1

These lecture notes are just prepared to help the students of Phys 300. There are not original; they
contain lots of material from several different sources

CH.1 THE WAVE FUNCTION

In this Ch. we will start studying the mathematical description of the wave nature
of moving subatomic/microscopic bodies,

In classical physics, there exist two distinct entities: particles and waves.

Particles are localized. They have definite position energy and momentum.
They are confined in space.
Imagine a classical particle of mass m, moving under a force ⃗𝑭, at any instant.
⃗ (𝒕) of the particle at any given
Its state is specified by the position (trajectory) 𝒓
time. Once we know that, we can figure out the any other physical quantities
⃗ (𝒕) = 𝒅𝒓
relevant to the system, like the velocity 𝒗 ⃗ /𝒅𝒕 , the momentum
⃗ (𝒕) = 𝑚𝑑𝑟/𝑑𝑡 , the kinetic energy 𝑇 = 𝑚𝑣 2 /2 ,etc.
𝒑
One needs an equation, or a set of equations of motion that will predict the
state 𝑟(𝑡) . These are usually differential equations.
The basic equation of the classical mechanics for the motion of a particle under
⃗ 𝑉 is
the influence of a conservative force 𝐹 = −∇
𝑑2 𝑟(𝑡) 𝑑𝑝
𝑚 = ⃗𝑉=
−∇ ∶ the Newton’s equation
𝑑𝑡 2 𝑑𝑡
and this equation also implies energy conservation,
𝑝2
𝐸= +𝑉 (∗)
2𝑚

This equation cannot be derived from some higher principles. It was

invented by Newton and works very well.
 2

A wave, in contrast, is a disturbance spread over space. It is described by

a wave function 𝜓(𝑟, 𝑡) which characterizes the disturbance at the point 𝑟 at
time t. In the case of electromagnetic waves, 𝜓(𝑟, 𝑡) can be any component
of the electric field vector 𝐸⃗ .
𝜓 obeys a second-order wave equation in time given by
1 𝜕2𝜓
⃗∇ 𝜓 =
2
𝑐 2 𝜕𝑡 2
which describes waves propagating at the speed of light, c. Given 𝜓(𝑟, 0)
and 𝜓̇(𝑟, 0) one can get the wave function 𝜓(𝑟, 𝑡) for all future times by
solving the wave equation
In Quantum mechanics, there is a wave-particle duality for matter and
radiation.

 Davisson-Germer experiment demonstrates conclusively that quantum

particles, like electrons, possess wave-like properties, exhibiting interference
and diffraction effects. (wave property of matter)

Like in classical mechanics, in quantum mechanics now, we need a

mathematical description of the state of these quantum particles (or systems)
at any time and an equation of motion to predict the evolution of the state with
time.
Schrödinger invented the wave equation for the mathematical description of
the quantum systems.
Schrödinger equation has been extremely successful in comparison with
experiments.
 3

 Reasonable assumptions for the properties of the desired quantum

mechanical wave equation:
1. It must be consistent with the de Broglie-Einstein postulates
ℎ 𝐸
𝜆= and 𝜈 =
𝑝 ℎ
2. It must be consistent with the conservation of energy
𝑝2
𝐸= +𝑉
2𝑚
Now consider a free electron (as observed in the electron diffraction
experiments) and suppose that the state of such an electron is described by
a monochromatic plane wave of angular frequency 𝝎 =
2𝜋
𝟐𝝅𝝂, wavelength 𝜆 = , and direction of motion 𝑘̂ can be described by a
𝑘

wave function
Ψ(𝑟, 𝑡) = 𝐴 𝑒𝑥𝑝[𝑖(𝑘⃗ ∙ 𝑟 − 𝜔𝑡)]
where
ℎ
𝐸 = ℎ𝜈 = 2𝜋𝜈 => 𝐸 = ℏ𝜔,
2𝜋
Now,
𝑓𝑟𝑜𝑚 (𝑘⃗, 𝜔) 𝑑𝑒𝑠𝑐𝑟𝑖𝑝𝑡𝑖𝑜𝑛 𝑡𝑜 (𝑝, 𝐸 )
ℎ
𝜆=
𝑝
ℎ
=> 𝑝=𝑘 => 𝑝 = ℏ𝑘
2𝜋
2𝜋
𝜆=
𝑘
𝑝2
Since for a free particle 𝐸 = =>
2𝑚

𝐸 𝑝2
𝜔= =
ℏ 2𝑚ℏ

Yields
 4

𝑝∙𝑟 𝑝2
Ψ(𝑟, 𝑡) = 𝐴 𝑒𝑥𝑝 [𝑖 ( − 𝑡)]
ℏ 2𝑚ℏ

Due to the wave-particle duality we can now assume that this wave
function must somehow be related to the wave properties of free
electrons (as observed in the electron diffraction experiments).
Note that, this wave function satisfies a differential equation

𝜕Ψ(𝑟, 𝑡) 𝑝2
𝑖ℏ = Ψ(𝑟, 𝑡) = 𝐸Ψ(𝑟, 𝑡)
𝜕𝑡 2𝑚
And also
ℏ2 2 𝑝2
− ∇ Ψ(𝑟, 𝑡) = Ψ(𝑟, 𝑡) = 𝐸Ψ(𝑟, 𝑡)
2𝑚 2𝑚
OR
𝝏𝚿(𝒓
⃗ , 𝒕) ℏ𝟐 𝟐
𝒊ℏ =− 𝛁 𝚿(𝒓
⃗ , 𝒕) ∶ 𝒇𝒐𝒓 𝒂 𝒇𝒓𝒆𝒆 𝒑𝒂𝒓𝒕𝒊𝒄𝒍𝒆
𝝏𝒕 𝟐𝒎

What happens when we the particle is not free?

𝑉 (𝑟, 𝑡) ≠ 0

All the above attempts do not give us any clue about the answer.
However, comparison of this differential equation with the classical energy
equation (*) can give us the idea to try the equation below as a starting
point for the calculation of wave functions for particles moving in a
potential
𝑉 (𝑟, 𝑡):

𝝏𝚿(𝒓
⃗ , 𝒕) ℏ𝟐 𝟐
𝒊ℏ =− 𝛁 𝚿(𝒓
⃗ , 𝒕) + 𝑽(𝒓
⃗ , 𝒕 ) 𝚿( 𝒓
⃗ , 𝒕)
𝝏𝒕 𝟐𝒎
 5

The basic postulate of QM:

The state of a system at any instant of time is represented by a state
function or a wave function 𝚿(𝒓
⃗ , 𝒕), which is a complex function, in
general. All the information regarding the state of the system is
contained in the wave function. All needed physical quantities can be
computed from this wave function. The basic equation for determining
the wave function 𝚿(𝒓
⃗ , 𝒕) is the Schrödinger equation,

𝜕Ψ(𝑟, 𝑡) ℏ2 2
𝑖ℏ =− ∇ Ψ(𝑟, 𝑡) + 𝑉 (𝑟, 𝑡) Ψ(𝑟, 𝑡)
𝜕𝑡 2𝑚

ℎ
where 𝑖 = √−1 and ℏ = = 1.0545 × 10−34 J.s, where
2𝜋

ℎ = 6.626 × 10−34 J.s is the Planck’s constant.

𝚿( 𝒓
⃗ , 𝒕): wave function in coordinate space

The Schrödinger equation is linear in the wave function Ψ(𝑟, 𝑡):

So that, if Ψ1 (𝑟, 𝑡) and Ψ2 (𝑟, 𝑡) are two different solutions to the equation
then Ψ(𝑟, 𝑡) = 𝑐1 Ψ1 (𝑟, 𝑡) + 𝑐2 Ψ2 (𝑟, 𝑡) is also a solution. (This combination
is said to be linear since it involves the 1.st power of 𝜓1 and 𝜓2 ; it is said to
be arbitrary since 𝑐1 and 𝑐2 are arbitrary.)
The linearity requirement ensures that we account for the interference and
diffraction patterns in DG experiment: We shall be able to add together wave
functions representing particles scattered from different planes of the atomic
layers in the crystal to produce the constructive and destructive
interferences.

The Schrödinger equation plays a role analogous to Newton’s second law:

Given the suitable initial conditions (typically, Ψ(𝑟, 0)), the Schrödinger
equation determines Ψ(𝑟, 𝑡) for all future time, just as in classical
mechanics Newton’s law determines 𝑟(𝑡) for all future time.
 6

The Statistical Interpretation: What exactly is this “wave function” and

what does it do for you ones you have got it?
After all, a particle, by its nature, is localized at a point, whereas the wave
function (as its name suggests) is spread out in space: it’s a function of 𝑟
for any given time t.
How can such an object represent the state of a particle?
The answer is provided by Max Born’s statistical interpretation of the
wave function, which says that
|𝜓(𝑟, 𝑡)|2 𝑑3 𝑟 = 𝑃(𝑟, 𝑡)𝑑3 𝑟

which is the probability that a measurement of the particle’s position at

time t will be in a volume element 𝑑3 𝑟 centered at a 𝑟.

𝑑3 𝑟

Or in one dimension
|Ψ(𝑥, 𝑡)|2 𝑑𝑥 ≡ 𝑃(𝑥, 𝑡) 𝑑𝑥
which gives the probability that the particle will be found at time t at the
point 𝑥, which is between (𝑥, 𝑥 + 𝑑𝑥 ). Thus,
|Ψ(𝑥, 𝑡)|2 ≡ 𝑃(𝑥, 𝑡) ∶ 𝑡ℎ𝑒 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑑𝑒𝑛𝑠𝑖𝑡𝑦
Such that
 7

EX: Consider a particle represented by a wave function Ψ and suppose that

it has a probability density |Ψ|2 given in Fig.1.2 below.

 At which point, A or C, is the particle most likely to be found for this

wave function? What about point B?
It is more likely to find it in the vicinity of point A, where |Ψ|2 is large, and
relatively unlikely to find it near point B.
 What is the probability of finding the particle anywhere between and b?

This probability is the area under the graph of |Ψ(𝑥, 𝑡)|2 .

𝒃

∫|Ψ(𝑥, 𝑡)|2 𝑑𝑥 = probability of finding the particle between a and b

𝒂

at time t.

The statistical interpretation brings up a kind of indeterminacy into quantum

mechanics: Even if you know the wave function which is the everything the
theory gives you about the particle, still you cannot predict with certainty the
outcome of a simple experiment to measure the position of the particle!
All quantum mechanics offers is statistical information about the possible
results.
It is a fact of nature, it is how nature works, rather than being a defect in
the theory.
 8

 Suppose I measure the position of the particle and I find it to be at point

C.
Question: Where was the particle just before I made the measurement?
Although Schrödinger equation was extremely successful in comparison with
experiments (not one prediction from quantum mechanics is found to be
contradicted by experiments), there exist a number of alternative schools of
thought over their interpretation. There are different answers to this question
from different schools of thought. The most widely accepted one is the
Copenhagen interpretation.
According to the Copenhagen interpretation:

 physical systems generally do not have definite properties before the

measurements are performed.
 Quantum mechanics can only predict the probability distribution of a
given measurement's possible results.
 The act of measurement affects the system in such a way that
immediately after the measurement the set of probabilities is reduced
to only one of the possible values. This feature is known as wave
function collapse.
 9

Therefore, Copenhagen interpretation answers the question above as “ the

particle was not really anywhere”.
The measurement radically alters the wave function so that it is now sharply
peaked about C. We say that upon measurement wave function collapses to a
spike at the point C.

What if we make a second measurement, immediately after the first one?

On the answer to this question everyone is in agreement: A repeated
measurement (on the same particle) must return the same value (repeatability
of the measurement !).
 10

Normalization
What is the probability of finding the particle anywhere in the x axis
(between −∞ and ∞ ) at time t.?
This probability is the area under the graph of |Ψ(𝑥, 𝑡)|2 .
∞
∫ |Ψ(𝑥, 𝑡)|2 𝑑𝑥 = probability of finding the particle
−∞

between − ∞ and ∞.
Now, the statistical interpretation of the wave function suggests that the total
probability of finding the particle somewhere in the x-axis must be equal to 1.
That is, the sum of the probabilities for all of the x-axis must be equal to one.
This is expressed by the integral
∞
∫ |Ψ(𝑥, 𝑡)|2 𝑑𝑥 = 1 (∗∗)
−∞

(which means that if the particle exits it must be somewhere at all times)
.Without this, the statistical interpretation would be nonsense.
Note that if Ψ(𝑥, 𝑡) is a solution to the Schrödinger equation (*),
𝜕Ψ(𝑥, 𝑡) ℏ2 𝜕 2 Ψ(𝑥, 𝑡)
𝑖ℏ =− + 𝑉 (𝑥, 𝑡) Ψ(𝑥, 𝑡)
𝜕𝑡 2𝑚 𝜕𝑥 2
so too is 𝐴 Ψ(𝑥, 𝑡), where 𝐴 is any complex constant. What we must do, then,
is to pick this undetermined factor so as to ensure that eq.(**) is satisfied. This
process is called normalizing the wave function and a wave function that
obeys this equation is said to be normalized. Non-normalizable solutions
cannot represent the particle and they must be rejected. A wave function must
be square integrable if it is a normalizable function.

∞
∫ |Ψ(𝑥, 𝑡)|2 𝑑𝑥 < ∞ => lim Ψ(𝑥, 𝑡) → 0
−∞ 𝑥→±∞
 11

Suppose that we have normalized the wave function at time t=0. How do
we know that it will stay normalized as time goes on and 𝚿 evolves ?
Schrödinger equation has the property that it preserves the normalization
of the wave function.
Proof:

𝑑 ∞ 2
∞
𝜕
∫ |Ψ(𝑥, 𝑡)| 𝑑𝑥 = ∫ |Ψ(𝑥, 𝑡)|2 𝑑𝑥 = 0
𝑑𝑡 −∞ −∞ 𝜕𝑡

By the product rule

𝜕 2
𝜕 ∗ ∗
𝜕Ψ 𝜕Ψ∗
|Ψ(𝑥, 𝑡)| = (Ψ Ψ) = Ψ + Ψ
𝜕𝑡 𝜕𝑡 𝜕𝑡 𝜕𝑡
Now, the Schrödinger eq
𝜕Ψ(𝑥, 𝑡) ℏ2 𝜕 2 Ψ(𝑥, 𝑡)
𝑖ℏ =− + 𝑉 (𝑥, 𝑡) Ψ(𝑥, 𝑡)
𝜕𝑡 2𝑚 𝜕𝑥 2
Says that
𝜕Ψ(𝑥, 𝑡) ℏ 𝜕 2 Ψ(𝑥, 𝑡) 1
=− + 𝑉 (𝑥, 𝑡) Ψ(𝑥, 𝑡)
𝜕𝑡 2𝑚𝑖 𝜕𝑥 2 𝑖ℏ
𝜕Ψ ∗ (𝑥, 𝑡) ℏ 𝜕 2 Ψ∗ (𝑥, 𝑡) 1
= 2
+ 𝑉 (𝑥, 𝑡) Ψ ∗ (𝑥, 𝑡)
𝜕𝑡 2𝑚𝑖 𝜕𝑥 −𝑖ℏ
Hence,

𝜕 2 ∗
ℏ 𝜕 2 Ψ(𝑥, 𝑡) 1
|Ψ(𝑥, 𝑡)| = Ψ [− + 𝑉 (𝑥, 𝑡) Ψ(𝑥, 𝑡)]
𝜕𝑡 2𝑚𝑖 𝜕𝑥 2 𝑖ℏ
ℏ 𝜕 2 Ψ∗ (𝑥, 𝑡) 1
+[ 2
− 𝑉 (𝑥, 𝑡) Ψ ∗ (𝑥, 𝑡)] Ψ
2𝑚𝑖 𝜕𝑥 𝑖ℏ
𝑖ℏ ∗
𝜕 2 Ψ(𝑥, 𝑡) 𝜕 2 Ψ∗ (𝑥, 𝑡)
= [Ψ −Ψ ]
2𝑚 𝜕𝑥 2 𝜕𝑥 2
𝜕 𝑖ℏ ∗
𝜕Ψ(𝑥, 𝑡) 𝜕Ψ ∗ (𝑥, 𝑡)
= [ (Ψ −Ψ )] (∗)
𝜕𝑥 2𝑚 𝜕𝑥 𝜕𝑥
 12

∞ ∞
𝜕 2
𝜕 𝑖ℏ ∗
𝜕Ψ(𝑥, 𝑡) 𝜕Ψ ∗ (𝑥, 𝑡)
∫ |Ψ(𝑥, 𝑡)| 𝑑𝑥 = ∫ [ (Ψ −Ψ )] 𝑑𝑥
−∞ 𝜕𝑡 −∞ 𝜕𝑥 2𝑚 𝜕𝑥 𝜕𝑥
∞
𝑖ℏ ∗
𝜕Ψ(𝑥, 𝑡) 𝜕Ψ∗ (𝑥, 𝑡)
=[ (Ψ −Ψ )]
2𝑚 𝜕𝑥 𝜕𝑥 −∞

But Ψ(𝑥, 𝑡) must go to zero as x goes to ±∞; otherwise the wave function
would not be normalizable. (Such functions are said to be square integrable.) It
follows that
∞
𝜕
∫ |Ψ(𝑥, 𝑡)|2 𝑑𝑥 = 0
−∞ 𝜕𝑡

And hence that the integral is constant (independent of time); if Ψ is

normalized at t=0, it stays normalized for all future time.

Now, from (*), write

𝜕 2
𝜕 𝜕 𝑖ℏ ∗
𝜕Ψ(𝑥, 𝑡) 𝜕Ψ ∗ (𝑥, 𝑡)
|Ψ(𝑥, 𝑡)| ≡ 𝑃(𝑥, 𝑡) = [ (Ψ −Ψ )] (∗∗)
𝜕𝑡 𝜕𝑡 𝜕𝑥 2𝑚 𝜕𝑥 𝜕𝑥
Note that the term on the right hand side of eq. (**) above, is defined as the
flux or equivalently the probability current
𝑖ℏ ∗
𝜕Ψ 𝜕Ψ∗
𝐽(𝑥, 𝑡) = − (Ψ − Ψ)
2𝑚 𝜕𝑥 𝜕𝑥
because it tells us the rate at which probability is flowing past the point x.
We get
𝜕 𝜕
𝑃(𝑥, 𝑡) + 𝐽(𝑥, 𝑡) = 0 (∗∗∗): 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
𝜕𝑡 𝜕𝑥
OR, in 3.dim,
𝜕
𝑃(𝑟, 𝑡) + ⃗∇ ∙ 𝐽(𝑟, 𝑡) = 0
𝜕𝑡
This is a conservation law analogous to the charge conservation equation in
classical electrodynamics;
 13

Continuity Equation in classical electrodynamics

𝑑𝑄
𝐼=− ∶ Electric current (1)
𝑑𝑡

𝐼 : instantaneous current flowing outwards

through S into the exterior space
Q : the instantaneous charge in the enclosed volume V.
Express the charge Q as a volume integral of the charge density, 𝜌

𝑄 = ∫ 𝜌 𝑑𝑉 (2)
𝑉

Therefore,

𝑑𝑄 𝑑 𝜕𝜌
= ∫ 𝜌 𝑑𝑉 = ∫ 𝑑𝑉 (3)
𝑑𝑡 𝑑𝑡 𝜕𝑡
𝑉 𝑉

Now we can express the current I as a surface integral of the current

density J:
divergence
theorem
𝐼 = ∫ 𝐽 ∙ 𝑑𝑆 =
⏞ ⃗ ∙ 𝐽 𝑑𝑉
∫∇ (4)
𝑆 𝑉

Now, combining (3) and (4), we conclude that

𝜕𝜌 𝜕𝜌
∫ ⃗ ∙ 𝐽 𝑑𝑉 => ∫ ( + ∇
𝑑𝑉 = − ∫ ∇ ⃗ ∙ 𝐽) 𝑑𝑉 = 0 =>
𝜕𝑡 𝜕𝑡
𝑉 𝑉 𝑉

𝜕𝜌
( + ⃗∇ ∙ 𝐽) = 0
𝜕𝑡
Any variation in the total charge within a closed surface must be due to
charges that flow across the surface.
This is called the equation of continuity. It is a direct expression of the
local law of conservation of charge.
 14

Now, in QM,

𝜕 𝜕
𝑃(𝑥, 𝑡) + 𝐽(𝑥, 𝑡) = 0 (∗∗∗): 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛
𝜕𝑡 𝜕𝑥
The continuity equation expresses the fact that a change in the probability
density in a region, say 𝑎 ≤ 𝑥 ≤ 𝑏, with time is compensated by a net
change in probability current (flux) into the region.
Let’s integrate (***)
𝑏 𝑏
𝜕 𝜕
∫ 𝑃(𝑥, 𝑡) 𝑑𝑥 = − ∫ 𝐽(𝑥, 𝑡)𝑑𝑥 = 𝐽(𝑎, 𝑡) − 𝐽(𝑏, 𝑡)
𝜕𝑡 𝜕𝑥
⏟
𝑎 𝑎
𝑃𝑎𝑏

This is the solution to Probl. 1.14

𝜕
𝑃 = 𝐽(𝑎, 𝑡) − 𝐽(𝑏, 𝑡)
𝜕𝑡 𝑎𝑏
If 𝑃𝑎𝑏 is increasing then more probability is flowing into the region at one
end (x=a) that flows out at the other.

OR, in 3.dim,
𝜕
𝑃(𝑟, 𝑡) + ⃗∇ ∙ 𝐽(𝑟, 𝑡) = 0
𝜕𝑡
𝑃(𝑟, 𝑡) = |Ψ(𝑥, 𝑡)|2 ∶ 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑑𝑒𝑛𝑠𝑖𝑡𝑦
𝐽(𝑟, 𝑡) ∶ 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑐𝑢𝑟𝑟𝑒𝑛𝑡

The time rate of change of probability of finding the particle within the
finite volume must be the negative of the probability of the net outflow of
the particle.
 15

Expectation value of the position :

⃗ (𝒕), for the motion

In QM, we cannot give an exact trajectory description, 𝒓
of a particle. Then

Q) How to relate a quantum mechanical calculation to something you can

observe in the laboratory?

Ans) the probability theory average value of measurable dynamical

parameters, such as x, as functions of the time is calculated.

The average value of an observable in QM is called its expectation

value.
Average Values of Classical Measurements
Suppose that we measure the location x of a classical object.
Measuring this quantity several times, one will in general obtain different
values 𝑥𝑖 , where each value occurs with the frequency of occurrence 𝑁𝑖 :
(𝑁𝑖 is the number of times that one gets the result 𝑥𝑖 ).( If the measuring
instruments were ideal, we would obtain the same value every time.)

Numerical result Number of times obtained

𝑥1 𝑁1
𝑥2 𝑁2
𝑥3 𝑁3
⋮ ⋮
𝑥𝑘 𝑁𝑘

If the total number of measurement is N we must have

𝑘

∑ 𝑁𝑖 = 𝑁
𝑖=1

The statistical mean (average) is defined to be

𝑁1 𝑥1 + 𝑁2 𝑥2 + ⋯ ∑𝑘𝑖=1 𝑁𝑖 𝑥𝑖 ∑𝑘𝑖=1 𝑁𝑖 𝑥𝑖
𝑥̅ = = 𝑘 =
𝑁1 + 𝑁2 + ⋯ ∑𝑖=1 𝑁𝑖 𝑁
 16

Here, when N>>1

𝑁𝑖
𝑃𝑖 = ∶ probability that measurement of 𝑥 finds the value 𝑥𝑖
𝑁
Thus,

𝑥̅ = ∑ 𝑃𝑖 𝑥𝑖 ∶ 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑜𝑓 𝑥
𝑖

with
∑ 𝑃𝑖 = 1
𝑖

This averaging concept is also applicable to sets of continuous data. We

need to perform a transition of going from a sum to an integral,

∑ → ∫
𝑖

and introduce the concept of a probability density function

𝜌(𝑥 ) ∶ probability density function
such that
𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛 𝑥
𝜌(𝑥 )𝑑𝑥 = { }
𝑙𝑖𝑒𝑠 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑥 𝑎𝑛𝑑 (𝑥 + 𝑑𝑥)
and
𝑏

𝑃(𝑎 ≤ 𝑥 ≤ 𝑏) = ∫ 𝜌(𝑥 )𝑑𝑥 : 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑥 𝑖𝑠 𝑠𝑜𝑚𝑒𝑤ℎ𝑒𝑟𝑒 𝑖𝑛 (𝑎, 𝑏)

𝑎

The rules for the discrete distribution translate in the obvious way:
∞

∫ 𝜌(𝑥 )𝑑𝑥 = 1
−∞
∞

𝑥̅ = ∫ 𝑥 𝜌(𝑥 )𝑑𝑥 (∗)

−∞
 17

Average Values of Quantum Measurements

In quantum mechanics, the average value of an observable is called its

expectation value.

Consider now a quantum particle in a state Ψ(𝑥, 𝑡). Let’s consider the
particle’s position observable, 𝑥. In analogy with (*), Its expectation value
is written as
∞ ∞
〈𝑥 〉 = ∫ 𝑥𝑃(𝑥, 𝑡)𝑑𝑥 = ∫ 𝑥 |Ψ(𝑥, 𝑡)|2 𝑑𝑥 = (∗∗)
−∞ −∞

What exactly does this mean?

It doesn’t mean that if you measure the position of one particle over and
over again and 〈𝑥 〉 is the average of the results you get. Since on the first
measurement of its position the wave function collapses resulting a peak at
the value you just obtained and the later measurements necessarily give the
same result. Therefore, the expectation value is rather an ensemble
average. One gets an ensemble of identically prepared systems, at the same
instant records the position of the particle and then calculate the average of
these records.
 18

Expectation value of the momentum :

We assume that the expectation value of momentum 〈𝑝〉 obeys the
equation
𝑑 〈𝑥 〉 1
= 〈𝑝〉 ≡ 〈𝑣〉 (∗)
𝑑𝑡 𝑚

With
∞
〈𝑥 〉 = ∫ 𝑥 |Ψ(𝑥, 𝑡)|2 𝑑𝑥
−∞

(We will return to this relation in CH3 when we study the time dependence
of expectation values.)
This is an assumption at this point, which has to prove itself in the following.
It is the “velocity” of the expectation value of x, which is not the same as the
velocity of the particle. It is not even clear what velocity means in quantum
mechanics: if the particle doesn’t have a determinate position (prior to the
measurement), neither does it have a well-defined velocity. For our present
purposes it will be Ok to postulate that the expectation value of the velocity is
equal to the time derivative of the expectation value of the position:
Since Ψ = Ψ(𝑥, 𝑡), as time goes on, 〈𝑥 〉 will change.

Let’s calculate the LHS of (*):

∞
𝑑 〈𝑥 〉 𝜕 2
𝑖ℏ ∞ 𝜕 ∗
𝜕Ψ 𝜕Ψ∗
= ∫ 𝑥 |Ψ(𝑥, 𝑡)| 𝑑𝑥 = ∫ 𝑥 (Ψ − Ψ) 𝑑𝑥
𝑑𝑡 −∞ 𝜕𝑡 2𝑚 −∞ 𝜕𝑥 𝜕𝑥 𝜕𝑥

Integration by parts=>

𝑑 〈𝑥 〉 𝑖ℏ ∞ 𝜕 ∗
𝜕Ψ 𝜕Ψ ∗ 𝑖ℏ ∞
= ∫ ⏟ 𝑥 (Ψ − Ψ) 𝑑𝑥 = ∫ 𝑢 𝑑𝑉
𝑑𝑡 2𝑚 −∞ 𝑢 𝜕𝑥 ⏟ 𝜕𝑥 𝜕𝑥 2𝑚 −∞
𝑉
 19

∞ ∞
𝑖ℏ ∗
𝜕Ψ 𝜕Ψ∗ ∗
𝜕Ψ 𝜕Ψ∗
= {𝑥 (Ψ − Ψ)| − ∫ 𝑑𝑥 (Ψ − Ψ)}
2𝑚 ⏟ 𝜕𝑥 𝜕𝑥 −∞ −∞ 𝜕𝑥 𝜕𝑥
=0

𝑑 〈𝑥 〉 𝑖ℏ ∞ ∗
𝜕Ψ 𝜕Ψ∗
=− ∫ 𝑑𝑥 (Ψ − Ψ)
𝑑𝑡 2𝑚 −∞ 𝜕𝑥 𝜕𝑥

𝑖ℏ ∞ ∗
𝜕Ψ 𝑖ℏ ∞ 𝜕Ψ ∗
=− ∫ 𝑑𝑥 Ψ + ∫ 𝑑𝑥 Ψ
2𝑚 −∞ 𝜕𝑥 2𝑚 −∞ 𝜕𝑥

Performing another integration by parts on the second term,

𝜕Ψ 𝜕Ψ ∗
u = Ψ , du = 𝑑𝑥 ; dV = 𝑑𝑥 , 𝑉 = Ψ∗
𝜕𝑥 𝜕𝑥

𝑑 〈𝑥 〉 𝑖ℏ ∞ 𝜕Ψ 𝑖ℏ ∞
𝜕Ψ
=− ∫ 𝑑𝑥 Ψ∗ + (Ψ ∗ Ψ)|∞
{⏟ −∞ − ∫ 𝑑𝑥 Ψ ∗
}
𝑑𝑡 2𝑚 −∞ 𝜕𝑥 2𝑚 −∞ 𝜕𝑥
=0

𝑑 〈𝑥 〉 𝑖ℏ ∞ 𝜕Ψ 1
= − ∫ 𝑑𝑥 Ψ∗ = 〈𝑝〉 =>
𝑑𝑡 𝑚 −∞ 𝜕𝑥 𝑚
∞
𝜕Ψ
〈𝑝〉 = −𝑖ℏ ∫ 𝑑𝑥 Ψ ∗
−∞ 𝜕𝑥
Let’s rewrite 〈𝑥 〉 and 〈𝑝〉 in the form below
∞
〈𝑥 〉 = ∫ Ψ ∗ (𝑥 ) Ψ 𝑑𝑥
−∞

∞
ℏ 𝜕
〈𝑝 〉 = ∫ Ψ∗ ( ) Ψ 𝑑𝑥
−∞ 𝑖 𝜕𝑥

We say that in QM,

The operator 𝑥 represents position :

ℏ 𝜕
The operator ( ) represents momentum
𝑖 𝜕𝑥
 20

Operators can be written in a variety of representations.

In the position representation where 𝑥 is the independent variable the
operators for the position and momentum have the form
ℏ 𝜕
𝑥̂ = 𝑥 , 𝑝̂ = [𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛]
𝑖 𝜕𝑥
In the momentum representation where 𝑝 is the independent variable the
operators for the position and momentum have the form
ℏ 𝜕
𝑝̂ = 𝑝 , 𝑥̂ = [𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛]
𝑖 𝜕𝑝
The form of the momentum operator above raises the question whether the

expectation value of the momentum in some states could be imaginary. We

can in fact show that expectation value of the momentum is always real.

We write
∞ ∞
ℏ 𝜕 ℏ 𝜕
〈𝑝 〉 − 〈𝑝 〉∗ =∫ Ψ ( ∗
) Ψ 𝑑𝑥 − ∫ Ψ (− ) Ψ∗ 𝑑𝑥
−∞ 𝑖 𝜕𝑥 −∞ 𝑖 𝜕𝑥

ℏ ∞ ∗
𝜕Ψ 𝜕Ψ ∗ ℏ ∞ 𝜕 ℏ
= ∫ (Ψ +Ψ ) 𝑑𝑥 = ∫ (Ψ ∗ Ψ) 𝑑𝑥 = (Ψ∗ Ψ)∞
−∞ = 0
𝑖 −∞ 𝜕𝑥 𝜕𝑥 𝑖 −∞ 𝜕𝑥 𝑖

Last step follows from the square integrability of the wave function.

