5.2.29)-(5.2.30) form a basis of CN 1 ⊗ C . We denote these
elements as
E = {Ψ1 , · · · , ΨN k }, (5.2.38)
with each Ψ j as in (5.2.30). The irreducible representation (5.2.23):

AH(U)A−1 = H1 (U) ⊕ · · · ⊕ HK (U), ∀U ∈ SU(N)

implies that if we take the basis transformation for (5.2.38)

⎛ ⎞ ⎛ ⎞
Ψ
51 Ψ1
⎜ .. ⎟ ⎜ . ⎟
⎝ . ⎠ = A ⎝ .. ⎠ , (5.2.39)
Ψ
5 k
N
ΨN k

then under the new basis E5 = {Ψ

5 1, · · · , Ψ
5 k }, the linear mapping
N

H(U) : X → X (X as in (5.2.22))

becomes the block diagonal form

⎛ ⎞
H1 (U) 0
H(U) = ⎝
⎜ .. ⎟
⎠, ∀U ∈ SU(N).
.
0 HK (U)

This means that the new basis E5 of X is divided into K sub-bases

G1 = {E51 , · · · , E5m1 }, · · · , GK = {E5JK +1 , · · · , E5Jk +mK },

with JK + mK = N K , such that each subspace XJ of X, spanned by the j-th sub-basis G j as

X j = span G j (1 ! j ! K)

is an invariant subspace of the mapping H(U) for all U ∈ SU(N). In particular, the block
matrix H j (U) is the restriction of H(U) on X j (1 ! j ! K):

H(U)|X j = H j (U) : X j → X j . (5.2.40)

Hence, when we take any linear transformation on CN as

U : CN → CN , U ∈ SU(N),

then the subspaces X j will undergo themselves a linear transformation in the fashion as
given by (5.2.40).
Thus, from the mathematical viewpoint, the Physical Explanation 5.5 is sound, and
provides the mathematical foundation for the quark model.
5.2 Quark Model 289

5.2.4 Computations for irreducible representations

We have understood the physical implications of the irreducible representations of SU(N).
The remaining crucial problem is how to compute the irreducible representations. Namely,
for an N-dimensional representation as (5.2.21), we need to determine its irreducible de-
composition (5.2.24). In other words, we need to determine m 1 , · · · , mK and K in

N · · ⊗ N4 ⊗ N
1 ⊗ ·23 · · ⊗ N4 = m1 ⊕ · · · ⊕ mK .
1 ⊗ ·23 (5.2.41)
k1 k2

A very effective method to compute (5.2.41) is the Young tableaux, which uses square
diagrams to deduce these numbers m j (1 ! j ! K). The method is divided in two steps.
The first step is to obtain the rule to group together square diagrams, and to obtain the
irreducible representation (5.2.41) in the form

N · · ⊗ N4 ⊗ N
1 ⊗ ·23 · · ⊗ N4 = 1 ⊕ 2 ⊕ · · · ⊕ K .
1 ⊗ ·23 (5.2.42)
k1 k2

The second step provides the rule and method to compute the dimensional m j from the j-th
square diagrams j on the right-hand side of (5.2.42):

Computation of m j from j . (5.2.43)

1. Rule to group the square diagram in (5.2.42). First of all, in the Young tableau we
use a square to stand for a fundamental representation N of SU(N), and use a column of
N − 1 squares to stand for the conjugate representation N in the right-hand side of (5.2.42):
⎫
⎬
N= , N= N − 1.
⎭

For example, we can write 3 ⊗ 3 ⊗ 3 as

3⊗3⊗3= ⊗ ⊗ .

By this rule, the left-hand side of (5.2.42) can be expressed as

N · · ⊗ N4 ⊗ N
1 ⊗ ·23 · · ⊗ N4 = ⊗ · · · ⊗ ⊗
1 ⊗ ·23 ⊗ ···⊗ (5.2.44)
1 23 4
k1 k2 k1
1 23 4
k2

Now, we give rules to group the square diagrams of the right-hand side of (5.2.42) for
different k1 and k2 .
290 Chapter 5 Elementary Particles

(a) Case k1 = 2, k2 = 0: By the rule (5.2.44), the left-hand side is

N ⊗N = ⊗ .

Then the right-hand side is defined as

⊗ α = α ⊕ = ⊕ . (5.2.45)
α

(b) Case k1 = 1 and k2 = 1: We have

⎫
⎫ α ⎪ ⎫
⎪ ⎪
⎪ α ⎪
⎪
⎪ ⎪
⎪ ⎪
⎪
⎬ ⎬ ⎬
α ⊗ N −1 = N+ N −1 (5.2.46)
⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎭ ⎪
⎪ ⎪
⎭
⎪
⎭

(c) Case k1 = 3 and k2 = 0: We define the right-hand side of (5.2.42) as shown

N ⊗N⊗N = ⊗ ⊗ (5.2.47)
; <
= ⊗ ⊗
! "
= ⊕ ⊗ (by (5.2.45))

= ⊗ ⊕ ⊗

= ⊗ α ⊕ ⊗ (use (5.2.46) for 2nd term)

β
β
= α ⊕ ⊕ ⊕ .
α

(d) Case k1 = 2 and k2 = 1: We have

⎫
⎪
⎬
N ⊗N⊗N = ⊗ ⊗ N −1
⎪
⎭
⎛ ⎫ ⎞
⎪
⎪
⎜ ⎬ ⎟
=⎜
⎝ ⊗ N − 1⎟
⎠⊗ (use (5.2.46))
⎪
⎪
⎭
5.2 Quark Model 291

⎡ ⎫ ⎤
⎪ ⎫
⎪
⎪ ⎪
⎢ ⎬ ⎬ ⎥
⎢ ⎥
=⎢ N+ N − 1⎥ ⊗ α
⎣ ⎪
⎪ ⎪
⎭ ⎦
⎪
⎭
⎫
⎫ (5.2.48)
⎪
⎪
⎪
⎬ ⎪
⎬
= α ⊗ N⊕ β ⊗ N −1
⎪
⎪ ⎪
⎭
⎪
⎭

⎫ ⎫ ⎫ ⎫
α ⎪ ⎪
⎪ ⎪
⎪ β ⎪
⎪
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎬ ⎬ β ⎬ ⎬
= N⊕ N⊕ N −1⊕ N − 1.
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎭ ⎪
⎪ ⎪
⎪ ⎪
⎪
β ⎭ ⎭ ⎭

We know that the representations of SU(N) for particles and antiparticles are conjugate
relation, i.e. if
N ⊗ ···⊗ N ⊗N ⊗ ···⊗ N (5.2.49)
1 23 4 1 23 4
k1 k2

represents a particle system, then its conjugate

N ⊗ ···⊗ N ⊗N ⊗ ···⊗ N (5.2.50)

1 23 4 1 23 4
k1 k2

represents the antiparticle system. Due to the symmetry of particles and antiparticles,
(4.2.49) and (4.2.50) have the same irreducible representations. In addition each composite
particle is composed of N particles with N ! 3. Hence it is enough to give the four cases
1)-4) above for the physical purpose.
2. Computation of (5.2.43). In the right-hand side of the Young tableaux (4.2.45)-
(4.2.48), the number of square diagrams is the K as in (4.2.42), and each of which represents
a dimension m j (1 ! j ! K) of an irreducible representation. The rule to compute m j is as
follows.
For example, for the following square diagram

(5.2.51)

the dimension m is given by the formula

αN
m= , (5.2.52)
βN
292 Chapter 5 Elementary Particles

where αN and βN can be computed by the diagram as (5.2.51).

(a) Computation of αN . Fill the blanks of (4.2.51) in numbers in the following fashion:

N N+1 N+2 N+2

N−1 N N+1 (5.2.53)
N−2

The αN equals to the multiplication of all numbers in (5.2.53),

αN = N(N + 1)(N + 2)(N + 3)(N − 1)N(N + 1)(N − 2). (5.2.54)

(b) Computation of βN . Fill the blanks of (5.2.51) in the fashion:

(5.2.55)

where the data in a square equals to the number of all squares on its right-hand side and
below it adding one. For example, for the square marked 6, there are 3 squares on its right-
hand side, and 2 square below it. Hence the number k in this blank is

k = 3 + 2 + 1 = 6.

Then, the number βN is the multiplication of all numbers in (4.2.55)

βN = 6 × 4 × 3 × 1 × 4 × 2 × 1 × 1. (5.2.56)

Thus, by (4.2.54) and (4.2.56) we can get the value of (4.2.52). In the following, we
give a few examples to show how to use the Young tableau to compute the irreducible
representations of SU(N).

Example 5.6 For N = 3, consider the two cases given by

3⊗3 and 3 ⊗ 3 ⊗ 3, (5.2.57)

which are the most important cases in particle physics. By (5.2.46) and (5.2.47), the Young
tableaux of (5.2.57) are as follows

3⊗3 = ⊗ = ⊕ (5.2.58)

3⊗3⊗3= ⊗ ⊗ = ⊕ ⊕ ⊕ (5.2.59)
5.2 Quark Model 293

According to (5.2.52) and the methods to compute αN and βN (N = 3), we infer from
(5.2.58) and (5.2.59) that

αN = 3 × 2 × 1, βN = 3 × 2 × 1, m = 1, for

αN = 3 × 4 × 2, βN = 3 × 1 × 1, m = 8, for

αN = 3 × 4 × 5, βN = 3 × 2 × 1, m = 10, for
Consequently, we derive, from (5.2.58) and (5.2.59), the following irreducible representa-
tions of (5.2.57):
3 ⊗ 3 = 1 ⊕ 8,
(5.2.60)
3 ⊗ 3 ⊗ 3 = 1 ⊕ 8 ⊕ 8 ⊕ 10.
Example 5.7 The other important cases in physics are the two irreducible representa-
tions for N = 4:
4 ⊗ 4 and 4 ⊗ 4 ⊗ 4.
Their Young tableaux are given by

⊗ = ⊕

⊗ ⊗ = ⊕ ⊕ ⊕

Then we can compute that

m = 1, for

m = 15, for

m = 4, for

m = 20, for and

Hence we deduce that

4 ⊗ 4 = 1 ⊕ 15,
(5.2.61)
4 ⊗ 4 ⊗ 4 = 4 ⊕ 20 ⊕ 20 ⊕ 20.
294 Chapter 5 Elementary Particles

5.2.5 Sakata model of hadrons

In 1950’s, many hadrons were discovered, leading to many attempts to investigate the deeper
hadron structure. Based on the irreducible representation (5.2.60) of SU(3), i.e..

3 ⊗ 3 = 8 ⊕ 1, (5.2.62)

Sakata presented a model for hadron structure, called the Sakata model. This was an early
precursor to the quark model, and also resulted in the physical implications of irreducible
representations as stated by Physical Explanation 5.5.
Sakata model proposed three particles

p, n, Λ (5.2.63)

as the fundamental particles for all strong interacting particles. In his scheme, Sakata used
the three particles in (5.2.63) as a basis of SU(3), such that each hadron consists of a funda-
mental particle and an antiparticle as

hadron = Si S j for 1 ! i, j ! 3, (5.2.64)

where (S1 , S2 , S3 ) = (p, n, Λ) are called the sakataons.

In (5.2.62), the left-hand side represents the pairs S i S j , and the right-hand side represents
the following eight mesons:
0
π +, π −, π 0 , K + , K − , K 0 , K , η . (5.2.65)

It is a coincidence that the eight particles just constitute an eight multiple state of hadrons,
and can be illustrated by the Eightfold Way; see Figure 5.3 and Figure 5.6.

Figure 5.6
5.2 Quark Model 295

Table 5.6
Sakataons I I3 Q S B
p 1/2 +1/2 +1 0 +1
n 1/2 −1/2 0 0 +1
Λ 0 0 0 −1 +1

In addition, from the quantum numbers of the three particles(5.2.63), we can deduce the
quantum numbers for the eight mesons in(5.2.65); see Table 5.6 and 5.7.
In Table 5.7, J is the spin and π is the parity. Hence J π = 0− represents that spin J = 0
and parity π = −1.
The Sakata model explains the mesons well. However, it encounters difficulties if we
use this model to describe baryons. In fact, if we use the three sakataons to form a baryon,
then the baryon number B = 3 for a baryon. Hence, we have to use two sakataons and one
anti-sakataon to form a baryon to ensure B = 1. Unfortunately those combinations do not
agree with experiments.

Table 5.7
mesons I I3 Q S Jπ m(MeV )
pn π+ 1 1 1 0 0− ∼ 140
−pn π− 1 −1 −1 0 0− ∼ 140
nn − pp
√ π0 1 0 0 0 0− ∼ 140
2
1 1
pΛ K+ 1 1 0− ∼ 495
2 2
1 1
nΛ K0 − 0 1 0− ∼ 495
2 2
0 1 1
Λn K 0 −1 0− ∼ 495
2 2
− 1 1
Λp K − −1 −1 0− ∼ 495
2 2
nn + pp − 2ΛΛ
√ η 0 0 0 0 0− ∼ 548
6
nn + pp + ΛΛ
√ η′ 0 0 0 0 0− ∼ 548
3

5.2.6 Gell-Mann-Zweig’s quark model

Based on the Eightfold Way, as shown in Figures 5.2-5.5, the hadrons are classified as
follows:
.
Mesons (J = 0) : 8 particles,
Mesons (J = 1) : 8 particles,
⎧ ! "
⎨ Baryons J = 1 : 8 particles,
⎪
2
⎪
⎩ Baryons (J = 3/2) : 10 particles.
296 Chapter 5 Elementary Particles

In view of the irreducible representations of SU(3) :

3 ⊗ 3 = 8 ⊕ 1,
3 ⊗ 3 ⊗ 3 = 10 ⊕ 8 ⊕ 8 ⊕ 1,

it is natural to guess the relations (5.2.18) or (5.2.19), i.e.

mesons = 3 ⊗ 3 = 8 ⊕ 1,
(5.2.66)
baryons = 3 ⊗ 3 ⊗ 3 = 10 ⊕ 8 ⊕ 8 ⊕ 1.

According to the Physical Explanation 5.5, we infer immediately, from (5.2.66), that the
hadrons in the Eightfold Way are composed of three fundamental particles, denoted by

q1 , q2 , q3 (5.2.67)

and mesons consist of a fundamental particle and an antiparticle. In a nutshell, baryons

consist of three fundamental particles:

mesons = qi q j for 1 ! i, j ! 3,
(5.2.68)
baryons = qi q j qk for 1 ! i, j, k ! 3.

Physicists Gell-Mann and Zweig did this work independently in 1964, and presented
the celebrated Quark Model. The three fundamental particles (5.2.67) were termed the up,
down and strange quarks by Gell-Mann, denoted by

q1 = u, q2 = d, q3 = s.

Zweig called these particles the aces.

By using the three quarks (u, d, s) to replace the sakataons (p, n, Λ), we can perfectly
explain all hadrons described by the Eightfold Way, i.e. the hadrons given by (5.2.1)-(5.2.4).
An important step to establish the quark model is to determine the quantum numbers of
quarks, which are introduced in the following several procedures:

1) The spins of quarks have to be J = 21 , as required by

spins of mesons: J = 0(↑↓), J = 1(#), J = −1($),

1 1
spins of baryons: J = (#↓), J = − ($↑), (5.2.69)
2 2
3 3
J = (↑↑↑), J = − (↓↓↓).
2 2
1
Namely, the unique choice to ensure (5.2.69) is that J = 2 for quarks.
5.2 Quark Model 297

2) The success of (p, n, Λ) for describing mesons suggests that for the strange number
S and the isospin (I, I3 ), u and p, d and n, s and Λ should be the same respectively.
Hence we have
! "
1 1
u : (S, I, I3 ) = 0, , ,
2 2
! "
1 1
d : (S, I, I3 ) = 0, , − ,
2 2
s : (S, I, I3 ) = (−1, 0, 0).

3) Since the baryon numbers of all baryons are B = 1, by the constituents (5.2.68) of
baryons, it is natural that
1 1 1
u: B= , d: B= , s: B= .
3 3 3
4) For all hadrons, the following formula, well known as the Gell-Mann-Nishijima rela-
tion, holds true:
B S
Q = I3 + + . (5.2.70)
2 2
This relation should also be valid for quarks. Hence, we deduce from (5.2.70) that
the electric charges of quarks are
2 1 1
u: Q= , d: Q=− , S: Q=− .
3 3 3
The data derived in 1)-4) above are collected in Table 5.3. According to the quantum
numbers of hadrons and quarks, we can determine the quark constituents of all hadrons.
For example, for uud its quantum numbers are derived from those of u and d as follows
1 1 1
uud : B = 1, Q = 1, S = 0, J = , I = , I3 = , (5.2.71)
2 2 2
which dictate that uud is the proton:

uud = p.

The quark constituents of the main hadrons are listed in Subsection 5.1.1.
In 1974, Ting and Richter discovered independently J/ψ particle, which implies the
existence of a new quark, named as the c quark. Thus, the quark family then was extended
to four members:
u, d, s, c.
As fundamental particles of SU(4), the irreducible representations (5.2.61), written as

4 ⊗ 4 = 15 ⊕ 1,
(5.2.72)
4 ⊗ 4 ⊗ 4 = 20 ⊕ 20 ⊕ 20 ⊕ 4
298 Chapter 5 Elementary Particles

suggest that the meson groups with 8 particles should be extended to 15 particles, and the
baryons be extended to 20 particles in a group. The extended group of mesons with 15
particles has been verified, which is given by
0
π ± , π 0 , K ± , K 0 , K , η , η ′ , D0 , D0 , D± , D±
s .

However, the 20 baryons corresponding to the irreducible representation (5.2.72) have not
been discovered now.
Up to now, the quark family has six members:

u, d, s, c, b,t.

Their irreducible representation is given by

6 ⊗ 6 = 35 ⊕ 1,
(5.2.73)
6 ⊗ 6 ⊗ 6 = 70 ⊕ 70 ⊕ 56 ⊕ 20.
The classification scheme corresponding to (5.2.73) does not seem to be realistic. The irre-
ducible representation of SU(N) is only a phenomenological theory in elementary particle
physics.

5.3 Weakton Model of Elementary Particles

5.3.1 Decay means the interior structure
As we addressed in Section 5.1.3, all charged leptons, quarks and mediators can undergo
decay as follows:

1) Charged lepton radiation and decay:

e− → e − + γ ,
µ − → e− + ν e + ν µ , (5.3.1)
τ → µ + ν µ + ντ .
− −

2) Quark decay:

d → u + e− + ν e ,
s → d + γ, (5.3.2)
c → d + s + u.

3) Mediator decay:

2 γ → e + + e− ,
W ± → l ± + ν l± , (5.3.3)
+ −
Z → l +l ,
5.3 Weakton Model of Elementary Particles 299

where l ± are the charged leptons.

All leptons, quarks and mediators are currently regarded as elementary particles. How-
ever, the decays in (5.3.1)-(5.3.3) show that these particles must have an interior structure,
and consequently they should be considered as composite particles rather than elementary
particles:

Decay Means Interior Structure.

5.3.2 Theoretical foundations for the weakton model

Subatomic decays and electron radiations indicate that there must be interior structure for
charged leptons, quarks and mediators. The main objective of this section is to propose an
elementary particle model, which we call weakton model, based on the weak and strong
interaction theories developed in the last chapter.
Angular momentum rule
It is known that the dynamic behavior of a particle is described by the Dirac equations:
∂ψ
ih̄ = Hψ, (5.3.4)
∂t
where ψ = (ψ1 , ψ2 , ψ3 , ψ4 )T is the Dirac spinor, H is the Hamiltonian:

H = −ih̄c(α k ∂k ) + mc2α 0 + V (x), (5.3.5)

V is the potential energy, α k (1 ! k ! 3) are the matrices as given by (2.2.48), and α 0 is the
matrix as ! " ! "
I 0 1 0
α0 = , I= .
0 −I 0 1
By the conservation laws in quantum mechanics, if an Hermitian operator L commutes with
H in (5.3.5):
LH = HL,
then the physical quantity L is conserved.
Consider the total angular momentum J⃗ of a particle as

J⃗ = ⃗L + s⃗S,

where ⃗L is the orbital angular momentum

⃗L =⃗r × ⃗p, ⃗p = −ih̄∇,

s is the spin, and ! "

⃗S = (S1 , S2 , S3 ), σk 0
Sk = h̄ ,
0 σk
300 Chapter 5 Elementary Particles

and σk (1 ! k ! 3) are the Pauli matrices; see (2.2.47).

It is readily to check that for H in (5.3.5),
1
J⃗1/2 = ⃗L + ⃗S commutes with H,
2 (5.3.6)
1
J⃗s = ⃗L + s⃗S does not commute with H for s ̸= in general.
2
Also, we know that

s⃗S commutes with H in straight line motion for any s. (5.3.7)

1
The properties in (5.3.6) imply that only particles with spin s = can make a rotational mo-
2
tion in a center field with free moment of force. However, (5.3.7) implies that the particles
1
with s ̸= will move in a straight line, i.e. ⃗L = 0, unless they are in a field with nonzero
2
moment of force.
In summary, we have derived the following angular momentum rule for subatomic par-
ticle motion, which is important for our weakton model established in the next subsection.
The more general form of the angular momentum rule will be addressed in Section 6.2.4.
1
Angular Momentum Rule 5.8 Only the fermions with spin s = can rotate around
2
1
a center with zero moment of force. The fermions with s ̸= will move on a straight line
2
unless there is a nonzero moment of force present.
For example, the particles bounded in a ball rotating around the center, as shown in
Figure 5.7, must be fermions with s = 1/2.

Figure 5.7 (a) Two particles A, B rotate around the center 0, amd (b) three particles A, B, C rotate
around the center 0.

Mass generation mechanism

For a particle moving with velocity v, its mass m and energy E obey the Einstein relation
D
C v2
2
E = mc 1− 2. (5.3.8)
c
5.3 Weakton Model of Elementary Particles 301

Usually, we regard m as a static mass which is fixed, and energy is a function of velocity v.
Now, taking an opposite viewpoint, we regard energy E as fixed, mass m as a function
of velocity v, and the relation (5.3.8) is rewritten as
D
v2 E
m= 1− . (5.3.9)
c2 c2
Thus, (5.3.9) means that a particle with an intrinsic energy E has zero mass m = 0 if it
moves at the speed of light v = c, and will possess nonzero mass if it moves with a velocity
v < c.
All particles including photons can only travel at the speed sufficiently close to the speed
of light. Based on this viewpoint, we can think that if a particle moving at the speed of light
(approximately) is decelerated by an interaction force ⃗F, obeying
D
d ⃗P v2 ⃗
= 1− F,
dt c2
then this massless particle will generate mass at the instant. In particular, by this mass
generation mechanism, several massless particles can yield a massive particle if they are
bound in a small ball, and rotate at velocities less than the speed of light.

Interaction charges

In the unified field model introduced in the last chapter, we derived that both weak and
strong interactions possess charges, as for gravity and electromagnetism; see Section 4.3.4:

gravitation: mass charge m,

electromagnetism: electric charge e,
(5.3.10)
weak interaction: weak charge gw ,
strong interaction: strong charge gs .

If Φ is a charge potential corresponding to an interaction, then the interacting force gener-

ated by its charge g is given by
F = −g∇Φ,

where ∇ is the spatial gradient operator.

It is very crucial to introduce both weak and strong interaction charges for us to develop
the weakton model. The charges in (5.3.10) possess the following physical properties:

1) Electric charge Qe , weak charge Qw , strong charge Qs are conservative. The energy
is a conserved quantity, but the mass M is not a conserved quantity due to the mass
generation mechanism as mentioned in (5.3.9).
302 Chapter 5 Elementary Particles

2) There is no interacting force between two particles without common charges. For
example, if a particle A possesses no strong charge, then there is no strong interacting
force between A and any other particles.

3) Only the electric charge e can take both positive and negative values, and other
charges take only nonnegative values.

Layered formulas of strong interaction potentials

The layered properties of strong and weak interaction potentials are also crucial for us
to establish the weakton model. We recall briefly the strong interaction potentials in the
general form as: E F
1 A
Φs = gs (ρ ) − (1 + kr)e−kr ,
r ρ
! "3 (5.3.11)
ρw
gs (ρ ) = N gs ,
ρ
where ρ is the particle radius, N is the number of strong charges.
There are about five levels of strong interactions between particles: w∗ -weaktons, quarks,
gluons, hadrons, atoms/molecules. Hence, based on (5.3.11) we can derive the layered for-
mulas for these five level particles as follows:
E F
s 1 A0 −k0 r
Φw∗ = gs − (1 + k0r)e ,
r ρw
! "3 E F
ρw 1 Aq
Φsq = gs − (1 + k1r)e−k1 r ,
ρq r ρq
! "3 E F
ρw 1 Ag
Φsg = 2 gs − (1 + k1r)e−k1 r , (5.3.12)
ρg r ρg
! " E F
ρw 3 1 An
Φsn = 3 gs − (1 + knr)e−kn r ,
ρn r ρn
! "3 E F
ρw 1 Aa
Φsa = N gs − (1 + kar)e−ka r .
ρa r ρa

where ρw , ρq , ρg , ρn , ρa are the radii of weakton, quark, gluon, nucleon, atom/molecule, and
gs is the strong charge of w∗ -weakton. By (4.5.70), gs can be expressed as
! "6
1 ρn
g2s = β 2 g2 , (5.3.13)
9 ρw
√ −1/4
where β = 2e /(8 − e1/2)1/2 and g is the Yukawa charge.
5.3 Weakton Model of Elementary Particles 303

Layered formulas for weak interaction potentials

The weak interaction potential for a particle with radius ρ and with N weak charges g w
is given by
E F
1 B
Φw = gw (ρ )e−kr − (1 + 2kr)e−kr ,
r ρ
! "3 (5.3.14)
ρw
gw (ρ ) = N gw ,
ρ
1
where k = = 1016 cm−1 , and r0 = 10−16 cm.
r0
For the weak interaction, there are about four levels: weaktons, mediators, quarks and
charged leptons. By (5.3.14), the layered formulas for these four level particles are given by
E F
1 Bw
Φw0 = gw e −kr
− (1 + 2kr)e −kr
,
r ρw
! "3 E F
ρw 1 Bm
Φwm = 2 gw e−kr − (1 + 2kr)e−kr ,
ρm r ρm
! "3 E F (5.3.15)
w ρw −kr 1 Bq
Φq = 3 gw e − (1 + 2kr)e −kr
,
ρq r ρq
! "3 E F
w ρw −kr 1 Bl
Φl = 3 gw e − (1 + 2kr)e −kr
.
ρl r ρl

By (4.6.37), the weak charge gw of weakton can be expressed as

! "6
1 ρn
g2w = α2 h̄c, (5.3.16)
9 ρw
! "
2 mw
where α = ' √ × 10−2.
5 2 mp

Duality of mediators

Based on the unified field theory, there exists a natural duality between the mediators:

tensor graviton gG ↔ vector graviton φG ,

vector photon γ ↔ scalar photon γ0
(5.3.17)
vector bosons W ± , Z ↔ Higgs H ± , H 0 ,
gluons {gk |1 ! k ! 8} ↔ scalar gluons {gk0 | 1 ! k ! 8}.

We shall see that the prediction (5.3.17) are in perfect agreement with the consequences of
the weakton model.
304 Chapter 5 Elementary Particles

5.3.3 Weaktons and their quantum numbers

The observation of subatomic decays and electron radiations leads us to propose a set of
1
elementary particles, which we call weaktons. They are massless, spin − particles with
2
one unit of weak charge gw .
The introduction of weaktons is based on the following theories, observational facts and
considerations:

1) The interior structure of charged leptons, quarks and mediators demonstrated by the
decays, scatterings and radiations, as as shown in (5.3.1)-(5.3.3);

2) The new quantum numbers of weak charge gw and strong charge gs introduced in
(5.3.10);

3) The mass generation mechanism presented in Section 5.3.2;

4) The weakton confinement theory given by the layered formulas of weak interaction
potentials (5.3.15)-(5.3.16); and

5) The duality (5.3.17) for the mediators.

The weaktons consist of the following 6 elementary particles and their antiparticles:

w∗ , w1 , w2 , νe , νµ , ντ ,
(5.3.18)
w , w1 , w2 , ν e , ν µ , ν τ ,
∗

where νe , νµ , ντ are the three flavor neutrinos, and w∗ , w1 , w2 are three new elementary
particles, which we call w-weaktons.
These weaktons in (5.3.18) are endowed with the quantum numbers: electric charge Q e ,
weak charge Qw , strong charge Qs , weak color charge Qc , baryon number B, lepton numbers
Le , Lµ , Lτ , spin J, and mass m. The quantum numbers of weaktons are listed in Table 5.8.

Table 5.8 Weakton quantum numbers

Weakton Qe Qw Qs Qc Le Lµ Lτ B J m
w∗ 2/3 1 1 0 0 0 0 1/3 ±1/2 0
w1 −1/3 1 0 1 0 0 0 0 ±1/2 0
w2 −2/3 1 0 −1 0 0 0 0 ±1/2 0
νe 0 1 0 0 1 0 0 0 −1/2 0
νµ 0 1 0 0 0 1 0 0 −1/2 0
ντ 0 1 0 0 0 0 1 0 −1/2 0

A few remarks are now in order.

Remark 5.9 For the weaktons and antiweaktons, the quantum numbers Q e , Qc ,
B, Le , Lµ , Lτ have opposite signs, and Qw , Qs , m have the same values. The neutrinos νe , νµ , ντ
5.3 Weakton Model of Elementary Particles 305

1
possess left-hand helicity with spin J = − , and the antineutrinos ν e , ν µ , ν τ possess right-
2
1
hand helicity with spin J = .
2
Remark 5.10 The weak color charge Qc is a new quantum number introduced for the
weaktons only, which will be used to rule out some unrealistic combinations of weaktons.

Remark 5.11 Since the fundamental composite particles as quarks only contain one
w∗ -weakton, there is no strong interaction between the constituent weaktons of a composite
particle, except the gluons. Therefore, for the weaktons (5.3.18), there is no need to intro-
duce the classical strong interaction quantum numbers as strange number S, isospin (I, I3 )
and parity π etc.

Remark 5.12 It is known that the quark model is based on the irreducible representa-
tions of SU(3) as

meson = 3 ⊗ 3 = 8 ⊕ 1,
baryon = 3 ⊗ 3 ⊗ 3 = 10 ⊕ 8 ⊕ 8 ⊕ 1.

However, the weakton model is based on the aforementioned theories and observational
facts 1)-5), different from the quark model.

5.3.4 Weakton constituents and duality of mediators

In this section we introduce the weakton compositions of charged leptons, quarks and me-
diators. Meanwhile, in the weakton compositions of mediators there exists a natural duality
in the spin arrangements, which give rise to the same conclusions as derived in (5.3.17). In
addition, the neutrinos νl (l = e, µ , τ ) form a new particle: the ν -mediator ν = ∑ αl νl ν l ,
which is of special importance because it can not only explain many decays, but also pro-
vide a more reasonable explanation for the well-known solar neutrino problem in Section
6.3.
1. Charged leptons and quarks. The weakton constituents of charged leptons and quarks
are given by
e = νe w1 w2 , µ = νµ w1 w2 , τ = ντ w1 w2 ,
u = w∗ w1 w1 , c = w∗ w2 w2 , t = w∗ w2 w2 , (5.3.19)
∗ ∗ ∗
d = w w1 w2 , s = w w1 w2 , b = w w1 w2 ,
where c,t and d, s, b are distinguished by their spin arrangements. We suppose that

u = w∗ w1 w1 (#↓, $↑, ↑↓↑, ↓↑↓, ↑$, ↓#),

c = w∗ w2 w2 (#↓, $↑, ↑$, ↓#), (5.3.20)
∗
t = w w2 w2 (↑↓↑, ↓↑↓),
306 Chapter 5 Elementary Particles

and
d = w∗ w1 w2 (#↓, $↑),
s = w∗ w1 w2 (↑$, ↓#), (5.3.21)
∗
b = w w1 w2 (↑↓↑, ↓↑↓).
These arrangements (5.3.20) and (5.3.21) are speculative, and the true results will have to
determined by experiments.

2. Mediators. According to the matched quantum numbers, the mediators γ ,W ± , Z, gk

should have the following weakton constituents:

γ = α1 w1 w1 + α2 w2 w2 (α12 + α22 = 1),

Z = β1 w1 w1 + β2 w2 w2 (β12 + β22 = 1),
W + = w1 w2 , (5.3.22)
W − = w1 w2 ,
gk = w∗ w∗ (k = color index).

In view of the W S electroweak theory (Quigg, 2013):

γ = cos θw Bµ − sin θwWµ3 ,

Z = sin θw Bµ + cos θwWµ3 ,
sin2 θw = 0.23,

we take α1 , α2 , β1 , β2 in (5.3.22) as follows

α1 = cos θw , α2 = − sin θw , β1 = sin θw , β2 = cos θw .

There is a natural duality in the spin arrangements:

(#, $) ↔ (↑↓, ↓↑), (5.3.23)

which not only yields new mediators with spin J = 0, but also gives the same conclusions
as in (5.3.17).
Thus, based on (5.3.23), the weakton model also leads to the dual mediators as follows:

γ = cos θw w1 w1 − sin θw w2 w2 (#, $),

Z = sin θw w1 w1 + cos θw w2 w2 (#, $),
W − = w1 w2 (#, $), (5.3.24)
+
W = w1 w2 (#, $),
gk = w∗ w∗ (#, $),
5.3 Weakton Model of Elementary Particles 307

and their dual particles

γ0 = cos θw w1 w1 − sin θw w2 w2 (↑↓, ↓↑),

H 0 = sin θw w1 w1 + cos θw w2 w2 (↑↓, ↓↑),
H − = w1 w2 (↑↓, ↓↑), (5.3.25)
+
H = w1 w2 (↑↓, ↓↑),
gk0 = w∗ w∗ (↑↓, ↓↑).

3. The ν -mediator. By the weak interaction potential formula Φwm in (5.3.15), the neu-
trino pairs
νe ν e , νµ ν µ , ντ ν τ (↓↑) (5.3.26)

should be bounded by the weak interacting force to form a new mediator, although they
have not been discovered. The three pairs in (5.3.26) may be indistinguishable. Hence, they
will be regarded as a particle, i.e. their linear combination

ν = ∑ αl νl ν l (↓↑), ∑ αl = 1, (5.3.27)
l l

is an additional mediator, and we call it the ν -mediator. We believe that ν is an independent

new mediator.

5.3.5 Weakton confinement and mass generation

Since the weaktons are assumed to be massless and no free w-weaktons are found, we have
to explain: i) the w-weakton confinement, and ii) the mass generation mechanism for the
massive composite particles, including the charged leptons e, µ , τ , the quarks u, d, s, c,t, b,
and the bosons W ± , Z, H ± , H 0 .

1. Weakton confinement. The weak interaction potentials (5.3.15) and the weak charge
formula (5.3.16) can help us to understand why no free w∗ , w1 , w2 are found, while single
neutrinos νe , νµ , ντ can be detected.
In fact, by (5.3.15) the weak interaction potential reads
E F
−kr 1 B
Φw = gs e − (1 + 2kr)e −kr
,
r ρw

The bound energy to hold particles together is negative. Hence for weaktons, their weak
interaction bound energy E is the negative part of gs Φw , i.e.

B 2
E =− g (1 + 2kr)e−2kr . (5.3.28)
ρw s
308 Chapter 5 Elementary Particles

By (5.3.16),
! "6
ρn
g2w = 0.63 × !c. (5.3.29)
ρw
It is estimated that
ρn
= 104 ∼ 106 .
ρw
Therefore, the bound energy E given by (5.3.28) and (5.3.29) is very large provided the
weak interaction constant B > 0.
Thus, by the sufficiently large bound energy, the weaktons can form triplets confined in
the interior of charged leptons and quarks as (5.3.19), and doublets confined in mediators
as (5.3.24)-(5.3.26). They cannot be opened unless the exchange of weaktons between the
composite particles.
The free neutrinos νe , νµ , ντ and antineutrinos ν e , ν µ , ν τ can be found in Nature. The
reason is that in the weakton exchange process there appear pairs of different types of neu-
trinos such as νe and ν µ , and between which the governing weak force is repelling because
the constant B in (5.3.28) is non-positive, i.e.

B!0 for different types of neutrinos. (5.3.30)

2. Massless mediators. For the mass problem, we know that the mediators:

γ , gk , ν and their dual particles, (5.3.31)

have no masses. To explain this, we note that these particles in (5.3.31) consist of pairs as

w1 w1 , w2 w2 , w∗ w∗ , νl ν l . (5.3.32)

The weakton pairs in (5.3.32) are bound in a circle with radius R 0 as shown in Figure 5.8.
Since the interacting force on each weakton pair is in the direction of their connecting line,
they rotate around the center 0 without resistance. As ⃗F = 0 in the moving direction, by the
relativistic motion law: D
d⃗ v2
P = 1 − 2 ⃗F, (5.3.33)
dt c

Figure 5.8
5.3 Weakton Model of Elementary Particles 309

the massless weaktons rotate at the speed of light. Hence, the composite particles formed
by the weakton pairs in (5.3.32) have no rest mass.
3. Mass generation. For the massive particles

e, µ , τ , u, d, s, c, t, b, (5.3.34)

by (5.3.19), they are made up of weakton triplets with different electric charges. Hence the
weakton triplets are not arranged in an equilateral triangle as shown in Figure 5.7 (b), and in
fact are arranged in an irregular triangle as shown in Figure 5.9. Consequently, the weakton
triplets rotate with nonzero interacting forces ⃗F ̸= 0 from the weak and electromagnetic
interactions. By (5.3.33), the weaktons in the triplets at a speed less than the speed of light
due to the resistance force. Thus, by the mass generating mechanism introduced in Section
5.3.2, the weaktons become massive. Hence, the particles in (5.3.34) are massive.

Figure 5.9

4. Massive mediators. Finally, we need to explain the masses for the massive mediators:

W ± , Z, H ± , H 0 , (5.3.35)

which consist of the weakton doublets

w1 w2 , w1 w2 , w1 w1 , w2 w2 . (5.3.36)

Actually, in the weakton exchange theory in the next section, we can see that the particles
in (5.3.36) are in some transition states in the weakton exchange procedure. At the moment
of exchange, the weaktons in (5.3.36) are at a speed v with v < c. Hence, the particles in
(5.3.35) are massive, and their life-times are very short (τ ≃ 10 −25s).

5.3.6 Quantum rules for weaktons

By carefully examining the quantum numbers of weaktons, the composite particles in (5.3.19)
and (5.3.24)-(5.3.26) are well-defined.
310 Chapter 5 Elementary Particles

In the last subsection, we solved the free weakton problem and the mass problem. In
this subsection, we propose a few rules to resolve the remainning problems.

1. Weak color neutral rule: All composite particles by weaktons must be weak color
neutral.

Based on this rule, many combinations of weaktons are ruled out. For example, it is
clear that there are no particles corresponding to the following www and ww combinations,
because they all violate the weak color neutral rule:

νe w1 w2 , w∗ w2 w2 , w∗ w1 w2 , etc., νe w1 , w∗ w1 , w∗ w2 etc.

2. BL = 0, Li L j = 0 (i ̸= j), where B is the baryon number, and L = L j = Le , Lµ , Lτ are

the lepton numbers.

The following combinations of weaktons

w∗ νi , νi ν j , νi ν k (i ̸= k),

are not observed in Nature, and are ruled out by this rule.

3. L + Qe = 0, |B + Qe | ! 1 for L, B ̸= 0.

The following combinations of weaktons

νi w1 w1 , νi w2 w2 , ν i w1 w2 , w∗ w∗ etc (5.3.37)

cannot be found in Nature, and are ruled out this rule.

3
4. Spin selection. In reality, there are no weakton composites with spin J = 2 as

w∗ w1 w1 (↑↑↑, ↓↓↓), w∗ w2 w2 (↑↑↑, ↓↓↓), w∗ w1 w2 (↑↑↑, ↓↓↓), (5.3.38)

and as
ν w1 w2 (↑↑↑, ↓↓↓). (5.3.39)

The cases (5.3.38) are excluded by the Angular Momentum Rule 5.8. The reasons for
this exclusion are two-fold. First, the composite particles in (5.3.38) carry one strong charge,
and consequently, will be confined in a small ball by the strong interaction potential as the
quarks confined in hadrons, as shown in Figure 5.7 (b). Second, due to the uncertainty
principle, the bounding particles will rotate, at high speed with almost zero moment of
1
force, which must be excluded for composite particles with J ̸= based on the angular
2
momentum rule.
5.3 Weakton Model of Elementary Particles 311

The exclusion for (5.3.39) is based on the observation that by the left-hand helicity of
1 1
neutrinos with spin J = − , one of w1 and w2 must be in the state with J = + to combine
2 2
with ν j (1 ! j ! 3), i.e. in the manner as

ν j w1 w2 (↓#, ↓↑↓, $↑).

In summary, under the above rules 1-4, only the weakton constituents in (5.3.19) and
(5.3.24)-(5.3.26) are allowed.
5. Eight quantum states of gluons. It is known that the gluons have eight quantum states

gk : g1 , · · · , g8 .

In (5.3.24) and (5.3.25), the vector and scalar gluons have the forms

w∗ w∗ (#, $) and w∗ w∗ (↑↓, ↓↑).

According to QCD, quarks have three colors

red(r), green(g), blue(b),

and anticolor r, g, b. They obey the rules

bb = rr = gg = w ( white).

Based on (5.3.19), w∗ is endowed with three colors and anticolors:

w∗r , w∗g , w∗b and wr∗ , wg∗ , wb∗ ,

which form the bases of SU(3) and SU(3). By the irreducible representation:

3 × 3 = 8 ⊕ 1,

we can derive the eight-multiple states as

g1 = (w∗ w)w , g2 = w∗b w∗r , g3 = w∗b w∗g , g4 = w∗r w∗g , (5.3.40)

5 ∗ 6 7 8
g = (w w)w , g = w∗r w∗b , g = w∗g w∗b , g = w∗g w∗r . (5.3.41)

where (w∗ w∗ )w is a linear combination of w∗b w∗b , w∗r w∗r , w∗g w∗g . Namely, the gluons in (5.3.41)
are the antigluons of these in (5.3.40).
In summary, the quantum rules presented above can be simply expressed in the follow-
ing formulas for quantum numbers Qc , Qe , L and B:
weak color neutral: Qc = 0,
mutual exclusion of leptons and baryons: BL = 0, Li L j = 0 (i ̸= j),
(5.3.42)
relation of leptons and charges: L + Qe = 0 (L ̸= 0),
relation of baryons and charges: |B + Qe| ! 1 (B ̸= 0).
312 Chapter 5 Elementary Particles

These relations are summed up from realistic particle data. Based on (5.3.42), together with
the mass generation mechanism, the Angular Momentum Rule 5.8 and the layered formulas
of weak and strong interaction potentials, all of the most basic problems in the weakton
model have a reasonable explanation.

5.4 Mechanisms of Subatomic Decays and Electron Radiations

5.4.1 Weakton exchanges

We conclude that all particle decays are caused by weakton exchange. The exchanges occur
between composite particles as mediators, charged leptons, and quarks.

1. Weakton exchange in mediators. First we consider one of the most important decay
processes in particle physics, the electron-positron pair creation and annihilation:

2 γ → e + + e− ,
(5.4.1)
e+ + e − → 2 γ .

In fact, the reaction formulas in (5.4.1) are not complete, and the correct formulas should
be as follows
2 γ + ν % e + + e− , (5.4.2)

where ν is the ν -mediator as in (5.3.27).

Note that the weakton components of γ and ν are

γ = cos θw w1 w1 − sin θw w2 w2 ,
ν = α1 νe ν e + α2 νµ ν µ + α3 ντ ν τ ,

which imply that the probability of the photon γ at the state w1 w1 is cos2 θw and at the state
−w1 w2 is sin2 θw , and the probability of ν at νi ν i is αi2 . Namely, for photon, the densities
of the w1 w1 (#) and −w2 w2 ($) particle states are cos2 θw and sin2 θw , and the densities of
the states νe ν e , νµ ν µ , ντ ν τ are α12 , α22 , α32 . Hence the formula (5.4.2) can be written as

w1 w1 (#) + w2 w2 ($) + νe ν e (↓↑) % νe w1 w2 (↓↑↓) + ν e w1 w2 (#↓). (5.4.3)

It is then clear to see from (5.4.3) that the weakton constituents w1 , w1 , w2 , w2 , νe , ν e can
regroup due to the weak interaction, and we call this process weakton exchange. The mech-
anism of this exchanging process can be explained using the layered formulas in (5.3.15) of
weak interaction potentials.
The layered potential formulas in (5.3.15) indicate that each composite particle has an
exchange radius R, which satisfies
ρ < R < ρ1 , (5.4.4)
5.4 Mechanisms of Subatomic Decays and Electron Radiations 313

where ρ is the radius of this particle and ρ1 is its repelling radius of weak force. If two
composite particles A and B are in a distance less than their common exchange radius,
then the weaktons in A and B are attracting to each other by their weak interacting forces
generated by Φw0 in (5.3.15), given by
E F
w −kr 1 Bw
Φ0 = gw e − (1 + 2kr)e −kr
. (5.4.5)
r ρw
There is a probability for these weaktons in A and B to recombine and form new particles.
Then, after the new particles have been formed, in the exchange radius R, the weak interact-
ing force between them is governed by the potentials of the new particles, and is repelling,
driving the newly formed particles apart.
For example, in Figure 5.10 we can clearly see how the weaktons in (5.4.3) undergo the
exchange process. When the randomly moving photons and ν -mediators, i.e. w 1 w1 , w2 w2
and νe ν e come into their exchange balls, they recombine to form an electron ν e w1 w2 and
a positron ν e w1 w2 under the weak interaction attracting forces generated by the weakton
potentials (5.4.5). Then, the weak interacting force between e − = νe w1 w2 and e+ = ν e w1 w2
governed by Φwl in (5.3.15) is repelling, and pushes them apart, leading to the decay process
(5.4.2)
We remark here that in this range the weak repelling force between e + and e− is much
stronger than the Coulomb attracting force. In fact, by (5.3.29), g 2w ∼ 1030 h̄c and the electric
1
charge square e2 = h̄c. Hence the weak repelling force in Figure 5.10 is (3g s )2 /r2 , much
137
stronger than e2 /r2 .

Figure 5.10

2. Weakton exchanges between leptons and mediators. The µ -decay reaction formula
is given by
µ − → e− + ν e + ν µ . (5.4.6)
314 Chapter 5 Elementary Particles

The complete formula for (5.4.6) is

µ − + ν → e− + ν e + ν µ ,

which is expressed in the weakton components as

νµ w1 w2 + νe ν e → νe w1 w2 + ν e + νµ . (5.4.7)

By the exclusive rule Le Lµ = 0, the two particles ν µ and ν e cannot be combined to form a
particle. Hence, ν e and νµ appear as independent particles, leading to the exchange of ν µ
and νe as in (5.4.7).

3. Weakton exchanges between quarks and mediators. The d-quark decay in (5.3.2) is
written as
d → u + e− + ν e . (5.4.8)

The correct formula for (5.4.8) is

d + γ + ν → u + e− + ν e ,

which, in the weakton components, is given by

w∗ w1 w2 + w1 w1 + νe ν e → w∗ w1 w1 + νe w1 w2 + ν e . (5.4.9)

In (5.4.9), the weakton pair w2 and w1 is exchanged, and νe is captured by the new doublet
w1 w2 , which is the vector boson W − , to form an electron νe w1 w2 .

5.4.2 Conservation laws

The weakton exchanges must obey certain conservation laws, which are listed in the fol-
lowing.

1. Conservation of weakton numbers. The weaktons given in (5.3.18) are elementary

particles, which cannot undergo any decay. Also, the w-weaktons cannot be converted
between each other. Although the neutrino oscillation, which is unconfirmed, may convert
one flavour of neutrino to another, for a particle decay or scattering, the neutrino number is
still conserved. Namely, the lepton numbers Le , Lµ , Lτ are conserved.
Therefore, for any particle reaction:

A1 + · · · + A N → B1 + · · · + B K , (5.4.10)

the number of each weakton type is invariant. Namely, for any type of weakton w,
5 its
number is conserved in (5.4.10):
NwA! = NwB! ,
5.4 Mechanisms of Subatomic Decays and Electron Radiations 315

where NwA! and NwB! are the numbers of the w

5 weaktons in both sides of (5.4.10).
2. Spin conservation. The spin of each weakton is invariant. The conservation of
weakton number implies that the spin is also conserved:

JA1 + · · · + JAN = JB1 + · · · + JBK ,

where JA is the spin of particle A.

In classical particle theories, the spin is not considered as a conserved quantity. The
reason for the non-conservation of spin is due to the incompleteness of the reaction formulas
given in Section 5.1.3. Hence, spin conservation can also be considered as an evidence for
the incompleteness of those reaction formulas. The incomplete decay reaction formulas can
be made complete by supplementing some massless mediators, so that the spin becomes a
conserved quantum number.
3. Energy conservation. Energy conservation is a universal physical law for all particle
systems. Hence, the energy conservation has also to be satisfied by the weakton exchanges.
This law is manifested in the following prohibitions for decays:

e− " µ − + ν µ + νe ,
e− " τ − + ν τ + νe , (5.4.11)
µ " τ + ν τ + νµ .
− −

The complete reaction formulas for (5.4.11) are

νe w1 w2 + νµ ν µ → νµ w1 w2 + ν µ + νe ,
νe w1 w2 + ντ ν τ → ντ w1 w2 + ν τ + νe , (5.4.12)
νµ w1 w2 + ντ ν τ → ντ w1 w2 + ν τ + νµ .

From the viewpoints of quantum rules given by (5.3.42) and weakton number conservation,
the decays in (5.4.12) are allowed. However, due to the mass relations

me < m µ < m τ ,

these weakton exchange processes (5.4.12) violate the energy conservation. Therefore these
decays cannot occur. However, if ν -mediators have high energy to hit electrons, then the
reactions (5.4.12) may occur.
In a weakton exchange process, the energy conservation law can be explicitly expressed
as follows:

Energy Rule 5.13 The composite particles with the lower masses cannot undergo
weakton exchanges with massless mediators in lower energy to decay into the composite
particles with higher masses.
316 Chapter 5 Elementary Particles

We remark that the Energy Rule 5.13 is sharper than the energy conservation law. In
fact, the reaction formulas from lower masses to higher masses do not imply that the energy
conservation must be violated. Hence, it is possible that there exist more basic unknown
physical laws under the Energy Rule 5.13.
4. Other conservative quantum numbers. From the invariance of weakton numbers, we
deduce immediately the following conserved quantum numbers:

electric charge Qe , weak charge Qw , strong charge Qs ,

baryon number B, lepton numbers Le , Lµ , Lτ .

5.4.3 Decay types

In particle physics, the reactions listed in Section 5.1.3 are classified into two types: the
weak interacting type and the strong interacting type. However, there are no clear definitions
to distinguish them. Usual methods are by experiments to determine reacting intensity, i.e.
the transition probability. In general, the classification is derived based on

weak type: i) presence of leptons in the reactions,

ii) change of strange numbers,
strong type: otherwise.

With the weakton model, all decays are carried out by exchanging weaktons. Hence
decay types can be fully classified into three types: the weak type, the strong type, and
the mixed type, based on the type of forces acting on the final particles after the weakton
exchange process.
For example, the reactions

ν µ + e− → µ − + νe ,
n → p + e− + ν e , (5.4.13)
0
π → γ + γ.

are weak decays,

∆++ → p+ + π + (5.4.14)
is a strong decay, and

Λ → p+ + π − ( i.e. Λ + gk + 3γ → p+ + π − + γ ) (5.4.15)

is a mixed decay.
In view of (5.4.13)-(5.4.15), the final particles contain at most one hadron in a weak
decay, contain no leptons and no mediators in a strong decay, and contain at least two
hadrons and a lepton or a mediator in a mixed decay.
5.4 Mechanisms of Subatomic Decays and Electron Radiations 317

In summary, we give the criteria for decay types based on the particles appearing in the
final step of the decay process:
weak decay: at most one hadron,
strong decay: no leptons and no mediators, (5.4.16)
mixed decay: others.
Remark 5.14 The new classification is easy to understand, and the criterion (5.4.16)
provides a convenient method for us to distinguish the different types of decays and scat-
terings. In fact, this classification truly reflects the roles of interactions in particle reaction
processes.

5.4.4 Decays and scatterings

Decays and non-elastic scatterings are caused by weakton exchanges. The massless media-
tors
γ , gk , γ0 , gk0 , ν , (5.4.17)
spread over the space or around the charged leptons and quarks at various energy levels, and
most of them are at low energy states. It is these random mediators in (5.4.17) entering the
exchange radius of matter particles that generate decays. In the following we shall discuss
these reaction processes for various types of decays and scatterings.
Weak types
1. First we consider the e-µ scattering
νµ + e− → µ − + νe ,
which is rewritten in the weakton components as
νµ + νe w1 w2 → w1 w2 (W − ) + νµ + νe (5.4.18)
→ νµ w1 w2 + νe .
In (5.4.18) we can see that the transient vector boson W − appears, and then captures ν µ to
form the muon µ . The Feynman digram of (5.4.18) is given by Figure 5.11.

Figure 5.11
318 Chapter 5 Elementary Particles

Replacing the Feynman diagram, we describe the scattering (5.4.18) by using Figure
5.12. It is clear that the scattering (5.4.18) is achieved by exchanging weaktons ν µ and νe .

2. β -decay. Consider the classical β -decay process

n → p + e− + ν e . (5.4.19)

with the quark constituents of n and p:

n = udd, p = uud,

the β -decay (5.4.19) is equivalent to the following d-quark decay:

d → u + e− + ν e ,

whose complete form should be given by

w∗ w1 w2 (d) + νe ν e (ν ) + w1 w1 (γ ) → w∗ w1 w1 (u) + w1 w2 (W − ) + νe ν e (ν ) (5.4.20)

→ w w1 w1 (u) + νe w1 w2 (e ) + ν e .
∗ −

In the β -decay (5.2.20), w∗ in d-quark and photon γ = w1 w1 recombine to form u-quark

and charged vector boson W − , then W − captures νe from ν -mediator to yield an electron
e− and an ν e .

Figure 5.12 e − µ scattering

3. Quark pair creations. Consider

gk + γ0 + γ → u + u,
gk0 + γ0 + γ0 → d + d.
5.4 Mechanisms of Subatomic Decays and Electron Radiations 319

They are rewritten in the weakton constituent forms as

w∗ w∗ # (gk ) + w1 w1 ↓↑ (γ0 ) + w1 w $ (γ ) (5.4.21)

∗ ∗
→ w w1 w1 ↑↓↑ (u) + w w1 w1 ↑$ (u),
w∗ w∗ ↑↓ (gk0 ) + w1 w1 ↑↓ (γ0 ) + w2 w2 ↓↑ (γ0 ) (5.4.22)
→ w w ↑↓ (gk0 ) + w1 w2 ↑↓ (H − ) + w1 w2
∗ ∗ +
↓↑ (H )
→ w∗ w1 w2 #↓ (d) + w∗ w1 w2 $↑ (d).

In (5.4.21), w∗ and w∗ in a gluon are captured by a scalar photon γ0 and a photon γ to create
a pair u and u. In (5.4.21), w1 and w2 in the two scalar photons γ0 are exchanged to form
a pair of charged Higgs H + and H − , then H + and H − capture w∗ and w∗ respectively to
create a pair of d and d.
4. Lepton decays. The lepton decays

µ − + ν → e− + ν e + ν µ ,
τ − + ν → µ − + ν µ + ντ ,

are rewritten in the weakton constituents as

νµ w1 w2 + νe ν e → νe w1 w2 + ν e + νµ ,
(5.4.23)
ντ w1 w2 + νµ ν µ → νµ w1 w2 + ν µ + ντ .
Here the neutrino exchanges form leptons in the lower energy states and a pair of neutrino
and antineutrino with different lepton numbers. By the exclusive rule L i L j = 0 (i ̸= j), the
generated neutrino and antineutino cannot be combined together, and are separated by the
weak interaction repelling force.
Strong types
1. ∆++ -decay. Consider the following decay

∆++ → p + π +.

The complete decay process should be

∆++ + gk0 + 2γ0 → p + π +,
(5.4.24)
∆++ + gk + 2γ → p + π +.
It is clear that the final particles are the proton p and meson π + . Hence by the criterion
(5.4.16), the ∆++ decay is a strong type. Recalling the weakton constituents in (5.3.19) and
the quark constituents of hadrons in Section 5.1.1, the first reaction formula in (5.4.24) is
rewritten as

3w∗ w1 w1 (∆++ ) + w∗ w∗ (gk0 ) + w1 w1 (γ0 ) + w2 w2 (γ0 ) (5.4.25)

→ (2w w1 w1 )(w w1 w2 )(p) + (w w1 w1 )(w w1 w2 )(π ).
∗ ∗ ∗ ∗ +
320 Chapter 5 Elementary Particles

This reaction process in (5.4.25) consists of two steps:

weakton exchanges: gk0 + 2γ0 → d + d, (5.4.26)

quark exchanges: uuu + dd → uud + ud. (5.4.27)

The exchange mechanism of (5.4.26) was discussed in (5.4.22), which is a weak interaction,
and the quark exchange (5.4.27) is both weak and strong interaction types. But, the final
particles p and π + are driven apart by the strong hadron repelling force.
After the weakton exchange to yield d-quark pair in (5.4.26), the mechanism of the
quark exchange and decay in (5.4.27) can be interpreted as follows:

1) When the quark pair dd is formed in the exchange radius R of ∆ ++ , the strong and
weak interactions between the quarks d, d and u, u, u in ∆++ are governed by Φsq in
(5.3.12) and Φwq in (5.3.15), which are attracting and recombine these quarks to form
two new hadrons p and π + , as shown in Figure 5.13(a); and

2) the two newly formed hadrons p and π + are controlled by the strong interaction
potential Φn for hadrons, which is repulsive in the exchange radius R and pushes
them apart; see Figure 5.13(b).

