The Two-Body Problem

3.1
The two-body problem
• The two-body problem: two point objects in 3D
interacting with each other (closed system)

• Interaction between the objects depends only on the

distance between them

• The number of degrees of freedom: 6

• Phase space dimensions: 12

 3.1
The two-body problem
• The Lagrangian of the system in Cartesian
coordinates:
3 2 m j (rij ) 2  3 
L   V   (ri1  ri 2 ) 
2

2  
i 1 j 1  i 1 

 
 2  2
m1 (r1 ) m2 (r2 )   2
  V (r1  r2 )
2 2

• It is a very non-trivial problem if we try to deal with

the Lagrangian in this format: all the 6 independent
coordinates are entangled in the potential function

• Let us look for a different configuration space

 3.1
New generalized coordinates
• Let us introduce new system of coordinates:
  
   m1r1  m2 r2
r  r1  r2 ; R 
m1  m2
• R – center of mass vector

  m2 r
• Then r1  R 
m1  m2
 
 m1r
r2  R 
m1  m2
• And

 2 
 2
 2   m2 r   2   m1r 
(r1 )   R   ; (r2 )   R 
 


 m1  m 2   m1  m 2 
 3.1
New generalized coordinates
• The Lagrangian in the new coordinates:

 
 2  2
m1 (r1 ) m2 (r2 )   2
L   V (r1  r2 )
2 2
 2 
 2
  m2 r    m1r 
m1  R   m2  R  
  m1  m2 
2
  m1  m2 
2

 V (r ) 
 2

 
 2     2  
m1m2 Rr m1 m2 r m1m2 Rr
2
m1 ( R) m2 ( R)
    
2 m1  m2 2(m1  m2 ) 2
2 m1  m2
   2  2  2
(m1  m2 )( R)
V  r 

V r  
m2 m1r  m1m2 (r )
 
2(m1  m2 ) 2
2 2(m1  m2 )
 3.1
New generalized coordinates
• The Lagrangian in the new coordinates:
 2  2
(m1  m2 )( R)
V  r 
m1m2 (r ) 
L 
2 2(m1  m2 )
• The center of mass coordinates are cyclic!

• Three Euler-Lagrange equations for them can be

solved immediately
L d  L  L
    const  (m1  m2 ) Ri  Pi
Ri dt  Ri  R i
• Total momentum of the system is conserved: three
integrals of motion
 3.1
New generalized coordinates
• The Lagrangian in the new coordinates:
 2  2
(m1  m2 )( R)
V  r 
m1m2 (r ) 
L 
2 2(m1  m2 )
 2  2
V  r 
( P) m1m2 (r ) 
 
2(m1  m2 ) 2(m1  m2 )
• Let’s re-gauge the Lagrangian
 2  2
V  r 
( P) m1m2 (r ) 
L'  L  
2(m1  m2 ) 2(m1  m2 )

• Constant term
 3.1
New generalized coordinates
• The re-gauged Lagrangian:
 2  2
V r   V  r 
m1m2 (r )  m(r )  m1m2
L'  m
2(m1  m2 ) 2 m1  m2
• We reduced the two-body problem to a one-body
problem in a central potential (potential that depends
only on the distance from the origin)

• m: reduced mass

• The number of degrees of freedom: 3

• Phase space dimensions: 6

 8.1
Spherical coordinates
 2
 
2 2 2
  
V  r  
 m( r r rz )
 V  rx  ry  rz 
m(r )
L' 
x y 2 2 2

2 2  
• Central potential is spherically symmetric

• It is convenient to work in spherical coordinates

rx  r sin  cos  ; ry  r sin  sin  ; rz  r cos 

r  
2 2 
m(r  r   r sin   )
 2 2 2 2

• Then L'   V (r )
2
 3.2
Spherical coordinates
L'  m(r 2  r 2 2  r 2 sin 2  2 ) / 2  V (r )
• The Euler-Lagrange equation for φ
d  L'  L' L'   p  const
    r 2
m sin 2
  
dt      
 p 
• The φ coordinate is cyclic  2
r m sin 2 
• Since the system is spherically symmetric, we have
a freedom of choosing the reference frame