An operator whose expectation value for all admissible wave functions is real

is called Hermitian operators. We see that momentum operator is a Hermitian

operator.
 21

Postulate) Observables and Operators: To any self-consistently and well-

defined observable in physics there corresponds an (Hermitian) operator such
that measurement of the observable yields values which are eigenvalues of the
corresponding operator.
A: an observable (e.g. Energy, mass, momentum, etc.)
 : operator corresponding to A

𝐴̂ 𝜓 = 𝑎 𝜓
An operator is an instruction to do something to the function that follows
it.
EX: Some mathematical operators which are not necessarily connected to
physics.
1)
𝑑 𝑑𝑓(𝑥)
̂=
𝐷 such that ̂ 𝑓(𝑥) =
𝐷
𝑑𝑥 𝑑𝑥
If f.ex. 𝑓(𝑥 ) = 𝑥 2 then
𝑑𝑓 (𝑥 )
̂ 𝑓(𝑥 ) =
𝐷 =2𝑥
𝑑𝑥
If 𝑓(𝑥 ) = 𝑒 −𝑖𝜔𝑥 , where 𝜔 is constant, then;:
𝜕𝑓 (𝑥 )
̂ 𝑓(𝑥 ) =
𝐷 = −𝑖𝜔𝑒 −𝑖𝜔𝑥 = −𝑖𝜔𝑓 (𝑥 ): 𝑒𝑖𝑔𝑒𝑛𝑣𝑎𝑙𝑢𝑒 𝑒𝑞.
𝜕𝑥
2)
𝐼̂ ∶ identity operator such that𝐼̂𝑓(𝑥 ) = 𝑓(𝑥 )for any function 𝑓(𝑥)
-------------------------------------------------------------------------------------------
 22

How to calculate the expectation values of other quantities? For this note that

any classical dynamical variable, 𝑸, can be expressed in terms of position

and/or momentum:

𝑸(𝒙, 𝒑)

EX:

Kinetic energy operator:

1 2
𝑝2
𝑇 = 𝑚𝑣 =
2 2𝑚

Angular Momentum:

𝐿⃗ = 𝑟 × 𝑝

In general, to calculate the expectation value of any such quantity,

𝑸(𝒙, 𝒑),

ℏ 𝜕
we simply replace every 𝑝 → 𝑝̂ = ( ) , insert the resulting operator
𝑖 𝜕𝑥

between Ψ ∗ and Ψ, and integrate

∞
ℏ 𝜕
〈𝑄(𝑥, 𝑝)〉 = ∫ Ψ ∗ 𝑄 (𝑥, ) Ψ 𝑑𝑥
−∞ 𝑖 𝜕𝑥

For example, the expectation value of the kinetic energy is

𝑝2 1 ∞ ∗ ℏ 𝜕 2 ℏ2 ∞ ∗ 𝜕 2 Ψ
〈𝑇 〉 = 〈 〉 = ∫ Ψ ( ) Ψ 𝑑𝑥 = − ∫ Ψ 𝑑𝑥
2𝑚 2𝑚 −∞ 𝑖 𝜕𝑥 2𝑚 −∞ 𝜕𝑥 2
 23

WAVE PACKETS AND UNCERTAINTY

We considered before the possibility that the state of a free electron (as
observed in the electron diffraction experiments) is described by a
monochromatic plane wave of angular frequency 𝜔 = 2𝜋𝜈 , wavelength
2𝜋
𝜆= ,
𝑘

Ψ(𝑥, 𝑡) = 𝐴 𝑒𝑥𝑝[𝑖(𝑘𝑥 − 𝜔𝑡)] = 𝐴 𝑒𝑥𝑝[𝑖 (𝑝𝑥/ℏ − 𝐸𝑡/ℏ)]

Note that the probability of finding this electron anywhere along the x-axis
is the same:
|𝜓(𝑥, 𝑡)|2 = |𝐴|2 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡.
It is clear therefore that this plane wave can not represent a localized electron; it is just
an indefinite series of wave with the same amplitude A. Instead, a particle which is
localized within a certain region of space can be described by a wave whose amplitude
is large in that region and zero outside it. In another word, the matter wave must be
localized around the region of space within which the particle is confined. Such a
localized wave is called a wave packet. We can build up localized wave packets by
adding up these free particle wave functions. A wave packet therefore consists of a
group of waves of slightly different wavelengths, with phases and amplitudes so chosen
that they interfere constructively over a small region of space and destructively
elsewhere. Mathematically, we can carry out this superposition by means of Fourier
transforms. For simplicity, we are going to consider a one-dimensional wave packet
∞
1
Ψ(𝑥, 𝑡) = ∫ 𝜙(𝑘, 𝑡)𝑒 𝑖(𝑘𝑥−𝜔𝑡) 𝑑𝑘
√2𝜋
−∞

𝜙(𝑘, 𝑡) is the amplitude of the packet.

∞
1
𝜙(𝑘, 𝑡) = ∫ Ψ(𝑥, 𝑡)𝑒 −𝑖(𝑘𝑥−𝜔𝑡) 𝑑𝑥
√2𝜋
−∞
 24

In physics, a wave packet is a short "burst" or "envelope" of wave action

𝑑𝜔
that travels as a unit with the group velocity v𝑔 = . Phase velocity 𝑣𝑝 =
𝑑𝑘
𝑐2
has no physical significance, because the motion of the wave group, not
v
the motion of the individual waves that make up the group, corresponds to
the motion of the body.
In summary, the particle is represented not by a single wave of well-defined
frequency and wavelength, but by a wave packet that is obtained by adding
a large number of such waves of different frequencies

There is a region, labeled Δ𝑥,

where the wave differs from
zero and the particle may be
found and outside this region
the wave is very small and the
particle is very unlikely to be found there.
 25

The Heisenberg Uncertainty Principle

To regard a moving particle as a wave implies that there are fundamental
limits to the accuracy with which we can measure the “particle” properties
such as position and momentum.

The uncertainty principle, developed by W. Heisenberg, is a statement of the

effects of wave-particle duality on the properties of subatomic objects.

A sine wave of 𝜆 implies that the

ℎ
momentum p is precisely known: 𝑝 = .
𝜆
But the wave function and the probability
of finding the particle is spread over all of
the space.
Momentum: precise
Ψ(𝑥, 𝑡) = 𝐴 𝑒𝑥𝑝[𝑖(𝑝𝑥/ℏ − 𝐸𝑡/ℏ)] Position: unknown (∞′ 𝑙𝑦 𝑚𝑎𝑛𝑦)

Take an extreme case of a particle localized at just

one point in space. Its matter wave is just a pulse at
that point in space.
what is the momentum of this spatially localized
particle?
The superposition given earlier answers this. We
found that when we took the matter waves of
particles with different momenta and added them, we
produced a matter wave that was spatially localized.
That set turns out to contain all possible values of
momenta.
Poisiton =precise
Momentum=unknown (∞′ 𝑙𝑦 𝑚𝑎𝑛𝑦)
∞
1
Ψ(𝑥, 𝑡) = ∫ 𝜙(𝑝, 𝑡)𝑒 𝑖(𝑝𝑥/ℏ−𝐸𝑡/ℏ) 𝑑𝑘
√2𝜋
−∞
 26

These two cases are the extremes:

 We have a matter wave with a definite momentum but all possible

positions;
 We have a matter wave with a definite position but all possible
momenta.

Free, propagating particles in quantum theory are represented by an

intermediate case:

We arrive at a wave packet by adding matter waves with a small range of

momenta. The resulting packet occupies a range of positions in space Δ𝑥 and
is associated with a range of momenta Δ𝑝.

The trade-off we have just seen between definiteness of position and

definiteness of momentum is quantified by Heisenberg's uncertainty
principle:

“It is impossible to know both the exact position and exact momentum
of an object at the same time. There is a minimum for the product of
uncertainties of these two measurements.”

This relation was discovered by Werner Heisenberg in 1927 and it is one of

the most significant of the physical laws. This is not a statement about the
inaccuracy of measurement instruments, nor a reflection on the quality of
experimental method; it arises from the wave properties inherent in the
quantum mechanical description of nature.
 1

Heisenberg Uncertainty Principle (HUP):

“It is impossible to know both the exact position and exact momentum
of an object at the same time. There is a minimum for the product of
uncertainties of these two measurements.”

In QM, HUP arises from the wave-particle duality. Each particle

actually exhibits wavelike behavior; every particle has a matter
wave associated with it .

A matter wave is represented by a wave packet.

We can build up localized wave packets

by adding up plane waves of slightly
different wavelengths, with phases and
amplitudes so chosen that they interfere
constructively over a small region of
space and destructively elsewhere
 2

A matter wave having a well-defined wavelength 𝝀 (and

𝒉
momentum 𝒑 = , ) is spread out; the associated particle, while having
𝝀

a rather precise momentum may be almost anywhere.

A strictly localized matter wave (Δ𝑥 is small) has

an indeterminate wavelength; its associated
particle, while having a definite position, has no
certain momentum.

 HUP

The more precisely the position is known the more uncertain the
momentum is and vice versa.
 3

Since the uncertainty in QM is the a measure of the spread of the distribution,

it is technically the standard deviation 𝜎 in statistics.
As a measure of how far each individual data deviates from its average,
𝑥𝑖 − 〈𝑥 〉
one sums the square of the deviations from the average

𝜎 2 = ∑ 𝑃𝑖 (𝑥𝑖 − 〈𝑥 〉)2 : the variance

𝑖

and then take the square root to find the standard deviation 𝜎:

𝜎 = √∑ 𝑃𝑖 (𝑥𝑖 − 〈𝑥 〉)2
𝑖

Lets first rewrite the variance:

𝜎 2 = ∑ 𝑃𝑖 (𝑥𝑖 − 〈𝑥 〉)2 = ∑ 𝑃𝑖 (𝑥𝑖2 − 2𝑥𝑖 〈𝑥 〉 + 〈𝑥 〉2 )

𝑖 𝑖

= ∑ 𝑃𝑖 𝑥𝑖2 − 2〈𝑥 〉 ∑ 𝑃𝑖 𝑥𝑖 + 〈𝑥 〉2 ∑ 𝑃𝑖
⏟𝑖 ⏟𝑖 ⏟𝑖
〈𝑥 2 〉 〈𝑥〉 1

Note that when the position data comprise a discrete set, then the expectation
value is given by

〈𝑥 〉 = ∑ 𝑃𝑖 𝑥𝑖
𝑖

with
∑ 𝑃𝑖 = 1
𝑖
Thus
 4

𝜎 2 = 〈𝑥 2 〉 − 2 〈⏟
𝑥 〉〈𝑥 〉 + 〈𝑥 〉2
〈𝑥〉2

OR
𝜎 2 = 〈𝑥 2 〉 − 〈𝑥 〉2 : 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒
And

𝜎 = √〈𝑥 2 〉 − 〈𝑥 〉2 ∶ 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

In QM, the standard deviation is called uncertainty

∆𝑥 = √〈𝑥 2 〉 − 〈𝑥 〉2 : 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑖𝑛 𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛

Similarly, for a general observable 𝐴, we calculate its uncertainty ∆𝐴 as

∆𝐴 = √〈𝐴2 〉 − 〈𝐴〉2 : 𝑢𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑖𝑛 𝐴

 5

In classical physics the standard deviation is a measure of the dispersion

of the measured values and it arises due to imperfection of experimental
set up. In QM, its meaning is quite different; ∆𝑥 or in general ∆𝐴 for any
observable 𝐴, is not due to instrumental errors, but it is an unavoidable
genuine quantum effect. Successive measurements can yield different
values even for ideal measuring equipment. For ex., that ∆𝑥 ≠ 0 does not
mean that each single position measurement always has an error of this
magnitude, but rather that the quantum object simply does not have a
position in the classical sense. The concept of “exact location” is not
appropriate to this quantum mechanical problem.

Under what conditions ∆𝐴 = 0 will be discussed later.

1
2
3
4
5
 6

∞ ∞ 0 ∞

1 = ∫ |Ψ|2 𝑑𝑥 = 𝐴2 ∫ 𝑒 −2𝜆|𝑥| 𝑑𝑥 = 𝐴2 ∫ 𝑒 −2𝜆|−𝑥| 𝑑𝑥 + 𝐴2 ∫ 𝑒 −2𝜆𝑥 𝑑𝑥

−∞ −∞ ⏟ −∞ ⏟ 0
𝐼1 𝐼2

For I1:
0 ∞

𝑦 = −𝑥, 𝑑𝑦 = −𝑑𝑥 => 𝐼1 = −𝐴2 ∫ 𝑒 −2𝜆𝑦 𝑑𝑦 = 𝐴2 ∫ 𝑒 −2𝜆𝑦 𝑑𝑦 = 𝐼2

∞ 0

Thus,
∞ ∞ 0 ∞

1 = ∫ |Ψ|2 𝑑𝑥 = 𝐴2 ∫ 𝑒 −2𝜆|𝑥| 𝑑𝑥 = 𝐴2 ∫ 𝑒 −2𝜆|−𝑥| 𝑑𝑥 + 𝐴2 ∫ 𝑒 −2𝜆𝑥 𝑑𝑥

−∞ −∞ ⏟ −∞ ⏟ 0
𝐼1 𝐼2
∞
2 −2𝜆𝑥 2
1 −2𝜆𝑥 ∞ 𝐴2 −2𝜆∞
= 2𝐴 ∫ 𝑒 𝑑𝑥 = 2𝐴 𝑒 |0 = − (𝑒⏟ 𝑒 −2𝜆0 ) =>
−⏟
−2𝜆 𝜆 0 1
0

𝐴 = √𝜆
 7

∞
𝑛 −𝑎𝑥
𝑛!
∫𝑥 𝑒 𝑑𝑥 =
𝑎𝑛+1
0
 8

|Ψ|2 = 𝜆𝑒 −2𝜆|𝑥|
Note that 〈𝑥 〉 = 0.
9
10
11
12
 1

CH 2 TIME-INDEPENDENT SCHRÖDINGER EQUATION

In Quantum Mechanics, we describe systems using their wave functions,

which is a complex function, in general. Wave function is used to find
the probability that a measurement of the particle’s position at time t will
be in a volume element 𝑑3 𝑟⃗ centered at a 𝑟⃗.
|𝜓(𝑟⃗, 𝑡)|2 𝑑3 𝑟⃗ = 𝑃(𝑟⃗, 𝑡)𝑑3 𝑟⃗
All the information regarding the state of the system is contained in the
wave function. All needed physical quantities can be computed from this
wave function.
The basic equation for determining the wave function 𝜓 is the

Schrödinger equation,

𝜕Ψ(𝑥, 𝑡) ℏ2 𝜕 2 Ψ(𝑥, 𝑡)
𝑖ℏ =− + 𝑉(𝑥, 𝑡) Ψ(𝑥, 𝑡)
𝜕𝑡 2𝑚 𝜕𝑥 2

Thus, we need to solve the Schrödinger equation.

Stationary States

In most of the problems, 𝑉(𝑥, 𝑡) = 𝑉 (𝑥 ), which is independent of time t.

In this case, we can apply the method of separation of variables: We
look for solutions that are given by

Ψ(𝑥, 𝑡) = ⏟(𝑥 )
𝝍 𝑻
⏟ (𝑡)
𝑜𝑛𝑙𝑦 𝑎 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑜𝑛𝑙𝑦 𝑎 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
𝑜𝑓 𝑥 𝑜𝑓 𝑡
 2

Substituting this into Schr.eq (*) and noting that

𝜕 2 Ψ(𝑥, 𝑡) 𝑑2 𝜓(𝑥 )
= 𝑇(𝑡)
𝜕𝑥 2 𝑑𝑥 2

𝜕Ψ(𝑥, 𝑡) 𝑑T(𝑡)
(
=𝜓 𝑥 )
𝜕𝑡 𝑑𝑡

Then,

𝑑𝑇 (𝑡) −ℏ2 𝑑2 𝜓(𝑥 )

𝑖ℏ𝜓(𝑥 ) = 𝑇(𝑡) + 𝑉(𝑥)𝜓(𝑥 )𝑇(𝑡)
𝑑𝑡 2𝑚 𝑑𝑥 2

Dividing both sides by 𝜓(𝑟⃗)𝑇(𝑡)

𝜓(𝑥 ) 𝑑𝑇(𝑡) −ℏ2 𝑇(𝑡) 𝑑2 𝜓(𝑥 ) 𝜓(𝑥 )𝑇(𝑡)

𝑖ℏ = + 𝑉(𝑥)
𝜓(𝑥 )𝑇(𝑡) 𝑑𝑡 2𝑚 𝜓(𝑥 )𝑇(𝑡) 𝑑𝑥 2 𝜓(𝑥 )𝑇(𝑡)

Then, we get

1 𝑑𝑇(𝑡) −ℏ2 1 𝑑2 𝜓(𝑥 ) !

𝑖ℏ = + 𝑉(𝑥) =
⏞ 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
⏟ 𝑇(𝑡) 𝑑𝑡 2𝑚 𝜓(𝑥 ) 𝑑𝑥 2
⏟
𝐿𝐻𝑆(𝑡) 𝑅𝐻𝑆(𝑥)

Let us postulate that (the reason will appear soon !)

𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 = 𝐸: The constant total energy

 3

Then, we have two equations:

1.

1 𝑑𝑇(𝑡) 𝑑𝑇(𝑡) 𝑖𝐸
𝑖ℏ =𝐸 ⟹ + 𝑇(𝑡) = 0
𝑇(𝑡) 𝑑𝑡 𝑑𝑡 ℏ

𝑇(𝑡) = 𝐴 𝑒 −𝑖𝐸𝑡/ℏ

2.

−ℏ2 1 𝑑2 𝜓(𝑥 )
+ 𝑉(𝑥) = 𝐸 ⟹
2𝑚 𝜓(𝑥 ) 𝑑𝑥 2

−ℏ2 𝑑 2
[ + 𝑉(𝑥)] 𝜓(𝑥)𝐸𝜓(𝑥) :Steady-state form of the Schrödinger Equation
2𝑚 𝑑𝑥 2

∶ Time − independent Schrödinger Equation

Therefore, solution to the Schrödinger Equation (*) is written as

𝑖𝐸𝑡
Ψ(𝑥, 𝑡) = 𝑒 − ℏ 𝜓(𝑥 )

which is called the stationary state.

Separation of variables has turned a partial differential equation into two
ordinary differential equations.
 4

Some properties of the stationary states:

1. In a stationary state, although the wave function
Ψ(𝑥, 𝑡) = 𝑒 −𝑖𝐸𝑡/ℏ 𝜓(𝑥 )
itself depends on time t, the probability density doesn’t depend on time:
|Ψ(𝑥, 𝑡)|2 = 𝑒⏟−𝑖𝐸𝑡/ℏ 𝑒 𝑖𝐸𝑡/ℏ |𝜓(𝑥 )|2 = |𝜓(𝑥 )|2
1

2. In a stationary state, the expectation value of any dynamical variable

𝑸(𝒙, 𝒑) is time independent:
∞
ℏ 𝜕
〈𝑄(𝑥, 𝑝)〉Ψ(x,t) = ∫ Ψ ∗ (𝑥, 𝑡)𝑄 (𝑥, ) Ψ(x, t)𝑑𝑥
−∞ 𝑖 𝜕𝑥
∞ ∗ ℏ 𝜕
=∫ (𝑒 −𝑖𝐸𝑡/ℏ 𝜓(𝑥 )) 𝑄 (𝑥, ) (𝑒 −𝑖𝐸𝑡/ℏ 𝜓(𝑥 )) 𝑑𝑥
−∞ 𝑖 𝜕𝑥
∞
ℏ 𝜕
=∫ 𝜓 ∗ (𝑥 )𝑄 (𝑥, ) 𝜓(𝑥 ) (𝑒
⏟
𝑖𝐸𝑡/ℏ
)(𝑒 −𝑖𝐸𝑡/ℏ ) 𝑑𝑥
−∞ 𝑖 𝜕𝑥
1
∞
ℏ 𝜕
〈𝑄(𝑥, 𝑝)〉Ψ(x,t) = ∫ 𝜓 ∗ (𝑥 )𝑄 (𝑥, ) 𝜓(𝑥 )𝑑𝑥 = 〈𝑄(𝑥, 𝑝)〉𝜓(𝑥)
−∞ 𝑖 𝜕𝑥

That is, every expectation value is constant in time.

𝑑〈𝑄(𝑥, 𝑝)〉
=0 𝑖𝑛 𝑎 𝑠𝑡𝑎𝑡𝑖𝑜𝑛𝑎𝑟𝑦 𝑠𝑡𝑎𝑡𝑒
𝑑𝑡
Since 〈𝑥 〉 is constant in time we have
𝑑 〈𝑥 〉
〈𝑝 〉 = 𝑚 =0
𝑑𝑡
Nothing ever happens in a stationary state.
 5

3. Stationary states are states of definite total energy.

In classical mechanics, the total energy is called the Hamiltonian
𝑝2
𝐻 = 𝑇 + 𝑉(𝑥) = + 𝑉(𝑥)
2𝑚
The corresponding Hamiltonian operator, obtained by the canonical
substitution
ℏ 𝑑
𝑝→
𝑖 𝑑𝑥
in the classical Hamiltonian. Thus
−ℏ2 𝑑 2
̂=
𝐻→𝐻 + 𝑉(𝑥)
2𝑚 𝑑 2
Recall the TISE
−ℏ2 𝑑2
[ 2
+ 𝑉(𝑥)] 𝜓(𝑥 ) = 𝐸𝜓(𝑥 )
⏟ 2𝑚 𝑑𝑥
̂
𝐻

Thus, the time-independent Schrödinger equation can be written

̂ 𝜓(𝑥 ) = 𝐸𝜓(𝑥 ) ∶ 𝑒𝑛𝑒𝑟𝑔𝑦 𝑒𝑖𝑔𝑒𝑛𝑣𝑎𝑙𝑢𝑒 𝑒𝑞
𝐻
The expectation value of the total energy is

𝑝𝑟𝑜𝑣𝑒𝑑 𝑎𝑙𝑟𝑒𝑎𝑑𝑦
𝑖𝑛 (2) ∞
〈𝐻 〉Ψ(x,t) =
⏞ 〈𝐻 〉𝜓(𝑥) = ∫ ̂⏟𝜓(𝑥 ) 𝑑𝑥
𝜓 ∗ (𝑥 ) 𝐻
−∞ 𝐸𝜓(𝑥)
=1
∞ ⏞∞
∗(
= 𝐸∫ 𝜓 𝑥 ) 𝜓(𝑥 )𝑑𝑥 = 𝐸 ∫ |𝜓(𝑥 )|2 𝑑𝑥
−∞ −∞

〈𝐻 〉 = 𝐸
 6

Moreover, lets calculate the uncertainty in energy:

∆𝐻 = √〈𝐻 2 〉 − 〈𝐻 〉2
∞ ∞
〈𝐻 2〉
=∫ ∗( ̂𝐻
𝜓 𝑥) 𝐻 ̂⏟𝜓(𝑥 ) 𝑑𝑥 = 𝐸 ∫ ̂⏟𝜓(𝑥 ) 𝑑𝑥 = 𝐸 2
𝜓 ∗ (𝑥 ) 𝐻
−∞ 𝐸𝜓(𝑥) −∞ 𝐸𝜓(𝑥)

So the uncertainty in energy is zero:

∆𝐻 = √𝐸 2 − 𝐸 2

∆𝐻 = 0

It means that the distribution has zero spread; so that every member of the

ensemble must share the same value.

Conclusion: A steady state of the system has the property that every

measurement of the total energy is certain to return the value E.

(This is why we have chosen E for the separation constant)

 7

The most general solution of the Schrödinger Equation:

The most general solution is a linear combination of the stationary
states (the separable solutions).
Note that the Schrödinger equation yields an infinite collection of
solutions, 𝜓1 (𝑥 ), 𝜓2 (𝑥 ), 𝜓3 (𝑥 ), …, each with its associated value of the
separation constant, 𝐸1 , 𝐸2 , 𝐸3 , …, . Thus, there is a different wave
function for each allowed energy :

𝑖𝐸1 𝑡 𝑖𝐸2 𝑡
Ψ1 (𝑥, 𝑡) = 𝜓1 (𝑥 ) 𝑒 − ℏ , Ψ2 (𝑥, 𝑡) = 𝜓2 (𝑥 ) 𝑒 − ℏ ,…

Now, the time-dependent Schrödinger equation has the property that

any linear combination of the solutions is itself a solution. Thus, once
we have found the stationary state solutions, then we can immediately
construct a much more general solution of the form
∞

Ψ(𝑥, 𝑡) = 𝑐1 Ψ1 (𝑥, 𝑡) + 𝑐2 Ψ2 (𝑥, 𝑡) + 𝑐3 Ψ3 (𝑥, 𝑡) + ⋯ = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡)

𝑛=1

Or explicitly writing
𝑖𝐸1 𝑡 𝑖𝐸2 𝑡 𝑖𝐸3 𝑡
Ψ(𝑥, 𝑡) = 𝑐1 𝜓1 (𝑥 ) 𝑒 − ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 − ℏ + 𝑐3 𝜓3 (𝑥 ) 𝑒 − ℏ +⋯
∞
𝑖𝐸𝑛 𝑡
= ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ
𝑛=1
 8

That is
∞ ∞
𝑖𝐸𝑛 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ
𝑛=1 𝑛=1

It so happens that every solution to the time-dependent Schrödinger

equation can be written in this form- it is simply a matter of finding the
right constants 𝑐1 , 𝑐2 , 𝑐3 , …
 9

SUMMARY:
The generic problem is as follows:
You are given a time-independent potential 𝑉(𝑥 ), and the starting wave function
Ψ(𝑥, 0):
∞ ∞
𝑖𝐸 0
− 𝑛
Ψ(𝑥, 0) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 0) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒⏟ ℏ
𝑛=1 𝑛=1 1

Your job is to find the wave function Ψ(𝑥, 𝑡) for any subsequent time t.
To do this you must solve the time-dependent Schrödinger equation
𝜕Ψ(𝑥, 𝑡) ℏ2 𝜕 2 Ψ(𝑥, 𝑡)
𝑖ℏ =− + 𝑉 (𝑥 ) Ψ(𝑥, 𝑡)
𝜕𝑡 2𝑚 𝜕𝑥 2
The strategy is first to solve the time-independent equation (Due to the
technique of separation of variables)
−ℏ2 𝑑2
[ + 𝑉(𝑥)] 𝜓(𝑥 ) = 𝐸𝜓(𝑥 )
2𝑚 𝑑𝑥 2
This yields in general an infinite set of solution 𝜓1 (𝑥 ), 𝜓2 (𝑥 ), 𝜓3 (𝑥 ), …, each
with its associated energy, 𝐸1 , 𝐸2 , 𝐸3 , …,.(separation constant). Thus, there is a
different wave function for each allowed energy:
𝑖𝐸1 𝑡 𝑖𝐸2 𝑡
Ψ1 (𝑥, 𝑡) = 𝜓1 (𝑥 ) 𝑒 − ℏ , Ψ2 (𝑥, 𝑡) = 𝜓2 (𝑥 ) 𝑒 − ℏ ,…
𝑖𝐸𝑛 𝑡
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 ) 𝑒 − ℏ

Now, the time-dependent Schrödinger equation has the property that any
linear combination of the solutions is itself a solution. Then the most
general solution is
 10

∞ ∞
𝑖𝐸𝑛 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ
𝑛=1 𝑛=1

To fit Ψ(𝑥, 0) you write down the general linear combination of these
solutions
∞

Ψ(𝑥, 0) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 )
𝑛=1

by appropriate choice of the constants 𝑐1 , 𝑐2 , 𝑐3 , …

The separable solutions themselves

𝑖𝐸 𝑡
− 𝑛
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 ) 𝑒 ℏ

are stationary solutions, in the sense that all probabilities and expectation
values are independent of time.
However, this property is not shared by the general solution Ψ(𝑥, 𝑡); the
energies are different for different stationary states, and the exponentials do
not cancel when |Ψ(𝑥, 𝑡)|2 is calculated
 11

The wave function at subsequent time is given by

2
𝑖𝐸𝑛 𝑡 𝑖𝐸1 𝑡 𝑖𝐸2 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ = 𝑐1 𝜓1 (𝑥 ) 𝑒 − ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 − ℏ
𝑛=1

Such that at t=0, we recover the initial state of the particle given by
Ψ(𝑥, 0) = 𝑐1 𝜓1 (𝑥 ) + 𝑐2 𝜓2 (𝑥 )
Here, 𝐸1 and 𝐸2 are energies associated with 𝜓1 and 𝜓2 .
The probability density is
|Ψ(𝑥, 𝑡)|2 =
𝑖𝐸 𝑡 𝑖𝐸 𝑡 ∗ 𝑖𝐸 𝑡 𝑖𝐸 𝑡
− 1 − 2 − 1 − 2
(𝑐1 𝜓1 (𝑥 ) 𝑒 ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 ℏ ) (𝑐1 𝜓1 (𝑥 ) 𝑒 ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 ℏ )
𝑖𝐸1 𝑡 𝑖𝐸2 𝑡 𝑖𝐸1 𝑡 𝑖𝐸2 𝑡
= (𝑐1 𝜓1 (𝑥 ) 𝑒 ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 ℏ ) (𝑐1 𝜓1 (𝑥 ) 𝑒− ℏ + 𝑐2 𝜓2 (𝑥 ) 𝑒 − ℏ )

𝑖(𝐸2 −𝐸1 )𝑡 𝑖(𝐸2 −𝐸1 )𝑡

−
= 𝑐12 𝜓12 + 𝑐22 𝜓22 + 𝑐1 𝜓1 𝑐2 𝜓2 𝑒 ℏ + 𝑐2 𝜓2 𝑐1 𝜓1 𝑒 ℏ

𝑖(𝐸2 −𝐸1 )𝑡 𝑖(𝐸2 −𝐸1 )𝑡

= 𝑐12 𝜓12 + 𝑐22 𝜓22 + 𝑐1 𝑐2 𝜓1 𝜓2 (𝑒 ℏ + 𝑒− ℏ )

(𝐸2 − 𝐸1 )𝑡
|Ψ(𝑥, 𝑡)|2 = 𝑐12 𝜓12 + 𝑐22 𝜓22 + 2𝑐1 𝑐2 𝜓1 𝜓2 cos
ℏ
 12

The Infinite Square Well

Let us consider the potential
∞ , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑉(𝑥) = {
0 0<𝑥<𝐿
illustrated in the next figure.
This means that the quantum object is limited to
a certain region between 𝑥 = 0 and 𝑥 = 𝐿 where
It moves freely but cannot leave.

A classical model would be a car on a

frictionless horizontal air track with
perfectly elastic bumpers, that just
keep the car bouncing back and forth
forever.

This is the simplest model that can be used for illustrating the quantum
mechanics. It is a hypothetical example, used to illustrate the differences
between classical and quantum systems.

Although it is quite a trivial one, it introduces various important

concepts, it may also itself be used as a first approximation to some
actual physical problems.
 13

In classical mechanics, the solution to the problem is trivial: We apply

the Newton’s laws of motion and initial conditions. Particle moves in a
straight line always at the same speed until it reflects from a bumper. If
the particle is given a velocity 𝑣, its motion between the walls is
x  x0  vt .

2L 2L 2m 2m
The period of motion is T  L L
v p/m 2mE E

 However, when the well becomes very narrow (on the scale of a
few nanometers), quantum effects become important. The problem
becomes very interesting when one attempts a quantum mechanical
solution with applying Schrodinger equation, since many fundamental
quantum mechanical concepts need to be introduced to find the solution.
Nevertheless, it remains a very simple and solvable problem.
The possible de Broglie wavelengths of the particle in the box are
determined by L.
 14

Solution to the Quantum Problem

∞ , 𝑥 ≤ 0, 𝑥 ≥ 𝐿 (DI,DIII)
𝑉(𝑥) = {
0 0 < 𝑥 < 𝐿 (DII)

This potential confines the particle to the

region 𝑥 ∈ [0, 𝐿]. Potential is discontinuous.

1-dimensional time-independent Schrodinger equation is

−ℏ2 𝑑2
̂
𝐻 𝜓(𝑥 ) = 𝐸𝜓(𝑥 ) => [ + 𝑉 (𝑥 )] 𝜓(𝑥 ) = 𝐸𝜓(𝑥 ) (1)
2𝑚 𝑑𝑥 2

 DI where 𝒙 ≤ 𝟎 ,we have

−ℏ2 𝑑 2
[ + ∞] 𝜓𝐼 (𝑥 ) = 𝐸𝐼 𝜓𝐼 (𝑥 ) when 𝑥 ≤ 0
2𝑚 𝑑𝑥 2

 DIII where 𝒙 ≥ 𝑳 ,we have

−ℏ2 𝑑2
[ + ∞] 𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐸𝐼𝐼𝐼 𝜓𝐼𝐼𝐼 (𝑥 ) when 𝒙 ≥ 𝑳
2𝑚 𝑑𝑥 2

For finite 𝐸𝐼 , 𝐸𝐼𝐼𝐼 and 𝜓𝐼 (𝑥 ), 𝜓𝐼𝐼𝐼 (𝑥 ), RHS is finite. This implies that
LHS must be finite too. This may be possible if

𝜓𝐼 (𝑥 ) = 0
𝜓𝐼𝐼𝐼 (𝑥 ) = 0
 15

This means that

| 𝜓𝐼 (𝑥 )|2 = | 𝜓𝐼𝐼𝐼 (𝑥 )|2 = 0

∶ The probability of finding the particle in DI and DIII is zero

DI and DIII are the examples for forbidden domains, where E<V.
(E≡ 𝐾 + 𝑉 𝑎𝑛𝑑 𝑖𝑓 𝐸 < 𝑉 𝑡ℎ𝑖𝑠 𝑚𝑒𝑎𝑛𝑠 𝐾 < 0 !)