Figure 5.13

2. D0 -decay. Let us discuss the D0 -decay, which is considered as the weak interacting
type in the classical theory, because it violates the strange number conservation. But in our
classification it belongs to strong type of interactions. The D0 -decay is written as

D0 → K − + π + .

The complete formula is

D0 + g k + 2 γ → K − + π + .
The weakton constituents of this decay is given by

(w∗ w2 w2 )(w∗ w1 w1 )(cu) + w∗ w∗ (gk ) + 2w1 w1 (γ )

→ (w∗ w1 w2 )(w∗ w1 w1 )(cu) + (w∗ w1 w1 )(w∗ w1 w2 )(ud).
5.4 Mechanisms of Subatomic Decays and Electron Radiations 321

This reaction is due to the c-quark decay

c + gk + 2γ → s + u + d,

which is given in the weakton constituents as

w∗ w2 w2 (c) + w∗ w∗ (gk ) + 2w1 w1 (γ ) → w∗ w1 w2 (s) + w∗ w1 w1 (u) + w∗ w1 w2 (d). (5.4.28)

The reaction (5.4.28) consists of two exchange processes:

w∗ w2 w2 (c) + w1 w1 (γ ) → w∗ w1 w2 (s) + w1 w2 (w− ), (5.4.29)

and
w1 w2 (W − ) + w1 w1 (γ ) + (w∗ w∗ )(gk ) → w∗ w1 w1 (u) + w∗ w1 w2 (d) (5.4.30)

It is clear that both exchanges (5.4.29) and (5.4.30) belong to weak interactions. However,
the final particles in the D0 -decay are K − and π + , which are separated by the strong hadron
repelling force. The principle for the quark exchange and hadron separation in the D 0 -decay
process is the same as shown in Figure 5.13.

Remark 5.15 In the mechanism of subatomic particle decays and scatterings, the
layered properties of the weak and strong interacting forces play a crucial role. Namely the
attracting and repelling radii of weak and strong forces for a particle depend on its radius ρ
and interaction constants A and B or equivalently the parameters

A B
in (5.3.11) and in (5.3.14). (5.4.31)
ρ ρ

These parameters determine the attracting and repelling regions of the particle. Hence, in the
exchange radius of particles, the weak and strong forces are attracting between weaktons
and quarks, and are repelling between the final particles in decays and scatterings. The
reason is that the parameters in (5.4.31) are different at various levels of subatomic particles.

Mixed decays

The typical mixed decay is the Λ-decay, written as

Λ → p + π −.

The correct form of Λ-decay should be

Λ + gk + 2γ + γ0 → p + π − + γ0 . (5.4.32)
322 Chapter 5 Elementary Particles

There are three exchange processes in (5.4.32):

gk + γ + γ0 → u + u, (5.4.33)
s + γ → d + γ0 (i.e. uds + γ → udd + γ0 ), (5.4.34)
udd(n) + uu → uud(p) + ud(π ). −
(5.4.35)

The process (5.4.33) was described by (5.4.21), the quark exchange process (5.4.35) is clear,
and (5.4.34) is the conversion from s quark to d quark, described by

w∗ w1 w2 ↑$ (s) + w1 w1 # (γ ) → w∗ w1 w2 #↓ (d) + w1 w1 ↓↑ (γ0 ). (5.4.36)

Namely, (5.4.36) is an exchange of two w1 with reversed spins.

5.4.5 Electron structure

The weakton constituents of an electron are νe w1 w2 , which rotate as shown in Figure 5.9.
Noting the electric charges and weak charges of νe , w1 , w2 are given by

1 2
electric charge: Qνe = 0, Qwe 1 = − , Qwe 2 = − ,
3 3
weak charge: Qνw = 1, Qww1 = 1, Qww2 = 1,

we see that the distribution of weaktons νe , w1 , w2 in an electron is in an irregular triangle

due to the asymmetric forces on the weaktons by the electromagnetic and weak interactions,
as shown in Figure 5.14.

Figure 5.14 Electron structure

5.4 Mechanisms of Subatomic Decays and Electron Radiations 323

In addition, the weak force formula between the naked electron and massless mediators
γ , γ0 , gk , gk0 , ν , is given by
G H
∂ 1 −kr B5
Fw = − gw (ρm )gw (ρe ) e − (1 + 2kr)e−2kr , (5.4.37)
∂r r ρ5
G H
−kr 1 1 4B5 r −kr
=gw (ρm )gw (ρe )e + − e ,
r2 rr0 ρ5 r02

where k = 1/r0 = 1016 cm−1 , gw (ρm ) and gw (ρe ) are the weak charges of mediators and the
naked electron, expressed as
! "3 ! "3
ρw ρw
gw (ρm ) = 2 gw , gw (ρe ) = 3 gw ,
ρm ρe
5 ρ5 is a parameter determined by the naked electron and mediators.
and B/
By the weak force formula (5.4.37), there is an attracting shell region of weak interaction
between naked electron and mediators

ρ1 < r < ρ 2 (5.4.38)

as shown in Figure 5.13, with small weak force, where ρ j ( j = 1, 2) are the zero points of
(5.4.37). Namely, they satisfy that
! "
−kρ j 3 ρ5 ρj
e ρj = 1+ , for 1 ! j ! 2.
4B5 r0
In the region (5.4.38), the weak force is attracting, and outside this region the weak force is
repelling: .
< 0 for ρ1 < r < ρ2 ,
Fw (5.4.39)
> 0 for r < ρ1 and ρ2 < r.
Since the mediators γ , γ0 , gk , gk0 and ν contain two weak charges Qw = 2gw , by (5.4.39)
they are attached to the electron in the attracting shell region (5.4.38), forming a cloud of
mediators. The irregular triangle distribution of the weaktons νe , w1 , w2 generate a small
moment of force on the mediators. Meanwhile there also exist weak forces between them.
Therefore the bosons will rotate at a speed less than the speed of light, and generate a small
mass attached to the naked electron νe w1 w2 .

5.4.6 Mechanism of bremsstrahlung

It is known that an electron emits photons as its velocity changes, which is called the
bremsstrahlung. The reasons why bremsstrahlung can occur is unknown in classical the-
ories. Based on the electron structure theory established in the last subsection, we present
here a mechanism of this phenomenon.
324 Chapter 5 Elementary Particles

Figure 5.15 (a) The naked electron is accelerated or decelerated in electromagnetic field; amd (b)
the mediators (photons) fly away from the attracting shell region under a perturbation of moment of
force.

In fact, if an electron is situated in an electromagnetic field, then the electromagnetic

field exerts a Coulomb force on ithe naked electron νe w1 w2 , but not on the attached neutral
mediators. Thus, the naked electron changes its velocity, which draws the mediator cloud to
move as well, causing a perturbation to moment of force on the mediators. As the attracting
weak force in the shell region (5.4.38) is small, under the perturbation, the centrifugal force
makes some mediators in the cloud, such as photons, flying away from the attracting shell
region, and further accelerated by the weak repelling force outside this shell region to the
speed of light, as shown in Figure 5.15.

5.5 Structure of Mediator Clouds Around Subatomic Particles

5.5.1 Color quantum number
3
In the baryon family with spin J = 2 there are three members ∆++ , ∆− , Ω− , whose quark
constituents are as follows

∆++ = uuu(↑↑↑, ↓↓↓),

∆− = ddd(↑↑↑, ↓↓↓), (5.5.1)
Ω = sss(↑↑↑, ↓↓↓).
−

These quarks have the same spin arrangements and energy levels. The phenomenon that
there are three identical fermions at one quantum state violates the Pauli exclusion principle.
Thus, there exist two possibilities:

1) Pauli exclusion principle is invalid in the interior of baryons; or

2) there is a new quantum number, such that the three same flavour of quarks in the
baryons of (5.5.1) possess different values for the new quantum number. Namely,
they are not identical, and the Pauli exclusion principle is still valid.
5.5 Structure of Mediator Clouds Around Subatomic Particles 325

To solve this problem, O. W. Greenberg presented the color quantum number in 1964.
According to the constituents of (5.5.1), Greenberg thought that each flavor of quarks has
three colors:
red r, green g, blue b, (5.5.2)

and antiparticles have three anti-colors:

anti-red r, anti-green g, anti-blue b. (5.5.3)

Based on the quantum numbers (5.5.2) and (5.5.3), each flavor of quarks is endowed with
three different colors
qr , qg , qb , qr , qg , qb .

The color indices of hadrons are given by

baryon = q1r q2g q3b ,

(5.5.4)
meson = q1r q2r , q1g q2g , q1b q2b .

Since hadrons are color neutral, the color quantum number should obey the following mul-
tiplication rule:
rr = w, gg = w, bb = w,
(5.5.5)
rgb = w, r g b = w, w = white color,
and the multiplication is commutative.
The color quantum number has attained many experimental supports. Quantum Chro-
modynamics (QCD) for the strong interaction is based on this theory. In fact, it is natural to
think that the three color quantum states of each flavour of quarks

q = (qr , qg , qb ) (5.5.6)

are indistinguishable in the strong interaction. This gives rise to the SU(3) gauge theory of
QCD, i.e. the QCD action is the Yang-Milli functional of SU(3) gauge fields, given by

1
LQCD = − Skµν S µν k + q(iγ µ Dµ − m)q, (5.5.7)
4
where q is the quark triplet as in (5.5.6), m is the mass of the quark,
j
Skµν = ∂µ Sνk − ∂ν Skµ + gs fikj Siµ Sν ,
(5.5.8)
Dµ = ∂µ + igs Skµ λk ,

and λk (1 ! k ! 8) are the generators of SU(3). In QCD, λk are taken as the Gell-Mann
matrices as defined in (3.5.38).
326 Chapter 5 Elementary Particles

5.5.2 Gluons
In the last subsection we saw that the three color quantum numbers defined in (5.5.2)-(5.5.3)
lead to the QCD action (5.5.7) of SU(3) gauge fields. In the following, we introduce the
gluons in a few steps.

Gluons derived from QCD

It was known that the SU(3) gauge theory for the strong interaction is oriented toward
two directions: 1) describing the field particles, i.e. the strong interaction mediators, and 2)
providing strong interaction potentials. In Section 4.5, we have discussed problem 2). Here,
we consider problem 1).
Experiments showed that there are eight field particles for the strong interaction, called
gluons, which are massless and electric neutral, denoted by

gk , 1 ! k ! 8. (5.5.9)

The QCD theory shows that the gluons are vector bosons with spin J = 1.
Based on both the unified field theory and the weakton model, corresponding to the
vector gluons (5.5.9) there are eight dual field particles, which are scalar bosons with spin
J = 0, denoted by
gk0 , 1 ! k ! 8. (5.5.10)

In addition, by PID, it follows from (5.5.7)-(5.5.8) that the field equations describing the
eight vector gluons (5.5.9) read

gs k αβ i j gs 1
∂ ν Skµν − f g Sα µ Sβ − gsQkµ = (∂µ + αl Slµ − α0 xµ )φsk , (5.5.11)
h̄c i j h̄c 4
for 1 ! k ! 8, and the field equations describing the eight scalar gluons (5.5.10) are
E! " F
gs 1 gs
µ k
∂ ∂µ φs + ∂ µ
αl Sµ − α0 xµ φsk = −gs ∂ µ Qkµ − fikj gαβ ∂ µ (Sαi µ Sβj ), (5.5.12)
l
h̄c 4 h̄c

for 1 ! k ! 8, where αl (1 ! l ! 8) and α0 are parameters, and

Qkµ = qγµ λ k q (λ k = λk , γµ = g µν γ ν ).

The field equations (5.5.11) and (5.5.12) are as in (4.4.34) and (4.4.37).
From (5.5.11) and (5.5.12) we can deduce the following theoretical conclusions for the
gluons (5.5.9) and dual gluons (5.5.10):

1) The field functions Wµk (1 ! k ! 8) describing the gluons (5.5.9) are vector fields,
and, therefore, gk in (5.5.9) are vector bosons with spin J = 1;
5.5 Structure of Mediator Clouds Around Subatomic Particles 327

2) The field function φsk (1 ! k ! 8) describing the gluons (5.5.10) are scalar fields, and
therefore gk0 in (5.5.10) are scalar bosons with J = 0;

3) The field equations (5.5.11) and (5.5.12) are nonlinear. Consequently, the gluons g k
and gk0 are not in free states, and in their bound states the masses of g k and gk0 may
appear;

4) In the bound states, the masses can be generated by the spontaneous symmetry break-
ing in (5.5.11) and (5.5.12), and the mass terms for g k and gk0 are as follows
gs 5k l
gk : αl φs Sµ (φ5lk are the ground states),
h̄c
gk0 : α0 Φks ;

5) Free gluons, if there exist, are massless.

Remark 5.16 Gluons and photons are massless only in their free states. When they
are confined in mediator clouds around electrons and quarks, they may attain masses. In
this case, massive photons are described by the PID electromagnetic equations as follows:
Vector photons:
! "
e 1
∂ ν (∂ν Aµ − ∂µ Aν ) − eJµ = ∂µ + β Aµ − kxµ φ . (5.5.13)
h̄c 4

Scalar photons:
e 1
∂ µ ∂ µ φ − kφ + β ∂ µ (Aµ φ ) − kxµ ∂ µ φ = 0. (5.5.14)
h̄c 4
The masses of bound photons are created in (5.5.13) and (5.5.14) by the spontaneous gauge-
symmetry breaking.

Color charge
In the QCD theory, color quantum numbers are regarded as color charges. Namely, each
quark carries one of the color charges:

red charge r, green charge g, blue charge b,

(5.5.15)
anti-red charge r, anti-green charge g, anti-blue charge b.

Each gluon is considered to carry a pair of a color charge and an anti-color charge in QCD.
Thus, the three pair of color charges and anti-color charges in (5.5.15) constitute fundamen-
tal bases of SU(3) and SU(3), and the gluons gk (1 ! k ! 8) form an octet of the irreducible
representation
SU(3) ⊗ SU(3) = 8 ⊕ 1,
328 Chapter 5 Elementary Particles

which are expressed as

1
g1 = rg, g2 = br, g3 = gb, g4 = √ (rr − bb), (5.5.16)
2
5 6 7 8 4
g = gr, g = rb, g = bg, g = g . (5.5.17)

We remark that the conclusions (5.5.16) and (5.5.17) of QCD for gluons are completely
consistent with those in the weakton model in (5.3.40) and (5.3.41). However, the QCD
version is based on the color charges of (5.5.15), and the version of weakton model is based
on the color quantum number of the w∗ -weakton:

w∗r , w∗g , w∗b , wr∗ , wg∗ , wb∗ .

Gluon radiation
Based on the Standard Model, quarks can emit and absorb gluons in the same fashion as
electrons emitting and absorbing photons. For example, a red u-quark u r emits an rg gluon
grg , then ur becomes a ug quark, as shown in Figure 5.16.

Figure 5.16

The reaction formula of the gluon radiation shown by Figure 5.16 is as

ur → ug + grg , (5.5.18)

and the color index operation for (5.5.18) is given by

r = g(rg) = rgg = r. (5.5.19)

Color index operation

However, if qr -quark radiates a gbq gluon, due to the lack of color algebra, then it is
difficult to determine the type of quark to which qr transforms, i.e.

qr → gbg + qX , X =?.
5.5 Structure of Mediator Clouds Around Subatomic Particles 329

In addition, the lack of color algebra also leads to another serious problem. It is known
that hadrons are color neutral:

the color of a hadron = w.

But for the color index ci defined as

c1 = r, c2 = g, c3 = b, c1 = r, c2 = g, c3 = b, (5.5.20)

they is no algebraic operation, i.e.

ci c j ̸= any one of (5.5.20) for i ̸= j. (5.5.21)

Hence the color indices of all non-neutral gluons cannot be defined. Thus, if a proton p
contains three quarks and N gluons (N ̸= 0):

p = uci + uc j + dck + ∑ nk gk , ∑ nk = N,
k k

then we cannot make sure whether p is color neutral.

In the following two subsections we shall establish the color algebra, a new mathemat-
ical theory, to solve above mentioned problems associated with the color index operation
and the QCD.

5.5.3 Color algebra

The main objective of this and next subsections is to introduce a consistent color algebraic
structure, and to establish a color index formula for hadrons and color transformation ex-
change for gluon radiation.
The color algebra for quantum chromodynamics (QCD) established here is based on
color neutral principle of hadrons, and is uniquely determined. Hence it serves as the math-
ematical foundation of QCD.
Color algebra of color quantum numbers
First we examine the crucial problems encountered in the existing theory for color al-
gebra. The color neutral principle of hadrons requires that the three colors must obey the
following laws:

rgb = w, r g b = w, (5.5.22)
rr = gg = bb = w. (5.5.23)

Basic physical considerations imply that the product operation of color indices is com-
mutative and associative:

rg = gr, rb = br, bg = gb, rgb = (rg)b = r(gb).

330 Chapter 5 Elementary Particles

Hence we infer from (5.5.22) and (5.5.23) that

rg = b, rb = g, bg = r,
(5.5.24)
r g = b, rb = g, bg = r.

Notice that the white color w is the unit element, i.e.

wr = r, wg = g, wb = b, wr = r, wg = g, wb = b.

Then again, we infer from (5.5.22) and (5.5.23) that

rr = r, gg = g, bb = b,
(5.5.25)
r r = r, g g = g, b b = b.

Multiplying (5.5.22) by b and using (5.5.25), we deduce that

r(gb) = b, (5.5.26)

which leads to inconsistency, no matter what color we assign for gb. For example, if we
assign gb = r, then we derive from (5.5.26) that rr = b, which is inconsistent with rr = r in
(5.4.25). If we assume gb = b, then rb = b, which, by multiplying by b, leads to r = bb = b,
a contradiction again.
The above inconsistency demonstrates that in addition to the six basic color indices

r, g, b, r, g, b,

we need to incorporate the following color extensions:

rg, rg, rb, rb, gb, gb.

Only two of these added color indices are independent, and in fact, we can derive from
(5.5.24) that
rg = br = gb, gr = rb = bg.

Hence we define them as yellow y and anti-yellow y as follows:

y = rg = br = gb,
(5.5.27)
y = gr = rb = bg.

In a nutshell, in order to establish a consistent color algebra, it is necessary to add two

quantum numbers yellow y and anti-yellow y to the six color quantum numbers, give rise to
a consistent and complete mathematical theory: color algebra.

Definition 5.17 A color algebra is defined by the following three basic ingredients:
5.5 Structure of Mediator Clouds Around Subatomic Particles 331

1) The generators of color algebra consists of quarks and gluons, which possesses nine
color indices as
r, g, b, y, r, g, b, y, w,

which form a finite commutative group. Here y and y are given by (5.5.27), and the
group product operation is defined by

(a) w is the unit element, i.e.

cw = c f or any color index c;

(b) c is the inverse o f c:

cc = w f or c = r, g, b, y;

(c) in addition to the basic operations given by (5.5.22)-(5.5.25) and (5.5.27), we

have

yr = b, yg = r, yb = g, yy = y,
yr = g, yg = b, yb = r, y y = y.

2) Color algebra is an algebra with quarks and gluons as generators with integer coef-
ficients, and its space is given by
. / 0
22 /
P = ∑ nk ek // nk ∈ Z ,
k=1

where ek (1 ! k ! 19) are 18 colored quarks and 4 colored gluons, and −e k represent
anti-quarks and anti-gluons.
22
3) The color index of ω = ∑ nk ek ∈ P is defined by
k=1

22
Indc (ω ) = ∏ ck k ,
n

k=1

n
where ck is the color index of ek and ck k = w if nk = 0.

Two remarks are now in order.

Remark 5.18 Each element ω = Σnk ek ∈ P represents a particle system, and nk is the
difference between the number of particles with color index c k and the number of antiparti-
cles with color index ck . In particular, particles with colors r, g, b must be quarks, particles
with colors r, g, b must be anti-quarks, and particles with colors y, y must be gluons.
332 Chapter 5 Elementary Particles

Remark 5.19 In (5.5.27), rg, br, gb are all yellow. Consequently the gluons g rg , gbr , ggb
have the same color. However, they represent different quantum states. In particular, in the
weakton model,
grg = w∗r w∗g , gbr = w∗b w∗r , ggb = w∗g w∗b ,
which represent different quantum states.

Color index formula of hadrons

We now study color neutral problem for hadrons and the radiation and absorption of
gluons for quarks.
Let us start with color neutral problem for hadrons. Consider the constituents of a proton

p = uc1 + uc2 + uc3 + ∑ nk gk ∈ P, ∑ nk = N,

whose color index is given by
8
Indc (p) = c1 c2 c3 ∏ ( Indc (gk ))nk ,
k=1

where Indc (gk ) is the color index of the gluon gk . The color neutral law requires that
Indc (p) = w, which does not necessarily lead to c1 c2 c3 = w. For example, for the following
constituents of p:
p = ur + ur + dg + 2grg + grb ,
we have

Indc (p) = r2 gy2 y = rgy = rr = w,

c1 c2 c3 = rrg = rg = y ̸= w.

In summary, the hadron color quantum numbers based on color algebra is very different
from the classical QCD theory.
For a baryon B with its constituents given by
3 3
B = ∑ qci + ∑ (nk gk + mk gk ) + K1 g4 + K2 g4 ,
i=1 k=1

where gk and gk (1 ! k ! 4) are as in (5.5.16) and (5.5.17), its quantum number distribution
satisfies the following color index formula:

c1 c2 c3 = yN1 yN2 , (5.5.28)

and N1 , N2 are as
3 3
N1 = ∑ nk , N2 = ∑ mk . (5.5.29)
k=1 k=1
5.5 Structure of Mediator Clouds Around Subatomic Particles 333

For a meson M with constituents:

3
M = qc1 + qc2 + ∑ (nk gk + mk gk ) + K1 g4 + K2 g4 ,
k=1

its color indices satisfy that

c1 c2 = yN1 yN2 , (5.5.30)

where N1 , N2 are as in (5.5.29).

Both equalities (5.5.28) and (5.5.29) are the color index formulas for baryons and mesons,
which ensure the neutral law of hadrons.

Color transformation of gluon radiation

Consider the transformation of a quark qc with color c to another quart qc3 after emitting
a gluon gc1 c2 :
qc → gc1 c2 + qc3 ,

then the color index c3 of then induced quark qc3 is given by

c3 = cc1 c2 .

Also, for the transformation of a quark qc to another quark qc4 after absorbing a gluon gc1 c2 :

qc + gc1 c2 → qc4 ,

then the color c4 of transformed quark qc4 is as follows

c4 = cc1 c2 .

General color algebra

In an abstract sense, a color algebra is a triplet

{Gc , PN , Indc },

which consists of

1) a finite group Gc , called color group;

2) an integer modular algebra with generators e1 , · · · , eN :

. / 0
N /
P = ∑ n k ek / n k ∈ Z ;
N /
k=1
334 Chapter 5 Elementary Particles

3) a homomorphism
Indc : PN → Gc ,
such that for each element w = ∑Nk=1 nk ek ∈ PN , we can define a color index for w by
N
Indc (w) = ∏ ( Indc (ek ))nk ,
k=1

where Indc (ek ) ∈ Gc is the image of ek under the homomorphism Indc .

The color algebra introduced earlier is such a triplet {Gc , PN , Indc } with N = 22, and

Gc = multiplication group generated by {r, g, b, y},

. / 0
22 /
PN = ∑ nk ek // nk ∈ Z ,
k=1

where ek are colored quarks and gluons given by

e1 = u r , e2 = u g , e3 = u b , e4 = d r , e5 = d g , e6 = d b ,
e7 = sr , e8 = sg , e9 = sb , e10 = cr , e11 = cg , e12 = cb ,
(5.5.31)
e13 = br , e14 = bg , e15 = bb , e16 = tr , e17 = tg , e18 = tb ,
e19 = g1 , e20 = g2 , e21 = g3 , e22 = g4 ,

and gk (1 ! k ! 4) are as in (5.5.16). The homomorphism Indc is naturally defined by the

assignment in (5.5.31), i.e.

Indc (e1 ) = r, Indc (e2 ) = g, ··· , Indc (e18 ) = b,

Indc (e19 ) = Indc (e20 ) = Indc (e21 ) = y, Ind(e22 ) = w.

5.5.4 w∗ -color algebra

Based on the weakton model, the weakton constituents of a quark q and a gluon g k are given
by

quark : q = w∗ ww,
gluon: gk = w∗ w∗ .

The only weakton that has colors is w∗ , which has three colors:

w∗r , w∗g , w∗b , (5.5.32)

and three anti-colors:

w∗r , wg∗ , wb∗ . (5.5.33)
With the three pairs of colored and anti-colored weaktons in (5.5.32) and (5.5.33), we
can define N = 3 color algebra, which we call the w∗ -color algebra.
5.5 Structure of Mediator Clouds Around Subatomic Particles 335

Definition 5.20 The w∗ -color algebra is the triplet {Gc , P3 , Indc }, which is defined as
follows:
1) the color group is
Gc = {r, g, b, y, r, g, b, y, w},
with group operation given in Definition 5.17;
2) the Z-modular algebra P3 given by
. / 0
/
P3 = ∑ nk w∗k // nk ∈ Z ;
k=r,g,b

3) the color homomorphism Indc : P3 → Gc defined by

Indc (w∗k ) = k, Indc (−w∗k ) = k, k = r, g, b.

The w∗ -color algebra given by Definition 5.20 is based on the weakton model, which is
much simpler than the QCD color algebra introduced in the last subsection, and is readily
used to study the structure of subatomic particles.
Consider a particle system ω , which consists of nk quarks qk , nk antiquarks qk , m1 gluons
g1 = grg , m2 gluons g2 = gbr , m3 gluons g3 = ggb , m4 color-neutral gluons g4 , and mk gluons
gk (1 ! k ! 4):
4
ω= ∑ (nk qk + nk qk ) + ∑ (mi gi + mi gi ). (5.5.34)
k=r,g,b i=1

Then ω corresponds to an element Xω ∈ P3 expressed as

Xω = ∑ Nk w∗k , (5.5.35)
k=r,g,b

where
Nr = (nr − nr ) + (m1 − m2 ) − (m1 − m2 ),
Ng = (ng − ng ) + (m3 − m1 ) − (m3 − m1 ), (5.5.36)
Nb = (nb − nb ) + (m2 − m3 ) − (m2 − m3 ),
and, consequently, the color index for ω is defined by

Indc (ω ) = Indc (Xω ) = rNr gNg bNb , (5.5.37)

where for Nk < 0 we defined

−Nk
kNk = k (k = r, g, b).

It is then clear that

s
Indc (ω1 + · · · + ωs ) = ∏ Indc (ωi ).
i=1
∗
The following is a basic theorem for w -color algebra, providing the needed foundation
for the structure of charged leptons and quarks.
336 Chapter 5 Elementary Particles

Theorem 5.21 The following assertions hold true for the w∗ -color algebra:

1) For any gluon particle system with no quarks and antiquarks

4
π = ∑ (mi gi + mi gi ), (5.5.38)
i=1

the color index of this system satisfies that

⎧
⎪ 3
⎪
⎪
⎪
⎪
∑
⎪ w for (m − m ) = ±3n,
i i
⎪
⎪ i=1
⎨ 3
Indc (π ) = y for ∑ (mi − mi ) = ±3n + 1, (5.5.39)
⎪
⎪ i=1
⎪
⎪ 3
⎪
⎪
⎩ y for ∑ (mi − mi ) = ±3n + 2,
⎪
⎪
i=1

for some integer n = 0, 1, 2, · · · ;

2) For any single quark system as

ω = q+π
with π as given by (5.5.38), (5.5.40)
ω = q+π

the color index of (5.5.40) satisfies that

Indc (ω ) = r, g, b,
(5.5.41)
Indc (ω ) = r, g, b,

3) For the hadronic systems

M = q + q+ π meson system,
(5.5.42)
B = q+q+q+π baryon system,

with π given by (5.5.38), the color indices of (5.5.42) must be as

Indc (M) = w, y, y, Indc (B) = w, y, y. (5.5.43)

Two remarks are now in order.

First, in particle physics, all basic and important particle systems are given by the parti-
cle systems in Assertions 1)-3) in this theorem. System (5.5.38) represents a gluon system
attached to quarks and hadrons, (5.5.40) represents a cloud system of gluons around a quark
or an anti-quark, and (5.5.42) represents a cloud system of gluons around a hadron (a meson
or a baryon).
5.5 Structure of Mediator Clouds Around Subatomic Particles 337

Second, physically, the numbers

3
∑ (mi − mi ) = ±3n (n = 0, 1, 2, · · · ),
i=1

indicate that through exchange of weaktons, a cloud system (5.5.38) of gluons can become
a system consisting of white gluons and the same number of yellow and anti-yellow gluons.
Proof of Theorem 5.21 We proceed in the following three steps.

Step 1. By the basic properties (5.5.34)-(5.5.37) of color index,

Indc (π ) = rMr gMg bMb ,

where

Mr = (m1 − m2 ) − (m1 − m2 ),
Mg = (m3 − m1) − (m3 − m1 ),
Mb = (m2 − m3) − (m2 − m3 ).
m
Consequently, using C −m = C , we have

Indc (π ) = (rg)m1 (rb)m2 (gb)m3 (rg)m1 (rb)m2 (gb)m3 (5.5.44)

m1 m2 m3 m1 m2 m3
=y y y y y y
3
= yM , M = ∑ (mi − mi ).
i=1

Notice that y2 = y, y2 = y. Then the equality (5.5.39) follows from (5.5.44), and Assertion
1) is proved.

Step 2. For Assertion 2), with the above argument, for the particle system (5.5.40) it is
easy to see that
Indc (ω ) = Indc (q) Indc (π ) = Indc (q)yM ,
(5.5.45)
Indc (ω ) = Indc (q) Indc (π ) = Indc (q)yM .
Due to the facts that
Indc (q) = r, g, b, Indc (q) = r, g, b,

and by the multiplication rule given in Definition 5.17, the conclusion (5.5.41) follows from
(5.5.45).

Step 3. With the same arguments as above, for the meson and baryon system (5.5.42),
we can derive that
Indc (M) = ci c j yM , Indc (B) = ci c j ck yM , (5.5.46)
338 Chapter 5 Elementary Particles

for 1 ! i, j, k ! 3, where c1 = r, c2 = g, c3 = b. By the basic rules for the color operation

given in (5.5.24), (5.5.25) and (5.5.27), we have

ci c j = w, y, y, ci c j ck = ci cl = w, y, y.

Therefore, (5.5.43) follows from (5.5.46). &

5.5.5 Mediator clouds of subatomic particles

Subatomic particles include charged leptons, quarks and hadrons. As demonstrated in
(5.3.12) and (5.3.15), strong and weak interactions consist of different layers, leading to
different particle structures in various levels: weakton level (elementary particle level), me-
diator level, charged lepton and quark level, and hadron level.
In this section, we address the structure of mediator clouds for charged leptons, quarks
and hadrons, based on the the color algebra and the layered properties of strong and weak
interactions.
Mediator clouds for charged leptons
For simplicity and due to similarities, we consider only the case for electrons. In Section
5.4.5, we have introduced the structure of photon cloud around electrons and the mechanism
of photon radiations. Here we shall again discuss the electron structure in more details.
1. Clouds of photons and ν -mediators. First, an electron consists of a naked electron
and several layers of mediator clouds. Since electron does not contain strong charges, the
mediator clouds only contain photons and ν -mediators (each gluon contains 2g s ). Namely

mediator cloud of electron = photon layer + the ν -layer. (5.5.47)

According to the layered formulas (5.4.37) of weak forces between naked electron and
mediators, the radius r of a mediator cloud is approximately given by

ρ5 2
r3 ≃ r0 , r0 = 10−16 cm, (5.5.48)
4B5
and ρ5/B5 depends on the type of mediators. In fact, for gluons and naked electron, the
parameter B5 ! 0, therefore electrons have no gluon clouds.
It is reasonable to think that the parameter ρ5/B5 for photon and ν -mediator is different

ρ5γ ρ5ν
̸= .
B5γ B5ν

Therefore, by (5.5.48) the radius rγ of the photon layer is different from the radius rν of the
ν -layer, i.e.
rγ ̸= rν . (5.5.49)
5.5 Structure of Mediator Clouds Around Subatomic Particles 339

Due the weak repelling forces between γ and ν , only the exterior layer of cloud can
emit mediators. Hence, the bremsstrahlung implies that the photon cloud should be in the
exterior layer. Namely (5.5.49) should be written as

rγ > rν . (5.5.50)

The relation (5.5.40) is only a theoretical consideration.

2. Particle number of each layer. Due to the weak interaction between mediators, the
mediators must maintain a distance r0 between each other, with r0 being approximately the
weak repelling radius of the mediators. Consequently, the total number N of particles in
each layer of mediators satisfies that

N ! 4π r2 /r02 , (5.5.51)

where r is the radius of the layer as in (5.5.48). The inequality (5.5.51) means that all
particles on an r-sphere are arranged with distance r0 .
Based on (5.5.51), if r = kr0 , then N ! 4π k2 .
3. Spin number of vector photons. An electron refers to the system consisting of the
naked electron νe w1 w2 and its mediator cloud as shown in (5.5.47). The photon layer con-
1
tains both vector and scalar photons. As the total spin of an electron is J = , both the
2
number NJ=1 of vector photons with spin J = 1 and the number NJ=−1 of photons with spin
J = −1 should be the same, i.e. the vector photons in an electron satisfy

NJ=1 = NJ=−1 . (5.5.52)

The condition (5.5.52) implies that the emitting and absorbing of vector photons always
in pairs with one spin J = 1 and another spin J = −1. Scalar photons and ν -mediators have
spin J = 0.
In fact, by the Angular Momentum Rule 6.14, only bosons with J = 0 can rotate around
a center without the presence of force moment. Hence the photons and ν -mediators in the
mediator cloud may all be scalar particles.
Mediator clouds of quarks
Different from the weakton constituents of charged leptons, the weakton constituents of
a quark include a w∗ -weakton with a strong charge, leading to different structure of mediator
clouds.
1. As a quark curries three weak charges 3gw and a strong charge gs , the mediator cloud
of a quark possesses three layers:
gluon layer: vector and scalar gluons,
photon layer: vector and scalar photons, (5.5.53)
ν -layer: ν -mediators.
340 Chapter 5 Elementary Particles

2. The radii of the 3 layers in (5.5.53) are different:

rg ̸= rγ , rγ ̸= rν , rν ̸= rg . (5.5.54)

Similar to the case for an electron, the particle number in each layer of (5.5.53) satisfies
a relation as (5.5.51).
1
3. The total spin of a quark system is also J = . Therefore the vector photons and
2
vector gluons in the mediator cloud of a quark have to obey the spin number rules as

NJ=1 = NJ=−1 , 5J=1 = N

N 5J=−1 , (5.5.55)

5J=1
where NJ=1 and NJ=−1 are the numbers of photons with spins J = 1 and J = −1, and N
5J=−1 are that of gluons with spins J = 1 and J = −1.
and N
We remark that two scalar mediators can be transformed into a pair of vector particles
with J = 1 and J = −1 respectively.

4. By Assertion 2) of Theorem 5.21, the color indices of a quark system q and an

anti-quark system q, carrying gluon clouds, must take the values as follows

Indc (q) = r, g, b,
(5.5.56)
Indc (q) = r, g, b.

5. The gluons in the cloud layers of quarks are confined in a hadron. However, gluons
between quarks in a hadron can be exchanged. The gluon exchange process between quarks
in a hadron is called gluon radiation. In fact, a quark can emit gluons which will be absorbed
by other quarks in the hadron.

6. Consider a quark q1 , which is transformed to q3 after emitting a gluon g0 , and con-

sider a quark q2 , which is transformed to q4 after absorbing the gluon g0 . The exchange
process is expressed as
q1 → q3 + g0 , q2 + g0 → q4 .

Equivalently,
q1 + q2 → q3 + q4 .

The corresponding color transformation is given by

Indc (q1 ) = Indc (q3 ) Indc (g0 ),

Indc (q4 ) = Indc (q2 ) Indc (g0 ),

which lead to
Indc (q1 + q2) = Indc (q3 + q4). (5.5.57)
5.5 Structure of Mediator Clouds Around Subatomic Particles 341

The relations (5.5.53)-(5.5.57) provide the basic properties of mediator clouds of quarks.
Further discussion will be given in the next chapter.
Gluon clouds of hadrons
Different from electrons and quarks, there is no photon and ν -mediator cloud layer
around a hadron, due to the fact that the radius of a naked hadron ρ H is greater than the
range of the weak interaction:
ρH " 10−16 cm.
Namely, a hadron can only have a strong attraction to vector and scalar gluons, forming a
gluon cloud with radius about the same as the radius of a hadron:

rH ∼ 10−16 ∼ 10−14 cm.

Also, hadrons are colorless. For a baryon B and a meson M given by (5.5.42), we have that

Indc (M) = ci ci Indc (π ) = w

Indc (B) = ci c j ck Indc (π ) = w, 1 ! i, j, k ! 3.

Consequently, if the naked hadron is white, then the gluon cloud π is white as well: Indc (π ) =
w.
Free gluons
If there exist free gluons, then they must be white color:

Indc (g) = w for free gluons g.

It is now an unknown problem that if there are free gluons.

Chapter 6
Quantum Physics

6.1 Introduction
Quantum physics is the study of the behavior of matter and energy at molecular, atomic,
nuclear, and sub-atomic levels.
Quantum physics was initiated and developed in the first half of the 20th century, fol-
lowing the pioneering work of (Planck, 1901) on blackbody radiation, of (Einstein, 1905)
on photons and energy quanta, of Niels Bohr on structure of atoms, of (de Broglie, 1924) on
matter-wave duality. Quantum physics and general relativity have become the two corner-
stones of modern physics. We refer interested readers to (Sokolov, Loskutov and Ternov,
1966; Greiner, 2000; Sakurai, 1994), among many others, for the basics and history of
quantum physics.
Our recent work on the field theory of the four fundamental interactions and on the
weakton model of elementary particles has lead to new insights to a few issues and chal-
lenges in quantum physics, including in particular 1) the basic laws for interacting multi-
particle quantum systems, 2) energy levels of sub-atomic particles, and 3) solar neutrino
problem.
Field theory for interacting multi-particle systems
Based PID and PRI, we know now that the fundamental interactions are due the cor-
responding charges of the particles involved, with the mass charge for gravitational effect,
the electric charge for electromagnetism, the strong and the weak charges for the strong and
weak interactions respectively. The geometric interaction mechanism implies that the dy-
namic matter distribution of the charged particles changes the geometry of the space-time
manifold as well as the corresponding vector-bundles, leading to the dynamic interaction
laws of the system.
Also, due to PRI, for a multi-particle system consisting of subsystems of different levels,
the coupling can only be achieved through PRI and the principle of symmetry-breaking.
The above new insights from the unified field theory in Chapter 4 give rise to the fol-
lowing
6.1 Introduction 343

Postulates for interacting multi-particle systems

1) the Lagrangian action for an N-particle system satisfy the SU(N) gauge
invariance;
2) gGaµ represent the interaction potentials between the particles.
3) for an N-particle system, only the interaction field G µ can be measured,
and is the interaction field under which this system interacts with other
external systems.

With this postulate, field equations for a given multi-particle system can be naturally
established using PID and PRI. In particular, one can achieve the unification so that the
matter field can be geometrized as hoped by Einstein and Nambu, as stated explicitly in as
stated in Nambu’s Nobel lecture (Nambu, 2008).

Solar neutrino problem

Neutrino was first proposed by Wolfgang Pauli in 1930 in order to guarantee the energy
and momentum conservation for β -decay. In the current standard model of particle physics,
there are three flavors of neutrinos: the electron neutrino νe , the tau neutrino ντ and the mu
neutrino νµ . The solar neutrino problem is referred to the discrepancy of the number of
electron neutrinos arriving from the Sun are between one third and one half of the number
predicted by the Standard Solar Model, and was first discovered by (Davis, Harmer and
Hoffman, 1968).
The current dominant theory to resolve the solar neutrino problem is the neutrino os-
cillation theory, which are based on three basic assumptions: 1) the neutrinos are massive,
and, consequently, are described by the Dirac equations, 2) the three flavors of neutrinos
νe , νµ , ντ are not the eigenstates of the Hamiltonian, and 3), instead, the three neutrinos are
some linear combinations of three distinct eigenstates of the Hamiltonian. However, the
massive neutrino assumption gives rise two serious problems. First, it is in conflict with
the known fact that the neutrinos violate the parity symmetry. Second, the handedness of
neutrinos implies their velocity being at the speed of light.
To resolve these difficulties encountered by the classical theory, we argue that there
is no physical principle that requires that neutrino must have mass to ensure oscillation.
The Weyl equations were introduced by H. Weyl in 1929 to describe massless spin- 12 free
particles (Weyl, 1929), which is now considered as the basic dynamic equations of neutrino
(Landau, 1957; Lee and Yang, 1957; Salam, 1957); see also (Greiner, 2000). One important
property of the Weyl equations is that they violate the parity invariance. Hence by using the
Weyl equations, we are able to introduce a massless neutrino oscillation model. With this
massless model, we not only deduce the same oscillation mechanism, but also resolve the
above two serious problems encountered in the massive neutrino oscillation model.
344 Chapter 6 Quantum Physics

Despite of the success of neutrino oscillation models and certain level of experimental
support, the physical principles behind the neutrino oscillation are still entirely unknown.
Recently, the authors developed a phenomenological model of elementary particles, called
the weakton model (Ma and Wang, 2015b). The ν mediator in the weakton model leads to
an alternate explanation to the solar neutrino problem. When the solar electron neutrinos
collide with anti electron neutrinos in the atmosphere, which are abundant due to the β -
decay of neutrons, they can form ν mediators, causing the loss of electron neutrinos. Note
that ν mediator can also have the following elastic scattering
ν + e− −→ ν + e− .
Also ν participates only the weak interaction similar to the neutrinos, and consequently
possesses similar behavior as neutrinos. Consequently, the new mechanism proposed here
does not violate the existing experiments (SNO and KamLAND).
Energy levels of sub-atomic particles
The classical atomic energy level theory demonstrates that there are finite number of
energy levels for an atom given by En = E0 + λn , n = 1, · · · , N, where λn are the nega-
tive eigenvalues of the Schrödinger operator, representing the bound energies of the atom,
holding the orbital electrons, due to the electromagnetism.
The concept of energy levels for atoms can certainly be generalized to subatomic parti-
cles. The key ingredients and the main results are given as follows.
1
1. The constituents of subatomic particles are spin- fermions, which are bound to-
2
gether by either weak or strong interactions. Hence the starting point of the study is the
layered weak and strong potentials as presented earlier, which play the similar role as the
Coulomb potential for the electromagnetic force which bounds the orbital electrons moving
around the nucleons.
2. The dynamic equations of massless particles are the Weyl equations, and the dynamic
equations for massive particles are the Dirac equations. The bound energies of all subatomic
particles are the negative eigenvalues of the corresponding Dirac and Weyl operators, and
the bound states are the corresponding eigenfunctions.
1
The Weyl equations were introduced by H. Weyl in 1929 to describe massless spin-
2
free particles (Weyl, 1929), which is now considered as the basic dynamic equations of
neutrino (Landau, 1957; Lee and Yang, 1957; Salam, 1957); see also (Greiner, 2000).
3. With bound state equations for both massless and massive particles, we derive the
corresponding spectral equations for the bound states. We show that the energy levels of
each subatomic particle are finite and discrete:
0 < E1 < · · · < EN < ∞,
6.2 Foundations of Quantum Physics 345

and each energy level En corresponds to a negative eigenvalue λn of the related eigenvalue
problem. Physically, λn represents the bound energy of the particle, and are related to the
energy level En with the following relation:

En = E0 + λ n , λn < 0 for 1 ! n ! N. (6.1.1)

Here E0 is the intrinsic potential energy of the constituents of a subatomic particle such as
the weaktons.

4. One important consequence of the above derived energy level theory is that there are
both upper and lower bounds of the energy levels for all sub-atomic particles, and the largest
and smallest energy levels are given by

0 < Emin = E0 + λ1 < Emax = E0 + λN < ∞. (6.1.2)

In particular, it follows from the energy level theory that the frequencies of mediators
such as photons and gluons are also discrete and finite, and are given by ω n = En /h̄ (n =
1, · · · , N). In the Planck classical quantum assumption that the energy is discrete for a fixed
frequency, and the frequency is continuous. Our results are different in two aspects. One is
that the energy levels have an upper bound. Two is that the frequencies are also discrete and
finite.

Outline of Chapter 6

Section 6.2 presents the basic postulate and facts from quantum physics from the Hamil-
tonian dynamics and Lagrangian dynamics points of view. We prove also a new angular mo-
mentum rule, which is useful for describing the weaktons constituents of charged leptons
and quarks.
Section 6.3 presents new alternative approaches for solar neutrino problem, and is based
on the recent work of authors (Ma and Wang, 2014f).
Section 6.4 introduces energy levels of subatomic particles, and is based on (Ma and
Wang, 2014g).
The field theory and basic postulates for interacting multi-particle systems are estab-
lished in Section 6.5, which is based entirely on (Ma and Wang, 2014d).

6.2 Foundations of Quantum Physics

6.2.1 Basic postulates

The main components of quantum physics include quantum mechanics and quantum field
theory, which are based on the following basic postulates.
346 Chapter 6 Quantum Physics

Postulate 6.1 A quantum system consists of some micro-particles, which are described
by a set of complex value functions ψ = (ψ1 , · · · , ψN )T , called wave functions. In other
words, each quantum system is identified by a set of wave functions ψ :

a quantum system = ψ , (6.2.1)

which contain all quantum information of this system.

Postulate 6.2 For a single particle system described by a wave function ψ , its modular
square
|ψ (x,t)|2

represents the probability density of the particle being observed at point x ∈ R 3 and at time
t. Hence, ψ satisfies that
!
|ψ |2 dx = 1.
R3

Postulate 6.3 Each observable physical quantity L corresponds to an Hermitian

operator L̂, and the values of the physical quantity L are given by eigenvalues λ of L̂:

L̂ψλ = λ ψλ ,

and the eigenfunction ψλ is the state function in which the physical quantity L takes value
λ . In particular, the Hermitian operators corresponding to position x, momentum p and
energy E are given by

position operator : x̂ψ = xψ ,

momentum operator : p̂ψ = −ih̄∇ψ , (6.2.2)
∂ψ
energy operator : Ê ψ = ih̄ .
∂t

Postulate 6.4 For a quantum system ψ and a physical Hermitian operator L̂, ψ can
be expanded as !
ψ = ∑ αk ψk + αλ ψλ d λ , (6.2.3)

where ψk and ψλ are the eigenfunctions of L̂ corresponding to discrete and continuous

eigenvalues respectively. In (6.2.3) for the coefficients αk and αλ , their modular square
|αk |2 and |αλ |2 represent the probability of the system ψ in the states ψk and ψλ . In addi-
tion, the following integral, denoted by
!
⟨ψ |L̂|ψ ⟩ = ψ † (L̂ψ )dx, (6.2.4)

represents the average value of physical quantity L̂ of system ψ .

6.2 Foundations of Quantum Physics 347

Postulate 6.5 For a quantum system with observable physical quantities l 1 , · · · , lN , if

they satisfy a relation
R(l1 , · · · , lN ) = 0,
then the quantum system ψ (see (6.2.1)) satisfies the equation

R(L̂1 , · · · , L̂N )ψ = 0,

where L̂k are the Hermitian operators corresponding to lk (1 ! k ! N), provided that
R(L̂1 , · · · , L̂N ) is a Hermitian.

Two remarks are in order. First, in Subsection 2.2.5, Postulates 6.3and 6.5 are introduced
as Basic Postulates 2.22 and 2.23.
Second, in addition to the three basic Hermitian operators given by (6.2.2), the other
Hermitian operators often used in quantum physics are as follows:

angular momentum: L̂ = x̂ × p̂ = −ih̄⃗r × ∇,

spin operator: Ŝ = sh̄⃗σ ,
scalar momentum: h̄p0 = ih̄(⃗σ · ∇) ( massless fermion), (6.2.5)
scalar momentum: p̂1 = −ih̄(⃗α · ∇) ( massive fermion),
Hamiltonian energy : Ĥ = K̂ + V̂ + M̂,

where s is the spin, ⃗σ and ⃗α are as in (2.2.47) and (2.2.48), and K̂, V̂ , M̂ are the kinetic
energy, potential energy, mass operators.
In addition to these Hermitians in (6.2.2) and (6.2.5), the following 5 types of particle
current operators are very important in quantum field theory:

scalar current operator: I (identity),

pseudo-scalar current operator: γ 5 ,
vector current operator: γµ, (6.2.6)
µ 5
axial vector current operator: γ γ ,
i
tensor current operator: σ µν = [γ µ , γ ν ],
2
and the corresponding currents are

scalar current: ρ = ψ †ψ ,
pseudo-scalar current: P = ψ † γ 5 ψ ,
vector current: V µ = ψγ µ ψ , (6.2.7)
µ µ 5
axial vector current : A = ψγ γ ψ ,
tensor current: T µν = ψσ µν ψ ,
348 Chapter 6 Quantum Physics

where ψ = ψ † γ 0 , and γ µ (0 ! µ ! 3) and γ 5 are the Dirac matrices, which are expressed in
the forms:
! " ! "
I 0 0 σk
γ0 = , γk = for 1 ! k ! 3,
0 −I −σ k 0
! " (6.2.8)
5 0 1 2 3 0 I
γ = iγ γ γ γ = .
I 0
Remark 6.6 The particle currents defined in (6.2.7) are very important in the transition
theory of particle decays and scatterings. In fact, the general form of particle currents is
written as
JAB = ψ A γψB + h · c ( Hermitian conjugate),
where γ is a current operator in (6.2.7), ψA and ψB are wave functions of particles A and B.
For example, for the β -decay
n → p + e− + ν e ,
by the Fermi theory, the transition amplitude is
Gf
M = √ (ψ e γ µ ψνe )(ψ p γµ ψn ) + h · c,
2
where G f is the Fermi constant, and

γµ = g µν γ ν = (−γ 0 , γ 1 , γ 2 , γ 3 ).

6.2.2 Quantum dynamic equations

In quantum mechanics, the following are the four basic dynamic equations:
Schrödinger equation, governing particles at lower velocity
Klein-Gordon equations, describing bosons,
(6.2.9)
Weyl equations, describing massless fermions,
Dirac equations, governing massive fermions.
These four equations were initially derived in the spirit of Postulate 6.5. They can also
be equivalently obtained by the Principle of Lagrangian Dynamics (PLD) or by the Principle
of Hamiltonian Dynamics (PHD).
Although the three fundamental principles: Postulate 6.5, PLD, and PHD are equiva-
lent in describing quantum mechanical systems, they offer different perspectives. In the
following, we introduce the four dynamic equations based on the three principles.
Quantum dynamics based on Postulate 6.5
1. Schrödinger equation. In classical mechanics we have the energy-momentum relation
1 2
E= p + V, V is potential energy.
2m
6.2 Foundations of Quantum Physics 349

By the Hermitian operators in (6.2.2), this relation leads to the Schrödinger equation, written
as
∂ψ h̄2
ih̄ = − ∆ψ + V (x)ψ , (6.2.10)
∂t 2m
which is clearly is non-relativistic.
2. Klein-Gordon equation. The relativistic energy-momentum relation is given by

E 2 − c 2 p 2 = m2 c4 .

From this relation we can immediately derive the following Klein-Gordon equation:
! "
1 ∂2 mc
− 2 2 + ∆ ψ − ( )2 ψ = 0. (6.2.11)
c ∂t h̄
In the 4-dimensional vectorial form, the equation (6.2.11) is expressed as
I mc J2
∂ µ ∂µ ψ − ψ = 0,
h̄
which is clearly Lorentz invariant.
3. Weyl equations. For a massless particle, the de Broglie matter-wave relation gives

E = h̄ω , p0 = h̄/λ , c = ωλ ,

where p0 is a scalar momentum. It follows that

E = cp0 , (6.2.12)

By (6.2.5), corresponding to p0 the Hermitian operator reads

p̂0 = ih̄(⃗σ · ∇).

Thus, the relation (6.2.12) leads to the following Weyl equation:

∂ψ
= c(⃗σ · ∇)ψ , (6.2.13)
∂t
where ψ = (ψ1 , ψ2 )T is a two-component Weyl spinor. The equation (6.2.13) describes free
neutrinos.
4. Dirac equations. For a massive particle, the de Broglie matter-wave relation (6.2.12)
should be rewritten in the form
E = cp1 ± mc2 , (6.2.14)
where p1 is a scalar momentum for massive fermions, and by (6.2.5) the Hermitian operators
for p1 and m are given by
p̂1 = −ih̄(⃗α · ∇), m̂ = mα0 ,
350 Chapter 6 Quantum Physics

where ⃗α = (α1 , α2 , α3 ),
! " ! "
I 0 0 σk
α0 = , αk = , k = 1, 2, 3,
0 −I σk 0

and σk (1 ! k ! 3) are the Pauli matrices as in (3.5.36).

Thus, the de Broglie matter-wave relation (6.2.14) for massive fermions leads to the
following Dirac equations:

∂ψ
ih̄ = −ih̄c(⃗α · ∇)ψ + mc2α0 ψ , (6.2.15)
∂t
where ψ = (ψ1 , ψ2 , ψ3 , ψ4 )T is the Dirac spinor.
Multiplying both sides of (6.2.15) by the matrix α0 , then the Dirac equation is rewritten
in the usual form I mc J
iγ µ ∂µ − ψ = 0, (6.2.16)
h̄
where γ µ = (γ 0 , γ 1 , γ 2 , γ 3 ) is the Dirac matrices as in (6.2.8), with γ 0 = α0 , γ k = α0 αk (1 !
k ! 3).

Remark 6.7 Based on the spinor theory, the Weyl equation (6.2.13) and the Dirac
equations (6.2.16) are Lorentz invariant (see Section 2.2.6) and space rotation invariant. In
addition, the Dirac equations are invariant under the space reflection

x5 = −x, t = t.
5 (6.2.17)

In fact, under the reflection transformation (6.2.17), the Dirac spinor ψ transforms as

5 = γ 0ψ
ψ ( or ψ = γ 0 ψ
5 ). (6.2.18)

Thus (6.2.16) becomes

mc I mc J I mc J
iγ µ ∂5µ ψ
5− 5 = γ 0 iγ 0 ∂0 ψ − iγ 0 γ k γ 0 ∂k ψ −
ψ ψ = γ 0 iγ µ ∂µ − ψ = 0.
h̄ h̄ h̄
Here we used the identities

γ 0 γ 0 = I, γ 0 γ k = −γ k for 1 ! k ! 3.

Remark 6.8 Because the Weyl spinor has two-components, under the reflection trans-
formation (6.2.17), it is invariant:

x → −x ⇒ ψ → ψ (ψ the Weyl spinor).

Hence the Weyl equation is not invariant under the space reflection (6.2.17), which leads to
violation of parity conservation for decays and scatterings involving neutrinos. The viola-
tion of parity conservation was discovered by (Lee and Yang, 1956).
6.2 Foundations of Quantum Physics 351

Quantum dynamics based on PLD

Due to PLD, as the Lagrangian action of a quantum system ψ is given by

!
L= L (ψ , Dψ )dxµ ,

then its dynamic equation is derived by

δ
L = 0. (6.2.19)
δψ

Hence, we only need to give the Lagrangian action for each of (6.2.9)

1. Schrödinger systems. For the Schrödinger equation (6.2.10), its action takes the form
! " ! "
1 ∂ ψ ∗ ∂ ψ∗ 1 h̄2
Ls = ih̄ ψ − ψ − |∇ψ |2 + V |ψ |2 . (6.2.20)
2 ∂t ∂t 2 2m

2. Klein-Gordon systems. The action for the Klein-Gordon equation (6.2.11) is as fol-
lows
1 1 I mc J2
LKG = ∂ µ ψ ∗ ∂µ ψ + |ψ |. (6.2.21)
2 2 h̄

3) Weyl systems. The action for the Weyl equation (6.2.13) reads

Lw = ψ † σ µ ∂ µ ψ , (6.2.22)

where σ 0 = −I, (σ 1 , σ 2 , σ 3 ) = ⃗σ .