• We chose it as follows: the initial velocity vector

belongs to a plane φ = const

• Then 0  0  p  0   0

 3.2
Spherical coordinates
L'  m(r 2  r 2 2  r 2 sin 2  2 ) / 2  V (r )
• The Euler-Lagrange equation for θ
d  L'  L'
 
dt    
• The θ coordinate is also cyclic

• Momentum conjugate to the θ coordinate

L
p    r 2 m  const

• Angular momentum in the plane of motion relative
to the origin is conserved

r 2 m  rmr  rmv  rp  const

 3.2
Spherical coordinates 8.1

L'  m(r 2  r 2 2  r 2 sin 2  2 ) / 2  V (r )

• The Euler-Lagrange equation for r
d  L'  L'  V (r )
  m r  rm 2

dt  r  r r
• Momentum conjugate to the r coordinate
L'
pr   mr
r
• Now we can write a Hamiltonian
1  p
2

H  pr r  p  p  L'   pr  2   V ( r )
2 
2m  r 
 3.2
The effective potential
2 2
pr p
H  2
 V (r ) p  const
2m 2mr 2
p
• The effective potential Veff ( r )  2
 V (r )
2mr
• The Hamiltonian effectively depends only on 1
coordinate now

• We reduced the two-body problem to a 1D problem

of a particle with a reduced mass m in the effective
potential

• The number of degrees of freedom: 1

• Phase space dimensions: 2

 3.2
The effective potential
2 2
pr p
H  2
 V (r )  E
2m 2mr
• Hamilton equations of motion:
H pr  H p
2
dV
r   p r    

pr m r mr 3
dr
 H p H
  p    0
p mr 2

dH H
 0 H  const  E
dt t
 p
2

rm  pr  2m E  2
 V (r ) 
 2mr 
 3.2
The orbit equation dr 3.5

dt 
2 
2
r  E 
p
  2  p
2

 V ( r )  E   V ( r ) 
m  
2
m 2mr  2mr 2

• On the other hand
 p
 2 mr 2 d dr
mr dt  mr d 
2

p 2  p
2

p E   V ( r ) 
• Orbit equation m  2mr 2 

r
p dr
  0  
 p
2

2m E   V (r ) 
r0 2
r 2
 2mr 
 3.5
The orbit equation
r
p dr
  0  
 p
2

2m E  
r0
r 2
2
 V ( r ) 
 2 mr 
• The orbit equation can be integrated for potentials
with the power dependence on the distance
V (r )  ar n

If n = 2, - 1, - 2, the integral can be expressed in

trigonometric functions

• If n = 6, 4, 1, - 3, - 4, - 6, the integral can be

expressed in elliptic functions
 3.5
The orbit equation
• From Hamilton’s equations of motion:
2
p pr p dV

 2 r  pr  rm
 p r  3
  rm
mr m mr dr
2
1 p 1 p 1 dV
 r  2 3 
dt mr d 2
m r m dr
d  dr  p d  p dr 
2
p 1 dV
   2  2 3 
dt  dt  mr d  mr d  m r
2
m dr
 1 dr  1 mr dV
2
d
 2   2
d  r d  r p dr
• If the orbit is known, the potential can be calculated
 Example
• Restore a potential for a spiral orbit:

r  r0 e

 
2
d 1 dr 1 mr dV
 2   2
d  r d  r p dr

1  1
 e   
m r0e    2
dV 1 1 mr 2 dV
   2
r0 r0e p
2
dr r r p dr
2 2 2
dV 2 p 2 p dr p
 V   V0  V0 
dr mr 3 mr 3
mr 2
 3.6
Stable circular orbits
• For a circular orbit: r  const r  0
H pr  pr  0
• On the other hand r  
pr m
2
H
2
p dV p dV
p r     0 3

r mr 3
dr mr dr

• For the extremum of the effective potential

d  p 
2 2
p dV (r )
  V ( r )   0  3 0
dr  2mr 2 
 mr dr
• Extremum of the of the effective potential
corresponds to a circular orbit
 3.6
Stable circular orbits
• For a stable circular orbit, the second derivative of
the of the effective potential should be positive:
d  p 
2 2
2 2