 DII where 𝟎 ≤ 𝒙 ≤ 𝑳 ,we have V(x)=0

−ℏ2 𝑑2
[ + 0] 𝜓𝐼𝐼 (𝑥 ) = 𝐸𝜓𝐼𝐼 (𝑥 ) when 0 < 𝑥 < 𝐿 (2)
2𝑚 𝑑𝑥 2

At the boundaries x=0 and x=L, potential rises abruptly to ∞ and

To find 𝜓𝐼𝐼 (𝑥 ), rewrite eq.(2):

𝜕 2 𝜓𝐼𝐼 (𝑥 ) 2𝑚𝐸 2𝑚𝐸

+ 2 𝜓𝐼𝐼 (𝑥 ) = 0 , or with 𝑘2 ≡
𝜕𝑥 2 ℏ ℏ2

𝜕 2 𝜓𝐼𝐼 (𝑥 )
2
+ 𝑘 2 𝜓𝐼𝐼 (𝑥 ) = 0
𝜕𝑥

This is a well-studied differential equation with a general solution

𝜓𝐼𝐼 (𝑥 ) = 𝐴 sin 𝑘𝑥 + 𝐵 cos 𝑘𝑥

OR,

𝜓𝐼𝐼 (𝑥 ) = 𝐶 eikx + 𝐷 e−ikx

Here A,B (or C,D) are constants, and k can be any real number.
 16

Now, to find the special solution, we must specify the boundary

conditions, and find A and B that satisfy these conditions:

Boundary Conditions:
1. 𝜓𝐼 (𝑥 )|𝑥=0 = 𝜓𝐼𝐼 (𝑥 )|𝑥=0 = 0 ⟹ 𝐵 cos 𝑘0 = 0 ⟹ 𝐵 = 0

Thus ; 𝜓𝐼𝐼 (𝑥 ) = 𝐴 sin 𝑘𝑥 .

2. 𝜓𝐼𝐼 (𝑥 )|𝑥=𝐿 = 𝜓𝐼𝐼𝐼 (𝑥 )|𝑥=𝐿 = 0 ⟹ 𝐴 sin 𝑘𝐿 = 0 ⟹

𝑘𝐿 = ±𝜋, ±2𝜋, ±3𝜋, …

However, note that negative solutions give noting new since

sin(−𝜃 ) = − sin(𝜃 ),

and this minuse sign can be absorbed into the constant A. (Also, −sin(𝜃 )
gives the same probability when squared)

Therefore the distinct solutions are

𝑛𝜋
𝜓𝐼𝐼 (𝑥 ) ≡ 𝜓𝑛 (𝑥 ) = 𝐴 sin 𝑘𝑛 𝑥 with 𝑘𝑛 = , 𝑛 = 1,2,3, …
𝐿

Note that n=0 solution is ruled out because in this case 𝜓𝐼𝐼 (𝑥 ) = 0 and
the whole solution to Sch.eq, is zero. This corresponds to the case
where there is no the particle in the box !.
 17

 To find the constant A , we must normalize the wave function 𝜓𝑛 (𝑥 ):

𝐿 𝐿
𝑛𝜋
1 = ∫ | 𝜓𝑛 (𝑥 )|2 𝑑𝑥 = |𝐴 ∫ sin2 ( 𝑥) 𝑑𝑥 = 1
|2
0 0 𝐿

𝑥 sin(2𝛽𝑥)
Using ∫ sin2 (𝛽𝑥 ) 𝑑𝑥 = − , evaluate the integral above
2 4𝛽

𝐿
𝑥 𝑛𝜋 1 𝐿 2
|2
|𝐴 [ − sin (2 𝑥) ] = 1 ⟹ |𝐴|2 [ − 0] = 1 ⟹ 𝐴 = √
2 𝐿 4𝑛𝜋/𝐿 0 2 𝐿

Finally, substituting this result

2 𝑛𝜋
𝜓𝑛 (𝑥 ) = √ sin ( 𝑥) ∶ eigenfunction of H
𝐿 𝐿

ℏ2 𝑘𝑛2 ℏ2 𝜋 2 2
𝐸𝑛 = = 𝑛 , 𝑛 = 1,2,3, … ∶ energy eigenvalues of the H
2𝑚 2𝑚𝐿2

Each permitted energy is called an energy level, and the integer n

that specifies an energy level 𝑬𝒏 is called its quantum number.

The ground state energy: 𝑛 = 1:

ℏ2 𝜋 2
𝐸1 = ∶ the ground state energy
2𝑚𝐿2

Then, in terms of ground state energy, the quantized energy levels are

written as 𝐸𝑛 = 𝐸1 𝑛2 , 𝑛 = 1,2,3, …
 18

The solution to the Schrödinger eq. for the particle in the infinite well
problem reveals some quantum behavior of the particle that contrasts
sharply with the predictions of classical mechanics.

The quantum behavior includes:

 Energy quantization: Quantized energy levels are possible. It is not
possible for the particle to have any arbitrary energy. Instead, only
discrete energy levels are allowed. The spacing between successive
levels gets larger as the energy increases.
 19

 Ground State Energy: is the lowest energy of the particle and it is

obtained when n=1:
ℏ2 𝜋 2
𝐸1 = ≠0
2𝑚𝐿2
Note that classically the lowest energy would be that of a particle at rest
and for p=0 and V=0, E=0. However, in quantum solution since n  0 ,
the lowest energy is non-zero.

This non-zero ground state energy can be also explained in terms of

the uncertainty principle:
When, then
𝑝2
𝑛 = 0 => 𝐸 = 𝐾𝐸 = =0
2𝑚
( since V=0 already). This implies that particle is at rest:
𝑝=0
which means that particle has a well definite momentum =>
∆p = 0 => ∆𝑥 = ∞
However, particle is confined in a limited region so that we at most have
∆𝑥 = 𝐿
=> CONTRADICTION! Thus, particle can not be at rest !
 20

 The waves are in the form of standing waves on a string of length L with
wavelength such that it fits into the space available (Figure). We see that
the wavelength of the wave is restricted to one of the values corresponding
to a whole number of half wavelengths fitting into the box:
2𝜋 𝑛𝜋 2𝐿
𝜆= and 𝑘𝑛 = => 𝜆𝑛 =
𝑘 𝐿 𝑛
This means that only these particular values of the wavelength are
allowed and, as the electron momentum is determined by the wavelength
through the de Broglie relation, the momentum is also restricted to a
particular set of values.
2𝐿 ℎ
𝜆𝑛 = =
𝑛 𝑝𝑛

 Spatial Nodes: In contrast to the classical mechanics, the Sch. Eq.

predicts that for some energy levels there are nodes (zero-crossing),
implying positions at which the particle can never be found. As you go
up in energy, each successive state has one more node: 𝜓1 has none, 𝜓2
has one, 𝜓3 has two, and so on.
Therefore at a particular place in the box, the probability of the particle
being present may be very different for different n. For instance,  1 has
2

its maximum at x=L/2, while  2 =0 there:

2

A particle in the lowest energy level n=1 is most likely to be in the

middle of the box, while a particle in the next higher state of n=2 is
never there.
 21

𝐿
 In all quantum states, 〈𝑥 〉 =
2

𝐿 𝐿
2 𝑛𝜋𝑥
〈𝑥 〉 = ∫ 𝑥 | 𝜓𝑛 (𝑥 )|2 𝑑𝑥 = ∫ 𝑥 sin2 ( ) 𝑑𝑥
0 𝐿 0 𝐿

𝐿 𝐿 𝐿
= + (1 − cos
⏟ 2𝑛𝜋 ) − sin 2𝑛𝜋
⏟
2 4(𝑛𝜋)2 1
2𝑛𝜋 0

𝐿
〈𝑥 〉 =
2

EX: Find the probability that a particle trapped in a box L wide can be
found between (0.45L,0.55L) for the ground state and for the 1.st excited
state.
In general,
𝑥2
2 𝑥2 2 𝑛𝜋𝑥
𝑃𝑛 (𝑥1 , 𝑥2 ) = ∫ 𝜓𝑛∗ (𝑥 )
𝜓𝑛 (𝑥 ) 𝑑𝑥 = ∫ sin ( ) 𝑑𝑥
𝑥1 𝐿 𝑥1 𝐿
𝑥2
2𝐿 𝑥 1 2𝜋𝑛
𝑃𝑛 (𝑥1 , 𝑥2 ) = [ − sin ( 𝑥)]
𝐿 2 𝐿 2𝑛𝜋 𝐿 𝑥1

0.55𝐿
𝑥 1 2𝜋
𝑃1 (0.45𝐿, 0.55𝐿) = [ − sin ( 𝑥)] = 0.198 ⟹
𝐿 2𝜋 𝐿 0.45𝐿

𝑷𝟏 (𝟎. 𝟒𝟓𝑳, 𝟎. 𝟓𝟓𝑳)~𝟐𝟎 %

0.55𝐿
𝑥 1 4𝜋
𝑃2 (0.45𝐿, 0.55𝐿) = [ − sin ( 𝑥)] = 0.006 ⟹
𝐿 4𝜋 𝐿 0.45𝐿

𝑷𝟐 (𝟎. 𝟒𝟓𝑳, 𝟎. 𝟓𝟓𝑳)~𝟎. 𝟔 %

 22

The most general solution to the Schrödinger equation is a linear

combination of stationary states Ψ𝑛 (𝑥, 𝑡):
∞

Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡)
𝑛=1

The stationary states of the infinite square well are written from
𝑖𝐸 𝑡
− 𝑛
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 ) 𝑒 ℏ =>
as
2 𝑛𝜋 ℏ𝜋2 𝑛2
−𝑖 𝑡
Ψ𝑛 (𝑥, 𝑡) = √ sin ( 𝑥) 𝑒 2𝑚𝐿2
𝐿 𝐿

Therefore, the most general solution is given by

∞
2 𝑛𝜋 ℏ𝜋2 𝑛2
−𝑖 𝑡
Ψ(𝑥, 𝑡) = √ ∑ 𝑐𝑛 sin ( 𝑥) 𝑒 2𝑚𝐿2 (∗)
𝐿 𝐿
𝑛=1

How to calculate the coefficients 𝑐𝑛 ?

-------------------------------------------------------------------------
Orthogonality of the eigenfunctions

𝜓𝑚 (𝑥 )′𝑠 are mutually orthogonal, in the sense that

𝐿
∫ 𝜓𝑛∗ (𝑥 ) 𝜓𝑚 (𝑥 ) 𝑑𝑥 = 0 𝑤ℎ𝑒𝑛𝑒𝑣𝑒𝑟 𝑛 ≠ 𝑚
0

Proof:
𝐿
𝐿
2 𝑛𝜋 𝑚𝜋
∫ 𝜓𝑛∗ (𝑥 ) 𝜓𝑚 (𝑥 ) 𝑑𝑥 = ∫ sin ( 𝑥) sin ( 𝑥) 𝑑𝑥
0 𝐿 𝐿 𝐿
0
𝐿
12 (𝑚 − 𝑛)𝜋 (𝑚 + 𝑛)𝜋
= ∫ Cos ( 𝑥) − Cos ( 𝑥) 𝑑𝑥
2𝐿 𝐿 𝐿
0
 23

𝐿
1 𝐿 (𝑚 − 𝑛)𝜋 𝐿 (𝑚 + 𝑛)𝜋
= { Sin ( 𝑥) − Sin ( 𝑥)}
𝐿 (𝑚 − 𝑛)𝜋 𝐿 (𝑚 + 𝑛)𝜋 𝐿 0

1 (𝑚 − 𝑛)𝜋 (𝑚 + 𝑛)𝜋
= {Sin ( ) − Sin ( )} = 0
𝜋 (𝑚 − 𝑛) (𝑚 + 𝑛)
where we used
1
sin(𝑎𝑥 ) sin(𝑏𝑥 ) = [cos(𝑎 − 𝑏) 𝑥 − cos(𝑎 + 𝑏) 𝑥 ]
2
When 𝑛 ≠ 𝑚, normalization tells us that integral is unity. In fact we can
combine orthogonality and normalization into a single statement:

𝐿
0 𝑖𝑓 𝑛 ≠ 𝑚
∫ 𝜓𝑛∗ (𝑥 ) 𝜓𝑚 (𝑥 ) 𝑑𝑥 = 𝛿𝑛𝑚 = { }
0
1 𝑖𝑓 𝑛 = 𝑚

𝜓𝑚 (𝑥 )′𝑠 are complete, in the sense that any other function 𝑓(𝑥) can be
expressed as a linear combination of them
∞ ∞
2 𝑛𝜋
( ) ( )
𝑓 𝑥 = ∑ 𝑐𝑛 𝜓𝑛 𝑥 = √ ∑ 𝑐𝑛 sin ( 𝑥)
𝐿 𝐿
𝑛=1 𝑛=1
This is nothing but the Fourier series for 𝑓(𝑥 ) and the fact that any
function can be expanded in this way is called Dirichlet’s theorem.
∗ ( )
The coefficients 𝑐𝑛 can be obtained by multiplying both sides by 𝜓𝑚 𝑥
and integrating:
∞ ∞
∗ ( ) ( ) ∗ ( )
∫ 𝜓𝑚 𝑥 𝑓 𝑥 𝑑𝑥 = ∑ 𝑐𝑛 ∫ 𝑑𝑥 𝜓𝑛 (𝑥 )𝜓𝑚 𝑥 = ∑ 𝑐𝑛 𝛿𝑛𝑚 = 𝑐𝑚
𝑛=1 𝑛=1

Note that 𝛿𝑛𝑚 kills every term in the sum except the one for which n=m.
Therefore
𝑐𝑛 = ∫ 𝜓𝑛∗ (𝑥 ) 𝑓(𝑥 )𝑑𝑥
------------------------------------------------------------------
 24

 If the initial wave function is given 𝚿(𝒙, 𝟎) and we are asked to find the
most general solution to the time dependent Schrödinger equation:
1. Find 𝑐𝑛 . For this start from (*) and get Ψ(𝑥, 0):
∞ ∞ ∞
2 𝑛𝜋
Ψ(𝑥, 0) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 0) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) = √ ∑ 𝑐𝑛 sin ( 𝑥)
𝐿 𝐿
𝑛=1 𝑛=1 𝑛=1

Thus,

2 𝑛𝜋
𝑐𝑛 = ∫ 𝜓𝑛∗ (𝑥 ) Ψ(𝑥, 0)𝑑𝑥 = √ ∫ sin ( 𝑥) Ψ(𝑥, 0)𝑑𝑥 (∗∗)
𝐿 𝐿

2. Plug this into eq.(*)

 25

Example 2.2. A particle in the infinite square well has the initial wave function.
At time t=0, a particle is represented by the wave function
𝐴 𝑥 (𝐿 − 𝑥 ) 𝑖𝑓 0 ≤ 𝑥 ≤ 𝐿
Ψ(𝑥, 0) = {
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Find Ψ(𝑥, 𝑡).
Since
∞
2 𝑛𝜋 ℏ𝜋2 𝑛2
−𝑖 𝑡
Ψ(𝑥, 𝑡) = √ ∑ 𝑐𝑛 sin ( 𝑥) 𝑒 2𝑚𝐿2 (∗)
𝐿 𝐿
𝑛=1

We need to find 𝑐𝑛 .
First normalize it:
𝐿 𝐿

∫|𝜓(𝑥, 0)|2 𝑑𝑥 = 1 => 𝐴2 ∫ 𝑥 2 (𝐿 − 𝑥)2 𝑑𝑥 = 1 =

0 0
𝐿

> 𝐴2 ∫ 𝑥 2 (𝑥 2 − 2𝐿𝑥 + 𝐿2 ) 𝑑𝑥 = 1
0
5 𝐿 𝐿 𝐿
2
𝑥 𝑥4 2
𝑥3 2
𝐿5 𝐿5 𝐿5
1 = 𝐴 { | − 2𝐿 | + 𝐿 | }=𝐴 { − + }
5 0 4 0 3 0 5 2 3
𝐿5 2
=𝐿
30
30
𝐴=√
𝐿5

Then use (**):

𝐿
30 2 𝑛𝜋
𝑐𝑛 = ∫ 𝜓𝑛∗ (𝑥 ) Ψ(𝑥, 0)𝑑𝑥 = √ √ ∫ sin ( 𝑥) 𝑥(𝐿 − 𝑥)𝑑𝑥
𝐿5 𝐿 𝐿
0
𝐿 𝐿
2√15 𝑛𝜋 𝑛𝜋
= 3
{𝐿 ∫ sin ( 𝑥) 𝑥𝑑𝑥 − ∫ sin ( 𝑥) 𝑥 2 𝑑𝑥}
𝐿 𝐿 𝐿
0 0
 26

2√15 𝐿 2 𝑛𝜋 𝑛𝜋 𝑛𝜋
= 3 {𝐿 [( ) (sin ( 𝑥) − 𝑥 Cos ( 𝑥))]
𝐿 𝑛𝜋 𝐿 𝐿 𝐿
𝐿
𝐿 3 𝑛𝜋 𝑛𝜋 𝑛𝜋 𝑛𝜋 2 𝑛𝜋
− ( ) (2 Cos ( 𝑥) + 2 𝑥 Sin ( 𝑥) − ( 𝑥) Cos ( 𝑥))}
𝑛𝜋 𝐿 𝐿 𝐿 𝐿 𝐿 0

2√15 𝐿 2 𝑛𝜋
= 3 {[𝐿 [( ) (0 − 𝐿 Cos(𝑛𝜋))]
𝐿 𝑛𝜋 𝐿
𝐿 3 𝑛𝜋 2
− ( ) (2 Cos(𝑛𝜋) + 0 − ( 𝐿) Cos(𝑛𝜋))]
𝑛𝜋 𝐿
𝐿 2 𝐿 3
− [𝐿 [( ) (0 − 0)] − ( ) (2 Cos(0) + 0 − 0)]}
𝑛𝜋 𝑛𝜋

2√15 𝐿3 𝐿 3 𝐿 3
= 3 { Cos(𝑛𝜋) + ( ) ((𝑛𝜋)2 − 2) Cos(𝑛𝜋) + ( ) 2 Cos(0)}
𝐿 𝑛𝜋 𝑛𝜋 𝑛𝜋
0 𝑖𝑓 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛
4√15
= {Cos(0) − Cos(𝑛𝜋)} => 𝑐𝑛 = { 8√15 }
(𝑛𝜋)3 𝑖𝑓 𝑛 𝑖𝑠 𝑜𝑑𝑑
(𝑛𝜋)3
Put into (*):
∞
2 8√15 1 𝑛𝜋 ℏ𝜋2 𝑛2
−𝑖 𝑡
Ψ(𝑥, 𝑡) = √ ∑ sin ( 𝑥) 𝑒 2𝑚𝐿2
𝐿 (𝜋)3 𝑛3 𝐿
𝑛=1,3,5,…
OR
∞
30 2 3 1 𝑛𝜋 −𝑖
ℏ𝜋2 𝑛2
𝑡
Ψ(𝑥, 𝑡) = √ ( ) ∑ sin ( 𝑥) 𝑒 2𝑚𝐿2
𝐿 𝜋 𝑛3 𝐿
𝑛=1,3,5,…
 27

Some useful formula for integration:

1
∫ sin(𝑏𝑥 ) 𝑥𝑑𝑥 = 2 (𝑆𝑖𝑛(𝑏𝑥 ) − 𝑏𝑥𝐶𝑜𝑠(𝑏𝑥 ))
𝑏
1
∫ sin(𝑏𝑥 ) 𝑥 2 𝑑𝑥 = 3 (2𝐶𝑜𝑠(𝑏𝑥 ) + 2𝑏𝑥𝑆𝑖𝑛(𝑏𝑥 ) − (𝑏𝑥 )2 𝐶𝑜𝑠(𝑏𝑥 ))
𝑏

sin((𝑎 − 𝑏)𝑥 ) sin((𝑎 + 𝑏)𝑥 )

∫ sin(𝑎𝑥 ) sin(𝑏𝑥 ) 𝑑𝑥 = −
2(𝑎 − 𝑏) 2(𝑎 + 𝑏)

1
sin(𝑎𝑥 ) sin(𝑏𝑥 ) = [cos(𝑎 − 𝑏) 𝑥 − cos(𝑎 + 𝑏) 𝑥 ]
2

-------------------------------------------------------------------------------
The meaning of 𝑐𝑛 :

Loosely speaking, 𝑐𝑛 tells us the “amount of 𝜓𝑛 (𝑥 ) that is contained in

Ψ.

|𝑐𝑛 |2 :
𝑡ℎ𝑒 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑎 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑛𝑒𝑟𝑔𝑦 𝑤𝑜𝑢𝑙𝑑 𝑦𝑖𝑒𝑙𝑑
𝑡ℎ𝑒 𝑣𝑎𝑙𝑢𝑒 𝐸𝑛

The sum of probabilities should be one,

∞

∑ |𝑐𝑛 |2 = 1
𝑛=1

Indeed this follows from the normalization of Ψ.

∗
∞ ∞

1 = ∫|Ψ(𝑥, 0)|2 𝑑𝑥 = ∫ (∑ 𝑐𝑛 𝜓𝑛 (𝑥 )) ( ∑ 𝑐𝑚 𝜓𝑚 (𝑥 )) 𝑑𝑥
𝑛=1 𝑚=1
 28

∞ 𝐿 ∞ ∞

= ∑ 𝑐𝑛∗ 𝑐𝑚 ∫ 𝜓𝑛∗ (𝑥 ) 𝜓𝑚 (𝑥 ) 𝑑𝑥 = ∑ 𝑐𝑛∗ 𝑐𝑚 𝛿𝑛𝑚 = ∑ |𝑐𝑛 |2

𝑛,𝑚 0 𝑛,𝑚 𝑛=1

Moreover, the expectation value of the energy in the most general state
is calculated directly:
∗
∞ ∞

〈𝐻 〉 = ∫ Ψ ∗ 𝐻Ψdx = ∫ (∑ 𝑐𝑛 𝜓𝑛 (𝑥 )) 𝐻 ( ∑ 𝑐𝑚 𝜓𝑚 (𝑥 )) 𝑑𝑥
𝑛=1 𝑚=1
∞ 𝐿 ∞ 𝐿
= ∑ 𝑐𝑛 𝑐𝑚 ∫ 𝜓𝑛∗ (𝑥 ) 𝐻𝜓
∗
⏟ 𝑚 (𝑥 ) 𝑑𝑥 = ∑ 𝑐𝑛 𝑐𝑚 𝐸𝑚 ∫ 𝜓𝑛∗ (𝑥 )𝜓𝑚
∗
𝑑𝑥
𝑛,𝑚 0 =𝐸𝑚 𝜓𝑚 𝑛,𝑚
⏟0
𝛿𝑛𝑚
∞ ∞

= ∑ 𝑐𝑛∗ 𝑐𝑛 𝐸𝑛 => 〈𝐻 〉 = ∑ |𝑐𝑛 |2 𝐸𝑛

𝑛 𝑛=1

Note that
|𝑐𝑛 |2 = probability of getting a particular energy is independent of time

〈𝐻 〉: independent of time
This is a manifestation of conservation of energy
 29

Expectation value of momentum of the particle in the infinite well:

We have shown that
𝐿 𝑑 〈𝑥 〉
〈𝑥 〉 = => 〈𝑝〉 = 𝑚 =0
2 𝑑𝑡
-------------------------------------------------------------------------

OR, we can show this by evaluating the following integral:

∞ ∞
𝑑
〈𝑝〉 = ∫ 𝜓 𝑝̂ 𝜓 𝑑𝑥 = ∫ 𝜓 ∗ (−𝑖ℏ
∗
) 𝜓 𝑑𝑥
−∞ −∞ 𝑑𝑥

2 𝑛𝜋 𝑑 2 𝑛𝜋 𝑛𝜋
𝜓𝑛 = 𝜓𝑛∗ = √ sin ( 𝑥) ; 𝜓𝑛 = √ Cos ( 𝑥)
𝐿 𝐿 𝑑𝑥 𝐿 𝐿 𝐿
2
2 𝑛𝜋 𝐿 𝑛𝜋 𝑛𝜋
〈𝑝〉 = −𝑖ℏ (√ ) ∫ sin ( 𝑥) Cos ( 𝑥) 𝑑𝑥
𝐿 𝐿 0 𝐿 𝐿
1
Using ∫ sin 𝛼𝑥 cos 𝛼𝑥 𝑑𝑥 = 𝑆𝑖𝑛2 𝛼𝑥 ,
2𝛼

2 𝑛𝜋 𝐿 2
𝑛𝜋 𝐿 −𝑖ℏ
〈𝑝〉 = −𝑖ℏ 𝑆𝑖𝑛 ( 𝑥) = [𝑆𝑖𝑛2 𝑛𝜋 − 0] = 0
𝐿 𝐿 2𝑛𝜋 𝐿 0 𝐿
〈𝑝 〉 = 0
----------------------------------------------------------------------
Note that for any real function 𝑅 ∗ (𝑥 ) = 𝑅 (𝑥 ),
∞
𝑑
〈𝑝〉 = ∫ 𝑅 ∗ (−𝑖ℏ ) 𝑅 𝑑𝑥
−∞ 𝑑𝑥
is imaginary (because of i), which is incompatible with the requirement
that 〈𝑝〉 = 〈𝑝〉∗ unless 〈𝑝〉 = 0
-------------------------------------------------------------------------
 30

Note also that since

𝑝𝑛2 2
ℏ2 𝜋 2 2
𝐸𝑛 = , and 𝐸𝑛 = 𝐸1 𝑛 = 𝑛
2𝑚 2𝑚𝐿2

𝑛𝜋ℏ
𝑝𝑛 = ±√2𝑚𝐸𝑛 = ±
𝐿

The particle is moving back and forth , so its average is

𝑛𝜋ℏ 𝑛𝜋ℏ
(+ ) + (− )
𝑝̅ = 𝐿 𝐿 =0
2
 31

Momentum Eigenvalue:

What is the solution to the momentum eigenvalue equation ?

𝑝̂ 𝜙(𝑥) = 𝜆𝜙(𝑥)
We know that in the infinite well problem, energy eigenvalue equation
̂ 𝜓𝑛 = 𝐸𝑛 𝜓𝑛
𝐻
has the solution

2 𝑛𝜋 ℏ2 𝜋 2 2
𝜓𝑛 (𝑥) = √ sin ( 𝑥) , 𝐸𝑛 = 𝑛 , 𝑛 = 1,2,3, ….
𝐿 𝐿 2𝑚𝐿2

Let us check if 𝜓𝑛 (𝑥) is also an eigenfunction of the momentum

operator :
?
𝑝̂ 𝜓𝑛 (𝑥) =
⏞ 𝑝𝑛 𝜓𝑛 (𝑥)

𝑑 𝑑 2 𝑛𝜋
𝐿𝐻𝑆 = 𝑝̂ 𝜓𝑛 = −𝑖ℏ 𝜓𝑛 = −𝑖ℏ [√ sin ( 𝑥)]
𝑑𝑥 𝑑𝑥 𝐿 𝐿

2 𝑛𝜋 𝑛𝜋
= −𝑖ℏ√ Cos ( 𝑥) ≠ (a constant)𝜓𝑛
𝐿 𝐿 𝐿

NO. 𝜓𝑛 is not an eigenfunction of 𝑝̂ operator.

 32

However, note that using

𝑒 𝑖𝜃 − 𝑒 −𝑖𝜃
sin 𝜃 =
2𝑖
We can write 𝜓𝑛 (𝑥) as

2 𝑛𝜋 2 1 𝑖𝑛𝜋𝑥 1 −𝑖𝑛𝜋𝑥
𝜓𝑛 (𝑥 ) = √ sin ( 𝑥) = √ ( 𝑒 𝐿 − 𝑒 𝐿 ) ≡ 𝜓𝑛+ − 𝜓𝑛−
𝐿 𝐿 𝐿 2𝑖 2𝑖

𝜓𝑛± ≡

𝑑 + 𝑑 2 1 𝑖𝑛𝜋𝑥
𝑝̂ 𝜓𝑛+ = 𝑝𝑛+ 𝜓𝑛+ ⟹ −𝑖ℏ 𝜓𝑛 = −𝑖ℏ [√ 𝑒 𝐿 ]
𝑑𝑥 𝑑𝑥 𝐿 2𝑖

𝑖𝑛𝜋 2 1 𝑖𝑛𝜋𝑥 𝑛𝜋ℏ + 𝑛𝜋ℏ

= −𝑖ℏ √ 𝑒 𝐿 = 𝜓𝑛 = 𝑝𝑛+ 𝜓𝑛+ ⟹ 𝑝𝑛+ =
𝐿 ⏟𝐿 2𝑖 𝐿 𝐿
+
𝜓𝑛

Similarly,

𝑑 − 𝑑 2 1 𝑖−𝑛𝜋𝑥
𝑝̂ 𝜓𝑛− = 𝑝𝑛− 𝜓𝑛− ⟹ −𝑖ℏ 𝜓𝑛 = −𝑖ℏ [−√ 𝑒 𝐿 ]
𝑑𝑥 𝑑𝑥 𝐿 2𝑖

−𝑖𝑛𝜋 2 1 𝑖 −𝑛𝜋𝑥 −𝑛𝜋ℏ −

= −𝑖ℏ ( ) (−√ 𝑒 𝐿 )= 𝜓𝑛 = 𝑝𝑛− 𝜓𝑛−
𝐿 𝐿 2𝑖 𝐿
⏟
−
𝜓𝑛

−𝑛𝜋ℏ
⟹ 𝑝𝑛− =
𝐿
 33

We conclude that 𝜓𝑛+ and 𝜓𝑛− are indeed the momentum

eigenfunctions for a particle in a box:
𝑖
𝑛𝜋𝑥 𝑛𝜋ℏ
𝑝̂ 𝜓𝑛+ = 𝑝𝑛+ 𝜓𝑛+ ; 𝜓𝑛+ ~ 𝑒 𝐿 with the eigenvalue 𝑝𝑛+ =
𝐿
−𝑖
𝑛𝜋𝑥 𝑛𝜋ℏ
𝑝̂ 𝜓𝑛− = 𝑝𝑛− 𝜓𝑛− ; 𝜓𝑛− ~ 𝑒 𝐿 with the eigenvalue 𝑝𝑛− = −
𝐿
 1

THE HARMONIC OSCILLATOR

A study of the Simple Harmonic Oscillator (SHO) is important in QM. Since there are only
limited number of continuous potentials for which it is possible to obtain solutions to the
Schrödinger equation by analytical techniques, and SHO potential is one of them.
 Coulomb potential : 𝑉(𝑟)~𝑟 −1
 SHO potential : 𝑉(𝑟)~𝑟 2
In addition, SHO is the prototype for any system involving oscillations. The system may be an
object supported by a spring or a diatomic molecule

The condition for harmonic motion is the presence of a restoring force that acts to return the
system to its equilibrium configuration when it is disturbed.