4. Dirac systems. The action for the Dirac equation (6.2.16) is

I mc J
LD = ψ iγ µ ∂µ − ψ. (6.2.23)
h̄
Remark 6.9 Historically the four basic quantum dynamic equations were discovered
all based on Postulate 5.5, and then their actions (6.2.20)-(6.2.23) were established by the
known corresponding equations. However, the Lagrange actions have played crucial roles
in developing the interaction field theories.

Quantum dynamics based on PHD

In Section 2.6.4, we derived all quantum dynamic equations from the Principle of Hamil-
tonian Dynamics (PHD). PHD amounts to saying that each conservation system is charac-
terized by a set of conjugate fields

ψk , ϕk , 1 ! k ! N, (6.2.24)
352 Chapter 6 Quantum Physics

and the conjugate fields (6.2.24) define the Hamiltonian energy

!
H= H (ψ , ϕ )dx. (6.2.25)
R3

Then the dynamic equations of this system are given by

∂ ψk δ
= H,
∂t δ ϕk
(6.2.26)
∂ ϕk δ
=− H,
∂t δ ψk
where ψ = (ψ1 , · · · , ψN ), ϕ = (ϕ1 , · · · , ϕN ) are as in (6.2.24), and H is the Hamiltonian
energy.
In the classical quantum mechanics, if the dynamic equation of a quantum system is in
the form
∂ψ
ih̄ = Ĥ ψ , Ĥ an Hermitian operator, (6.2.27)
∂t
then Ĥ is called the Hamiltonian operator of this system, and the physical quantity
!
H = ⟨ψ |Ĥ|ψ ⟩ = ψ † Ĥ ψ dx (6.2.28)
R3

is the Hamilton energy of the system.

Three systems: the Schrödinger system, the Weyl system and the Dirac system, are in
the form of (3.127) and (3.128). Namely the equations can be expressed in the form (3.127),
with the Hamiltonian operators given by

h̄2
Ĥ = − ∆ + V(x) for Schrödinger
2m
Ĥ = ih̄c(⃗σ · ∇) for Weyl spinor fields, (6.2.29)

Ĥ = −ih̄c(⃗α · ∇) + mc2α0 for Dirac spinor fields,

and the associated Hamiltonian energies given by

! E 2 F
h̄ 2 2
H= |∇ψ | + V |ψ | dx for Schrödinger system,
R3 2m
! K L
H= ih̄cψ † (⃗σ · ∇)ψ dx for Weyl System, (6.2.30)
3
!R K L
H= −ih̄cψ † (⃗α · ∇)ψ + mc2 ψ † α0 ψ dx for Dirac system.
R3

Here the conjugate fields are the real and imaginary parts of ψ = ψ 1 + iψ 2 .
By applying the PHD, from the Hamiltonian energies in (6.2.30) we can derive the
dynamic equations of the three systems, which are equivalent to the form (6.2.27); see
(2.6.49)-(2.6.55).
6.2 Foundations of Quantum Physics 353

However the Klein-Gordon system is different. In fact, we can write the equation
(6.2.11) in the form ! " ! "
∂ ψ ψ
ih̄ = Ĥ , (6.2.31)
∂t ϕ ϕ

where Ĥ is given by
! "
0 1
Ĥ = . (6.2.32)
−h̄2 c2 ∆ + m2c4 0

However, it is clear that the operator Ĥ of (6.2.32) is not Hermitian, and therefore Ĥ is not
a Hamiltonian. Consequently the quantity
! K L
⟨Φ|Ĥ|Φ⟩ = ψ ∗ ϕ + ϕ ∗ (−h̄2 c2 ∆ψ + m2 c4 ψ ) dx
R3

is also not a physical quantity because it is not a real number in general. In other words,
under the theoretic frame based on Postulate 6.5 and PLD, the Klein-Gordon equation can
not be regarded as a model to describe a conservation quantum system.
With PHD, we can, however, show that the Klein-Gordon equation is a model for a con-
served system. As seen in (2.6.57), we take a pair conjugate fields (ψ , ϕ ) and the Hamilto-
nian energy
! E F
1 2 2 2 m2 c 4 2
H= ϕ + c |∇ψ | + 2 |ψ | dx. (6.2.33)
2 R3 h̄
Then the Klein-Gordon equation (6.2.11) can be rewritten as
⎛ ⎞
δ
⎜ δψ H 0
! " ! "
∂ ψ ⎟ ψ
= J⎜ ⎟ , (6.2.34)
∂t ϕ ⎝ δ ⎠ ϕ
0 H
δϕ

where H is as in (6.2.33), and ! "

0 1
J= . (6.2.35)
−1 0

The Hamiltonian operator Ĥ for the Klein-Gordon system reads

⎛ 2 4 ⎞
−c 2∆ + m c 0 ⎠
Ĥ = δ H = ⎝ h̄2 . (6.2.36)
0 1

The model of (6.2.33)-(6.2.36) is in a standard form of PHD. In fact, all conservation

quantum systems, including the four classical systems (6.2.9), can be expressed in the stan-
dard PHD from, which we call the Quantum Hamiltonian Dyanmics (QHD), which is stated
in the following.
354 Chapter 6 Quantum Physics

Principle 6.10 A conserved quantum system can be described by a set of conjugate

fields
Ψ = (ψ1 , · · · , ψN ), Φ = (ϕ1 , · · · , ϕN ),

and an associated Hamiltonian energy H = H(Ψ, Φ), such that the dynamic equations of
this system are in the form
! " ! "
∂ Ψ 0 I
= J Ĥ(Ψ, Φ), J= , (6.2.37)
∂t Φ −I 0

where Ĥ = δ H is the variational derivative operator of the Hamiltonian energy H, I is

the N-th order identity matrix. In particular, if the Hamiltonian H is invariant under the
transformation of conjugate fields
! " ! "! "
ψk cos θ − sin θ ψk
→ for 1 ! k ! K,
ϕk sin θ cos θ ϕk

then the conjugate fields constitute complex valued wave functions ψ k +iϕk (or ϕk +iψk ), 1 !
k ! K, for this system.

Remark 6.11 In classical quantum mechanics, the Klein-Gordon equation encounters

a difficulty as mentioned in (6.2.31) that the operator Ĥ of (6.2.32) is not Hermitian, incon-
sistent with quantum mechanical principles for describing bosonic behaviors. But, under the
QHD model (6.2.37) this difficulty is solved. In the Angular Momentum Rule in Section
6.2.4 and the spinor BEC (nonlinear quantum system) in Chapter 7, we can see this point
clearly.

6.2.3 Heisenberg uncertainty relation and Pauli exclusion principle

The Heisenberg uncertainty relation and Pauli exclusion principle are two important quan-
tum physical laws.

Uncertainty principle

We first give this relation, which is stated as follows.

Uncertainty Principle 6.12 In a quantum system, the position x and momentum p, the
time t and energy E satisfy the uncertainty relations given by

1 1
∆x∆p " h̄, ∆t∆E " h̄, (6.2.38)
2 2
where ∆A represents a measuring error of A value. Namely, (6.2.38) implies that x and p,t
and E can not be precisely observed at the same moment.
6.2 Foundations of Quantum Physics 355

The relations (6.2.58) can be deduced from Postulates 6.3 and 6.4. The average values
of position and momentum are
!
⟨x⟩ = ψ ∗ xψ dx,
!
∂ψ
⟨p⟩ = − ψ ∗ ih̄ dx.
∂x
In statistics, the error to an average value is expressed by the squared deviation. Namely,
the errors to ⟨x⟩ and ⟨p⟩ are
!
⟨(∆x)2 ⟩ = ψ ∗ (x − ⟨x⟩)2 ψ dx = ⟨x2 ⟩ − ⟨x⟩2 ,
!
⟨(∆p)2 ⟩ = ψ ∗ (px − ⟨px ⟩)2 ψ dx = ⟨p2x ⟩ − ⟨px ⟩2 .

Assume that ⟨x⟩ = 0 and ⟨px ⟩ = 0. Then we have

!
⟨(∆x)2 ⟩ = ⟨x2 ⟩ = x2 |ψ |2 dx,
! ! " (6.2.39)
∂ 2
⟨(∆p)2 ⟩ = ⟨p2x ⟩ = ψ ∗ −ih̄ ψ dx.
∂x
To get the relation (6.2.38) we consider the integral
! ! "! "
∂ ψ∗ ∂ψ
I(α ) = α xψ ∗ + α xψ + dx, (6.2.40)
∂x ∂x
where α is a real number. The integral (6.2.40) can be written as

I(α ) = Aα 2 − Bα + C,

where
!
A= x2 |ψ |2 dx = ⟨(∆x)2 ⟩ (by (6.2.39)),
! !
∂
B=− x |ψ |2 dx = |ψ |2 dx = 1,
∂x
! ! ! "
∂ ψ∗ ∂ ψ 1 ∂ 2 1
C= dx = 2 ψ ∗ −ih̄ ψ dx = 2 ⟨(∆p)2 ⟩ ( by (6.2.39)).
∂x ∂x h̄ ∂x h̄
(6.2.41)
It is clear that A, B,C > 0, and by (6.2.40),

I(α ) " 0, ∀α ∈ R1 . (6.2.42)

Let α0 be the minimal of I(α ). Then α0 satisfies

B
I ′ (α0 ) = 2Aα0 − B = 0 ⇒ α0 = .
2A
356 Chapter 6 Quantum Physics

Inserting α0 = B/4A in (6.2.42) we get that

1
AC " B2 .
4
It follows from (6.2.41) that
h̄2
⟨(∆x)2 ⟩⟨(∆p)2 ⟩ " ,
4
which is the first relation of (6.2.38). The second relation (6.2.38) can be derived in the
same fashion.
The Heisenberg uncertainty relation (6.2.38) has profound physically implications, some
of which are listed as follows:
1) If a particle A consists of N more fundamental particles Ai (1 ! i ! N), in a small ball
with radius r,
A = A1 + · · · + AN ,
then the momentum p of each particle Ai is at least
p " h̄/4r.
Hence, the smaller the composite particle, the greater the bounding energy is needed
to hold its constituents together.

2) Most particles are unstable, and their lifetime τ is very short. In addition, the energy
distribution of each particle is in a range, called the energy width Γ. By the uncertainty
relation, τ and Γ satisfy that
τ Γ " h̄/2.
This relation is very important in experiments, because the width Γ is observable
which determines the lifetime of a particle by the uncertainty relation τ ≃ h̄/2Γ.

3) Uncertainty relations (6.2.38) also imply that the energy and momentum conserva-
tions may be violated in a small scale of time and space. Both conservations are only
the averaged results in larger scale ranges of time and space.

Pauli exclusion principle

We recall that particles are classified two types:
n
fermions = particles with spin J = for odd n,
2
bosons = particles with spin J = n for integer n.
Fermions and bosons display very different characteristics. The fermions do not like to
live together with the same fermions, but bosons are sociable particles. This difference is
characterized by the Pauli exclusion principle.
Pauli Exclusion Principle 6.13 In a quantum system, there are no two or more fermions
living in the same quantum states, i.e. possessing entirely the same quantum numbers.
6.2 Foundations of Quantum Physics 357

6.2.4 Angular momentum rule

In Section 5.3.2 we introduced the Angular Momentum Rule, which was first proved in
1
(Ma and Wang, 2015b). It says that only the fermions with spin J = can rotate around a
2
central force field. In fact, this rule can be generalized to scalar bosons, i.e. particles with
spin J = 0. In this subsection we shall discuss the rule in more details. To this end, we first
introduce conservation laws of quantum systems based on Principle 6.10.
Conservation laws based on quantum Hamiltonian dynamics (QHD)
Let H be the Hamiltonian energy of a conservative quantum system, which can be de-
scribed by the following Hamiltonian equations:
∂Ψ
= ĤΦ (Ψ, Φ),
∂t
(6.2.43)
∂Φ
= −ĤΨ (Ψ, Φ),
∂t
where
δH δH
ĤΦ = , ĤΨ = .
δΦ δΨ
Let L be an observable physical quantity with the corresponding Hermitian operator L̂
for the conjugate fields (Ψ, Φ)T of (6.2.43), and L̂ is expressed as
! "
L̂11 L̂12
L̂ = , L̂T12 = L̂∗21 .
L̂21 L̂22

Then the physical quantity L of system (6.2.43) is given by

! ! " ! K
† † Ψ L
L = (Ψ , Φ )L̂ dx = Ψ† L̂11 Ψ + Φ† L̂22 Φ + 2Re(Ψ† L̂12 Φ) dx. (6.2.44)
Φ

It is clear that the quantity L of (6.2.44) is conserved if for the solution (Ψ, Φ) T of
(6.2.43) we have
dL
= 0,
dt
which is equivalent to
! M
ĤΦ† L̂11 Ψ + Ψ†L̂11 ĤΦ − ĤΨ† L̂22 Φ − Φ† L̂22 ĤΨ (6.2.45)

+ 2Re(ĤΦ† L̂12 Φ − Ψ† L̂12 ĤΨ )]dx = 0.

We remark here that if the QHD is described by a complex valued wave function:

ψ = Ψ + iΦ,
358 Chapter 6 Quantum Physics

and its dynamic equation is linear, then (6.2.43) can be written as

!
∂ψ
ih̄ = Ĥ ψ , H = ψ † Ĥ ψ dx. (6.2.46)
∂t
In this case, the physical quantity L in (6.2.44) is in the form
!
L= ψ † L̂ψ dx, (6.2.47)

and the conservation law (6.2.45) of L is equivalent to

L̂Ĥ − Ĥ L̂ = 0. (6.2.48)

The formulas (6.2.46)-(6.2.48) are the conservation laws of the classical quantum mechan-
ics.
Hence, the conservation laws in (6.2.45) are the generalization to the classical quantum
mechanics, and are applicable to all conservative quantum systems, including the Klein-
Gordon systems and nonlinear systems.

Angular momentum rule

From the conservation laws (6.2.45) and (6.2.48) we can deduce the following Angular
Momentum Rule.
1
Angular Momentum Rule 6.14 Only the fermions with spin J = and the bosons
2
with J = 0can rotate around a center with zero moment of force. The particles with J ̸=
1
0, will move on a straight line unless there is a nonzero moment of force present.
2
In the following we give a mathematical derivation of the Angular Momentum Rule.

1. Fermions. Consider fermions which obey the Dirac equations as (6.2.46) with the
Hamiltonian
Ĥ = −ih̄c(α k ∂k ) + mc2 α 0 + V (r), (6.2.49)

where V is the potential energy of a central field, and α 0 , α k (1 ! k ! 3) are the Dirac
matrices ! " ! "
I 0 0 σk
α0 = , αk = for 1 ! k ! 3, (6.2.50)
0 −I σk 0

and σ k are the Pauli matrices.

The total angular momentum Jˆ of a particle is

Jˆ = L̂ + sŜ,
6.2 Foundations of Quantum Physics 359

where s is the spin, L̂ is the orbital angular momentum

L̂ = (L̂1 , L̂2 , L̂3 ) = r̂ × p̂, p̂ = −ih̄∇,

L̂1 = −ih̄(x2 ∂3 − x3 ∂2 ),
(6.2.51)
L̂2 = −ih̄(x3 ∂1 − x1 ∂3 ),
L̂3 = −ih̄(x1 ∂2 − x2 ∂1 ),

and Ŝ is the spin operator

! k "
σ 0
Ŝ = (Ŝ1 , Ŝ2 , Ŝ3 ), Ŝk = h̄ for 1 ! k ! 3. (6.2.52)
0 σk

By (6.2.49)-(6.2.52), we see that

Ĥ L̂1 − L̂1 Ĥ =h̄2 c[(x2 ∂3 − x3 ∂2 )(α 2 ∂2 + α 3 ∂3 ) − (α 2 ∂2 + α 3 ∂3 )(x2 ∂3 − x3∂2 )]

=h̄2 c[α 2 ∂3 (x2 ∂2 − ∂2 x2 ) − α 3 ∂2 (x3 ∂3 − ∂3 x3 )].

Notice that
x2 ∂2 − ∂2 x2 = x3 ∂3 − ∂3 x3 = −1.

Hence we get
Ĥ L̂1 − L̂1 Ĥ = h̄2 c(α 3 ∂2 − α 2 ∂3 ). (6.2.53)

Similarly we have
Ĥ L̂2 − L̂2 Ĥ = h̄2 c(α 1 ∂3 − α 3 ∂1 ),
(6.2.54)
Ĥ L̂3 − L̂3 Ĥ = h̄2 c(α 2 ∂1 − α 1 ∂2 ).
On the other hand, we infer from (6.2.49) and (6.2.52) that
M N
Ĥ Ŝ j − Ŝ j Ĥ = −ih̄2 cγ 5 ∂k (σ k σ j − σ j σ k ) = −ih̄2 cγ 5 (2iεk jl σ l )∂k = 2h̄2 cεk jl α l ∂k ,

where γ 5 is defined by ! "

5 0 1 2 30 I
γ = iγ γ γ γ = .
I 0
Hence we have
Ĥ Ŝ1 − Ŝ1 Ĥ = −2h̄2 c(α 3 ∂2 − α 2 ∂3 ),
Ĥ Ŝ2 − Ŝ2 Ĥ = −2h̄2 c(α 1 ∂3 − α 3 ∂1 ), (6.2.55)
Ĥ Ŝ3 − Ŝ3 Ĥ = −2h̄2 c(α 2 ∂1 − α 1 ∂2 ).

For Jˆ = L̂ + sŜ, we derive from (6.2.53)-(6.2.55) that

1
Ĥ Jˆ− JˆĤ = 0 ⇐⇒ spin s = . (6.2.56)
2
360 Chapter 6 Quantum Physics

When fermions move on a straight line,

Ĥ = cα 3 p3 , L̂ = 0.

In this case, by (6.2.53)-(6.2.54), for straight line motion,

Ĥ Jˆ− JˆĤ = 0 for any s. (6.2.57)

Thus, by the conservation law (6.2.48), the assertion of Angular Momentum Rule for fermions
follows from (6.2.56) and (6.2.57).
2. Bosons. Now, consider bosons which bey the Klein-Gordon equation in the form
(6.2.43). It is known that the spins J of bosons depend on the types of Klein-Gordon fields
(Ψ, Φ): ⎧
⎪
⎪ a scalar field ⇒ J = 0,
⎪
⎪
⎪
⎪ a 4-vector field ⇒ J = 1,
! " ⎪ ⎨
Ψ
= a 2nd-order tensor field ⇒ J = 2, (6.2.58)
Φ ⎪
⎪
⎪
⎪ a real valued field ⇒ neutral bosons,
⎪
⎪
⎪
⎩
a complex valued field ⇒ charged bosons.
For the Klein-Gordon fields (Ψ, Φ)T , the Hamiltonian for a central force field is given
by
! E F
1 2 2 1 2 4
2 2
H= |Φ| + c |∇Ψ| + 2 (m c + V (r))|Ψ| dx (6.2.59)
2 h̄
The Hamiltonian energy operator Ĥ of (6.2.59) is given by
! " E F
ĤΨ 0 1
Ĥ = , ĤΦ = Φ, ĤΨ = −c2 ∆ + 2 (m2 c4 + V ) Ψ. (6.2.60)
0 ĤΦ h̄
The angular momentum operator Jˆ is
! "
L̂ 0
Jˆ = + sh̄σ̂ , σ̂ = (σ 1 , σ 2 , σ 3 ). (6.2.61)
0 L̂

where s is the spin of bosons, and L̂ is as in (6.2.51).

For scalar bosons, spin s = 0 in (6.2.61) and the Hermitian operators in the conservation
law (6.2.45) are

L̂11 = L̂22 = L̂, L̂12 = L̂21 = 0, ĤΦ , ĤΨ as in (6.2.60)

Then by
! !
Φ† L̂Ψdx = − Ψ∗ L̂† Φdx,
! !
ĤΨ† L̂Φdx = − Φ† L̂ĤΨ dx,
6.3 Solar Neutrino Problem 361

we derive the conservation law (6.2.45), i.e.

! M N
Φ† L̂Ψ + Ψ† L̂Φ − ĤΨ† L̂Φ − Φ† L̂ĤΨ dx = 0.

However, it is clear that Ĥ and Jˆ in (6.2.60) and (6.2.61) don’t satisfy (6.2.45) for spin s ̸= 0.
Hence the quantum rule of angular momentum for bosons holds true.

Remark 6.15 The Angular Momentum Rule is very useful in the weakton model and
the theory of mediator cloud structure of charged leptons and quarks, which explain why all
stable fermions with mediator clouds are at spin J = 1/2.

6.3 Solar Neutrino Problem

6.3.1 Discrepancy of the solar neutrinos

The solar neutrino problem is known as that the number of electron neutrinos arriving from
the Sun are between one third and one half of the number predicted by the Standard Solar
Model. This important discovery was made in 1968 by R. Davis, D. S. Harmer and K. C.
Hoffmann (Davis, Harmer and Hoffman, 1968).
To understand this problem clearly, we begin with a brief introduction to the Standard
Solar Model, following (Griffiths, 2008).
In the nineteenth century, most physicist believed that the source of the Sun’s energy was
gravity. However, based on this assumption, Rayleigh showed that the maximum possible
age of the Sun was substantially shorter than the age of the earth estimated by geologists.
At the end of the nineteenth century, Bacquerel and Curies discovered radioactivity, and
they noted that radioactive substances release a large amounts of heat. This suggested that
nuclear fission, not gravity, might be the source of the Sun’s energy, and it could allow
for a much longer lifetime of the Sun. But, the crucial problem for this solar model was
that there were no heavier radioactive elements such as uranium or radium present in the
Sun, and from the atomic spectrum, it was known that the Sun is made almost entirely of
hydrogen.
Up to 1920, F. W. Aston gave a series of precise measurements of atomic masses. It
was found that four hydrogen atoms are more weight slight than one atom of helium-4.
This implied that the fusion of four hydrogens to form a 4 He would be more favorable, and
would release a substantial amounts of energy. A. Eddington proposed that the source of
the Sun’s energy is the nuclear fusion, and in essence he was correct.
In 1938, H. Bethe in collaboration with C. Critchfield had come up with a series of
subsequent nuclear reactions, which was known as the proton-proton p − p chain. The
p − p cycle well describes the reaction processes in the Sun, and consists of the following
four steps:
362 Chapter 6 Quantum Physics

Step 1: two protons yield a deuteron

p + p −→ 2 H + e+ + νe at 99.75%,
2
p + p + e −→ H + νe
−
at 0.25%,
Step 2: a deuteron and a proton produces a helium-3
2
H + p −→3 He + γ ,
Step 3: helium-3 makes helium-4 or beryllium
3
He + p −→ 4 He + e+ + νe ,
3
He + 3He −→ 4 He + p + p almost at 86%
3
He + 4He −→ 7 Be + γ at 14%,
Step 4: beryllium makes helium -4
7
Be + e− −→ 7 Li + νe at 99.89%,
7 4 4
Li + p −→ He + He,
7
Be + p −→ 8 Be + γ at 0.11%,
8 8
Be −→ Be + e + νe ,
∗ +

8
Be∗ −→ 4 He + 4 He.
In the p − p chain, it all starts out as hydrogen (proton), and it all ends up as 4 He plus some
electrons, positrons, photons and neutrinos. Because neutrinos interact so weakly, they are
the unique products in the p − p reactions reaching the earth’s surface.
In the p − p chain there are five reactions to yield neutrinos:
p + p −→ 2 H + e+ + νe , (6.3.1)
2
p + p + e −→ H + νe ,
−
(6.3.2)
3
He + p −→ 4 He + e+ + νe , (≃ 0%), (6.3.3)
7 7
Be + e −→ Li + νe ,
−
(6.3.4)
8 8
Be −→ Be + e + νe . ∗ +
(6.3.5)
But the problem is that the detection of the neutrinos have an effect threshold which will
lead to a nearly vanishing response to all neutrinos of lower energy. The energy spectras of
neutrinos in the five reactions are
Em ≃ 0.4 MeV for (6.3.1),
Em ≃ 1.44 MeV for (6.3.2),
Em ≃ 18 MeV for (6.3.3), (6.3.6)
Em ≃ 0.9 MeV for (6.3.4),
Em ≃ 14 MeV for (6.3.5),
6.3 Solar Neutrino Problem 363

where Em is the maximum energy of neutrinos, and the energy flux are

F ≃ 1011/cm2 · s for (6.3.1),

8 2
F ≃ 10 / cm · s for (6.3.2)
2 2
F ≃ 10 / cm · s for (6.3.3) (6.3.7)
10 2
F ≃ 10 / cm · s for (6.3.4),
6 2
F ≃ 10 / cm · s for (6.3.5).

Homestake experiments
The experimental search for solar neutrinos has been undertaken since 1965 by R. Davis
and collaborators in the Homestake goldmine in South Dakota. Since the neutrinos cannot
be directly detected by instruments, it is only by the reactions

νe + X −→ Y + e−

to detect the outgoing products that counte the neutrinos. The Homestake experiments take

νe + 37 Cl −→ 37 Ar + e− . (6.3.8)

The effective threshold of the reaction (6.3.8) is

Ec = 5.8MeV.

Thus, by (6.3.6) only these neutrinos from both reactions (6.3.3) and (6.3.5) can be ob-
served, which occur at a frequency of 0.015%. Theoretic computation showed that the
expected counting rate of solar neutrinos is at

NT h = (5.8 ± 0.7) snu, (6.3.9)

where snu stands for solar neutrino unit:

1 snu = 10−36 reactions/(37 Cl atom · s).

In 1968, R. Davis et al (Davis, Harmer and Hoffman, 1968) reported the experimental
results, their measuring rate is

N Exp = (2.0 ± 0.3) snu. (6.3.10)

the experimental value (6.3.10) is only about one third of the theoretically expected value
(6.3.9). It gave rise to the famous solar neutrino problem.
364 Chapter 6 Quantum Physics

Super-K experiment
In 2001, the Super-Kamiokande collaboration presented its results on solar neutrinos.
Unlike the Homestake experiment, Super-K uses water as the detector. The process is elastic
neutrino-electron scattering:
νx + e −→ νx + e,
where νx is one of the three flavors of neutrinos. This reaction is sensitive to µ and τ neu-
trinos as well as e-neutrinos, but the detection efficiency is 6.5 times greater for e-neutrinos
than for the other two kinds. The outgoing electron is detected by the Cherenkov radiation
it emits in water. They observed the rate at

r = 45% of the expected value.

The Super-Kamiokande detector is located in the Mozumi Mine near Kamioka section of
the city of Hida, Japan.
Sudbury Neutrino Observatory (SNO)
Meanwhile, in the summer of 2001 the SNO collaboration reported their observation
results. They obtained
r = 35% of the predicted value.
The SNO used heavy water (2 H2 O) instead of ordinary water (H2 O), and the SNO detection
method is based on the following reactions:

νe + 2 H −→ p + p + e−, (6.3.11)
νx + 2 H −→ p + n + νx, (6.3.12)
νx + e −→ νx + e .
− −
(6.3.13)

SNO detects electrons e− , but not τ − and µ − , as there is not enough energy in the solar
electron-neutrino such that the transformed tau and mu neutrino can excite neutrons in 2 H
to produce either τ − or µ − .
KamLAND
The loss of reactor electron anti-neutrino ν e is verified by the KamLAND experiment.
A potential alternative experiment
It is known that the following reaction

νµ + n −→ µ − + p (6.3.14)

may occur if the energy of ν µ satisfies

Eνµ > mµ c2 = 106 MeV.

6.3 Solar Neutrino Problem 365

By the energy spectrum (6.3.6), the maximum energy of solar neutrinos is about 14 ∼
18 MeV, which is much smaller than m µ c2 . Hence, assuming oscillation does occur for
solar neutrinos, the reaction
νµ + 2 H −→ µ − + p + p

does not occur for the transformed ν µ from solar electron-neutrinos.

However, based on the weakton model, the complete reaction for (6.3.14) should be

νµ + n + γ −→ µ − + p.

Consequently, the following reaction

νµ + 2 H + γ −→ µ − + p + p (6.3.15)

would occur if
Eνµ + Eγ > 106 MeV. (6.3.16)

Hence one may use high energy photons to hit the heavy water to create the situation in
(6.3.16), so that the reaction (6.3.15) may take place. From (6.3.15), we can detect the µ −
particle to test the neutrino oscillation.
Alternatively, by the µ -decay:

µ − → e− + ν e + ν µ ,

we may measure the electrons to see if there are more electrons than the normal case to test
the existence of mu-neutrinos.

6.3.2 Neutrino oscillations

In order to explain the solar neutrino problem, in 1968 B. Pontecorvo (Pontecorvo, 1957,
1968) introduced the neutrino oscillation mechanism, which amounts to saying that the
neutrinos can change their flavors, i.e. an electron neutrino may transform into a muon or a
tau neutrino. According to this theory, a large amount of electron neutrinos ν e from the Sun
have changed into the ν µ or ντ , leading the discrepancy of solar electron neutrinos. This
neutrino oscillation mechanism is based on the following assumptions:

• The neutrinos are massive, and, consequently, are described by the Dirac equations.

• The three types of neutrinos νe , νµ , ντ are not the eigenstates of the Hamiltonian (i.e.
the Dirac operator)
Ĥ = −ih̄c(⃗α · ∇) + mc2α0 . (6.3.17)
366 Chapter 6 Quantum Physics

• There are three discrete eigenvalues λ j of (6.3.17) with eigenstates:

Ĥ ν j = λ j ν j for 1 ! j ! 3, (6.3.18)

such that νe , νµ , ντ are some linear combinations of {ν j | 1 ! j ! 3}:

⎛ ⎞ ⎛ ⎞
νe ν1
⎝νµ ⎠ = A ⎝ν2 ⎠ , (6.3.19)
ντ ν3

where A ∈ SU(3) is a third-order complex matrix given by (6.3.26) below.

Remark 6.16 The formulas (6.3.17)-(6.3.19) constitute the current model of neutrino
oscillation, which requires the neutrinos being massive. However, the massive neutrino
assumption gives rise two serious problems. First, it is in conflict with the known fact that
the neutrinos violate the parity symmetry. Second, the handedness of neutrinos implies their
velocity being at the speed of light.
In fact, by using the Weyl equations as the neutrino oscillation model we can also deduce
the same conclusions and solve the two mentioned problems. Moreover, the ν mediator
introduced by the authors in (Ma and Wang, 2015b) leads to an alternate explanation to the
solar neutrino problem.

Under the above three hypotheses (6.3.17)-(6.3.19), the oscillation between ν e , νµ and
ντ are given in the following fashion. For simplicity we only consider two kinds neutrinos
νe , νµ , i.e. ντ = 0. In this case, (6.3.19) becomes

ν1 = cos θ νµ − sin θ νe ,
(6.3.20)
ν2 = sin θ νµ + cos θ νe .

By the Dirac equations (6.2.15) and (6.3.18), ν1 and ν2 satisfy

∂ νk
ih̄ = λk νk for k = 1, 2.
∂t
The solutions of these equations read

νk = νk (0)e−iλkt/h̄ , k = 1, 2. (6.3.21)

Assume that the initial state is at νe , i.e.

νe (0) = 1, νµ (0) = 0.

Then we derive from (6.3.20) that

ν1 (0) = − sin θ , ν2 (0) = cos θ . (6.3.22)

6.3 Solar Neutrino Problem 367

It follows from (6.3.21) and (6.3.22) that

ν1 = − sin θ e−iλ1t/h̄ , ν2 = cos θ e−iλ2t/h̄ . (6.3.23)

Inserting (6.3.23) into (6.3.20) we deduce that

νµ (t) = cos θ ν1 (t) + sin θ ν2 (t) = sin θ cos θ (−e−iλ1t/h̄ + e−iλ2t/h̄ ).

Hence, the probability of νe transforming to νµ at time t is

E ! "F2
2 λ2 − λ1
P(νe → νµ ) = |νµ (t)| = sin 2θ sin t . (6.3.24)
2h̄
Also, we derive in the same fashion that

νe (t) = cos θ ν2 − sin θ ν1 = cos2 θ e−iλ1t/h̄ + sin2 θ e−iλ2t/h̄ ,

and the probability of ν µ to νe is given by

! " ! "
λ2 − λ1 λ2 − λ1
P(νµ → νe ) = |νe (t)|2 = cos2 t + cos2 2θ sin2 t (6.3.25)
2h̄ h̄
From formulas (6.3.24) and (6.3.25), we derive the oscillation between ν e and νµ , the energy
difference λ2 − λ1 , and the angle θ , if the discrepancy probability P(νe → νµ ) is measured.

6.3.3 Mixing matrix and neutrino masses

As mentioned in Remark 6.16, the current neutrino oscillation requires mass matrix A de-
fined in (6.3.19). In this subsection we shall discuss these two topics.
Mixing matrix
The matrix A given in (6.3.19) is called the MNS matrix, which is due to Z. Maki, M.
Nakagawa and S. Sakata for their pioneering work in (Maki, Nakagawa and Sakata, 1962).
This can be considered as an analog for leptons as the Cabibbo-Kobayashi-Maskawa (CKM)
matrix for quarks. The MNS matrix is written as
⎛ ⎞
c12 c13 s12 c13 s13 e−iδ
A = ⎝−s12 c23 − c12s23 s13 eiδ c12 c23 − s12 s23 s13 eiδ s23 c13 ⎠ , (6.3.26)
s12 s23 − c12 c23 s13 e iδ −c12 c23 − s12 c23 s13 e iδ c23 c13

where δ is the phase factor, and

ci j = cos θi j , si j = sin θi j ,

with the values θi j being measured as

θ12 ≃ 34◦ ± 2◦, θ23 ≃ 45◦ ± 8◦, θ13 ≃ 10◦ .

368 Chapter 6 Quantum Physics

The matrix A of (6.3.26) is a unitary matrix: A† = A−1 . Therefore, (6.3.19) can be also
rewritten as ⎛ ⎞ ⎛ ⎞
ν1 νe
⎝ν2 ⎠ = A† ⎝νµ ⎠ .
ν3 ντ
Neutrino masses
As masses are much less than kinetic energy c|p|, by the Einstein triangular relation of
energy-momentum
E 2 = p 2 c2 + m2 c4 ,
we obtain an approximate relation:
1 m2 c 3
E ≃ |p|c + .
2 |p|
The eigenvalues λk of (6.3.18) and E satisfy
1 m2k c3
λk = Ek ≃ |p|c + for k = 1, 2, 3. (6.3.27)
2 |p|
Then we have
(m2i − m2j )
λi − λ j = Ei − E j ≃ c4 , E ≃ |p|c. (6.3.28)
2E
By (6.3.28), if we can measure the energy difference λ i − λ j , then we get the mass square
difference of νi and ν j :
∆i j = m2i − m2j .
There are three mass square differences for ν1 , ν2 , ν3 :

∆21 = m22 − m21 , ∆32 = m23 − m22 , ∆31 = m23 − m21 , (6.3.29)

only two of which are independent (∆31 = ∆32 + ∆21).

Now, we consider the mass relation between νe , νµ , ντ and ν1 , ν2 , ν3 . Applying the
Dirac operator Ĥ on both sides of (6.3.19), by (6.3.18), we have
⎛ ⎞ ⎛ ⎞⎛ ⎞
νe λ1 0 0 ν1
Ĥ ⎝νµ ⎠ = A ⎝ 0 λ2 0 ⎠ ⎝ν2 ⎠ , (6.3.30)
ντ 0 0 λ3 ν3
where A is the MNS matrix (6.3.26). By Postulate 5.4 (i.e. (6.2.4)), the energies E e , Eµ , Eτ
of νe , νµ , ντ are given by
! !
Ee = νe∗ Ĥ νe dx = A∗1k A1 j νk∗ Ĥ ν j dx (Ĥ ν j = λ j ν j ),
! !
Eµ = νµ∗ Ĥ νµ dx = A∗2k A2 j νk∗ Ĥ ν j dx, (6.3.31)
! !
Eτ = ντ∗ Ĥ ντ dx = A∗3k A3 j νk∗ Ĥ ν j dx,
6.3 Solar Neutrino Problem 369

where Ai j are the matrix elements of A, A∗i j are the complex conjugates of Ai j . The masses
me , mµ , mτ of νe , νµ , ντ are as follows

Ee2 = p2 c2 + m2e c4 , Eµ2 = p2 c2 + m2µ c4 , Eτ2 = p2 c2 + m2τ c4 . (6.3.32)

It is very difficult to compute Ee , Eµ , Eτ by (6.3.31). However, since A ∈ SU(3) is norm-

preserving:
Ee2 + Eµ2 + Eτ2 = E12 + E22 + E32 ,

by (6.3.32) and Ek2 = p2 c2 + m2k c4 , we deduce that

m2e + m2µ + m2τ = m21 + m22 + m23 ,

which leads to
m2e + m2µ + m2τ = ∆32 + 2∆21 + 3m21, (6.3.33)

where ∆32 and ∆21 are as in (6.3.29).

If neutrinos have masses, then only the mass square differences ∆ i j in (6.3.29) can be
measured by current experimental methods. Hence, the only mass information of ν e , νµ , ντ
is given by the relation (6.3.33).

6.3.4 MSW effect

In 1978, L. Wolfenstein (Wolfenstein, 1978) first noted that as neutrinos pass through matter
there are additional effects due to elastic scattering

νe + e −→ νe + e.

This phenomenon was also observed and expanded by S. Mikheyev and A. Smirnov (Mikheev
and Smirnov, 1986), and is now called the MSW effect.
The MSW effect can be reflected in the neutrino oscillation model. We recall the oscil-
lation model without MSW effect expressed as

νk = ϕk (x)e−iλk t/h̄ ,
[−ih̄c(⃗α · ∇) + mc2α0 ]ϕk = λk ϕk k = 1, 2, 3,
⎛ ⎞ ⎛ ⎞ (6.3.34)
νe ν1
⎝νµ ⎠ = A ⎝ν2 ⎠ , A is as in (6.3.26).
ντ ν3

To consider the MSW effect, we have to add weak interaction potentials in the Hamilto-
nian operator Ĥ for neutrinos νe , νµ , ντ . The weak potential energy is as given in (4.6.32):
E F
1 B
Vν = gs (ρν )ρs (ρe )Ne e−kr − (1 + 2kr)e−kr , (6.3.35)
r ρ
370 Chapter 6 Quantum Physics

where ρν , ρe are the radii of neutrinos and electron, gs is the weak charge, and Nw is the
weak charge density. Namely the Hamiltonian with MSW effect for (νe , νµ , ντ ) is
⎛ ⎞ ⎛ ⎞⎛ ⎞
νe Ĥ + Ve 0 0 νe
Hˆ ⎝νµ ⎠ = ⎝ 0 Ĥ + Vµ 0 ⎠ ⎝ν µ ⎠ , (6.3.36)
ντ 0 0 Ĥ + Vτ ντ

where Ĥ = −ih̄c(⃗α · ∇) + mc2α0 , and V is as in (6.3.35).

The equations in (6.3.34) are also in the form
⎛ iλ t/h̄ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
e 1 νe λ1 ϕ1
⎝ eiλ2t/h̄ ⎠ A† Ĥ ⎝νµ ⎠ = ⎝ λ2 ⎠ ⎝ϕ2 ⎠ . (6.3.37)
e iλ3 t/h̄ ντ λ3 ϕ3

Replacing Ĥ by Hˆ in (6.3.37), we infer from (6.3.34) that

⎛ ⎞ ⎛ ⎞⎛ ⎞
ν1 β1 ν1
A† Hˆ A ⎝ν2 ⎠ = ⎝ β2 ⎠ ⎝ν2 ⎠ , (6.3.38)
ν3 β3 ν3

where Hˆ is as in (6.3.36).
The equation (6.3.38) is the neutrino oscillation model with the MSW effect, where the
eigenvalues βk and eigenstates νk (1 ! k ! 3) are different from that of (6.3.34). In fact, the
MSW effect is just the weak interaction effect.

6.3.5 Massless neutrino oscillation model

There are several serious problems in the massive neutrino oscillation model (6.3.34), which
we briefly explain as follows.

1. Parity problem. It is known that all weak interaction decays and scatterings involving
neutrinos violate the parity symmetry, discovered by Lee and Yang in 1956 and experimen-
tally verified by C. Wu (Lee and Yang, 1956; Wu, Ambler, Hayward, Hoppes and Hudson,
1957). It means that the neutrinos are parity non-conserved. Hence it requires that under
the space reflective transformation
x −→ −x, (6.3.39)

the equations governing neutrinos should violate the reflective invariance. Based on Re-
marks 6.7 and 6.8, the Dirac equations are invariant under the reflection (6.3.39), while
the Weyl equations are not invariant. Hence, the massive neutrino oscillation model is in
conflict with the violation of parity symmetry.

2. Handedness and speed of neutrinos. Experiments showed that all neutrinos possess
1
only the left-handed spin J = − , and anti-neutrinos possess the right-handed spin J =
2
6.3 Solar Neutrino Problem 371

1
. It implies that the velocity of free neutrinos must be at the speed of light, which is a
2
contradiction with massive neutrino assumption.
In fact, the handedness is allowed only for massless particles. Otherwise, there exist two
coordinate systems A and B satisfying

vA < v p < v B ,

where vA , vB and v p are the velocities of A, B and the particle. When we look at the particle
ν from A and B, the spins would be reversed. Therefore, all massive particles must have
both left-handed and right-handed spins.
In addition, all experiments measuring neutrino velocity had found no violation to the
speed of light.
3. Infinite number of eigenvalues and eigenstates. The neutrino oscillation theory faces
the problem of the existing of infinite number of eigenvalues. In the massive model (6.3.34),
the wave functions are the Dirac spinors

ϕ = (ϕ 1 , ϕ 2 , ϕ 3 , ϕ 4 )T .

For free neutrinos moving on a straight line, ϕ depends only on z. Thus the eigenvalue
equations in (6.3.34) become
! " ! 1" ! 1"
d ϕ3 2 ϕ ϕ
− ih̄cσ3 + mc =λ ,
dz ϕ 4 ϕ2 ϕ2
! 1" ! 3" ! 3" (6.3.40)
d ϕ 2 ϕ ϕ
− ih̄cσ3 − mc =λ ,
dz ϕ 2 ϕ4 ϕ4
where !"
1 0
σ3 = . (6.3.41)
0 −1
The equations (6.3.40) possess infinite number of eigenvalues
O
4π 2 n2 h̄2 c2
λ = m2 c4 + , ∀l > 0, n = 0, 1, 2, · · · , (6.3.42)
l2
and each eigenvalue has two eigenstates
⎛' ⎞
1 + mc2/λ
ei2π nz/l ⎜' 0 ⎟
ϕ1 = √ 3/2 ⎜ ⎝
⎟,
2l 1 − mc /λ ⎠
2

0
⎛ ⎞ (6.3.43)
' 0
ei2π nz/l ⎜ 1 + mc2/λ ⎟
ϕ2 = √ 3/2 ⎜ ⎟.
2l ⎝ ' 0 ⎠
− 1 − mc2/λ
372 Chapter 6 Quantum Physics

The problem is that which eigenvalues and eigenstates in (6.3.42) and (6.3.43) are the ones
in the neutrino oscillation model (6.3.34), and why only three of (6.3.42)-(6.3.43) stand for
the flavors of neutrinos.
The Weyl equations (6.2.13) can replace the Dirac equations to describe the neutrino
oscillation, which we call massless neutrino oscillation model, expressed as follows

νk = ϕk (x)e−iλk t/h̄ ,
ih̄c(⃗σ · ∇)ϕk = λk ϕk for k = 1, 2, 3,
⎛ ⎞ ⎛ ⎞ (6.3.44)
νe ν1
⎝νµ ⎠ = A ⎝ν2 ⎠ , A is as in (6.3.26),
ντ ν3

where νk (1 ! k ! 3) are the two-component Weyl spinors, and ⃗σ = (σ1 , σ2 , σ3 )

Based on the massless model (6.3.44), both problems of parity and handedness of neutri-
nos have been resolved, and we can derive in the same conclusions as given in (6.3.24) and
(6.3.25). In this case, the differences λi − λ j of eigenvalues in the transition probabilities
such as (6.3.24) and (6.3.25) stand for the differences of frequencies:

λ i − λ j = ωi − ω j , (6.3.45)

where ωk (1 ! k ! 3) are the frequencies of νk .

However, the massless model also faces the problem of infinite number of eigenvalues
as mentioned above. The eigenvalue equations in (6.3.44) for the straight line motion on the
y-axis is written as
! 1" ! 1" ! "
2 d ϕ ϕ 2 0 −i
ih̄cα =λ with α = . (6.3.46)
dy ϕ 2 ϕ2 i 0

The eigenvalues of (6.3.46) are

λk = kh̄c, ∀k > 0, (6.3.47)

and each eigenvalue of (6.3.47) has two eigenstates

! 1" ! " ! 1" ! "
ϕ1 sin ky ϕ2 cos ky
= , = . (6.3.48)
ϕ12 − cos ky ϕ22 sin ky

The eigenvalues of (6.3.44) at x-axis and z-axis are all the same as in (6.3.47), and the
eigenstates at the x and z axes are
! 1 " ! −ikx " ! " ! −ikz " ! "
ϕ e ϕ1 e 0
= −ikx and = or . (6.3.49)
ϕ2 e ϕ2 0 eikz
6.3 Solar Neutrino Problem 373

6.3.6 Neutrino non-oscillation mechanism

Although the massless neutrino oscillation model can solve the parity and the handedness
problems appearing in the massive neutrino oscillation mechanism, the problem of infinite
numbers of eigenvalues and eigenstates still exists in the model (6.3.44).
In fact, the weakton model first introduced in (Ma and Wang, 2015b) can provide an
alternative explanation to the solar neutrino problem. Based on the weakton model, there
exists a ν -mediator, whose weakton constituents are given by

ν = αe νe ν e + αµ νµ ν µ + ατ ντ ν τ , (6.3.50)

where αe2 + αµ2 + αµ2 = 1. The values αe2 , αµ2 , ατ2 represent the ratio of the neutrinos νe , νµ
and ντ in our Universe
In view of (6.3.50), we see the reaction

νe + ν e −→ ν (νe ν e ),
νµ + ν µ −→ ν (νµ ν µ ), (6.3.51)
ντ + ν τ −→ ν (ντ ν τ ),

which are generated by the weak interaction attracting force, as demonstrated in the weak
charge potentials
E ! " F
−r/r0 1 Bi 2r −r/r0
Φi = gw e − 1+ e for 1 ! i ! 3, (6.3.52)
r ρν r0

where r0 = 10−16 cm, B1 , B2 , B3 > 0 are the weak interaction constants for νe , νµ , ντ respec-
tively, and ρν is the neutrino radius.
The formula (6.3.52) defines attractive radii Ri for the neutrinos and antineutrinos of the
same flavors. Namely, when νi and ν i are in the radius Ri , the reaction (6.3.51) may occur:

νi + ν i −→ ν (νi ν i ) if dist(νi , ν i ) < Ri for 1 ! i ! 3, (6.3.53)

where ν1 = νe , ν2 = νµ , ν3 = ντ , and dist(νi , ν i ) is the distance between νi and ν i . The

condition (6.3.53) implies that the transition probability Γi depends on Ri :

Γi = Γi (Ri ) for 1 ! i ! 3. (6.3.54)

The attracting radius Ri satisfies that

d
Φi (Ri ) = 0. (6.3.55)
dr
Thus we can give a non-oscillation mechanism of neutrinos to explain the solar neutrino
problem. Namely, due to the β -decay, there are large amounts of electronic anti-neutrinos
374 Chapter 6 Quantum Physics

ν e around the earth, which generate the reaction (6.3.51) with νe , leading to the discrepancy
of the solar neutrino.
In addition, there are three axis eigenvalues of the Weyl equations given by (6.3.48)
and (6.3.49). We believe that they are the three flavors of neutrinos ν e , νµ , ντ . Namely, the
following three wave functions
! " ! "
sin ky cos ky
ψ1 = c 1 + c2 ,
− cos ky sin ky
! −ikx "
e
ψ2 = −ikx , (6.3.56)
e
! −ikz " ! "
e 0
ψ3 = c 3 + c4 ikz
0 e
represent the three flavors of neutrinos. In fact, massive particles in field equations are
distinguished by different masses, and flavors of neutrinos by different axis eigenstates.
Also, neutrinos have significant scattering effect of the weak interaction. In fact, they
can have weak interactions with all subatomic particles. This can certainly cause deficits of
neutrinos, and may be one of the main reasons for the loss of solar neutrinos.

6.4 Energy Levels of Subatomic Particles

6.4.1 Preliminaries
For subatomic particles we can establish energy levels, and this theory is based on two basic
ingredients:

1) each subatomic particle consisting of two or three weaktons or quarks bound by weak
or strong interaction, and

2) the weak and strong interaction potential formulas.

For convenience, we here briefly recall them.

1. Elementary particles. There are six pairs of elementary particles, called weaktons:

w∗ , w1 , w2 , νe , νµ , ντ ,
w∗ , w1 , w2 , ν e , ν µ , ν τ .

Each weakton carries one unit of weak charge gw , and both w∗ and w∗ are only weaktons
carrying a strong charge gs .
2. Six classes of subatomic particles. There are six types of subatomic particles:
charged leptons, quarks, barons, mesons, intermediate bosons, mediators, whose members
are listed as follows.
6.4 Energy Levels of Subatomic Particles 375

1) Charged leptons:
e± , µ ± , τ ± .

2) Quarks:

u, d, s, c, b, t,
u, d, s, c, b, t.

3) Baryons:
p± , n, Λ, Σ± , Σ0 , ∆++ , ∆± , ∆0 , Ξ± , Ξ0 , etc.

4) Mesons:
π ± , π 0 , K ± , K 0 , η , ρ ± , ρ 0 , K ∗± , K ∗0 , etc.

5) Intermediate bosons:
W ±, Z0 , H ±, H 0 .

6) Mediators:
γ , γ0 , gk , gk0 , ν .

3. Constituents of subatomic particles.

1) Weakton constituents of charged leptons:

e− = νe w1 w2 , µ − = νµ w1 w2 , τ − = ντ w1 w2 .
e+ = ν e w1 w2 , µ + = ν µ w1 w2 , τ + = ν τ w1 w2 .

2) Weakton constituents of quarks:

u = w∗ w1 w1 , c = w∗ w2 w2 , t = w∗ w2 w2 ,
d = w∗ w1 w2 , s = w∗ w1 w2 , b = w∗ w1 w2 .

3) Quark constituents of baryons:

Baron = qqq.

4) Quark constituents of mesons:

Meson = qq.

5) Weakton constituents of intermediate bosons:

W + = w1 w2 (#, $), W − = w1 w2 (#, $), Z 0 = α1 w1 w1 + α2 w2 w2 (#, $),

H + = w1 w2 (↑↓, ↓↑), H − = w1 w2 (↑↓, ↓↑), H 0 = α1 w1 w1 + α2 w2 w2 (↑↓, ↓↑).
376 Chapter 6 Quantum Physics

6) Weakton constituents of mediators:

γ = α1 w1 w1 − α2 w2 w2 (#, $), gk = w∗ w∗ (#, $),

γ0 = α1 w1 w1 − α2 w2 w2 (↑↓, ↓↑), gk0 = w∗ w∗ (↑↓, ↓↑),
ν = αe νe ν e + αµ νµ ν µ + ατ ντ ν τ .

4. Bound forces of subatomic particles. The main forces holding weaktons and quarks
to form subatomic particles are the weak and strong interactions, and their force sources are
from the interaction charges, i.e.
weak charge gw , strong charge gs .
The 4-dimensional interaction potentials are
weak potential Wµ = ωaWµa = (W0 ,W1 ,W2 ,W3 ),
(6.4.1)
strong potential S µ = ρk Skµ = (S0 , S1 , S2 , S3 ),
The acting forces are
weak force =− gw ∇W0 ,
strong force =− gs∇S0 ,
(6.4.2)
weak magnetism =− gw curl W ⃗,
strong magnetism=− gscurl⃗S,
where W⃗ = (W1 ,W2 ,W3 ), ⃗S = (S1 , S2 , S3 ).
Since each weakton carries one weak charge gw , and w∗ , w∗ carry one strong charge gs ,
by the constituents of subatomic particles, the main bound energy for various particles takes
charged leptons and quarks: weak interaction,
hadrons (barons and mesons) : strong interaction,
intermediate bosons: weak interaction, (6.4.3)
vector and scalar gluons: weak and strong interactions,
other mediators: weak interaction.
Remark 6.17 We need to explain that although each quark has three weak charges, due
to the weak force range r ! 10−16 cm and the distance r > 10−16 cm between the quarks in a
hadron, the main bound force of hadrons is strong interaction, and the weak forces between
the quarks an be ignored.

5. Weak interaction potentials. In Section 4.6.2, we deduced the weak interaction

potential (4.6.17). For convenience, here we again write this formulas as
E ! " F
1 B 2r −r/r0
W0 = Ngw (ρ )e−r/r0 − 1+ e ,
r ρ r0
! "3 (6.4.4)
ρw
gw (ρ ) = gw ,
ρ
6.4 Energy Levels of Subatomic Particles 377

where W0 is the weak charge potential of a composite particle with N weak charges g w and
with radius ρ , and ρw is the weakton radius, B the weak interaction constants of this particle,
r0 = 10−16 cm.
Based on (6.4.4), the weak potential energy generated by two particles with N1 and N2
weak charges and with radii ρ1 , ρ2 is expressed as follows
E ! " F
−r/r0 1 B12 2r −r/r0
Vw = N1 N2 gw (ρ1 )gw (ρ2 )e − 1+ e , (6.4.5)
r ρ12 r0

where gw (ρ1 ) and gw (ρ2 ) are as in (6.4.4), and B12 /ρ12 depends on the types of the two
particles.
The basic weak charge gw satisfies the relation (4.6.37), i.e.
! "6
ρn
g2w = 5 × 10−3 h̄c. (6.4.6)
ρw

6. Strong interaction potentials. The strong interaction potential given by (4.5.39),

reads as E F
1 AI r J −r/R
S0 = gs (ρ ) − 1+ e ,
r ρ R
! "3 (6.4.7)
ρw
gs (ρ ) = gs ,
ρ
where R is as P
10−16 cm for w∗ and quarks,
R=
10−13 cm for hadrons.
By (6.4.7), the strong potential energy for two particles with N1 , N2 strong charges and
with radii ρ1 , ρ2 is given by
E F
1 A12 I r J −r/R
Vs = N1 N2 gs (ρ1 )gs (ρ2 ) − 1+ e , (6.4.8)
r ρ12 R

where gs (ρ1 ) and gs (ρ2 ) are as in (6.4.7), and A12 /ρ12 depends on the types of the two
particles.
The basic strong charge gs satisfies the relation (4.5.66), i.e.
! "6
ρn
g2s = 2 × 10 −2
g2 (g2 ≃ 1h̄c). (6.4.9)
ρw

6.4.2 Spectral equations of bound states

In the last subsection we see that the subatomic particles have six classes, in where the
mediators are massless and others are massive. By the mass generation mechanism given in
Subsection 5.3.2, the weaktons in massive subatomic particles possess masses, however the
378 Chapter 6 Quantum Physics

weaktons in mediators are massless. The spectral equations for both massive and massless
bound states are very different. in the following we shall discuss them respectively.

Massive bound states

The subatomic particles consist of two or three fermions, their wave functions are the
Dirac spinors
Ψ1 , · · · , ΨN , N = 2 or 3. (6.4.10)

As N = 2 or 3, for each particle its bound energy can be approximatively regarded as the
superposition of the remaining N − 1 particles. Thus the bound potential for each fermion
takes the form
⎧
⎪ (N − 1)gwWµ for weak interaction,
⎨
gAµ = (N − 1)gsSµ for strong interaction, (6.4.11)
⎪
⎩
(N − 1)(gwWµ + gsSµ ) for weak and strong interactions

where Wµ and S µ are as in (6.4.1).

Let the masses of the N particles are
⎛ ⎞
m1 0
⎜
m=⎝ .. ⎟
⎠.
.
0 mN

Then the N wave functions of (6.4.10) satisfy the Dirac equation

(ih̄cγ µ Dµ − c2 m)Ψ = 0, (6.4.12)

where Ψ = (Ψ1 , · · · , ΨN ), and

g
Dµ = ∂µ + i Aµ , gAµ as in (6.4.11). (6.4.13)
h̄c
It is known that each wave function is a 4-components spinor

Ψk = (Ψk1 , Ψk2 , Ψk3 , Ψk4 ), 1 ! k ! N.

Therefore, the equation (6.4.12) takes the equivalent form

! k " ! "
Ψ1 Ψk3
(ih̄ ∂∂t − gA0 − c2 mk ) k = −ih̄c(⃗σ · ⃗
D) ,
! Ψ2k " Ψk4 "
! (6.4.14)
∂ 2 Ψ3 Ψk1
(ih̄ ∂ t − gA0 + c mk ) = −ih̄c(⃗σ · D)
⃗ ,
Ψk4 Ψk2

where ⃗σ = (σ 1 , σ 2 , σ 3 ) is the Pauli matrix operator, ⃗D is as in (6.4.13).

6.4 Energy Levels of Subatomic Particles 379

We now derive spectral equations for massive bound states from (6.4.14). Let the solu-
tions of (6.4.14) be in the form
2 )t/h̄
Ψk = e−i(λ +mk c ψ k.