2  2
 V (r )   0 3 p

d V (r )
0
dr  2mr  mr 4
dr 2

2 2 2
d V (r ) 3 p d V (r ) 3 dV p
2
dV
  
dr 2
mr 4
dr 2
r dr mr 3 dr
• For potentials with the power dependence on the
distance V (r )  ar
n

3n n 1
n(n  1)ar n2
  ar n 1  3 n  2
r
 10.5
Application of the Hamilton-Jacobi
theory
• The problem of spherically symmetric potential can
be neatly treated employing Hamilton-Jacobi theory
L'  m(r 2  r 2 2  r 2 sin 2  2 ) / 2  V (r )
H  pr r  p  p  L'

1  2 p
2
p
2

 pr  2  2 2   V ( r )
2m  r r sin  
• Then equation for Hamilton’s characteristic function


1   W 
2
1  W 
2
1  W  
2

   2   2 2    V  1
2m   r  r    r sin     
 10.5
Application of the Hamilton-Jacobi
theory

1   W 
2
1  W 
2
1  W  
2

   2   2 2    V  1
2m   r  r    r sin     
• Let us assume that the variables can be separated
W (r ,  ,  )  Wr (r )  W ( )  W ( )
• Then
1  2
 Wr '  2 W '  2 2 W '   V  1
2 1 2 1
2m  r r sin  
•The φ coordinate is cyclic, therefore

W '    W   
 10.5
Application of the Hamilton-Jacobi
theory
  
2

Wr '  2  W '  2   2mV  2m1

2 1 2

r  sin  
• The circled part should be constant, since it 2
contains only the θ dependence 
W '
2
  
2

• Then, finally sin 

2

 2
Wr '2
 2
 2mV  2m1
r
•The variables are completely separated!

• The resulting equation can be integrated in

quadratures and is equivalent to the orbit equation
 3.7
The Kepler problem
1
• Kepler potential: V (r )  r
• Mediating gravitational and electrostatic
Johannes Kepler
interactions (1571-1630)
k
• Attraction: V (r )   ; k  0
r
• Repulsion: k
V (r )  ; k  0
r
• Integral orbit equation:
p dr
  0  
 p
2
k
r 2
2m E  2
 
 2mr r 
 3.7
The Kepler problem
k
• Let us consider an attractive potential: V (r )  
r
p dr
  0   k 0
 k p 
2
1
r 2
2m E   
2 
u
 r 2 mr  r
du
  0   2mE
 2
2m( E  ku) / p  u
2 2
p
• Table integral:

du 2u   2mk
 2
   u  u 2
 arccos
  4
2 p
 3.7
The Kepler problem 2mk
2u  2
p
   0  arccos
2
up
2  2mk  2mE
1  2  4 2
 p 
   0  arccos km    p
2
2 Ep
1 2
2
mk 2
2 Ep up 1
1 1 2
cos(   0 )  u
mk km r

km  2  1
1  1 
2 Ep
cos(   ) 
p   r
2 2 0
mk

 3.7
The Kepler problem
km
1 km  2  C
 2 1 1
2 Ep
cos(    )  p
2

r p  mk 2 0

  2
2 Ep
1 e
 C 1  e cos(   0 ) 
1 2
mk
r
• We obtained an explicit expression for the orbit

• Depending on the values of C and e, the orbits can

assume qualitatively different shapes

• For a positive C (attraction), the shapes of the orbits

represent all possible conic sections
 3.3
Classification of Kepler’s orbits 3.7

• Effective potential for the attractive Kepler case:

2
p k
Veff (r )  2

2mr r
2
p k
Veff (0)  2
  
2m0 0
2
p k Veff (0)  
Veff (r )  0 2

2mr r
 p 2 
  Veff    0
2
r
p Veff  
2km r  2 km 
 3.3
Classification of Kepler’s orbits 3.7

• Effective potential for the attractive Kepler case:

p
2
k k 2m
Veff (r )  2
 Veff (rmin )   2
2mr r 2 p

• Minimum point of the effective potential

2 2
dVeff (r ) p k p
0  3  2  0 rmin 
dr mr r km
2 2 2 2
p k k m k m k m
Veff (rmin )  2
  2
 2  2
2mrmin rmin 2 p p 2 p
 3.3
Classification of Kepler’s orbits 3.7