Note that any restoring force 𝐹(𝑥) can be expressed in a Maclaurin’s series about the
equilibrium position (take x=0) as
𝑑𝐹 1 𝑑𝐹 2
𝐹(𝑥) = 𝐹(0) + ( ) 𝑥+ ( ) 𝑥 2 + ⋯.
𝑑𝑥 𝑥=0 2 𝑑𝑥 𝑥=0

Since x=0 is the equilibrium position, 𝐹(0) = 0. For small x, keep the 1.st order term only
(higher orders are smaller). Then,
𝑑𝐹
𝐹(𝑥) ≅ ( ) 𝑥
𝑑𝑥 𝑥=0
𝑑𝐹
which is the Hooke’s law when ( ) < 0.
𝑑𝑥 𝑥=0

𝐹(𝑥) = −𝑘𝑥
 2

From the 2.nd law of motion

𝐹(𝑥) = 𝑚𝑎
we have
𝑑2𝑥 𝑑2𝑥 𝑘
−𝑘𝑥 = 𝑚 2 => 2 + 𝑥 = 0
𝑑𝑡 𝑑𝑡 𝑚
With 𝜔2 = 𝑘/𝑚 , the angular frequency,
𝑑2𝑥
2
+ 𝜔2 𝑥 = 0
𝑑𝑡
The general solution to this equation :
𝑥(𝑡) = 𝐴 cos(𝜔𝑡 + 𝛿)
where 𝐴 is the amplitude and 𝛿 is a phase factor.
The potential energy that corresponds to a Hooke’s Law force may be found by calculating the
work needed to bring a particle from 𝑥 = 0 to any point 𝑥 against such a force:
𝑥 𝑥
𝑑𝑉(𝑥) 1
𝐹(𝑥) = − => 𝑉(𝑥) = − ∫ 𝐹(𝑥 ′ )𝑑𝑥 ′ = 𝑘 ∫ 𝑥 ′ 𝑑𝑥 ′ => 𝑉(𝑥) = 𝑘𝑥 2
𝑑𝑥 2
0 0

which has the form

If the energy of the oscillator is E,

particle vibrates back and forth between
𝑥 = −𝐴 and 𝑥 = 𝐴, where E and A are
related by
1
𝐸 = 𝑘𝐴2
2

The system has continuous energy starting from zero. Energy can have any value since A is
arbitrary.
However, quantum mechanics predicts that the total energy 𝐸 can assume only a discrete set of
values because the particle is bound by the potential to a region of finite extent. To find the
allowed energy values predicted by Schrödinger quantum mechanics, the time-independent
Schrödinger equation for the SHO potential must be solved.
 3

 Classical probability of finding the particle in the interval (𝑥, 𝑥 + 𝑑𝑥) is proportional to time
𝑑𝑡 which it takes to pass through this interval :
𝑑𝑡
𝑃(𝑥)𝑑𝑥 =
𝑇
Where
2𝜋
𝑇= : period of oscillation
𝜔
Then,
𝑑𝑡 𝑑𝑡 𝑑𝑥 1 𝑑𝑥 𝜔 𝑑𝑥
𝑃(𝑥)𝑑𝑥 = = = => 𝑃(𝑥)𝑑𝑥 =
2𝜋/𝜔 2𝜋/𝜔 𝑑𝑥 2𝜋/𝜔 𝑑𝑥/𝑑𝑡 2𝜋 𝑣
To find the velocity, write total energy:
1 1 1
𝐸 = 𝑚𝑣 2 + 𝑚𝜔2 𝑥 2 = 𝑚𝜔2 𝐴2 => 𝑣 = 𝜔√(𝐴2 − 𝑥 2 )
2 2 2
Then
1 𝑑𝑥 1
𝑃(𝑥)𝑑𝑥 = => 𝑃(𝑥) =
2𝜋 √(𝐴2 − 𝑥 2 ) 2𝜋√(𝐴2 − 𝑥 2 )

𝑃(𝑥) is largest near the turning points 𝑥 = ±𝐴, where the speed of the particle vanishes.
 4

The Quantum Harmonics Oscillator

The quantum problem is to solve the Schrödinger equation for the potential
1
𝑉(𝑥) = 𝑚𝜔2 𝑥̂ 2
2

As we have seen, it suffices to solve the time independent Schrödinger equation

−ℏ2 𝑑 2 1
[ 2
+ 𝑚𝜔2 𝑥̂ 2 ] 𝜓(𝑥) = 𝐸𝜓(𝑥) (∗)
2𝑚 𝑑𝑥 2
There are two different approaches to this problem:

1. Using the algebraic technique, using the so-called ladder operators.

2. Solving the Scrödinger equation for quantum SHO (*) using the power series method.

In contrast to the infinite square well problem, in SHO problem, the differential equation
that needs to be solved is not so trivial, and one reason for discussing this problem is to learn
something about the techniques for solving such equations.

 The introduction of operators brings in a new concern: the order they act is important.
Consider

𝜕 𝜕𝜓(𝑥)
𝒙
̂𝒑̂ 𝜓(𝑥) = 𝑥 (−𝑖ℏ ) 𝜓(𝑥) = −𝑖ℏ𝑥 (𝑎)
𝜕𝑥 𝜕𝑥

And

𝜕 𝜕𝜓(𝑥)
̂𝒙
𝒑 ̂ 𝜓(𝑥) = (−𝑖ℏ ) (𝑥𝜓(𝑥)) = −𝑖ℏ {𝑥 + 𝜓(𝑥)} (𝑏)
𝜕𝑥 𝜕𝑥

Note that
𝒙
̂𝒑̂ 𝜓(𝑥) ≠ 𝒑
̂𝒙̂ 𝜓(𝑥)

Let’s find the difference.

(𝒙
̂𝒑̂− 𝒑
̂𝒙̂)𝜓(𝑥) ≡ [𝑥̂, 𝑝̂ ]𝜓(𝑥) = 𝑖ℏ𝜓(𝑥)

Since this is true for all 𝜓(𝑥), we conclude that we have an operator relation

[𝑥̂, 𝑝̂ ] = 𝑖ℏ ∶ a commutator relation

The difference between classical and quantum mechanics lies in that physical variables are
described by operators in quantum mechanics and these do not necessarily commute.
 5

Commutator Relations in Quantum Mechanics

The commutator between two operators, 𝐴̂ and 𝐵̂ are defined as

[𝐴̂, 𝐵̂] = 𝐴̂𝐵̂ − 𝐵̂𝐴̂

 If [𝐴̂, 𝐵̂] = 𝟎 => 𝐴̂𝐵̂ = 𝐵̂𝐴̂ , two operators are said to commute (𝐴̂ and 𝐵̂ are compatible)

with each other.

 [𝐴̂, 𝐵̂] = −[𝐵̂, 𝐴̂]

 [𝐴̂, 𝐴̂] = 𝟎

 [𝐴̂, 𝑎] = 𝟎, where 𝐴̂ is any operator and 𝑎 is any constant.

 [𝐴̂, 𝑎𝐵̂] = [𝒂𝐴̂, 𝐵̂] = 𝑎[𝐴̂, 𝐵̂]

 [𝐴̂, 𝐵̂ + 𝐶̂ ] = [𝐴̂, 𝐵̂] + [𝐴̂, 𝐶̂ ]

 [𝐴̂, 𝐵̂𝐶̂ ] = [𝐴̂, 𝐵̂]𝐶̂ + 𝐵̂[𝐴̂, 𝐶̂ ]

 𝐴̂ commutes with any function of 𝐴̂, 𝑓(𝐴̂): [𝐴̂, 𝑓(𝐴̂)] = 0

 6

I) Algebraic (Operator) Method:

Quantum mechanical Hamiltonian for the harmonic oscillator is
𝑝̂ 2 1
̂=
𝐻 + 𝑚𝜔2 𝑥̂ 2
2𝑚 2
Now, we define two dimensionless operators
𝛽 𝑝̂ 𝛽 𝑝̂ 𝑚𝜔
𝑎̂− ≡ [(𝑥̂ + 𝑖 )] 𝑎𝑛𝑑 𝑎̂+ ≡ [(𝑥̂ − 𝑖 )] with 𝛽 ≡ √
√2 𝑚𝜔 √2 𝑚𝜔 ℏ

Let’s find the commutator [𝑎̂− , 𝑎̂+ ]:

𝛽 𝑝̂ 𝛽 𝑝̂
[𝑎̂− , 𝑎̂+ ] = [ (𝑥̂ + 𝑖 ), (𝑥̂ − 𝑖 )]
√2 𝑚𝜔 √2 𝑚𝜔
2
𝛽 1 1 𝑚𝜔 1
=( ) ([𝑥̂, 𝑥̂] + 2
[𝑝̂ , 𝑝̂ ] + 𝑖 ([𝑝̂
⏟ , 𝑥̂] − [𝑥
⏟̂, 𝑝̂ ])) = 𝑖 (−2𝑖ℏ) = 1
√2 (𝑚𝜔) 𝑚𝜔 2ℏ 𝑚𝜔
−𝑖ℏ 𝑖ℏ

[𝑎̂− , 𝑎̂+ ] ≡ 𝑎̂− 𝑎̂+ − 𝑎̂+ 𝑎̂− = 1 => 𝑎̂− 𝑎̂+ ≠ 𝑎̂+ 𝑎̂− (∗)
We can write 𝑥̂ and 𝑝̂ in terms of 𝑎̂ and 𝑎̂+ :
𝑎̂− + 𝑎̂+ 𝑚𝜔 (𝑎̂− − 𝑎̂+ )
𝑥̂ = , 𝑝̂ =
√2𝛽 𝑖 √2𝛽
Therefore,
2 2
1 𝑚𝜔 (𝑎̂− − 𝑎̂+ ) 1 𝑎̂− + 𝑎̂+
̂=
𝐻 [ 2
] + 𝑚𝜔 [ ]
2𝑚 𝑖 √2𝛽 2 √2𝛽
2
(𝑎̂− ± 𝑎̂+ )2 = 𝑎̂− 2 + 𝑎̂+ ± (𝑎̂− 𝑎̂+ + 𝑎̂+ 𝑎̂−)
From eq.(*)
𝑎̂− 𝑎̂+ = 1 + 𝑎̂+ 𝑎̂−
Then,
2
(𝑎̂− ± 𝑎̂+ )2 = 𝑎̂− 2 + 𝑎̂+ ± (1 + 𝑎̂+ 𝑎̂− )

𝑚2 𝜔2 (−1) − 2 +2
𝑚𝜔2 − 2 2
̂=
𝐻 + −
(𝑎̂ + 𝑎̂ − (1 + 𝑎̂ 𝑎̂ )) + (𝑎̂ + 𝑎̂+ + (1 + 𝑎̂+ 𝑎̂− ))
2𝑚2𝛽 2 2 2𝛽 2

𝑚𝜔2 −2 +2 + − −2 +2
= 2
̂
[(−𝑎 − 𝑎
̂ + 1 + 𝑎
̂ 𝑎
̂ ) + (𝑎
̂ + 𝑎
̂ + 1 + 𝑎̂+ 𝑎̂− )]
4𝛽
 7

1 ℏ
̂ = 𝑚𝜔2 (
𝐻 ) 2(1 + 𝑎̂+ 𝑎̂− ) ⟹
4 𝑚𝜔
1
̂ = 𝜔ℏ (𝑎̂+ 𝑎̂− + )
𝐻
2
OR
1
̂ = 𝜔ℏ (𝑎̂− 𝑎̂+ − )
𝐻
2

̂ , 𝑎̂∓ ]:
Let’s calculate the commutators [𝐻

1 1
̂ , 𝑎̂− ] = 𝜔ℏ [𝑎̂+ 𝑎̂− + , 𝑎̂− ] = 𝜔ℏ ([𝑎̂+ 𝑎̂− , 𝑎̂− ] + [ , 𝑎̂− ])
[𝐻
2 ⏟2
0

[𝑎̂+ , 𝑎̂− ] 𝑎̂− + 𝑎̂+ ⏟

= 𝜔ℏ (⏟ [𝑎̂− , 𝑎̂− ]) ⇒
−1 0

̂ , 𝑎̂− ] ≡ 𝐻
[𝐻 ̂ 𝑎̂− − 𝑎̂− 𝐻
̂ = −𝜔ℏ𝑎̂− (∗)

Similarly, we can show that (Homework!)

̂ , 𝑎̂+ ] ≡ 𝐻
[𝐻 ̂ 𝑎̂+ − 𝑎̂+ 𝐻
̂ = +𝜔ℏ𝑎̂+ (∗∗)
 8

The energy eigenvalue problem for the quantum harmonic oscillator:

Let
̂ for the harmonic oscillator
𝜓 ∶ energy eigen function of H
and
𝐸 ∶ the corresponding energy eigenvalue
Such that
̂ 𝜓 = 𝐸𝜓
𝐻
 Let’s prove that the state (𝑎̂+ 𝜓) satisfies the Schrödinger equation too with the energy
(𝐸 + 𝜔ℏ). That is, we can write a new energy eigenvalue equation
̂ (𝑎
𝐻 ⏟̂ + 𝜓) = (𝐸 + 𝜔ℏ)(𝑎̂+ 𝜓)
𝑎 𝑛𝑒𝑤 𝑠𝑡𝑎𝑡𝑒
̂ acts on
Consider what happens when 𝐻
̂ (𝑎̂+ 𝜓) = (𝐻
𝐻 ̂ 𝑎̂+ )𝜓 = ̂ + 𝜔ℏ𝑎̂+ )𝜓 = 𝑎̂+ ⏟
⏟ (𝑎̂+ 𝐻 ̂ 𝜓 + 𝜔ℏ𝑎̂+ 𝜓 = (𝐸 + 𝜔ℏ)(𝑎̂+ 𝜓)
𝐻
𝑒𝑞.(∗∗) 𝐸𝜓

That is
̂ (𝑎
𝐻 ⏟̂ + 𝜓) = (𝐸 + 𝜔ℏ)(𝑎̂+ 𝜓) ∶ also an eigenvalue eq.
𝑎 𝑚𝑒𝑤 𝑠𝑡𝑎𝑡𝑒
̂ but, with a new egenvalue (𝐸 + 𝜔ℏ) that is
The new state (𝑎̂+ 𝜓) is also an eigenfunction of 𝐻
greater than E of the original state 𝜓 by one quantum of energy 𝜔ℏ.
 Let’s prove that the state (𝑎̂− 𝜓) satisfies the Schrödinger equation too with the energy
(𝐸 − 𝜔ℏ). That is, we can write a new energy eigenvalue equation
̂ (𝑎
𝐻 ⏟̂ − 𝜓) = (𝐸 − 𝜔ℏ)(𝑎̂− 𝜓)
𝑎 𝑛𝑒𝑤 𝑠𝑡𝑎𝑡𝑒

̂ acts on (𝑎̂− 𝜓):

Consider what happens when 𝐻
̂ (𝑎̂− 𝜓) = (𝐻
𝐻 ̂ 𝑎̂− )𝜓 = ̂ − 𝜔ℏ𝑎̂− )𝜓 = 𝑎̂− ⏟
⏟ (𝑎̂− 𝐻 ̂ 𝜓 − 𝜔ℏ𝑎̂− = (𝐸 − 𝜔ℏ)(𝑎̂− 𝜓)
𝐻
𝑒𝑞.(∗) 𝐸𝜓

That is
̂ (𝑎
𝐻 ⏟̂ − 𝜓) = (𝐸 − 𝜔ℏ)(𝑎̂− 𝜓) ∶ also an eigenvalue eq.
𝑎 𝑛𝑒𝑤 𝑠𝑡𝑎𝑡𝑒
̂ but, with a new egenvalue (𝐸 − 𝜔ℏ) that is
The new state (𝑎̂− 𝜓) is also an eigenfunction of 𝐻
smaller than E of the original state 𝜓 by one quantum of energy 𝜔ℏ.
 9

Therefore we refer to these operators as ladder operators because they take us up and down a
ladder of energy eigenstates, as shown in the figures.

The 𝑎̂− is called the lowering operator , and 𝑎̂+ is called the raising operator.
𝑎̂− (𝑎̂+ ) lowers (raises) the energy eigenvalue by the amount of 𝜔ℏ.
 10

̂ − ? The energy eigenvalue will

Q) What happens if we keep operating upon 𝝍 with 𝒂
decrease and decrease. Can it decrease indefinitely? NO! The lowering cannot go on forever!
(Eventually, we reach a state with E<0, which doesn’t exist) This means that there is a state of
lowest energy, the ground state with energy 𝐸0 and wave function 𝜓0 such that
𝑎̂− 𝜓0 = 0
(Even the operation by 𝑎̂− can not lower the energy further).

We now use this condition to find the energy of the ground state:
1 1
̂ 𝜓0 = 𝐸0 𝜓0 ⟹ 𝐻
𝐻 ̂ 𝜓0 = 𝜔ℏ (𝑎̂+ 𝑎̂− + ) 𝜓0 = 𝜔ℏ (𝑎̂+ 𝑎
̂ − 𝜓0 + 𝜓0 ) ⟹
⏟
2 2
0

.
1 1
̂ 𝜓0 = 𝜔ℏ 𝜓0 = 𝐸0 𝜓0 ⟹ 𝐸0 = 𝜔ℏ ∶ Ground state Energy
𝐻
2 2

Note that ground state does not have zero energy, in contrast to the classical harmonics
oscillator. Rather, quantum mechanical ground state has a zero-point energy of 𝜔ℏ/2, that is
also consistent with the uncertainty principle.
 11

To generate the next energy eigenstate up the ladder of states, we act with the raising operator
𝑎̂+ on the ground state 𝜓0 , which produces a new energy eigenstate with the energy increased
by one quantum 𝜔ℏ:
𝑎̂+ 𝜓0 ~𝜓1 such that 𝐻 ̂ 𝜓1 = 𝐸1 𝜓1
where
1 3
𝐸1 = 𝐸0 + 𝜔ℏ = 𝜔ℏ + 𝜔ℏ = 𝜔ℏ
2 2
We repeat the action of the raising operator
𝑎̂+ 𝜓1 ~𝜓2 such that ̂ 𝜓2 = 𝐸2 𝜓2
𝐻
where
3 5
𝐸2 = 𝐸1 + 𝜔ℏ = 𝜔ℏ + 𝜔ℏ = 𝜔ℏ
2 2

We write the energy spectrum compactly as

1
𝐸𝑛 = 𝜔ℏ (𝑛 + ) , 𝑛 = 0,1,2,3, …
2
The quantum number n is used to label the energy.
 12

Ground State Eigenfunction:

Again consider
𝛽 𝑝̂ 𝑑
𝑎̂− 𝜓0 = 0 , with 𝑎̂− = [(𝑥̂ + 𝑖 )] and 𝑝̂ = −𝑖ℏ
√2 𝑚𝜔 𝑑𝑥
Thus
𝑚𝜔 ℏ 𝑑 𝑑𝜓0 𝑚𝜔
√ [(𝑥̂ + )] 𝜓0 = 0 => =− 𝑥𝜓0
2ℏ 𝑚𝜔 𝑑𝑥 𝑑𝑥 ℏ
𝑑𝜓0 𝑚𝜔 𝑚𝜔 2
∫ =− ∫ 𝑥𝑑𝑥 => ln 𝜓0 = − 𝑥 +𝑐
𝜓0 ℏ 2ℏ
𝑚𝜔 2
𝜓0 = 𝐴𝑒 2ℏ 𝑥
−

Normalize
∞
𝑚𝜔 2 𝜋ℏ 𝑚𝜔 1/4
1= |𝐴|2 ∫ 𝑒− ℏ 𝑥 𝑑𝑥 = |𝐴|2 √ => 𝐴 = ( )
𝑚𝜔 𝜋ℏ
−∞

𝑚𝜔 1/4 −𝑚𝜔𝑥 2
𝜓0 = ( ) 𝑒 2ℏ
𝜋ℏ
 13

Example 2.4: Find the 1.st excited state of the SHO

𝑎̂+ 𝜓0 ~𝜓1
Let’s write as
𝐴𝑎̂+ 𝜓0 = 𝜓1
Let’s calculate RHS:

𝑚𝜔 𝑑 1 𝑚𝜔 ℏ 𝑑
𝑎̂+ ≡ √ [(𝑥̂ − 𝑖 (−𝑖ℏ ) )] = √ [(𝑥̂ − )]
2ℏ 𝑑𝑥 𝑚𝜔 2ℏ 𝑚𝜔 𝑑𝑥
𝑚𝜔 ℏ 𝑑 𝑚𝜔 1/4 −𝑚𝜔𝑥 2
𝐴𝑎̂+ 𝜓0 = 𝐴 √ [(𝑥̂ − )] {( ) 𝑒 2ℏ }
⏟ 2ℏ 𝑚𝜔 𝑑𝑥 ⏟ 𝜋ℏ
𝑎̂+ 𝜓0

𝑚𝜔 ℏ 𝑑 𝑚𝜔 1/4 −𝑚𝜔𝑥 2 𝑚𝜔 1/4 𝑚𝜔 ℏ 𝑑 𝑚𝜔 2

= 𝐴√ [(𝑥̂ − )] ( ) 𝑒 2ℏ = 𝐴( ) √ [(𝑥̂ − )] 𝑒 − 2ℏ 𝑥
2ℏ 𝑚𝜔 𝑑𝑥 𝜋ℏ 𝜋ℏ 2ℏ 𝑚𝜔 𝑑𝑥

𝑑 −𝑚𝜔𝑥 2 𝑚𝜔 −𝑚𝜔𝑥 2
𝑒 2ℏ = − 𝑥𝑒 2ℏ
𝑑𝑥 ℏ

𝑚𝜔 1/4 𝑚𝜔 ℏ 𝑚𝜔 −
𝑚𝜔 2
𝑥 𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2
+
𝐴𝑎̂ 𝜓0 = ( ) √ [(𝑥̂ − [− 𝑥])] 𝑒 2ℏ = 𝐴( ) √ 𝑥𝑒 2ℏ 𝑥
−
𝜋ℏ 2ℏ ⏟ 𝑚𝜔 ℏ 𝜋ℏ ℏ
2𝑥

Thus,

𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2
𝜓1 = 𝐴 ( ) √ 𝑥𝑒 2ℏ 𝑥
−
𝜋ℏ ℏ

We can normalize the state:

∞
? 𝑚𝜔 1/2 2𝑚𝜔 𝑚𝜔 2 𝑚𝜔 1/2 2𝑚𝜔 1 𝜋
⏞ 𝐴2 (
1= ) ( ) ∫ 𝑥 2 𝑒 − ℏ 𝑥 𝑑𝑥 = 𝐴2 ( ) ( ) 𝑚𝜔 √ 𝑚𝜔 = 𝐴2
𝜋ℏ ℏ 𝜋ℏ ℏ 2(
−∞ ℏ ) ( ℏ )

𝐴=1
 14

Consider, for any two functions, 𝑓(𝑥) and 𝑔(𝑥). Then, we have
∞ ∞

∫ 𝑓 ∗ (𝑎̂− 𝑔)𝑑𝑥 = ∫ (𝑎̂+ 𝑓)∗ 𝑔 𝑑𝑥

−∞ −∞
Proof:
∞ ∞ ∞ ∞
𝛽 ℏ 𝑑 𝛽 𝛽 ℏ 𝑑
∫ 𝑓 ∗ (𝑎̂− 𝑔)𝑑𝑥 = ∫ 𝑓 ∗ (𝑥̂ + ) 𝑔𝑑𝑥 = ∫ 𝑥 𝑓 ∗ 𝑔𝑑𝑥 + ∫ 𝑓∗ 𝑔𝑑𝑥
√2 𝑚𝜔 𝑑𝑥 √2 √2 𝑚𝜔 𝑑𝑥
−∞ −∞ −∞ ⏟ −∞
𝐼𝐼

Integration by parts to the 2.nd term II

𝑑𝑔
𝑢 = 𝑓∗ , 𝑑𝑣 =
𝑑𝑥
∞ ∞
𝛽 ℏ 𝑑 ∗ 𝛽 ℏ 𝑑 ∗
𝐼𝐼 = ⏟∗ 𝑔|∞
(𝑓 −∞ − ∫ 𝑓 𝑔𝑑𝑥) => 𝐼𝐼 = − ∫ 𝑓 𝑔𝑑𝑥
√2 𝑚𝜔 0
𝑑𝑥 √2 𝑚𝜔 𝑑𝑥
−∞ −∞

Then
∗
∞ ∞ ∞
∗ −
𝛽 ℏ 𝑑𝑓 ∗∗
𝛽 ℏ 𝑑
∫ 𝑓 (𝑎̂ 𝑔)𝑑𝑥 = ∫ (𝑥̂𝑓 𝑔 − 𝑔) 𝑑𝑥 = ∫ [ (𝑥̂ − ) 𝑓] 𝑔𝑑𝑥
√2 𝑚𝜔 𝑑𝑥 ⏟
√2 𝑚𝜔 𝑑𝑥
−∞ −∞ −∞
𝑎̂ +

∞ ∞

∫ 𝑓 ∗ (𝑎̂− 𝑔)𝑑𝑥 = ∫ (𝑎̂+ 𝑓)∗ 𝑔𝑑𝑥

−∞ −∞

Similarly,(HW!)
∞ ∞

∫ 𝑓 ∗ (𝑎̂+ 𝑔)𝑑𝑥 = ∫ (𝑎̂− 𝑓)∗ 𝑔𝑑𝑥

−∞ −∞
 15

We have found a scheme of generating a sequence of unnormalized eigenfunctions and

̂:
eigenvalues of 𝐻
𝑎̂+ 𝜓0 ~𝜓1 , 𝑎̂+ 𝜓1 ~𝜓2
Thus,
𝑎̂+ 𝜓𝑚 ~𝜓𝑚+1
Let
𝑎̂+ 𝜓𝑚 = 𝑐𝑚 𝜓𝑚+1 (𝑎)

And,
𝑎̂− 𝜓0 = 0 , 𝑎̂− 𝜓1 ~𝜓0

𝑎̂− 𝜓𝑚 ~𝜓𝑚−1
Let
𝑎̂− 𝜓𝑚 = 𝑑𝑚 𝜓𝑚−1 (𝑏)

𝑐𝑚 , 𝑑𝑚 =?
Consider (a).
∞ ∞ ∞

∫ (𝑎̂+ 𝜓𝑚 )∗ (𝑎̂+ 𝜓𝑚 ) 𝑑𝑥 = ∫ (𝑐𝑚 𝜓𝑚+1 )∗ (𝑐𝑚 𝜓𝑚+1 ) 𝑑𝑥 = |𝑐𝑚 |2 ∫ (𝜓𝑚+1 )∗ 𝜓𝑚+1 𝑑𝑥 = |𝑐𝑚 |2 (∗)
−∞ −∞ −∞

∞ ∞

∫ (𝑎̂+ 𝜓𝑚 )∗ (𝑎̂+ 𝜓𝑚 ) 𝑑𝑥 = ∫ (𝑎̂− 𝑎̂+ 𝜓𝑚 )∗ 𝜓𝑚 𝑑𝑥

−∞ −∞
Since
1 𝐻̂ 1 1
̂ − + − +
𝐻 = 𝜔ℏ (𝑎̂ 𝑎̂ − ) => 𝑎̂ 𝑎̂ = + , 𝑤𝑖𝑡ℎ 𝐸𝑚 = 𝜔ℏ (𝑚 + )
2 𝜔ℏ 2 2

Then
1
̂
𝐻 1 𝐸 1 𝜔ℏ (𝑚 + ) 1
− +
𝑎̂ 𝑎̂ 𝜓𝑚 = ( + )𝜓 = (
𝑚
+ )𝜓 = ( 2 + ) 𝜓𝑚 = (𝑚 + 1)𝜓𝑚
𝜔ℏ 2 𝑚 𝜔ℏ 2 𝑚 𝜔ℏ 2

∞ ∞
(∗)
(𝑎+
∫ ̂ 𝜓𝑚 )∗ (𝑎+ )∗ ⏞ |𝑐𝑚 |2
̂ 𝜓𝑚 ) 𝑑𝑥 = (𝑚 + 1) ∫ (𝜓𝑚 𝜓𝑚 𝑑𝑥 = (𝑚 + 1) =
−∞ −∞

𝑐𝑚 = √𝑚 + 1
 16

Therefore
𝑎̂+ 𝜓𝑚 = √𝑚 + 1𝜓𝑚+1

Similarly,
∞ ∞ ∞

∫ (𝑎̂− 𝜓𝑚 )∗ (𝑎̂− 𝜓𝑚 ) 𝑑𝑥 = ∫ (𝑑𝑚 𝜓𝑚−1 )∗ (𝑑𝑚 𝜓𝑚−1 ) 𝑑𝑥 = |𝑑𝑚 |2 ∫ (𝜓𝑚−1 )∗ 𝜓𝑚−1 𝑑𝑥

−∞ −∞ −∞
= |𝑑𝑚 |2 (∗∗)
∞ ∞

∫ (𝑎̂− 𝜓𝑚 )∗ (𝑎̂− 𝜓𝑚 ) 𝑑𝑥 = ∫ (𝑎̂+ 𝑎̂− 𝜓𝑚 )∗ 𝜓𝑚 𝑑𝑥

−∞ −∞
Since
1 𝐻̂ 1 1
̂ = 𝜔ℏ (𝑎̂+ 𝑎̂− + ) => 𝑎̂+ 𝑎̂− =
𝐻 − , 𝐸𝑚 = 𝜔ℏ (𝑚 + )
2 𝜔ℏ 2 2

Then
1
̂ 1
𝐻 𝐸 1 𝜔ℏ (𝑚 − ) 1
𝑎̂+ 𝑎̂− 𝜓𝑚 = ( − ) 𝜓𝑚 = (
𝑚
− ) 𝜓𝑚 = ( 2 + ) 𝜓 = 𝑚𝜓
𝜔ℏ 2 𝜔ℏ 2 𝜔ℏ 2 𝑚 𝑚

∞ ∞
(∗)
(𝑎−
∫ ̂ 𝜓𝑚 )∗ (𝑎− )∗ ⏞ |𝑑𝑚 |2 => 𝑑𝑚 = √𝑚
̂ 𝜓𝑚 ) 𝑑𝑥 = 𝑚 ∫ (𝜓𝑚 𝜓𝑚 𝑑𝑥 = 𝑚 =
−∞ −∞

Therefore

𝑎̂− 𝜓𝑚 = √𝑚𝜓𝑚−1

𝑎̂− 𝜓𝑚 = √𝑚𝜓𝑚−1

𝑎̂+ 𝜓𝑚 = √𝑚 + 1𝜓𝑚+1
 17

Now:
1
𝜓𝑚+1 = 𝑎̂+ 𝜓𝑚
√𝑚 + 1

𝑚 = 0:
1
𝜓1 = 𝑎̂+ 𝜓0
√1
𝑚 = 1:
1 1
𝜓2 = 𝑎̂+ 𝜓1 = 𝑎̂+ 𝑎̂+ 𝜓0
√2 √2.1

𝑚 = 2:
1 1
𝜓3 = 𝑎̂+ 𝜓2 = 𝑎̂+ 𝑎̂+ 𝑎̂+ 𝜓0
√3 √3.2.1
Clearly

1
𝜓𝑚 = (𝑎̂+ )𝑚 𝜓0
√𝑚!
 18

Orthogonality of Eigenfunctions of Quantum SHO

∞
∗ (𝑥) 0 𝑖𝑓 𝑚 ≠ 𝑛
∫ 𝜓𝑚 𝜓𝑛 𝑑𝑥 = 𝛿𝑚𝑛 = { }
−∞
1 𝑖𝑓 𝑚 = 𝑛

Proof: Consider the integral

∞
∗
𝐼 ≡ ∫ 𝜓𝑚 ( 𝑎̂+ 𝑎̂− )𝜓𝑛 𝑑𝑥
−∞
Now
𝑎̂− 𝜓𝑛 = √𝑛𝜓𝑛−1 (𝑎)

𝑎̂+ 𝜓𝑛 = √𝑛 + 1𝜓𝑛+1 (𝑏)

𝑎̂+ 𝑎
̂ − 𝜓𝑛 = √𝑛(𝑎̂+ 𝜓𝑛−1 ) =
⏟ ⏟ √𝑛(√𝑛 − 1 + 1𝜓𝑛−1+1 ) = √𝑛√𝑛𝜓𝑛
(𝑎)=> (𝑏)
√𝑛𝜓𝑛−1
Thus
𝑎̂+ 𝑎̂− 𝜓𝑛 = 𝑛𝜓𝑛 (𝑐)

Then

∞ ∞
∗ + − ∗
𝐼=∫ 𝜓𝑚 (⏟𝑎̂ 𝑎̂ )𝜓𝑛 𝑑𝑥 => 𝐼 = 𝑛 ∫ 𝜓𝑚 𝜓𝑛 𝑑𝑥 (∗)
−∞ 𝑛𝜓𝑛 −∞

Integral 𝐼 can also be written in the form

∞ ∞ ∞
∗ (
𝐼≡∫ 𝜓𝑚 𝑎̂+ 𝑎̂− )𝜓𝑛 𝑑𝑥 = ∫ (𝑎̂− 𝜓𝑚 )∗ (𝑎̂− 𝜓𝑛 ) 𝑑𝑥 = ∫ ( 𝑎̂+ 𝑎̂− 𝜓𝑚 )∗ 𝜓𝑛 𝑑𝑥
−∞ −∞ −∞
From (c)
𝑎̂+ 𝑎̂− 𝜓𝑚 = 𝑚 𝜓𝑚
Thus,

∞
∗
𝐼 = 𝑚 ∫ 𝜓𝑚 𝜓𝑛 𝑑𝑥 (∗∗)
−∞
Compare (*) and (**)=>
∞
∗
∫ 𝜓𝑚 𝜓𝑛 𝑑𝑥 = 0 𝑖𝑓 𝑚 ≠ 𝑛
−∞
 19

Orthonormality means that we can expand any function 𝑓(𝑥) 𝑎s a linear combination of
stationary states
∞

𝑓(𝑥) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥)
𝑛=1

Therefore, we can write the initial state in this way;

∞

Ψ(𝑥, 0) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥)
𝑛=1

With

𝑐𝑛 = ∫ 𝜓𝑛∗ (𝑥) Ψ(𝑥, 0)𝑑𝑥

|𝑐𝑛 |2 : 𝑡ℎ𝑒 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑎 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑚𝑡 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑛𝑒𝑟𝑔𝑦 𝑤𝑜𝑢𝑙𝑑 𝑦𝑖𝑒𝑙𝑑 𝑡ℎ𝑒 𝑣𝑎𝑙𝑢𝑒 𝐸𝑛
∞ ∞
𝑖𝐸𝑚 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑚 Ψ𝑚 (𝑥, 𝑡) = ∑ 𝑐𝑚 𝜓𝑚 (𝑥) 𝑒 − ℏ
𝑚=1 𝑚=1
 20