Then equations (6.4.14) become

! k " ! k "
ψ1 ψ3
(λ − gA0) k = −ich̄(⃗σ · D)
⃗ , (6.4.15)
ψ2 ψ4k
! k " ! k "
ψ3 ψ1
(λ − gA0 + 2mk c2 ) = σ
−ich̄(⃗ · ⃗
D) , (6.4.16)
ψ4k ψ2k

for 1 ! k ! N. The equation (6.4.16) can be rewritten as

! " ! " ! k "
ψ3k −ih̄ λ − gA0 −1 ψ1
= 1+ (⃗σ · D)
⃗ . (6.4.17)
ψ4k 2mk c 2mc2 ψ2k

In physics, λ is the energy, and λ − gA0 is the kinetic energy

1
λ − gA0 = mk ν 2 .
2

For massive particles, ν 2 /c2 ≃ 0. Hence, (6.4.17) can be approximatively expressed as

! k " ! k "
ψ3 −ih̄ ψ1
= (⃗σ · D)
⃗ .
ψ4k 2mk c ψ2k

Inserting this equation into (6.4.15), we deduce that

! " ! "
ψ1k h̄2 ψ1k
(λ − gA0) =− (⃗σ · ⃗D)2 . (6.4.18)
ψ2k 2mk ψ2k

Now, we need to give the expression of (⃗σ · ⃗D)2 . To this end, note that the Pauli matrices
satisfy
(σ k )2 = 1, σ k σ j = −σ j σ k = iεljk σ l .

Here εljk is the arrange symbal:

⎧
⎨ 1 as ( jkl) the even arrange,
εijk = −1 as ( jkl) the odd arrange,
⎩
0 otherwise.

Hence we obtain Q R2
3
(⃗σ · ⃗D) =2
∑σ k
Dk = D2 + i⃗σ · (⃗D × ⃗D). (6.4.19)
k=1
380 Chapter 6 Quantum Physics

g⃗
with ⃗D = ∇ + i h̄c A, we derive that
M N
⃗D × ⃗D = i g ∇ × ⃗A + ⃗A × ∇ .
h̄c
Note that as an operator we have

∇ × ⃗A = curl ⃗A − ⃗A × ∇.

Hence we get
⃗D × ⃗D = i g curl⃗A.
h̄c
Thus, (6.4.19) is written as
g
(⃗σ · ⃗D)2 = D2 − ⃗σ · curl⃗A. (6.4.20)
h̄c
By (6.4.20), the spectral equation (6.4.18) is in the form
E F! k " ! k " ! k "
h̄2 2 ψ1 ψ1 ψ1
− D + gA0 +⃗ µ k · curl⃗
A = λ , (6.4.21)
2mk ψ2k ψ2k ψ2k

where D = (D1 , D2 , D3 ), (A0 , A1 , A2 , A3 ) is as in (6.4.11), and

h̄g g
⃗µk = ⃗σ , D k = ∂k + i Ak (1 ! k ! 3). (6.4.22)
2mk h̄c

Since the fermions are bound in the interior Ω of subatomic particle, ψ k (1 ! k ! N) are zero
outside Ω. Therefore the equation (6.4.21) are supplemented with the Dirichlet boundary
conditions:
(ψ1k , ψ2k )|∂ Ω = 0 (1 ! k ! 3), (6.4.23)
where Ω ⊂ R3 is a bounded domain.
The boundary value problem (6.4.21)-(6.4.23) is the model for the energy level theory
of massive subatomic particles.
Massless bound states
In order to obtain the spectral equations for massless subatomic particle, we have to
derive their wave equations, which are based on the basic quantum mechanics principle: the
Postulate 5.5.
We first recall the Weyl equation
∂ψ
= c(⃗σ · ∇)ψ , (6.4.24)
∂t
which describes massless and free fermions. The Weyl equation (6.4.24) is derived from the
de Broglie relation
E = cp ( see (6.2.12)) (6.4.25)
6.4 Energy Levels of Subatomic Particles 381

with
∂
, p̂ = ih̄(⃗σ · ∇).
Ê = ih̄ (6.4.26)
∂t
As consider the massless bound states in weak and strong interactions, the Hermitian
operators in (6.4.26) are replaced by
∂
Ê = ih̄ − gA0, p̂ = ih̄(⃗σ · ⃗D), (6.4.27)
∂t
where ⃗D = (D1 , D2 , D3 ) is as in (6.4.22).
In Section 6.4.1, we knew that the mediators such as photons and gluons consist of two
massless weaktons, which are bound in a small ball Br by the weak and strong interactions.
Hence the Weyl spinor ψ of each weakton is restricted in a small ball B r , i.e.

ψ = 0, ∀x ̸∈ Br ,

which implies the boundary condition

ψ |∂ Br = 0. (6.4.28)

However, in mathematics the boundary problem for the Weyl equations generated by
(6.4.25) and (6.4.27) given by

ih̄ ∂∂ψt = ih̄c(⃗σ · ⃗D)ψ + gA0ψ ,

(6.4.29)
ψ |∂ Br = 0,
is in general not well-posed, i.e (6.4.29) has no solution for a given initial value ψ (0) = ψ 0
in general. Hence, for a massless fermion system with the boundary condition (6.4.28), we
have to consider the relation
pE = cp2 , (6.4.30)
which is of first order in time t. It is known that the operator p̂Ê is Hermitian if and only if

p̂Ê = Ê p̂.

Note that Ê = ih̄∂ /∂ t − gA0, and in general

p̂A0 ̸= A0 p̂.

Hence, in order to ensure p̂Ê being Hermitian, we replace p̂A0 by 12 ( p̂A0 + A0 p̂), i.e. take
∂ g
− ( p̂A0 + A0 p̂), p̂ = ih̄c(⃗σ · ⃗D).
p̂Ê = ih̄ p̂ (6.4.31)
∂t 2
Then by Postulate 6.5, from (6.4.30) and (6.4.31) we derive the boundary problem of
massless system in the following form
∂ψ ig
(⃗σ · ⃗D) = c(⃗σ · ⃗D)2 ψ − {(⃗σ · ⃗D), A0 }ψ ,
∂t 2h̄ (6.4.32)
ψ |∂ Ω = 0,
382 Chapter 6 Quantum Physics

where Ω ⊂ Rn is a bounded domain, and {A, B} = AB + BA is the anti-commutator.

Now, we derive the spectral equations from (6.4.32) for the massless bound states. Let
the solutions ψ of (6.4.32) be in the form
! "
−iλ t/h̄ ϕ1
ψ =e ϕ, ϕ= ,
ϕ2

Then equation (6.4.32) are reduced to the eigenvalue problem

ig
−h̄c(⃗σ · ⃗D)2 ϕ + {(⃗σ · ⃗D), A0 }ϕ = iλ (⃗σ · ⃗D)ϕ ,
2
and by (6.4.20) which can be rewritten as
! " ! "
ϕ1 ig ϕ1
[−h̄cD2 + g⃗σ · curl⃗A] + {(⃗σ · ⃗D), A0 }
ϕ2 2 ϕ2
! 1 "
ϕ
= iλ (⃗σ · ⃗D) , (6.4.33)
ϕ2
(ϕ 1 , ϕ 2 )|∂ Ω = 0,

where {(⃗σ · ⃗D), A0 } is the anti-commutator defined by

{(⃗σ · ⃗D), A0 } = (⃗σ · ⃗D)A0 + A0 (⃗σ · ⃗D). (6.4.34)

The eigenvalue equations (6.4.33)-(6.4.34) are taken as the model for the energy levels
of massless bound states. The mathematical theory (Theorem 2.42) established in Subsec-
tion 3.6.5 laid a solid foundation for the energy level theory provided by (6.4.33).

Remark 6.18 In the equations (6.4.21)-(6.4.22) and (6.4.33) for bound states, we see
that there are terms

⃗µ · curl⃗A for massive particle systems, (6.4.35)

g⃗σ · curl⃗A for massless particle systems. (6.4.36)

In (6.4.35), the physical quantity ⃗µ = h̄g⃗σ /2m represents magnetic moment, and the term
in (6.4.36) is magnetic force generated by the spin coupled with ether weak or strong in-
teraction. In other words, in the same spirit as the electric charge e producing magnetism,
the weak and strong charges gw , gs can also produce similar effects, which we also call
magnetism.
Indeed, all three interactions: electromagnetic, weak, strong interactions, enjoy a com-
mon property that moving charges yield magnetism, mainly due to the fact that they are all
gauge fields.
6.4 Energy Levels of Subatomic Particles 383

6.4.3 Charged leptons and quarks

According to structure and interaction types, we shall discuss the energy levels for three
groups of particles. charged leptons and quarks, hadrons, mediators. In this subsection we
only consider the case of charged leptons and quarks.
In Section 6.4.1 we see that charged leptons and quarks are made up of three weaktons,
with masses caused by the deceleration of the constituent weaktons. Let the masses of the
constituent weaktons be m1 , m2 , m3 , and the wave functions of these weaktons be given by
! k "
k ψ1
ψ = for k = 1, 2, 3.
ψ2k

Here ψ1k and ψ2k represent the left-hand and right-hand states. The bound states are due to
the weak interaction, and the potential in (6.4.11) takes the form

⃗ ).
gAµ = 2gwWµ = (2gwW0 , 2gwW

By (6.4.21)-(6.4.23), the spectral equations for charged leptons and quarks are as follows
! "
h̄2 2gw ⃗ 2 j ⃗ )ψ j
− ∇+i W ψ + 2(gww0 +⃗µ j · curlW
2m j h̄c
=λψ j in ρw < |x| < ρ , 1! j!3 (6.4.37)
1 2 3
ψ = (ψ , ψ , ψ ) = 0 at |x| = ρw , ρ ,

where ρw is the weakton radius, ρ the attracting radius of weak interaction, W0 is given by
(6.4.4) with N = 1, gw (ρ ) = gw , and
h̄gw
⃗µ j = ⃗σ is the weak magnetic moment.
2m j
We are in position now to derive a few results on the energy levels for charged leptons
and quarks based on (6.4.37).
1. Bound states and energy levels. We know that the negative eigenvalues and eigen-
functions of (6.4.37) correspond to the bound energy and bound states. Let

−∞ < λ1 ! · · · ! λN < 0

be all negative eigenvalues of (6.4.37) with eigenfunctions

ψ1 , · · · , ψ N for ψk = (ψk1 , ψk2 , ψk3 )T .

Each bound state ψk satisfies

!
|ψkj |2 dx = 1 for 1 ! j ! 3, 1 ! k ! N, (6.4.38)
Ω
384 Chapter 6 Quantum Physics

where Ω = {x ∈ R3 | ρw < |x| < ρ }. Then, by (6.4.37) and (6.4.38) we get

! /I
h̄ / gw ⃗ J j //2
λk = / ∇+i W ψk / dx (6.4.39)
2m j Ω h̄c
! !
+2 ⃗ |ψ j |2 dx + 2
⃗µ j · curlW gwW0 |ψkj |2 dx.
k
Ω Ω

λk is independent of j (1 ! j ! 3), i.e. each weakton has the same bound energy λ k but in
different bound state ψkj .
In the right-hand side of (6.4.39), the first term stands for the kinetic energy, the second
term for the weak magnetic energy, and the third term for the weak potential energy, the
potential energy in (6.4.39) is negative. Hence, the bound energy can be written as

λk = kinetic energy + magnetic energy + potential energy.

In addition, the energy distributions of charged leptons and quarks are discrete and finite:

0 < E1 < · · · < EN0 (N0 ! N), (6.4.40)

and N is the number of negative eigenvalues. Each energy level E k can be expressed as

Ek = 3(E0 + λk ), E0 = g2w /ρw is the intrinsic energy.

2. Masses. At an energy level Ek of (6.4.40), the mass Mk of a lepton or a quark satisfies

the relation
3 3
3g2 3λk
Mk = ∑ m j + Ek /c2 = ∑ m j + w2 + 2 .
j=1 j=1 ρ w c c

3. Parameters of electrons. In all charged leptons and quarks, only the electrons are long
life-time and observable. Hence the physical parameters of electrons are important. By the
spectral equation (6.4.37) we can derive some information for electronic parameters.
To this end, we recall the weak interaction potential for the weaktons, which is written
as E ! " F
1 Bw 2r −r/r0 −r/r0
W0 = gw − 1+ e e , (6.4.41)
r ρw r0
where Bw is the constant for weaktons.
Assume that the masses of three weaktons are the same. We ignore the magnetism, i.e.
⃗ = 0. Then (6.4.37) is reduced in the form
let W
h̄2
− ∆ψ + 2gwW0 ψ = λ ψ for ρw < |x| < ρ , (6.4.42)
2m
ψ = 0, for |x| = ρw , ρ .
We shall apply (6.4.41) and (6.4.42) to derive some basic parameters and their relations
for the electrons.
6.4 Energy Levels of Subatomic Particles 385

Let λe and ψe be the spectrum and bound state of an electron, which satisfy (6.4.42). It
follows from (6.4.41) and (6.4.42) that
! "
1 2 2 1 κe
λe = mv + 2gw − (6.4.43)
2 re ρw
1 2
where mv is the kinetic energy of each weakton in electron re the radius of the naked
2
electron, κe is the bound parameter of electron, and they are expressed as
!
1 −r/r0
re = e |ψe |2 dx,
Bρ r
! ! "
2r −2r/r0
κe = Bw 1+ e |ψe |2 dx,
Bρ r0
I m v J2 !
w
= |∇ψe |2 dx.
h̄ Bρ

These three parameters are related with the energy levels of an electron, i.e. with the λ k
and ψk . However, the most important case is the lowest energy level state. We shall use the
spherical coordinate to disscuss the first eigenvalue λ1 of (6.4.42). Let the first eigenfunction
ψe be in the form
ψe = ϕ0 (r)Y (θ , ϕ ).
Then ϕ0 and Y satisfy
! "
h̄2 1 d 2 d
− r ϕ0 + 2gsW0 ϕ0 + βr2k ϕ0 = λe ϕ0 ,
2m r2 dr dr (6.4.44)
ϕ0 (ρw ) = ϕ0 (ρ ) = 0,
and E ! " F
1 ∂ ∂ 1 ∂2
sin θ + 2 Yk = βkYk , (6.4.45)
sin θ ∂ θ ∂θ sin θ ∂ ϕ 2
where βk = k(k + 1), k = 0, 1, · · · .
Because λe is the minimal eigenvalue, it implies that βk = β0 = 0 in (6.4.44). The
eigenfunction Y0 of (6.4.45) is given by
1
Y0 = √ .
4π
Thus ψe is as follows
1
ψe = √ ϕ0 (r),
4π
and λe , ϕ0 are the first eigenvalue and eigenfucntion of the following equation
h̄2 1 d ; 2 d <
− r dr ϕ0 + 2gwW0 ϕ0 = λ1 ϕ0 (6.4.46)
2m r2 dr
ϕ0 (ρw ) = ϕ0 (ρ ) = 0.
386 Chapter 6 Quantum Physics

In this case, the parameters in (6.4.43) are simplified as

! ρ
re = rϕ02 (r)e−r/r0 dr,
ρw
! ρ ! "
2r −2r/r0 2
κe = Bw r2 1 + e ϕ0 (r)dr, (6.4.47)
ρw r0
I mv J2 ! ρ ! d ϕ "2
0
= r2 dr,
h̄ ρw dr
where ϕ0 satisfies (6.4.46).

6.4.4 Baryons and mesons

Hadrons include baryons and mesons, their spectral equations are given respectively in the
following.
1. Spectral equations of baryons. Baryons consist of three quarks: B = qqq, and each
quark q consists of three w-weaktons

q = w∗ ww and q = w∗ ww.

Hence, each quark possesses one strong charge g s and three weak charges 3gw . It looks as
if the bound energy of baryons is provided by both weak and strong interactions. However,
since the weak interaction is short-ranged, i.e.

range of weak force ! 10−16 cm,

and the radii of baryons are

r > 10−16 cm.
Hence the main interaction to hold three quarks together is the strong force. Let m 1 , m2 , m3
be the masses of three quarks in a baryon, and ψ k = (ψ1k , ψ2k )T (1 ! k ! 3) be the wave
functions. Then the spectral equations (6.4.21)-(6.4.23) for baryons are in the form
! "
h̄2 2gs 2 k
− ∇ + i ⃗s ψ + 2gsS0 ψ k + 2⃗µk · curl⃗Sψ k
2mk h̄c
= λ ψk in ρ0 < |x| < ρ1 , for 1 ! k ! 3, (6.4.48)
1 2 3
ψ = (ψ , ψ , ψ ) = 0 at |x| = ρ0 , ρ1 ,

where ρ0 is the quark radius, ρ1 the strong attracting radius, S µ = (S0 , ⃗S) as in (6.4.1) is 4-
dimensional strong potential, and ⃗µk = h̄gs⃗σ /2mk the strong magnetic moment. ψ k satisfy
the nomalization
!
|ψ k |2 dx = 1, in Ω = {x ∈ R3 | ρ0 < |x| < ρ1 }.
Ω
6.4 Energy Levels of Subatomic Particles 387

2. Spectral equations of mesons. Mesons consist of a quark and a antiquark: M =

qq. Hence, the bound energy of mesons is mainly provided by strong interaction potential
Sµ = (S0 , ⃗S). Let m1 , m2 be the masses of quark and antiquark, ψ 1 and ψ 2 are their wave
functions. Then the spectral equations for mesons take the form
h̄2 I gs J2
− ∇ + i ⃗S ψ k + gs S0 ψ k +⃗µk · curl⃗Sψ k
2mk h̄c
= λ ψ k in Ω = {x ∈ R3 | ρ0 < |x| < ρ1 }, k = 1, 2, (6.4.49)
(ψ 1 , ψ 2 ) = 0 on ∂ Ω,

and ψ k (k = 1, 2) satisfy the normalization.

3. Physical parameters of nucleons. In hadrons, only nucleons (protons and neutrons)
are long life-time. Therefore the parameters of nucleons are very important.
By (6.4.7), the strong interaction potential for quarks is taken as
! "3 E ! " F
ρw 1 Aq r
S0 = gs − 1+ e−r/r0 , (6.4.50)
ρ0 r ρ0 r0
where r0 = 10−16 cm, ρ0 is the quark radius, Aq the constant of quarks.
Similar to the relation (6.4.43), from (6.4.48) and (6.4.50) we can obtain the parameter
relation of nucleons as follows
! "6 ! "
1 2 ρw 2 1 κn
λ n = mq v + 2 gs − (6.4.51)
2 ρ0 rn ρ0
where λn is the bound energy of nucleons, mq the quark masses, v the average velocity of
quarks, rn the nucleon radius, κn the bound parameter of nucleons, these parameters are
expressed as
! ρ1
rn = rϕn2 (r)dr,
ρ0
! ρ1 ! "
2 r −r/r0 2
κn = Aq r 1+ e ϕn (r) dr, (6.4.52)
ρ0 r0
I m v J2 ! ρ1 ! d ϕ "2
q n
= r2 dr,
h̄ ρ0 dr
where λn and ϕn are the first eigenvalue and eigenfucntion of the following equation
! "
h̄2 1 d 2 d
− r ϕn + 2gsS0 ϕn = λ ϕn ,
2mq r2 dr dr (6.4.53)
ϕn (ρ0 ) = ϕn (ρ1 ) = 0.
Remark 6.19 The two sets of parameters (6.4.47) and (6.4.52) can characterize the
physical properties of naked electrons and naked nucleons respectively. For the real elec-
trons and nucleons, we have to consider their mediator clouds.
388 Chapter 6 Quantum Physics

6.4.5 Energy spectrum of mediators

Mediators are massless bosons, which consists of a weakton and its anti-particle. From
the viewpoint of bound energy, gluons are generated by both weak and strong interactions,
and others are bound only by weak interaction. Hence, the spectral equations of gluons are
different from those the other mediators.

1. Gluons. The weakton constituent of gluons is given by

g = w∗ w∗ . (6.4.54)

Based onthe weakton model, w∗ and w∗ contain a weak charge gw and a strong charge gs .
By (6.4.6) and (6.4.9),
g2w
= 0.25,
g2s
i.e. gw and gs have the same order. Therefore, the interactions for the gluon are both weak
and strong forces, i.e. the 4-dimensional potential A µ = (A0 , ⃗A) is as

gA0 = gwW0 + gsS0 , g⃗A = gwW

⃗ + gs⃗S. (6.4.55)

In (6.4.54), we only need to consider the bound state for a single weakton. Then the
spectral equation is provided by (6.4.33)-(6.4.34), and for (6.4.55) which is written as

− h̄cD2 ψ + [gw⃗σ · curlW⃗ + gs⃗σ · curl⃗S]ψ

i 6 7
+ (⃗σ · ⃗D), gwW0 + gs S0 ψ = iλ (⃗σ · ⃗D)ψ for ρw < |x| < ρg , (6.4.56)
2
ψ =0 for |x| = ρw , ρg ,

where ρw and ρg are the radii of weaktons and gluons, and

⃗D = ∇ + i g⃗A, g⃗A as in (6.4.55).

h̄c

2. Photons and ν -mediators. The other mediators such as photons and ν -mediator
consist of a pair of weakton and anti-weakton:

γ = α1 w1 w1 + α2 w2 w2 ,
(6.4.57)
ν = αe νe ν e + αµ νµ ν µ + ατ ντ ν τ .

The weaktons in (6.4.57) only contain a weak charge g w , hence the bound energy of γ and
ν is given by the weak force, i.e.

⃗ ).
gAµ = gwWµ = (gwW0 , gwW
6.4 Energy Levels of Subatomic Particles 389

In this case, the spectral equations are in the form

igw
− h̄cD2 ψ +gw⃗σ · curlW
⃗ ψ+ {(⃗σ · ⃗D),W0 }ψ
2
= iλ (⃗σ · ⃗D)ψ , for ρw < |x| < ρm , (6.4.58)
ψ = 0, on |x| = ρw , ρm ,

where ρm is the radius of the mediator γ or ν , and

⃗D = ∇ + i gwW
⃗.
h̄c

3. Energy levels of mediators. By the spectral theory of the Weyl operator established
in Subsection 2.6.5, the negative eigenvalues of (6.4.56) and (6.4.57) are finite, i.e.

−∞ < λ1 ! · · · ! λN < 0,

which stand for bound energy of mediators. It shows that the energy levels of each kind
mediator are finite
0 < E1 ! · · · ! EN . (6.4.59)
Each energy level Ek can be expressed as

Ek = E0 + λ k (1 ! k ! N), (6.4.60)

where E0 is the intrinsic energy of mediators.

It follows from (6.4.59) and (6.4.60) that the frequencies of mediators are finite and
discrete
1
ωk = Ek (1 ! k ! N), (6.4.61)
h̄
and the difference of two adjacent frequencies is
1
∆ωk = ωk+1 − ωk = (λk+1 − λk ). (6.4.62)
h̄
Remark 6.20 In the classical quantum mechanics, the frequencies of particles are con-
tinuously distributed. Here we derive from the energy spectrum theory that the frequencies
are finite and discrete. In fact, by the estimate (3.6.51) of the number of negative eigenval-
ues, we can verify that the frequency difference (6.4.62) is too small to measure in the next
subsection.

6.4.6 Discreteness of energy spectrum

Based on the spectral theory developed in Section 3.6, the energy levels of all subatomic
particles are finite and discrete:

0 < E1 < · · · < EN , (6.4.63)

390 Chapter 6 Quantum Physics

where the number N of energy levels depends on the particle type. Each subatomic particle
lies in an energy state of Ek (1 ! k ! N) in (6.4.63). In traditional conception, the energy
distribution of all particles is infinite and continuous, i.e. a particle can lie in any state of
energy E with 0 < E < ∞. Hence the energy level theory established here arrive at a very
different conclution:

Physical Conclusion 6.21 The energy distribution of subatomic particles is finite and
discrete.

Usually, we cannot observe the discreteness of energy because the number N of energy
levels is so large that the average difference of adjacent energy level:
EN − E1
∆Ek = Ek+1 − Ek ≃ ≃ 0,
N
is very small. In the following discussion we shall show this point.
To this end, we first give the estimates of the number N of energy levels for various
types of subatomic particles.
1. Energy level number of electrons and quarks. To consider the approximatively com-
putation of number N of energy levels, we always ignore the magnetic effects. In this case,
the spectral equations for charged leptons and quarks are given by (6.4.42), and which are
written as
h̄2
− ∆ψ + 2gwW0 ψ = λ ψ , in ρw < |x| < ρ , (6.4.64)
2mw
ψ = 0, on |x| = ρw , ρ ,
where mw is the masses of weaktons in leptons and quarks, ρ is the attracting radius of weak
interaction.
For the weak interaction potential W0 , we approximatively take
g w Bw
W0 = − .
ρw
Then take the dimensional transformation

x → ρx (ρ as in (6.4.64)).

Note that the weakton radius ρw for smaller than ρ ,

ρw ≪ ρ .

Hence, the problem (6.4.64) can be approximatively expressed as

4mw Bw ρ 2 2
−∆ψ − gw ψ = λ ψ for |x| < 1,
h̄2 ρw (6.4.65)
ψ =0 on |x| = 1.
6.4 Energy Levels of Subatomic Particles 391

It is clear that for the equation (6.4.65), the parameters r, α , θ as in (3.6.28) of Theorem
2.37 are as follows
4mw Bw ρ 2 2
r = 1, α = 0, θ = gw
h̄2 ρw
Thus the number N of the energy levels of charged leptons and quarks is approximatively
given by
E F3
4 Bw ρ 2 mw c g2w 2
N= , (6.4.66)
λ1 ρw h̄ h̄c
where λ1 is the first eigenvalue of −∆ in unit ball B1 ⊂ R3 .
By the de Broglie relation,
mw c 1
∼ ,
h̄ ̸λ
where λ
̸ is the wave length of weaktons, i.e. λ
̸ = nrm (n = 1, 2, · · · ), rm the mediator radius.
We take ̸λ = rm ≃ ρ , and

4 Bw ρ mw c
≃ 1, ≃ 102 , ρ ≃ 1.
λ1 ρw h̄

Then, by (6.4.6) the number N in (6.4.66) has the estimate

! "9
ρn
N∼ " 1045 . (6.4.67)
ρw

This number is very large because (ρn /ρw ) " 105 .

We remark that the estimate (6.4.67) only for the naked leptons and quarks. For the
5 should take as
natural leptons and quarks, their energy level number N

5 = N · N1 ,
N

where N is as in (6.4.67) and N1 the energy level number of mediator cloud around the
charged leptons and quarks.

2. Energy level number of hadrons. We only consider the case of baryons, and the case
of mesons is similar. For the baryons, spectral equation (6.4.48) can be approximatively
reduced in the form

4mq Aq ρ12 2
−∆ψ − gs ψ = λ ψ , for |x| < 1,
h̄2 ρq (6.4.68)
ψ = 0, on |x| = 1.

where mq is the quark mass, ρq the quark radius, ρ1 the strong attracting radius, and Aq the
strong interaction constant of quarks.
392 Chapter 6 Quantum Physics

By Theorem 3.38, from (6.4.68) we can get the estimates of energy level number N of
baryons as follows
E F3
4 ρ12 Aq mq c g2s 2
N= , (6.4.69)
λ1 ρq h̄ h̄c
where λ1 is as in (6.4.66). By (6.4.69) and (6.4.9) we can get the same estimate as in
(6.4.67).
3. Energy level number of mediators. Likewise, we only consider the energy level
number of photons. In this case, the spectral equation (6.4.58) can be reduced as
! "
Bw ρ r 2
−∆ϕ = i λ + gw (⃗σ · ∇)ϕ , for |x| < 1,
h̄cρw (6.4.70)
ϕ =0 on |x| = 1,
where ργ is the photon radius.
By (6.4.70) we can see that the parameter K as in (3.6.51) takes
Bw ργ g2w
K= .
ρw h̄c
Thus, by (3.6.51) the energy level number N of photons is given by
! "3 E F3
K 1 Bw ργ g2w
N= = , (6.4.71)
β1 β1 ρw h̄c
where β1 is as in (3.6.51), and β1 ∼ o(1).
From the physical significance, ργ is approximatively the weak attracting radius of
weaktons, and ρw /Bw = ρ satisfying
P
> 0 for 0 < r < ρ ,
Fw
< 0 for ρ < r < ργ .
Hence, it is natural to think that
Bw ρ γ ργ
= = 102 ∼ 104 .
ρw ρ
Thus, by (6.4.71) and (6.4.6) we get hat
! "18
ρn
N∼ " 1090. (6.4.72)
ρw
Remark 6.22 In the estimates (6.4.66), (6.4.69) and (6.4.71), the energy level numbers
N for various subatomic particles are counting the multiplicities of eigenvalues. However,
the numbers N have the same order as the real energy level numbers. It is because that the
multiple eigenvalues are unstable, under the perturbation of electromagnetism together with
weak and strong magnetism, most multiple eigenvalues become the simple eigenvalues.
6.4 Energy Levels of Subatomic Particles 393

4. Energy level gradient of photons. Each energy level E k (1 ! k ! N) of photons can

be written as
Ek = E0 + λ k (1 ! k ! N).

It is clear that the largest and smallest energy levels are given by

Emax = E0 + λN , Emin = E0 + λ1.

The total energy level difference is

Emax − Emin = λN − λ1 .

Since |λ1 | ≫ |λN |, the average energy level gradient (for two adjacent energy levels) is
approximatively given by
Emax − Emin |λ1 |
∆E = ≃ . (6.4.73)
N N
The first eigenvalue λ1 of (6.4.70) is
! "
h̄c
λ1 ≃ −K the unit is .
ργ

Hence, by (6.4.71) and (6.4.73)

h̄c Bw ργ g2w
∆E = β13 K −2 , K= .
ργ ρw h̄c

Then we get the estimates

! "12
ρw h̄c
∆E = . (6.4.74)
ρn ργ
As we take
ρw
= 10−5 , ργ = 10−20cm,
ρn
then from (6.4.74) we derive that

∆E = 10−40 h̄c/cm = 2 × 10−45eV.

This is a very small value, and it is impossible for experiments to measure.

Remark 6.23 The physical conclusion that all particles have finite and discrete energy
distribution is of very important significance for the quantum field theory. It implies that
the infinity appearing in the field quantization does not exist, and the renormalization theory
needs to be reconsidered.
394 Chapter 6 Quantum Physics

6.5 Field Theory of Multi-Particle Systems

6.5.1 Introduction
We start with the known model of multi-particle systems. Consider an N-particle system
with particles
A1 , · · · , AN . (6.5.1)

Let xk = (x1k , x2k , x3k ) ∈ R3 be the coordinate of Ak , and

ψ = ψ (t, x1 , · · · , xk ) (6.5.2)

be the wave function describing the N-particle system (6.5.1). Then, the classical theory for
(6.5.1) is provided by the Schrödinger equation

∂ψ N
h̄2
ih̄ =−∑ ∆k ψ + ∑ V (x j , xk )ψ , (6.5.3)
∂t k=1 2mk j̸=k

where V (x j , xk ) is the potential energy of interactions between A j and Ak , mk is the mass of

Ak , and
∂2 ∂2 ∂2
∆k = + + .
(∂ xk )
1 2 (∂ xk )
2 2 (∂ x3k )2
The wave function ψ satisfies the normalization condition
! !
··· |ψ |2 dx1 · · · dxN = 1.
R3 R3

Namely, the physically |ψ (t, x1 , · · · , xN )|2 represents the probability density of A1 , · · · ,

AN appearing at x1 , · · · , xN at time t.
It is clear that the Schrödinger equation (6.5.3) for an N-particle system is only an ap-
proximate model:

• It is non-relativistic model;

• The model does not involve the vector potentials ⃗A of the interactions between parti-
cles.

• By using coordinate xk to represent the particle Ak amounts essentially to saying that

the wave function ψ satisfying (6.5.3) can only describe the statistic properties of the
system (6.5.1), and contains no information for each individual particle A k (1 ! k !
N).

• The model is decoupled with interaction fields, i.e. the interaction fields in the model
are treated as given functions.
6.5 Field Theory of Multi-Particle Systems 395

In fact, the most remarkable characteristic of interacting multi-particle systems is that

both particle fields and interaction fields are closely related. Therefore, a complete field
model of multi-particle systems have to couple both the particle field equations and the
interaction field equations. In particular, a precise unified field theory should be based on
the field model of the multi-particle system coupled with the four fundamental interactions.

6.5.2 Basic postulates for N-body quantum physics

As mentioned in the last subsection, the dynamic models for multi-particle quantum systems
have to couple both particle and interaction fields. Therefore there should be some added
quantum rules for the systems. In the following we propose the basic postulates for N-
particle quantum systems.
First of all, the physical systems have to satisfy a few fundamental physical principles
introduced below.
Postulate 6.24 Any N-particle quantum system has to obey the physical fundamental
principles such as:
Einstein General Relativity,
Lorentz Invariance,
Gauge Invariance,
(6.5.4)
Gauge Representation Invariance (PRI),
Principle of Lagrange Dynamics (PLD),
Principle of Interaction Dynamics (PID),
where the gauge invariance means the invariance of the Lagrangian action under corre-
sponding gauge transformations.

We note that in general multi-particle systems are layered, and may consist of numerous
sub-systems. In particular, we know that the weak and strong interactions are also layered.
Hence, here we consider the same level systems, i.e. the systems which consist of identical
particles or sub-systems possessing the same level of interactions.
For multi-particle systems with N same level subsystems Ak (1 ! k ! N), the energy
contributions of Ak are indistinguishable. Hence, the Lagrangian actions for the N-particle
systems satisfy SU(N) gauge invariance. Thus we propose the following basic postulate:
Postulate 6.25 An N-particle system obeys the SU(N) gauge invariance, i.e. the
Lagrangian action of this system is invariant under the SU(N) gauge transformation
⎛ ⎞ ⎛ ⎞
ψ
51 ψ1
⎜ .. ⎟ ⎜ .. ⎟
⎝ . ⎠ = Ω ⎝ . ⎠ , Ω ∈ SU(N), (6.5.5)
ψ
5N ψN
where ψ1 , · · · , ψN are the wave functions of the N particles.
396 Chapter 6 Quantum Physics

We now need to explain the physical significance of the SU(N) gauge fields induced by
Postulate 6.25.
Let each particle of the N-particle system carry an interaction charge g (for example
a weak charge g = gw ). Then, there are interactions present between the N particles. By
the SU(N) gauge theory, the gauge invariant 4-dimensional energy-momentum operator is
given by
Dµ = ∂µ + igGaµ τa for 1 ! a ! N 2 − 1, (6.5.6)
and the interaction energy generated by the N particles is
.
Ψ(iγ µ Dµ Ψ) for fermions,
E= (6.5.7)
|Dµ Ψ|2 for bosons,

where Ψ = (ψ1 , · · · , ψN )T , and Dµ is as in (6.5.6). From (6.5.6) and (6.5.7) we obtain the
physical explanation to the SU(N) gauge fields Gaµ , stated in the following postulate:

Postulate 6.26 For a system of N-particles in the same level with each particle carrying
an interaction charge g, the N particles induce dynamic interactions between them, and the
SU(N) gauge fields
gGaµ for 1 ! a ! N 2 − 1 (6.5.8)
stand for the interaction potentials between the N particles.

The N particles induce dynamic interactions between them in terms of the SU(N) gauge
fields (6.5.8). These interaction fields cannot be measured experimentally because they
depend on the choice of generator representation τa of SU(N). By the SU(N) geometric
theory in Section 3.5, there is a constant SU(N) tensor

αaN = (α1N , · · · , αNN ), (6.5.9)

such that the contraction field using PRI

Gµ = αaN Ga (6.5.10)

is independent of the SU(N) representation τa . The field (6.5.10) is the interaction field
which can be experimentally observed. Thus we propose the following basic postulate.

Postulate 6.27 For an N-particle system, only the interaction field given by (6.5.10)
can be measured, and is the interaction field under which this system interacts with other
external systems.

Remark 6.28 Postulates 6.24-6.27, together with the Principle of Symmetry-Breaking

2.14 and the Postulates 6.1-6.5, form a complete foundation for quantum physics. In fact,
without Postulates 6.24-6.27, we cannot establish the quantum physics of multi-particle
systems.
6.5 Field Theory of Multi-Particle Systems 397

The main motivation to introduce Postulates 6.25 and 6.26 are as follows. Consider an
N-particle system with each particle carrying an interaction charge g. Let this be a fermionic
system, and the Dirac spinors be given by

Ψ = (ψ1 , · · · , ψN )T .

By Postulates 6.3 and 6.5, the Dirac equations for this system can be expressed in the general
form
iγ µ Dµ Ψ + MΨ = 0, (6.5.11)

where M is the mass matrix, and

⎛ ⎞ ⎛ 11 ⎞⎛ ⎞
ψ1 Gµ ··· G1N
µ ψ1
⎜ .. ⎟ ⎜ .. .. ⎟ ⎜ .. ⎟ ,
Dµ Ψ = ∂µ ⎝ . ⎠ + ig ⎝ . . ⎠⎝ . ⎠ (6.5.12)
ψN GN1
µ ··· NN
Gµ ψN

where G = (Giµj ) is an Hermitian matrix, representing the interaction potentials between the
N particles generated by the interaction charge g.
Notice that the space consisting of all Hermitian matrices

H(N) = {G| G is an N-th order Hermitian matrix}

is an N 2 −dimensional linear space with basis

τ0 , τ1 , · · · , τK with K = N 2 − 1, (6.5.13)

where τ0 = I is the identity, and τa (1 ! a ! N 2 − 1) are the traceless Hermitian matrices.

ij
Hence, the Hermitian matrix G = (G µ ) ∈ H(N) in (6.5.12) can be expressed as

G = G0µ I + Gaµ τa with τa as in (6.5.13).

Thus, the differential operator in (6.5.12) is in the form

Dµ = ∂µ + igG0µ + igGaµ τa . (6.5.14)

The equations (6.5.11) with (6.5.14) are just the Dirac equations in the form of SU(N)
gauge fields {Gaµ | 1 ! a ! N 2 − 1} with a given external interaction field G0µ . Thus, based
on Postulate 6.24, the gauge invariance of an N-particle system and the expressions (6.5.11)
and (6.5.14) of the N fermionic particle field equations dictate Postulates 6.25 and 6.26.
The derivation here indicates that Postulates 6.25 and 6.26 can be considered as the
consequence of 1) the gauge invariance stated in Postulate 6.24, and 2) the existence of
interactions between particles as stated in (6.5.12), which can be considered as an axiom.
398 Chapter 6 Quantum Physics

6.5.3 Field equations of multi-particle systems

Based on the basic axioms given by Postulates 6.24-6.27, we can establish field equations
for various levels of N-particle systems. We proceed in several different cases.

Fermionic systems

Consider N fermions at the same level with interaction charge g, the wave functions
(Dirac spinors) are given by

Ψ = (ψ1 , · · · , ψN )T , ψk = (ψk1 , ψk2 , ψk3 , ψk4 )T for 1 ! k ! N, (6.5.15)

with the mass matrix ⎛ ⎞

m1 0
⎜
M=⎝ .. ⎟
⎠. (6.5.16)
.
0 mN
By Postulates 6.24 and 6.25, the Lagrangian action for the N-particle system (6.5.15)-
(6.5.16) must be in the form !
L= (LG + LD )dx, (6.5.17)

where LG is the sector of the SU(N) gauge fields, and LD is the Dirac sector of particle
fields:
1
LG = − Gab gµα gνβ Gaν µ Gbαβ ,
4h̄c
E ! " F (6.5.18)
ig ig c
LD = Ψ iγ µ ∂µ + G0µ + Gaµ τa − M Ψ,
h̄c h̄c h̄
where Gaµ (1 ! a ! N 2 −1) are the SU(N) gauge fields representing the interactions between
the N particles, τa (1 ! a ! N 2 − 1) are the generators of SU(N), and

1
Gab = Tr(τa τb† ),
2
g a b c
Gaµν = ∂µ Gaν − ∂ν Gaµ + λ G G .
h̄c bc µ ν
According to PID and PLD, for the action (6.5.17) the field equations are given by

δL
= Dµ φa by PID,
δ Gaµ
(6.5.19)
δL
=0 by PLD,
δΨ
where Dµ is the PID gradient operator given by
! "
1 1 gα gβ 0
Dµ = ∂ µ − k2 x µ + Gµ + Gµ ,
h̄c 4 h̄c h̄c
6.5 Field Theory of Multi-Particle Systems 399

Gµ is as in (6.5.10), α and k are parameters, k −1 stands for the range of attracting force of
I gα J−1
the interaction, and is the range of the repelling force.
h̄c
Thus, by (6.5.18) and (6.5.19) we derive the field equations of the N-particle system
(6.5.15)-(6.5.16) as follows
M g b αβ c d N
Gab ∂ ν Gbν µ − λcd g Gα µ Gβ − gΨγµ τa Ψ (6.5.20)
h̄c
E F
1 gα gβ 0
= ∂ µ − k2 x µ + Gµ + G φa for 1 ! a ! N 2 − 1,
4 h̄c h̄c µ
⎛ ⎞ ⎛ ⎞
E F ψ1 ψ1
ig ig ⎜ . ⎟ c ⎜ ⎟
⎟ − M ⎜ .. ⎟ = 0,
iγ µ ∂µ + G0µ + Gaµ τa ⎜ .
.
⎝ ⎠ h̄ ⎝ ⎠ . (6.5.21)
h̄c h̄c
ψN ψN

where γµ = g µν γ ν , and G0µ is the interaction field of external systems. It is by this field G0µ
that we can couple external sub-systems to the model (6.5.20)-(6.5.21).

Remark 6.29 In the field equations of multi-particle systems there is a gauge fixing
problem. In fact, we know that the action (6.5.17)-(6.5.18) is invariant under the gauge
transformation
I J ! "
Ψ,
5 G5aµ τa = eiθ a τa Ψ, Gaµ eiθ b τb τa e−iθ b τb − 1 ∂µ θ b τb . (6.5.22)
g

Hence if (Ψ, Gaµ ) is a solution of

δ L = 0, (6.5.23)

then (Ψ,
5 G 5aµ ) is a solution of (6.5.23) as well. In (6.5.22) we see that G
5aµ have N 2 − 1 free
functions
θ a (x) with 1 ! a ! N 2 − 1. (6.5.24)

In order to eliminate the N 2 − 1 freedom of (6.5.24), we have to supplement N 2 − 1 gauge

fixing equations for the equation (6.5.23). Now, as we replace the PLD equation (6.5.23).
By the PID equations (6.5.19), (6.5.22) breaks the gauge invariance. Therefore the N 2 − 1
freedom of (6.5.24) is eliminated. However, in the PID equations (6.5.19) there are addi-
tional N 2 − 1 new unknown functions φa (1 ! a ! N 2 − 1). Hence, the gauge fixing problem
still holds true. There are two possible ways to solve this problem:

1) there might exist some unknown fundamental principles, which can provide the all or
some of the N 2 − 1 gauge fixing equations; and

2) there might be no general physical principles to determine the gauge fixing equations,
and these equations will be determined by underlying physical system.
400 Chapter 6 Quantum Physics

Bosonic systems

Consider N bosons with charge g, the Klein-Gordon fields are

Φ = (ϕ1 , · · · , ϕN )T ,

and the mass matrix is given by (6.5.16). The action is

!
L= (LG + LKG )dx (6.5.25)

where LG is as given by (6.5.18), and LKG is the Klein-Gordon sector given by

1 1 I c J2
LKG = |Dµ Φ|2 + |MΦ|2
2 2 h̄
ig ig
Dµ = ∂µ + G0µ + Gaµ τa .
h̄c h̄c
Then, the PID equations of (6.5.25) are as follows
M g b αβ c d N ig K L
Gab ∂ ν Gbν µ − λcd g Gα µ Gβ + (Dµ Φ)† (τa Φ) − (τa Φ)† (Dµ Φ) (6.5.26)
E h̄c 2 F
1 2 g g
= ∂µ − k xµ + α Gµ + β G0µ φa for 1 ! a ! N 2 − 1,
4 h̄c h̄c
⎛ ⎞ ⎛ ⎞
ϕ1 I c J2 ϕ1
µ ⎜ .. ⎟ 2 ⎜ .. ⎟
D Dµ ⎝ . ⎠ − M ⎝ . ⎠ = 0. (6.5.27)
h̄
ϕN ϕN

Mixed systems

Consider a maxed system consisting of N1 fermions with n1 charges g and N2 bosons

with n2 charges g, and the fields are

Dirac fields: Ψ = (ψ1 , · · · , ψN1 )T ,

Klein-Gordon fields: Φ = (ϕ1 , · · · , ϕN2 )T .

The interaction fields of this system are SU(N1 )× SU(N2 ) gauge fields, SU(N1 ) gauge fields
are for fermions, and SU(N2 ) for bosons:

{Gaµ |1 ! a ! N12 − 1} for Dirac fields Ψ,

5kµ | 1 ! k ! N22 − 1}
{G for Klein-Gordon fields Φ.

The action is given by ! K L

L= LG1 + LG2 + LD + LKG dx, (6.5.28)
6.5 Field Theory of Multi-Particle Systems 401

where LG1 and LG2 are the sectors of SU(N1 ) and SU(N2 ) gauge fields as given in (6.5.18)
with N = N1 and N = N2 respectively.
Define the two total gauge fields of SU(N1 ) and SU(N2 ), as defined by (6.5.9)-(6.5.10):

Gµ = αaN1 Gaµ for 1 ! a ! N12 − 1,

(6.5.29)
5µ = α N2 G
G 5kµ for 1 ! k ! N22 − 1.
k

Namely, LD and LKG are given by

E ! " F
in1 g 0 in1 g 5 in1 g a 1 c
LD = Ψ iγ µ ∂µ + Gµ + Gµ + Gµ τa − M1 Ψ,
h̄c h̄c h̄c h̄
! " /
1 // in2 g 0 in2 g in2 g 5k 2 / 2 1 I c J2
LKG = / ∂µ + Gµ + Gµ + Gµ τk Φ/ + |M2 Φ|2 ,
2 h̄c h̄c h̄c 2 h̄

where G0µ is the external field, and G 5µ and Gµ are as in (6.5.29).

Thus, we derive the field equations for mixed multi-particle systems expressed in the
following form
M n1 g b αβ c d N n1 g
1
Gab ∂ ν Gbν µ − λ g Gα µ Gβ − Ψγµ τa1 Ψ (6.5.30)
h̄c 1cd h̄c
in2 g K L
+ (Dµ Φ)∗ (αaN1 Φ) − (αaN1 Φ)∗ (Dµ Φ)
E h̄c F
1 n1 g n2 g 5
= ∂µ − k12 xµ + α1 Gµ + α2 Gµ φa for 1 ! a ! N 2 − 1,
4 h̄c h̄c
M N
Gkl2 ∂ ν G 5lν µ − n2 g λ2il j gαβ G 5 j − n1 g α N2 Ψγµ Ψ
5iα µ G (6.5.31)
h̄c β h̄c k
in1 g K L
+ (Dµ Φ)† (τk2 Φ) − (τk2 Φ)† (Dµ Φ)
E 2h̄c F
1 2 n1 g n2 g 5 5
= ∂ µ − k2 x µ + β1 G µ + β2 Gµ φk for 1 ! k ! N22 − 1,
4 h̄c h̄c
! "
µ in1 g 0 in1 g 5 in1 g a 1 c
iγ ∂µ + Gµ + Gµ + Gµ τa Ψ − M1 Ψ = 0, (6.5.32)
h̄c h̄c h̄c h̄
I c J2
gµν Dµ Dν Φ − M22 Φ = 0, (6.5.33)
h̄
5µ are as in (6.5.29), and D µ is defined by
where Gµ and G
in2 g 0 in2 g in1 g 5k 2
Dµ = ∂µ + G + Gµ + G τ .
h̄c µ h̄c h̄c µ k
We remark here that the coupling interaction between fermions and bosons is directly
represented on the right hand side of gauge field equations (6.5.30) and (6.5.31), due to
the presence of the dual interaction fields based on PID. Namely, the interactions between
particles in an N-particle system are achieved through both the interaction gauge fields and
the corresponding dual fields. This fact again validates the importance of PID.
402 Chapter 6 Quantum Physics

Another remark is that the gauge actions LG1 and LG2 in (6.5.28) obey the gauge invari-
ance, but sectors LD and LKG break the gauge symmetry, due to the coupling of different
level physical systems. Namely, the Principle of Symmetry-Breaking 2.14 holds true here.
In addition, the field equations (6.5.30) and (6.5.31) spontaneously break the gauge symme-
5µ on the right-sides of the field equations.
try, due essentially to the fields G µ and G

Layered systems

Let a system be layered consisting of two levels: 1) level A consists of K sub-systems

A1 , · · · , AK , and 2) level B is level inside of each sub-system A j , which consists of N particles
j j
B1 , · · · , BN :
at level A : A = {A1 , · · · , AK },
(6.5.34)
at level B : A j = {B1j , · · · , BNj } for 1 ! j ! K.

Each particle Bij carries n charges g.

Let the particle field functions be

at level A : ΨA = (ψA1 , · · · , ψAK ),

at level B : ΨB j = (ψB j1 , · · · , ψB jN ) for 1 ! j ! K.

The interaction is the SU(K) × SU(N) gauge fields:

at level A : SU(K) gauge fields Aaµ 1 ! a ! K 2 − 1,

at level B : SU(N) gauge fields (B j )kµ 1 ! k ! N 2 − 1.

Without loss of generality, we assume A and B are the fermion systems. Thus the action
of this layered system is
!
G H
K K
L= LAG + ∑ LB j G + LAD + ∑ LB j D dx, (6.5.35)
j=1 j=1

where

LAG = the sector of SU(K) gauge fields,

E ! " F
inN 0 inN inN a K c
LAD = ΨA iγ µ ∂µ + gGµ + Bµ + gAµ τa − MA ΨA ,
h̄c h̄c h̄c h̄
(6.5.36)
LB j G = the j-th the sector of SU(N) gauge fields,
E ! " F
ing 0 ing ing c
LB j D = ΨB j iγ µ ∂µ + Gµ + Aµ + (B j )kµ τkN − MB j ΨB j ,
h̄c h̄c h̄c h̄

where G0µ is the external field. The corresponding PID field equations of the layered multi-
particle system (6.5.34) follow from (6.5.35) and (6.5.36), and here we omit the details.
6.5 Field Theory of Multi-Particle Systems 403

Remark 6.30 Postulate 6.27 is essentially another expression of PRI, which is very
crucial to couple all sub-systems together to form a complete set of field equations for a
given multi-particle system. In particular, this approach is natural and unique to derive
models for multi-particle systems, satisfying all fundamental principles of (6.5.4), the Prin-
ciple of Symmetry-Breaking 2.14, and the gauge symmetry breaking principle (Principle
4.4). It is also a unique way to establish a unified field theory coupling the gravity and other
interactions in various levels of multi-particle systems. In the next subsection we discuss
this topic.

6.5.4 Unified field model coupling matter fields

In Chapter 4, we have discussed the unified field theory, in which we consider two aspects:
1) the interaction field particles, and 2) the interaction potentials. Hence, it restricted the
unified field model to be the theory based on

Einstein relativity + U(1) × SU(2) × SU(3) symmetry. (6.5.37)

However, if we consider the interaction potentials between the particles of N-particle

systems, then the unified field theory has to be based on the following symmetries instead
of (6.5.37):
Einstein relativity + SU(N1) × · · · × SU(NK ) symmetry, (6.5.38)

where N1 , · · · , NK are the particle numbers of various sub-systems and layered systems.
The two types of unified field models based on (6.5.37) and (6.5.38) are mutually com-
plementary. They have different roles in revealing the essences of interactions and particle
dynamic behaviors.
In this subsection, we shall establish the unified field model of multi-particle systems
based on (6.5.38), which matches the vision of Einstein and Nambu. In his Nobel lecture
(Nambu, 2008), Nambu stated that
Einstein used to express dissatisfaction with his famous equation of gravity

Gµν = 8π Tµν

His point was that, from an aesthetic point of view, the left hand side of the
equation which describes the gravitational field is based on a beautiful geo-
metrical principle, whereas the right hand side, which describes everything
else, . . . looks arbitrary and ugly.

... [today] Since gauge fields are based on a beautiful geometrical prin-
ciple, one may shift them to the left hand side of Einstein s equation. What
is left on the right are the matter fields which act as the source for the gauge
fields ... Can one geometrize the matter fields and shift everything to the left?
404 Chapter 6 Quantum Physics

The gravity will be considered only in systems possessing huge amounts of particles,
which we call gravitational systems. Many gravitational systems have very complicated
structures. But they are composites of some simple systems. Here we only discuss two
cases.

Systems with gravity and electromagnetism

Consider the system consisting of N1 fermions with n1 electric charges n1 e and N2

bosons with n2 charges n2 e:

Ψ = (ψ1 , · · · , ψN1 ) for fermions,

Φ = (ϕ1 , · · · , ϕN2 ) for bosons.

The action is given by

! E F
c4 N1 N2 √
L= R + LA + LA + h̄cLD + h̄cLKG −gdx (6.5.39)
8π G

where R is the scalar curvature, G is the gravitational constant, g = det(g µν ), LAN1 and LAN2
are the sectors of SU(N1 ) and SU(N2 ) gauge fields for the electromagnetic interaction

1
LAN1 = − Gab gµα gνβ Aaµν Abαβ 1 ! a, b ! N12 − 1,
4
1
LAN2 = − G5kl gµα gνβ A 5kµν A
5l
αβ 1 ! k, l ! N22 − 1,
4 (6.5.40)
n1 e a b c
Aaµν = ∂µ Aaν − ∂ν Aaµ + λ A A n1 ∈ Z,
h̄c bc µ ν
5kµν = ∂µ A
A 5kµ + n2 e 5
5kν − ∂ν A λkA 5j
5i A n2 ∈ Z,
h̄c i j µ ν

and LD , LKG are the Dirac and Klein-Gordon sectors:

E I F
µ in1 e 0 in1 e a J c
LD = Ψ iγ ∂µ + A + A τa − M1 Ψ,
h̄c µ h̄c µ h̄
1 1 cI J2
LKG = gµν (Dµ Φ)† (Dν Φ) + |M2 Φ|2 , (6.5.41)
2 2 h̄
in2 e 0 in2 e 5k
Dµ = ∇µ + A + A τ5k ,
h̄c µ h̄c µ

where M1 and M2 are the masses, ∇µ is the covariant derivative, and A0µ is the external
electromagnetic field.
6.5 Field Theory of Multi-Particle Systems 405

Based on PID and PLD, the field equations of (6.5.39) are given by
δ c4 G g
L= D φ , (PID)
δ g µν 8π G µ ν
δ
L = DAµ φa , (PID)
δ Aaµ
δ ! (6.5.42)
L = DAµ φ5k , (PID)
δ Aµ
5 k

δ
L = 0, (PLD)
δΨ
δ
L = 0, (PLD)
δΦ
where
n1 e n2 e 5
DG
µ = ∇µ + Aµ + Aµ ,
h̄c h̄c
1 n1 e n2 e 5
DAµ = ∂µ − k12 xµ + α Aµ + α
5 Aµ , (6.5.43)
4 h̄c h̄c
! 1 n1 e n2 e 5 5
DAµ = ∂µ − k22 xµ + β Aµ + β Aµ .
4 h̄c h̄c
Here Aµ = αaN1 Aaµ and A 5µ = α N2 A
k
5kµ are the total electromagnetic fields generated by the
fermion system and the boson system.
By (6.5.39)-(6.5.41), the equations (6.5.42)-(6.5.43) are written as
1 8π G I n1 e n2 e 5 J g
Rµν − gµν R = − 4 Tµν + ∇µ + Aµ + Aµ φν , (6.5.44)
M 2 c
n1 e b αβ c d N h̄c h̄c
Gab ∂ ν Abν µ − λ g Aα µ Aβ − n1 eΨγµ τa Ψ (6.5.45)
E h̄c cd F
1 n1 e n2 e 5
= ∂µ − k12 xµ + α Aµ + α
5 Aµ φa ,
4 h̄c h̄c
M N i
5lν µ − n2 e 5
K L
G5kl ∂ ν A λilj gαβ A5iα µ A5βj + n2 e (Dµ Φ)† (τ5k Φ) − (τ5k Φ)† (Dµ Φ) , (6.5.46)
E h̄c 2
F
1 n1 e n2 e 5 5 5
= ∂µ − k22 xµ + β Aµ + β Aµ φk ,
4 h̄c h̄c
E F
in 1 e in 1 e c
iγ µ ∂µ + A0µ + Aaµ τa Ψ − M1 Ψ = 0, (6.5.47)
h̄c h̄c h̄
I c J2
gµν Dµ Dν Φ − M22 Φ = 0, (6.5.48)
h̄
where the energy-momentum tensor Tµν in (6.5.44) is
1 1
Tµν = − gµν (LAN1 + LAN2 + h̄cLD + h̄cLKG ) + (Dµ Φ)† (Dν Φ) (6.5.49)
2 2
1 αβ a b 1 5 αβ 5k 5l
− Gab g Aµα Aνβ − Gkl g Aµα Aνβ .
4 4
406 Chapter 6 Quantum Physics

The energy-momentum tensor Tµν contains the masses M1 , M2 , the kinetic energy and elec-
tromagnetic energy.
It is clear that both sides of the field equations (6.5.44)-(6.5.48) are all generated by
the fundamental principles. It is the view presented by Einstein and Nambu and shared by
many physicists that the Nature obeys simple beautiful laws based on a few first physical
principles. In other words, the energy-momentum tensor Tµν is now derived from first
principles and is geometrized as Einstein and Nambu hoped.
Systems with four interactions
The above systems with gravity and electromagnetism in general describe the bodies in
lower energy density. For the systems in higher energy density, we have to also consider
the weak and strong interactions. The interactions are layered as shown below, which were
derived in (Ma and Wang, 2015b, 2014g):

The layered systems and sub-systems above determine the action of the system with
four interactions as follows:
!
c4 √
L= R −gdx + actions of all levels, (6.5.50)
8π G
6.5 Field Theory of Multi-Particle Systems 407

and the action of each layered level is as given by the manner as used in (6.5.35)-(6.5.36).
Hence, the unified field model of a multi-particle system is completely determined by
the layered structure of this system, as given by (6.5.50). It is very natural that a rationale
unified field theory must couple the matter fields and interaction fields together.