• Effective potential for the attractive Kepler case:

p
2
k k 2m
Veff (r )  2
 Veff (rmin )   2
2mr r 2 p
 C 1  e cos(   0 ) 
1
r
• The simplest case e  0
1 1
• Circular orbit C r
r C
2
2 Ep
2 k m
e  1  0 E 2
 V eff ( rmin )
mk 2 2 p
 3.3
Classification of Kepler’s orbits 3.7

• Circular orbit e0

2 2
1 p k m
r  E 2
C km 2 p
• Circle is one of conic sections
 3.3
Classification of Kepler’s orbits 3.7

rx
 C 1  e cos(   0 )  cos(   0 ) 
• If e  0
1
r r
1  rx  1
 r 1  e   r  erx r   erx  h  erx
C  r C
rx  ry  r  h  2erx h  e rx
2 2 2 2 h
a
2 2

1 e 2
(1  e )rx  2erx h  ry  h
2 2 2 2

2 2 
2
 2 eh
1 e   eh   1  e  rx 0
2 2
   rx    r 1
2  1 e 2

 h   1  e   h  y
  h
 rx  rx 0   ry 
2 2
b
      1 1  e2
 a  b
 3.3
Classification of Kepler’s orbits 3.7
2
 rx  rx 0   ry 
2

      1 h
b
• If 0  e 1  a   b  1  e2
1 e 2 2
• Then is real and b is positive

• The orbit is an ellipse with its center shifted from

the origin by rx 0 and two semi-major axes a and b

• Ellipse is also a conic section

 3.3
Classification of Kepler’s orbits 3.7
2
 rx  rx 0   ry 
2

      1 V (r )   mk 2

 a  b eff min

2p 
2

2 2
2 Ep mk
0  e 1 0  1 2
1  2
E0
mk 2 p
Veff (rmin )  E  0
• Elliptic motion is limited by two
values of r Aphelion
Perihelion
 3.3
Classification of Kepler’s orbits 3.7
2
 rx  rx 0   ry 
2

      1
 a  b
h h
b a 2
 b 2
 a e
1 e 2
1 e 2
a
• This parameter is known as an eccentricity of an
ellipse
eh h
 rx 0 a  rx 0 
1 e 2
1 e
• For a constant energy, perihelion
is decreasing with increasing eccentricity
 3.3
Classification of Kepler’s orbits 3.7
22
 rxx  rxx 00   ryy 
22

       1 h

• If e 1  a   b'  1  e2
b

1 e 2 2
• Then is imaginary and b is negative

• The orbit is a hyperbola b  ib'

• Hyperbola is a conic section as well
 3.3
Classification of Kepler’s orbits 3.7
2
 rx  rx 0   ry 
2

      1
 a   b' 
2
e 1 2 Ep
1 2
1 E 0
mk
• Hyperbolic motion is limited by one
value of r - perihelion
 3.3
Classification of Kepler’s orbits 3.7

(1  e )rx  2erx h  ry  h
2 2 2 2

• Finally, if e  1
2
h r
rx   
y
2rx h  ry  h
2 2
• Then
2 2h
• The orbit is a parabola with its center shifted from
the origin by h / 2

• Parabola is a conic section

 3.3
Classification of Kepler’s orbits 3.7
2
h ry
rx   
2 2h
2
e 1 2 Ep
1 2
1 E 0
mk
• Parabolic motion is also limited by one value of r -
perihelion
Synopsys for orbit classification
 3.8
Motion in time dr
dt 
• In Kepler’s case: 2  p
2

r E   V ( r ) 
m  
dr
t 2mr 2 
2 k
2
p
E  
r0

m  2mr 2 r  2
k 2 Ep
• Substitution: r  (1  1  2
cos )
2E mk
 a(1  e cos )
3
ma ma3
t
k 0  (1  e cos )d 
k
(  e sin  )

ma3
• For ψ = 2π (one period): t    2
k
 3.8
A bit of history: Kepler’s laws

Tycho Brahe/ Johannes Kepler

Tyge Ottesen (1571-1630)
Brahe de Knudstrup
(1546-1601)
• First law: “The planets move in elliptical orbits with
the sun at one focus”

 mk / 2 p  E  0
2 2
0  e 1
 3.8
A bit of history: Kepler’s laws
• Second law: “The radius vector to a planet sweeps
out area at a rate that is independent of its position in
the orbit” dA  d 
dA  r (rd ) / 2  r r /2
dt  dt 
2 
A    p  const
r
2 m
p  r m  const
2