Problem 2.10
a) Construct 𝜓2

From
1
𝜓𝑚+1 = 𝑎̂+ 𝜓𝑚
√𝑚 + 1

𝑚 = 1:
1
𝜓2 = 𝑎̂+ 𝜓1
√2
From Ex.2.4, we have
𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2
𝜓1 = ( ) √ 𝑥𝑒 − 2ℏ 𝑥
𝜋ℏ ℏ

𝑚𝜔 𝑑 1 𝑚𝜔 ℏ 𝑑
𝑎̂+ ≡ √ [(𝑥̂ − 𝑖 (−𝑖ℏ ) )] = √ [(𝑥̂ − )]
2ℏ 𝑑𝑥 𝑚𝜔 2ℏ 𝑚𝜔 𝑑𝑥

1 1 𝑚𝜔 ℏ 𝑑
𝜓2 = 𝑎̂+ 𝜓1 = √[(𝑥̂ − )] 𝜓1
√2 √2 2ℏ 𝑚𝜔 𝑑𝑥

𝑑 𝑚𝜔 1/4 2𝑚𝜔 𝑑 𝑚𝜔 2 𝑚𝜔 1/4 2𝑚𝜔 𝑑 𝑚𝜔 2 −𝑚𝜔𝑥 2

𝜓1 = ( ) √ {𝑥𝑒 − 2ℏ 𝑥 } = ( ) √ {1 − 𝑥 } 𝑒 2ℏ
𝑑𝑥 𝜋ℏ ℏ 𝑑𝑥 𝜋ℏ ℏ 𝑑𝑥 ℏ

1 𝑚𝜔 𝑚𝜔 1/4 2𝑚𝜔 ℏ 𝑚𝜔 2 𝑚𝜔 2
𝜓2 = √ ( ) √ [(𝑥̂𝑥 − {1 − 𝑥 })] 𝑒 − 2ℏ 𝑥
√2 2ℏ 𝜋ℏ ℏ 𝑚𝜔 ℏ

1 𝑚𝜔 𝑚𝜔 1/4 2𝑚𝜔 ℏ 2𝑚𝜔 2 𝑚𝜔 2

𝜓2 = √
( ) √ [ 𝑥 − 1] 𝑒 − 2ℏ 𝑥
√2 2ℏ 𝜋ℏ ℏ 𝑚𝜔 ℏ
1 𝑚𝜔 1/4 2𝑚𝜔 2 𝑚𝜔 2
𝜓2 = ( ) [ 𝑥 − 1] 𝑒 − 2ℏ 𝑥
√2 𝜋ℏ ℏ

(Show that it is normalized)

 21

c)
∞ 𝑜𝑑𝑑
∗ (𝑥) ⏞
∫ 𝜓 ⏟0 𝜓1 𝑑𝑥 =0
−∞ 𝑒𝑣𝑒𝑛
∞ 𝑜𝑑𝑑
∗ (𝑥) ⏞
∫ 𝜓 ⏟2 𝜓1 𝑑𝑥 =0
−∞ 𝑒𝑣𝑒𝑛

𝑚𝜔 1/4 −𝑚𝜔𝑥 2
𝜓0 = ( ) 𝑒 2ℏ
𝜋ℏ
∞
1 𝑚𝜔 1/4 𝑚𝜔 1/4 ∞ −𝑚𝜔𝑥 2 2𝑚𝜔 2 𝑚𝜔 2
∫ 𝜓0∗ (𝑥)𝜓2 𝑑𝑥 = ( ) ( ) ∫ 𝑒 2ℏ [ 𝑥 − 1] 𝑒 − 2ℏ 𝑥 𝑑𝑥
−∞ √2 𝜋ℏ 𝜋ℏ −∞ ℏ

1 𝑚𝜔 1/2 ∞ 2𝑚𝜔 2 𝑚𝜔 2
= ( ) ∫ [ 𝑥 − 1] 𝑒 − ℏ 𝑥 𝑑𝑥 =
√2 𝜋ℏ −∞ ℏ

1 𝑚𝜔 1/2 2𝑚𝜔 ∞ 2 −𝑚𝜔𝑥 2 ∞ 𝑚𝜔 2

= ( ) ∫ 𝑥 𝑒 ℏ 𝑑𝑥 − ∫ 𝑒 ℏ 𝑥 𝑑𝑥 = 0
−
√2 𝜋ℏ ℏ ⏟−∞ ⏟−∞
1 ℏ √ 𝜋ℏ √ 𝜋ℏ
{ 2𝑚𝜔 𝑚𝜔 𝑚𝜔 }
 22

Problem 2.13

Solution
a)
∞

1 = |𝐴|2 ∫ (3𝜓0 + 4𝜓1 )∗ (3𝜓0 + 4𝜓1 )𝑑𝑥

−∞
∞

= |𝐴|2 9 ∫ |𝜓0 |2 𝑑𝑥
⏟
−∞
1
(𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛)
∞

+|𝐴|2 16 ∫ |𝜓1 |2 𝑑𝑥
⏟
−∞
1
(𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜)
∞

+|𝐴|2 12 ∫ 𝜓0∗ (𝑥)𝜓1 (𝑥) 𝑑𝑥

⏟
−∞
0
(𝑜𝑟𝑡𝑜𝑔𝑜𝑛𝑎𝑙𝑖𝑡𝑦)
∞

+|𝐴|2 12 ∫ 𝜓1∗ (𝑥)𝜓0 (𝑥) 𝑑𝑥

⏟
−∞
0
(𝑜𝑟𝑡𝑜𝑔𝑜𝑛𝑎𝑙𝑖𝑡𝑦)

1
= |𝐴|2 (9 + 16) => 𝐴 =
5
 23

b)
∞
𝑖𝐸𝑚 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑚 𝜓𝑚 (𝑥) 𝑒 − ℏ
𝑚=1

Here;
𝑐0 = 3/5, 𝑐1 = 4/5 , 𝑐𝑚 = 0 𝑤ℎ𝑒𝑛 𝑚 ≠ 0,1
1 𝑖𝐸0 𝑡 𝑖𝐸1 𝑡
Ψ(𝑥, 𝑡) = [3𝜓0 𝑒 − ℏ + 4𝜓1 𝑒 − ℏ ]
5

1 ∗
𝑖(𝐸0 −𝐸1 )𝑡
∗
−𝑖(𝐸0 −𝐸1 )𝑡
|Ψ(𝑥, 𝑡)|2 = |𝜓 | 2 | 2
[9 0 + 16|𝜓1 + 12𝜓0 𝜓1 𝑒 ℏ + 12𝜓1 𝜓0 𝑒 ℏ ]
25
𝜓𝑛 = 𝜓𝑛∗ , 𝑓𝑜𝑟 𝐻𝑂
1 𝑖(𝐸0 −𝐸1 )𝑡 −𝑖(𝐸0 −𝐸1 )𝑡
|Ψ(𝑥, 𝑡)|2 = [9|𝜓0 |2 + 16|𝜓1 |2 + 12𝜓1 𝜓0 (𝑒 ℏ +𝑒 ℏ )]
25
1 (𝐸0 − 𝐸1 )𝑡
|Ψ(𝑥, 𝑡)|2 = [9|𝜓0 |2 + 16|𝜓1 |2 + 24𝜓1 𝜓0 cos ( )]
25 ℏ

Note that
1 1
(𝐸0 − 𝐸1 ) = 𝜔ℏ (0 + ) − 𝜔ℏ (1 + ) = −𝜔ℏ
2 2
Then
1
|Ψ(𝑥, 𝑡)|2 = [9|𝜓0 |2 + 16|𝜓1 |2 + 24𝜓1 𝜓0 cos(𝜔𝑡)]
25
The probability density oscillates with an angular frequency 𝜔.
If 𝜓1 is replaced by 𝜓2 , then we get

1
|Ψ(𝑥, 𝑡)|2 = [9|𝜓0 |2 + 16|𝜓2 |2 + 24𝜓2 𝜓0 cos(2𝜔𝑡)]
25
since
1 1
(𝐸0 − 𝐸2 ) = 𝜔ℏ (0 + ) − 𝜔ℏ (2 + ) = −2𝜔ℏ
2 2
 24

c)
∞
∗ (𝑥,
𝑎̂− + 𝑎̂+
〈𝑥〉 = ∫ Ψ 𝑡)𝑥̂ Ψ(𝑥, 𝑡)𝑑𝑥 , 𝑤ℎ𝑒𝑟𝑒 𝑥̂ =
−∞ √2𝛽
∞
1 ∞ ∗
〈𝑥〉 = ∫ ∫ (3𝜓0 𝑒 −𝑖𝐸0 𝑡/ℏ + 4𝜓1 𝑒 −𝑖𝐸1 𝑡/ℏ ) 𝑥̂ (3𝜓0 𝑒 −𝑖𝐸0 𝑡/ℏ + 4𝜓1 𝑒 −𝑖𝐸1 𝑡/ℏ )𝑑𝑥
25 −∞
−∞
∞ ∞ ∞
1 𝑖(𝐸0 −𝐸1 )𝑡 −𝑖(𝐸0 −𝐸1 )𝑡
= 9 ∫ 𝜓0 𝑥𝜓0 𝑑𝑥 + 16 ∫ 𝜓1 𝑥𝜓1 𝑑𝑥 + 12 (𝑒 ℏ +𝑒 ℏ ) ∫ 𝜓0 𝑥𝜓1 𝑑𝑥
25
⏟
−∞ ⏟
−∞ −∞
[ 0 0 ]
∞
24 1 (𝐸1 − 𝐸2 )𝑡
= cos ( ) ∫ 𝜓0 (𝑎̂− + 𝑎̂+ )𝜓1 𝑑𝑥
25 √2𝛽 ℏ
−∞
∞ ∞
24 1
= 𝑎̂− 𝜓1 𝑑𝑥 + ∫ 𝜓0 ⏟
cos(𝜔𝑡) [ ∫ 𝜓0 ⏟ 𝑎̂+ 𝜓1 𝑑𝑥]
25 √2𝛽
−∞ √1𝜓0 −∞ √2𝜓2

∞ ∞
24 1
= cos(𝜔𝑡) ∫ |𝜓0 |2 𝑑𝑥 + √2 ∫ 𝜓0∗ (𝑥)𝜓2 (𝑥)𝑑𝑥
25 √2𝛽
⏟
−∞ ⏟ −∞
1 0
[(𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛) (𝑜𝑟𝑡𝑜𝑔𝑜𝑛𝑎𝑙𝑖𝑡𝑦) ]

24 ℏ
〈𝑥〉 = √ cos(𝜔𝑡)
25 2𝑚𝜔

𝑑 24 ℏ 24
〈𝑝〉 = 𝑚 〈𝑥〉 = −𝑚𝜔 √ sin(𝜔𝑡) => 〈𝑝〉 = − √𝑚𝜔ℏ/2 sin(𝜔𝑡)
𝑑𝑡 25 2𝑚𝜔 25

d)
32 9
𝑃(𝐸 = 𝐸0 ) = | | =
5 25
2
4 16
𝑃(𝐸 = 𝐸1 ) = | | =
5 25
 25

Example2.5: Find the expectation value of the potential energy in the nth state of HO.,
∞
1 1
〈𝑉〉 = 𝑚𝜔2 〈𝑥̂ 2 〉 = 𝑚𝜔2 ∫ 𝜓𝑚
∗ 2
𝑥̂ 𝜓𝑚
2 2
−∞

𝑎̂− + 𝑎̂+ 𝑚𝜔 (𝑎̂− − 𝑎̂+ )

𝑥̂ = , 𝑝̂ =
√2𝛽 𝑖 √2𝛽

2
(𝑎̂− + 𝑎̂+ )2 ℏ
𝑥̂ = = (𝑎̂− + 𝑎̂+ )2
2𝛽 2 2𝑚𝜔
2
(𝑎̂− + 𝑎̂+ )2 = 𝑎̂− 2 + 𝑎̂+ + 𝑎̂− 𝑎̂+ + 𝑎̂+ 𝑎̂−
Use
[𝑎̂− , 𝑎̂+ ] = 𝑎̂− 𝑎̂+ − 𝑎̂+ 𝑎̂− = 1
2 2
(𝑎̂− + 𝑎̂+ )2 = 𝑎̂− 2 + 𝑎̂+ + 𝑎̂− 𝑎̂+ + 𝑎
̂ + 𝑎̂− = 𝑎̂− 2 + 𝑎̂+ + 2𝑎̂− 𝑎̂+ − 1
⏟
𝑎̂− 𝑎̂+ −1

∞
1 ℏ 2
〈𝑉〉 = 𝑚𝜔2 ∗
∫ 𝜓𝑚 (𝑎̂− 2 + 𝑎̂+ + 2𝑎̂− 𝑎̂+ − 1)𝜓𝑚
2 2𝑚𝜔
−∞
∞ ∞
∗ −2 ∗
∫ 𝜓𝑚 𝑎̂ 𝜓𝑚 ~ ∫ 𝜓𝑚 𝜓𝑚−2 = 0
−∞ −∞
∞ ∞
∗ + 2 ∗
∫ 𝜓𝑚 𝑎̂ 𝜓𝑚 ~ ∫ 𝜓𝑚 𝜓𝑚+2 = 0
−∞ −∞
∞ ∞ ∞
∗ − + ∗ − ∗
2 ∫ 𝜓𝑚 𝑎̂ 𝑎̂ 𝜓𝑚 = 2√𝑚 + 1 ∫ 𝜓𝑚 𝑎̂ 𝜓𝑚+1 = 2√𝑚 + 1√𝑚 + 1 ∫ 𝜓𝑚 𝜓𝑚 = 2(𝑚 + 1)
−∞ −∞ −∞
∞
∗
− ∫ 𝜓𝑚 𝜓𝑚 = −1
−∞
1 ℏ 1 1 1
〈𝑉〉 = 𝑚𝜔2 (0 + 0 + 2(𝑚 + 1) − 1 = 𝜔ℏ (𝑚 + ) = 〈𝐻〉
2 2𝑚𝜔 2 2 2
 26

Problem 2.11
a)
∞ ∞
〈𝑥〉0 = ∫ 𝑥 𝜓0∗ (𝑥)𝜓0 𝑑𝑥 = ∫ (𝑜𝑑𝑑) (𝑒𝑣𝑒𝑛)2 𝑑𝑥 = 0
−∞ −∞
∞ ∞
〈𝑥〉1 = ∫ 𝑥 𝜓1∗ (𝑥)𝜓1 𝑑𝑥 = ∫ (𝑜𝑑𝑑) ⏟
(𝑜𝑑𝑑)2 𝑑𝑥 = 0
−∞ −∞ 𝑒𝑣𝑒𝑛

𝑑
〈𝑝〉0,1 = 𝑚 〈𝑥〉 = 0
𝑑𝑡 0,1

In general, for any stationary state of SHO

〈𝑥〉𝑚 = 〈𝑝〉𝑚 = 0
n=0
∞
𝑚𝜔 1/2 ∞ 2 −𝑚𝜔𝑥 2 𝑚𝜔 1/2 1 ℏ 𝜋ℏ
〈𝑥 2 〉0 = ∫ 𝑥 2 𝜓0∗ (𝑥)𝜓0 𝑑𝑥 = ( ) ∫ 𝑥 𝑒 ℏ 𝑑𝑥 = ( ) √ =
−∞ 𝜋ℏ −∞ 𝜋ℏ 2 𝑚𝜔 𝑚𝜔

1 ℏ
〈𝑥 2 〉0 =
2 𝑚𝜔

∞
𝑑2
〈𝑝2 〉0 = −ℏ2 ∫ 𝜓0∗ (𝑥) 𝜓 𝑑𝑥
−∞ 𝑑𝑥 2 0

𝑑2 𝑑 𝑑 𝑚𝜔 1/4 𝑑 𝑑 −𝑚𝜔𝑥 2 𝑚𝜔 1/4 𝑑 𝑚𝜔 −

𝑚𝜔 2
𝑥
2
𝜓0 = ( 𝜓 0 ) = ( ) 𝑒 2ℏ = ( ) (− 𝑥) 𝑒 2ℏ
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝜋ℏ 𝑑𝑥 𝑑𝑥 𝜋ℏ 𝑑𝑥 ℏ
𝑚𝜔 𝑚𝜔 1/4 𝑑 𝑚𝜔 2 𝑚𝜔 𝑚𝜔 1/4 𝑚𝜔 𝑚𝜔 2
= (− )( ) (𝑥𝑒 − 2ℏ 𝑥 ) = (− )( ) (1 − 𝑥 2 ) 𝑒 − 2ℏ 𝑥
ℏ 𝜋ℏ 𝑑𝑥 ℏ 𝜋ℏ ℏ

𝑚𝜔 𝑚𝜔 1/4 𝑚𝜔 1/4 ∞ 𝑚𝜔 2 −𝑚𝜔𝑥 2

〈𝑝2 〉0 = −ℏ (−2
)( ) ( ) ∫ (1 − 𝑥 ) 𝑒 ℏ 𝑑𝑥
ℏ 𝜋ℏ 𝜋ℏ −∞ ℏ

∞
𝑚𝜔 𝑚𝜔 1/2
2 −
𝑚𝜔 2
𝑥 𝑚𝜔 ∞ 2 −𝑚𝜔𝑥 2
=ℏ ( )( ) ∫ 𝑒 ℏ 𝑑𝑥 − ∫ 𝑥 𝑒 ℏ 𝑑𝑥
ℏ 𝜋ℏ ⏟−∞ ℏ ⏟−∞

{ √ 𝜋ℏ 1 ℏ √ 𝜋ℏ
}
𝑚𝜔 2𝑚𝜔 𝑚𝜔
 27

𝑚𝜔 𝑚𝜔 1/2 𝜋ℏ 1 𝑚𝜔
〈𝑝2 〉0 = ℏ2 ( )( ) √ {1 − } = ℏ2 ( )
ℏ 𝜋ℏ 𝑚𝜔 2 2ℏ

𝑚𝜔ℏ
〈𝑝2 〉0 =
2
Short-cut Method:
From
𝑝̂ 2 1
̂=
𝐻 + 𝑚𝜔2 𝑥̂ 2
2𝑚 2

We write

̂ − (𝑚𝜔)2 𝑥̂ 2 => 〈𝑝2 〉𝑛 = 2𝑚〈𝐻

𝑝̂ 2 = 2𝑚𝐻 ̂ 〉𝑛 − (𝑚𝜔)2 〈𝑥 2 〉𝑛
n=0
1 ℏ 1 1 ℏ 1
〈𝑝2 〉0 = 2𝑚𝐸0 − (𝑚𝜔)2 ( ) = 2𝑚 (𝜔ℏ (0 + )) − (𝑚𝜔)2 ( ) = 𝑚𝜔ℏ − 𝑚𝜔ℏ
2 𝑚𝜔 2 2 𝑚𝜔 2
1
〈𝑝2 〉0 = 𝑚𝜔ℏ
2

n=1
2
∞ ∞
𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2
〈𝑥 2 〉1 =∫ 𝑥 2
𝜓1∗ (𝑥)𝜓1 𝑑𝑥 = [( ) √ ] ∫ (𝑥)4 𝑒 − 2ℏ 𝑥 𝑑𝑥
−∞ 𝜋ℏ ℏ ⏟−∞
3 ℏ 2 √ 𝜋ℏ
( )
4 𝑚𝜔 𝑚𝜔
∞ 1 1
𝑚𝜔 2+1−2−2 3 ℏ 3
〈𝑥 2 〉1 = ∫ 𝑥 2 𝜓1∗ (𝑥)𝜓1 𝑑𝑥 = 2( ) =
−∞ ℏ 4 𝑚𝜔 2
3 ℏ
〈𝑥 2 〉1 =
2 𝑚𝜔

∞
𝑑2
〈𝑝2 〉1 2
= −ℏ ∫ 𝜓1∗ (𝑥) 𝜓 𝑑𝑥
−∞ 𝑑𝑥 2 1
𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2
𝜓1 = ( ) √ 𝑥𝑒 − 2ℏ 𝑥
𝜋ℏ ℏ
 28

𝑑2 𝑑 𝑑 𝑚𝜔 1/4 2𝑚𝜔 𝑑 𝑑 𝑚𝜔 2
2
𝜓1 = ( 𝜓1 ) = ( ) √ (𝑥𝑒 2ℏ 𝑥 )
−
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝜋ℏ ℏ 𝑑𝑥 𝑑𝑥

𝑚𝜔 1/4 2𝑚𝜔 𝑑 𝑚𝜔 2 −𝑚𝜔𝑥 2

=( ) √ (1 − 𝑥 ) 𝑒 2ℏ
𝜋ℏ ℏ 𝑑𝑥 ℏ

𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 2 𝑚𝜔 𝑚𝜔 𝑚𝜔 2
=( ) √ {(1 − 𝑥 ) (− 𝑥) + (−2 𝑥)} 𝑒 − 2ℏ 𝑥
𝜋ℏ ℏ ℏ ℏ ℏ

𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 𝑚𝜔 2 −𝑚𝜔𝑥 2

=( ) √ (− 𝑥) (3 − 𝑥 ) 𝑒 2ℏ
𝜋ℏ ℏ ℏ ℏ

2
𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 ∞ 2 𝑚𝜔 2 −𝑚𝜔𝑥 2
〈𝑝2 〉1 = −ℏ [(2
) √ ] (− ) ∫ 𝑥 (3 − 𝑥 ) 𝑒 2ℏ 𝑑𝑥
𝜋ℏ ℏ ℏ −∞ ℏ

2
∞
𝑚𝜔 1/4 2𝑚𝜔 𝑚𝜔 𝑚𝜔 2
2 − ℏ 𝑥
𝑚𝜔 ∞ 𝑚𝜔 2
2
= −ℏ [( ) √ ] (− ) 3∫ 𝑥 𝑒 𝑑𝑥 − ( ) ∫ (𝑥)4 𝑒 − 2ℏ 𝑥 𝑑𝑥
𝜋ℏ ℏ ℏ ⏟−∞ ℏ ⏟−∞
1 ℏ √ 𝜋ℏ 3 ℏ 2 √ 𝜋ℏ
{ ( ) }
2𝑚𝜔 𝑚𝜔 4 𝑚𝜔 𝑚𝜔

1
𝑚𝜔 1/2 2𝑚𝜔 𝑚𝜔 𝑚𝜔 −1−1/2 3 𝑚𝜔 −2−2+1 3
〈𝑝2 〉1 =ℏ ( 2
) ( )( ) {( ) √𝜋 − ( ) √𝜋}
𝜋ℏ ℏ ℏ ℏ 2 ℏ 4
3𝑚𝜔ℏ
〈𝑝2 〉1 = ( )
2
Short-cut Method:
From
̂ 〉𝑛 − (𝑚𝜔)2 〈𝑥 2 〉𝑛
〈𝑝2 〉𝑛 = 2𝑚〈𝐻
n=1
3 ℏ 1 3 ℏ 3
〈𝑝2 〉1 = 2𝑚𝐸1 − (𝑚𝜔)2 ( ) = 2𝑚 (𝜔ℏ (1 + )) − (𝑚𝜔)2 ( ) = 3𝑚𝜔ℏ − 𝑚𝜔ℏ
2 𝑚𝜔 2 2 𝑚𝜔 2
3
〈𝑝2 〉1 = 𝑚𝜔ℏ
2
 29

b) n=0

1 ℏ
𝜎𝑥 = √〈𝑥 2 〉0 − 〈𝑥〉20 = √〈𝑥 2 〉0 − 0 = √
2 𝑚𝜔

𝑚𝜔ℏ
𝜎𝑝 = √〈𝑝2 〉0 − 〈𝑝〉20 = √〈𝑝2 〉0 − 0 = √
2

1 ℏ 𝑚𝜔ℏ ℏ
𝜎𝑥 𝜎𝑝 = √ =
2 𝑚𝜔 2 2
n=1
3 ℏ
𝜎𝑥 = √〈𝑥 2 〉1 − 〈𝑥〉12 = √〈𝑥 2 〉1 − 0 = √
2 𝑚𝜔

3𝑚𝜔ℏ
𝜎𝑝 = √〈𝑝2 〉1 − 〈𝑝〉12 = √〈𝑝2 〉1 − 0 = √
2

3 ℏ 3𝑚𝜔ℏ ℏ
𝜎𝑥 𝜎𝑝 = √ =3
2 𝑚𝜔 2 2
 30

∞ (2𝑛)!
2 𝜋
∫ 𝑥 2𝑛 𝑒 −𝑎𝑥 𝑑𝑥 = 2𝑛
√ 2𝑛+1 𝑓𝑜𝑟 𝑎 > 0
−∞ 𝑛! 2 𝑎
 1

I) Analytic Method:
We return to Schrödinger eq. for the harmonic oscillator:

𝑑 2 𝜓 2𝑚 1
+ (𝐸 − 𝑚𝜔2 𝑥 2 ) 𝜓 = 0
𝑑𝑥 2 ℏ 2 ⏟
2
𝑉(𝑥)

𝑑 2 𝜓 2𝑚𝐸 𝑚𝜔 2 2 𝑚𝜔 1/2
+ 𝜓 − ( ) 𝑥 𝜓 = 0 ; 𝛽 = ( )
𝑑𝑥 2 ℏ2 ⏟ℏ ℏ
𝛽4

𝑑 2 𝜓 2𝑚𝐸
2
+ 2 𝜓 − 𝛽 4 𝑥 2 𝜓 = 0 (∗)
𝑑𝑥 ℏ
We will solve this equation directly by the series method.
Now, introduce the dimensionless quantity:
𝑦 ≡ 𝛽𝑥 ,
To write eq.(*) in terms of 𝑦, find
𝑑 2 𝜓 𝑑 2 𝜓 𝑑𝑦 2 𝑑 2 𝜓 2 𝑑 2 𝜓 2 2𝑚𝐸 4
𝑦2
= ( ) = 𝛽 => 𝛽 + 2 𝜓−𝛽 2𝜓 = 0
𝑑𝑥 2 𝑑𝑦 2 𝑑𝑥 𝑑𝑦 2 𝑑𝑦 2 ℏ 𝛽
OR,
𝑑 2 𝜓 2𝑚𝐸 1
+ 𝜓 − 𝑦2𝜓 = 0
𝑑𝑦 2 ⏟ℏ2 𝛽 2
2𝐸
ℏ𝜔
2𝐸
Let 𝛼 ≡ . Note that 𝛼 is dimensionless too.
ℏ𝜔
2𝐸 𝐸
𝛼≡ =
ℏ𝜔 1 ℏ𝜔
2
1
So that it is the energy in units of ℏ𝜔. Simplify the above eq.;
2

𝑑2𝜓
2
+ (𝛼 − 𝑦 2 )𝜓 = 0 (∗∗)
𝑑𝑦
To find the solution we first analyze the behavior of 𝜓(𝑦) in the asymptotic region |𝑦| → ∞.
In the limit |𝑦| → ∞ , for any finite value of the energy, the quantity 𝛼 becomes negligible with
respect to 𝑦 2 term In this case, the DE. is approximately given by
 2

𝑑2𝜓
2
− 𝑦2𝜓 ≈ 0
𝑑𝑦
with the solution
𝜓(𝑦) ≈ 𝐴 𝑒𝑥𝑝[−𝑦 2 /2] + 𝐵 𝑒𝑥𝑝[𝑦 2 /2]
However,
𝑒𝑥𝑝[𝑦 2 /2] → ∞ 𝑎𝑠 𝑦 → ±∞
Thus the 2.nd term is not normalizable. Therefore, we do not take it and
𝑦2
𝜓(𝑦) ≈ 𝑒𝑥𝑝 [− ]
2
This suggest that we make a guess about the solution as if we separate out the exponential part
𝜓(𝑦) = ℎ(𝑦)𝑒𝑥𝑝[−𝑦 2 /2]
To rewrite (**) in terms of ℎ(𝑦), note that
𝑑𝜓 𝑑H 2 2 𝑑ℎ 2
= [ 𝑒 −𝑦 /2 + ℎ(−𝑦𝑒 −𝑦 /2 )] = [ − ℎ𝑦] 𝑒 −𝑦 /2
𝑑𝑦 𝑑𝑦 𝑑𝑦
𝑑2𝜓 𝑑 2 ℎ 𝑑ℎ −𝑦 2 /2
𝑑ℎ 2
2
= [ 2
− 𝑦 − ℎ] 𝑒 + [ − ℎ𝑦] (−𝑦𝑒 −𝑦 /2 )
𝑑𝑦 𝑑𝑦 𝑑𝑦 𝑑𝑦
𝑑2𝜓 𝑑2ℎ 𝑑ℎ 2 −𝑦 2 /2
= [ − 2 𝑦 − ℎ(1 − 𝑦 )] 𝑒
𝑑𝑦 2 𝑑𝑦 2 𝑑𝑦
Plugging this into the DE (**), we get
𝑑2ℎ 𝑑ℎ 2 2
[ 2 − 2 𝑦 − ℎ(1 − 𝑦 2 )] 𝑒 −𝑦 /2 + (𝛼 − 𝑦 2 ) ℎ 𝑒 −𝑦 /2 = 0
𝑑𝑦 𝑑𝑦

𝑑2ℎ 𝑑ℎ
− 2 𝑦 + (𝛼 − 1) ℎ = 0
𝑑𝑦 2 𝑑𝑦
which is called the Hermite equation.
We look for solutions to this equation for ℎ(𝑦) in the form of a power series in 𝑦:
∞

ℎ(𝑦) = ∑ 𝐶𝑛 𝑦 𝑛 (∗∗∗)
𝑛=0

(Note that according to Taylor’s theorem, any reasonable well-behaved function can be
expressed as a power series, so eq. (***) involves no loss of generality.)
 3
∞ ∞
𝑑ℎ 𝑛−1
𝑑2ℎ
= ∑ 𝐶𝑛 𝑛𝑦 , 2
= ∑ 𝐶𝑛 𝑛(𝑛 − 1)𝑦 𝑛−2
𝑑𝑦 𝑑𝑦
𝑛=0 𝑛=0
∞ ∞ ∞

∑ 𝐶𝑛 𝑛(𝑛 − 1)𝑦 𝑛−2 − 2 ∑ 𝐶𝑛 𝑛𝑦 𝑛−1 𝑦 + (𝛼 − 1) ∑ 𝐶𝑛 𝑦 𝑛 = 0

𝑛=0 𝑛=0 𝑛=0

Simplifying
∞ ∞

∑ 𝐶𝑛 𝑛(𝑛 − 1)𝑦 𝑛−2 + ∑ 𝐶𝑛 ((𝛼 − 1) − 2𝑛)𝑦 𝑛 = 0

𝑛=0 𝑛=0

Replacing (𝑛 − 2) → 𝑚 in the first term, and 𝑛 → 𝑚 in the second term:

∞ ∞

∑ 𝐶𝑚+2 (𝑚 + 2)(𝑚 + 1)𝑦 𝑚 + ∑ 𝐶𝑚 ((𝛼 − 1) − 2𝑚)𝑦 𝑚 = 0

𝑚=−2 𝑚=0

But 𝑚 = −2, −1 in the 1.st sum is zero. Therefore, we have

∞

∑ {𝐶𝑚+2 (𝑚 + 2)(𝑚 + 1) + 𝐶𝑚 ((𝛼 − 1) − 2𝑚)} 𝑦 𝑚 = 0

𝑚=0

It follows that coefficient of each power of 𝑦 must vanish. So, we arrive at the recursion
relation:
(2𝑚 + 1 − 𝛼)
𝐶𝑚+2 = 𝐶
(𝑚 + 2)(𝑚 + 1) 𝑚
This recursion formula generates all the even-numbered coefficients, starting with 𝐶0 ;

(1 − 𝛼) (5 − 𝛼) (5 − 𝛼)(1 − 𝛼)
𝐶2 = 𝐶0 ; 𝐶4 = 𝐶2 = 𝐶0 ;
2 12 12 ∙ 2

(9 − 𝛼) (9 − 𝛼)(5 − 𝛼)(1 − 𝛼)
𝐶6 = 𝐶4 = 𝐶0 ; . 𝑒𝑡𝑐
30 30 ∙ 12 ∙ 2

And all the odd-numbered coefficients, starting with 𝐶1

(3 − 𝛼) (7 − 𝛼) (7 − 𝛼)(3 − 𝛼)
𝐶3 = 𝐶1 ; 𝐶5 = 𝐶3 = 𝐶1 ;
6 20 20 ∙ 6

(11 − 𝛼) (11 − 𝛼)(7 − 𝛼)(3 − 𝛼)

𝐶7 = 𝐶5 = 𝐶1 ; . 𝑒𝑡𝑐
42 42 ∙ 20 ∙ 6
 4

We write the complete solution as

ℎ(𝑦) = ℎ𝑒𝑣𝑒𝑛 (𝑦) + ℎ𝑜𝑑𝑑 (𝑦)
∞

ℎ𝑒𝑣𝑒𝑛 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶0 + 𝐶2 𝑦 2 + 𝐶4 𝑦 4 + ⋯
𝑚:𝑒𝑣𝑒𝑛
∞

ℎ𝑜𝑑𝑑 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶1 𝑦 + 𝐶3 𝑦 3 + 𝐶5 𝑦 5 + ⋯
𝑚:𝑜𝑑𝑑

These equations together with the recursion relation determine ℎ(𝑦) in terms of two arbitrary
constants 𝐶0 and 𝐶1 , which is just what we would expect for a second-order differential
equation.
In theory we have solved the problem. However, there is a problem here: Not all the solutions
so obtained are normalizable.
In the limit of large m, we find that
𝐶𝑚+2 2
lim →
𝑚→∞ 𝐶𝑚 𝑚
2
It is possible to show that this is the behavior exhibited by 𝑒 𝑦 as 𝑦 gets larger:

𝑦2
𝑦4 𝑦6
2
𝑦𝑚 𝑦 𝑚+2
𝑒 =1+𝑦 + + + ⋯+ 𝑚 + +⋯
2! 3! (2)! ( 𝑚 + 2
!
2 )
Thus
1 1
𝐶𝑚 = 𝑚 , 𝐶𝑚+2 = 𝑚
(2)! ( 2 + 1) !