Remark 6.31 Once again we emphasize that, using PRI contractions as given by
(6.5.10) and proper gauge fixing equations, from the unified field model (6.5.50) coupling
matter fields for multi-particle system, we can easily deduce that the total electromagnetic
field Aµ obtained from (6.5.50) satisfies the U(1) electromagnetic gauge field equations,
and derive the weak and strong interaction potentials as given in (Ma and Wang, 2015a,
2014h).

6.5.5 Atomic spectrum

Classical quantum mechanics is essentially a subject to deal with single particle systems.
Hence, the hydrogen spectrum theory was perfect under the framework of the Dirac equa-
tions. But, for general atoms the spectrum theory was defective due to lack of precise field
models of multi-particle systems.
In this subsection, we shall apply the field model of multi-particle systems to establish
the spectrum equations for general atoms.

1. Classical theory of atomic shell structure. We recall that an atom with atomic number
Z has energy spectrum
Z 2 me4
En = − 2 2 , n = 1, 2, · · · . (6.5.51)
n 2h̄
If we ignore the interactions between electrons, the orbital electrons of this atom have the
idealized discrete energies (6.5.51). The integers n in (6.5.51) are known as principal quan-
tum number, which characterizes the electron energy levels and orbital shell order:

n: 1 2 3 4 5 6 7
(6.5.52)
shell symbol: K L M N O P Q.

Each orbital electron is in some shell of (6.5.52) and possesses the following four quantum
numbers:

1) principle quantum number n = 1, 2, · · · ,

2) orbital quantum number l = 0, 1, 2, · · · , (n − 1),

3) magnetic quantum number m = 0, ±1, · · · , ±l,

1
4) spin quantum number J = ± .
2
408 Chapter 6 Quantum Physics

For each given shell n, there are sub-shells characterized by orbital quantum number l,
whose symbols are:
l: 0 1 2 3 4 ···
(6.5.53)
sub-shell: s p d f g · · · .

By the Pauli exclusion principle, at a give sub-shell nl, there are at most the following
electron numbers
Nnl = Nl = 2(2l + 1) for 0 ! l ! n − 1.

Namely, for the sub-shells s(l = 0), p(l = 1), d(l = 2), f (l = 3), g(l = 4), their maximal
electron numbers are

Ns = 2, N p = 6, Nd = 10, N f = 14, Ng = 18.

Thus, on the n-th shell, the maximal electron number is

n−1
Nn = ∑ Nl = 2n2. (6.5.54)
l=0

2. Atomic field equations. Based on the atomic shell structure, the electron system of an
atom consists of shell systems as (6.5.52), which we denote by

Sn = the n-th shell system for n = 1, 2, · · · . (6.5.55)

Each shell system Sn has n sub-shell systems as in (6.5.53), denoted by

Snl = the l-th sub-shell system of Sn for 0 ! l ! n − 1. (6.5.56)

Thus, we have two kinds of classifications (6.5.55) and (6.5.56) of sub-systems for
atomic orbital electrons, which lead to two different sets of field equations.

A. F IELD EQUATION OF SYSTEM Sn . If we ignore the orbit-orbit interactions, then we

take (6.5.55) as an N-particle system. Let Sn have Kn electrons:

Sn : Ψn = (ψn1 , · · · , ψnKn ), Kn ! Nn , 1 ! n ! N, (6.5.57)

where Nn is as in (6.5.54). Hence, the model of (6.5.57) is reduced to the SU(K1 ) × · · · ×

SU(KN ) gauge fields of fermions. Referring to the single fermion system (6.5.15)-(6.5.21),
the action of (6.5.57) is
! N
L= ∑ (LSU(Kn ) + LDn )dx, (6.5.58)
n=1
6.5 Field Theory of Multi-Particle Systems 409

where
1 µα νβ an an
LSU(Kn ) = − g g Aµν Aαβ 1 ! an ! Kn ,
I 4h̄c J
me c
LDn = Ψn iγ µ Dµ − Ψn 1 ! n ! N,
h̄
e (6.5.59)
Aaµν
n
= ∂µ Aaνn − ∂ν Aaµn − λbanncn Abµn Acνn ,
! h̄c "
ie 0 ie an
Dµ Ψn = ∂µ − Aµ − Aµ τan Ψn ,
h̄c h̄c
where Aaµn are the SU(Kn ) gauge fields representing the electromagnetic (EM) potential of
the electrons in Sn , λbanncn are the structure constants of SU(Kn ) such that Gan bn = 21 tr(τan τb†n ) =
δan bn , A0µ is the EM potential generated by the nuclear, g = −e (e > 0) is the charge of an
electron, and me is the electron mass.
The PID gradient operators for SU(K1 ) × · · · × SU(KN ) in (6.5.19) are given by
G H
1 e
h̄c k̸∑
Dnµ =
(k)
∂µ + Aµ for 1 ! n ! N, (6.5.60)
h̄c =n

(k) K a
where A µ = αaKk Aµk is the total EM potential of Sk shell as defined in (6.5.10).
Then by (6.5.58)-(6.5.60), the field equations of (6.5.57) can be written in the following
form
e
∂ ν Aaνnµ + λbanncn gαβ Abαnµ Acβn + eΨγµ τ an Ψn (6.5.61)
G h̄c H
e
h̄c k̸∑
A µ φ an for 1 ! an ! Kn2 − 1, 1 ! n ! N,
(k)
= ∂µ +
=n
E F
µ ie 0 ie an me c
iγ ∂µ − Aµ − Aµ τan Ψn − Ψn = 0. (6.5.62)
h̄c h̄c h̄

B. F IELD EQUATION OF SYSTEM Snl . The precise model of atomic spectrum should
n−1
take (6.5.56) as an N-particle system. Also, Sn = ∑ Snl is again divided into n sub-systems
l=0

Sn : Sn0 , · · · , Snn−1.

Hence, the system Snl has more sub-systems than Sn , i.e. if Sn has N sub-systems, then Snl
1
has N(N + 1) sub-systems.
2
Let Snl have Knl electrons with wave functions:
K
Snl : Ψnl = (ψnl1 , · · · , ψnlnl ), 1 ! n ! N, 0 ! l ! n − 1, (6.5.63)

and Knl ! 2(2l + 1). Then the action of (6.5.63) takes as

! n−1 N
L= ∑ ∑ (LSU(Knl ) + LDnl )dx, (6.5.64)
l=0 n=1
410 Chapter 6 Quantum Physics

where LSU(Knl ) and LDnl are similar to that of (6.5.59). Thus, the field equation of the system
(6.5.63) is determined by (6.5.64).

Remark 6.32 The reason why atomic spectrum can be divided into two systems
(6.5.57) and (6.5.63) to be considered is that in the system (6.5.63) the electrons in each
Snl have the same energy, and in (6.5.57) the electons in each S n have the same energy if we
ignore the interaction energy between different l-orbital electrons of S nl . Hence, the system
of Snl is precise and the system of Sn is approximative.

3. Atomic spectrum equations. For simplicity, we only consider the system S n , and for
Snl the case is similar. Since the electrons in each Sn have the same energy λn , the wave
functions in (6.5.57) can take as

ψnj = ϕnj (x)e−iλnt/h̄ for 1 ! j ! Kn . (6.5.65)

It is known that the EM fields Aaµ in atomic shells are independent of time t, i.e. ∂t Aaµ = 0.
Therefore, inserting (6.5.65) into (6.5.61) and (6.5.62) we derive the spectrum equation in
the form

λn Φn = ich̄(⃗α · D)Φn − eV Φn + me c2 α0 Φn + eAa0n τan Φn for 1 ! n ! N, (6.5.66)

e e e
h̄c k̸∑
∆Aa0n − λbanncn ⃗Abn · (∇Ac0n + λdcnn fn Adn ⃗A fn ) − eΦ†n τan Φn = A0 φ an ,
(k)
(6.5.67)
h̄c h̄c =n
Q R
e e
∆⃗Aan − ∇( div ⃗Aan ) + λbanncn gαα ⃗Abαn Acαn +eΦ)n⃗γτan Φn = ∇+ ∑ ⃗A(k) φ an , (6.5.68)
h̄c h̄c k̸=n

where Φn = (ϕn1 , · · · , ϕnKn )T , Aaµn = (Aa0n , ⃗Aan ),⃗α = (α1 , α2 , α3 ) and α0 are as in (3.1.15),⃗γ =
(γ 1 , γ 2 , γ 3 ) is as in (6.2.8), V = ze/r is the Coulomb potential of the nuclear, ⃗A = (A1 , A2 , A3 )
is the magnetic potential of the nuclear, and
! "
ie ⃗ ie ⃗ an
DΦn = ∇ − A − A τan Φn ,
h̄c h̄c
⃗Abαn = ∂α ⃗Abn − ∇Abαn − e λ bn Acαn ⃗Adn .
h̄c cn dn
The equations (6.5.66)-(6.5.68) need to be complemented with some gauge fixing equa-
tions; see Remark 6.29.
Chapter 7
Astrophysics and Cosmology

The aim of this chapter is to study fundamental issues of astrophysics and cosmology based
on the first principles dictating the law of gravity, the cosmological principle, and the prin-
ciple of symmetry-breaking. The study and the results presented in this chapter are based
on recent papers (Ma and Wang, 2014e,a; Hernandez, Ma and Wang, 2015).
First, we have rigorously proved a basic Blackhole Theorem on the nature and structure
of black holes, Theorem 7.15:
Assume the validity of the Einstein theory of general relativity, then black holes
are closed, innate and incompressible.
This theorem was originally discovered and proved in (Ma and Wang, 2014a, Theorem
4.1). One important part of the theorem is that all black holes are closed: matters can neither
enter nor leave their interiors. Classical view was that nothing can get out of black holes,
but matters can fall into blackholes. We show that nothing can get inside the blackhole
either. This theorem offers a very different views on the geometric structure and the origin
of our Universe, on the formation and stability of stars and galaxies, and on the mechanism
of supernovae and active galactic nucleus (AGN) jets.
In particular, we rigorously show that our Universe is not originated from a Big-Bang,
and is static, under the assumption of the Einstein general relativity and the cosmological
principle. The redshift is then due mainly to the black hole effect, and the CMB is caused
by the blackbody equilibrium radiation.
Also, we show that the dark energy is associated with the the negative pressure, the
effect of the gravitational repelling force, and the dark matter is caused the space curved
energy.
This chapter is organized as follows. Section 7.1 introduces the momentum represen-
tation of astrophysical fluid dynamical models. One important aspect of the models is the
coupling between the gravitational field equations and the fluid dynamics equations, using
the principle of symmetry-breaking. Section 7.4 proves the black hole theorem.
Stellar circulations are studied in Section 7.2, leading to the mechanism of supernovae
explosion. Section 7.3 addresses galactic circulations, and provides the mechanism of AGN
jets.
412 Chapter 7 Astrophysics and Cosmology

Based on the black hole theorem, theorems on the structure and origin of the Universe
are proved in Section 7.5.
Finally, in Section 7.6, we have derived 1) the PID cosmological model of our Universe,
2) the nature of dark energy and dark matter, and 3) the gravitational force formulas.

7.1 Astrophysical Fluid Dynamics

The main objective of this section is to establish fluid dynamical models for astrophysical
and cosmological objects such as stars, galaxies, and clusters of galaxies, based on the
Principle of Symmetry-Breaking 2.14.

7.1.1 Fluid dynamic equations on Riemannian manifolds

To consider astrophysical fluid dynamics, we first need to discuss the Navier-Stokes equa-
tions on Riemannian manifolds.
Let (M , gi j ) be an n-dimensional Riemannian manifold. The fluid motion on M are
governed by the Navier-Stokes equations given by
∂u 1
+ (u · ∇)u = ν ∆u − ∇p + f for x ∈ M ,
∂t ρ (7.1.1)
divu = 0,
where u = (u1 , · · · , un ) is the velocity field, p is the pressure, f is the external force, ρ is
the mass density, ν is the dynamic viscosity, and the differential operator ∆ is the Laplace-
Beltrami operator defined as ∆u = (∆u1 , · · · , ∆un ) with
∆ui = div(∇ui ) + gi j R jk uk , (7.1.2)
  i   j 
∂ ∂u ∂u j s
div(∇ui ) = gkl + Γ i
kj u j
+ Γ i
lj + Γ ks u (7.1.3)
∂ xl ∂ xk ∂ xk
 i 
j ∂u
−Γkl + Γijs us .
∂xj
Here Ri j is the Ricci curvature tensor and Γik j the Levi-Civita connection:
 2 
1 ∂ gkl ∂ 2 gi j ∂ 2 gk j ∂ 2 gli
Ri j = gkl + − − (7.1.4)
2 ∂ xi ∂ x j ∂ xk ∂ xl ∂ xi ∂ xl ∂ x j ∂ xk
 
+ gkl grs Γrkl Γsij − Γril Γsk j ,
 
i 1 il ∂ gkl ∂ g jl ∂ gk j
Γk j = g + k − . (7.1.5)
2 ∂xj ∂x ∂ xl
The nonlinear convection term (u · ∇)u in (7.1.1) is defined by
(u · ∇)u = (ui Di u1 , · · · , ui Di un ),
∂ uk (7.1.6)
ui Di uk = u i + Γkij ui u j
∂ xi
7.1 Astrophysical Fluid Dynamics 413

the pressure term is  

∂p ∂p
∇p = g1k k , · · · , gnk k , (7.1.7)
∂x ∂x
the divergence of u is
√
∂ uk k j 1 ∂ ( guk )
divu = k + Γ jk u = √ , (7.1.8)
∂x g ∂ xk
and g = det(gi j ).
By (7.1.2) and (7.1.6)-(7.1.8), the Navier-Stokes equations (7.1.1) can be equivalently
written as
∂ ui ∂ ui h i 1 ∂p
+ uk k + Γik j uk u j = ν div(∇ui ) + gi j R jk uk − gi j j + f i ,
∂t ∂x ρ ∂x
(7.1.9)
∂ uk k j
+ Γ jk u = 0.
∂ xk
Remark 7.1 In the Navier-Stokes equations (7.1.1), the Laplace operator ∆ can be
taken in two forms:

∆ = dδ + δ d the Laplace-Beltrami operator, (7.1.10)

∆ = div · ∇ the Laplace operator. (7.1.11)

Here we choose (7.1.10) instead of (7.1.11) to represent the viscous term in (7.1.1). The
reason is that the Laplace-Beltrami operator

(d δ + δ d)ui = div · ∇ui + gi j R jk uk

gives rise to an additional term gi j R jk uk . In fluid dynamics, the term µ div · ∇u represents
the viscous (frictional) force, and the term gi j R jk uk is the force generated by space curvature
and gravitational interaction. Hence physically, it is more natural to take (7.1.10) instead of
(7.1.11).
Remark 7.2 In the fluid dynamic equations (7.1.9), the symmetry of general relativity
breaks, and the space and time are treated independently.

7.1.2 Schwarzschild and Tolman-Oppenheimer-Volkoff (TOV) metrics

We recall in this section the classical Schwarzschild and TOV metrics for centrally symmet-
ric gravitational fields.
Schwarzschild metric
Many stars in the Universe are spherically-shaped, generating centrally symmetric gravi-
tational fields. It is known that the Riemannian metric of a spherically symmetric gravitation
field takes the following form:

ds2 = −eu c2 dt 2 + ev dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ), (7.1.12)

414 Chapter 7 Astrophysics and Cosmology

where (r, θ , ϕ ) is the spherical coordinate system, and u = u(r,t) and v = v(r,t) are functions
of r and t, which are determined by the gravitational field equations.
In the exterior of a ball, Schwarzschild first obtained an exact solution of the Einstein
field equations in 1916, which describes the gravitational fields for the external vacuum
state of a static spherically symmetric matter field.
Let m be the total mass of a centrally symmetric ball. Then the classical Newtonian
gravitational potential of the ball reads
mG
ϕ =− . (7.1.13)
r
Based on the Einstein general theory of relativity, the time-component g 00 of gravitational
potential gµν and the Newton potential ϕ have the following relation
 
2
g00 = − 1 + 2 ϕ (7.1.14)
c
Hence, by (7.1.13) and (7.1.14) for the ball we have
2mG
g00 = −1 + . (7.1.15)
c2 r
Now we consider the gravitational field equations in the exterior of the ball. In the
vacuum state,
Tµν = 0. (7.1.16)
On the other hand, by R = g µν Rµν , the Einstein gravitational field equations
1 8π G
Rµν − gµν R = − 4 Tµν
2 c
can be equivalently written as
8π G 1
Rµν = − (Tµν − gµν T ), T = gkl Tkl .
c4 2
Thus, by (7.1.16) the Einstein field equations become
Rµν = 0 (7.1.17)
The nonzero components of the metric (7.1.12) g µν are
g00 = −eu , g11 = ev , g22 = r2 , g33 = r2 sin2 θ . (7.1.18)
Since the gravitational resource is static, u and v only depend on r. By (7.1.18), all
nonzero components of the Levi-Civita connection (7.1.5) are given by
1 1
Γ100 = eu−v u0 , Γ111 = v0 , Γ122 = −re−v ,
2 2
1 1
Γ133 = −re−v sin2 θ , Γ010 = u0 , Γ212 = , (7.1.19)
2 r
1 cos θ
Γ233 = − sin θ cos θ , Γ313 = , Γ323 = .
r sin θ
7.1 Astrophysical Fluid Dynamics 415

Thus, by (7.1.4) and (7.1.19) we get R µν as

 
1 1 1
R00 = eu−v − u00 − u0 + u0 (v0 − u0) ,
2 r 4
1 1 1
R11 = u00 − v0 − u0 (v0 − u0 ),
2 r 4
h r i (7.1.20)
R22 = e−v 1 − ev + (u0 − v0 ) ,
2
R33 = R22 sin2 θ ,

Rµν = 0, ∀µ 6= ν .

Therefore, the vacuum Einstein field equations (7.1.17) become the following system of
ordinary differential equations
1 00 1 0 1 0
u + u − (v − u0)u0 = 0, (7.1.21)
2 r 4
1 00 1 0 1 0
u − v − (v − u0 )u0 = 0, (7.1.22)
2 r 4
r 0
(u − v0 ) − ev + 1 = 0. (7.1.23)
2
By the Bianchi identity, only two equations of (7.1.22)-(7.1.23) are independent. The
difference of (7.1.21) and (7.1.22) leads to

u0 + v0 = 0,

which implies that

u+v = β (constant), (7.1.24)

and (7.1.23) becomes

ev + rv0 − 1 = 0.

Namely
d
(re−v ) = 1.
dr
It follows that
b
e−v = 1 − ,
r
where b is a to-be-determined constant.
Then it follows from (7.1.24) that
 
b
eu = eβ e−v = eβ 1 − .
r
416 Chapter 7 Astrophysics and Cosmology

By scaling time t, we can take eβ = 1. Hence the solution of the Einstein field equations
(7.1.17) is given by  
u b
g00 = −e = − 1 − ,
r
 −1 (7.1.25)
b
g11 = ev = 1 − .
r
Comparing (7.1.25) with (7.1.15), we can obtain
2mG
b= .
c2
Thus, we get the solution of (7.1.17) as
 
2mG
g00 = − 1 − 2 ,
c r
 
2mG −1
g11 = 1 − 2 ,
c r

and the metric reads

   
2 2mG 2 2 2mG −1 2
ds = − 1 − 2 c dt + 1 − 2 dr + r2 d θ 2 + r2 sin2 θ d ϕ 2 , (7.1.26)
c r c r

We have in particular
2mG
e−v = eu = 1 − .
c2 r
which is called the Schwarzschild solution or metric.
TOV metric
The Schwarzschild metric (7.1.26) describes the exterior gravitational fields of a cen-
trally symmetric ball. For the interior gravitational fields, the metric is given by the TOV
solution.
Let m be the mass of a centrally symmetric ball, and R be the radius of this ball. In the
interior of the ball, the variable r satisfies 0 6 r < R. Let the ball be a static liquid sphere
consisting of idealized fluid, an approximation of stars. The energy-momentum tensor of an
idealized fluid is in the form

T µν = (c2 ρ + p)uµ uν + pgµν ,

where p is the pressure, ρ is the density, and u µ is the 4-velocity. For a static fluid, u µ is
given by
1
uµ = √ (1, 0, 0, 0).
−g00
7.1 Astrophysical Fluid Dynamics 417

Hence, the (1,1)-type of the energy-momentum tensor is in the form

 
−ρ 2 0 0 0
 0 p 0 0 
Tµν = 
 0
.
0 p 0 
0 0 0 p
The Einstein gravitational field equations of an interior ball read
1 8π G
Rνµ − δµν R = − 4 Tµν ,
2 c (7.1.27)
Dν Tµν = 0.
By (7.1.20) we have
2
R = gµν Rµν = −e−u R00 + e−ν R11 + R22 .
r2
Then, by (7.1.20) and Rνµ = gνα Rα µ we get
 
1 1 v0 1
R00 − R = −e−v 2 − + 2,
2 r r r
 0 
1 u 1 1
R11 − R = −e−v + 2 + 2,
2 r r r
 
1 1 1 1 1 1
R22 − R = − e−v u00 − u0 v0 + u02 + u0 − v0
2 2 2 2 r r
In addition, we know that

Dα Tµα = D0 Tµ0 + D1 Tµ1 + D2 Tµ2 + D3 Tµ3 ,

β
β ∂ Tµ β β
Dα Tµ = + Γαν Tµν − Γνα µ Tν .
∂ xα
Thus, by (7.1.19) we have
dp 1
Dα T1α = + (p + c2 ρ )u0 ,
dr 2
Dα Tµα = 0 for µ 6= 1.

Hence, the field equations (7.1.27) can be written as

 
1 v0 1 8π G
e −v
− − 2 = − 2 ρ, (7.1.28)
r2 r r c
 
1 u 0 1 8π G
e−v 2 + − 2 = 4 p, (7.1.29)
r r r c
1 1 1 1 16π G
e−v (u00 − u0 v0 + u02 + u0 − v0 ) = p, (7.1.30)
2 2 r r c4
1
p0 + (p + c2ρ )u0 = 0. (7.1.31)
2
418 Chapter 7 Astrophysics and Cosmology

By the Bianchi identity, only three equations of (7.1.28)-(7.1.31) are independent. Here
we also regard p and ρ as unknown functions. Therefore, for the four unknown functions
u, v, p, ρ , we have to add an equation of state to the system of (7.1.28)-(7.1.31):

ρ = f (p), (7.1.32)

and the function f will be given according to physical conditions.

On the surface r = R of the ball, p = 0 and u and v are given in terms of the Schwarzschild
solution:  
2Gm
p(R) = 0, u(R) = −v(R) = ln 1 − . (7.1.33)
Rc2
We are now in position to discuss the solutions of problem (7.1.28)-(7.1.33). Let

c2 r
M(r) = (1 − e−v). (7.1.34)
2G
Then the equation (7.1.28) can be rewritten as
1 dM
= 4πρ ,
r2 dr
whose solution is given by
Z r
M(r) = 4π r2 ρ dr for 0 < r < R. (7.1.35)
0

By (7.1.35), we see that M(r) is the mass, contained in the ball B r . It follows from (7.1.34)
that
2GM(r)
e−v = 1 − . (7.1.36)
c2 r
Inserting (7.1.36) in (7.1.29) we obtain
 
1 8π G 3
u0 = pr + 2GM(r) . (7.1.37)
r(c2 r − 2MG) c2
Putting (7.1.37) into (7.1.31) we get
 
0 p + c2 ρ 8π G 3
p =− pr + 2GM(r) . (7.1.38)
2r(c2 r − 2MG) c2
Thus, it suffices for us to derive the solution p, M and ρ from (7.1.32)-(7.1.34) and
(7.1.38), and then v and u will follow from (7.1.36)-(7.1.37) and (7.1.33).
The equation (7.1.38) is called the TOV equation, which was derived to describe the
structure of neutron stars.
We note that (7.1.36) is the interior metric of a blackhole provided that 2GM(r)/
2
(c r) = 1. Thus the TOV solution (7.1.36) gives a rigorous proof of the following theo-
rem for the existence of black holes.
7.1 Astrophysical Fluid Dynamics 419

Theorem 7.3 If the matter field in a ball BR of radius R is spherically symmetric, and
the mass MR and the radius R satisfy
2GMR
= 1,
c2 R
then the ball must be a blackhole.

An idealized model is that the density is homogeneous, i.e. (7.1.32) is given by

ρ = ρ0 a constant.

In this case, we have

4π
M(r) = ρ0 r 3 for 0 6 r 6 R,
3
3 m
ρ0 = .
4π R3
Thus we obtain the following solution of (7.1.36)-(7.1.38) with (7.1.33):
  1/2   
2Gmr2 2Gm 1/2
 1− 2 3 − 1− 2 
 c R c R 
p(r) = ρ0   1/2  1/2  , (7.1.39)
 2Gm 2Gmr 2 
3 1− 2 − 1− 2 3
c R c R
"  1/2  1/2 #2
u 3 2Gm 1 2Gmr2
e = 1− 2 − 1− 2 3 , (7.1.40)
2 c R 2 c R
 −1
2Gmr2
ev = 1 − 2 3 . (7.1.41)
c R

The functions (7.1.39)-(7.1.41) are the TOV solution. By (7.1.12), the solution (7.1.40) and
(7.1.41) yields the metric
"    1/2 #2
2 3 2Gm 1/2 1 2Gmr2
ds = − 1− 2 − 1− 2 3 c2 dt 2 (7.1.42)
2 c R 2 c R
 −1
2Gmr2
+ 1− 2 3 dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ),
c R

which is called the TOV metric.

7.1.3 Differential operators in spherical coordinates

In Subsection 7.1.1, we gave the Navier-Stokes equations on general Riemannian manifolds.
For astrophysical fluid dynamics, we mainly concern the equations on 3D spheres. Hence in
420 Chapter 7 Astrophysics and Cosmology

this subsection we discuss the basic differential operators (7.1.2)-(7.1.8) under the spherical
coordinate systems (θ , ϕ , r).
For a 3D sphere M, the Riemannian metric is given by

ds2 = α (r)dr2 + r2 d θ 2 + r2 sin2 θ d ϕ 2 (7.1.43)

where α (r) > 0 represents the relativistic effects:



 1 no relativistic effect,

  −1

 2Gm

1− 2 for the Schwarzschild metric (7.1.26),
α= c r (7.1.44)

  −1



 2Gmr2
 1− 2 3 for the TOV metric (7.1.42).
c R

In (7.1.43) we have

g11 = α (r), g22 = r2 , g33 = r2 sin2 θ , gi j = 0 for i 6= j.

By (7.1.5) we can get the Levi-Civita connection as

1 1
Γ221 = Γ212 = , Γ233 = − sin θ cos θ , Γ331 = Γ313 = ,
r r
3 3 cos θ 1 r r
Γ32 = Γ32 = , Γ22 = − , Γ33 = − sin2 θ ,
1 (7.1.45)
sin θ α α
1 1 dα k
Γ11 = , Γi j = 0 for others.
2α dr
We deduce from (7.1.45) the explicit form of the Ricci curvature tensor (7.1.4):

1 dα 1 r dα
R11 = − , R22 = − 2 − 1,
α r dr α 2α dr (7.1.46)
R33 = R22 sin2 θ , Ri j = 0 6 j.
∀i =

Based on (7.1.45) and (7.1.46) we can obtain the expressions of the differential operators
(7.1.2)-(7.1.8) as follows:

1) The Laplace-Beltrami operator ∆uk = (∆ur , ∆uθ , ∆uϕ ):

  
1 1 ∂ ∂ uθ 1 ∂ 2 uθ
∆uθ = 2 sin θ + 2 (7.1.47)
r sin θ ∂ θ ∂θ sin θ ∂ ϕ 2
  
1 ∂ 2 ∂ uθ 2 ∂ ur 2 cos θ ∂ uϕ 1
+ r + − − 2 uθ
α ∂r ∂r r ∂θ sin θ ∂ ϕ sin θ
 
1 ∂ α0 ∂ 2
+ 2 2 (ruθ ) − (r uθ ) ,
αr ∂r 2α ∂ r
7.1 Astrophysical Fluid Dynamics 421

    
1 1 ∂ ∂ uϕ 1 ∂ 2 uϕ 1 ∂ 2 ∂ uϕ
∆uϕ = sin θ + + r (7.1.48)
r2 sin θ ∂ θ ∂θ sin2 θ ∂ ϕ 2 α ∂r ∂r

2 cos θ ∂ uϕ 2 cos θ ∂ uθ 2 ∂ ur
+ − 2uϕ + +
sin θ ∂ θ sin3 θ ∂ ϕ r sin2 θ ∂ ϕ
 
1 ∂ α0 ∂ 2
+ 2 2 (ruϕ ) − (r uϕ ) ,
αr ∂r 2α ∂ r
    
1 1 ∂ ∂ ur 1 ∂ 2 ur 1 ∂ 2 ∂ ur
∆ur = 2 sin θ + 2 + r (7.1.49)
r sin θ ∂ θ ∂θ sin θ ∂ ϕ 2 α ∂ r ∂r
  
2 cos θ ∂ uθ ∂ uϕ r ∂ 2 α 0
− 2 ur + r uθ + r +r − ur .
αr sin θ ∂θ ∂ϕ 2 ∂ r 2α

2) By (7.1.6) and (7.1.45), (u · ∇)uk can be written as

∂ uθ ∂ uθ ∂ uθ 2
uk Dk uθ = u r + uθ + uϕ + uθ ur − sin θ cos θ uϕ2 (7.1.50)
∂r ∂θ ∂ϕ r
∂ uϕ ∂ uϕ ∂ uϕ 2 cos θ 2
uk Dk uϕ = u r + uθ + uϕ + uθ uϕ + uϕ ur , (7.1.51)
∂r ∂θ ∂ϕ sin θ r
∂ ur ∂ ur ∂ ur r 2 α0
uk Dk ur = u r + uθ + uϕ − (uθ + sin2 θ uϕ2 − u2r ). (7.1.52)
∂r ∂θ ∂ϕ α 2r

3) The gradient operator:

 
1 ∂p 1 ∂p 1 ∂p
∇p = , , . (7.1.53)
α ∂ r r2 ∂ θ r2 sin2 θ ∂ ϕ

√ √
4) By (7.1.8) and g = r2 sin θ α , the divergent operator divu is

1 ∂ ∂ uϕ 1 ∂ 2√
divu = (sin θ uθ ) + + 2√ (r α ur ). (7.1.54)
sin θ ∂ θ ∂ϕ r α ∂r

Remark 7.4 The expressions (7.1.47)-(7.1.54) are the differential operators appearing
in the fluid dynamic equations describing the stellar fluids. However, we need to note that
the two components uθ and uϕ are the angular velocities of θ and ϕ , i.e.

dθ dϕ
uθ = , uϕ = .
dt dt
In classical fluid dynamics, the velocity field v = (vθ , vϕ , vr ) is the line velocity. The relation
of u and v is given by

1 1
u θ = vθ , uϕ = vϕ , ur = v r . (7.1.55)
r r sin θ
422 Chapter 7 Astrophysics and Cosmology

Hence, inserting (7.1.55) into (7.1.1) with the expressions (7.1.47)-(7.1.54), we derive
the Navier-Stokes equations in the usual spherical coordinate form as follows
∂ vθ 1 ∂p
+ (u · ∇)vθ = ν ∆vθ − + fθ ,
∂t ρr ∂ θ
∂ vϕ 1 ∂p
+ (u · ∇)vϕ = ν ∆vϕ − + fϕ ,
∂t ρ r sin θ ∂ ϕ (7.1.56)
∂ vr 1 ∂p
+ (u · ∇)vr = ν ∆vr − + fr ,
∂t ρα ∂ r
div v = 0,

where
2 ∂ vr 2 cos θ ∂ vϕ vθ 1 dα ∂
∆vθ = ∆v
e θ+ − − 2 2 − 2 (rvθ ),
r2 ∂ θ r2 sin2 θ ∂ ϕ r sin θ 2 α r dr ∂ r
e ϕ + 2 ∂ vr + 2 cos θ ∂ vθ − vϕ
∆vϕ = ∆v −
1 dα ∂
(rvϕ ),
r2 sin θ ∂ ϕ r2 sin2 θ ∂ ϕ r2 sin2 θ 2α 2 r dr ∂ r
   
e r − 2 vr + ∂ vθ + cos θ vθ + 1 ∂ vϕ + 1 ∂
∆vr = ∆v
1 dα
v r ,
α r2 ∂θ sin θ sin θ ∂ ϕ 2α ∂ r α dr
(7.1.57)
∆
e is the Laplace operator for scalar fields given by
   
1 ∂ ∂T 1 ∂ 2T 1 ∂ 2 ∂T α0 ∂ T
∆T = 2
e sin θ + 2 2 + r − , (7.1.58)
r sin θ ∂ θ ∂θ r sin θ ∂ ϕ 2 α r2 ∂ r ∂r 2rα 2 ∂ r

the nonlinear term (u · ∇)v is

vθ ∂ v θ vϕ ∂ v θ ∂ vθ vθ vr cos θ vϕ2
(v · ∇)vθ = + + vr + − ,
r ∂θ r sin θ ∂ ϕ ∂r r r sin θ
vθ ∂ v ϕ vϕ ∂ v ϕ ∂ vϕ vϕ vr cos θ vϕ vθ
(v · ∇)vϕ = + + vr + + , (7.1.59)
r ∂θ r sin θ ∂ ϕ ∂r r r sin θ
vθ ∂ v r vϕ ∂ v r ∂ vr 1 1 dα 2
(v · ∇)vr = + + vr − (v2 + v2ϕ ) + v ,
r ∂ θ r sin θ ∂ ϕ ∂ r αr θ 2α dr r
and the divergent term div v reads
√
1 ∂ (sin θ vθ ) 1 ∂ vϕ 1 ∂ (r2 α vr )
div v = + + 2√ . (7.1.60)
r sin θ ∂θ r sin θ ∂ ϕ r α ∂r
7.1.4 Momentum representation
The Universe, galaxies and galactic clusters are composed of stars and interstellar nebulae.
Their velocity fields are not continuous. Hence it is not appropriate that we model cosmic
objects using continuous velocity field v(x,t) as in the Navier-Stoks equations or by discrete
position variables xk (t) as in the N-body problem.
7.1 Astrophysical Fluid Dynamics 423

The idea is that we use the momentum density field P(x,t) to replace the velocity field
v(x,t) as the state function of cosmic objects. The main reason is that the momentum density
field P is the energy flux containing the mass, the heat, and all interaction energy flux, and
can be regarded as a continuous field. The aim of this section is to establish the momentum
form of astrophysical fluid dynamics model.
The physical laws governing the dynamics of cosmic objects are as follows

Theory of General Relativity,

Newtonian Second Law,
Heat Conduction Law, (7.1.61)
Energy-Momentum Conservation,
Equation of State.

The mathematical expressions of these laws are given respectively in the following:
1. Gravitational field equations.
1 8π G 1 e
Rµν − gµν R = − 4 Tµν + (D µ Φν + D
e ν Φµ ),
2 c 2 (7.1.62)
e µ = Dµ + e Aµ ,
D
h̄c
where Aµ is the electromagnetic potential, the time components of g µν are as
 
2
g00 = − 1 + 2 ψ , g0k = gk0 = 0 for 1 6 k 6 3,
c
and ψ is the gravitational potential.
2. Fluid dynamic equations. The Newton’s Second Law can be expressed as
dP
= Force, (7.1.63)
dτ
where τ is the proper time given by
√
dτ = −g00 dt, (7.1.64)

P is the momentum density field, formally defined by

dx 1
= P,
dτ ρ
with ρ being the energy density,

dP ∂ P ∂ P dxk ∂P 1
= + k = + (P · ∇)P,
dτ ∂ τ ∂ x dτ ∂τ ρ
424 Chapter 7 Astrophysics and Cosmology

and
ν ∆P + µ ∇(divP) the frictional force,
− ∇p the pressure gradient,
c2
ρ (1 − β T)∇g00 = −ρ (1 − β T)∇ψ the gravitational force.
2
Hence, the momentum form of the fluid dynamic equations (7.1.63) is written as
∂P 1
+ (P · ∇)P = ν ∆P + µ ∇(divP) − ∇p − ρ (1 − β T)∇ψ , (7.1.65)
∂τ ρ
where the differential operators ∆, ∇ and (P · ∇) are with respect to the space metric g i j (1 6
i, j 6 3) determined by (7.1.62), as defined in (7.1.2)-(7.1.8).
3. Heat conduction equation:
∂T 1
+ (P · ∇)T = κ ∆T
e + Q, (7.1.66)
∂τ ρ
where ∆
e is defined as  
e = − √1 ∂
∆T
√ ij ∂T
gg ,
g ∂ xi ∂xj
and g = det(gi j ), 1 6 i, j 6 3.
4. Energy-momentum conservation:
∂ρ
+ divP = 0, (7.1.67)
∂τ
where ρ is the energy density:
ρ = mass + electromagnetism + potential + heat.

5. Equation of state:
p = f (ρ , T ). (7.1.68)
Remark 7.5 Both physical laws (7.1.63) and (7.1.67) are the more general form than
the classical ones:
dv
m = Force the Newton’s Second Law,
dτ (7.1.69)
∂m
+ div(mv) = 0 the continuity equation,
∂τ
where m is the mass density. Hence the momentum representation equations (7.1.65)-
(7.1.67) can be applicable in general. The momentum P represents the energy density flux,
consisting essentially of
P = mv + radiation flux+heat flux.
Hence in astrophysics, the momentum density P is a better candidate than the velocity field
v, to serve as the continuous-media type of state function.
7.1 Astrophysical Fluid Dynamics 425

7.1.5 Astrophysical Fluid Dynamics Equations

Dynamic equations of stellar atmosphere

Different from planets, stars are fluid spheres. Like the Sun, most of stars possess at-
mospheric layers. The atmospheric dynamics of stars is an important topic, and we are now
ready to present the stellar atmospheric model.
The spatial domain is a spherical shell:

M = {x ∈ R3 | r0 < r < r1 }.

The stellar atmosphere consists of rarefied gas. For example, the solar corona has mass
density about ρm = 10−9 ρ0 where ρ0 is the density of the earth atmosphere. Hence we use
the Schwarzschild solution in (7.1.44) as the metric:
 
2mG −1 2mG
α (r) = 1 − 2 , r0 > . (7.1.70)
c r c2

where m is the total mass of the star, and the condition r0 > 2mG/c2 ensures that the star is
not a black hole.
The stellar atmospheric model is the momentum form of the astrophysical fluid dynam-
ical equations defined on the spherical shell M :

∂P 1 mGρ
+ (P · ∇)P = ν ∆P + µ ∇(divP) − ∇p − 2 (1 − β T )~k,
∂τ ρ r
∂T 1
+ (P · ∇)T = κ ∆T,
e (7.1.71)
∂τ ρ
∂ρ
+ divP = 0,
∂τ
where P = (Pr , Pθ , Pϕ ) is the momentum density field, T is the temperature, p is the pressure,
ρ is the energy density, ν and µ is the viscosity coefficient, β is the coefficient of thermal
expansion, κ is the thermal diffusivity, α is as in (7.1.70), ∆P, (P · ∇)P, ∆T,
e div P are as in
(7.1.57)-(7.1.60), and

Pθ ∂ T Pϕ ∂ T ∂T
(P · ∇)T = + + Pr . (7.1.72)
r ∂ θ r sin θ ∂ ϕ ∂r
The equations (7.1.71) are supplemented with the boundary conditions:

∂ Pθ ∂ Pϕ
Pr = 0, = 0, =0 at r = r0 , r1 ,
∂r ∂r
T = T0 at r = r0 , (7.1.73)

T = T1 at r = r1 ,
426 Chapter 7 Astrophysics and Cosmology

where T0 and T1 are approximatively taken as constants and satisfy the physical condition

T0 > T1 .

A few remarks are now in order:

Remark 7.6 First, there are three important parameters: the Rayleigh number Re, the
Prandtl number Pr and the δ -factor δ , which play an important role in astrophysical fluid
dynamics:
mGr0 r1 β T0 − T1 ν 2mG
Re = , Pr = , δ= 2 . (7.1.74)
κν h κ c r0
The δ -factor δ reflects the relativistic effect contained the Laplacian operator.

Remark 7.7 Astronomic observations show that the Sun has three layers of atmo-
spheres: the photosphere, the chromosphere, and the solar corona, where the solar atmo-
spheric convections occur. It manifests that the thermal convection is a universal phe-
nomenon for stellar atmospheres. In the classical fluid dynamics, the Rayleigh number
dictates the Rayleign-Bénard convection. Here, however, both the Rayleigh number Re
and the δ -factor defined by (7.1.74) play an important role in stellar atmospheric convec-
tions.

Remark 7.8 For rotating stars with angular velocity ~Ω, we need add to the right hand
side of (7.1.71) the Coriolis term:

−2~Ω × P = 2Ω(sin θ Pr − cos θ Pθ , cos θ Pϕ , − sin θ Pϕ ),

where Ω is the magnitude of ~Ω.

Fluid dynamical equations inside open balls

As the fluid density in a stellar atmosphere is small, the equations (7.1.71) can be re-
garded as a precise model governing the stellar atmospheric motion. However, for a fluid
sphere with high density, the fluid dynamic equations have to couple the gravitational field
equations.
The Universe and all stars are in the momentum-flow state, i.e. they are fluid spheres.
To investigate the interiors of the Universe, galaxies and stars, we need to develop dynamic
models for fluid spheres.
The precise equations of fluid sphere should be defined in the Riemannian metric space
as follows:
ds2 = g00 (x,t)c2 dt 2 + gi j (x,t)dxi dx j for x ∈ M 3 , (7.1.75)
where M 3 is the spherical space. The gravitational field equations are expressed as
1 8π G
Rµν − gµν R = − 4 Tµν − Dµ Φµ , (g j0 = g0 j = 0, 1 6 j 6 3), (7.1.76)
2 c
7.1 Astrophysical Fluid Dynamics 427

where h i
Tµν = gµα gνβ ε α ε β + pgαβ ,
and ε µ is the 4D energy-momentum vector.
For the fluid component of the system, it is necessary to simplify the model by making
some physically sound approximations.
Approximation Hypothesis 7.9 The metric (7.1.75) and the stationary solutions of
the fluid dynamical equations are radially symmetric.

Under Hypothesis 7.9, the metric (7.1.75) is as in (7.1.12), or is written in the following
form
ds2 = −ψ (r)c2 dt 2 + α (r)dr2 + r2 d θ 2 + r2 sin2 θ d ϕ 2 , (7.1.77)
and the fluid dynamic equations are rewritten as
∂P 1 c2 ρ d ψ
+ (P · ∇)P = ν ∆P + µ ∇(divP) − ∇p − (1 − β (T − T0 ))~k,
∂τ ρ 2α dr
∂T 1 (7.1.78)
+ (P · ∇)T = κ ∆T
e + Q(r),
∂τ ρ
∂ρ
+ divP = 0,
∂τ
where P = (Pr , Pθ , Pϕ ), ∇P, ∆T,
e (P·∇)P, div P are as in (7.1.57)-(7.1.60), ∇p is as in (7.1.56),
~k = (1, 0, 0), and
∂ T Pθ ∂ T Pϕ ∂ T
(P · ∇)T = Pr + + .
∂r r ∂ θ r sin θ ∂ ϕ
The gravitational field equation (7.1.76) for the metric (7.1.77) is radially symmetric,
therefore
Φν = Dν φ , φ = φ (r).
Thus we have
1 1 00 1
D0 D0 φ = ψ 0φ 0, D1 D1 φ = φ − 2 α 0φ 0,
2αψ α 2α
1 0
D2 D2 φ = D 3 D3 φ = φ, D µ Dν φ = 0 for µ 6= ν .
rα
Then, in view of (7.1.28)-(7.1.31), the equation (7.1.76) can be expressed by
 
1 1 α0 1 8π G 1
− − 2 = − 2 ρ0 + ψ 0φ 0,
α r2 rα r c 2αψ
 
1 1 ψ0 1 8π G 1 1
+ − 2 = 2 p + φ 00 − 2 α 0 φ 0 , (7.1.79)
α r2 rψ r c α 2α
"    #
1 ψ 00 1 ψ 0 2 α 0 ψ 0 1 ψ 0 α 0 16π G 2
− − + − = p + φ 0,
α ψ 2 ψ 2αψ r ψ α c2 rα
428 Chapter 7 Astrophysics and Cosmology

where the pressure p satisfies the stationary equations of (7.1.78) with P = 0 as follows
c2 0
p0 = − ψ ρ [1 − β (T − T0 )] ,
2
  (7.1.80)
κ d 2 dT
r = −Q(r).
α r2 dr dr
The functions ψ and α satisfy the boundary conditions (7.1.33), i.e.
 
2Gm 2Gm −1
ψ (r0 ) = 1 − 2 , α (r0 ) = 1 − 2 . (7.1.81)
c r0 c r0
In addition, for the ordinary differential equations (7.1.79)-(7.1.81), we also need the bound-
1
ary conditions for ψ 0 , φ 0 and T . Since − c2 ψ 0 represents the gravitational force, the condi-
2
tion of ψ 0 at r = r0 is given by
2mG
ψ 0 (r0 ) = 2 2 . (7.1.82)
c r0
Based on the Newton gravitational law, φ 0 is very small in the external sphere; also see (Ma
and Wang, 2014e). Hence we can approximatively take that

φ 0 (r0 ) = 0. (7.1.83)

Finally, it is rational to take the temperature gradient in the boundary condition as follows
∂T
(r0 ) = −A (A > 0). (7.1.84)
∂r
Let the stationary solution of the problem (7.1.79)-(7.1.81) be given by pe, Te, ψ , α , φ 0 .
Make the translation transformation

P → P, p → p + pe, T → T + Te.

Then equations (7.1.78) are rewritten in the form

∂P 1 c2 ρ d ψ ~
+ (P · ∇)P = ν ∆P − ∇p + β kT,
∂τ ρ 2α dr
∂T 1 e
e − 1 d T Pr , (7.1.85)
+ (P · ∇)T = κ ∆T
∂τ ρ ρ dr
div P = 0,
supplemented with the boundary conditions:
∂T ∂ Pθ ∂ Pϕ
= 0, Pr = 0, = =0 at r = r0 . (7.1.86)
∂r ∂r ∂r
The model (7.1.85)-(7.1.86), we just derived describes interior dynamics of the Universe,
galaxies and stars.
7.2 Stars 429

7.2 Stars
7.2.1 Basic knowledge
The large scale structure of the Universe consists of mainly the following levers:

stars, stellar clusters, galaxies, clusters of galaxies.

Star is the most elementary constituent of the Universe, and we now explore some of their
basic properties.

1. Mass m. The mass of the Sun is m = 2 × 1030kg, and the range of the masses of the
main-sequence stars is about
0.1m ∼ 40m .

A few extreme stars have masses m ' 60m , and the least massive stars have masses around
m ' 0.07m .

2. Radius R. The radius of the Sun is R = 7 × 105 km, and the radii of the main-
sequence stars are
0.3R ∼ 25R .

A neutron star has radius R ' 10km, and a red giant star has R = 10 3R .

3. Surface temperature T . The surface temperature of the Sun is T = 5800◦K, and the
range of surface temperatures of stars in general is

2600◦K ∼ 35000◦K.

4. Luminosity L. The Sun’s luminosity is L = 1(4 × 1033erg/s), and the luminosities

of the main-sequence stars have ranges in

8 × 10−3L ∼ 3.2 × 105L .

5. Parameter relation. Based on the radiation theory of black bodies, the above three
parameters R, T, L enjoy
L = 4πσ R2 T 4 , (7.2.1)

where σ is the Stefan-Boltzmann constant:

4
σ = 5.7 × 10−5erg/cm2 · s · k .

For the main-sequence stars, we have the following the relation

L/L = (m/m )α , (7.2.2)

430 Chapter 7 Astrophysics and Cosmology

where α takes different values as


 1.8 for m < 0.3m ,

α= 4 for 0.3m < m < 3m ,


2.8 for 3m < m.

6. Main-sequence stars. In (7.2.1) we see that the luminosity L depends on T and R.

However, the radius R is also related with the temperature T . In 1910, two astronomers E.
Hertzsprung and H. R. Russell independently obtained the statistical law between L and T .
They discovered that for most stars, called the main-sequence stars, their luminosity and
temperature possess certain relation. This relation is illustrated by a diagram, called the HR
diagram, which is shown in Figure 7.1.

Figure 7.1 The HR diagram in which the luminosity L and the surface temperature T of stars are
plotted.

Each star is plotted as a point in the HR diagram in Figure 7.1. The stars in the band
are the main-sequence stars. The red giants are in the right-top region, and have lower
temperatures and higher brightness, and the white dwarfs are in the left-bottom region, and
have higher temperatures and low brightness.
The main-sequence stars are young and middle aged stars, the red giants are old aged,
and the white dwarfs are dead stars.
7. Variable stars. Most stars shine with almost constant brightness, but a small number
of stars, called variable stars, change periodically their brightness. Since the luminosity
of a star depends on its radius (see (7.2.1)), variable stars are rhythmically expanding and
contracting, pulsating in size and brightness.
7.2 Stars 431

8. Red giants. Stars are held together by gravity. The gravitational force pulling inward
is opposed by a force pushing outward, consisting mainly the thermal pressure in the inte-
rior. Stars maintain their balance of pressure and gravity by the heat energy produced by
burning hydrogen.
Stars less massive than the Sun evolve more slowly and stay on the main-sequence
longer than 1010 years. But more massive stars evolve more quickly, and end their life
becoming white dwarfs and neutron stars (or possibly black holes) as their final fate deter-
mined by their masses.
After a more massive star consumes its central supply of hydrogen, it leaves the main
sequence and enters into the group of red giants in the HR diagram.
When the central nuclear reaction has ceased to generate heat, the interior pressure re-
duces and the star begins to contract, leading to the release of gravitational energy. Hence,
this contraction causes the temperature to rise. Thus, the hydrogen outside the core begins
to burn. The burning shell causes the star to expend. Thus, the star is luminous, large and
cool, and becomes a red giant.
9. White dwarfs. For a star with mass in the range

1m 6 m < 5m , (7.2.3)

when it is in the red-giant phase, its expanding shell is thrown off into space, and the naked
core is all that remains. Contraction ceases, nuclear burning ends, and the core cools down
as a white dwarf. A white dwarf has a size approximatively equal to that of the earth, with
mass m 6 1.44M .
10. Neutron stars and pulsars. A star with mass

5m < m, (7.2.4)

will eventually evolve to a neutron star. A neutron star has a radius of about 10km with mass
m > 3m , and has a high density of nearly 109 ton/cm3 . The neutron star has a magnetic
field of 1012 gauss that is 1012 times stronger than the earth’s magnetic field.
The pulsar is a special neutron star, which emits a pulse-like energy message.
11. Supernovae. When a large massive red giant (m > 5m ) exhausts its nuclear fuels,
it begins to collapse. This contraction will lead to explosion, and the explosive star is called
a supernova. A neutron star is born after a huge explosion of a supernova.

7.2.2 Main driving force for stellar dynamics

Stars can be regarded as fluid balls. To investigate the stellar interior dynamic behavior,
we need to use the fluid spherical models coupling the heat conductivity equation. There
are two types of starts: stable and unstable. The sizes of stable stars do not change. The
432 Chapter 7 Astrophysics and Cosmology

main-sequence stars, white dwarfs and neutron stars are stable stars. The radii of unstable
stars may change; variable stars and expanding red giants are unstable stars. The dynamic
equations governing the two types of stars are different, and will be addressed hereafter
separately.
We note that the fluid dynamic equations (7.1.78) represent the Newton’s second law,
and their left-hand sides are the acceleration and their right-hand sides are the total force.
The total force consists of four parts: the viscous friction, the pressure gradient, the rela-
tivistic effect, and the thermal expansion force, which are given as follows:

• The viscous friction force is caused by the electromagnetic, the weak and the strong
interactions between the particles and the pressure, and is given by

Fν P = ν 4P = ν (∆Pθ , ∆Pϕ , ∆Pr ), (7.2.5)

as defined by (7.1.57).

• The pressure gradient is defined by:

 
1 ∂p 1 ∂p 1 ∂p
− ∇p = − , , . (7.2.6)
r ∂ θ r sin θ ∂ ϕ α ∂ r

• The relativistic effect is reflected in the following terms:

  
ν dα ∂ ν dα ∂ ν ∂ 1 dα
FG P = − 2 (rPθ ), − 2 (rPϕ ), Pr , (7.2.7)
2α r dr ∂ r 2α r dr ∂ r 2α ∂ r α dr
which can be regarded as the coupling interaction between the gravitational potential
α and the electromagnetic, the weak and the strong potentials represented by the
viscous coefficient ν .
We shall see that the force (7.2.7) is responsible to the supernova’s huge explosion.

• The thermal expansion force is due to the coupling between the gravity ∇ψ and the
heat Q:  
c2 d ψ
FT = 0, 0, βT , (7.2.8)
2α dr
which is the main driving force for generating stellar interior circulations and nebular
matter spurts of red giants.

The two forces (7.2.7) and (7.2.8) are the main driving forces for the stellar motion, and
hereafter we derive their explicit formulas.
1. Formula for thermal expansion force. The thermal expansion force (7.2.8) is radially
symmetric, which is simply written in the r-component form
c2 d ψ
fT = β T.
2α dr
7.2 Stars 433

In its nondimensional form, f T is expressed as

fT = σ (r)T, (7.2.9)

and σ (r) is called the thermal factor given by

c2 r04 β 1 dT d ψ
σ (r) = − . (7.2.10)
2κ 2 α dr dr
Here α , ψ , T satisfy equations (7.1.79)-(7.1.80) with the boundary conditions (7.1.81)-
(7.1.84). The detailed derivation of (7.2.9)-(7.2.10) will be given hereafter.
The σ -factor (7.2.10) can be expressed in the following form, to be deduced later:
 
c2 r3 (1 − δ )β eζ (r) 2 1 δ r2 + η
σ (r) = 0 2 2 · (1 − δ r − η ) · + r ξ (7.2.11)
2κ r eζ (1) r2 1 − δ r2 − η
 Z 
1 1 r2 Q
· A− dr for 0 6 r 6 1,
κ r 1 − δ r2 − η
where Z r 2 0 0
1 r ψφ
η= dr,
2r 0 αψ
Z r 
α
ζ= + rξ dr, (7.2.12)
0 r
8π G α 0φ 0
ξ= α p + φ 00
− for 0 6 r 6 1,
c2 2α
δ is called the δ -factor given by
2mG
δ= 2 , (7.2.13)
c r0
and m, r0 are the mass and radius of the star.
2. Formula for the relativistic effect. The term FG P in (7.2.7) can be expressed in the
following form:
   
η0 ∂
−ν δ + (rPθ )

 2r ∂ r 

   
 η0 ∂ 
FG P =  −ν δ + (rPϕ ) , (7.2.14)

  2r ∂ r 

 ν (2δ r + η 0 )2 
00 P + ν (2δ r + η 0 ) ∂ P

+ 2 δ + η r r
2 1−δr −η 2 2 ∂r
where η , δ are as in (7.2.12) and (7.2.13).
The force FG P is of special importance in studying supernovae, black holes, and the
galaxy cores. In fact, by the boundary conditions (7.1.81)-(7.1.83), we can reduce that the
radial component of the force (7.2.14) on the stellar shell is as
 
2νδ 2 ∂ Pr
fr = + δ + φ 00 (1) Pr + δ ν ,
1−δ ∂r
434 Chapter 7 Astrophysics and Cosmology

which has
2νδ 2
fr ∼ Pr → ∞ as δ → 1 (for Pr > 0). (7.2.15)
1−δ
The property (7.2.15) will lead to a huge supernovae explosions as they collapse to the radii
r0 → 2mG/c2 . It is the explosive force (7.2.15) that prevents the formation of black holes;
see Sections 7.2.6 and 7.3.3.
3. Derivation of formula (7.2.14). To deduce (7.2.14) we have to derive the gravitational
potential α . The first equation of (7.1.79) can be rewritten as
 
dM 2 c2 r 2 ψ 0 φ 0 c2 r 1
= 4π r ρ 0 − , M= 1− ,
dr 4G αψ 2G α
It gives the solution as
Z r 2 0 0  −1
4 c2 r ψφ 2MG
M = π r 3 ρ0 − dr, α = 1− 2 .
3 4G 0 αψ c r
4
By ρ0 = m/ π r03 , and in the nondimensional form (r → r0 r), we get
3
α = (1 − δ r2 − η )−1 for 0 6 r 6 1, (7.2.16)

where η , δ are as in (7.2.12) and (7.2.13). By (7.1.81) we have

η (1) = 0, (i.e. η (r0 ) = 0). (7.2.17)

Then, the formula (7.2.14) follows from (7.2.16).