• Third law: “The square of the period of an orbit is

proportional to the cube of the semi-major axis
length”
ma 3 4 2 ma3
 2 2 
k k
 10.8
The Kepler problem in action-angle
variables
• We consider periodic motion in the case of
attraction

1   W 
2
1  W 
2
1  W   k
2

   2   2 2     E
2m   r  r    r sin      r
• By definition, the action variables are

W W
J r   pr dr   dr J   p d   d
r 
W
J   p d   d

 10.8
The Kepler problem in action-angle
variables
W
• We found earlier for the two-body problem  

 W 
2
 2
 W 
2
 k  
2

     2    2m E    2
2

   sin   r   r r
• Therefore
W
J   d    d  2

W  2

J   d     2 d  2 (   )
2

 sin 
 10.8
The Kepler problem in action-angle
variables
W  k  
2
Jr   dr   2m E    2 dr
r  r r

k  J   J  
2
 2m
 2m E    dr   J  J  k
 r 4 r
2 2
E
2 2 mk 2
E 2  H
( J r  J  J )
• Frequencies
H H H 4 mk2 2
vr   v   v  
J r J  J  ( J r  J   J  ) 3
 10.8
The Kepler problem in action-angle
variables
H H H 4 mk
2 2
vr   v   v  
J r J  J  ( J r  J   J  ) 3

• Degenerate case: the frequencies for all three

variables coincide

• Hence it takes the same time for all three variables

to return to the same value – the same point on the
orbit

• Therefore, a completely degenerate solution

corresponds to a closed orbit

• We did not have to obtain an explicit expression for

the orbit to realize that it is closed
 10.8
The Kepler problem in action-angle
variables
• To lift the degeneracy, we can introduce another
canonical transformation employing the following
generating function:
F  ( w  w ) J1  ( w  wr ) J 2  wr J 3
• Then
w1  w  w ; w2  w  wr ; w3  wr
J1  J  ; J 2  J   J  ; J 3  J r  J   J 
• And the Hamiltonian

2 mk
2 2
2 mk 2 2
H  
( J r  J  J ) 2
J3
2
 10.8
The Kepler problem in action-angle
variables
2 2 mk 2
H  2
J3
• This Hamiltonian is cyclic in 5 variables, therefore
their 5 corresponding conjugates are conserved:
J1  J   2p  const
J 2  J   J  2p  2 ( p  p )  2p  const
2m w2  const
J 3  k  const w1  const
E
• We obtained 5 constants of motion for a system with
6 degrees of freedom (the last two can be shown to
be related to certain orbit parameters)
 3.10
Repulsive Kepler potential
k
• Let us consider a repulsive potential: V ( r )  ; k  0
r
• Total mechanical energy: E  T  V

T 0 V  0 E  0
p dr
• Orbit equation    
0  

2

k p
1
 C e cos(   0 )  1 r 2
2 m  E    
 r 2 mr 2 
r  
2
km 2 Ep
C 2 e  1 2
1
p mk
 3.10
Repulsive Kepler potential
 C e cos(   0 )  1
1 k e
r a  r 
C (1  e )
x 0 2
2 2 E
 rx  rx 0   ry 
2

      1 1
 a  b b
C e2 1
• This is a hyperbola

• Therefore, hyperbolic orbits correspond to the case

of a positive total energy for both the attractive and
the repulsive interactions
2
 rx  rx 0   ry 
2

      1
 a  b
• Scattering – the orbit is never closed for E > 0
 3.10
Scattering
2
 rx  rx 0   ry 
2

      1
 a  b
• Scattering angle:  s

• Impact parameter: b

 s    2   
2
1 a
 s  2  2 tan
b k
• On the other hand: a 
k 2E
 s  2 tan 1

2 Eb
 3.10
Scattering
• For a beam of (noninteracting) particles incident on
the scattering center, intensity (flux density) is the
number of particles crossing unit area normal to the
beam in unit time