For large m, the ratio of the coefficients of successive powers of y gives

1
( 𝑚 ) 𝑚 𝑚
𝐶𝑚+2 ( 2 + 1) ! (2)! (2)! 𝑓𝑜𝑟 𝑙𝑎𝑟𝑔𝑒 𝑚 2
lim = = 𝑚 = 𝑚 𝑚 →
𝑚→∞ 𝐶𝑚 𝑚
1 ( 2 + 1) ! ( 2 + 1) ( 2 ) !
( 𝑚 )
(2)!
 5

Therefore, for large m we have

2
ℎ(𝑦)~𝑒 𝑦
In that case
2 2 /2
𝜓(𝑦)~𝑒 𝑦 𝑒𝑥𝑝[−𝑦 2 /2]~𝑒 𝑦 → ∞ for large values of 𝑦
which is not physically acceptable. Therefore for normalizable solutions there is only one way
to avoid this behavior: we must demand that 𝒉(𝒚) is a series that must terminate.
Lets assume that at 𝑚 = 𝑛, series terminates and 𝐶𝑛 is the last term. Then all 𝐶𝑝 with 𝑝 > 𝑛
must be zero =>
(2𝑛 + 1 − 𝛼)
𝐶𝑛+2 = 𝐶 = 0 => (2𝑛 + 1 − 𝛼) = 0 => 𝛼 = 2𝑛 + 1
(𝑛 + 2)(𝑛 + 1) 𝑛
Note that
2𝐸
𝛼≡
ℏ𝜔
Therefore
2𝐸 1
= 2𝑛 + 1 => 𝐸 = (𝑛 + ) ℏ𝜔 , 𝑛 = 0,1,2,3, … . .
ℏ𝜔 2

From the requirement that the power series converge, and therefore truncate, we find that the
energy levels for the harmonic oscillator are quantized.
 6

Wave Functions:
The recursion relation
(2𝑚 + 1 − 𝛼)
𝐶𝑚+2 = 𝐶
(𝑚 + 2)(𝑚 + 1) 𝑚
For the allowed values of 𝛼 = 2𝑛 + 1, it takes the form
2(𝑚 − 𝑛)
𝐶𝑚+2 = 𝐶
(𝑚 + 2)(𝑚 + 1) 𝑚
This recursion formula generates all the even-numbered coefficients, starting with 𝐶0 ;

−2𝑛
𝑚 = 0; 𝐶2 = 𝐶
2 0

2(2 − 𝑛) 2(2 − 𝑛)(−𝑛)

𝑚 = 2; 𝐶4 = 𝐶2 = 𝐶0 ;
12 12

2(4 − 𝑛) 4(4 − 𝑛)(2 − 𝑛)(−𝑛)

𝑚 = 6; 𝐶6 = 𝐶4 = 𝐶0 ; . 𝑒𝑡𝑐
30 30 ∙ 12

And all the odd-numbered coefficients, starting with 𝐶1

2(1 − 𝑛)
𝑚=1; 𝐶3 = 𝐶1
6

2(3 − 𝑛) 4(3 − 𝑛)(1 − 𝑛)

𝑚 = 3; 𝐶5 = 𝐶3 = 𝐶1 ;
20 20 ∙ 6

2(5 − 𝑛) 8(5 − 𝑛)(3 − 𝑛)(1 − 𝑛)

𝑚=5; 𝐶7 = 𝐶5 = 𝐶1 ; . 𝑒𝑡𝑐
42 42 ∙ 20 ∙ 6

Therefore, when n= even, ℎ(𝑦) = ℎ𝑒𝑣𝑒𝑛 (𝑦), survives and we separately insist that 𝑐1 = 0 to
kill ℎ𝑜𝑑𝑑 (𝑦).
Similarly, when n= odd, ℎ(𝑦) = ℎ𝑜𝑑𝑑 (𝑦), survives and we separately insist that 𝑐0 = 0 to kill
ℎ𝑒𝑣𝑒𝑛 (𝑦).
 7

Let us construct some solutions for this. Note that 𝑛 ≡ 𝑚𝑚𝑎𝑥 and

𝜓(𝑦) = ℎ(𝑦)𝑒𝑥𝑝[−𝑦 2 /2]

 𝑛 = 0. (even)
0

ℎ(𝑦) = ℎ𝑒𝑣𝑒𝑛 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶0

𝑚:𝑒𝑣𝑒𝑛

=> We insist that 𝑐1 = 0.

2(𝑚 − 0)
𝐶𝑚+2 = 𝐶
(𝑚 + 2)(𝑚 + 1) 𝑚

𝑚 = 0; 𝐶2 = 0
𝜓0 = 𝐶0 𝑒𝑥𝑝[−𝑦 2 /2]
With
𝑚𝜔 1/2
𝑦=( ) 𝑥
ℏ
𝜓0 (𝑥) = 𝐶0 𝑒𝑥𝑝[−𝑚𝜔𝑥 2 /2ℏ]
Normalization=>

∞ ∞
2 /ℏ 𝜋ℏ 𝑚𝜔 1/4
1=∫ 𝜓0∗ (𝑥) 𝜓0 (𝑥)𝑑𝑥 = 2
𝐶0 ∫ 𝑒 −𝑚𝜔𝑥 𝑑𝑥 = 𝐶02 √ => 𝐶0 = ( )
−∞ −∞ 𝑚𝜔 𝜋ℏ

𝑚𝜔 1/4 −𝑚𝜔𝑥 2/2ℏ

𝜓0 (𝑥) = ( ) 𝑒
𝜋ℏ
 𝑛 = 1 (odd)
1

ℎ(𝑦) = ℎ𝑜𝑑𝑑 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶1 𝑦

𝑚=𝑜𝑑𝑑

=> We insist that 𝑐0 = 0.

2(𝑚 − 1)
𝐶𝑚+2 = 𝐶 => 𝑚 = 1 => 𝐶3 = 0
(𝑚 + 2)(𝑚 + 1) 𝑚
𝜓1 = 𝐶1 𝑦 𝑒𝑥𝑝[−𝑦 2 /2]
 8

OR
𝑚𝜔 1/2 2
𝜓1 = 𝐶1 ( ) 𝑥 𝑒 −𝑚𝜔𝑥 /2ℏ
ℏ

Normalization=>

∞
𝑚𝜔 ∞ 2 −𝑚𝜔𝑥 2 𝑚𝜔 1 ℏ 𝜋ℏ
1=∫ 𝜓1∗ (𝑥) 𝜓1 (𝑥)𝑑𝑥 = 𝐶12 ( ) ∫ 𝑥 𝑒 ℏ 𝑑𝑥 = 𝐶12 ( ) ( )√
−∞ ℏ −∞ ℏ 2 𝑚𝜔 𝑚𝜔

𝜋ℏ 1/2 1 𝑚𝜔 1/4
= 𝐶12 ( ) => 𝐶1 = ( ) √2
𝑚𝜔 2 𝜋ℏ
𝑚𝜔 1/4 𝑚𝜔 1/2 2
𝜓1 = ( ) (2 ) 𝑥 𝑒 −𝑚𝜔𝑥 /2ℏ
𝜋ℏ ℏ
4 1/4 𝑚𝜔 3/4 2
𝜓1 = ( ) ( ) 𝑥 𝑒 −𝑚𝜔𝑥 /2ℏ
𝜋 ℏ
 n=2. (even)

ℎ(𝑦) = ℎ𝑒𝑣𝑒𝑛 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶0 + 𝐶2 𝑦 2

𝑚:𝑒𝑣𝑒𝑛

=> We insist that 𝑐1 = 0.

2(𝑚 − 2) 4
𝐶𝑚+2 = 𝐶𝑚 : 𝑚 = 0 => 𝐶2 = − 𝐶0 𝑎𝑛𝑑 𝑚 = 2 => 𝐶4 = 0
(𝑚 + 2)(𝑚 + 1) 2
ℎ(𝑦) = ℎ𝑒𝑣𝑒𝑛 (𝑦) = 𝐶0 (1 − 2𝑦 2 )
𝜓2 = 𝐶0 (1 − 2𝑦 2 ) 𝑒𝑥𝑝[−𝑦 2 /2]
OR

𝜓2 = 𝐶0 (2 − 4𝑦 2 ) 𝑒𝑥𝑝[−𝑦 2 /2]
 n=3 (odd)
3

ℎ(𝑦) = ℎ𝑜𝑑𝑑 (𝑦) = ∑ 𝐶𝑚 𝑦 𝑚 = 𝐶1 𝑦 + 𝐶3 𝑦 3

𝑚=0𝑑𝑑
=> We insist that 𝑐0 = 0.
 9

2(𝑚 − 3)
𝐶𝑚+2 = 𝐶
(𝑚 + 2)(𝑚 + 1) 𝑚
2(1 − 3) −4
𝑚 = 1: 𝐶3 = 𝐶1 => 𝐶3 = 𝐶1 ,
(1 + 2)(1 + 1) 6
2 1
ℎ(𝑦) = ℎ𝑜𝑑𝑑 (𝑦) = 𝐶1 (𝑦 − 𝑦 3 ) = 𝐶1 (12𝑦 − 8𝑦 3 )
3 12

1
𝜓3 = 𝐶1 (12𝑦 − 8𝑦 3 )
12
In general ℎ(𝑦) will be a polynomial of degree n in y, involving even powers only, if n is an even
integer, and odd powers only, if n is an odd integer. Apart from the overall factor (𝐶0 or 𝐶1 ), they
are the so called Hermite polynomials , 𝐻𝑛 (𝑦).

By tradition, the arbitrary multiplicative factor is chosen so that the coefficient of the highest
power of y is 2𝑛 . With this convention, the normalized stationary states for the HO are

𝑚𝜔 1/4 1
𝜓𝑛 = ( ) 𝐻𝑛 (𝑦) 𝑒𝑥𝑝[−𝑦 2 /2]
𝜋ℏ 𝑛
√2 𝑛!
 10

The quantum oscillator is strikingly different from its classical counterpart:

1. Energies are quantized
2. The position distributions have some bizarre features: For instance;
 Probability of finding the particle outside the classically allowed region (where
x>classical amplitude for the energy in question) is not zero.
 In all odd states the probability of finding the particle at the center is zero.
 11

At large n, we begin to see some resemblance to the classical case. Classical probability of
finding the particle in the interval (x,x+dx) is proportional to time dt which it takes to pass
through this interval :

dt 2
P(x) dx  where T : the period of oscillation.
T 
dt dt dx  dx  dx
Then, P(x) dx    
T 2/ dx 2 dx/dt 2 v
1 1 1 k
To find v, E  m v 2  k x 2 = k A 2  v 2  (A 2  x 2 )  v   2 (A 2  x 2 )
2 2 2 m
 dx 1
Then, P(x) dx   P(x) 
2  A 2  x 2 2 A 2  x 2
 12

Comparison of Classical and Quantum Probabilities for Harmonic Oscillator

http://hyperphysics.phy-
astr.gsu.edu/hbase/quantum/hosc6.html#:~:text=The%20fact%20that%20the%20overall,is%20called%20the%20correspondence%20
principle.&text=The%20quantum%20probabilities%20do%20extend,exponentially%20decaying%20into%20that%20region.
 13

It has been superimposed the classical position distribution on the quantum one (for
n=100). If you smoothed out the bumps, the two would fit pretty well.
The fact that the overall picture of probability of finding the oscillator at a given
value of x converges for the quantum and classical pictures is called
the correspondence principle.
14
 15

Problem 2.17:
a) The Rodrigues formula says that

2 𝑑 𝑚 −𝑦 2
𝐻𝑚 (𝑦) = (−1)𝑚 𝑒 𝑦 𝑒
𝑑𝑦 𝑚
2 𝑑 0 −𝑦 2
𝐻0 (𝑦) = (−1)0 𝑒 𝑦 𝑒 =1
𝑑𝑦 0
2 𝑑 −𝑦 2 2 2
𝐻1 (𝑦) = (−1)1 𝑒 𝑦 𝑒 = −𝑒 𝑦 (−2𝑦)𝑒 −𝑦 = 2𝑦
𝑑𝑦
2 𝑑2 −𝑦 2 𝑦2
𝑑 2 2 2 2
𝐻2 (𝑦) = (−1)2 𝑒 𝑦 𝑒 = 𝑒 (−2𝑦𝑒 −𝑦 ) = 𝑒 𝑦 (−2𝑒 −𝑦 − 2𝑦(−2𝑦)𝑒 −𝑦 )
𝑑𝑦 2 𝑑𝑦
𝐻2 (𝑦) = −2 + 4𝑦 2

2 𝑑3 −𝑦 2 3 𝑦2
𝑑2 −𝑦 2 3 (−2)𝑒 𝑦 2
𝑑1 2
𝐻3 (𝑦) = (−1)3 𝑒 𝑦 𝑒 = (−1) 𝑒 [(−2𝑦)𝑒 ] = (−1) [(1−2𝑦 2 )𝑒 −𝑦 ]
𝑑𝑦 3 𝑑𝑦 2 𝑑𝑦 1
2 2
= 2𝑒 𝑦 [(−4𝑦) + (1−2𝑦 2 )(−2𝑦)]𝑒 −𝑦

𝐻3 (𝑦) = −12𝑦 + 8𝑦 3
Similarly,

2 𝑑 4 −𝑦 2
𝐻4 (𝑦) = (−1)4 𝑒 𝑦 4
𝑒 = 16𝑦 4 − 48𝑦 2 + 12
𝑑𝑦

b)

𝐻𝑛+1 (𝑦) = 2𝑦𝐻𝑛 (𝑦) − 2𝑛𝐻𝑛−1 (𝑦)

𝐹𝑜𝑟 𝑛 = 4
𝐻5 (𝑦) = 2𝑦𝐻4 (𝑦) − 2 (4)𝐻3 (𝑦) = 2𝑦[16𝑦 4 − 48𝑦 2 + 12] − 8[−12𝑦 + 8𝑦 3 ]

𝐻5 (𝑦) = 32𝑦 5 − 160𝑦 3 + 120𝑦

𝐹𝑜𝑟 𝑛 = 5
𝐻6 (𝑦) = 2𝑦𝐻5 (𝑦) − 2 (4)𝐻4 (𝑦) = 2𝑦[32𝑦 5 − 160𝑦 3 + 120𝑦] − 2(5)[16𝑦 4 − 48𝑦 2 + 12]

𝐻6 (𝑦) = 64𝑦 6 − 480𝑦 4 + 720𝑦 2 − 120

 1

THE FREE PARTICLE

V(x)=0 everywhere.
Classically, this would just mean motion at constant velocity, but in QM the problem
is surprisingly tricky.
Time-independent Schrödinger equation reads for a free particle
−ℏ2 𝑑2
[ ⏟ ] 𝜓(𝑥 ) = 𝐸𝜓(𝑥 )
+ 𝑉(𝑥)
2𝑚 𝑑𝑥 2
0

OR
𝑑2 𝜓(𝑥 ) 2𝑚𝐸
+ 2 𝜓(𝑥 ) = 0
𝑑𝑥 2 ⏟ℏ
𝑘2

Solution
𝜓(𝑥 ) = 𝐴𝑒 𝑖𝑘𝑥 + 𝐵𝑒 −𝑖𝑘𝑥
Unlike the infinite well, there are no boundary conditions to restrict the possible
values of k and hence E: The fee particle can carry any (+) energy.
Tacking on the standard time dependence
𝑖𝐸𝑡 𝑖𝐸𝑡
− 𝑖𝑘𝑥 −𝑖𝑘𝑥 −
Ψ(𝑥, 𝑡) = 𝜓(𝑥 )𝑒 ℏ = [𝐴𝑒 + 𝐵𝑒 ]𝑒 ℏ

and
Ψ(𝑥, 𝑡) = 𝐴𝑒 𝑖𝑘𝑥 𝑒 −𝑖𝐸𝑡/ℏ + 𝐵𝑒 −𝑖𝑘𝑥 𝑒 −𝑖𝐸𝑡/ℏ
OR with
ℏ2 𝑘 2
𝐸=
2𝑚
 2

ℏ𝑘 ℏ𝑘
Ψ(𝑥, 𝑡) = 𝐴 𝑒𝑥𝑝 [𝑖𝑘 (𝑥 − 𝑡)] + 𝐵𝑒𝑥𝑝 [−𝑖𝑘 (𝑥 + 𝑡)]
2𝑚 2𝑚
The structure of this solution is characteristic of a propagating wave.
Any function of 𝑥 and 𝑡 that depends on these variables in the special combination
ℏ𝑘
(𝑥 ∓ 𝑣𝑡) , where 𝑣 = = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡, represents a wave of fixed profile travelling
2𝑚

in the ± direction at speed 𝑣.

Ψ+ (𝑥, 𝑡) = 𝐴𝑒 𝑖𝑘(𝑥−𝑣𝑡)
Ψ− (𝑥, 𝑡) = 𝐵𝑒 −𝑖𝑘(𝑥+𝑣𝑡)
Function Ψ+ represents a wave propagating in +x direction with velocity v.
Function Ψ− represents a wave propagating in -x direction with velocity v.

Since every point on the waveform is moving along with the same velocity, its shape
doesn’t change as it propagates.

At any instant, one may plot the x-dependence of

Ψ+ . If t increases to 𝑡 + ∆𝑡, this curve is
displaced to the right (as a rigid body) by the
amount 𝑣∆𝑡.

To see this, let’s check the following:

?
Ψ+ (𝑥, 𝑡) =
⏞ Ψ+ (𝑥 + ∆𝑥, 𝑡 + ∆𝑡)

𝑥 → 𝑥 + ∆𝑥 ≡ 𝑥 + 𝑣∆𝑡

Ψ+ (𝑥 + 𝑣∆𝑡, 𝑡 + ∆𝑡) = 𝐴 𝑒𝑥𝑝[𝑖𝑘 (𝑥 + 𝑣∆𝑡 − 𝑣(𝑡 + ∆𝑡)] = 𝑒𝑥𝑝[𝑖𝑘 (𝑥 − 𝑣𝑡)] = Ψ+ (𝑥, 𝑡)

 3

Since Ψ+ (𝑥, 𝑡) and Ψ− (𝑥, 𝑡) only differ by the sign of k, we might as well write
ℏ𝑘 2
𝑖(𝑘𝑥− 𝑡)
Ψk (𝑥, 𝑡) = 𝐴 𝑒 2𝑚

with

2𝑚𝐸 𝑘 > 0 => 𝑡𝑟𝑎𝑣𝑒𝑙𝑙𝑖𝑛𝑔 𝑡𝑜 𝑡ℎ𝑒 𝑟𝑖𝑔ℎ𝑡

𝑘 ≡ ±√ 𝑤𝑖𝑡ℎ {
ℏ2 𝑘 < 0 => 𝑡𝑟𝑎𝑣𝑒𝑙𝑙𝑖𝑛𝑔 𝑡𝑜 𝑡ℎ𝑒 𝑙𝑒𝑓𝑡

Conclusion:
 Stationary states of the free particle are propagating waves.
 Their wavelength is
2𝜋
𝜆=
|𝑘|
From De Broglie formula
ℎ
𝜆=
𝑝

Therefore, we have
𝑝 = ℏ𝑘

 The speed of these waves representing the free particles are

ℏ𝑘 1 𝑝
𝑣𝑞𝑢𝑎𝑛𝑡𝑢𝑚 = =
2𝑚 2 𝑚

On the other hand, the classical speed of a free particle with mass m is
𝑝
𝑣𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 = = 2 𝑣𝑞𝑢𝑎𝑛𝑡𝑢𝑚 : Discrepancy‼
𝑚
We will return this paradox soon.
In addition, the plane wave solution is not normalizable.
∞ ∞

∫ |Ψk (𝑥, 𝑡)|2 𝑑𝑥 = |A|2 ∫ 𝑑𝑥 = ∞

−∞ −∞
 4

Then, in case of the free particle the stationary state solutions do not represent
physically realizable states.
A free particle cannot exist in a stationary state; or
There is no such thing as a free particle with a definite energy.

But this doesn’t mean that the stationary state solutions are of no use to us, for they
play a mathematical role that is entirely independent of their physical interpretation.

The way out of this difficulty is to write the most general solution of the time-
dependent Schrödinger equation for the free particle as a superposition of stationary
state solutions:
1 1 ℏ𝑘 2
𝑖(𝑘𝑥− 2𝑚 𝑡)
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)Ψk (𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 (∗)
√2𝜋 √2𝜋
1
The factor is for convenience.
√2𝜋
Compare this with the one for ∞ well:
∞

Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡)
𝑛=0

∑ → ∫ 𝑑𝑘
𝑛
𝑐𝑛 → 𝜙(𝑘)

The most general solution of the free particle necessarily carries a range of k’s and
hence a range of energies and speeds. We call it a wave packet; it is a superposition
of sinusoidal functions whose amplitudes are modulated by 𝜙(𝑘).

Now, this wave function can be normalized if 𝜙(𝑘) is normalized.

 5

DIRAC DELTA FUNCTION

Consider the function 𝛿 (𝜀) (𝑥) with the following properties:

1 𝜀 𝜀
𝑤ℎ𝑒𝑛 − <𝑥<
𝛿 (𝜀) (𝑥 ) = { 𝜀 2
𝜀
2 (1)
0 𝑓𝑜𝑟 |𝑥 | >
2

which is given in the following curve, whose area is one:

 The Dirac delta function 𝛿(𝑥) can be defined as the limit of 𝛿 (𝜀) (𝑥) as 𝜀 → 0:

𝛿 (𝑥 ) = lim 𝛿 (𝜀) (𝑥) (∗)

𝜀→0
with the property that
∞ 𝑓𝑜𝑟 𝑥=0
𝛿 (𝑥 ) = { (2)
0 𝑓𝑜𝑟 𝑥≠0
and from (*)
∞ 𝜀/2
1
∫ 𝛿 (𝑥 ) 𝑑𝑥 = lim ∫ 𝛿 (𝜀) (𝑥) 𝑑𝑥 = lim ( 𝜀) = 1 (3)
−∞ 𝜀→0 −𝜀/2 𝜀→0 𝜀

Dirac Delta function is a very peculiar kind of function. Actually, it is not a function
of the usual mathematical sense (it is not finite at x=0!) , but it is rather “a generalized
function” or “a distribution” , It is an infinitesimally narrow spike at the origin. But,
it is an extremely useful construct in theoretical physics.
 6

 Note that if one multiplies 𝛿 (𝑥 − 𝑥0 ) by an arbitrary function 𝑓 (𝑥 ), it is the same

as multiplying by 𝑓(𝑥0 ):

𝛿 (𝑥 − 𝑥0 ) 𝑓(𝑥0 ) = 𝛿 (𝑥 − 𝑥0 ) 𝑓(𝑥 ) (4)

since from (2), 𝛿 (𝑥 − 𝑥0 ) = 0 when 𝑥 ≠ 𝑥0 . Integrating (4)

∞ ∞
∫ 𝛿 (𝑥 − 𝑥0 ) 𝑓(𝑥0 )𝑑𝑥 = ∫ 𝛿 (𝑥 − 𝑥0 ) 𝑓(𝑥 )𝑑𝑥
−∞ −∞
∞
𝐿𝐻𝑆 = 𝑓 (𝑥0 ) ∫ 𝛿 (𝑥 − 𝑥0 ) 𝑑𝑥 => 𝐿𝐻𝑆 = 𝑓(𝑥0 )
⏟−∞
=1
Therefore we get
∞
∫ 𝛿 (𝑥 − 𝑥0 ) 𝑓(𝑥 )𝑑𝑥 = 𝑓(𝑥0 ) (5)
−∞
This is the most important property of Dirac delta function. Of couse the limits of
the integral need not go from −∞ to +∞; all that matters is that the domain of
integration include the point 𝑥0 , so 𝑥0 − ∞ to 𝑥0 +∞ would do for any 𝜖 > 0
 Now let’s prove that
∞
1
𝛿 (𝑥 ) = ∫ 𝑑𝑥 𝑒 𝑖𝑘𝑥
2𝜋
−∞

To prove this we use Fourier integral given by Plancherel’s theorem

F(k) is called the Fourier transform of f(x); f(x) is the inverse Fourier transform
of F(x).

Let
𝑓 (𝑥 ) = 𝛿 (𝑥 )
 7

Put into the 2.nd integral

∞
(5)
1 −𝑖𝑘𝑥
1 1
𝐹 (𝑘) = ∫ 𝑑𝑥 𝛿 (𝑥 ) ⏟
𝑒 =
⏞ 𝑔(0) = 1
√2𝜋 𝑔(𝑥) √2𝜋 √2𝜋
−∞

Then, put into the 1.st integral, now we have 𝑓 (𝑥 ) = 𝛿 (𝑥 ):

∞ ∞
1 1 1
𝛿 (𝑥 ) = ∫ 𝑑𝑘 𝑒 𝑖𝑘𝑥 => 𝛿 (𝑥 ) = ∫ 𝑑𝑘 𝑒 𝑖𝑘𝑥 (𝑎)
√2𝜋 √2𝜋 2𝜋
−∞ −∞

OR,
∞
1
𝛿 (𝑘) = ∫ 𝑑𝑥 𝑒 −𝑖𝑘𝑥 (𝑏)
2𝜋
−∞

--------------------------------------------------------------------------------------
The most general solution of the time-dependent Schrödinger equation for the free
particle as a superposition of the stationary state solutions:
1 1 ℏ𝑘 2
𝑖(𝑘𝑥− 𝑡)
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)Ψk (𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 2𝑚 (∗)
√2𝜋 √2𝜋

Now, this wave function can be normalized if 𝜙(𝑘) is normalized.

Proof:
∞ ∞
1
∫ 𝑑𝑥 |Ψ(𝑥, 𝑡)|2 = ∫ 𝑑𝑥 {(∫ 𝑑𝑘′ 𝜙(𝑘′)𝑒 𝑖(𝑘′𝑥−𝜔′𝑡) ) (∫ 𝑑𝑘 𝜙 ∗ (𝑘)𝑒 −𝑖(𝑘𝑥−𝜔𝑡) )}
2𝜋
∞ ∞
∞
1
= ∫ 𝑑𝑥 𝑒 𝑖(𝑘′𝑥−𝜔′𝑡) 𝑒 −𝑖(𝑘𝑥−𝜔𝑡) {(∫ 𝑑𝑘′ 𝜙(𝑘′)) (∫ 𝑑𝑘 𝜙 ∗ (𝑘))}
2𝜋
∞

Note that
∞ ∞
1 1
∫ 𝑑𝑥 𝑒 −𝑖(𝑘𝑥−𝜔𝑡) 𝑒 𝑖(𝑘′𝑥−𝜔′𝑡) = 𝑒 𝑖(𝜔−𝜔′) ∫ 𝑑𝑥 𝑒 𝑖(𝑘′−𝑘)𝑥
2𝜋 2𝜋
−∞ −∞
 8

Since
∞
1
∫ 𝑑𝑥 𝑒 𝑖(𝑘′−𝑘)𝑥 = 𝛿(𝑘 ′ − 𝑘)
2𝜋
−∞

Then
∞

∫ 𝑑𝑥 |Ψ(𝑥, 𝑡)|2
∞
𝑘=𝑘 ′
𝜔=𝜔′
= ∫ 𝑑𝑘 ∫ 𝑑𝑘′ 𝜙(𝑘′)𝜙 ∗ (𝑘)𝛿(𝑘 − 𝑘′)𝑒 𝑖(𝜔−𝜔′) =
⏞ ∫ 𝑑𝑘 𝜙(𝑘)𝜙 ∗ (𝑘)
∞ ∞

∫ 𝑑𝑥 |Ψ(𝑥, 𝑡)|2 = ∫ 𝑑𝑘 |𝜙(𝑘)|2

∞ ∞

Therefore, as long as 𝜙(𝑘) is normalized , so does Ψ(𝑥, 𝑡).

-------------------------------------------------------------------------------------------

In a generic quantum problem, we are given Ψ(𝑥, 0) and we are asked to find
Ψ(𝑥, 𝑡). For the free particle it is given by (*). But what is 𝜙(𝑘) there to match with
the initial packet ?
We write from (*)
1
⏟(𝑥, 0) =
Ψ ∫ 𝑑𝑘 𝜙(𝑘)𝑒 𝑖𝑘𝑥
𝑔𝑖𝑣𝑒𝑛 √2𝜋
To find 𝜙(𝑘)
1
Ψ(𝑥, 0)𝑒 −𝑖𝑘′𝑥 = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 𝑖𝑘𝑥 𝑒 −𝑖𝑘′𝑥
√2𝜋

Then integrate
 9

∞ ∞
1
∫ 𝑑𝑥 Ψ(𝑥, 0)𝑒 −𝑖𝑘′𝑥 = ∫ 𝑑𝑘 𝜙(𝑘) ∫ 𝑑𝑥 𝑒 𝑖𝑘𝑥 𝑒 −𝑖𝑘′𝑥
√2𝜋 ⏟
−∞ −∞
2𝜋𝛿(𝑘−𝑘′)
2𝜋
= ∫ 𝑑𝑘 𝜙(𝑘)𝛿(𝑘 − 𝑘′)
√2𝜋
∞
′𝑥
∫ 𝑑𝑥 Ψ(𝑥, 0)𝑒 −𝑖𝑘 = √2𝜋𝜙(𝑘 ′ ) =>
−∞
∞
1
𝜙(𝑘′) = ∫ 𝑑𝑥 Ψ(𝑥, 0)𝑒 −𝑖𝑘′𝑥
√2𝜋
−∞
So the solution to the generic QM problem for the free particle is to replace 𝜙 in (*).

𝜙(𝑘) is called the Fourier transform of Ψ(𝑥, 0); Ψ(𝑥, 0)is the inverse Fourier
transform of 𝜙(𝑘).
 10

Now, we return the paradox:

𝑝
𝑣𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 = = 2 𝑣𝑞𝑢𝑎𝑛𝑡𝑢𝑚 : Discrepancy‼
𝑚

Ψ(𝑥, 𝑡) is a wave packet; it is a superposition of sinusoidal functions whose

amplitude is modulated by 𝜙(𝑘)

It consists of “ripples” contained within an “envelope”. What corresponds to the

particle velocity is not the speed of the individual ripples (the so-called phase
velocity), but rather the speed of the envelope (the group velocity).
 11

Example: For the wave function of a free particle, we can show that 𝑣𝑔 = 2𝑣𝑝 . To
show this, we determine the group velocity of a wave packet with the general form
1
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 𝑖(𝑘𝑥−𝜔(𝑘)𝑡)
√2𝜋
Let us assume that 𝜙(𝑘) is narrowly peak about some particular point 𝑘0 .