4. Derivation of σ -factor (7.2.11). By (7.2.10) we need to calculate T 0 and ψ 0 . By
(7.1.80), T 0 can be expressed in the form
Z r
dT 1 a
=− 2 r2 α Qdr + ,
dr κr 0 r2
where a is a determined constant. By (7.1.84) we obtain
Z r0
1
a = −Ar02 + r2 α Qdr.
κ 0

In the nondimensional form, we have

Z 1
dT A 1
=− 2 + 2 r2 α Qdr for 0 6 r 6 1, A > 0. (7.2.18)
dr r κr r

To consider ψ 0 , by the second equation of (7.1.79) we obtain

k
ψ = eζ (r) , (7.2.19)
r
7.2 Stars 435

where ζ (r) is as in (7.2.12), k is a to-be-determined constant. In view of (7.1.81), i.e.

ψ (1) = 1 − δ , we have
k = (1 − δ )e−ζ (1).

Then, it follows from (7.2.19) that

 
dψ (1 − δ )eζ (r) α −1
= + rξ for 0 6 r 6 1, (7.2.20)
dr eζ (1)r0 r2

where ξ is as in (7.2.12). By (7.1.82) and (7.2.16)-(7.2.17), we can deduce that

ξ (1) = 0, (i.e. ξ (r0 ) = 0). (7.2.21)

Thus, the σ -factor (7.2.11) follows from (7.2.16), (7.2.18) and (7.2.20).

5. Thermal Force and (7.2.21). The thermal expansion force acting on the stellar shell
(i.e. at r = r0 ) can be deduced from (7.2.9) and (7.2.11) in the following (nondimensional)
form
c2 r03 β (1 − δ )δ
fT = σ0 T, σ0 = A (A > 0). (7.2.22)
2κ 2
By 0 < δ < 1, we have

σ0 > 0 (σ0 = σ (1), i.e. σ (r0 )).

Hence, it follows that there is an ε > 0 such that

σ (r) > 0, for ε < r 6 1. (7.2.23)

The positiveness of σ (r) in (7.2.23) shows that the thermal force f T of (7.2.22) is an
outward expanding force. It is this power that causes the swell and the nebular matter spurt
of a red giant. We also remark that the temperature gradient A on the boundary is maintained
by the heat source Q.

7.2.3 Stellar interior circulation

Recapitulation of dynamic transition theory

First we briefly recall the dynamic transition theory developed by the authors in (Ma and
Wang, 2013b) and the references therein. Many dissipative systems, both finite and infinite
dimensional, can be written in an abstract operator equation form as follows

du
= Lλ u + G(u, λ ), (7.2.24)
dt
where Lλ is a linear operator, G is a nonlinear operator, and λ is the control parameter.
436 Chapter 7 Astrophysics and Cosmology

It is clear that u = 0 is a stationary solution of (7.2.24). We say that (7.2.24) undergoes

a dynamic transition from u = 0 at λ = λ1 if u = 0 is stable for λ < λ1 , and unstable for
λ > λ1 . The dynamic transition of (7.2.24) depends on the linear eigenvalue problem:

Lλ ϕ = β (λ )ϕ .

Mathematically this eigenvalue problem has eigenvalues βk (λ ) ∈ C such that

Reβ1 (λ ) > Reβ2 (λ ) > · · · .

The following are the main conclusions for the dynamic transition theory; see (Ma and
Wang, 2013b) for details:

• Dynamic transitions of (7.2.24) take place at (u, λ ) = (0, λ1 ) provided that λ1 satisfies
the following principle of exchange of stability (PES):

 <0 for λ < λ1 (or λ > λ1 ),

Reβ1 = 0 for λ = λ1 ,

 (7.2.25)
>0 for λ > λ1 (or λ < λ1 ),
Reβk (λ1 ) < 0, ∀k > 2.

• Dynamic transitions of all dissipative systems described by (7.2.24) can be classified

into three categories: continuous, catastrophic, and random. Thanks to the universal-
ity, this classification is postulated in citeptd as a general principle called principle of
dynamic transitions.

• Let uλ be the first transition state. Then we can also use the same stratege outlined
above to study the second transition by considering PES for the following linearized
eigenvalue problem
Lλ ϕ + DG(uλ , λ )ϕ = β (2) (λ )ϕ .

Also we know that successive transitions can lead to chaos.

Stellar interior circulation

The governing fluid component equations are (7.1.85). We first make the nondimen-
sional. Let

(r, τ ) = (r0 r0 , r02 τ 0 /κ ),

!
d Te 2 0 2
(P, T, p, ρ ) = κρ0 P /r0 , − r0 T , ρ0 κ p /r0 , ρ0 ρ .
0 0
dr
7.2 Stars 437

Then the equations (7.1.85) are rewritten as (drop the primes):

∂P 1 1
+ (P · ∇)P = Pr∆P + FG P + σ (r)T~k − ∇p,
∂τ ρ κ
∂T 1 (7.2.26)
+ (P · ∇)T = ∆T
e + Pr ,
∂τ ρ
divP = 0,

where P = (Pθ , Pϕ , Pr ),~k = (0, 0, 1), σ (r) and FG P are as in (7.2.11) and (7.2.14), Pr = ν /κ
is the Prandtl number, and the ∆ is given by
2 ∂ Pr 2 cos θ ∂ Pϕ Pθ
∆Pθ = ∆P
e θ+ − 2 2 − 2 2 ,
r2 ∂ θ r sin θ ∂ ϕ r sin θ
e ϕ + 2 ∂ Pr + 2 cos θ ∂ Pθ − Pϕ ,
∆Pϕ = ∆P (7.2.27)
r2 sin θ ∂ ϕ r2 sin θ ∂ ϕ r2 sin2 θ
 
e r − 2 Pr + ∂ Pθ + cos θ Pθ + 1 ∂ Pϕ ,
∆Pr = ∆P
α r2 ∂θ sin θ sin θ ∂ ϕ
Based on the dynamic transition theory introduced early in this section, the stellar cir-
culation depends on the following three forces:
1
Pr∆P, FG P, σ T~k, (7.2.28)
κ
where in general Pr∆P prevents/slows-down the circulation.
By (7.2.14) we see that FG P depends on the δ -factor. The Sun’s δ -factor is δ =
10 /2, and in general the δ -factors for stars are of the order:
−5

δ ∼ 10−8 for red giants,

δ ∼ 10 −5
for main-sequence stars,
δ ∼ 10−3 for white dwarfs, (7.2.29)
δ ∼ 10−1 for neutron stars,
δ =1 for black holes.

Hence it is clear that all stars, except black holes, have small δ -factors.
On the other hand, for a small δ -factor, it follows from equations (7.1.79) and (7.1.80)
that α , ψ , φ has the order

α ∼ 1 + δ + η, η , η 0 ∼ δ 2, α0 ∼ δ ,
ψ ∼ rδ , ψ 0 /ψ ∼ δ , φ 0 , φ 00 ∼ δ .

Hence, we deduce from (7.2.14) that

FG P ∼ δ P for δ  1.
438 Chapter 7 Astrophysics and Cosmology

Thus, in view of (7.2.29) we conclude that the relativistic effect FG P is negligible on the
stellar interior motion for all stars except supernovae.
Hence the main driving force for stellar circulations is the thermal expansion force char-
acterized by the σ -factor σ0 in (7.2.22). Due to δ  1, σ0 can be approximately given
by
r2 mGβ
σ0 = 0 2 A, (7.2.30)
κ
which plays the similar role as the Rayleigh number Re in the earth’s atmospheric circu-
lation. The value σ0 of (7.2.30) is large enough to generate thermal convections for main-
sequence stars and red giants.
We remark that the heat source Q is caused not only by nuclear reactions, but also by the
pressure gradient, the density and the gravitational potential energy. Based on the σ -factor
in (7.2.30), we obtain the following physical conclusions:

1. Main-sequence stars. Based on the dynamic transition theory, by (7.2.23), we deduce

that there is a critical number σc > 0, independent of the parameter σ0 in (7.2.30), such that
equations (7.2.26) undergo no dynamic transition if σ0 < σc , and a dynamic transition if
σ0 > σ c : (
< 0 there is no stellar circulation,
σ0 − σ c (7.2.31)
> 0 there exists stellar circulation.
In particular, σc has the same order of magnitude as the first eigenvalue λ1 of the the fol-
lowing equations in the unit ball B1 :

− Pr∆P + ∇p = λ1 P for x ∈ B1 ⊂ R3 ,
 
∂ Pθ ∂ Pϕ
Pr , , =0 at r = 1,
∂r ∂r

where ∆P is as in (7.2.27).
For the main-sequence stars, the σ -factors are much larger than the first eigenvalue λ 1
of (7.2.32). For example, the Sun consists of hydrogen, and

r0 = 7 × 108m, m = 2 × 1030kg, G = 6.7 × 10−11m3 /(kg · s2 ).

Using the average temperature T = 106 K, the parameter κ is given by

 1.72
T
κ = 0.18 10−4 m2 /s ' 50m2 /s.
190k

With thermal expansion coefficient β in the order β ∼ 10−4/K, the σ -factor of (7.2.30) for
the Sun is about
σ ∼ 1030A [m/K]. (7.2.32)
7.2 Stars 439

Due to nuclear fusion, stars have a constant heat supply, which leads to a higher bound-
ary temperature gradient A. Referring to (7.2.32), we conclude that there are always interior
circulations and thermal motion in main-sequence stars and red giants, which has large σ -
factors.

2. Red giants. The nuclear reaction of a red giant stops in its core, but does take place
in the shell layer, which maintains a larger temperature gradient A on the boundary than
the main-sequence phase. Therefore, the greater σ -factor makes the star to expand, and the
increasing radius r0 raises the σ -factor (7.2.30). The increasingly larger σ -factor provides
a huge power to hurl large quantities of gases into space at very high speed.

3. Neutron stars and pulsars. Neutron stars are different from other stars, which have
bigger δ -factors, higher rotation Ω and lower σ -factor (as the nuclear reaction stops). In-
stead of (7.2.26) the dynamic equations governing neutron stars are

∂P 1 1
+ (P · ∇)P = Pr∆P + FG P − 2~Ω × P − ∇p + σ T~k,
∂τ ρ κ
∂T 1 (7.2.33)
+ (P · ∇)T = ∆T
e + Pr ,
∂τ ρ

divP = 0.

As the nuclear reaction ceases, the temperature gradient A tends to zero as time t → ∞,
and consequently
σ → 0 as t → ∞. (7.2.34)

Based on the dynamic transition theory briefly recalled earlier in this section, we derive
from (7.2.33) and (7.2.34) the following assertions:

• By (7.2.34), neutron stars will eventually stop convection.

• Due to the high rotation Ω, the convection of (7.2.33) for the early neutron star is time
periodic, and its period T is inversely proportional to Ω, and its convection intensity
√
B is proportional to σ − σc , i.e.

C1
period T ' ,
Ω (7.2.35)
√
intensity B ' C2 σ − σc ,

where σc is the critical value of the transition, and c1 , c2 are constants. The properties
of (7.2.35) explain that the early neutron stars are pulsars, and by (7.2.34) their pulsing
√ √
radiation intensities decay at the rate of σ − σc or kA − σc (k is a constant).
440 Chapter 7 Astrophysics and Cosmology

7.2.4 Stellar atmospheric circulations

The model describing stellar atmosphere circulation without rotation is given by (7.1.71)-
(7.1.73), and the eigenvalue equations read as

e θ + 2 ∂ Pr − 2 cos θ ∂ Pϕ − Pθ − c0 Pθ
Pr ∆P
r2 ∂ θ r2 sin2 θ ∂ ϕ r2 sin θ r2 ν

δ ∂ 1 ∂ϕ
+ 2 (rPθ ) − = β Pθ ,
2r ∂ r r ∂θ

e ϕ + 2 ∂ Pr + 2 cos θ ∂ Pθ − Pϕ − c0 Pϕ
Pr ∆P
r2 sin θ ∂ ϕ r2 sin2 θ ∂ ϕ r2 sin2 θ r2 ν

δ ∂ 1 ∂p
+ 2 (rPϕ ) − = β Pϕ ,
2r ∂ r r sin θ ∂ ϕ (7.2.36)

e r − 2 (1 − δ )(Pr + ∂ Pθ + cos θ Pθ + 1 ∂ Pϕ ) − c1 Pr
Pr ∆P
r2 ∂θ sin θ sin θ ∂ ϕ r2 v

δ2 1 δ ∂ Pr 1 √ ∂p
+ Pr − + ReT − (1 − δ ) = β Pr ,
2(1 − δ ) r2 2r ∂ r r2 ∂r
√
e + 1 RePr = β T,
∆T
r 2

divP = 0,

with the boundary conditions

∂ Pθ ∂ Pϕ
T = 0, Pr = 0, = = 0 at r = r0 , r0 + 1, (7.2.37)
∂r ∂r
where
   
1 ∂ ∂f 1 ∂2 f 1−δ ∂ ∂f
∆
ef = sin θ + + 2 r2 ,
r2 sin θ ∂ θ ∂θ r2 sin θ ∂ ϕ 2
2 r ∂r ∂r

and δ = 2mG/c2 rh is the δ -factor with r0 < r < r0 + 1 being nondimensional.

For stellar atmosphere circulations, the star radius r0 is much greater than the convection
height h = r1 − r0 . Hence, we can approximatively take δ as the δ -factor as defined in
(7.1.74).
Thus the eigenvalues β of (7.2.36)-(7.2.37) depend on δ and the Rayleigh number Re
in (7.1.74):
β = β (δ , Re), (0 < δ < 1, 0 < Re).

For (7.2.36)-(7.2.37), the following conclusions hold true:

7.2 Stars 441

1. The differential operator ∆ in (7.2.36) is the Laplace-Beltrami operator, which is

symmetric. Hence, the differential operator in (7.2.36)-(7.2.37) is symmetric. It implies
that all eigenvalues β of (7.2.36)-(7.2.37) are real. Hence, the stellar atmospheric system
(7.1.71)-(7.1.73) undergo the first dynamic transition to stationary solutions.
2. Let L be the differential operator on the left-hand side of (7.2.36); then the first
eigenvalue β1 satisfies
1
β1 = max hLu, ui. (7.2.38)
||u||=1 2

As the δ -factor of (7.1.74) is smaller than one, then by (7.2.38), β 1 satisfies the inequality
√
β1 6 −k1 + k2 Re, (7.2.39)

for some constants k1 , k2 > 0. By (7.1.74) we see that

Re → ∞ as m → ∞, or as r0 → ∞, for (T0 − T1 ) > 0.

By (7.2.39) we get
β1 → +∞ as m → ∞, or as r0 → ∞.
In addition, it is known that
(k)
β1 → +∞ (k > 2) as β1 → +∞.

Hence based on the dynamic transition theory, we derive the following physical conclu-
sions.
Physical Conclusion 7.10 For the stars with small δ  1, their atmosphere have
thermal convections to occur. In particular for the stars with large mass and radius, the
convections are in turbulent state.

3. In the third equation of (7.2.36), the term

δ2 1
F= Pr (7.2.40)
2(1 − δ ) r2
represents the radially expanding force which is the relativistic gravitational effect. The
force (7.2.40) satisfies
F → +∞ as δ → 1 for Pr > 0.
It implies that the force F in (7.2.40) is explosive as δ → 1. In addition, by (7.2.38) the first
eigenvalue β1 satisfies the following inequality
√ k3
β1 6 −k1 + k2 Re + for δ → 1.
1−δ
Thus we can deduce the following conclusions.
442 Chapter 7 Astrophysics and Cosmology

Physical Conclusion 7.11 Due to the explosive force F of (7.2.40), the stars with
δ w 1(δ = 1 is the black hole) has no the stellar atmospheres, which are erupted into the
outer space. However, as 1 − δ > 0 is small, the stellar atmospheres are in the turbulent
convection state,caused by the relatistic gravitational effect.

4. Let lτ and lr be the convection scales in the horizontal direction and radial direction.
The ratio
lτ
r=
lr
depends on the coefficient ratio of the horizontal components of the momentum (Pθ , Pϕ )
and the radial momentum component Pr in (7.2.36), i.e. the ratio

δ − 1 − c0/ν
η= . (7.2.41)
−2(1 − δ ) − c1/ν + δ 2 /2(1 − δ )

As |η |  ∞, γ andη have the qualitative relation


 η −1 for η > 0,
γ∝

|η | for η < 0.

By (7.2.41) we see that

η > 0 and η −1 ∼ o(1) for δ  1 and large ν ,

η < 0 and |η |  1 for δ → 1.

Hence we deduce the following conclusions.

Physical Conclusion 7.12 The convection scale ratio γ depends on δ for δ  1 and
δ → 1, and on the ratio c0 /c1 for δ < 1 and v  1. Moreover, γ possesses the following
properties 
 o(1) for δ  1 and v large,
γ= (7.2.42)
 1 for δ → 1.

1
The sun atmosphere convections satisfy the relation (7.2.42). For the Sun, δ w ×10−5 .
2
The observations show that
.
the photosphere convection r = 1 ∼ 2,
.
the chromosphere convection r = 2 ∼ 3.

For the solar corona, v  1, we need to know the ratio c0 /c1 .

7.2 Stars 443

5. For the stars with rotation ~Ω, the corresponding eigenvalue equations with the Cori-
olis force are " √ #
f Re~
Pr ∆P − 2 + 2 kT − 2~Ω × P = β P,
r r
√ (7.2.43)
Re
∆T
f+ Pr = β T,
r2
div P = 0,
with the boundary conditions (7.2.37).
In (Ma and Wang, 2013b) we showed that there is a lower bound Ω 0 such that as Ω > Ω0
the first eigenvalue β1 of (7.2.43) is complex in the critical state (7.2.25). Therefore, for the
stars with the bigger rotation their atmosphere convections in the first phase transition are
time periodic.

7.2.5 Dynamics of stars with variable radii

For stars with varying sizes and for supernovae, their radii expand and shrink periodically.
Therefore, the metric in the interior of such stars is as follows:
 
ds2 = −ψ c2 dt 2 + R2 (t) α dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ) ,

where ψ = ψ (r,t), α = α (r,t), and R(t) is the scalar factor representing the star radius. For
convenience, we denote

ψ = eu(r,t) , α = ev(r,t) , R2 (t) = ek(t) , 0 6 r 6 1.

Then the metric is rewritten as

 
ds2 = −eu c2 dt 2 + ek ev dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ) . (7.2.44)

The stars with variable radii are essentially in radial motion. Hence, the horizontal
momentum (Pθ , Pϕ ) is assumed to be zero:

(Pθ , Pϕ ) = 0. (7.2.45)

In the following we develop dynamic models for astronomical objects with variable sizes.
1. Gravitational field equations. We recall the gravitational field equations (Ma and
Wang, 2014e):
8π G 1 1
Rµν = − 4 (Tµν − gµν T ) − (Dµν φ − gµν Φ). (7.2.46)
c 2 2
The nonzero components of the metric (7.2.44) are

g00 = −eu , g11 = ek+v , g22 = ek r2 , g33 = ek r2 sin2 θ ,

444 Chapter 7 Astrophysics and Cosmology

the nonzero components of the Levi-civita connections are

1 1 v−u 1
Γ000 = ut , Γ011 = e (kt + vt ), Γ010 = ur ,
2c 2c 2
r2 r2 −u 1
Γ022 = e−u kt , Γ033 = e kt sin2 θ , Γ100 = eu−v ur
2c 2c 2
1 1 1
Γ11 = vr , Γ110 = (kt + vt ), Γ122 = −re−v ,
2 2c
1
Γ133 = −re−v sin2 θ , Γ221 = , Γ233 = sin θ cos θ ,
r
1 cos θ
Γ331 = , Γ332 = ,
r sin θ
and the nonzero components of the Ricci curvature tensor read
 
1 3 2 1 2
R00 = 2 3ktt + kt + vtt + vt + kt vt − ut (kt + vt )
2c 2 2
 
1 u−k−v 1 2 1 2
− e urr + ur − ur vr + ur ,
2 2 2 r
 
ek+v−u 3 2 2 1
R11 = − k tt + k + vtt + v + 3k t vt − ut (kt + vt )
2c2 2 t t
2
 
1 1 2 1 2
+ urr + ur − ur vr − vr ,
2 2 2 r
2 k−u
  h i
r e 3 2 1 r
R22 = − 2
ktt + kt + k (v
t t − ut ) − e−v ev + (kr + vr − ur ) − 1 ,
2c 2 2 2
R33 = R22 sin2 θ ,
1 h r i
R10 = − (1 + ur )kt + vr .
cr 2
The energy-momentum tensor is in the form
 
ρ g00 g11 Pr c 0 0
g00 g11 Pr c g11 p 0 0 
Tµν = 

,
0 0 g22 p 0 
0 0 0 g33 p

where ρ is the energy density, Pr is the radial component of the momentum density. Then
direct computations imply that
1 1
T = gµν Tµν = −ρ + 3p, T00 − g00 T = (ρ + 3p),
2 2
1 1 1 1
T11 − g11 T = ek+v (ρ − p), T22 − g22 T = ek r2 (ρ − p),
2 2 2 2
1 1 1
T33 − g33 T = (T22 − g33 T ) sin2 θ , T10 − g10 T = g00 g11 Pr c.
2 2 2
7.2 Stars 445

To derive an explicit expression of (7.2.46), we need to compute the covariant derivatives of

the dual gravitational field φ :

∂ 2φ ∂φ
Dµν φ = − Γλµν λ .
∂ x µ ∂ xν ∂x
Let φ = φ (r,t). Then we have

1 1 1
D00 φ = φtt − 2 ut φt − eu−v ur φr ,
c2 2c 2
1 v−u 1
D11 φ = φrr − 2 e (kt + vt )φt − vr φr ,
2c 2
r2 −u
D22 φ = − 2 e kt φt + re−v φr ,
2c
D33 φ = D22 φ sin2 θ ,
1
D10 φ = φrt − (ur φt + φr kt + φr vt ).
2c
Thus, the field equations (7.2.46) are written as

R10 = −D1 D0 φ ,
8π G 1 1
Rkk = − 4 (Tkk − gkk T ) − (Dkk φ − gkk Φ) for k = 0, 1, 2,
c 2 2
which are expressed as
 rur  8π Gr r
1+ kt + vt = 2 eu+k+v Pr + crφrt − (ur φt + φr kz + φr vt ), (7.2.47)
2 c 2
3 2 1 2
3ktt + kt + vtt + vt + kt vt − ut (kt + vt ) (7.2.48)
2 2  
2 u−k−v 1 1 2
−c e urr + u2r − ur vr + ur
2 2 r
 
8π G 2 u−k−v 2eu−k
= − 2 (ρ + 3p) − c D00 φ + e D11 φ + 2 D22 φ ,
c r
3 1
ktt + kt2 + vtt + vt2 + 3kt vt − ut (kt + vt ) (7.2.49)
2  2 
2 u−k−v 1 2 1 2
−c e urr + ur − ur vr − vr
2 2 r
8π G u 2 u−k−v 2eu−k
= e ( ρ − p) + c (e D 11 φ − D 00 φ − D22 φ ),
c2 r2
3 1 2c2 eu−k−v h v r i
ktt + kt2 + kt (vt − ut ) + 2
e + (kr + vr − ur ) − 1 (7.2.50)
2 2 r 2
8π G u 2 u−k−v
= 2 e (ρ − p) + c (D00 φ − e D11 φ ),
c
446 Chapter 7 Astrophysics and Cosmology

The equations (7.2.47)-(7.2.50) have seven unknown functions u, v, k, φ , Pr , ρ , p, in which

Pr , ρ , p satisfy the fluid dynamic equations and the equation of state introduced hereafter.
2. Fluid dynamic model. The fluid dynamic model takes the momentum representation
equations coupling the heat equation. Under the condition (7.2.45) and the radial symmetry,
they are given as follows:
∂ Pr 1 ∂ Pr 1 ∂ v 2
+ Pr + P (7.2.51)
∂τ ρ ∂r 2 ∂r r
    
−v 1 ∂ 2 ∂ Pr 2 1 ∂ ∂v
= νe r − 2 Pr + Pr
r2 ∂ r ∂r r 2 ∂r ∂r
   
∂ e−v/2 ∂ 2 v/2 −v ∂ p ρ ∂ eu
+ γ e−v (r e Pr ) − e − (1 − β T ) ,
∂ r r2 ∂ r ∂r 2 ∂r
 
∂T 1 ∂T κ e−v ∂ ∂T
+ Pr = 2 r2 + Q(r), (7.2.52)
∂τ ρ ∂r r ∂r ∂r
∂ ρ e−v/2 ∂ 2 v/2
+ 2 (r e Pr ) = 0. (7.2.53)
∂τ r ∂r

3. Equation of state. We know that the gravitational field equations represent the law of
gravity, which essentially dictates the gravity related unknowns: e u , ev , R = ek/2 , φ .
The laws for describing the matter field are the motion equation (7.2.51), the heat equa-
tion (7.2.52), and the energy conservation equation (7.2.53). To close the system, one needs
to supplement an equation of state given by thermal dynamics, which provides a relation
between temperature T , pressure p, and energy density ρ :

f (T, p, ρ ) = 0, (7.2.54)

which depends on the underlying physical system.

In summary, we have derived a consistent model coupling the gravitational field equa-
tions, the fluid dynamic equations and the equation of state consists of eight equations solv-
ing for eight unknowns: ψ = eu , α = ev , R = ek/2 , φ , Pr , T , p and ρ .
4. Energy conservation formula. From the energy conservation equation (7.2.53), we
can deduce energy conservation in the following form
1 d
R3 r2 ev/2 Pr + Er = 0 for 0 < r < 1, (7.2.55)
4π dt
where r = 1 stands for the boundary R = ek/2 of the star, Er is the total energy in the ball Br
with radius r.
To see this, we first note that the volume differential element of the Riemannian manifold
is given by
√
dV = g11 g22 g33 drd θ d ϕ = e3k/2 r2 ev/2 sin θ drd θ d ϕ .
7.2 Stars 447

Taking volume integral for (7.2.53) on Br implies that

Z Z r Z Z
d ∂  2 v/2  π 2π
ρ dV + R3 r e Pr sin θ d θ d ϕ = 0,
dt Br 0 ∂r 0 0

which leads to
dMr
+ 4π R3r2 ev/2 Pr = 0,
dt
and (7.2.55) follows.

5. Shock wave. As the total energy ER of the star is invariant, we have

d
ER = 0.
dt
It follows from (7.2.55) that

Pr = 0 on r = 1 (i.e. on the boundary R). (7.2.56)

On the other hand, the physically sound boundary condition for the star with variable radius
is
∂ Pr
= 0 on r = 1, (7.2.57)
∂r
which means that there is no energy exchange between the star and its exterior. Thus,
(7.2.56) and (7.2.57) imply that there is a shock wave outside the star near the boundary.

Remark 7.13 Formula √ (7.2.55) is very important. In fact, due to the boundary con-
dition (7.1.81) and ev/2 ' 1/ 1 − δ , in the star shell layer, (7.2.55) can be approximately
written as √
1−δ d
ρ Pr = − Mr for R − r > 0 small, (7.2.58)
4π R2 dt
where δ = 2Mr0 G/c2 R. This shows that a collapsing supernova is prohibited to shrink into
a black hole (δ = 1). In fact, the strongest evidence for showing that black holes cannot
be created comes from the relativistic effect of (7.2.14), which provides a huge explosive
power in the star shell layer given by

νδ 2
Pr → ∞ as δ → 1 (Pr 6= 0). (7.2.59)
1−δ
Here Pr is the convective momentum different from the contracting momentum Pr in (7.2.58);
see Section 7.3.3 for details.

Remark 7.14 One difficulty encountered in the classical Einstein gravitational field
equations is that the number of unknowns is less than the number of equations, and con-
sequently the coupling between the field equations and fluid dynamic and heat equations
become troublesome.
448 Chapter 7 Astrophysics and Cosmology

7.2.6 Mechanism of supernova explosion

In its late stage of life, a massive red giant collapses, leading to a supernova’s huge explo-
sion. It was still a mystery where does the main source of driving force for the explosion
come from, and the current viewpoint, that the blast is caused by the large amount of neu-
trinos erupted from the core, is not very convincing.
The stellar dynamic model (7.1.78)-(7.1.84) provides an alternative explanation for su-
pernova explosions, and we proceed in a few steps as follows:
1. When a very massive red giant completely consumes its central supply of nuclear
fuels, its core collapses. Its radius r0 begins to decrease, and consequently the δ -factor
increases:
2mG
r0 decreases ⇒ δ = 2 increases.
c r0
2. The huge mass m and the rapidly reduced radius r0 make the δ -factor approaching
one:
δ → 1 as r0 → Rs
where Rs = 2mG/c2 is the Schwarzschild radius.
3. By (7.2.58), the shrinking of the star slows down:
p
Pr ∼ 1 − δ ,

and nearly stops as δ → 1.

4. Then the model (7.2.26) is valid, and the eigenvalue equations of (7.2.26) are given
by
1
Pr∆P + FG P + σ T~k − ∇p = β P,
κ
e + Pr = β T,
∆T (7.2.60)

divP = 0.
The first eigenvalue β depends on the δ -factor, and by (7.2.14)
 1/2
Prδ 2
β1 ∼ as δ → 1. (7.2.61)
1−δ
Based on the transition criterion (7.2.25), the property (7.2.61) implies that the star has
convection in the shell layer, i.e., the radial circulation momentum flux Pr satisfies

Pr > 0 in certain regions of the shell layer.

5. The radial force (7.2.15) in the shell layer is

2Pr δ 2
fr ' Pr → ∞ as δ → 1 and Pr > 0,
1−δ
7.2 Stars 449

which provides a very riving force, resulting in the supernova explosion, as shown in Figure
7.2.

Figure 7.2 Circulation in a shell layer causing blast

6. Since Pr = 0 at r = r0 , the radial force of (7.2.14) is zero:

fr = 0 at r = r0 .

Here r0 is the radius of the blackhole core. Hence, the supernova’s huge explosion preserves
an interior core of smaller radius containing the blackhole core, which yields a neutron star.
In particular, the huge explosion has no imploding force, and will not generate a new black
hole.
The analysis in Steps 1-6 above provides the supernova exploding mechanism, and
clearly provides the power resource of the explosion.
In addition, by (7.2.12) and (7.2.16) we have
Z r 2 0 0
1 1 r ψφ
α= , η (r) = dr for 0 6 r 6 r0 .
1 − δ r2/r02 − η (r) 2r 0 αψ

We can verify that

η (r) > 0 for 0 < r < r0 . (7.2.62)
In fact, by (7.2.17) and (7.2.12) we have

η (0) = 0, η (r0 ) = 0. (7.2.63)

Therefore, η has an extremum r (0 < r < r) satisfying

η 0 (r) = 0.
1
Let η = f . Then
r
450 Chapter 7 Astrophysics and Cosmology

η 0 (r) = 0 ⇒ f (r) = ea r (a = constant).

Hence, at the extremum r, η takes a positive value
1
η (r) = f (r) = ea > 0 for 0 < r < r0 . (7.2.64)
r
Thus, (7.2.62) follows from (7.2.63) and (7.2.64).
The fact (7.2.62) implies that the critical δ -factor δc for the supernova explosion is less
than one, i.e. δc < 1.

7.3 Black Holes

7.3.1 Geometric realization of black holes
The concept of black holes was originated from the Einstein general theory of relativity.
Based on the Einstein gravitational field equations, K. Schwarzschild derived in 1916 an
exact exterior solution for a spherically symmetrical matter field, and Tolman-Oppenheimer-
Volkoff derived in 1939 an interior solution; see Section 7.1.2. In both solutions if the radius
R of the matter field with mass M is less than or equal to a critical radius Rs , called the
Schwarzschild radius: 2MG
R 6 Rs = 2 , (7.3.1)
c
then the matter field generates a singular spherical surface with radius R s , where time stops
and the spatial metric blows-up; see Figure 7.3. The spherical region with radius R s is called
the black hole.

Figure 7.3 The spherical region enclosing a matter field with mass M and radius R satisfying (7.3.1)
is called black hole.

We recall again the Schwarzschild metric in the exterior of a black hole written as

ds2 = g00 c2 dt 2 + g11 dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ),

    (7.3.2)
2MG 2MG −1
g00 = − 1 − 2 , g11 = 1 − 2 ,
c r c r

where r > Rs when the condition (7.3.1) is satisfied.

7.3 Black Holes 451

In (7.3.2) we see that at r = Rs , the time interval is zero, and the spatial metric blows up:
 
√ Rs 1/2
−g00 dt = 1 − dt = 0 at r = Rs , (7.3.3)
r
 
√ Rs −1/2
g11 dr = 1 − dr = ∞ at r = Rs . (7.3.4)
r

Physically, the proper time and distance for (7.3.2) are

√
proper time = −g00 t,
p
proper distance = g11 dr2 + r2 d θ 2 + r2 sin θ d ϕ 2 .

The coordinate system (t, x) with x = (r, θ , ϕ ) represents the projection of the real world to
the coordinate space. Therefore the radial motion speed dr/dt in the projected world differs
p
from the proper speed vr by a factor −g00/g11 , i.e.
dr p
= −g00 /g11vr .
dt
Hence, the singularity (7.3.3) and (7.3.4) means that for an object moving toward to the
boundary of a black hole, its projection speed vanishes:
dr
=0 at r = Rs .
dt
This implies that any object in the exterior of the black hole cannot pass through its boundary
and enter into the interior. In the next subsection we shall rigorously prove that a black hole
is a closed and innate system.
Mathematically, a Riemannian manifold (M , gi j ) is called a geometric realization (i.e.
isometric embedding) in RN , if there exists a one to one mapping

~r : M → RN ,

such that
d~r d~r
gi j = · .
dxi dx j
The geometric realization provides a “visual” diagram of M , the real world of our Universe.
In the following we present the geometric realization of a 3D metric space of a black
hole near its boundary. By (7.3.2), the space metric of a black hole is given by
 
Rs −1 2 2MG
ds2 = 1 − dr + r2 (d θ 2 + sin2 θ d ϕ 2 ) for r > Rs = 2 . (7.3.5)
r c

It is easy to check that a geometric realization of (7.3.5) is given by~r : M → R 4 :

n p o
~rext = r sin θ cos ϕ , r sin θ sin ϕ , r cos θ , 2 Rs (r − Rs ) for r > Rs . (7.3.6)
452 Chapter 7 Astrophysics and Cosmology

In the interior of a black hole, the Riemannian metric near the boundary is given by the
TOV solution (7.1.42), and its space metric is in the form
 −1
2 r2
ds = 1 − 2 dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ) for r < Rs , (7.3.7)
Rs
A geometrical realization of (7.3.7) is
 q 
~rint = r sin θ cos ϕ , r sin θ sin ϕ , r cos θ , ± Rs − r .
± 2 2 (7.3.8)

The diagrams of (7.3.6) and (7.3.8) are as shown in Figure 7.4, where case (a) is the embed-
ding (
~rext for r > Rs ,
~r+ = +
~rint for r < Rs ,
and case (b) is the embedding
(
−
~rext for r > Rs ,
~r = −
~rint for r < Rs .

The base space marked as R3 in (a) and (b) are taken as the coordinate space (i.e. the
projective space), and the surfaces marked by M represent the real world which are separated
into two closed parts by the spherical surface of radius Rs : the black hole (r < Rs ) and the
exterior world (r > Rs ).

Figure 7.4 M is the real world with the metric (7.3.5) for r > Rs and the metric (7.3.7) for r < Rs ,
and in the base space R3 the coordinate system is taken as spherical coordinates (r, θ , ϕ ).

In particular, the geometric realization of (7.3.7) for a black hole clearly manifests that
the real world in the black hole is a hemisphere with radius Rs embedded in R4 ; see Fig-
ure 7.4(a):
x21 + x22 + x23 + x24 = R2s for 0 6 |x4 | 6 Rs ,
where
 q 
(x1 , x2 , x3 , x4 ) = r sin θ cos ϕ , r sin θ sin ϕ , r cos θ , ± Rs − r .
2 2
7.3 Black Holes 453

We remark that the singularity of M at r = Rs , where the tangent space of M is perpen-

dicular to the coordinate space R3 , is essential, and cannot be removed by any coordinate
transformations. The coordinate transformations such as those given by Eddington and
Kruskal possess the singularity as well, and, consequently, cannot be used as the coordinate
systems for the metrics (7.3.5) and (7.3.7).

7.3.2 Blackhole theorem

The main objective of this section is to prove the following blackhole theorem.
Theorem 7.15 (Blackhole Theorem) Assume the validity of the Einstein theory of
general relativity, then the following assertions hold true:
1) black holes are closed: matters can neither enter nor leave their interiors;
2) black holes are innate: they are neither born to explosion of cosmic objects, nor born
to gravitational collapsing; and
3) black holes are filled and incompressible, and if the matter field is non-homoge-
neously distributed in a black hole, then there must be sub-blackholes in the interior
of the black hole.

We prove this theorem in three steps as follows.

Step 1. Closedness of black holes. First, it is classical that all matter, including photons,
cannot escape from a black hole when they are within the Schwarzschild radius.
Step 2. Now we need to show that all external energy cannot enter into the interior of a
black hole. By the energy-momentum conservation, we have
∂E
+ div P = 0, (7.3.9)
∂τ
where E and P are the energy and momentum densities. Take the volume integral of (7.3.9)
on B = {x ∈ R3 | Rs < |x| < R1 }:
Z  
∂E √
+ divP dΩ = 0, dΩ = gdrd θ d ϕ , (7.3.10)
B ∂ τ
where divP is as in (7.1.60), and
 
2MG −1
g = det(gi j ) = g11 g22 g33 = α r4 sin2 θ , α = 1− 2 .
c r
Let E be the total energy in B, then the Gauss formula, we have
Z
∂E d
dΩ = E ,
ZB∂ τ dt
Z p Z
√
divPdΩ = α (R1 )Pr dSR1 − lim α Pr dSr .
B S R1 r→Rs Sr
454 Chapter 7 Astrophysics and Cosmology

Here Sr = {x ∈ R3 | |x| = r}. In view of (7.3.10) we deduce that

Z p Z
dE √
= lim α Pr dSr − α (R1 ) Pr dSR1 . (7.3.11)
dt r→Rs Sr S R1

The equality (7.3.11) can be rewritten as

Z
" Z
#
1 dE p
lim Pr dSr = lim p + α (R1 ) Pr dSR1 = 0. (7.3.12)
r→Rs Sr r→Rs α (r) dt S R1

This together with no escaping of particles from the interior of the black hole shows that

lim Pr = 0.
r→R+
s

In other words, there is no energy flux Pr on the Schwarzschild surface, and we have shown
that no external energy can enter into a black hole.
In conclusion, we have shown that black holes are closed: no energy can penetrate the
Schwarzschild surface.

Step 3. Innateness of black holes. The explosion mechanism introduced in Subsection

7.2.6 clearly manifests that any massive object cannot generate a new black hole. In other
words, we conclude that black holes can neither be created nor be annihilated, and the total
number of black holes in the Universe is conserved.

Step 4. Assertion 3) follows by applying conclusion (7.5.55) and the fact that sub black-
holes are incompressible. The theorem is therefore proved.

We remark again that the singularity on the boundary of black holes is essential and can-
not be removed by any differentiable coordinate transformation with differentiable inverse.
The Eddington and Kruskal coordinate transformations are non-differentiable, and are not
valid.

Remark 7.16 The gravitational force F generated by a black hole in its exterior is
given by
mc2 ∂ψ
F= ∇g00 = −mg11 ,
2 ∂r
where ψ is the gravitational potential. By (7.3.2) we have the following gravitational force:
 
2MG mMG
F = − 1− 2 . (7.3.13)
c r r2

Consequently, on the boundary of a black hole, the gravitational force is zero:

F =0 at r = Rs .
7.3 Black Holes 455

7.3.3 Critical δ -factor

Black holes are a theoretical outcome. Although we cannot see them directly due to their
invisibility, they are, however, strong evidences from many astronomical observations and
theoretic studies.
In the following, we first briefly recall the Chandrasekhar limit of electron degeneracy
pressure and the Oppenheimer limit of neutron degeneracy pressure; then we present new
criterions to classify pure black holes, which do not contain other black holes in their inte-
rior, into two types: the quark and weakton black holes, by using the δ -factor.

1. Electron and neutron degeneracy pressures. Classically we know that there are two
kinds of pressure to resist the gravitational pressure, called the electron degeneracy pres-
sure and the neutron degeneracy pressure. These pressures prevent stars from gravitational
collapsing with the following mass relation:
(
1.4M for electron pressure,
m< (7.3.14)
3M for neutron pressure.

Hence, by (7.3.14), we usually think that a dead star is a white dwarf if its mass m < 1.4M ,
and is a neutron star if its mass m < 3M . However, if the dead star has mass m > 3M ,
then it is regarded as a black hole. Hence the neutron pressure gradient is thought to be
a final defense to prevent a star from collapsing into a black hole. Thus, 3M becomes a
critical mass to determine the possible formation of a black hole.

2. Interaction potential pressure. However, thanks to the strong and weak interaction
potentials established in (Ma and Wang, 2015a, 2014c), there still exist three kinds of po-
tential pressures given by

neutron potential, quark potential, weakton potential. (7.3.15)

These three potential pressures maintain three types of astronomical bodies:

neutron stars,
quark black holes if they exist, (7.3.16)
weakton black holes if they exist.

We are now in position to discuss these potential pressures. By the theory of elementary
particles, a neutron is made up of three quarks n = uud, and u, d quarks are made up of
three weaktons as u = w∗ w1 w1 , d = w∗ w1 w2 . The three levels of particles possess different
potentials distinguished by their interaction charges:
456 Chapter 7 Astrophysics and Cosmology

 
ρw 3
neutron charge gn = 3 gs ,
ρn
 3
ρw (7.3.17)
quark charge gq = gs ,
ρq
weakton weak charge gw ,

where ρn , ρq , ρw are the radii of neutron, quark and weakton.

Let g be a specific charge in (7.3.17). Then by the interaction potentials obtained in (Ma
and Wang, 2015a), the particle with charge g has a repulsive force:

g2
f= .
r2
The force acts on particle’s cross section with area S = π r 2 , which yields the interaction
potential pressure as
f g2
P= = 4. (7.3.18)
S πr
Let each ball Br with radius r contain only one particle. Then the mass density ρ is given
by
3m0
ρ= , (7.3.19)
4π r3
where m0 is the particle mass. By the uncertainty relation, in Br the particle energy ε0 is
h̄
ε0 = ,
2t
and t = r/v, where v is the particle velocity. Replacing v by the speed of light c, we have

h̄c
ε0 = .
2r
By m0 = ε0 /c2 , the density ρ of (7.3.19) is written as

3h̄ 3h̄
ρ= or equivalently r4 = . (7.3.20)
8π cr4 8π cρ

Inserting r4 of (7.3.20) into (7.3.18), we derive the interaction potential pressure P in the
form
8cρ g2
P= . (7.3.21)
3h̄

3. Critical δ -factors. It is known that the central pressure of a star with mass m and
radius r0 can be expressed as

Gm2 2π c2 2mG
PM = = ρδ , δ= , (7.3.22)
r04 3 c 2 r0
7.3 Black Holes 457

where δ is the δ -factor.

We infer from (7.3.21) and (7.3.22) the critical δ -factor as
4 g2
δc = , (7.3.23)
π h̄c
where g is one of the interaction charges in (7.3.17).
The critical δ -factor in (7.3.23) provides criterions for the three types of astronomical
bodies of (7.3.16).
4. Physical significance of δc . It is clear that for a star with m > 1.4M if
4 g2n
δ< , (7.3.24)
π h̄c
then the neutron potential pressure Pn in (7.3.21) is greater than the star pressure PM in
(7.3.22):
Pn > PM .
In this case, neutrons in the star cannot be crushed into quarks. Hence, (7.3.24) should be a
criterion to determine if the body is a neutron star. It is known that

g2n ∼ h̄c.

Thus, we take
π
g2n = h̄c, (7.3.25)
4
and (7.3.24) is just the black hole criterion.
If the δ -factor satisfies that
4 g2n 4 g2q
6δ < , (7.3.26)
π h̄c π h̄c
then the neutrons will be crushed to become quarks and gluons. The equality (7.3.25) shows
that the star satisfying (7.3.26) must be a black hole which is composed of quarks and
gluons, and is called quark black hole.
If δ satisfies
4 g2q 4 g2w
6δ < , (7.3.27)
π h̄c π h̄c
then the quarks are crushed into weaktons, and the body is called weakton black hole.
In summary, we infer from (7.3.24), (7.3.26) and (7.3.27) the following conclusions:

 a neutron star if δ < δnc and m > 1.4M ,

a body = a quark black hole if δnc < δ < δqc , (7.3.28)

 c c
a weakton black hole if δq < δ < δw ,
where
2
4 gj
δ jc = for j = n, q, w.
π h̄c
458 Chapter 7 Astrophysics and Cosmology

5. Upper limit of the radius. Weaktons are elementary particles, which cannot be
crushed. Therefore, there is no star with δ -factor greater than δwc . Thus there exists an
upper limit for the radius rc for astronomical bodies with mass m, determined by

2MG
= δwc .
c 2 rc
Namely, the upper limit of the radius rc reads

π mG h̄c
rc = . (7.3.29)
2c2 g2w

Finally we remark that since black holes cannot be compressed, the δ -factor of any
astronomical object cannot be less than one: δ > 1. Therefore, by (7.3.25) and (7.3.26),
there exist no weakton black holes.

7.3.4 Origin of stars and galaxies

The closeness and innateness of black holes provide an excellent explanation for the origin
of planets, stars and galaxies.
In fact, all black holes are inherent. Namely, black holes exist at the very beginning
of the Universe. During the evolution of the Universe, each black hole forms a core and
adsorbs a ball of gases around it. The globes of gases eventually evolve into planets, stars
and galaxy nuclei, according to the radii or masses of the inner cores of black holes. Of
course, it is possible that several black holes can bound together to form a core of a bulk of
gases.
Due to the closedness of black holes, planets, stars and galaxy nuclei are stable, which
cannot be absorbed into the inner cores of black holes and vanish.

1. Jeans theory on the origin of stars and galaxies. In the beginning of the twentieth
century, J. Jeans presented a general theory for the formation of galaxies and stars. He
thought that the Universe in the beginning was filled with chaotic gas, and various astro-
nomical objects were formed in succession by a process of gas decomposition into bulks of
clouds, consequently forming galaxies, stars, and planets.
According to the Jeans theory, a ball of clouds with homogeneous density ρ can be held
together only if
V + K 6 0, (7.3.30)

where V is the total gravitational potential energy, and K is the total kinetic energy of all
particles. The potential energy V is
Z R
GMr 3GM 2
V =− × 4π r2ρ dr = − , (7.3.31)
0 r 5R
7.3 Black Holes 459

where M is the mass of the cloud, Mr = 4π r3 ρ /3, and R is the radius. The kinetic energy K
is expressed as the sum of thermal kinetic energies of all particles:

3
K = NkT,
2
where N is the particle number, T is the temperature, and k is the Boltzmann constant.
Assume that all particles have the same mass m, then N = M/m, and we have

3M
K= kT. (7.3.32)
2m
Thus, by (7.3.30)-(7.3.32) we obtain that

GM 5
> kT. (7.3.33)
R 2m
The inequality (7.3.33) is called the Jeans condition.

2. Masses of astronomical objects. The Jeans condition (7.3.33) guarantees only the
gaseous clouds being held together, and does not imply that the gas clouds can contract to
form an astronomical object. However, a black hole must attract the nebulae around it to
form a compact body.
We consider the mass relation between an astronomical object and its black hole core.
The mass M of the object is
M = Mb + M1 , (7.3.34)

where Mb is the mass of the black hole, and M1 is the mass of the material attracted by this
black hole. The total binding potential energy V of this object is given by
Z R
GMr
V =− × 4π r2ρ dr, (7.3.35)
Rs r

where R is the object radius, Rs is the radius of the black hole, ρ is the mass density outside
the core, and
Z r
Mr = Mb + 4π r2 ρ dr. (7.3.36)
Rs

Since Rs  R, we take Rs = 0 in the integrals (7.3.35) and (7.3.36). We assume that the
density ρ is a constant. Then, it follows from (7.3.35) and (7.3.36) that
Z R   
4π 4 Mb R2 4π
V = −4π Gρ Mb r + ρ r dr = −4π Gρ + ρ R5 .
0 3 2 3×5

4
By ρ = M1 / π R3 and Mb = M − M1 , we have
3
460 Chapter 7 Astrophysics and Cosmology

 
G 3 9
V =− MM1 − M12 . (7.3.37)
R 2 10
The stability of the object requires that −V takes its maximum at some M1 such that
dV /dM1 = 0. Hence we derive from (7.3.37) that
5 1
M1 = M, Mb = M. (7.3.38)
6 6
The relation (7.3.38) means that a black hole with mass Mb can form an astronomical object
with mass M = 6Mb .
3. Relation between radius and temperature. A black hole with mass Mb determines the
mass M of the corresponding astronomical system: M = 6Mb . Then, by the Jeans relation
(7.3.33), the radius R and average temperature T satisfy
2 × 6GmMb
T= (7.3.39)
5kR
where T is expressed as
Z
3
T= τ (x)dx,
4π R3 BR
where BR is the ball of this system, and τ (x) is the temperature distribution. Let τ = τ (r)
depend only on r, then we have
Z R
3
T= r2 τ (r)dr. (7.3.40)
R3 0

4. Solar system. For the Sun, M = 2 × 1030kg and R = 7 × 108m. Hence the mass of the
solar black hole core is about
1
M b = × 1030kg,
3
and the average temperature has an upper limit:
4 6.7 × 10−11m3 /kg · s2 × 1030kg × 1.7 × 10−27kg
T= × ' 108 K.
5 1.4 × 10−24kg · m2 /s2 · K × 7 × 108m
For the earth, M = 6 × 1024kg, R = 6.4 × 106m. Thus we have

Meb = 1024kg, T = 3.3 × 104.

5. The radii of the solar and earth’s black hole cores. The radius of solar black hole is
given by
Rs = 500 m,
and the radius of black hole of the earth is as
3
Res = cm.
2
7.4 Galaxies 461

7.4 Galaxies
7.4.1 Introduction
Galaxy is an elementary unit in the large scale structure of our Universe, which is composed
of star clusters, stars, gases and dusts. In the following we introduce this topic.
1. Levels of cosmical structure. There are mainly five levels of cosmic structure: stars,
star clusters, galaxies, clusters of galaxies, and the Universe, and their basic scales are given
inTable 7.1.

Table 7.1 Levels of Cosmic Structure

star clusters galaxies clusters of
levels Sun Universe
(globular) (spiral) galaxies
radius(pc) 10−8 10 105 106 1010
distance(pc) 1 103 106 108
mass(M ) 1 106 1011 1014 1021
density
103 10−18 10−20 10−23 10−27
(kg/m3 )

2. Galaxy types. Galaxies consist of the following three types:

Elliptical (E), Spiral (S), irregular (Irr).
The spiral galaxies are divided into two different sequences:

normal spirals, denoted by S,

barred spirals, denoted by SB.

Elliptical galaxies have an oval appearance, which are classified into eight grades, de-
noted by E0 , E1 , · · · , E7 , according to their apparent ellipticity e = (a − b)/a, where a is the
major radius and b is the minor radius. The E0 -type of elliptical galaxies are spherical in
shape with e0 = 0, and the E7 -types are of the most ablate appearance, i.e. the apparent
ellipticity ek of Ek is arranged in order

e0 < e 1 < · · · < e 7 .

The S-spirals form a sequence of three types: Sa , Sb , Sc . The Sa have large galaxy nuclei
and tightly wound arms, the Sb have moderate galaxy nuclei and less tightly wound arms,
and the Sc have the smallest nuclear bulges and the least wound arms. Our galaxy (the Milky
Way) is type Sb .
About one third of all spiral galaxies are the SB-type. They also consist of three types:
SBa , SBb , SBc , according to the size of the galaxy nuclei and tightness of the spiral arms,
exactly as in S-spiral sequence.
The galaxy classification given above was introduced by E. Hubble, who arranged the
galaxies in an orderly diagram as shown in Figure 7.5.
462 Chapter 7 Astrophysics and Cosmology

Figure 7.5 Hubble classification of galaxies, the ellipticals arranged on the left, and the spirals ar-
ranged on the right in two parallel sequences.

The SO (S-zero) galaxies have the characteristicsof ellipticals and spirals. They are disk
shaped, like spirals, but lack spiral structure, and therefore like flat ellipticals.
The amount of interstellar gas, also called nebula, in galaxies increases from left to right
in the diagram given by Figure 7.2. In ellipticals the amount of gas is very small, and it is
also small in SO galaxies. However, the nebula increases progressively in the spirals from
Sa and SBa to Sc and SBc . in addition, another important distinction between ellipticals
and spirals is that clusters of galaxies contain 80% elliptics, and outside clusters, 80% of
galaxies are spirals.
3. Galaxy properties. Various properties of galaxies are listed in Tables 7.2-7.4.

Table 7.2 Galaxy distribution

Types E S0 S + SB Irr Undetermined
% 13.0 21.5 61.1 3.1 0.9

Table 7.3 Average physical quantities

Types E Sa Sb Sc Irr
mass (M ) 2.0 × 1011 1.6 × 1011 1.3 × 1011 1.6 × 1010 1 × 109
mass/luminosity
20-70 6.6 3.6 1.4 0.9
(m /L )
density (M /pc3 ) 0.16 0.08 0.025 0.013 0.003
gas mass (%) 6 0.2 1.3 3 20 40

Table 7.4 Properties of ellipticals and spirals

Properties ellipticals spirals
interstellarmatter non plentiful
young stars non yes
appearance elliptical shape disk shape
stellar motion random rotation
color red blue

4. Active galaxies. According to their active extent, galaxies are divided into two types:
normal galaxies and active galaxies. Active galaxies mainly include
7.4 Galaxies 463

Starbursts, Seyferts, Quasars, Radio-galaxies

Active galaxies are usually characterized by their extraordinary energy emision. The
starburst galaxies emit a large amount of infrared radiation, which is caused by a great
number of newly formed stars. But the three types of galaxies: Seyferts, Quasars and
Radio-galaxies are known as active galaxies because they possess a compact region at their
centers that has a much higher radiation than normal luminosity, which is called the active
galactic nucleus (AGN).
AGN generates and emits immense quantities of energy over a wide range of electro-
magnetic wavelengths. Today, it is widely believed that AGN is an accretion disk generated
by a supermassive black hole. Though we have not completely confirmed evidence to show
the existence of black holes, many astronomical observations strongly suggest their exis-
tence. The main features of active galaxies are listed in Table 7.5.

Table 7.5 Features of active galaxies

Galaxy Type Active nuclei strong radiation Jets
Normal no no no
Starburst no yes no
Seyfert yes yes yes
Quasar yes yes yes
Radio galaxy yes yes yes

5. The Milky Way. Our own galaxy is known as the Milky Way, which is of S b type and
has 1011 stars and an enormous quantities of clouds of gases and dusts. Its radius is about
105 ly.
The Milky Way consists of two parts: the disk and the halo. The disk has the radius
of about 5 × 105 ly and the thick of about 5 × 104 ly. The disk is composed of stars and
nebulae, rotating around the center of the galaxy. The Sun is 3 × 10 4 ly from the nucleus, it
moves at 300km/s and has a period of 2 × 108 years.
The halo of the Milky Way is spherically-shaped, centered on the nucleus of the galaxy,
and has the radius of 105 ly. The halo consists of less gas and roughly 120 globular clusters,
each of which has hundreds of thousands of stars and moves in an elliptical orbit around the
nucleus of the galaxy. The halo does not rotate with the disk.