• Scattering cross-section in a given direction is the

ratio of the number of particles scattered into a solid
angle per unit time to the incident intensity

dN 
d 
I
• Differential scattering cross-section: 
 3.10
Scattering cross-section
dN 
d  2 sin  s d s d 
I
• Conservation of the number of particles:

I (2 bdb)  dN  Id  I 2 sin  s d s

b db k k
  s  2 tan 1 b
sin  s d s 2 Eb s
2 E tan
2
 3.10
Scattering cross-section
 
b db k d  k 
   
sin  s d s s d  s 
2 E tan sin  s s
 2 E tan 
2  2 
2
k k
 b
2 4 s 
4 E sin 2 E tan s
2 2
• Rutherford scattering cross-section

• It is independent of the sign of k!

Ernest Rutherford
(1871 – 1937)
 3.10
Total scattering cross-section
• Total scattering cross-section

k 2 sin  s d s
 2

 T   d   
2 4 
0 0 4 E sin s
2
• It diverges because of the long-range nature of
Kepler’s potential

• All the particles in an incident beam of infinite lateral

extent will be scattered to some extent and must be
included in the total scattering cross-section
 3.11
Laboratory coordinates
• Let us recall the initial transformation of
coordinates:     
m2 r  m1r
r1  R  ; r2  R 
m1  m2 m1  m2
• Let us recall the re-gauged Lagrangian:
 2  2
V  r 
( P) m1m2 (r ) 
L'  L  
2(m1  m2 ) 2(m1  m2 )
• All the results obtained so far are in the re-gauged
center-of-mass system, in which the center of mass
is at rest:
 
 m2 r  m1r
r1  ; r2  
m1  m2 m1  m2
 3.11
Laboratory coordinates
• In the center-of-mass system, the scattering
process of two particles will look like this:

• Often, while the incident particle is

moving, the second one is initially
at rest

• We introduce the laboratory system of coordinates,

in which the center of mass is moving with a constant
velocity
   
 m2 r  m1r    m2 r  
 m1r
r1  ; r2   r1  R  ; r2  R 
m1  m2 m1  m2 m1  m2 m1  m2
 3.11
Laboratory coordinates
• In the laboratory system, the scattering process of
two particles will look like this:

• Let us introduce notations:

 
 m2 r  m1r
r1 '  ; r2 '  
m1  m2 m1  m2
     
• Then r  R  r ' ; r  R  r '
1 1 2 2

• Initially
 m1r10  m2 r20
R  const
m1  m2    
  m2 r   m1r
r1  R  ; r2  R 
m1  m2 m1  m2
 3.11
Laboratory coordinates
   
m1r10 
   m1r10  mr10 
R r1  R  r1 '   r1 '   r1 '
m1  m2 m1  m2 m2

 mv10 
v1   v1 '
m2
mv10 v1 sin   v1 ' sin  s
v1 cos   v1 ' cos  s 
m2
• Taking the ratio of these two equations:

sin  s
tan  
mv10
cos  s 
m2 v1 '
 3.11
Laboratory coordinates
• Now we can write the differential scattering cross
section expressed in laboratory system

• Conservation of the number of particles:

I ( s )2 sin  s d s  I L ()2 sin d

 L ( )   ( s )
sin  s d s 
  ( s )
1  2  cos  s   
2 3/ 2

sin  d 1   cos  s
sin  s
tan   mv10
mv10 
cos  s  m2 v1 '
m2 v1 '
 3.12
The three-body problem
• The Lagrangian of the system in Cartesian
coordinates:
 2  2  2
m1 (r1 ) m2 (r2 ) m3 (r3 )
L  
2 2 2
   2
    2
 
 V (r1  r2 )  V (r1  r3 )  V
  2
(r2  r3 ) 
• This problem has 9 independent coordinates
entangled by the 3 potential functions

• This Lagrangian cannot be re-gauged to a one-

particle Lagrangian

• No general explicit solution is known

 3.12
The three-body problem
• Constants of motion (not independent): total energy,
three components of the center of mass linear and
angular momenta

• If the three objects are allowed to move freely in 3D

the orbits become very complex and sensitive to
initial conditions

• Even after fixing positions of two particles and

letting the third particle move in a plane, the orbit still
can not be found explicitly