We may then Taylor expand 𝜔(𝑘) about 𝑘0 , since the integrand is negligible
except at the vicinity of 𝑘0 .
𝑑𝜔(𝑘)
𝜔(𝑘) ≅ 𝜔0 + | (𝑘 − 𝑘0 )
𝑑𝑘 𝑘=𝑘0
𝑑𝜔(𝑘)
𝜔0′ ≡ |
𝑑𝑘 𝑘=𝑘0
1 ′
Ψ(𝑥, 𝑡) ≅ ∫ 𝑑𝑘 𝜙(𝑘, 𝑘0 ) 𝑒 𝑖(𝑘𝑥−[𝜔0 +𝜔0 (𝑘− 𝑘0 )]𝑡) 𝑒 −𝑖𝑘0 𝑥 𝑒 𝑖𝑘0 𝑥
√2𝜋

∞
1 𝑑𝜔𝑘
Ψ(𝑥, 𝑡) ≅ 𝑒
⏟𝑖(𝑘0 𝑥−𝜔0 𝑡)
∫ 𝜙(𝑘, 𝑘0 ) 𝑒 𝑖(𝑘−𝑘0 )(𝑥− 𝑑𝑘 𝑡)
𝑑𝑘
√2𝜋 𝑎 𝑤𝑎𝑣𝑒 𝑜𝑓 ⏟
−∞
𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦
𝜔0 𝑤𝑎𝑣𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 ⏟ 𝐴 𝑚𝑜𝑑𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑓𝑎𝑐𝑡𝑜𝑟
⏟ 𝑘0 𝑑𝜔
𝐹(𝑥− 𝑑𝑘 𝑡)
𝜔
𝑓(𝑥− 𝑘 0 𝑡)
0
𝜔0 𝑑𝜔
Ψ(𝑥, 𝑡) ≅ 𝑓 (𝑥 − 𝑡) 𝐹 (𝑥 − 𝑡)
𝑘0 𝑑𝑘
Apart from the phase factor, which won’t affect |Ψ|2 , the wave packet moves along
𝑑𝜔𝑘
at a speed
𝑑𝑘

𝑑𝜔(𝑘)
𝑣𝑔𝑟𝑜𝑢𝑝 = 𝜔0′ = |
𝑑𝑘 𝑘=𝑘0
The ordinary phase velocity is
𝜔0 ℏ𝑘 2
𝑣𝑝ℎ𝑎𝑠𝑒 = ; 𝜔=
𝑘0 2𝑚
 12

Thus
ℏ𝑘0
𝑣𝑝ℎ𝑎𝑠𝑒 =
2𝑚
And
𝑑𝜔(𝑘) ℏ𝑘0
𝑣𝑔𝑟𝑜𝑢𝑝 = | = = 2𝑣𝑝ℎ𝑎𝑠𝑒
𝑑𝑘 𝑘=𝑘0 𝑚

This confirms that it is the group velocity of the wave packet , not the phase
velocity of the stationary states that matches the classical velocity:,

𝑣𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 = 𝑣𝑔𝑟𝑜𝑢𝑝 = 2𝑣𝑝ℎ𝑎𝑠𝑒

 13

Example 2.6: A free particle which is initially localized in the range −𝑎 < 𝑥 < 𝑎 is
released at t=0
𝐴 −𝑎 < 𝑥 < 𝑎
𝚿(𝒙, 𝟎) = {
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

where A and a are positive constants. Find Ψ(𝑥, 𝑡).

Solution:
Normalization:
∞ 𝑎
1
∫ |Ψ(𝑥, 0)|2 𝑑𝑥 = |A|2 ∫ 𝑑𝑥 = 2𝑎|A|2 = 1 => 𝐴 =
√2𝑎
−∞ −𝑎
We have
1 1 ℏ𝑘 2
𝑖(𝑘𝑥− 2𝑚 𝑡)
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)Ψk (𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 (∗)
√2𝜋 √2𝜋
1 1
𝚿(𝒙, 𝟎) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 𝑖𝑘𝑥 => 𝜙(𝑘) = ∫ 𝑑𝑥 𝚿(𝒙, 𝟎)𝑒 −𝑖𝑘𝑥
√2𝜋 √2𝜋

Next, find 𝜙(𝑘).

𝑎 𝑎
1 1−𝑖𝑘𝑥
𝑒 −𝑖𝑘𝑥 1 1 𝑒 𝑖𝑘𝑎 − 𝑒 −𝑖𝑘𝑎
𝜙(𝑘) = ∫𝑒 𝑑𝑥 = | = ( )
√2𝜋 √2𝑎 2 √𝑎𝜋 −𝑖𝑘 −𝑎 𝑘 √ 𝑎𝜋 2𝑖
−𝑎
𝑎 sin(𝑘𝑎)
𝜙(𝑘) = √ ( )
𝜋 (𝑘𝑎)
Finally, plug this into (*)
1 ℏ𝑘 2
𝑖(𝑘𝑥− 𝑡)
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 2𝑚
√2𝜋
∞ 2
1𝑎 sin(𝑘𝑎) 𝑖(𝑘𝑥−ℏ𝑘
2𝑚 𝑡) 𝑑𝑘
Ψ(𝑥, 𝑡) = √ ∫( )𝑒
√2𝜋 𝜋 (𝑘𝑎)
−∞
Unfortunately, this integral cannot be solved in terms of elementary functions. But
it is instructive to see the limiting cases:
 14

1. a is very small. Then the starting wave function is a localized spike

𝑎 sin(𝑘𝑎) 𝑎
sin(𝑘𝑎) ≈ 𝑘𝑎 => 𝜙(𝑘) = √ ( ) ≈ √ = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
𝜋 (𝑘𝑎) 𝜋

This is an example of uncertainty principle:

If the spread in position (~𝒂) is small, the spread in momentum must be large.

2. a is large. The spread in the position is broad (a) and from

𝑎 sin(𝑘𝑎)
𝜙(𝑘) = √ ( )
𝜋 (𝑘𝑎)
sin(𝑧)
𝑀𝑎𝑥 [ ]| =1
(𝑧) 𝑧=0
sin(𝑧)
=0 𝑎𝑡 𝑘𝑎 = ±𝜋
(𝑧)

Therefore, for large a, 𝜙(𝑘) is a sharp spike about k=0 (b). This time it’s got a well
defined momentum but an ill-defined position.
 15

Bound States and Scattering States:

We have seen two very different kinds of solutions to the time-independent

Schrödinger equation:
1. Infinite square well and harmonic oscillator:
The stationary states are normalizable, they are labelled by a discrete index n:
𝑖𝐸 𝑡 2 𝑛𝜋 ℏ𝜋2 𝑛2
− 𝑛 −𝑖 𝑡
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 ) 𝑒 ℏ = √ sin ( 𝑥) 𝑒 2𝑚𝐿2
𝐿 𝐿

𝑚𝜔 1/4 1 𝑚𝜔 1/2 𝑚𝜔 2 −𝑖𝐸𝑛 𝑡

Ψ𝑛 (𝑥, 𝑡) = ( ) 𝐻𝑛 (( ) ) 𝑒𝑥𝑝 [− 𝑥 ]𝑒 ℏ
𝜋ℏ 𝑛
√2 𝑛! ℏ 2ℏ
These solutions represent physically realizable states.
The general solution to the time dependent Schrödinger equation is a linear
combination of stationary states:
∞ ∞
𝑖𝐸𝑛 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ
𝑛=1 𝑛=1
2. The free particle: The stationary states are not normalizable, they are labelled by
continuous index k
ℏ𝑘 2
𝑖(𝑘𝑥− 2𝑚 𝑡)
Ψk (𝑥, 𝑡) = 𝐴 𝑒
Free particle stationary states do not represent a physically realizable solution.
The general solution to the time dependent Schrödinger equation is a superposition
of stationary state solutions:
1 1 ℏ𝑘 2
𝑖(𝑘𝑥− 𝑡)
Ψ(𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)Ψk (𝑥, 𝑡) = ∫ 𝑑𝑘 𝜙(𝑘)𝑒 2𝑚
√2𝜋 √2𝜋

What is the physical significance of this distinction?

The two different kinds of solutions to the time-independent Schrödinger equation
correspond precisely to two different kinds of motions and states:
 16

In classical mechanics:
In Fig. (a), let [𝑎, 𝑏] represents two classical turning
points.
 If

𝑉 (𝑥 ) ≤ 𝐸 𝑓𝑜𝑟 𝑥 ∈ [𝑎, 𝑏]
𝑉 (𝑥 ) > 𝐸 𝑓𝑜𝑟 𝑥 ∉ [𝑎, 𝑏]
then the system is in a bound state for 𝑥 ∈ [𝑎, 𝑏]

The particle is stuck in the potential well; it rocks back and forth between two
turning points.
 If, on one side or everywhere (b)
𝑉(𝑥 ) ≤ 𝐸
The particle comes in from infinity,
slows down or speeds up due to V(x).
It can’t get trapped in the potential, but
returns to infinity. This is called
a scattering state.

In quantum domain, the only thing that matters is the potential at infinity (because
of the phenomenon of tunneling that allows the particle to leak through any finite
potential barrier)
 17

In reality, the most potentials go to zero at infinity and this further simplifies
criterion:
 18

Delta-Function Potential
Consider a potential of the form
𝑉 (𝑥 ) = −𝛼 𝛿(𝑥)
where 𝛼 is a positive constant.
Here,
∞ 𝑓𝑜𝑟 𝑥=0
𝛿 (𝑥 ) = {
0 𝑓𝑜𝑟 𝑥≠0
and
∞
∫ 𝛿 (𝑥 ) 𝑑𝑥 = 1
−∞

This is an artificial potential, but it is very simple to work with and illuminates the
basic theory with a minimum mathematics.
Solution:
The Schrödinger equation reads
ℏ2 𝑑2 𝜓
− + 𝑉(𝑥)𝜓 = 𝐸𝜓
2𝑚 𝑑𝑥 2

1. 𝐸 < 0 (𝑏𝑜𝑢𝑛𝑑 𝑠𝑡𝑎𝑡𝑒𝑠). Let 𝐸 = −|𝐸 |

I.region 𝑥 < 0 ≔> 𝑉 (𝑥 ) = 0
𝑑2 𝜓𝐼 2𝑚
− 2 |𝐸 | 𝜓𝐼 = 0 => 𝜓𝐼 (𝑥 ) = 𝐴𝑒 −𝜅𝑥 + 𝐵𝑒 𝜅𝑥
𝑑𝑥 2 ⏟
ℏ
≡𝐾 2 >0

But the
lim 𝐴𝑒 −𝜅𝑥 → ∞
𝑥→−∞

Therefore, we must choose A=0:

𝜓𝐼 (𝑥 ) = 𝐵𝑒 𝜅𝑥
 19

II.region 𝑥 > 0 ≔> 𝑉 (𝑥 ) = 0

𝑑2 𝜓𝐼𝐼 2𝑚
− 2 |𝐸 | 𝜓𝐼𝐼 = 0 => 𝜓𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 + 𝐺𝑒 𝜅𝑥
𝑑𝑥 2 ⏟
ℏ
≡𝐾2 >0

But the
lim 𝐺𝑒 𝜅𝑥 → ∞
𝑥→∞

Therefore, we must choose G=0:

𝜓𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥
In summary:
𝜓𝐼 (𝑥 ) = 𝐵𝑒 𝜅𝑥 𝑥 < 0
𝜓(𝑥 ) = {
𝜓𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 𝑥 > 0
 20

Boundary conditions: The standard boundary conditions for 𝜓(𝑥 ) are:

1. The continuity of 𝝍(𝒙) at x=0, says

𝜓𝐼 (𝑥 = 0) = 𝜓𝐼𝐼 (𝑥 = 0) => 𝐵𝑒 𝜅0 = 𝐹𝑒 −𝜅0 => 𝐵 = 𝐹
Therefore
𝜓𝐼 (𝑥 ) = 𝐵𝑒 𝜅𝑥 𝑥 < 0
𝜓(𝑥 ) = {
𝜓𝐼𝐼 (𝑥 ) = 𝐵𝑒 −𝜅𝑥 𝑥 > 0

2. The continuity of 𝒅𝝍(𝒙)/𝒅𝒙 at x=0,

𝑑𝜓(𝑥)
at x=0 is discontinuous and the second boundary condition says nothing:
𝑑𝑥
this is the case where V is infinite at the boundary (like the infinite well).
Here, we have relax one of our mathematical expectations just as we relaxed
the square integrability condition for the plane wave.
 21

The energy can be obtained from the discontinuity condition of the first
derivative of the wave function. It can be obtained by integrating the
Schrödinger eq,
ℏ2 𝑑2 𝜓
− + 𝑉(𝑥)𝜓 = 𝐸𝜓
2𝑚 𝑑𝑥 2
from −𝜖 to +𝜖, and then take the limit 𝜖 → 0: (𝜖 > 0, but arbitrarily small
which also will enable us to evaluate the discontinuity)

+𝜖 +𝜖 +𝝐
2 2
ℏ 𝑑 𝜓
− ∫ 𝑑𝑥 + ∫ 𝑉(𝑥)𝜓(𝑥) 𝑑𝑥 = 𝐸 ∫ 𝝍(𝒙) 𝒅𝒙
2𝑚 𝑑𝑥 2
−𝜖 −𝜖 −𝝐

The 1.st term:

+𝜖
𝑑2 𝜓 𝑑𝜓 𝑑𝜓 𝑑𝜓
∫ 𝑑𝑥 = | − | ≡ Δ ( )
𝑑𝑥 2 𝑑𝑥 +𝜖 𝑑𝑥 −𝜖 𝑑𝑥
−𝜖

RHS:
+𝜖

lim ∫ 𝜓(𝑥) 𝑑𝑥 = 0
𝜖→0
−𝜖

Since it is the area of a sliver with vanishing width and finite height.
Thus,
+𝜖
𝑑𝜓 𝑑𝜓 𝑑𝜓 2𝑚
| − | ≡ Δ ( ) = 2 lim ∫ 𝑉(𝑥)𝜓(𝑥) 𝑑𝑥
𝑑𝑥 +𝜖 𝑑𝑥 −𝜖 𝑑𝑥 ℏ 𝜖→0
−𝜖

Now, since 𝑉(𝑥 ) = −𝛼𝛿(𝑥), we have

+𝜖
𝑑𝜓 𝒅𝝍 𝒅𝝍 2𝑚𝛼 2𝑚𝛼
Δ( ) = | − | = − 2 lim ∫ 𝛿 (𝑥 )𝜓(𝑥 ) 𝑑𝑥 = − 2 𝜓(0)
𝑑𝑥 𝒅𝒙 +𝝐 𝒅𝒙 −𝝐 ℏ 𝜖→0 ℏ
−𝜖
2𝑚𝛼
=− 𝐵
ℏ2
 22

𝒅𝝍 𝒅𝝍 2𝑚𝛼
| − | = − 2 𝐵 (∗)
𝒅𝒙 +𝝐 𝒅𝒙 −𝝐 ℏ

𝑑𝜓 𝑑𝜓𝐼𝐼 𝑑𝜓
When 𝑥 > 0, = = −𝐵𝐾𝑒 −𝜅𝑥 => | = −𝐵𝐾
𝑑𝑥 𝑑𝑥 𝑑𝑥 +𝜖→0
𝑑𝜓 𝑑𝜓𝐼 𝑑𝜓
When 𝑥 < 0, = = 𝐵𝐾𝑒 𝜅𝑥 => | = +𝐵𝐾
𝑑𝑥 𝑑𝑥 𝑑𝑥 −𝜖→0
(∗)
𝑑𝜓 𝑑𝜓 2𝑚𝛼 𝑚𝛼
| − | = −2𝐾𝐵 =⏞ − 2 𝐵 => 𝐾 = 2
𝑑𝑥 +𝜖 𝑑𝑥 −𝜖 ℏ ℏ
The allowed energy is

ℏ2 𝐾 2 𝑚𝛼 2
𝐸=− => 𝐸 = − 2
2𝑚 2ℏ
There is, therefore, only one bound state solution. As for the excited states, all of
them are unbound.

Normalize 𝝍:

+∞ 0 ∞ ∞

1 = ∫ |𝜓(𝑥)|2 𝑑𝑥 = 𝐵2 ∫|𝑒 𝜅𝑥 |2 𝑑𝑥 + 𝐵2 ∫ |𝑒 −𝜅𝑥 |2 𝑑𝑥 = 2𝐵 2 ∫ 𝑒 −2𝜅𝑥 𝑑𝑥

−∞ −∞ 0 0

−1
2 √𝑚𝛼
= 2𝐵 ( )
0 − 1 => 𝐵 = √𝐾 =
2𝐾 ℏ

√𝑚𝛼 𝑚𝛼 𝑥
𝜓𝐼 (𝑥 ) = 𝑒 ℏ2 𝑥<0
𝜓 (𝑥 ) = ℏ ;
√𝑚𝛼 −𝑚𝛼 𝑥
𝜓𝐼𝐼 (𝑥 ) = 𝑒 ℏ2 𝑥>0
{ ℏ
𝑚𝛼 2
𝐸 = − 2 ∶ the energy of the single bound state
2ℏ
 23

2. 𝐸 > 0 (scattering states).

I.region 𝑥 < 0 ≔> 𝑉 (𝑥 ) = 0
𝑑2 𝜓𝐼 2𝑚
+ 2 𝐸 𝜓𝐼 = 0 => 𝜓𝐼 (𝑥 ) = 𝑨𝒆𝒊𝒌𝒙 + 𝑩𝒆−𝒊𝒌𝒙 (∗)
𝑑𝑥 2 ⏟
ℏ
≡𝑘 2 >0

II.region 𝑥 > 0 ≔> 𝑉 (𝑥 ) = 0

𝑑2 𝜓𝐼𝐼 2𝑚
+ 2 𝐸 𝜓𝐼𝐼 = 0 => 𝜓𝐼𝐼 (𝑥 ) = 𝑭𝒆𝒊𝒌𝒙 + 𝑮𝒆−𝒊𝒌𝒙 (∗∗)
𝑑𝑥 2 ⏟
ℏ
≡𝑘 2 >0

Note that:
𝒊𝑬 𝒊𝑬
𝒊𝒌(𝒙− 𝒕) 𝒊𝒌(𝒙− 𝒕)
𝑨𝒆 𝒌ℏ and 𝑭𝒆 𝒌ℏ represent the wave of amplitude 𝐴 and 𝐹 incident from

the left, propagating to the right.

𝒊𝑬 𝒊𝑬
𝒊𝒌(𝒙+ 𝒕) 𝒊𝒌(𝒙+ 𝒕)
𝑩𝒆 𝒌ℏ and 𝑮𝒆 𝒌ℏ represent the wave of amplitude 𝐵 and G incident

from the right propagating to the left.

In a typical scattering experiment

particles are fired in from one direction,
say from the left. In this case, the
amplitude of a wave coming in from the
right will be zero:
G=0
Hence, we can write

𝜓𝑖 (𝑥 ) ≡ 𝐴𝑒 𝑖𝑘𝑥 ∶ incoming wave of amplitude A

𝜓𝑟 (𝑥 ) = 𝐵𝑒 −𝑖𝑘𝑥 ∶ reflected wave of amplitude B
𝜓𝑡 (𝑥 ) = 𝐹𝑒 𝑖𝑘𝑥 : transmitted wave of amplitude F
 24

Boundary conditions:
1. The continuity of 𝝍(𝒙) at x=0, says

𝜓𝐼 (𝑥 = 0) = 𝜓𝐼𝐼 (𝑥 = 0) => 𝐴 + 𝐵 = 𝐹 (𝑎)

2. The derivatives are

𝒅𝝍 𝒅𝝍 2𝑚𝛼
| − | = − 2 𝐵 (∗)
𝒅𝒙 +𝝐 𝒅𝒙 −𝝐 ℏ

𝑑𝜓 𝑑𝜓𝐼𝐼 𝑑𝜓
When 𝑥 > 0, = = 𝑖𝑘𝐹𝑒 𝑖𝑘𝑥 => | = 𝑖𝑘𝐹
𝑑𝑥 𝑑𝑥 𝑑𝑥 +𝜖→0
𝑑𝜓 𝑑𝜓𝐼 𝑑𝜓
When 𝑥 < 0, = = 𝑖𝑘(𝐴𝑒 𝑖𝑘𝑥 − 𝐵𝑒 −𝑖𝑘𝑥 ) => | = 𝑖𝑘(𝐴 − 𝐵)
𝑑𝑥 𝑑𝑥 𝑑𝑥 −𝜖→0
(∗)
𝑑𝜓 𝑑𝜓 2𝑚𝛼
| − | = 𝑖𝑘 (𝐹 − 𝐴 + 𝐵) =⏞− 2 𝜓
⏟ (0) =>
𝑑𝑥 +𝜖 𝑑𝑥 −𝜖 ℏ
𝐴+𝐵 𝑂𝑅 𝐹
2𝑚𝛼 2𝑚𝛼
𝑖𝑘(𝐹 − 𝐴 + 𝐵) = − (𝐴 + 𝐵 ) => 𝐹 = 𝑖 (𝐴 + 𝐵 ) + 𝐴 − 𝐵
ℏ2 𝑘ℏ2
𝑚𝛼
𝛽≡
𝑘ℏ2

𝐹 = 𝐴(1 + 2𝑖𝛽 ) − 𝐵(1 − 2𝑖𝛽) (𝑏)

Solve (a) and (b) for F and B:

𝑖𝛽 1
𝐵= 𝐴 ;𝐹= 𝐴
1 − 𝑖𝛽 1 − 𝑖𝛽
 25

The probability of finding the particle at a specified location is |𝜓|2 . However, this
is not a normalizable wave function, so the absolute probability of finding the
particle at a particular location is not well defined; nevertheless, the ratio of
probabilities for the incident and reflected waves is meaningful:

Introduce

𝑹: reflection coefficient: Probability that incident particle is reflected

𝑻: Transmission coefficient: Probability that incident particle is transmitted
𝑭𝟐 1 2 1
𝑻≡| | =| | = => ,
𝑨 1 − 𝑖𝛽 1 + 𝛽2

𝑩𝟐 𝑖𝛽 2 𝛽2 1
𝑹≡| | =| | = =
𝑨 1 − 𝑖𝛽 1 + 𝛽 2 1 + 1/𝛽 2

Now,

𝑚𝛼 2𝑚 𝑚𝛼
𝛽≡ , 𝑘 = √ 𝐸 => 𝛽 ≡
𝑘ℏ2 ℏ2 ℏ√2𝑚𝐸

1 1
𝑇= , 𝑅 =
1 + 𝑚𝛼 2 /2𝐸ℏ2 1 + 2𝐸ℏ2 /𝑚𝛼 2

The higher the energy, the greater the probability of transmission.

The sum of these probabilities should be 1: 𝑇 + 𝑅 = 1

 1

THE FINITE SQUARE WELL

Consider the finite square well potential

−𝑉0 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎

𝑉 (𝑥 ) = {
0 𝑓𝑜𝑟 |𝑥 | > 𝑎

where 𝑉0 is a positive constant.

The bound states: E<0 : Let 𝐸 = −|𝐸 |

Schrödinger eq:

𝑑2 𝜓 2𝑚
+ (𝐸 − 𝑉 (𝑥 ))𝜓 = 0
𝑑𝑥 2 ℏ2

In the region I: x<-a: V(x)=0

𝑑2 𝜓𝐼 2𝑚 2
2𝑚
− |𝐸 | 𝜓 𝐼 = 0; 𝜅 ≡ |𝐸 | > 0
𝑑𝑥 2 ⏟2
ℏ ℏ2
𝜅2 >0

The general solution

𝜓𝐼 (𝑥 ) = 𝐴𝑒 −𝜅𝑥 + 𝐵𝑒 𝜅𝑥
But the 1.st term blows up as 𝑥 → −∞:

lim 𝐴𝑒 −𝜅𝑥 → ∞ => 𝐴 = 0

𝑥→−∞

Thus the general solution is

𝜓𝐼 (𝑥 ) = 𝐵𝑒 𝜅𝑥 , 𝑓𝑜𝑟 𝑥 < −𝑎
 2

In the region II: −𝑎 < 𝑥 < 𝑎, 𝑉(𝑥) = −|𝑉0 |

𝑑2 𝜓𝐼𝐼 2𝑚 2
2𝑚
+ (−|𝐸 | + |𝑉0 |) 𝜓 𝐼𝐼 = 0 , ℓ ≡ (−|𝐸 | + |𝑉0 |) > 0
𝑑𝑥 2 ⏟
ℏ2 ℏ2
ℓ2

The general solution is

𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 + 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎

In the region III: x>a: V(x)=0

𝑑2 𝜓𝐼𝐼𝐼 2𝑚 2
2𝑚
− |𝐸 | 𝜓 𝐼𝐼𝐼 = 0; 𝜅 ≡ |𝐸 | > 0
𝑑𝑥 2 ⏟2
ℏ ℏ2
𝜅2

The general solution

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 + 𝐺𝑒 𝜅𝑥
But the 2.nd term blows up as 𝑥 → ∞.

lim 𝐺𝑒 𝜅𝑥 → ∞ => 𝐺 = 0
𝑥→∞

Thus the general solution is

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 , 𝑓𝑜𝑟 𝑥 > 𝑎
Summary:

𝜓𝐼 (𝑥 ) = 𝐵𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎
𝜓(𝑥 ) = {𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 + 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎
 3

The boundary conditions:

Note that this potential is an even function;

𝑉 (−𝑥 ) = 𝑉(𝑥)
Therefore the Hamiltonian is also even function. Thus, its eigenfunctions are the
eigenfunctions of the parity operator and therefore they have definite parities: the
solutions are then either antisymmetric (odd) or symmetric (even).
𝑃̂𝜓(𝑥 ) = 𝜓(−𝑥 ) = ±𝜓(𝑥)
The advantage of this is that we need only impose the boundary conditions on one
side, say at +a; the other side is the automatic,.

Note that:
BC at 𝑥 = −𝑎 gives
𝐵𝐾𝑒 −𝜅𝑎 = 𝐶𝑙 cos(−ℓ𝑎) − 𝑙𝐷 sin(−ℓ𝑎)
BC at 𝑥 = 𝑎 gives
−𝐹𝐾𝑒 −𝜅𝑎 = 𝐶𝑙 cos(ℓ𝑎) − 𝑙𝐷 sin(ℓ𝑎)
Divide them
𝐵 𝐶𝑙 cos(ℓ𝑎) + 𝑙𝐷 sin(ℓ𝑎)
− =
𝐹 𝐶𝑙 cos(ℓ𝑎) − 𝑙𝐷 sin(ℓ𝑎)
For even solutions, where 𝐶 = 0 => 𝐵 = 𝐹
𝜓𝐼 (𝑥 ) = 𝐹𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎
𝜓𝑒𝑣𝑒𝑛 (𝑥 ) = { 𝜓𝐼𝐼 (𝑥 ) = 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎
For odd solutions, where 𝐷 = 0 => 𝐵 = −𝐹
𝜓𝐼 (𝑥 ) = −𝐹𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎
𝜓𝑜𝑑𝑑 (𝑥 ) = { 𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎
 4

It is sufficient to match the solutions 𝝍𝒆𝒗𝒆𝒏 and 𝝍𝒐𝒅𝒅 only at 𝒙 = 𝒂 since the
same conditions will obtain when matching is done at 𝒙 = −𝒂. (Due to the
symmetry under reflection)

Lets work out the even solutions

The continuity of 𝜓(𝑥 ) at 𝑥 = 𝑎 says
𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) = 𝜓𝐼𝐼 (𝑥 = 𝑎) => 𝐹𝑒 −𝜅𝑎 = 𝐷 cos ℓ𝑎 (𝑎)
𝑑𝜓(𝑥)
The continuity of at 𝑥 = 𝑎 says
𝑑𝑥

𝑑𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) 𝑑𝜓𝐼𝐼 (𝑥 = 𝑎)
= => −𝜅𝐹𝑒 −𝜅𝑎 = −ℓ𝐷 sin ℓ𝑎 (𝑏)
𝑑𝑥 𝑑𝑥

Dividing eq.(b) by eq.(a)

−𝜅𝐹𝑒 −𝜅𝑎 −ℓ𝐷 sin ℓ𝑎
= => 𝜅 = ℓ tan ℓ𝑎
𝐹𝑒 −𝜅𝑎 𝐷 cos ℓ𝑎
OR
𝜅𝑎 = ℓ𝑎 tan ℓ𝑎 (∗)

Introduce the notation:

2𝑚𝑎2 2𝑚𝑎2 2𝑚𝑎2

𝒛 ≡ ℓ𝑎 = √ (|𝑉0 | − |𝐸 |) = √ |𝑉0 | − |𝐸 | ,
ℏ2 ℏ2 ℏ2

Now define

2𝑚𝑎2
𝑧0 ≡ √ 2 |𝑉0 |
ℏ

Since

2
2𝑚𝑎2
(𝜅𝑎) = |𝐸 |
ℏ2
 5

Then

2𝑚𝑎2 2𝑚𝑎2
𝒛 ≡ ℓ𝑎 = √ 2 |𝑉0 | − |𝐸 | => 𝑧 = √𝑧02 − (𝜅𝑎)2
⏟ℏ ⏟ℏ2
𝑧02 (𝜅𝑎)2

𝜿𝒂 = √𝑧02 − 𝑧 2

So, from (*)

(𝜅𝑎)
𝜅𝑎 = ℓ𝑎 tan ℓ𝑎 => 𝑡𝑎𝑛 ℓ𝑎 =
(ℓ𝑎)
OR

√𝑧02 − 𝑧 2 𝑧0 2
𝑡𝑎𝑛 𝑧 = √
= ( ) −1
𝑧 𝑧

Let

𝑧0 2 𝑧0 2
𝑦(𝑧) ≡ √( ) − 1 . Note that 𝑦(𝑧0 ) = √( ) − 1 = 0
𝑧 𝑧0

𝑦(𝑧) = 𝑡𝑎𝑛 𝑧
This is a transcendental eq. for z as a function of 𝑧0 . We can plot both sides of
this equation on the same graph for various values of to get an idea of
what happens.
 6

z0 8
10

6
y z tanz
4

0
0 2 4 6 8 10

z0 50
40 z0 50
40

30 30
y z

20
20

10
10
tanz
0 3 5 7 9 11
0 2 2
2 2
3 2
4 2
5 2
6
0
0 10 20 30 40 50
 7

In these plots, we show what happens for three different values of 𝑧0 . The blue

𝑧 2
curves show the plot of 𝑡𝑎𝑛 𝑧 ; the red curves that of 𝑦(𝑧) ≡ √( 0 ) − 1.
𝑧

1. In the first graph, with 𝑧0 = 2, we get only one intersection between the two
plots, around 𝑧 = 1 . Thus for 𝑧0 = 2, there is only one bound state, with an
energy that can be worked out from

1 2
𝑎 1 ℏ⏞
𝑧
𝒛 ≡ ℓ𝑎 = √2𝑚(−|𝐸| + |𝑉0 |) ≈ 1 => |𝐸 | ≈ |𝑉0 | − ( )
ℏ 2𝑚 𝑎

2. In the second graph, with 𝑧0 = 8, we get three intersections between the two
plots, around 𝑧1 = 1.40, 𝑧2 = 4.15, 𝑧3 = 6.80 . Thus for 𝑧0 = 8, there are three
bound states, with an energy that can be worked out from
𝑎
𝑧 ≡ ℓ𝑎 = √2𝑚(−|𝐸| + |𝑉0 |) ≈ 𝑧𝑛 𝑓𝑜𝑟 𝑛 = 1,2,3
ℏ
1 ℏ𝑧𝑛 2
|𝐸𝑛 | ≈ |𝑉0 | − ( ) ; 𝑧1 = 1.40, 𝑧2 = 4.15, 𝑧3 = 6.80
2𝑚 𝑎

3. In the 3.rd graph, with 𝑧0 = 50, we get 16 intersections between the two plots,
 8

 We can see that no matter how small we make 𝑧0 , we will always have
at least one bound state (since the 𝑡𝑎𝑛 𝑧 graph starts off from the origin).
 Note that
 When 0 ≤ 𝑧0 < 𝜋 => there is 1 bound state

In the 1.st graph: 0 ≤ (𝑧0 = 2 ≅ 0.64𝜋) < 𝜋

 When 𝜋 ≤ 𝑧0 < 2𝜋 => there are 2 bound states

 When 2𝜋 ≤ 𝑧0 < 3𝜋 => there are 3 bound states

In the 2.nd graph: 2𝜋 ≤ (𝑧0 = 8 ≅ 2.55𝜋) < 3𝜋

 When 𝑛𝜋 ≤ 𝑧0 < (𝑛 + 1)𝜋 => there are (n+1) bound states

In the 3.rd graph: 15𝜋 ≤ (𝑧0 = 50 ≅ 15.9𝜋) < 16𝜋

There are 16 bound states for 𝑧0 = 50.

 As |𝑉0 | → ∞, we would expect to get the infinite square well states.