6. Galaxy clusters. Galaxies are not uniformly distributed in the Universe, but aggregate
in clusters of different size. Clusters of galaxies are the largest known astronomical systems
bound by gravitational attraction. They form dense regions in the large scale structure of
our Universe. The clusters are associated with much larger, non-gravitationally bounded
groups, called superclusters.
The great regular clusters of galaxies are spherically-shaped, have roughly thousands
of galaxies, almost all of which are of E and S0 types. The regular clusters of galaxies
464 Chapter 7 Astrophysics and Cosmology

are typically 5 × 106ly in radius, and have no clear outer boundaries. Clusters are locked
as their galaxies held together by mutual gravitational attraction. However, their velocities
are too large, exceeding 103 km/s, for them to remain gravitationally bound by their mutual
attractions. It implies the presence of either an additional invisible mass component, or
an additional attractive force besides the Newtonian gravity. This is the so called dark
matter phenomenon. Astronomical observations manifest that there are large amounts of
intergalactic gas which is very hot between 107 ∼ 108 K. The total mass of the gas is greater
than that of all galaxies in the cluster. The wind of intergalactic gas streaming through these
fast moving galaxies is strong enough to strip away their interstellar gas. This explains
why the E and S0 types of galaxies have less gas because their gas has been swept out by
intergalactic winds.
The other galaxy clusters are irregular. They have various sizes ranged from thousands
of members to a few tens of members. The smaller clusters are also called galaxy groups.
For example, our galaxy is a member of a galaxy group known as the local group which
possesses about 40 galaxies. The irregular clusters lack spherical symmetry in shape and
contain a mixture of all types of galaxies.

7.4.2 Galaxy dynamics

Galaxies are mainly either spiral or elliptical. Each galaxy possesses a compact core, known
as galactic nucleus, which is supermassive and spherical-shaped. Thus, the galactic dynamic
model is defined in an annular domain:
r0 < r < r 1 ,

where r0 is the radius of galaxy nucleus and r1 the galaxy radius. In the following we
develop models for spiral and elliptical galaxies, and provide their basic consequences on
galactic dynamics.
1. Spiral galaxies. Spiral galaxies are disc-shaped, as shown in Figure 7.6. We model
the galaxy in a disc domain as
D = {x ∈ R2 | r0 < |x| < r1 }, (7.4.1)

Figure 7.6 A schematic diagram of spiral galaxy.

7.4 Galaxies 465

for which the spherical coordinates reduce to the polar coordinate system (ϕ , r):
π 
(θ , ϕ , r) = ,ϕ,r for 0 6 ϕ 6 2π , r0 < r < r1 . (7.4.2)
2
The metric satisfying the gravitational field equations (7.1.62) of the galaxy nucleus is the
Schwarzschild solution:
 
2 M0 G
g00 = − 1 + 2 ψ , ψ =− ,
c r
  (7.4.3)
δ r0 −1 2M0 G
g11 = α (r) = 1 − , δ= 2 ,
r c r0
where r0 < r < r1 and M0 is the mass of galactic nucleus.
With (7.4.2) and (7.4.3), the 2D fluid equations (7.1.65)-(7.1.68) are written as
∂ Pϕ 1 1 ∂p
+ (P · ∇)Pϕ = ν ∆Pϕ − ,
∂τ ρ r ∂ϕ
∂ Pr 1 1 ∂p Mr G
+ (P · ∇)Pr = ν ∆Pr − − ρ (1 − β T) 2 ,
∂τ ρ α ∂r αr
(7.4.4)
∂T 1
+ (P · ∇)T = κ ∆T
e + Q,
∂τ ρ
∂ρ
+ divP = 0,
∂τ
supplemented with boundary conditions:
Pϕ (r0 ) = ζ0 , Pr (r0 ) = 0, T (r0 ) = T0 ,
(7.4.5)
Pϕ (r1 ) = ζ1 , Pr (r1 ) = 0, T (r1 ) = T1 .
Here α is as in (7.4.3), and Mr is the total mass in the ball Br .
2. Elliptical galaxies. Elliptical galaxies are spherically-shaped, defined in a spherical-
annular domain, as shown in Figure 7.7:

Ω = {x ∈ R3 | r0 < |x| < r1 } (7.4.6)

Figure 7.7 A schematic diagram of elliptical galaxy.

466 Chapter 7 Astrophysics and Cosmology

The metric is as in (7.4.3), and the corresponding fluid equations (7.1.65)-(7.1.68) are
in the form:
∂P 1 M0 G
+ (P · ∇)P = ν ∆P − ∇p − ρ (1 − β T) 2 ~k,
∂τ ρ αr
∂T 1
+ (P · ∇)T = κ ∆T e + Q, (7.4.7)
∂τ ρ
∂ρ
+ div P = 0,
∂τ
supplemented with the physically sound conditions:
∂ Pθ ∂ Pϕ
Pr = 0, = 0, = 0 at r = r0 , r1 ,
∂r ∂r (7.4.8)
T (r0 ) = T0 , T (r1 ) = T1 .

3. Galaxy dynamics. Based on both models (7.4.4)-(7.4.5) and (7.4.7)-(7.4.8), we out-

line below the large scale dynamics of both spiral and elliptical galaxies.
Let the models be abstractly written in the following form
du
= F(u, ρ ), (7.4.9)
dt
where u = (P, T, p) is the unknown function, and ρ is the initial density distribution, which
is used as a control parameter representing different physical conditions.
First, we consider the stationary equation of (7.4.9) given by
F(u, ρ ) = 0. (7.4.10)
Let u0 be a solution of (7.4.10), and consider the deviation from u 0 as
u = v + u0 .
Thus, (7.4.9) becomes the following form
dv
= Lλ v + G(v, λ , ρ ), (7.4.11)
dt
where λ = (δ , Re), and the δ -factor and the Rayleigh number are defined by
2M0 G M0 Gr0 r1 β T0 − T1
δ= , Re = . (7.4.12)
c 2 r0 κν r1 − r 0
The Lλ is the derivative operator (i.e. the linearized operator) of F(u, p) at u 0 :
Lλ = DF(u0 , ρ ),
and G is the higher order operator.
Then, we consider the dynamic transition of (7.4.11). Let veλ be a stable transition
solution of (7.4.11). Then the function
ue = u0 + veλ (7.4.13)
provides the physical information of the galaxy.
7.4 Galaxies 467

7.4.3 Spiral galaxies

Spiral galaxies are divided into two types: normal spirals (S-type) and barred spirals (SB-
type). We are now ready to discuss these two sequences of galaxies by using the spiral
galaxy model (7.4.4)-(7.4.5).
Let the stationary solutions of (7.4.4)-(7.4.5) be independent of ϕ , given by

Pr = 0, Pϕ = Peϕ (r), T = Te(r), p = pe(r).

The heat source is approximatively taken as Q = 0. Then the stationary equations of (7.4.4)-
(7.4.5) are !
00 0 1 δ r 0 00 3 0
eϕ
P
rPeϕ + 2Peϕ − Peϕ − rPeϕ + Peϕ − = 0,
r r 2 2r
∂ pe 1 e2 1
= P − ρ (1 − β T)M
e r G,
∂r rρ ϕ r2 (7.4.14)
!
d e
2 dT
r = 0,
dr dr
Peϕ (r0 ) = ζ0 , Peϕ (r1 ) = ζ1 , Te(r0 ) = T0 , Te(r1 ) = T1 .
The first equation of (7.4.14) is an elliptic boundary value problem, which has a unique
solution Pϕ . Since δ r0 /r is small in the domain (7.4.1), the first equation of (7.4.14) can be
approximated by
2 1
Peϕ00 + Pϕ0 − 2 Peϕ = 0,
r r
which has an analytic solution as
√ √
5−1 5+1
Peϕ = β1 rk1 + β2rk2 , k1 = , k2 = − . (7.4.15)
2 2
By the boundary conditions in (7.4.14), we obtain that

r0k2 ζ1 − r1k2 ζ0 r1k1 ζ0 − r0k1 ζ1

β1 = , β2 = . (7.4.16)
r1k1 r0k2 − r0k1 r1k2 r1k1 r0k2 − r0k1 r1k2

Thus we derive the solution of (7.4.14) as

Z
" #
T0 − T1 r0 Peϕ2 ρ G
Peϕ , Te = T0 + r1 ( − 1), pe = − 2 (1 − β Te)Mr dr.
r1 − r 0 r rρ r

Make the translation

Pr → Pr , Pϕ → Pϕ + Peϕ , T → T + Te, p → p + pe;

468 Chapter 7 Astrophysics and Cosmology

then the equations (7.4.4) and boundary conditions (7.4.5) become

∂ Pϕ 1 Peϕ d Peϕ 1 ∂p
+ (P · ∇)Pϕ = ν ∆Pϕ − ( + )Pr − ,
∂τ ρ r dr r ∂ϕ
∂ Pr 1 2Peϕ ρβ Mr G 1 ∂p
+ (P · ∇)Pr = ν ∆Pr + Pϕ + T− ,
∂τ ρ αr αr 2 α ∂r
(7.4.17)
∂T 1 e + r0 r1 γ Pr − 1 Peϕ ∂ T ,
+ (P · ∇)T = κ ∆T
∂τ ρ ρ r2 ρr ∂ ϕ
div P = 0,
P = 0, T =0 at r = r0 , r1 ,

where r = (T0 − T1 )/(r1 − r0 ).

The eigenvalue equations of (7.4.17) are given by

1 Peϕ d Peϕ 1 ∂p
− ∆Pϕ + ( + )Pr + = λ Pϕ ,
ν r dr rν ∂ ϕ
2Peϕ ρβ Mr G 1 ∂p
− ∆Pr − Pϕ − T+ = λ Pr ,
αν r αν r2 αν ∂ r
(7.4.18)
1 ∂ T r0 r1 γ
− ∆T + Peϕ − Pr = λ T,
ρ r ∂ ϕ κρ r2
div P = 0,
P = 0, T =0 at r = r0 , r1 .

The eigenvalues λ of (7.4.18) are discrete (not counting multiplicity):

λ1 > λ 2 > · · · > λ k > · · · , λk → −∞ as k → ∞.

The first eigenvalue λ1 and first eigenfunctions

Φ = (Pϕ0 , Pr0 , T 0 ) (7.4.19)

dictate the dynamic behaviors of spiral galaxies, which are determined by the physical pa-
rameters:
Z r1
T0 − T1 2M0 G
ζ 0 , ζ 1 , r0 , r1 , κ , ν , β , γ = , δ = 2 , Mr = M0 + 4π r2 ρ dr. (7.4.20)
r1 − r 0 c r0 r0

Based on the dynamic transition theory in (Ma and Wang, 2013b), we have the following
physical conclusions:

• If the parameters in (7.4.20) make the first eigenvalue λ1 < 0, then the spiral galaxy
is of S0-type.
7.4 Galaxies 469

• If λ1 > 0, then the galaxy is one of the types Sa , Sb , Sc , SBa , SBb , SBc , depending on
the structure of (Pϕ0 , Pr0 ) in (7.4.19).

• Let λ1 > 0 and the first eigenvector (Pϕ0 , Pr0 ) of (7.4.19) have the vortex structure as
shown in Figure 7.8. Then the number of spiral arms of the galaxy is k/2, where k is
the vortex number of (Pϕ0 , Pr0 ). Hence, if k = 2, the galaxy is of the SBc -type.

Figure 7.8 The vortex structure of the first eigenvector (Pϕ0 , Pr0 ).

The reason behind the number of spiral arms being k/2 is as follows. First the number of
vortices in Figure 7.8 is even, and each pair of vortices have reversed orientations. Second, if
the orientation of a vortex matches that of the stationary solution Pϕ (r) of (7.4.14), then the
superposition of Pϕ (r) and Pϕ0 of (7.4.19) gives rise to an arm; otherwise, the counteraction
of Pϕ (r) and Pϕ0 with reversed orientations reduces the energy momentum density, and the
region becomes nearly void.
Remark 7.17 There are three terms in (7.4.18), which may generate the transition of
(7.4.17):
 
k1 T k2 Pr ρβ Mr G r0 r1 γ
F1 = 0, − 2 , − 2 , k2 = , k2 = ,
r r αν κρ
! !
1 Peϕ d Peϕ 2Peϕ
F2 = + Pr , − Pϕ , 0 ,
ν r dr αν r
 
1 dα ∂ 1 ∂ 1 dα
F3 = − 2 (rPϕ ), ( Pr ), 0 .
2α r dr ∂ r 2α ∂ r α dr
The term F1 corresponds to the Rayleigh-Bénard convection with the Rayleigh number
β Mr Gr0 r1 γ
R = k 1 k2 = ,
ανκ
the term F2 corresponds to the Taylor rotation which causes the instability of the basic flow
(Pϕ , Pr ) = (Peϕ , 0), and F3 is the relativistic effect which only plays a role in the case where
δ ' 1.
470 Chapter 7 Astrophysics and Cosmology

7.4.4 Active galactic nuclei (AGN) and jets

The black hole core of a galaxy attracts a large amounts of gases around it, forming a galactic
nucleus. The mass of a galactic nucleus is usually in the range
105 M ∼ 109 M . (7.4.21)
Galactic nuclei are divided into two types: normal and active. In particular, an active galactic
nucleus emits huge quantities of energy, called jets. We focus in this section the mechanism
of AGN jets.
1. Model for AGN. The domain of an galactic nucleus is a spherical annulus:

B = x ∈ R3 | Rs < |x| < R1 , (7.4.22)
where Rs is the Schwarzschild radius of the black hole core, and R1 is the radius of the
galaxy nucleus.
The model governing the galaxy nucleus is given by (7.4.7)-(7.4.8), defined in the do-
main (7.4.22) with boundary conditions:
∂ Pθ
Pr = 0, = 0, Pϕ = P0 , T = T0 for r = Rs ,
∂r (7.4.23)
∂ Pθ
Pr = 0, = 0, Pϕ = P1 , T = T1 for r = R1 .
∂r
Let the stationary solution of the model be as
Pθ = 0, Pr = 0, Pϕ = Pϕ (r, θ ),
and p, ρ , T be independent of ϕ . Then the stationary equations for the four unknown func-
tions Pϕ , T, p, ρ are in the form
∂p 1 cos θ 2
= P ,
∂θ ρ sin θ ϕ
∂p 1 2 Mb G
= Pϕ − ρ (1 − β T) 2 ,
∂r ρr r
(7.4.24)
Pϕ 1 d α ∂
− ν ∆P
e ϕ+
2 sin θ
+ 2 (rPϕ ) = 0,
r
  2α r dr ∂r
κ ∂ ∂
− 2 r2 T = Q(r),
αr ∂ r ∂r
where Mb is the mass of the black hole core, Q is the heat source generated by the nuclear
burning, and
 
1 ∂ ∂ 1 ∂ ∂
∆
e= sin θ + 2 (r2 ),
r2 sin θ ∂ θ ∂θ αr ∂ r ∂ r
 
2Mb G −1
α = 1− 2 .
c r
7.4 Galaxies 471

The boundary conditions of (7.4.23) become

Pϕ (Rs ) = Rs Ω0 , Pϕ (R1 ) = R1 Ω1 , T (Rs ) = T0 , T (R1 ) = T1 , (7.4.25)

where Ω0 , Ω1 only depend on θ , T0 , T1 are constants.

Make the translation

Pr → Pr , Pθ → Pθ , Pϕ → Pϕ + Peϕ , T → T + Te, p → p + pe,

where (Peϕ , Te, pe, ρ ) is the solution of (7.4.24)-(7.4.25). Then the equations (7.4.7) are rewrit-
ten as

∂ Pθ 1 Peϕ ∂ Pθ 2 cos θ Peϕ 1 ∂p

+ (P · ∇)Pθ = ν ∆Pθ − + Pϕ − ,
∂τ ρ ρ r sin θ ∂ ϕ ρ r sin θ r ∂θ
∂ Pϕ 1 1 ∂ Peϕ Peϕ ∂ Pϕ 1 ∂ Peϕ
+ (P · ∇)Pϕ = ν ∆Pϕ − Pθ − − Pr
∂τ ρ ρr ∂ θ ρ r sin θ ∂ ϕ ρ ∂r
Peϕ cos θ Peϕ 1 ∂p
− Pr − Pθ − ,
ρr ρ r sin θ r sin θ ∂ ϕ (7.4.26)
∂ Pr 1 Peϕ ∂ Pr 2Peϕ 1 ∂p Mb G
+ (P · ∇)Pr = ν ∆Pr − + Pϕ − +βρ T,
∂τ ρ ρ r sin θ ∂ ϕ ρα r α ∂r α r2
∂T 1 e − Pϕ ∂ T − 1 d T Pr ,
e e
+ (P · ∇)T = κ ∆T
∂τ ρ ρ r sin θ ∂ ϕ ρ dr
divP = 0,

with the boundary conditions

∂ Pθ
Pr = 0, Pϕ = 0, = 0, T =0 at r = Rs , R1 . (7.4.27)
∂r

2. Taylor instability. By the conservation of angular momentum and R 1  Rs , the

angular momentums Ω0 and Ω1 in (7.4.25) satisfy that

Ω0  Ω 1 , (7.4.28)

This property leads to the instability of the rotating flow represented by the stationary solu-
tion:
(Pr , Pθ , Pϕ ) = (0, 0, Peϕ ), (7.4.29)

which is similar to the Taylor-Couette flow in a rotating cylinder. The rotating instability can
generate a circulation in the galactic nucleus, as the Taylor vortices in a rotating cylinder,
472 Chapter 7 Astrophysics and Cosmology

as shown in Figure 7.9. The instability is caused by the force F = (Fr , Fθ , Fϕ , T ) in the
equations of (7.4.26) given by

2Peϕ Peϕ ∂ Pr
Fr = Pϕ − ,
ρα r ρ r sin θ ∂ ϕ
2 cos θ Peϕ Peϕ ∂ Pθ
Fθ = Pϕ − ,
ρ r sin θ ρ r sin θ ∂ ϕ
! ! (7.4.30)
1 Peϕ ∂ Peϕ 1 cos θ e ∂ Peϕ Peϕ ∂ Pϕ
Fϕ = − + Pr − Pϕ + Pθ − ,
ρ r ∂r ρ r sin θ ∂θ ρ r sin θ ∂ ϕ

Peϕ ∂ T
T =− .
ρ r sin θ ∂ θ

3. Rayleigh-Bénard instability. Due to the nuclear reaction (fusion and fission) and the
large pressure gradient, the galactic nucleus possesses a very large temperature gradient in
(7.4.25) as
DT = T0 − T1 , (7.4.31)
which yields the following thermal expansion force in (7.4.26), and gives rise to the Rayleigh-
Bénard convection:
Mb G 1 d Te
Fr = β ρ T, T= Pr . (7.4.32)
αr 2 ρ dr

4. Instability due to the gravitational effects. Similar to (7.2.7), there is a radial force in
the term ν ∆ur of the third equation of (7.4.26):
 
ν ∂ 1 dα
Fr = Pr , (7.4.33)
2α ∂ r α dr

where
α = (1 − Rs/r)−1 , Rs < r < R 1 . (7.4.34)
In (7.4.33) and (7.4.34), we see the term

ν R2s
fr = Pr , (7.4.35)
1 − Rs/r r4
which has the property that
(
+∞ for Pr > 0 at r = Rs ,
fr = (7.4.36)
−∞ for Pr < 0 at r = Rs .

It is the force (7.4.36) that not only causes the instability of the basic flow (7.4.29), but also
generates jets of the galaxy nucleus.
7.4 Galaxies 473

5. Latitudinal circulation. The above three types of forces: the rotating force (7.4.30),
the thermal expansion force (7.4.32), and the gravitational effect (7.4.35), cause the insta-
bility of the basic flow (7.4.29) and lead to the latitudinal circulation of the galactic nucleus,
as shown in Figure 7.9.

Figure 7.9 The latitudinal circulation with k = 2 cells.

6. Jets and accretions. Each circulation cell has an exit as shown in Figure 7.10, where
the circulating gas is pushed up by the radial force (7.4.35)-(7.4.36), and erupts leading to
a jet. The cell has an entrance as shown in Figure 7.10, where the exterior gas is pulled
into the nucleus, is cyclo-accelerated by the force (7.4.35), goes down to the inner boundary
r = Rs , and then is pushed by Fθ of (7.4.30) toward to the exit. Thus the circulation cells
form jets in their exits and accretions in their entrances. In Figure 7.11(a), we see that there
is a jet in the latitudinal circulation with k = 1 cell, and in Figure 7.11(b) there are two jets
in the circulation with k = 2 cells in its south and north poles, and an accretion disk near its
equator.

Figure 7.10
474 Chapter 7 Astrophysics and Cosmology

Figure 7.11 (a) A jet in the latitudinal circulation with k = 1 cell, two jets in the latitudinal circulation
with k = 2 cells.

7. Condition for jet generation. The main power to generate jets comes from the gravi-
tational effect of (7.4.35)-(7.4.36) by the black hole. The radial momentum Pr in (7.4.35) is
the bifurcated solution of (7.4.26), which can be expressed as

Pr = R2s Qr ,

where Qr is independent of Rs . Thus, the radial force (7.4.35) near r = Rs is approximatively

written as
ν
fr = QR , r = R s + e r for 0 < e
r  Rs . (7.4.37)
1 − Rs/r s
Let fE be the lower limit of the effective force, which is defined as that the total radial
force Fr in the third equation of (7.4.26) is positive provided f r > fE :

Fr > 0 if fr > fE .

Let RE be the effective distance:

fr > fE if Rs < r < R s + R E .

Then, it follows from (7.4.37) that

RE = kRs (k = ν QRs / fE ). (7.4.38)

It is clear that there is a critical distance Rc such that

a jet forms if RE > Rc or Rs > k−1 Rc ,

(7.4.39)
no jet forms if RE < Rc or Rs < k−1 Rc .

The criterion (7.4.39) is the condition for jet generation.

7.5 The Universe 475

The condition (7.4.39) can be equivalently rewritten as that there is a critical mass Mc
such that the galactic nucleus is active if its mass M is bigger than Mc , i.e. M > Mc . By
(7.4.21), we have
105 M < Mc or 106 M < Mc .

Remark 7.18 The jets shown in Figures 7.9 and 7.10 are column-shaped. If the cell
number k > 3 for the latitudinal circulation of galaxy nucleus, then there are jets which are
disc-shaped. We don’t know if there exist such galaxy nuclei which have the disc-shaped
jets in the Universe. Theoretically, it appears to be possible.

Remark 7.19 Galactic nucleus are made up of plasm. The precise description of AGN
jets requires to take into consideration of the magnetic effect in the modeling. However the
essential mechanism does not change and an explosive magnetic energy as in (7.4.37) will
contribute to the supernovae explosion.

7.5 The Universe

7.5.1 Classical theory of the Universe
In this section, we recall some basic aspects of modern cosmology, including the Hubble
Law, the expanding universe, and the origin of our Universe, together with their experimen-
tal justifications.

1. The Hubble Law. In 1929, American astronomer Edwin Hubble discovered an ap-
proximatively linear relation between the recession velocity v and the distance R of remote
galaxies, which is now called the Hubble Law:

v = HR, (7.5.1)

where H is called the Hubble constant, depends on time, and its present-time value is

H = 70 km/s · Mpc, Mpc = 106pc (1 pc = 3.26 ly). (7.5.2)

Formula (7.5.1) implies that the farther away the galaxy is from our galaxy, the greater
its velocity is.

2. Expansion of the Universe. An important physical conclusion from the Hubble Law
(7.5.1)-(7.5.2) is that our Universe is expanding.
If we regard our Universe as a 3-dimensional sphere:

M = Sr3 = {x ∈ R4 | x21 + x22 + x23 + x24 = r2 }. (7.5.3)

Each point on the sphere Sr3 can be regarded as a center, as the radius r increases, all points
on the sphere are moving away from the point. Moreover, the farther away a remote object
476 Chapter 7 Astrophysics and Cosmology

is from the point, the faster the object appears to be moving. For example, for any two
points p1 and p2 on a sphere Sr3 , θ is the angle between p1 and p2 ; see Figure 7.12. As the
radius r varies from r1 to r2 , the distance R between P1 and P2 varies from R1 to R2 :

R1 = θ r 1 → R 2 = θ r 2 .

Figure 7.12

The velocity of separation of these two points is

dR
v= = θ ṙ, (R = θ r). (7.5.4)
dt
Assume that ṙ is proportional to r, i.e.

ṙ = Hr, (H is a constant). (7.5.5)

Then the Hubble law (7.5.1) follows from (7.5.4) and (7.5.5).
The Hubble Law (7.5.1) also manifests that the recession velocity is independent of
the direction. Hence, it implies that the Universe is isotropic. This very property gives
rise to a strong restriction on the global structure of the Universe: there are only two types
of geometry for our Universe: either the 3D sphere (7.5.3) or the 3D Euclidian space R 3 .
Namely, we have reached the following physical conclusion.

Physical Conclusion 7.20 The Hubble law allows only two types of topological struc-
ture, either S3 or R3 , as our Universe. Further, the Hubble Law requires that the Universe
is expanding.

3. Big-Bang theory. In 1927, the Belgium cosmologist G. Lemaı̂tre first proposed that
the Universe begins with a big explosion, known as the Big-Bang theory. After the Hubble
law was discovered, in 1931, G. Lemaı̂tre regarded the recessions of remote galaxies as the
7.5 The Universe 477

expanding of the Universe, and he thought that as the time inverses, the early universe must
be in a high temperature and dense state, and the stage is considered as the beginning of the
Big-Bang.
The Big-Bang theory appears to be consistent with the three most important astronom-
ical facts: the Hubble law, the cosmic microwave background, discovered in 1965 by two
physicists A. Penzias and R. Wilson, and the abundance of helium.

4. Age of the Universe. Based on the Big-Bang theory, the age of the Universe is finite.
Let T be the age, then the present-day distance between any two points is

R = vT.

It follows from (7.5.1) that

1
T= , (7.5.6)
H
where H is the Hubble constant. We infer then from (7.5.2) and (7.5.6) the age T of the
Universe as
T = 1.4 × 1010 year. (7.5.7)

As we approximatively regard the recession velocity v of remote galaxies as the speed

of light, i.e. v ' c, the age (7.5.7) is also considered as the radius of the Universe.

5. Structure of universes. A central topic in modern cosmology is to investigate both

the global topological and geometrical structure of the Universe. A fundamental principle
of cosmology, called cosmological principle, states that ignoring local irregularities, the
Universe is homogeneous and isotropic.
The cosmological principle is compatible with the Hubble law: Physical Conclusion
7.20, and can be equivalently stated in the following form.

Theorem 7.21 (Topological Structure of the Universe) The global topological struc-
ture of the Universe is either S3 or R3 .

6. Global geometrical structure. Based on the Cosmological Principle 7.21, three physi-
cists A. Friedmann (1922), G. Lemaı̂tre (1927), H. P. Robertson (1935), and a mathemati-
cian A. G. Walker (1936), independently derived the globally geometrical structure of the
Universe.

Theorem 7.22 (Geometrical Structure of the Universe) The Riemannian metric of 4D

space-time manifold satisfying the Cosmological Principle 7.21 takes the following form
 
dr2
ds2 = −c2 dt 2 + R(t) + r 2
(d θ 2
+ sin2
θ d ϕ 2
) , (7.5.8)
1 − kr2
478 Chapter 7 Astrophysics and Cosmology

where R(t) is the scalar factor representing the radius of the Universe, which depends only
on time t, and k = 1, 0, −1 stand for the sign of space scalar curvature in the following
sense:
1
k=1: M = S3 the 3D sphere with scalar curvature = 2 ,
R
k=0: M = R3 the 3D flat Euclidean space, (7.5.9)
1
k = −1 : M = L3 the Lobachevsky space with scalar curvature = − 2 ,
R
where the Lobachevsky space L3 has the same topological structure as R3 .

Historically, Theorem 7.22 was rigorously proved by Robertson and Walker, and was
assumed by Friedmann and Lemaı̂tre to derive the field equations satisfied by R(t).
7. The Newton cosmology. The Newton cosmology is based on the Newton Gravitational
Law. By Cosmological Principle (Roos, 2003), the universe is spherically symmetric. For
any reference point p ∈ M , the motion equation of an object with distance r from p is

d2r GM(r)
=− 2 , (7.5.10)
dt 2 r
where M(r) = 4π r 3 ρ /3, and ρ is the mass density. Thus, (7.5.10) can be rewritten as
follows
4
r00 = − π Grρ . (7.5.11)
3
Make the nondimensional
r = R(t)r0 ,
where R(t) is the scalar factor, which is the same as in the FLRW metric (Ma and Wang,
2014e). Let ρ0 be the density at R = 1. Then we have

ρ = ρ0 /R3 . (7.5.12)

Thus, equation (7.5.11) is expressed as

4π G ρ0
R00 = − , (7.5.13)
3 R2
which is the dynamic equation of Newtonian cosmology.
Multiplying both sides of (7.5.13) by R0 we have
 
d 2 8π G ρ0
Ṙ − = 0.
dt 3 R
Hence, (7.5.13) is equivalent to the equation
8π G ρ0
Ṙ2 = − κ, (7.5.14)
3 R
7.5 The Universe 479

where κ is a constant, and we shall see that κ = kc2 , and k = −1, 0, or 1.

8. The Friedmann cosmology. The nonzero components of the Friedmann metric are
R2
g00 = −1, g11 = , g22 = R2 r2 , g33 = R2 r2 sin2 θ .
1 − kr2
Again by the Cosmological Principle (Roos, 2003), the energy-momentum tensor of the
Universe is in the form
 2 
ρc 0 0 0
 0 g11 p 0 0 
Tµν = 
 0
.
0 g22 p 0 
0 0 0 g33 p
By the Einstein gravitational field equations
8π G 1
Rµν = − (Tµν − gµν T ),
c4 2
Dµ Tµν = 0,

we derive three independent equations

 
4π G 3p
R̈ = − ρ + 2 R, (7.5.15)
3 c
 p
RR̈ + 2Ṙ + 2kc = 4π G ρ − 2 R2 ,
2 2
(7.5.16)
c
 
Ṙ p
ρ̇ = −3 ρ+ 2 , (7.5.17)
R c
where R, ρ , p are the unknown functions.
Equations (7.5.15)-(7.5.17) are called the Friedmann cosmological model, from which
we can derive the Newtonian cosmology equations (7.5.14). To see this, by (7.5.15) and
(7.5.16), we have
 2
Ṙ kc2 8π G
=− 2 + ρ. (7.5.18)
R R 3
By the approximate p/c2 ' 0, (7.5.12) follows from (7.5.17). Then we deduce (7.5.14)
from (7.5.18) and (7.5.12).
From the equation (7.5.18), the density ρc corresponding to the case k = 0 is
 2
3 Ṙ 3
ρc = = H 2, (7.5.19)
8π G R 8π G
where H = Ṙ/R is the Hubble constant, and by (7.5.2) we have

ρc = 10−26 kg/m3 . (7.5.20)

Thus, by the Friedmann model we can deduce the following conclusions.

480 Chapter 7 Astrophysics and Cosmology

Conclusions of Friedmann Cosmology 7.23

1) By (7.5.18) we can see that
ρ > ρc ⇔ k=1 the Universe is closed : M = S3 ,
ρ = ρc ⇔ k=0 the Universe is open : M = R3 , (7.5.21)
ρ < ρc ⇔ k = −1 the Universe is open : M = L3 .

2) Let E0 be the total kinetic energy of the Universe, M is the mass, then we have


 3 GM 2
 for k = 0,
E0 = 5 R 2 (7.5.22)
 2 GM − 1 Mc2 for k = 1,


3π R 2
where the first term represents the total gravitational bound potential energy, and the second
term is the energy resisting curvature tensor.
3) By (7.5.18), Ṙ 6≡ 0, and consequently the universe is dynamic.
4) By (7.5.15), R̈ < 0, the dynamic universe is decelerating.

9. The Lemaı̂tre cosmology. Consider the Einstein gravitational field equations with
the cosmological constant Λ term:
8π G 1
Rµν = − 4
(Tµν − gµν T ) + Λgµν , Λ > 0, (7.5.23)
c 2
then the metric (7.5.8) satisfies the following equations
 
4π G 3p Λc2
R̈ = − ρ + 2 R+ R, (7.5.24)
3 c 3
 2
Ṙ 8π G Λc2 kc2
= ρ+ − 2, (7.5.25)
R 3 3 R
  
Ṙ p
ρ̇ + 3 ρ + 2 = 0. (7.5.26)
R c
The equations (7.5.24)-(7.5.26) are known as the Lemaı̂tre cosmological model, or the Λ-
cosmological model, which leads to the following conclusions of Λ-cosmology:
Conclusions of Λ-Cosmology 7.24
1) The Universe is temporally open: R → ∞ as t → ∞;
2) There is a critical radius Rc , such that
the universe is decelerating for R < Rc ,
the universe is accelerating for R > Rc ,
 
4π Gρ0 1/3
where Rc ' ;
Λc2
7.5 The Universe 481

3) As t → ∞ and ρ → 0, then we deduce from (7.5.25) that the cosmological radius R

has the asymptotic relation
√ 2
R ∼ e Λc /3t as t → ∞. (7.5.27)

Namely √
Λc2 /3t
R(t)/e = const. as t → ∞;

4) The total kinetic energy E is given by



 3 GM 2 Λ 2 2
 5 R + 10 Mc R
 for k = 0,
E= (7.5.28)

 2 GM 2 Λ 2 2 1 2

 + Mc R − Mc for k = 1.
3π R 6 2
Remark 7.25 The field equations (7.5.23) with a cosmological constant Λ lead to a
special conclusion that in the expansion process, there are a large quantities of energy to be
created, and the added energy in (7.5.28) is generated by Λ is as
1 2 2 Λ
Mc ΛR (k = 1) and Mc2 R2 (k = 0).
6 10
It implies that the total energy is not conserved in the Λ-model.

7.5.2 Globular universe with boundary

If the spatial geometry of a universe is open, then by our theory of black holes developed
in Section 7.3, we have shown that the universe must be in a ball of a black hole with
a fixed radius. In fact, according to the basic cosmological principle that the universe is
homogeneous and isotropic (Roos, 2003), given the energy density ρ 0 > 0 of the universe,
by Theorem 7.3, the universe will always be bounded in a black hole of open ball with the
Schwarzschild radius: s
3c2
Rs = ,
8π Gρ0
as the mass in the ball BRs is given by MRs = 4π R3s ρ0 /3. This argument also clearly shows
that
there is no unbounded universe.
In addition, since a black hole is unable to expand and shrink, by property (7.5.55) of
black holes, all globular universes must be static.
Globular universe
We have shown that the universe is bounded, and suppose that the universe is open, i.e.
its topological structure is homeomorphic to R3 , and it begins with a ball. Let E be its total
energy:
E = mass + kinetic + thermal+ Ψ, (7.5.29)
482 Chapter 7 Astrophysics and Cosmology

where Ψ is the energy of all interaction fields. Let

M = E/c2 . (7.5.30)

At the initial stage, all energy is concentrated in a ball with radius R0 . By the theory of
black holes, the energy contained in the ball generates a black hole in R 3 with radius

2MG
Rs = , (7.5.31)
c2
provided Rs > R0 ; see (7.3.1) and Figure 7.3.
Thus, if the universe is born to a ball, then it is immediately trapped in its own black
hole with the Schwarzschild radius Rs of (7.5.31). The 4D metric inside the black hole of
the static universe is given by

ds2 = −ψ (r)c2 dt 2 + α (r)dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ), (7.5.32)

where ψ and α satisfy the equations (7.1.79) and (7.1.80) with boundary conditions:

ψ →ψ
e, α →α
e as r → Rs ,

where Rs is given by (7.5.31). Also, ψ

e, α
e are given via the TOV metric (7.1.40)-(7.1.41):
   −1
1 r2 r2
ψ
e= 1− 2 , α
e = 1− 2 for 0 6 r < Rs . (7.5.33)
4 Rs Rs

Basic problems

A static universe is confined in a ball with fixed radius Rs in (7.5.31), and the ball
behaves like a black hole. We need to examine a few basic problems for a static universe,
including the cosmic edge, the flatness, the horizon, the redshift, and the cosmic microwave
background (CMB) radiation problems.

1. The cosmic edge problem. In the ancient Greece, the cosmic-edge riddle was pro-
posed by the philosopher Archytas, a friend of Plato, who used “what happens when a
spear is thrown across the outer boundary of the Universe?” The problem appears to be very
difficult to answer. Hence, for a long time physicists always believe that the Universe is
boundless.
Our theory of black holes presented in Section 7.3 shows that all objects in a globular
universe cannot reach its boundary r = Rs . In particular, an observer in any position of the
globular universe looking toward to the boundary will see no boundary due to the openness
of the ball and the relativistic effect near the Schwarzschild surface. Hence the cosmic-edge
riddle is no longer a problem.
7.5 The Universe 483

2. The flatness problem. In modern cosmology, the flatness problem means that k = 0
in the FLRW metric (7.5.8)-(7.5.9). It is common to think that the flatness of the universe
is equivalent to the fact that the present energy density ρ must be equal to the critical value
given by (7.5.20). In fact, mathematically the flatness means that any geodesic triangle has
the inner angular sum π = 180◦ .
Measurements by the WMAP (Wilson Microwave Anisotropy Probe) spacecraft in the
last ten years indicated that the Universe is nearly flat. The present radius of the Universe is
about
R = 1026m. (7.5.34)
If the Universe is static, then (7.5.34) gives the Schwarzschild radius (7.5.31), from which
it follows that the density ρ of our Universe is just the critical density of (7.5.20):

ρ = ρc = 10−26 kg/m3 . (7.5.35)

Thus, we deduce that if the universe is globular, then it is static. In addition, we have shown
that any universe is bounded and confined in a 3D hemisphere of a black hole or in a 3D
sphere as shown in Figure 7.14. Hence as the radius is sufficiently large, the universe is
nearly flat.
3. The horizon problem. The cosmic horizon problem can be simply stated as that all
places in a universe look as the same. It seems as if the static Universe with boundary
violates the horizon problem. However, due to the gravitational lensing effect, the light
bents around a massive object. Hence, the boundary of a globular universe is like a concave
spherical mirror, and all lights reaching close to it will be reflected back, as shown in Figure
7.13. It is this lensing effect that makes the globular universe looks as if everywhere is the
same, and is horizontal. In Figure 7.13, if we are in position x, then we can also see a star
as if it is in position ye, which is actually a virtual image of the star at y.

Figure 7.13 Due to the lensing effect, one at x can also see the star at y as if it is another star at ye.

4. The redshift problem. Observations show that light coming from a remote galaxy is
redshifted, and the farther away the galaxy is, the larger the redshift is. In astronomy, it is
484 Chapter 7 Astrophysics and Cosmology

customary to characterize redshift by a dimensionless quantity z in the formula

λobserv
1+z = ,
λemit
where λobserv and λemit represent the observed and emitted wavelenths. There are three
redshift types:

Doppler effect, cosmological redshift, gravitational redshift.

The gravitational redshift in a black hole are caused by both the gravitational fields of the
emitting object and the black hole.
The first type of redshift, due to the gravitational field, is formulated as
r
2mG
1− 2
c r
1+z = r , (7.5.36)
2mG
1− 2
c r0
where m is the mass of the emitting object, r0 is its radius, and r is the distance between
the object and the observer.
The second type of redshifts, due to the cosmological effect or black hole effect, is
p
−g00(r0 )
1+z = p , (7.5.37)
−g00(r1 )
where g00 is the time-component of the black hole gravitational metric, r0 and r1 are the
positions of the observer and the emitting object (including virtual images).
If a universe is not considered as a black hole, then the gravitational redshift is simply
given by (7.5.36) and is very small for remote objects. Likewise, the cosmological redshift
is also too small to be significant. Hence, astronomers have to think the main portion of the
redshift is due to the Doppler effect:
p
1 + v/c
1+z = p . (7.5.38)
1 − v/c

When v/c is small, (7.5.38) can be approximatively expressed as

z ' v/c. (7.5.39)

In addition, Hubble discovered that the redshift has an approximatively linear relation with
the distance:
z ' kR, k is a constant. (7.5.40)
Thus, the Hubble Law (7.5.1) follows from (7.5.39) and (7.5.40). It is the Hubble Law
(7.5.1) that leads to the conclusion that our Universe is expanding.
7.5 The Universe 485

However, if we adopt the view that the globular universe is in a black hole with the
Schwarzschild radius Rs as in (7.5.31), the black hole redshift (7.5.37) cannot be ignored.
By (7.5.32) and (7.5.33), the time-component g00 for the black hole can approximatively
take the TOV solution as r near Rs :
 
1 r2
g00 = − 1− 2 , for r near Rs .
4 Rs
Hence, the redshift (7.5.37) is as
q
1 − r02/R2s
1+z = q , for r0 , r1 < Rs . (7.5.41)
1 − r12/R2s

It is known that for a remote galaxy, r1 is close to the boundary r = Rs . Therefore by (7.5.41)
we have
z → +∞ as r1 → Rs .
It reflects the redshifts observed from most remote objects. If the object is a virtual image
as shown in Figure 7.13, then its position is the reflection point r1 . Thus, we see that even
if the remote object is not moving, its redshift can still be very large.
5. CMB problem. In 1965, two physicists A. Penzias and R. Wilson discovered the
low-temperature cosmic microwave background (CMB) radiation, which fills our Universe,
and it is ever regarded as the Big-Bang product. However, for a static closed Universe, it
is the most natural thing that there exists a CMB, because the Universe is a black-body and
CMB is a result of black-body radiation.
6. None expanding Universe. As the energy of the Universe is given, the maximal
radius, i.e. the Schwarzschild radius Rs , is determined, and the boundary is invariant. In fact,
a globular universe must fill the ball with the Schwarzschild radius, although the distribution
of the matter in this ball may be slightly non-homogenous. The main reason is that if the
universe has a radius R smaller than Rs , then it must contain at least a sub-black hole with
radius R0 as follows r
R
R0 = R.
Rs
In Section 7.5.4 we shall discuss this topic.

7.5.3 Spherical Universe without boundary

Bounded universe has finite energy and space, and our Universe is bounded as we have
demonstrated in the last section. Besides the globular universe, another type of bounded
universe is the spherically-shaped corresponding to the k = 1 case in the Friedmann model
(7.5.15)-(7.5.17).
486 Chapter 7 Astrophysics and Cosmology

A globular universe must be static. With the same argument, a spherical closed universe
have to be static as well. In this subsection, we are devoted to investigate the spherical
cosmology.
1. Cosmic radius. For a static spherical universe, its radius Rc satisfies that

Ṙc = 0, R̈c = 0.

By the Friedmann equation (7.5.18), it leads to that

3c2
R2c = . (7.5.42)
8π Gρ
For a 3D sphere, its volume V is given by

V = 2π 2 R 3 .

Thus, ρ = M/2π 2 R3c , and by (7.5.43) we get the radius Rc as

4MG
Rc = . (7.5.43)
3π c2
This value (7.5.43) is also the maximal radius for a (possibly) oscillatory spherical universe.
2. Negative pressure. By (7.5.15) and R̈c = 0, the pressure is negative:
ρ c2
p=− . (7.5.44)
3
In order to resist the gravitational pulling, it is natural that there is a negative pressure in a
static universe, which originates from three sources:
thermal expanding, radiation pressure, and dark energy.
These three types of forces are repulsive, and therefore yield the negative pressure as given
by (7.5.44).
In fact, in our Universe both thermal and radiation (microwave radiation) pressures are
very small. The main negative pressure is generated by the so called “dark energy”. In (Ma
and Wang, 2014e), we have shown that the dark energy is the repulsive gravitational effect
for a remote object of great distance. From the field theoretical point view, dark energy is
an effect of the dual gravitational field Ψ µ in the PID-induced gravitational field equations
(4.4.10) discovered by the authors.
3. Equivalence. It seems that both spherical and globular geometries are very different.
However, in the following we show that they are equivalent in cosmology. In fact, as the
space-time curvature is caused by gravitation, a globular universe must be a 3D hemisphere
as shown in Figure 7.14(a), and a spherical universe is as shown in Figure 7.14(b), which is
a 3D sphere piecing the upper and lower hemispheres together.
7.5 The Universe 487

Figure 7.14 (a) A 3D hemisphere, and (b) a 3D sphere piecing the upper and lower hemispheres
together.

In cosmology, the globular universe is a black hole, which likes as a 3D hemisphere,

and the spherical universe can be regarded as if there were two hemispheres of black holes
attached together.
We show this version from the cosmological dynamics.
First, by the Newtonian cosmological equation (7.5.14), i.e.
 2
Ṙ 8π G kc2
= ρ− 2 . (7.5.45)
R 3 R

For a static universe in a black hole with maximal radius Rc , the equation (7.5.45) becomes

8π G kc2
Ṙc = 0 ⇔ ρ= 2. (7.5.46)
3 Rc

The volume of the hemisphere is

Z 2π Z π Z π
2
V0 = dθ dϕ R3c sin θ sin2 ψ d ψ = π 2 R3c ,
0 0 0

where x ∈ R4 takes the spherical coordinate:

(x1 , x2 , x3 , x4 ) = (Rc sin ψ sin θ cos ϕ , Rc sin ψ sin θ sin ϕ , Rc sin ψ cos θ , Rc cos ψ ). (7.5.47)

Thus, the mass density is

ρ = Mtotal /π 2 R3c , (7.5.48)

Then it follows from (7.5.46) that

8GMtotal
Rc = for k = 1. (7.5.49)
3π c2
Remark 7.26 The mass Mtotal in (7.5.48) contains the energy contributed by the space
curvature, i.e.
Mtotal = M + space curved energy,
488 Chapter 7 Astrophysics and Cosmology

where M is the mass of the flat space. By the invariance of density,

4π 3
M/ R = Mtotal /π 2R3c ,
3 c
we get the relation
3π
Mtotal = M. (7.5.50)
4
With the flat space mass (7.5.50), from (7.5.49) we get the Schwarzschild radius R s = Rc
for the cosmic black hole as follows

Rs = 2GM/c2 .

It means that the globular universe is essentially hemispherically-shaped. In particular the

relation (7.5.50) can be generated to an arbitrary region Ω ⊂ R 3 , i.e.

VΩ
MΩ;total = MΩ , (7.5.51)
|Ω|

where MΩ is the flat space mass in Ω, MΩ;total is the curved space mass, |Ω| is the volume of
flat Ω, Z
√
VΩ = gdx, g = det(gi j ),
Ω

and gi j (1 6 i, j 6 3) is the spatial gravitational metric.

Now, we return to the Friedmann model (7.5.18) with k = 1, which has the same form
as that of the globular dynamic equation (7.5.45), and is of the same maximal radius R c as
that in (7.5.49). Hence, it is natural that a static spherical universe is considered as if there
were two hemispherical black holes attached together. In fact, the static spherical universe
forms an entire black hole as a closed space.

4. Basic problems. Since a static spherical universe is equivalent to two globular uni-
verses to be pieced together along with their boundary, an observer in its hemisphere is as if
one is in a globular universe. Hence, the basic problems – the cosmic edge problem, flatness
problem, horizon problem, and CMB problem– can be explained in the same fashion.
The redshift problem is slightly different, and the gravitational redshift is given by

1
1+z = p , (7.5.52)
−g00 (r)

where r is the distance between the light source and the observer, and g 00 is the time-
component of the gravitational metric.
Due to the horizon of sphere, for an arbitrary point on a spherical universe, its opposite
hemisphere relative to the point plays a similar role as a black hole. Hence, in the redshift
7.5 The Universe 489

formula (7.5.52), g00 can be approximatively taken as the Schwarzschild solution for distant
objects as follows

Rs 2MG
−g00 = 1 − , Rs = , r = 2Rs − r
e for 0 6 r < Rs ,
r
e c2
where M is the cosmic mass of hemisphere, and e r is the distance from the light source to the
opposite radial point, and r is from the light source to the point. Hence, formula (7.5.52)
can be approximatively written as
√
1 2Rs − r
1+z = s   = p for 0 < r < Rs . (7.5.53)
Rs α (r)(R s − r)
α (r) 1 −
r
e

where
α (0) = 2, α (Rs ) = 1, α 0 (r) < 0.

We see from (7.5.53) that the redshift z → ∞ as r → Rs , and, consequently, we cannot see
objects in the opposite hemisphere. Intuitively, α (r) represents the gravitational effect of
the matter in the hemisphere of the observer.

5. Physical conclusions. In either case, globular or spherical, the universe is equivalent

to globular universe(s). It is not originated from a Big-bang, is static, and confined in a
black hole in the sense as addressed above. The observed mass M and the implicit mass
Mtotal have the relation
3π
Mtotal = 2 × M = 3π M/2, (7.5.54)
4
which is derived by (7.5.50) adding the mass of another hemisphere.
The implicit mass Mtotal of (7.5.54) contains the dark matter. In (Ma and Wang, 2014e),
both the dark matter Mtotal − M and the dark energy (i.e. the negative pressure (7.5.44)) are
just a property of gravity.

7.5.4 New cosmology

We start with two difficulties encountered in modern cosmology.

First, if the Universe were born to a Big-Bang and expanded continuously, then in the
expansion process it would generate successively a large number of black holes, whose radii
vary as follows:
r r
R0 R 2MG
R0 6 r 6 R, R0 < R 6 Rs = 2 , (7.5.55)
Rs Rs c
where M is the total mass in the universe, R0 is the initial radius, R is the expanding radius,
and r is the radius of sub-black holes.
490 Chapter 7 Astrophysics and Cosmology

To see this, we consider a homogeneous universe with radius R < R s . Then the mass
density ρ is given by
3M
ρ= . (7.5.56)
4π R3
On the other hand, by Theorem 7.3, the condition for a ball B r with radius r to form a black
hole is that the mass Mr in Br satisfies that

Mr c2
= . (7.5.57)
r 2G
By (7.5.56), we have
4π 3 r3
Mr = r ρ = 3 M.
3 R
Then it follows from (7.5.57) that r
R
r= R. (7.5.58)
Rs
Actually, in general for a ball Br in a universe with radius R < Rs , if its mass Mr satisfies
(7.5.57) then it will form a black hole, and its radius r satisfies that
r
R
r6 R.
Rs

In particular, there must exist a black hole whose radius r is as in (7.5.58). Thus, we derive
the conclusion (7.5.55).
Based on (7.5.55) we can deduce that if the Universe were born to a Big-Bang and
continuously expands, then it would contain many black holes with smaller ones being
embedded in the larger ones. In particular, the Universe would contain a huge black hole
whose radius r is almost equal to the cosmic radius Rs . This is not what we observed in our
Universe.
The second difficulty of modern cosmology concerns with the Hubble Law (7.5.1),
which is restated as v = HR, where c/H = Rs . Consider a remote object with mass M0 , then
M0
the observed mass Mobser is given by Mobser = r . Consequently, the corresponding
v2
1− 2
c
gravitational force F to the observer with mass m is

mMobser G mM0 G mM0 G mM0 G

F =− =− r =− r =− s .
r2 v 2 2
H 2 r2
2
r 1− 2 2
r 1− 2 r 2
r 1− 2
c c Rs

It is clear then that as r → Rs , F → −∞. This is clearly not what is observed.

In conclusion, we have rigorously derived the following new theory of cosmology:
7.5 The Universe 491

Theorem 7.27 Assume a) the Einstein theory of general relativity, and b) the prin-
ciple of cosmological principle that the universe is homogeneous and isotropic. Then the
following assertions hold true for our Universe:

1) All universes are bounded, are not originated from a Big-Bang, and are static; and

2) The topological structure of our Universe can only be the 3D sphere such that to each
observer, the corresponding equator with the observer at the center of the hemisphere
can be viewed as the black hole horizon.

Theorem 7.28 If we assume only a) the Einstein theory of general relativity, and b’)
the universe is homogeneous. Then all universes can only be either a 3D sphere as given in
Theorem 7.27, or a globular universe, which is a 3D open ball B Rs of radius Rs , forming the
interior of a black hole with Rs as its Schwarzschild radius. In the later case, the Universe
is also static, is not originated from a Big-Bang, and the matter fills the entire Universe.
Also, the following assertions hold true:
1) The cosmic observable mass M and the total mass Mtotal , which includes both M
and the non-observable mass due to the space curvature energy, satisfy the following
relation (
3π M/2 for the spherical structure,
Mtotal = (7.5.59)
3π M/4 for the globular structure.
The difference Mtotal − M can be regarded as the dark matter;

2) The static Universe has to possess a negative pressure to balance the gravitational
attracting force. The negative pressure is actually the effect of the gravitational re-
pelling force, also called dark energy; and

3) Both dark matter and dark energy are a property of gravity, which is reflected in
both space-time curvature, and the attracting and repulsive gravitational forces in
different scales of the Universe. This law of gravity is precisely described by the new
gravitational field equations (4.4.10); see also (Ma and Wang, 2014e).

We end this section with three remarks and observations.

First, astronomical observations have shown that the measurable mass M is about one
fifth of total mass Mtotal . By (7.5.59), for the spherical universe,

Mtotal = 4.7M.

This relation also suggest that the spherical universe case fits better the current understand-
ing for our Universe.
Second, due to the horizon of sphere, for an arbitrary point in a spherical universe, its op-
posite hemisphere relative to the point is as if it is a black hole. Hence the main contribution
to the redshifts is from the effect of the black hole, as explicitly given by (7.5.53).
492 Chapter 7 Astrophysics and Cosmology

Third, in modern cosmology, the view of expanding universe was based essentially on
the Friedmann model and the Hubble Law. The observations can accurately measure the
distances and redshifts for some galaxies, which allowed astronomers to get both measured
and theoretical data, and their deviation led to the conclusion that the expanding universe is
accelerating. The misunderstanding comes from the perception that the Doppler redshift is
the main source of redshifts.

7.6 Theory of Dark Matter and Dark Energy

7.6.1 Dark energy and dark matter phenomena
1. Dark matter and Rubin rotational curve. In astrophysics, dark matter is an unknown
form of matter, which appears only participating in gravitational interaction, but does not
emit nor absorb electromagnetic radiations.
Dark matter was first postulated in 1932 by Holland astronomer Jan Oort, who noted
that the orbital velocities of stars in the Milky Way don’t match their measured masses.
Namely, the orbital velocity v and the gravity should satisfy the equilibrium relation
v2 Mr G
= 2 , (7.6.1)
r r
where Mr is the total mass in the ball Br with radius r. But the observed mass M0 was
less than the theoretic mass Mr in (7.6.1), and the difference Mr − M0 was explained as the
presence of dark matter. The phenomenon was also discovered by Fritz Zwicky in 1933
for the missing mass in the orbital velocities of galaxies in clusters. Subsequently, other
observations have manifested the existence of dark matter in the Universe, including the
rotational velocities of galaxies, gravitational lensing, and the temperature distribution of
hot gaseous.
A strong support to the existence of dark matter is the Rubin rotational curves for galac-
tic rotational velocity. The rotational curve of a galaxy is the rotational velocity of visible
stars or gases in the galaxy on their radial distance from the center of the galaxy. The Ru-
bin rotational curve amounts to saying that most stars in spiral galaxies orbit at roughly
the same speed. If a galaxy had a mass distribution as the observed distribution of visible
astronomical objects, the rotational velocity would decrease at large distances. Hence, the
Rubin curve demonstrates the existence of additional gravitational effect to the gravity by
the visible matter in the galaxy.
More precisely, the orbital velocity v(r) of the stars located at radius r from the center
of galaxies is almost a constant:

v(r) ∼
= a constant for a given galaxy, (7.6.2)

as illustrated typically by Figure 7.15 (a), where the vertical axis represents the velocity
(Km/s), and the horizontal axis is the distance from the galaxy center (extending to the
7.6 Theory of Dark Matter and Dark Energy 493

galaxy radius). However, the calculation from (7.6.1) gives a theoretic curve as shown in
Figure f6.15(b), showing discrepancies between the mass determined from the gravita-
tional effect and the mass calculated from the visible matter. The missing mass suggests the
presence of dark matter in the Universe.

Figure 7.15 (a) Typical galactic rotational curve by Rubin, and (b) theoretic curve based on the
Newtonian gravitational law.

In fact, we have seen in Section 7.5.4 that the dark matter is a space curved energy, or
equivalently a gravitational effect, which is also reflected in the revised gravitational force
formula in which there is an additional attracting force to the classical Newtonian gravity.

2. Dark energy. Dark energy was first proposed in 1990’s, which was based on the
hypotheses that the Universe is expanding.
The High-z Supernova Search Team in 1998 and the Supernova Cosmology Project in
1999 published their precisely measured data of the distances of supernovas and the red-
shifts. The observations indicated that the measured and theoretical data have a deviation,
which was explained, based on the Hubble Law and the Friedmann model, as the accel-
eration of the expanding universe. The accelerating expansion is widely accepted as an
evidence of the existence of dark energy.
However, based on the new cosmology postulated in the last section, the dark energy is
a field energy form of gravitation which balances the gravitational attraction to maintain the
homogeneity and stability of the Universe.

7.6.2 PID cosmological model and dark energy

We have shown in (Ma and Wang, 2014e) that both dark matter and dark energy are a
property of gravity. Dark matter and dark energy are reflected in two aspects: a) the large
scale space curved structure of the Universe caused by gravity, and b) the gravitational
attracting and repelling aspects of gravity. In this section, we mainly explore the nature of
dark energy in aspect a) using the PID-induced cosmological model.
494 Chapter 7 Astrophysics and Cosmology

PID cosmological model

According to Theorem 7.22, the metric of a homogeneous spherical universe is of the
form  
2 2 2 2 dr2 2 2 2 2
ds = −c dt + R + r (d θ + sin θ d ϕ ) , (7.6.3)
1 − r2
where R = R(t) is the cosmic radius. The PID induced gravitational field equations are
given by    
8π G 1 1
Rµν = − 4 Tµν − gµν T − ∇µν φ − gµν Φ , (7.6.4)
c 2 2
where Φ = gαβ Dαβ φ , and φ depends only on t.
The nonzero components of R µν read as

3 1
R00 = Rtt ,
c2 R
1
Rkk = − 2 2 gkk (RRtt00 + 2Rt2 + 2c2 ) for 1 6 k 6 3,
c R
and by Tµν = diag(c2 ρ , g11 p, g22 p, g33 p), we have
 
1 c2 3p
T00 − g00 T = ρ+ 2 ,
2 2 c
1 c 2  p
Tkk − gkk T = gkk ρ − 2 for1 6 k 6 3,
2 2 c
 
1 1 3Rt
φ00 − g00 Φ = 2 φtt − φt ,
2 2c R
 
1 1 Rt
φkk − gkk Φ = 2 gkk φtt + φt for1 6 k 6 3.
2 2c R

Thus, we derive from (7.6.4) two independent field equations as

 
4π G 3p 1 1
00
R =− ρ + 2 R − φ 00 R + R0 φ 0 , (7.6.5)
3 c 6 2
 0 2 
R00 R 2c2 p 1 1 R0 0
+2 + 2 = 4π G ρ − 2 + φ 00 + φ. (7.6.6)
R R R c 2 2R

We infer from (7.6.5) and (7.6.6) that

8π G 2 1
(R0 )2 = R ρ + R2 φ 00 − c2 . (7.6.7)
3 3
By the Bianchi identity:
8π G
∇µ (∇µν φ + Tµν ) = 0,
c4
7.6 Theory of Dark Matter and Dark Energy 495

we deduce that  
3R0 00 3R0 3R0 p
φ +
000
φ = −8π G ρ +
0
ρ+ . (7.6.8)
R R R c2
It is known that the energy density ρ and the cosmic radius R (also called the scale
factor) satisfy the relation:
ρ0
ρ= , ρ0 the density at R = 1. (7.6.9)
R3
Hence, it follows from (7.6.9) that

ρ 0 = −3ρ R0/R.

Thus, (7.6.8) is rewritten as

3R0 00 24π G R0
φ 000 + φ =− 2 p. (7.6.10)
R c R
In addition, making the transformation
ψ
φ 00 = , (7.6.11)
R3
then, from (7.6.5), (7.6.7) and (7.6.9)-(7.6.11) we can deduce that

(R0 )2 φ 0 = 0. (7.6.12)

Denote ϕ = φ 00 , by (7.6.12), the equations (7.6.5), (7.6.7) and (7.6.10) can be rewritten
in the form  
4π G 3p ϕ
00
R =− ρ+ 2 + R,
3 c 8π G
1 (7.6.13)
(R0 )2 = (8π Gρ + ϕ )R2 − c2 ,
3
3R0 24π G R0
ϕ0 + ϕ =− 2 p.
R c R
Only two equations in (7.6.13) are independent. However, there are three unknown
functions R, ϕ , p in (7.6.13). Hence, we need to add an additional equation, the equation of
state, as follows:
p = f (ρ , ϕ ). (7.6.14)
Based on Theorem 7.27, the model describing the static Universe is the equation (7.6.14)
together with the stationary equations of (7.6.13), which are equivalent to the form
 
3p
ϕ = −8π G ρ + 2 ,
c
(7.6.15)
c 4
p=− .
8π GR2
496 Chapter 7 Astrophysics and Cosmology

The equations (7.6.14) and (7.6.15) provide a theoretic basis for the static Universe,
including the dark energy.
Now, we need to determine the explicit expression for the equation (7.6.14) of state. It
is natural to postulate that the equation of state is linear. Hence, (7.6.14) can be written as

c2
p= (α1 ϕ − α2 Gρ ), (7.6.16)
G
where α1 and α2 are nondimensional parameter, which will be determined by the observed
data.
The equations (7.6.15) and (7.6.16) are the PID cosmological model, where the cosmo-
logical significants of R, p, ϕ , ρ are as follows:

R the cosmic radius (of the 3D spherical universe),

p the negative pressure, generated by the repulsive aspect of gravity,
ϕ represents the dual gravitational potential, (7.6.17)
3M Mtotal
ρ the cosmic density, given by = 2 3,
4π R3 π R
where M and Mtotal are as in Remark 7.26.
Here, we remark that in the classical Einstein field equations where φ = 0, the relation
(7.6.9) still holds true, by which we can deduce that R0 p = 0.

Theory of dark energy

In the static cosmology, dark energy is defined in the following manner. Let E ob be the
observed energy, and R be the cosmic radius. We define the observable mass and the total
mass as follows:
Eob
Mob = , (7.6.18)
c2
Rc2
MT = . (7.6.19)
2G
If MT > Mob , then the difference
∆E = ET − Eob (7.6.20)

is called the dark energy.

The CMB measurement and the WMAP analysis indicate that the difference ∆E in
(7.6.20) is positive,
∆E > 0,

which is considered as another evidence for the presence of dark energy.

7.6 Theory of Dark Matter and Dark Energy 497

From the PID cosmological model (7.6.15)-(7.6.17), we see that the dark energy ∆E
in (7.6.20) is essentially due to the dual gravitational potential ϕ . In fact, we infer from
(7.6.15) that
ϕ = 0 ⇔ R = 2Mob G/c2 (i.e. ∆E = 0),
(7.6.21)
ϕ > 0 ⇔ ∆E > 0.
Hence, dark energy is generated by the dual gravitational field. This fact is reflected in the
PID gravitational force formula derived in subsections hereafter.
If we can measure precisely, with astronomical observations, the energy (7.6.18) and the
cosmic radius R (i.e. MT of (7.6.19)), then we can obtain a relation between the parameters
α1 and α2 in (7.6.16). In fact, we deduce from (7.6.15) and (7.6.16) that

βp 1 + 24πα1
ρ+ = 0, β= . (7.6.22)
c2 α2 + 8πα1
As we get
∆M MT − Mob
= =k (k > 0). (7.6.23)
Mob Mob
Then by (7.6.22) and
3Mob c4
ρ= , p=− ,
4π R3 8π GR2
we obtain from (7.6.22) that
3α2 = 24kπα1 + k + 1. (7.6.24)

By the relation (7.6.24) from (7.6.22), we can also derive, in the same fashion as above,
the dark energy formula (7.6.23).

7.6.3 PID gravitational interaction formula

Consider a central gravitational field generated by a ball Br0 with radius r0 and mass M. It
is known that the metric of the central field at r > r0 can be written in the form

ds2 = −eu c2 dt 2 + ev dr2 + r2 (d θ 2 + sin2 θ d ϕ 2 ), (7.6.25)

and u = u(r), v = v(r).

In the exterior of Br0 , the energy-momentum is zero, i.e.

Tµν = 0, for r > r0 .

Hence, the PID gravitational field equation for the metric (7.6.25) is given by
1
Rµν − gµν R = −∇µν φ , r > r0 . (7.6.26)
2
where φ = φ (r) is a scalar function of r.
498 Chapter 7 Astrophysics and Cosmology

By (7.1.25) and (7.1.26), we have

 
1 1 1
R00 − g00 R = − eu−v v0 + (ev − 1) ,
2 r r
 
1 1 0 1 v
R11 − g11 R = − u − (e − 1) ,
2 r r
2
   
1 r 1 0 1
R22 − g22 R = − e−v u00 + u+ (u0 − v0 ) ,
2 2 2 r
1 u−v 0 0
∇00 φ = − e u φ ,
2
1
∇11 φ = φ − v0 φ 0 ,
00
2
∇22 φ = −re−v φ 0 .

Thus, the fields equations (7.6.26) are as follows

1 r
v0 + (ev − 1) = − u0 φ 0 ,
r 2
1 v 1
u − (e − 1) = r(φ 00 − v0 φ 0 ),
0
(7.6.27)
r  2
1 0 1 2
00
u + u+ (u − v ) = − φ 0 .
0 0
2 r r
Now we are ready to deduce from (7.6.27) the PID gravitational interaction formula as
follows.
First, we infer from (7.6.27) that
rφ 00
u0 + v 0 = r ,
1 + φ0
2  
1 2 v
0 0
u −v = r (e − 1) + r φ 00
.
1 − φ0 r
2
Consequently,
1 1 v rφ 00
u0 = r 0 r (e − 1) + r . (7.6.28)
1− φ 1 − ( φ 0 )2
2 2
It is known that the interaction force F is given by
c2 u
F = −m∇ψ , ψ= (e − 1).
2
Then, it follows from (7.6.28) that
 
mc2  1 1 v rφ 00 
F= e u − r 0 r (e − 1) −  r 2  . (7.6.29)
2 1− φ
2 1 − φ0
2
7.6 Theory of Dark Matter and Dark Energy 499

The formula (7.6.29) provides the precise gravitational interaction force exerted on an
object with mass m in a spherically symmetric gravitation field.
In classical physics, the field functions u and v in (7.6.29) are taken by the Schwarzschild
solution:  
u 2GM v 2GM −1
e = 1− 2 , e = 1− 2 , (7.6.30)
c r c r
and φ 0 = φ 00 = 0, which leads to the Newton gravitation.
However, due to the presence of dark matter and dark energy, the field functions u, v, φ
in (7.6.29) should be an approximation of the Schwarzschild solution (7.6.30). Hence we
have
|rφ 0 |  1 for r > r0 . (7.6.31)
Under the condition (7.6.31), formula (7.6.29) can be approximatively expressed as
 
mc2 u 1 v
F= e − (e − 1) − rφ . 00
(7.6.32)
2 r
7.6.4 Asymptotic repulsion of gravity
In this section, we shall consider the asymptotic properties of gravity, and rigorously prove
that the interaction force given by (7.6.32) is repulsive at very large distance.
To this end, we need to make the following transformation to convert the field equations
(7.6.27) into a first order autonomous system:

r = es ,
w = ev − 1,
x1 (s) = es u0 (es ), (7.6.33)
s
x2 (s) = w(e ),
x3 (s) = es φ 0 (es ).

Then the equations (7.6.27) can be rewritten as

1 1 1 1
x01 = −x2 + 2x3 − x21 − x1 x3 − x1 x2 − x21 x3 ,
2 2 2 4
1 1
x02 = −x2 − x1 x3 − x22 − x1 x2 x3 , (7.6.34)
2 2
1 1
x3 = x1 − x2 + x3 − x2 x3 − x1 x23 .
0
2 4
The system (7.6.34) is supplemented with an initial condition

(x1 , x2 , x3 )(s0 ) = (α1 , α2 , α3 ) with r0 = es0 . (7.6.35)

We now study the problem (7.6.34)-(7.6.35) in a few steps as follows.

500 Chapter 7 Astrophysics and Cosmology

Step 1. Asymptotic flatness. For a globular matter distribution, its gravitational field
should be asymptotically flat, i.e.

g00 → −1, g11 → 1 if r → ∞.

It implies that x = 0 is the uniquely physical equilibrium point of (7.6.34) and the following
holds true:
x(s) → 0 if s → ∞ (i.e. r → ∞). (7.6.36)

Step 2. Physical initial values. The physically meaningful initial values α = (α 1 , α2 , α3 )

in (7.6.35) have to satisfy the following two conditions:

(a) The solutions x(s, α ) of (7.6.34)-(7.6.35) are asymptotically flat in the sense of (7.6.36).
Namely, the initial values α are in the stable manifold E s of x = 0, defined by

E s = {α ∈ R3 | x(s, α ) → 0 for s → ∞}; (7.6.37)

(b) The solutions x(s, α ) are near the Schwarzschild solution:

|x1 − es u0 |, |x2 + 1 − ev|, |x3 |  1, (7.6.38)

where u, v are as in (7.6.30).

In fact, by (7.6.31) and (7.6.33) we can see that all Schwarzschild solutions lie on the
line
L = {(x1 , x2 , 0)| x1 = x2 , x1 , x2 > 0}. (7.6.39)
In particular, the line L is on the stable manifold E s of (7.6.37).
Step 3. Stable manifold E s . The equations (7.6.34) can be written as

ẋ = Ax + O(|x|2),

where  
0 −1 2
A =  0 −1 0  . (7.6.40)
1 −1 1
The dimension of the stable manifold E s is the number of negative eigenvalues of the matrix
A. It is easy to see that the eigenvalues of A are given by

λ1 = −1, λ2 = −1, λ3 = 2.

Hence, the dimension of E s is two:

dim E s = 2.
7.6 Theory of Dark Matter and Dark Energy 501

Consequently, the initial value α of an asymptotically flat solution has only two independent
components due to α ∈ E s , which is of two dimensional. Namely, we arrive at the following
conclusion.

Physical Conclusion 7.29 In the gravitation formula (7.6.29) there are two free pa-
rameters to be determined by experiments (or by astronomical measurements).

In fact, the two free parameters will be determined by the Rubin rotational curve and the
repulsive property of gravity at large distance.
Step 4. Local expression of E s . In order to derive the asymptotic property of the grav-
itational force F of (7.6.29), we need to derive the local expression of the stable manifold
E s near x = 0. Since the tangent space of E s at x = 0 is spanned by the two eigenvectors
(1, 1, 0)t and (1, −1, −1)t corresponding to the two negative eigenvalues λ1 = λ2 = −1, the
coordinate vector (0, 0, 1) of x3 is not contained in E s . This implies that the stable manifold
can be expressed near x = 0 in the form

x3 = h(x1 , x2 ). (7.6.41)

Inserting the Taylor expansion for (7.6.41) into (7.6.34), and comparing the coefficients,
we derive the following local expression of (7.6.41) of the stable manifold function:
1 1 1 1
h(x1 , x2 ) = − x1 + x2 + x21 − x22 + O(|x|3 ). (7.6.42)
2 2 16 16
Inserting (7.6.41)-(7.6.42) into the first two equations of (7.6.34), we deduce that
1 1 3
x01 = −x1 − x21 − x22 − x1 x2 + O(|x|3 ),
8 8 4 (7.6.43)
1 1
x02 = −x2 + x21 − x22 − x1 x2 + O(|x|3 ).
4 4
The system (7.6.43) is the system (7.6.34) restricted on the stable manifold E s . Hence, its
asymptotic behavior represents that of the interaction force F in (7.6.29).
Step 5. Phase diagram of system (7.6.43). In order to obtain the phase diagram of
(7.6.43) near x = 0, we consider the ratio: x02 /x01 = dx2 /dx1 . Omitting the terms O(|x|3 ), we
infer from (7.6.43) that
1 1
dx2 x2 + x22 + x2 x1 − x21
= 4 4 . (7.6.44)
dx1 1 2 3 1
x1 + x2 + x2 x1 + x21
8 4 8
Let k be the slope of an orbit reaching to x = 0:
x2
k= as (x2 , x1 ) → 0.
x1
502 Chapter 7 Astrophysics and Cosmology

Then (7.6.44) can be expressed as

1 1
k + k2 x1 + kx1 − x1
k= 4 4 ,
1 2 3 1
1 + k x1 + kx1 + x1
8 4 8
which leads to the equation
k3 − 2k2 − k + 2 = 0.
This equation has three solutions:

k = ±1, k = 2,

giving rise to three line orbits:

x2 = x 1 , x2 = 2x1 , x2 = −x1 ,

which divide the neighborhood of x = 0 into six invariant regions. It is clear that all physi-
cally meaningful orbits can only be in the following three regions:
 
1
Ω1 = −x2 < x1 < x2 , x2 > 0 ,
2
 
1 (7.6.45)
Ω2 = x2 6 x 1 6 x 2 , x2 > 0 ,
2

Ω3 = {x2 < x1 , x2 > 0}.

On the positive x2 -axis (i.e. x1 = 0, x2 > 0), which lies in Ω1 , the equations (7.6.43) take
the form
1
x01 = − x22 + O(|x|3 ),
8
x02 = −x2 − x22 + O(|x|3 ).

It is easy to show that the orbits in Ω1 with x1 > 0 will eventually cross the x2 −axis. Thus,
using the three invariant sets in (7.6.45), we obtain the phase diagram of (7.6.43) on x 2 > 0
as shown in Figure 7.16. In this diagram, we see that, the orbits in Ω 2 and Ω3 will not cross
the x2 -axis, but these in Ω1 with x1 > 0 will do.
Step 6. Asymptotic repulsion theorem of gravity. We now derive an asymptotic repulsion
theorem of gravity, based on the phase diagram in Figure 7.16. In fact, by (7.6.28) and
(7.6.29), the gravitational force F reads as
mc2 u 0
F =− e u. (7.6.46)
2
7.6 Theory of Dark Matter and Dark Energy 503

Figure 7.16 Only the orbits on Ω1 with x1 > 0 will eventually cross the x2 -axis, leading to the sign
change of x1 , and to a repelling gravitational force corresponding to dark energy.

It is known that

F <0 represents attraction,

F >0 represents repelling.

Hence, by x1 = ru0 (r) and (7.6.46), the phase diagram shows that an orbit in Ω 1 , starting
with x1 > 0, will cross the x2 -axis, and the sign of x1 changes from positive to negative,
leading consequently to a repulsive gravitational force F. Namely, we have obtained the
following theorem.

Theorem 7.30 (Asymptotic Repulsion of Gravitation) For a central gravitational

field, the following assertions hold true:

1) The gravitational force F is given by (7.6.29), and is asymptotic zero:

F → 0 if r → ∞. (7.6.47)

2) If the initial value α in (7.6.35) is near the Schwarzschild solution (7.6.31) with 0 <
α1 < α2 /2, then there exists a sufficiently large r1 such that the gravitational force F
is repulsive for r > r1 :
F > 0 for r > r1 . (7.6.48)

We remark that Theorem 7.30 is valid provided the initial value α is small because
the diagram given by Figure 7.16 is in a neighborhood of x = 0. However, all physically
meaningful central fields satisfy the condition (note that any a black hole is enclosed by a
huge quantity of matter with radius r > 0  2MG/c2 ). In fact, the Schwarzschild initial
values are as
δ 2MG
x1 (r0 ) = x2 (r0 ) = , δ= 2 . (7.6.49)
1−δ c r0
504 Chapter 7 Astrophysics and Cosmology

For example see (7.6.29), where the δ -factors are of the order δ 6 10 −1 , sufficient for the
requirements of Theorem 7.30.
The most important cases are for galaxies and clusters of galaxies. For these two types
of astronomical objects, we have

galaxy : M = 1011 M , r0 = 3 × 105ly,

cluster of galaxies : M = 1014 M , r0 = 3 × 106ly.

Thus the δ -factors are

galaxies δ = 10−7 , cluster of galaxies δ = 10−5 . (7.6.50)

In fact, the dark energy phenomenon is mainly evident between galaxies and between clus-
ters of galaxies. Hence, (7.6.50) shows that Theorem 7.30 is valid for both central gravita-
tional fields of galaxies and clusters of galaxies. The asymptotic repulsion of gravity plays
the role to stabilize the large scale homogeneous structure of the Universe.

7.6.5 Simplified gravitational formula

We have shown that all four fundamental interactions are layered. Namely, each interaction
has distinct attracting and repelling behaviors in different scales and levels. The dark matter
and dark energy represent the layered property of gravity.
In this section, we simplify the gravitational formula (7.6.32) to clearly exhibit the lay-
ered phenomena of gravity.
In (7.6.32) the field functions u and v can be approximatively replaced by the
Schwarzschild solution (7.6.30). Since 2MG/c2 r is very small for r > r0 as indicted in
(7.6.29) and (7.6.50), the formula (7.6.32) can be expressed as
 
1 r 00
F = mMG − 2 − φ , r > r0 . (7.6.51)
r δ r0

By the field equation (7.6.26), we have

R = Φ for r > r0 , (7.6.52)

where R is the scalar curvature, and

 
1 2
Φ = gµν Dµν φ = e−v −φ 00 + (u0 − v0)φ 0 + ϕ 0 .
2 r

In view of (7.6.52), we obtain that

2 1
φ 00 = −ev R + φ 0 + (u0 − v0 )φ 0
r 2
7.6 Theory of Dark Matter and Dark Energy 505

Again by the Schwarzschild approximation, we have

 
2 δ r0
φ =
00
+ 2 φ 0 − R. (7.6.53)
r r

Integrating (7.6.53) and omitting e±δ r0 /r , we derive that

 Z 
2
φ = −r ε + r Rdr ,
0 −2

where ε is a constant. Thus (7.6.51) can be rewritten as

   Z 
1 r 2r R
F = mMG − 2 + R+ 1+ εr + r dr . (7.6.54)
r δ r0 δ r0 r2
The solutions of (7.6.34) can be Taylor expanded. Hence by (7.6.33) we see that

1 ∞
u0 (r) = ∑ ak (r − r0 )k .
r2 k=0

By (7.6.46), the gravitational force F takes the following form

1 ∞
F= ∑ bk r k ,
r2 k=0
b0 = −mMG.

In view of (7.6.54), it implies that R can be expanded as

ε0
R= − ε1 + O(r),
r
and by Physical Conclusion 7.29, ε0 and ε1 are two to-be-determined free parameters. In-
serting R into (7.6.54) we obtain that
 
1 k0
F = mMG − 2 − + ϕ (r) for r > r0 . (7.6.55)
r r
1
where k0 = ε0 , and
2
ε1
ϕ (r) = ε1 + k1 r + O(r), k1 = ε + .
δ r0
The nature of dark matter and dark energy suggests that

k0 > 0, k1 > 0.

Based on Theorem 7.30, ϕ (r) → 0 as r → ∞, and (7.6.55) can be further simplified as in the
form for r0 < r < r1 ,  
1 k0
F = mMG − 2 − + k1 r , (7.6.56)
r r
506 Chapter 7 Astrophysics and Cosmology

where k0 and k1 will be determined by the Rubin rotational curve and the astronomical data
for clusters of galaxies in the next section, where we obtain that

k0 = 4 × 10−18km−2 , k1 = 10−57km−3 . (7.6.57)

The formula (7.6.56) is valid only in the interval

r0 < r < r 1 ,

and r1 is the distance at which F changes its sign:

F(r1 ) = 0.

Both observational evidence on dark energy and Theorem 7.30 show that the distance r 1
exists. The formula (7.6.56) with (7.6.57) clearly displays the layered property of gravity:
attracting at short distance and repelling at large distance.

7.6.6 Nature of dark matter and dark energy

As mentioned in Section 7.6.2, both dark matter and dark energy are a property of gravita-
tional effect, reflected in two aspects, which will be addressed in detail in this section:

a) spatially geometrical structure, and

b) gravitational attracting and repelling as in (7.6.56).

Space curved energy and negative pressure

Gravitational potential causes space curvature and the spherical structure of the Uni-
verse, and displays two types of energies: a) dark matter contributed by the curvature of
space, and b) dark energy generated by the dual gravitational potential in (7.6.17). We
address each type of energy as follows.
1. Dark matter: the space curved energy. In Section 7.5.2, we have introduced the space
curved energy Mtotal for the 3D spherical Universe as follows:
3π
Mtotal = M, M is the observed mass in the hemisphere.
2
Now, we consider a galaxy with an observed mass MΩ . By (7.5.51), we have shown that the
space curved energy Mtotal;Ω is
VΩ
Mtotal;Ω = MΩ , (7.6.58)
|Ω|
where Ω is the domain occupied by the galaxy, VΩ and |Ω| are the volumes of curved and
flat Ω. VΩ contain two parts:

VΩ = cosmic spherical VΩ1 + local bump VΩ2 . (7.6.59)

7.6 Theory of Dark Matter and Dark Energy 507

It is known that
3π
VΩ1 = |Ω|.
4
For VΩ2 , we propose that

VΩ2 = π 2 r03 , r0 the galaxy radius.

In fact, the formula is precise for the galaxy nucleus.

4
By |Ω| = π r03 , we infer from (7.6.58) and (7.6.59) that
3
3π
Mtotal;Ω = MΩ , (7.6.60)
2
which gives rise to the relation between the masses of dark matter and observable matter.

2. Dark energy: the dual gravitational potential. The static universe is described by the
stationary solution of (7.6.13)-(7.6.14), which is given by (7.6.15)-(7.6.16). In the solution
a negative pressure presents, which prevents galaxies and clusters of galaxies from gravita-
tional contraction to form a void universe, and maintains the homogeneous distribution of
the Universe. The negative pressure contains two parts:

1 c2
p = − ρ c2 − ϕ (see (7.6.15)), (7.6.61)
3 24π G
where the first term is contributed by the observable energy, and the second term is the dark
energy generated by the dual gravitational potential ϕ ; see also (7.6.21).
By the Blackhole Theorem, Theorem 7.15, black holes are incompressible in their inte-
riors. Hence, in (7.6.61) the negative pressure

1
p = − ρ c2 , (7.6.62)
3
is essentially the incompressible pressure of the black hole generated by the normal energy.
By the cosmology theorem, Theorem 7.27, the Universe is a 3D sphere with a blackhole
radius. However, the CMB and the W MAP measurements manifest that the cosmic radius
R is greater than the blackhole radius of normal energy. By (7.6.21), the deficient energy is
compensated by the dual gravitational potential, i.e. by the second term of (7.6.61).

Attraction and repulsion of gravity

Based on Theorem 7.30, gravity possesses additional attraction and repelling to the
Newtonian gravity, as shown in the revised gravitational formula:
 
1 k0
F = mMG − 2 − + k1 r . (7.6.63)
r r
508 Chapter 7 Astrophysics and Cosmology

By using this formula we can explain the dark matter and dark energy phenomena. In
particular, based on the Rubin rotational curve and astronomical data, we can determine an
approximation of the parameters k0 and k1 in (7.6.63).

1. Dark matter: the additional attraction. Let Mr be the total mass in the ball with radius
r of galaxy, and V be the constant galactic rotational velocity. By the force equilibrium, we
infer from (7.6.63) that
 
V2 1 k0
= Mr G 2 + − k1 r , (7.6.64)
r r r
which implies that
V2 r
Mr = . (7.6.65)
G 1 + k0 r − k1 r 3
The mass distribution (7.6.65) is derived based on both the Rubin rotational curve and the
revised formula (7.6.63). In the following we show that the mass distribution Mr fits the
observed data.
We know that the theoretic rotational curve given by Figure 7.15(b) is derived by using
the observed mass Mob and the Newton formula

mMob G
F =− .
r2
Hence, to show that Mr = Mob , we only need to calculate the rotational curve vr by the
Newton formula from the mass Mr , and to verify that vr is consistent with the theoretic
curve. To this end, we have
v2r Mr G
= 2 ,
r r
which, by (7.6.65), leads to
V
vr = p .
1 − k 0 r − k1 r 2
As k1  k0  1, vr can be approximatively written as
 
1 1
vr = V 1 − k0 r + k02 r2 . (7.6.66)
2 4

It is clear that the rotational curve described by (7.6.66) is consistent with the theoretic
rotational curve as illustrated by Figure 7.15(b). Therefore, it implies that

Mr = Mob . (7.6.67)

The facts of (7.6.64) and (7.6.67) are strong evidence to show that the revised formula
(7.6.63) is in agreement with the astronomical observations.
7.6 Theory of Dark Matter and Dark Energy 509

We now determine the constant k0 in (7.6.63). According to astronomical data, the

average mass Mr1 and radius r1 of galaxies are about

Mr1 = 1011M ∼ = 2 × 1041Kg,

(7.6.68)
r1 = 104 ∼ 105 pc =
∼ 1018Km.

The observations show that the constant velocity V in the Rubin rotational curve is about
V = 300km/s. Then we have
V2
= 1024kg/km
G
Based on physical considerations,

k 0  k 1 r1 (r1 as in (7.6.68)). (7.6.69)

Then, we deduce from (7.6.65) that

V2 1 1
k0 = − = 4 × 10−18Km−1 . (7.6.70)
G Mr1 r1
We can explain the dark matter by the revised formula (7.6.63). As the matter distribu-
tion Mr is in the form
V2 r
Mr = ,
G 1 + k0 r
then the Rubin law holds true. In addition, the revised gravitation produces an excessive
mass Me as
2 2
Me = MT − Mr = V r1 − V r1
,
1
G G 1 + k 0 r1
where M1 = V 2 r1 /G is the Newton theoretic value of the total mass. Hence we have

Me k 0 r1 4 Me
= = or = 5.
MT 1 + k 0 r1 5 Mr1

Namely, the additional mass M e is four time the visible matter Mr = MT − M.e Thus, it gives
1
an explanation for the dark matter.
We remark that the ratio M/M e ob = 5 is essentially the same as in (7.6.60). It shows
that the dark matter is a gravitational effect, reflected in both the space curvature and the
additional gravitational attraction.
2. Dark energy: asymptotic repulsion of gravity. If gravity is always attracting as given
by the Newton formula, then the cosmic homogeneity is unstable. In fact, It is known that
the average mass M and distance for the clusters of galaxies are as

M = 1014M = ∼ 1044Kg
(7.6.71)
= 108 pc ∼
r∼ = 1020 ∼ 1021 Km.
510 Chapter 7 Astrophysics and Cosmology

Then the Newton gravitation between two clusters of galaxies is

M 2 G ∼ 29
F =− = 10 N = 1028Kg. (7.6.72)
r2
However, astronomical observations indicate that no gravitational interaction between
clusters of galaxies. The Universe is isotropic, therefore no rotation to balance the huge
force of (7.6.72) in the clusters.
Thus, the new cosmology theorem, Theorem 7.27, suggests that gravity should be
asymptotically repulsive. Theorem 7.30 offers a solid theoretic foundation for the prop-
erty, based on which we derive the simplified gravitational force formula (7.6.63).
Now we consider the constant k1 in (7.6.63). Due to the astronomical fact that no grav-
itational force between clusters of galaxies, we have

F(r) = 0, r = the average distance between galactic clusters.

By (7.6.71), we take r
2
r= × 1020km. (7.6.73)
5
Then we deduce from (7.6.63) that
k0 1
k1 r − − 2 = 0,
r r
which, by (7.6.70) and (7.6.73), leads to

k1 = 10−57 km−3 . (7.6.74)

In summary, we conclude that the dark matter and dark energy are essentially gravita-
tional effect generated by the gravitational potential field g µν , its dual field Φ µ and their
nonlinear interactions. They can be regarded as the gravitational field energy caused by g µν
and Φµ .
Bibliography

Davis, R., Harmer, D. S., and Hoffman, K. C. (1968). Search for neutrinos from the sun. Phys. Rev.
Lett. 20, 1205–1209, 10.1103/Phys. Rev. Lett., 20.1205.

de Broglie, L. (1924). Recherches sur la théorie des quanta, Thesis Paris.

Einstein, A. (1905). Über einen die Erzeugung und Verwandlung des Lichtes betreffenden heuristis-
chen Gesichtspunkt. Annalen der Physik, 322, 132–148, doi:10.1002/andp.19053220607.

Einstein, A. (1915). Die feldgleichungen der gravitation. Sitzungsberichte der Preussischen Akademie
der Wissenschaften zu Berlin, 844–847.

Einstein, A. (1916). The foundation of the general theory of relativity. Annalen der Physik, 354 (7),
769–822.

Englert, F. and Brout, R. (1964). Broken symmetry and the mass of gauge vector mesons. Phys. Rev.
Lett., 13 (9), 321–323.

Greiner, W. (2000). Relativistic Quantum Mechanics: Ware Equation. Berlin, Heidelberg, New York:
Springer.

Griffiths, D. (2008). Introduction to Elementary Particles. New York: Wiley-Vch.

Gross, D. J. and Wilczek, F. (1973). Ultraviolet behavior of non-abelian gauge theories. Phys. Rev.
Lett., 30, 1343–1346, 10.1103/Phys. Rev. Lett., 30.1343.

Guralnik, G., Hagen, C. R., and Kibble, T. W. B. (1964). Global conservation laws and massless
particles. Phys. Rev. Lett., 13 (20), 585–587.

Halzen, F. and Martin, A. D. (1984). Quarks and Leptons: An Introductory Course in Modern Particle
Physics. New York: John Wiley and Sons.

Harrison, E. (2000). Cosmology, the Science of the Universe. 2nd ed. Cambridge University Press.

Hernandez, M., Ma, T., and Wang, S. (2015). Theory of dark energy and dark matter. To appear.

Higgs, P. W. (1964). Broken symmetries and the masses of gauge bosons. Phys. Rev. Lett., 13, 508–
509.

Kaku, M. (1993). Quantum Field Theory, A Modern Introduction. London: Oxford University Press.

Kane, G. (1987). Modern Elementary Particle Physics, Vol. 2. Reading: Addison-Wesley.

512 Bibliography

Klein, O. (1938). New Theories in Physics, 77 (Intern. Inst. of Intellectural Co-operation, League of
nations).

Kutner, M. L. (2003). Astronomy: A Physical Perspective. Cambridge University Press.

Landau, L. (1957). On the conservation laws for weak interactions. Nuclear Physics, 3, 1, 127–131,
doi:[Link]

Landau, L. D. and Lifshitz, E. M. (1975). The Classical Theory of Fields. Oxford: Pergamon Press.
Translated from the Russian by Morton Hamermesh.

Lee, T. D. and Yang, C. N. (1956). Question of parity conservation in weak interactions. Phys. Rev.,
104, 254–258, doi:10.1103/PhysRev.104.254.

Lee, T. D. and Yang, C. N. (1957). Parity nonconservation and a two-component theory of the neutrino.
Phys. Rev., 105, 1671.

Ma, T. (2010). Manifold Topology. Beijing: Science Press in Chinese.

Ma, T. (2011). Theory and Methods of Partial Differential Equations. Beijing: Science Press in Chi-
nese.

Ma, T. (2012). Physical World From the Mathematical Point of View: Geometrical Analysis, Gravi-
tational Field, and Relativity. Beijing: Science Press in Chinese.

Ma, T. (2014). Physical World From the Mathematical Point of View: Theory of Elementary Particles
and Unified Fields. Beijing: Science Press in Chinese.

Ma, T. and Wang, S. (2005). Bifurcation theory and applications, World Scientific Series on Nonlinear
Science. Series A: Monographs and Treatises, Vol. 53 (World Scientific Publishing Co. Pte. Ltd.,
Hackensack, NJ),.

Ma, T. and Wang, S. (2013a). Duality theory of weak interaction, IU ISCAM Preprint Series, #1302:
[Link]

Ma, T. and Wang, S. (2013b). Phase Transition Dynamics. Springer-Verlag.

Ma, T. and Wang, S. (2013c). Structure and stability of matter, IU ISCAM Preprint Series, #1303:
[Link]

Ma, T. and Wang, S. (2014a). Astrophysical dynamics and cosmology. Journal of Mathematical Study,
47:4, 305–378.

Ma, T. and Wang, S. (2014b). Color algebra in quantum chromodynamics, Isaac Newton Institute
Preprint # NI14001: [Link]

Ma, T. and Wang, S. (2014c). Duality theory of strong interaction. Electronic Journal of Theoret-
ical Physics 11:31, 101–124, iU ISCAM Preprint Series, #1301: [Link]
~iscam/preprint/[Link].

Ma, T. and Wang, S. (2014d). Field theory for multi-particle system, Isaac Newton Institute Preprint
# NI14057: [Link]
Bibliography 513

Ma, T. and Wang, S. (2014e). Gravitational field equations and theory of dark matter and dark energy,
Discrete and Continuous Dynamical Systems, Ser. A 34:2, 335–366, preprint, [Link]
org/abs/1206.5078.

Ma, T. and Wang, S. (2014f). On solar neutrino problem, Isaac Newton Institute Preprint # NI14058:
[Link]

Ma, T. and Wang, S. (2014g). Spectral theory of differential operators and energy levels of subatomic
particles, Isaac Newton Institute Preprint # NI14002: [Link]
preprints/[Link].

Ma, T. and Wang, S. (2014h). Unified field theory and principle of representation invariance. Applied
Mathematics and Optimization 69:3, 359–392, doi:0.1007/s00245-013-9226-0, preprint, http:
//[Link]/abs/1212.4893.

Ma, T. and Wang, S. (2015a). Unified field equations coupling four forces and principle of interaction
dynamics. Discrete and Continuous Dynamical Systems, Ser. A, 35:3, 1103–1138, preprint, http:
//[Link]/abs/1210.0448.

Ma, T. and Wang, S. (2015b). Weakton model of elementary particles and decay mechanisms.
Electronic Journal of Theoretical Physics 12:32, 139–178, iU ISCAM Preprint Series, #1304:
[Link]

Maki, Z., Nakagawa, M., and Sakata, S. (1962). Remarks on the unified model of elementary particles.
Progress of Theoretical Physics 28, 870–880.

Mikheev, S. P. and Smirnov, A. Y. (1986). Resonant amplification of ν oscillation in matter and solar-
neutrino spectroscopy, IL Nuovo Cimento C, 9:1, 17–26.

Milgrom, M. (1983). A modification of the Newtonian dynamics as a possible alternative to the hidden
mass hypothesis. Astrophysical Journal 270, 365–370, doi:10.1086/161130.

Milgrom, M. (2014). MOND laws of galactic dynamics. MNRAS, 437, 2531–2541, doi:10.1093/
mnras/stt2066, eprint 1212.2568.

Nambu, Y. (1960). Quasi-particles and gauge invariance in the theory of superconductivity. Phys. Rev.,
117, 648–663, doi:10.1103/[Link].117.648.

Nambu, Y. (2008). Spontaneous symmetry breaking in particle physics: a case of cross fertilization,
Slides for his Nobel Lecture.

Nambu, Y. and Jona-Lasinio, G. (1961a). Dynamical model of elementary particles based on an anal-
ogy with superconductivity. Phys. Rev. I, 122, 345–358, doi:10.1103/PhysRev.122.345.

Nambu, Y. and Jona-Lasinio, G. (1961b). Dynamical model of elementary particles based on an anal-
ogy with superconductivity. Phys. Rev. II, 124, 246–254, doi:10.1103/PhysRev.124.246.

Perlmutter, S. and et al. (1999). Measurements of ω and λ from 42 high-redshift supernovae. Astro-
phys. J., 517, 565–586.

Planck, M. (1901). Ueber das Gesetz der Energieverteilung im Normalspectrum. Annalen der Physik,
309, 553–563, doi:10.1002/andp.19013090310.
514 Bibliography

Politzer, H. D. (1973). Reliable perturbative results for strong interactions? Phys. Rev. Lett., 30, 1346–
1349, doi:10.1103/PhysRevLett.30.1346.

Pontecorvo, B. (1957). Mesonium and anti-mesonium. [Link], 6, 429.

Pontecorvo, B. (1968). Neutrino experiments and the problem of conservation of leptonic charge.
[Link], 26, 984.

Popławski, N. (2012). Nonsingular, big-bounce cosmology from spinor-torsion coupling. Phys. Rev.
D, 85, 107502, doi:10.1103/PhysRevD.85.107502.

Quigg, C. (2013). Gauge Theories of the Strong, Weak, and Electromagnetic Interactions, 2nd ed.
Princeton Unversity Press.

Riess, A. G. and et al. (1989). Observational evidence from supernovae for an accelerating universe
and a cosmological constant. Astron. J., 116, 1009–1038.

Roos, M. (2003). Introduction to Cosmology, 3rd Ed.. Wiley.

Rubin, V. and Ford, J. W. K. (1970). Rotation of the andromeda nebula from a spectroscopic survey
of emission regions. Astrophysical Journal 159, 379–404.

Sakurai, J. J. (1994). Modern Quantum Mechanics, revised ed. Addison-Wesley Publishing Company,
Inc.

Salam, A. (1957). On parity conservation and neutrino mass, Nuovo Cimento, 5, 299–301.

Sokolov, A. A., Loskutov, Y. M., and Ternov, I. M. (1966). Quantum Mechanics. New York: Holt,
Rinehart & Winston, Inc.

Weisskopf, V. (1972). Physics in the Twentieth Century: Selected Essays. Cambridge, Massachusetts:
MIT Press.

Weyl, H. (1912). Das asymptotische Verteilungsgesetz der Eigenwerte linearer partieller Differential-
gleichungen (mit einer Anwendung auf die Theorie der Hohlraumstrahlung). Math. Ann. 71, 4,
441–479, doi:10.1007/BF01456804.

Weyl, H. (1919). Raum, Zeit, Materie. Verlag von Julius Springer.

Weyl, H. (1929). Electron and gravitation. Z. Phys., 56, 330–352.

Wolfenstein, L. (1978). Neutrino oscillations in matter. Phys. Rev. D, 17, 2369–2374, doi:10.1103/
PhysRevD.17.2369.

Wu, C. S., Ambler, E., Hayward, R. W., et al. (1957). Experimental test of parity conservation in beta
decay. Phys. Rev., 105, 1413–1415, doi:10.1103/PhysRev.105.1413.

Yang, C. N. and Mills, R. (1954). Conservation of isotopic spin and isotopic gauge invariance. Phys.
Rev. 96, 191–195.

Zwicky, F. (1937). On the masses of nebulae and of clusters of nebulae. Astrophysical Journal, 86,
217–246.
Index

Symbols e, the negative electric charge, 57, 180

a
fbc , structure constants for Gell-Mann Ma-
D0 -decay, 307 trices, 163
De , gradient operator, 202 gs , strong charge, 180
Dg , gradient operator, 202 gw , weak charge, 180
Ds , gradient operator, 202 mH , mass of φ w , 205
Dw , gradient operator, 202 mπ , mass of φ s , 205
G-parity, 260 ml , 201
H k Sobolev space, 125 mq , 201
L p spaces, 123 ms , 202
SU(N), 269 mw , 202
w∗ -color algebra, 320
SU(N) gauge transformation, 118
4-D gradient operator, 51
SU(N) representation, 73
m , solar mass, 412
SU(N) representation invariance, 156
SU(N) tensor, 161 A
Sn , n-sphere, 108
T M = M ⊗ p Tp M , tangent bundle, 113 accretion, 454
U(1) gauge invariance, 70 active galaxy, 443
W ± bosons, 211 AGN, 451
W k,p Sobolev space, 124 angular momentum conservation, 265
Z boson, 211 angular momentum rule, 342, 344
∆++ -decay, 306 arc-length, 111
Gab (A), SU(N) tensor, 161 Archytas, 463
Γrkl , Levi-Civita connection, 147 astrophysical fluid dynamics, 396, 408
M ⊗ p E N , vector bundle, 113 asymptotic flatness, 479
asymptotic freedom, 227
β -decay, 261, 305
atomic spectrum, 389
δ -factor, 436, 438
λk , Gell-Mann matrices, 163, 204 B
M ⊗ p Cn , complex bundle, 11
ν -mediator, 295 baryon, 256
ρw , radius of w∗ -weakton, 219 spectral equations, 370
σk , Pauli matrices, 163, 204 baryon decay, 263
{M , gµν }, Riemannian manifold, 11 baryon number, 259
516 Index

BEC conservation
Bose-Einstein condensation, 104 of weakton number, 302
Betti number, 141 conservation laws, 89, 264
Bianchi identity, 192, 474 conservative systems
Big-Bang, 457 dynamics of, 96
black hole, 402 control parameter, 418
existence of, 402 cosmic edge problem, 463
black holes, 432 cosmic radius, 466
geometric realization, 432 cosmological constant, 70
Blackhole Theorem, 395, 434 Cosmology Theorem, 470
Bohr atomic model, 166 Coulomb potential, 207
Bose-Einstein condensation covariant derivative, 63, 65, 66
Hamiltonian system for, 104 on a vector bundle, 119
boson, 254 critical δ -factor, 438
bound state, 367 current density, 51
bremsstrahlung, 310 curvature tensor, 67
mechanism of, 310
D
C
dark energy, 23, 24, 192, 471, 473
Chandrasekhar limit, 436 theory of, 476
charged Higgs, 211 dark matter, 23, 24, 192, 471
charged lepton, 366 dark matter and dark energy
mediator cloud, 324 the nature of, 485
weakton constituent, 294 de Broglie relation, 57
CKM matrix, 352 decay, 261, 304
classical electrodynamics, 82 decay type, 303
cluster of galaxies, 412 decoupling
CMB problem, 465 of the unified field theory, 209
color algebra, 316, 317 deep inelastic scattering, 227
color charge, 314 derivative on T M , 121
color index, 315, 317 derivative on T ∗ M , 121
formula for hadron, 318 derivative on Trk M , 122
color quantum number, 311 differential operators, 128
conjugate representation, 272 in spherical coordinates, 403
connection, 119 Dirac equations, 45, 57, 336
for GL(n) group, 121 Hamiltonian system for, 103
for SU(N) group, 120 Dirac matrices, 58, 334
for Lorentz group, 120 Dirac spinor, 58
Index 517

Dirac spinor fields, 87 electromagnetic potential, 50, 407

dissipative systems, 418 electromagnetism, 14, 181
divergence operator, 129, 397, 405 electron degeneracy pressure, 436
dual gluon, 313 electron structure, 309
dual gluons, 211, 216 electron-positron pair
dual gravitational potential, 486 creation and annihilation, 300
dual graviton, 211, 212 electroweak theory, 245
dual photon, 211, 214 problems in, 249
duality elementary particle
for the weak interaction, 218 strong interaction potential of,
of field particles, 210 218
of interacting forces, 211 elliptic operator
of mediators, 210 negative eigenvalues of, 169
duality of fundamental interactions, 209 elliptical galaxy, 442, 446
dynamic transition theory, 418, 450 energy conservation, 264, 303
classification, 419 energy level, 358, 367
energy levels, 166
E energy operator, 333
energy rule, 303
Eötvös, 63 energy spectrum, 371
Eddington coordinate transformation, 434 discreteness, 373
Eightfold Way, 267 energy-momentum conservation, 406,
Einstein energymomentum relation, 56 407
Einstein equation energy-momentum tensor, 400
non-well-posedness of, 193 of idealized fluid, 400
Einstein equations energy-momentum vector, 51
non well-posedness, 194 equation of state, 406, 408, 428
Einstein gravitational field equations, 35, essence of physics, 34
68, 398, 400, 429 expansion of the Universe, 456
Einstein mass-energy formula, 53
Einstein’s theory of general relativity, 60 F
Einstein-Hilbert action, 67
Einstein-Hilbert functional, 67, 68 fermion, 254
derivative operator, 147 field equations
elastic scatterring, 227 for the strong interaction, 216
elastic waves, 81 for the weak interaction, 217
electric charge, 180, 257 field theory
electric field, 54 multi-particle system, 377
electromagnetic force, 180, 215 first Betti number, 141
518 Index

flatness problem, 463 gluon, 312, 313

fluid dynamic equations, 396, 407 gluon radiation, 314
momentum representation of, 406 color transformation, 319
four forces, 180 gluons, 211, 216
four interactions, 180 gradient operator, 129, 405
Fredholm Alternative Theorem, 133 gravitational field equations, 68, 406, 425
Friedmann cosmology, 459 based on PID, 211
fundamental representation, 271 gravitational force, 180, 212
gravitational force formula, 212
G gravitational interaction formula, 477, 483
gravitational mass, 63
G-parity, 260
gravitational potential, 45, 407
G. Lemaı̂tre, 457
graviton, 211, 212
galaxies, 442
gravity, 181
galaxy, 412
asymptotic repulsion, 478
galaxy cluster, 445
guiding principle
galaxy types, 442
of theoretical physics, 34
Galilean invariance, 44
gauge invariance, 44, 70, 78 H
gauge symmetry
physical implication of, 188 Hölder space, 125
gauge symmetry breaking, 188 hadron, 254, 256, 283, 326
principle of, 189 gluon cloud, 326
gauge transformation, 118 Hamiltonian dynamics, 80, 97
SU(2), 187 Hamiltonian system
gauge-fixing problem, 208 energy-conservation of, 98
non well-posedness of, 209 for Bose-Einstein condensation,
Gauss formula, 131, 132 104
Gell-Mann matrices, 163 for Dirac equations, 103
Gell-Mann-Zweig for Klein-Gordon equation, 104
quark model, 284 for Schrödinger equation, 102
general invariants, 64 for Weyl equations, 102
general tensor, 63 Hamiltonian systems
generator representation in classical mechanics, 93
of SU(N), 72 heat conduction, 407
geometric interaction mechanism, 182 Heisenberg uncertainty relation, 340
geometry of unified fields, 185 Helmholtz decomposition, 135
Ginzburg-Landau free energy, 199 Higgs mechanism, 196
globular universe, 462 Hodge theory, 142
Index 519

Homestake experiments, 348 latitudinal circulation, 454

horizon problem, 463 Lemaı̂tre cosmology, 460
Hubble, 443 lepton, 254
Hubble Law, 456 lepton decay, 262, 306
hypercharge, 260 lepton numbers, 258
Leray decomposition, 136
I Levi-Civita connection, 62, 147, 396
Lie group, 270
Indc , color index, 317
irreducible representation, 270
inertial mass, 63
lifetime, 257
interaction potential pressure, 437
linear elliptic equation, 133
irreducible representation, 269
linear hyperbolic equation, 134
computation, 278
Lipschitz space, 125
of SU(N), 273
Lorentz force, 207
isospin, 260
Lorentz invariance, 44, 46
J Lorentz invariants, 51
Lorentz tensor, 47, 50
Jeans theory, 440 Lorentz transformation, 46
jets, 451, 454 Lorentz transformation group, 49

K M

KamLAND, 349
magnetic field, 54
KG equation, 56
magnetic potential, 207
Klein-Gordon equation, 56, 335
main driving force, 414
Hamiltonian system for, 104
pressure gradient, 415
Klein-Gordon fields, 87
relativistic effect, 415
Kruskal coordinate transformation, 434
stellar dynamics, 414
L thermal expansion force, 415
viscous friction, 415
Lagrangian action, 68 main-sequence stars, 413, 420
Lagrangian actions manifold with boundary, 143
in quantum mechanics, 86 mass, 257
Lagrangian dynamics, 80 mass charge, 180
Lagrangian for dynamics of charged par- mass generation, 196, 295
ticles, 84 mass of earth’s black hole core, 442
Laplace operator, 130 mass of solar black hole core, 441
Laplace-Beltrami operator, 131, 396, massive bound state, 362
404 massless bound state, 364
520 Index

Maxwell equations, 214 neutron star and pulsar, 414, 421

modified, 214 Newton cosmology, 458
Maxwell equations for electromagnetic fields, Newtonian dynamics, 80
54 Newtonian Gravitational Law, 35
mechanism of supernova explosion, 429 Newtonian gravitational potential, 398
mediator, 254, 255 Newtonian Second Law, 406
energy spectrum, 371 Noether Theorem, 90
weakton constituents, 294
mediator cloud, 311, 324 O
of quark, 325
observable, 332
structure of, 311
Oort, J., 472
mediator decay, 263
Oppenheimer limit, 436
meson, 256
origin of stars and galaxies, 439
spectral equations, 370
Orthogonal Decomposition Theorem
meson decay, 264
for tensor fields, 136
Milky Way galaxy, 444
Minkowski space, 47 P
mixed decay, 308
mixed type decay, 304 parity, 259
mixing matrix, 352 parity problem, 355
MNS matrix, 352 particle decay, 261
modified QED model, 213 particle radiation, 262
modified Yukawa potential, 229 particle scattering, 261
momentum conservation, 265 particle transition, 260
momentum operator, 333 Pauli exclusion principle, 311, 340, 342
MSW effect, 354 Pauli matrices, 163
multi-particle system, 377 Penzias, A., 465
PHD
N abbreviation for principle of Hami-
ltonian dynamics, 93
Navier-Stokes equations, 396 for electrodynamics, 100
in spherical coordinates, 405 photon, 211
negative pressure, 466, 485 PID
neutral Higgs, 211 principle of interaction dynamics, 44
neutrino PID cosmological model, 473
non-oscillation mechanism, 357 PID field equations, 190
neutrino mass, 352 PID gravitational field equations, 211
neutrino oscillation, 350 PID gravitational interaction formula, 477
neutron degeneracy pressure, 436 PLD, 208
Index 521

abbreviation for Principle of Lagrangian quantum physics, 332

Dynamics, 79 quantum system, 332
principle of Lagrangian dynamics, 36 quark, 254, 255, 366
quantum dynamics, 337 mediator cloud, 325
position operator, 333 weakton constituents, 294
position vector, 50 quark confinement, 226
Prandtl number, 419 quark decay, 263
pressure gradient, 415 quark model, 267, 284
PRI quark pari creation, 305
abbreviation of principle of represen-
tation invariance, 165 R
abbreviation for principle of repre-
sentation invariance, 191 radiation, 262
principle of equivalence, 36, 61, 62 radio-galaxy, 443
principle of gauge invariance, 78 radius of earth’s black hole core, 442
principle of general relativity, 36, 44, 60 radius of solar black hole core, 442
principle of Hamiltonian dynamics Rayleigh number, 451
abbreviated as PHD, 93 Rayleigh-Bénard convection, 451
principle of interaction dynamics, 190 Rayleigh-Bénard instability, 453
abbreviated as PID, 44 Rayleign-Bénard convection, 409
Principle of Lagrangian Dynamics, 37, 79 red giant, 414, 421
abbreviated asPLD, 79 redshift problem, 463
principle of Lagrangian dynamics, 208 relativistic effect, 415
abbreviated as PLD, 36 relativistic invariants, 51
principle of Lorentz invariance, 46 relativistic mechanics, 52
principle of representation invariance, 191 relativistic quantum mechanics, 56
abbreviated as PRI, 165 basic postulates, 56
principle of symmetry-breaking, 44, 45 representation invariance, 156
proper distance, 432 of gauge theory, 165
proper time, 432 representation transformation
SU(2), 188
Q SU(3), 188
Ricci curvature, 396
QCD, 216, 227, 312 Ricci curvature tensor, 147
QED, 213 Ricci tensor, 67
quantum chromodynamics, 216 Riemannian geometry, 109
quantum Hamiltonian system, 101 Riemannian manifold, 11, 108, 396
quantum mechanics Riemannian metric, 109, 110
Lagrangian actions, 86 Rubin rotational curve, 471
522 Index

S luminosity of, 412

mass of, 412
Sakata model parameter relation of, 412
of hadrons, 283 radius of, 412
Sakata’s explanation, 277 surface temperature of, 412
scalar curvature, 68 starburst, 443
scalar gluons, 216 stellar atmosphere, 408
scalar photon, 314 stellar atmospheric circulation, 422
scalar potential theorem, 154 stellar cluster, 412
scattering, 261, 304 stellar dynamics, 414
Schrödinger equation, 86, 335 main driving force for, 414
Hamiltonian system for, 102 stellar interior circulation, 418
Schwarzschild metric, 397, 400, 432 strange number, 260
Schwarzschild solution, 36 strong charge, 180
Seyfert galaxy, 443 strong decay, 306
Sobolev embedding theorem, 126 strong force, 180, 219
Sobolev space of tensor fields, 123 strong force potential, 207
solar neutrino problem, 346 strong interacting decay, 303
solar neutrinos strong interaction, 14, 182
discrepancy of, 346 bound energy, 226
space curved energy, 485 field equations for the, 216
spectral equations short-range nature of, 233
baryon, 370 strong interaction potential, 219
meson, 370 layered formula of, 223
Spectral Theorem of Elliptic Operators, strong interaction potentials, 218
168 strong magnetic potential, 207
spectral theory, 166 subatomic particle
spherical universe, 466 energy level, 358
spherically symmetric gravitational mediator cloud, 324
field, 397 Sudbury Neutrino Observatory, 349
spin, 258 Super-K experiment, 349
spin operator, 333 superconductivity, 199
spinor transformation, 118 supernova, 414, 429
spiral galaxy, 445, 448 symmetry-breaking, 45
spontaneous gauge symmetry breaking, 45
spontaneous symmetry-breaking T
PID mechanism of, 241
stable manifold, 480 tangent bundle, 113
star, 412 tangent space, 112
Index 523

Taylor instability, 452 layered formulas, 236

the Universe weak force potential, 207
age of, 457 weak interacting decay, 303
geometric structure of, 458 weak interaction, 14, 181
topological structure, 458 duality of the, 218
theory of dark matter and dark energy,
field equations for the, 217
471
theory, 234
thermal expansion force, 415
weak interaction potential, 234
TOV metric, 397, 400, 403
weak magnetic potential, 207
U weak solution, 133
weakly differentiable functions, 123
unification
weakton black hole, 439
the Einstein, 183
weakton color algebra, 320
based on PID, PRI and PSB, 184
weakton confinement, 295
unified field model, 200
weakton constituents, 294
coupling matter fields, 385
unified field theory, 179 weakton exchange
decoupling of the, 209 between leptons and mediators,
301
V between quarks and mediators,
302
variable star, 413
in mediators, 300
variation with divA -free constraint, 144,
weakton interaction potential, 437
150, 152
weakton model, 287
variational principle, 144
weakton number, 302
vector bundle, 113
connection on a, 119 Weyl asymptotic relation, 171
covariant derivative on a, 119 Weyl equations, 57, 335
linear transformations on, 116 Hamiltonian system for, 102
vector photon, 314 Weyl operator
Vierbein tensor, 188 spectrum of, 174
viscous friction, 415 white dwarf, 414
Wilson Microwave Anisotropy Probe
W
WMAP, 463
wave function, 332 Wilson, R., 465
weak charge, 180 WMAP
weak decay, 304 Wilson Microwave Anisotropy Probe,
weak force, 180, 218, 234 463
524 Index

Yang-Mills action, 74
Yang-Mills functional, 74
derivative of the, 146
Young tableaux, 278
Yukawa interaction mechanism, 182
Yukawa potential, 229
modified, 229, 230