To see this, note that
𝑧 2
The graph of 𝑦(𝑧) ≡ √( 0 ) − 1 intersects the horizontal axis at
𝑧
𝑧 = 𝑧0 , so as |𝑉0 | → ∞, 𝑧0 → ∞
The intersection point gets further and further along the axis, so the
number of intersections with branches of the tangent gets larger. Thus
the number of energy states gets larger and larger, eventually
becoming infinite. As to the locations of these intersections, we can
𝑧 2
notice that for any fixed, finite value of 𝑧, the quantity √( 0 ) − 1 tends
𝑧
to infinity as 𝑧0 → ∞,
 9

𝑧0 2
lim 𝑦(𝑧) ≡ (√( ) − 1) → ∞
𝑧0 →∞, 𝑧

so that means that the entire curve gets higher, so the intersections
with the tangent curve will occur at higher locations. The tangent is
𝜋
asymptotic to the vertical lines 𝑛 for odd n, so we would expect the
2
intersection points to eventually become
𝜋
𝑧𝑛 = 𝑛 , 𝑤ℎ𝑒𝑟𝑒 𝑛 = 𝑜𝑑𝑑
2
This means that
𝟐
2𝑚 𝜋 2
𝒛 ≡ (ℓ𝑎 )2 2
= 𝑎 ( 2 (−|𝐸 | + |𝑉0 |)) ≈ (𝑛 )
ℏ 2

𝑛2 𝜋 2 ℏ2
|⏟
𝑉0 | − |𝐸𝑛 | =
2𝑚 (2𝑎)2
>0

Since |𝑉0 | − |𝐸𝑛 | is the height of the bound state above the bottom of the well,
we can see that this formula does indeed give us the expected energy levels
for an infinite square well of width 2𝑎, or at least those corresponding to odd
n.
 10

Problem 2.29. Odd Bound states

𝜓𝐼 (𝑥 ) = −𝐹𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎

𝜓𝑜𝑑𝑑 (𝑥 ) = { 𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎
The continuity of 𝜓(𝑥 ) at x=a, says
𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) = 𝜓𝐼𝐼 (𝑥 = 𝑎) => 𝐹𝑒 −𝜅𝑎 = 𝐶 sin ℓ𝑎 (𝑎)
𝑑𝜓(𝑥)
The continuity of at x=a, says
𝑑𝑥

𝑑𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) 𝑑𝜓𝐼𝐼 (𝑥 = 𝑎)
= => −𝜅𝐹𝑒 −𝜅𝑎 = ℓ𝐶 cos ℓ𝑎 (𝑏)
𝑑𝑥 𝑑𝑥

Dividing eq.(b) by eq.(a)

−𝜅𝐹𝑒 −𝜅𝑎 ℓ𝐶𝑐𝑜𝑠ℓ𝑎
= => 𝜅 = −ℓ cot ℓ𝑎
𝐹𝑒 −𝜅𝑎 𝐶𝑠𝑖𝑛ℓ𝑎
OR
𝜅𝑎 = −ℓ𝑎 cot ℓ𝑎
With the same notation as before we get

𝑧0 2
√
−𝑐𝑜𝑡 𝑧 = ( ) − 1
𝑧

In the plots below, we show what happens for three different values of 𝑧0 . The
orange curves show the plot of −𝑐𝑜𝑡 𝑧 ; the green curves that of 𝑦(𝑧) ≡
2
√(𝑧0 ) − 1. (Together with 𝑡𝑎𝑛 𝑧, blue curves)
𝑧
 11

z0=2 z0=8
10 10

8 8

6 6

4 4

2 2

0 z 0
3 3 5
0
2 2 0 2 3
2 2 2

z0=50

40

30

20

10

0
3 5 7 9 11
0 2 3 4 5 6
2 2 2 2 2 2
 12

Note that
𝜋
 When 𝑧0 < No odd bound state; only one even bound state
2

𝜋 𝑎 𝜋 1 ℏ𝜋 2
𝑧0 < => √2𝑚|𝑉0 | < => 𝐼𝑓 |𝑉0 | < ( ) 𝑡ℎ𝑒𝑛 𝑛𝑜 𝑜𝑑𝑑 𝑏𝑜𝑢𝑛𝑑 𝑠𝑡𝑎𝑡𝑒
2 ℏ 2 2𝑚 2𝑎

 When 0 ≤ 𝑧0 < 𝜋/2 => there is 1 bound state (only even)

 When 𝜋/2 ≤ 𝑧0 < 𝜋 =>

there are 2 bound states ( one is even one is odd)

𝜋
In the 1.st graph: ≤ (𝑧0 = 2 ≅ 0.64𝜋) < 𝜋 => 2 𝑏𝑠
2

 When 𝜋 ≤ 𝑧0 < 3𝜋/2 => there are 3 bound states (2 even , one odd)

Generalize this:

𝝅 𝝅
When 𝒏 ≤ 𝒛𝟎 < (𝒏 + 𝟏) => there are (n+1) bound states
𝟐 𝟐

5𝜋
In the 2.nd graph: ≤ (𝑧0 = 8 ≅ 2.55𝜋) < 3𝜋 => there are 6 bound states
2

In the 3.rd graph: 31 𝜋/2 ≤ (𝑧0 = 50 ≅ 15.9 𝜋) < 16𝜋

There are 32 bound states for 𝑧0 = 50.

 13

3 5 7 9 11 13 15 17 19 21 23 25 27 29 31
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
 14

1 ℏ𝑧𝑛 2 10
|𝐸𝑛 | ≈ |𝑉0 | − ( )
2𝑚 𝑎
8
𝜓𝐼 (𝑥) = 𝐹𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎
𝜓𝑒𝑣𝑒𝑛 (𝑥) = {𝜓𝐼𝐼 (𝑥) = 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
6
𝜓𝐼𝐼𝐼 (𝑥) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎

𝜓𝐼 (𝑥) = −𝐹𝑒 𝜅𝑥 𝑓𝑜𝑟 𝑥 < −𝑎 4

𝜓𝑜𝑑𝑑 (𝑥) = {𝜓𝐼𝐼 (𝑥) = 𝐶 sin ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥) = 𝐹𝑒 −𝜅𝑥 𝑓𝑜𝑟 𝑥 > 𝑎 2

0
3 5
0 2 3
2 2 2

Ground State: (Even)

1.st excited state : (Odd)
2.nd Excited state (Even)
 15

Inside the well, in the classically allowed region, we have E>V(x) and the
differential equation admits only sinusoidal solutions (sin ℓ𝑥 or cos ℓ𝑥)
characterized by the wave vector ℓ . The oscillatory part of the wave function
(inside the well) has a characteristic wavelength
2𝜋 2𝜋
𝜆ℓ = =
ℓ 2𝑚
√ (|𝑉0 | − |𝐸 |)
ℏ2
So the larger the energy difference between the eigenvalue and the potential energy,
the smaller the wavelength, which is seen in fig. above)
 16

Outside the well, in the classically forbidden region, we have E<V(x) the
differential equation admits only real exponential solutions:
𝜓𝐼 (𝑥 ) = 𝐹𝑒 𝜅𝑥 𝑎𝑛𝑑 𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥
with a decay length of 1/𝜅, which is zero for for the infinite square well.

In the finite potential well, since the wave function is not zero in the classically
forbidden region, the quantum mechanical particle can be found where the classical
particle cannot:
|𝜓𝐼 (𝑥 )|2 = |𝐹 |2 𝑒 2𝜅𝑥 𝑎𝑛𝑑 |𝜓𝐼𝐼𝐼 (𝑥 )|2 = |𝐹 |2 𝑒 −2𝜅𝑥
This penetration of the wave function into the potential energy barrier leads to the
phenomenon of tunneling, which we explore in the problem 2.32.
The wave function plots in the above figure indicate that the barrier penetration is
more pronounced for higher energy levels and can become large for energies close
to the top of the well. To see this let’s calculate the decay length from uncertainty
relation:
1 1
Δ𝑥Δ𝑝~ℏ => Δ𝑥Δ(ℏ𝜅 )~ℏ => Δ𝑥~ =
𝜅 2𝑚
√ |𝐸 |
ℏ2
Since 𝜅 in the forbidden region decreases as the energy increases, which means that
the decay length Δ𝑥 becomes larger, so more of the wave function is outside the
well.
In comparing the finite and infinite well energies, we also note that a given finite
well energy eigenvalue 𝐸𝑛 lies below the corresponding infinite well energy
eigenvalue. This is in consistent with the longer wavelength of the finite well
eigenstate compared to the corresponding infinite well state.
 17

Problem 2.40
∞ 𝑥<0
𝑉 (𝑥 ) = {−32ℏ2 /𝑚𝑎2 0<𝑥<𝑎
0 𝑥>𝑎

Let, |𝑉0 | ≡ 32ℏ2 /𝑚𝑎2

Let 𝐸 = −|𝐸 |
Schrödinger eq:
𝑑2 𝜓 2𝑚
+ (𝐸 − 𝑉 (𝑥 ))𝜓 = 0
𝑑𝑥 2 ℏ2

In the region I: x<: 𝑉(𝑥 ) = ∞ => 𝜓𝐼 (𝑥 ) = 0

In the region II: 0 < 𝑥 < 𝑎, 𝑉(𝑥) = −|𝑉0 |

𝑑2 𝜓𝐼𝐼 2𝑚 2
2𝑚
+ (−|𝐸 | + |𝑉0 |) 𝜓 𝐼𝐼 = 0 , ℓ ≡ (−|𝐸 | + |𝑉0 |) > 0
𝑑𝑥 2 ⏟
ℏ2 ℏ2
ℓ2

The general solution is

𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 + 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎

In the region III: x>a: V(x)=0

𝑑2 𝜓𝐼𝐼𝐼 2𝑚 2
2𝑚
− |𝐸 | 𝜓 𝐼𝐼𝐼 = 0; 𝜅 ≡ |𝐸 | > 0
𝑑𝑥 2 ⏟2
ℏ ℏ2
𝜅2

The general solution

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 + 𝐺𝑒 𝜅𝑥
But the 2.nd term blows up as 𝑥 → ∞.
 18

lim 𝐺𝑒 𝜅𝑥 → ∞ => 𝐺 = 0
𝑥→∞

Thus the general solution is

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 −𝜅𝑥 , 𝑓𝑜𝑟 𝑥 > 𝑎

0 𝑥<0
𝜓(𝑥 ) = {𝐶𝑠𝑖𝑛ℓ𝑥 + 𝐷 cos ℓ𝑥 0<𝑥<𝑎
𝐹𝑒 −𝜅𝑥 𝑥>𝑎

The next step is to impose the boundary conditions.

The continuity of 𝜓(𝑥 ) at x=0, says

𝜓𝐼 (𝑥 = 0) = 𝜓𝐼𝐼 (𝑥 = 0) => 0 = 𝐷 cos ℓ0 = 𝐷 => 𝐷 = 0
Thus
𝜓𝐼𝐼 (𝑥 ) = 𝐶𝑠𝑖𝑛ℓ𝑥
0 𝑥<0
𝜓(𝑥 ) = {𝐶𝑠𝑖𝑛ℓ𝑥 0<𝑥<𝑎
𝐹𝑒 −𝜅𝑥 𝑥>𝑎

The continuity of 𝜓(𝑥 ) at x=a, says

𝜓𝐼𝐼 (𝑥 = 𝑎) = 𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) => 𝐶𝑠𝑖𝑛ℓ𝑎 = 𝐹𝑒 −𝜅𝑎 (𝑎)
𝑑𝜓(𝑥)
The continuity of at x=a, says
𝑑𝑥

𝑑𝜓𝐼𝐼 (𝑥 = 𝑎) 𝑑𝜓𝐼𝐼𝐼 (𝑥 = 𝑎)
= => ℓ𝐶 cos ℓ𝑎 = −𝜅𝐹𝑒 −𝜅𝑎 (𝑏)
𝑑𝑥 𝑑𝑥

Dividing eq.(b) by eq.(a)

−𝜅𝐹𝑒 −𝜅𝑎 ℓ𝐶 cos ℓ𝑎
= => −𝜅 = ℓ cot ℓ𝑎
𝐹𝑒 −𝜅𝑎 𝐶𝑠𝑖𝑛ℓ𝑎
 19

OR
−𝜅𝑎 = ℓ𝑎 cot ℓ𝑎
Introduce the notation:
𝑎 𝑎
𝒛 ≡ ℓ𝑎 = √2𝑚(|𝑉0 | − |𝐸 |) , 𝑧0 ≡ √2𝑚|𝑉0 |
ℏ ℏ
𝑎
𝜅𝑎 = √2𝑚|𝐸| = √𝑧02 − 𝑧 2
ℏ

So

√𝑧02 − 𝑧 2 𝑧0 2
−𝜅𝑎 = ℓ𝑎 cot ℓ𝑎 => − 𝑐𝑜𝑡 𝑧 = √
= ( ) −1
𝑧 𝑧

Let

𝑧0 2
𝑦(𝑧) ≡ √( ) − 1
𝑧

This is a transcendental eq. for z as a function of 𝑧0 . We can plot both sides of

this equation on the same graph for various values of to get an idea of
what happens.

𝑎 32ℏ2 𝑎 32ℏ2
𝑧0 ≡ √2𝑚|𝑉0 | , |𝑉0 | ≡ => 𝑧0 ≡ √ 2𝑚 => 𝑧0 = 8
ℏ 𝑚𝑎2 ℏ 𝑚𝑎2
 20

10

𝜋
5 ≅ 7.85
8 2
The number of bound states is 3

0 3 5 7
z
0 2 3
2 2 2 2

1 ℏ𝑧𝑛 2 32ℏ2 ℏ2 2
|𝐸𝑛 | ≈ |𝑉0 | − ( ) = − 𝑧
2𝑚 𝑎 𝑚𝑎2 2𝑚𝑎2 𝑛
Let
ℏ2
Δ≡
2𝑚𝑎2
Then
|𝐸𝑛 | = Δ(64 − 𝑧𝑛2 )

𝑛 𝑧𝑛 |𝐸𝑛 |/ Δ
1 2.785 56.2438
2 5.522 33.5075
3 7.9573 0.6813
 21

b)
+∞ +∞ +∞

𝑃(𝑥 > 𝑎) = ∫ |𝜓𝐼𝐼𝐼 (𝑥 )|2 𝑑𝑥 = ∫ |𝐹𝑒 −𝜅𝑥 |2 𝑑𝑥 = |𝐹 |2 ∫ 𝑒 −2𝜅𝑥 𝑑𝑥

𝑎 𝑎 𝑎
−2𝜅𝑥 ∞
𝑒
= |𝐹 |2 |
−2𝜅 𝑎
1 −2𝜅𝑎
𝑃(𝑥 > 𝑎) = |𝐹 |2 𝑒
2𝜅
Normalization=>
𝑎 +∞

1 = ∫|𝜓𝐼𝐼 (𝑥 )|2 𝑑𝑥 + ∫ |𝜓𝐼𝐼𝐼 (𝑥 )|2 𝑑𝑥 =>

0 𝑎
𝑎 +∞

1 = 𝐶 2 ∫ 𝑆𝑖𝑛2 ℓ𝑥 𝑑𝑥 + 𝐹 2 ∫ 𝑒 −2𝜅𝑥 𝑑𝑥 =>

0 𝑎

But continuity at at x=a, says

𝜓𝐼𝐼 (𝑥 = 𝑎) = 𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) => 𝐶𝑠𝑖𝑛ℓ𝑎 = 𝐹𝑒 −𝜅𝑎 => 𝐹 = 𝐶𝑒 𝜅𝑎 𝑠𝑖𝑛ℓ𝑎
𝑎 +∞

1 = 𝐶 2 [∫ 𝑆𝑖𝑛2 ℓ𝑥 𝑑𝑥 + 𝑒 2𝜅𝑎 𝑆𝑖𝑛2 ℓ𝑎 ∫ 𝑒 −2𝜅𝑥 𝑑𝑥] =

0 𝑎
𝑎
2
𝑥 1 2𝜅𝑎 2
1 −2𝜅𝑥 ∞
= 𝐶 [{ − 𝑠𝑖𝑛ℓ𝑥 𝑐𝑜𝑠ℓ𝑥} + 𝑒 𝑆𝑖𝑛 ℓ𝑎 { 𝑒 } ]
2 2ℓ 0 −2𝜅 𝑎

2
𝑎 1 2𝜅𝑎 2
𝑒 −2𝜅𝑎
= 𝐶 [( − 𝑠𝑖𝑛ℓ𝑎 𝑐𝑜𝑠ℓ𝑎) + 𝑒 𝑆𝑖𝑛 ℓ𝑎 ]
2 2ℓ 2𝜅
 22

2
𝑎 1 𝑆𝑖𝑛2 ℓ𝑎
= 𝐶 [( − 𝑠𝑖𝑛ℓ𝑎 𝑐𝑜𝑠ℓ𝑎) + ]
2 2ℓ 2𝜅
𝐶2 𝜅
= [(𝑎 − 𝑠𝑖𝑛ℓ𝑎 𝑐𝑜𝑠ℓ𝑎) + 𝑆𝑖𝑛2 ℓ𝑎]
2𝜅 ⏟
ℓ
−𝐶𝑜𝑡ℓ𝑎

𝐶2 2 2
𝐶2 2𝜅
= [(𝑎𝜅 + 𝑐𝑜𝑠 ℓ𝑎) + 𝑆𝑖𝑛 ℓ𝑎] = (1 + 𝑎𝜅 ) = 1 => 𝐶 = √
2𝜅 2𝜅 (1 + 𝑎𝜅 )

2𝜅
𝐹 = 𝐶𝑒 𝜅𝑎 𝑠𝑖𝑛ℓ𝑎 = √ 𝑒 𝜅𝑎 𝑠𝑖𝑛ℓ𝑎
(1 + 𝑎𝜅 )

1 −2𝜅𝑎 2𝜅 1
𝑃(𝑥 > 𝑎) = |𝐹 |2 𝑒 = 𝑒 2𝜅𝑎 𝑠𝑖𝑛2 ℓ𝑎 𝑒 −2𝜅𝑎
2𝜅 (1 + 𝑎𝜅 ) 2𝜅

𝑠𝑖𝑛2 ℓ𝑎
𝑃(𝑥 > 𝑎) =
(1 + 𝑎𝜅 )

The highest-energy bound state is 𝐸3 with 𝑧3 ≅ 7.9573. (Mathematica gives)

Z0=8=> P=0.542
 23

The scattering states: E>0

In the region I: x<-a: V(x)=0

𝑑2 𝜓𝐼 2𝑚 2
2𝑚
+ 𝐸 𝜓 𝐼 = 0; 𝑘 ≡ 𝐸>0
𝑑𝑥 2 ⏟ ℏ2 ℏ2
𝑘2

The general solution

𝜓𝐼 (𝑥 ) = 𝐴𝑒 𝑖𝑘𝑥 + 𝐵𝑒 −𝑖𝑘𝑥 , 𝑓𝑜𝑟 𝑥 < −𝑎

In the region II: −𝑎 < 𝑥 < 𝑎, 𝑉 = −|𝑉0 |

𝑑2 𝜓𝐼𝐼 2𝑚 2
2𝑚
+ (𝐸 + |𝑉0 |) 𝜓 𝐼𝐼 = 0 , ℓ ≡ (𝐸 + |𝑉0 |) > 0
𝑑𝑥 2 ⏟
ℏ2 ℏ2
ℓ2

𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 + 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎

In the region III: x>a: V(x)=0

𝑑2 𝜓𝐼𝐼𝐼 2𝑚 2
2𝑚
+ |𝐸 | 𝜓 𝐼𝐼𝐼 = 0; 𝑘 ≡ |𝐸 | > 0
𝑑𝑥 2 ⏟2
ℏ ℏ2
𝑘2

The general solution

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 𝑖𝑘𝑥 + 𝐺𝑒 −𝑖𝑘𝑥
But to the right there is no incoming wave in this region. Thus the general solution
is

𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 𝑖𝑘𝑥 , 𝑓𝑜𝑟 𝑥 > 𝑎

 24

𝜓𝐼 (𝑥 ) = 𝐴𝑒 𝑖𝑘𝑥 + 𝐵𝑒 −𝑖𝑘𝑥 𝑓𝑜𝑟 𝑥 < −𝑎

𝜓(𝑥 ) = {𝜓𝐼𝐼 (𝑥 ) = 𝐶 sin ℓ𝑥 + 𝐷 cos ℓ𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 𝑖𝑘𝑥 𝑓𝑜𝑟 𝑥>𝑎
Here
𝐴: the incident amplitude
𝐵: the reflected amplitude
𝐹: the transmitted amplitude

The transmission and reflection coefficients are defined as

𝑇: Probability that incident particle is transmitted
𝑅: Probability that incident particle is reflected.

𝐹2 𝐵2
𝑇≡| | , 𝑅≡| |
𝐴 𝐴
----------------------------------------------------------------------
 25

The continuity of 𝝍(𝒙) at x=-a, says

𝜓𝐼 (𝑥 = −𝑎) = 𝜓𝐼𝐼 (𝑥 = −𝑎) =>
𝐴𝑒 −𝑖𝑘𝑎 + 𝐵𝑒 𝑖𝑘𝑎 = 𝐶 sin(−ℓ𝑎) + 𝐷 cos(−ℓ𝑎)

𝐴𝑒 −𝑖𝑘𝑎 + 𝐵𝑒 𝑖𝑘𝑎 = −𝐶 sin(ℓ𝑎) + 𝐷 cos(ℓ𝑎) (1)

𝑑𝜓(𝑥)
The continuity of at x=-a, says
𝑑𝑥

𝑑𝜓𝐼 (𝑥 = −𝑎) 𝑑𝜓𝐼𝐼 (𝑥 = −𝑎)

= =>
𝑑𝑥 𝑑𝑥
𝑖𝑘[𝐴𝑒 −𝑖𝑘𝑎 − 𝐵𝑒 𝑖𝑘𝑎 ] = ℓ [𝐶 cos(ℓ𝑎) + 𝐷 Sin(ℓ𝑎) ] (2)

The continuity of 𝝍(𝒙) at x=a, says

𝜓𝐼𝐼 (𝑥 = 𝑎) = 𝜓𝐼𝐼𝐼 (𝑥 = 𝑎) =>

𝐶 sin(ℓ𝑎) + 𝐷 cos(ℓ𝑎) = 𝐹𝑒 𝑖𝑘𝑎 (3)

𝑑𝜓(𝑥)
The continuity of at x=a, says
𝑑𝑥

𝑑𝜓𝐼𝐼 (𝑥 = 𝑎) 𝑑𝜓𝐼𝐼𝐼 (𝑥 = 𝑎)
= =>
𝑑𝑥 𝑑𝑥
ℓ [𝐶 cos(ℓ𝑎) − 𝐷 Sin(ℓ𝑎) ] = 𝑖𝑘𝐹𝑒 𝑖𝑘𝑎 (4)
We use (3) and (4) to eliminate C and D :
Multiply (3) with sin(ℓ𝑎) and (4) with cos(ℓ𝑎):
𝐶 sin2 (ℓ𝑎) +𝐷sin(ℓ𝑎) cos(ℓ𝑎) = 𝐹𝑒 𝑖𝑘𝑎 sin(ℓ𝑎)
ℓ [𝐶 cos2 (ℓ𝑎) −𝐷 Sin(ℓ𝑎) cos(ℓ𝑎) ] = 𝑖𝑘𝐹𝑒 𝑖𝑘𝑎 cos(ℓ𝑎)
Add them:
𝑘
𝐶 = 𝐹𝑒 𝑖𝑘𝑎 (sin(ℓ𝑎) + 𝑖 cos(ℓ𝑎)) (5)
ℓ
 26

Multiply (3) with cos(ℓ𝑎) and (4) with sin(ℓ𝑎) :

𝐶sin(ℓ𝑎) cos(ℓ𝑎) + 𝐷 cos2 (ℓ𝑎) = 𝐹𝑒 𝑖𝑘𝑎 cos(ℓ𝑎)

ℓ [𝐶sin(ℓ𝑎) cos(ℓ𝑎) −𝐷sin2 (ℓ𝑎) ] = 𝑖𝑘𝐹𝑒 𝑖𝑘𝑎 sin(ℓ𝑎)
Subtract them:
𝑘
𝐷 = 𝐹𝑒 𝑖𝑘𝑎 (cos(ℓ𝑎) − 𝑖 sin(ℓ𝑎)) (6)
ℓ
Put (5) and (6) into (1):
𝑘
𝐴𝑒 −𝑖𝑘𝑎 + 𝐵𝑒 𝑖𝑘𝑎 = − sin(ℓ𝑎) 𝐹𝑒 𝑖𝑘𝑎 (sin(ℓ𝑎) + 𝑖 cos(ℓ𝑎))
ℓ
𝑘
+ cos(ℓ𝑎) 𝐹𝑒 𝑖𝑘𝑎 (cos(ℓ𝑎) − 𝑖 sin(ℓ𝑎))
ℓ
𝑘
= 𝐹𝑒 𝑖𝑘𝑎 {−sin
⏟ 2 (ℓ𝑎) + cos2 (ℓ𝑎) − 𝑖 2⏟cos(ℓ𝑎) sin(ℓ𝑎)}
cos(2ℓ𝑎)
ℓ
sin(2ℓ𝑎)

𝑘
𝐴𝑒 −𝑖𝑘𝑎 + 𝐵𝑒 𝑖𝑘𝑎 = 𝐹𝑒 𝑖𝑘𝑎 {cos(2ℓ𝑎) − 𝑖 sin(2ℓ𝑎)} (7)
ℓ
Put (5) and (6) into (2):
𝑖𝑘[𝐴𝑒 −𝑖𝑘𝑎 − 𝐵𝑒 𝑖𝑘𝑎 ]
𝑘
= ℓ [𝐹𝑒 𝑖𝑘𝑎 (sin(ℓ𝑎) + 𝑖 cos(ℓ𝑎)) cos(ℓ𝑎)
ℓ
𝑘
+ 𝐹𝑒 𝑖𝑘𝑎 (cos(ℓ𝑎) − 𝑖 sin(ℓ𝑎)) Sin(ℓ𝑎) ]
ℓ
𝑘
= ℓ𝐹𝑒 𝑖𝑘𝑎 [2
⏟ cos(ℓ𝑎) sin(ℓ𝑎) + 𝑖 ⏟ 2 (ℓ𝑎) + cos2 (ℓ𝑎))]
(−sin
ℓ cos(2ℓ𝑎)
sin(2ℓ𝑎)

ℓ
[𝐴𝑒 −𝑖𝑘𝑎 − 𝐵𝑒 𝑖𝑘𝑎 ] = 𝐹𝑒 𝑖𝑘𝑎 [cos(2ℓ𝑎) − 𝑖 sin(2ℓ𝑎)] (8)
𝑘
 27

(7)+(8)=>
ℓ 𝑘
2𝐴𝑒 −𝑖𝑘𝑎 = 𝐹𝑒 𝑖𝑘𝑎 [2 cos(2ℓ𝑎) − 𝑖 ( + ) sin(2ℓ𝑎)]
𝑘 ℓ
𝑖2𝑘𝑎
ℓ2 + 𝑘 2
𝐴 = 𝐹𝑒 [cos(2ℓ𝑎) − 𝑖 ( ) sin(2ℓ𝑎)] (9)
2𝑘ℓ
Put this into (7)
ℓ 𝑘
2𝐵𝑒 𝑖𝑘𝑎 = 𝐹𝑒 𝑖𝑘𝑎 [𝑖 ( − ) sin(2ℓ𝑎)]
𝑘 ℓ
ℓ2 − 𝑘 2
𝐵 = 𝐹 [𝑖 ( ) sin(2ℓ𝑎)]
2𝑘ℓ

Thus
sin(2ℓ𝑎) 2
𝐵=𝑖 (ℓ − 𝑘 2 )𝐹
(2𝑘ℓ)
And from (9)=>

𝑒 −2𝑖𝑘𝑎
𝐹= 𝐴
(ℓ2 + 𝑘 2 )
cos(2ℓ𝑎) − 𝑖 sin(2ℓ𝑎)
(2𝑘ℓ)

The transmission and reflection coefficients are defined as

𝑇(𝑅 ): Probability that incident particle is transmitted (reflected)
𝐹2 𝐵2
𝑇≡| | , 𝑅≡| |
𝐴 𝐴

𝐹 𝑒 −2𝑖𝑘1 𝑎
=
𝐴 (ℓ2 + 𝑘 2 )
(cos(2ℓ𝑎) − 𝑖 sin(2ℓ𝑎))
(2𝑘ℓ)
 28

Thus,

|𝐹 |2 1
𝑇= =
|𝐴|2 2
1 ℓ2 + 𝑘 2
⏟ (2ℓ𝑎) + 4 ( 𝑘ℓ ) sin2 (2ℓ𝑎))
(cos 2

1−sin2 (2ℓ𝑎)
1
=
2 1 𝑘 4 + ℓ4 + 2𝑘 2 ℓ2
(1 + sin (2ℓ𝑎) [−1 + ( )])
4 𝑘 2 ℓ2

1
𝑇= 2 (∗∗)
2 1 ℓ2 − 𝑘 2
(1 + sin (2ℓ𝑎) ( ) )
4 𝑘ℓ

2
2 2 2
2𝑚 2𝑚 2 2𝑚|𝑉0 |
(ℓ − 𝑘 ) = ( 2 (𝐸 + |𝑉0 |) − 2 𝐸) = ( ) (𝑎)
ℏ ℏ ℏ2
2𝑚 2𝑚
(𝑘ℓ)2 = ( 2 𝐸 2 (𝐸 + |𝑉0 |)) (𝑏)
ℏ ℏ
(𝑎) 𝑉02 𝑉02 1
= = =
(𝑏) 𝐸(𝐸 + |𝑉0 |) |𝑉0 |𝐸(1 + 𝐸/|𝑉0 |) 𝐸/|𝑉0 |(1 + 𝐸/|𝑉0 |)
Let
𝝐 ≡ 𝑬/|𝑽𝟎 |

2𝑚 2𝑚 2𝑚|𝑉0 |
2ℓ𝑎 = 2𝑎√ (𝐸 + |𝑉0 |) = 2𝑎 √ (𝐸 + |𝑉0 |) = 2𝑎 √ (1 + 𝜖 )
ℏ2 ℏ2 ℏ2
Therefore
−1
1 2𝑚|𝑉0 |
𝑇 = {1 + sin2 (2𝑎√ (1 + 𝜖 ))} (∗)
4𝜖(1 + 𝜖) ℏ2
 29

Total Transmission
−1
1 2𝑚|𝑉0 |
𝑇 = {1 + sin2 (2𝑎√ (1 + 𝜖 ))}
4𝜖(1 + 𝜖) ℏ2

 Note that we have total transmission at very high energies and/or at weak potential
well (potential well is completely transparent to the incoming particle):
𝑬 ≫ 𝑽𝟎 => 𝝐 ≫ 𝟏 ⟹ 𝑻 → 𝟏 𝒂𝒏𝒅 𝑹 → 𝟎

This means that particles with certain energies will be completely transmitted.
However, remember that the wave function associated with a particular, well
defined energy is not normalizable. This means that a wave packet whose energy is
peaked near the proper value will be mostly transmitted.
 30

 Note also that we have total transmission when

2𝑚 2𝑚
sin2 (2𝑎√ (𝐸 + |𝑉0 |)) = 0 => 2𝑎 √ (𝐸 + |𝑉0 |) = 𝑛𝜋 , 𝑛 = 0,1,2, …
ℏ2 ⏟ ℏ2

ℓ
The energies for which the finite potential well becomes transparent is given by
ℏ2 𝜋 2
𝐸𝑛 = 2
𝑛2 − |𝑉0 |, 𝑛 = 0,1,2, … : requirement for the perfect transition
2𝑚(2𝑎)
The right-hand side is the energy of the n-th bound state of the infinite square well
of width 2a.
We get full transmission for those energies 𝐸𝑛 > 0 that are in the spectrum of the
infinite-square-well extension of our finite square well. Since infinite square well
bound states are characterized by fitting an integer number of half wavelengths, we
have a resonance type situation in which perfect transmission is happening when
the scattering waves fit perfectly inside the finite square well. The phenomenon we
have observed is called resonant transmission which do not occur in classical
physic.
In figure below, transmission coefficient in Eq. (*), is plotted as a function of the
energy showing the positions 𝐸𝑅 of the resonances.
 31

We can formulate the resonance condition also, as follows:

The maximum transmission occur when
2𝑎ℓ = 𝑛𝜋
Note that De Broglie wavelength of the wave inside is given by
2𝜋
𝜆ℓ =
ℓ
Thus
2𝜋
𝜆ℓ = 2𝑎 => 𝑛𝜆ℓ = 4𝑎 𝑅𝑒𝑠𝑜𝑛𝑎𝑛𝑐𝑒 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜𝑛
𝑛𝜋

This means that the transmission is boosted whenever the distance 4a back and
forth that wave traverses inside the well due to reflection is equal to an integer
number of its de Broglie wave length.

In wave language, this resonance phenomenon is due to a destructive

interference between the wave reflected at x=-a and wave reflected once, twice,
thrice,…, at the edge x=a, which does not occur in classical physics.
 32

This phenomenon is observed experimentally in a number of cases such as when

scattering low-energy (𝐸~0.1 𝑒𝑉) electrons off noble atoms (known as the
Ramsauer–Townsend effect, a consequence of symmetry of noble atoms) and
neutrons off nuclei.

At low energies, the collusions are all elastic and classically

𝜎𝑒𝑙𝑎𝑠𝑡𝑖𝑐 ~𝐸
But instead what is observed is that 𝜎𝑒𝑙𝑎𝑠𝑡𝑖𝑐 drops almost to zero before increasing
again with increasing energy. This is due to the electron acting like a wave and we
can equate the atom to a potential well so that at certain wavelengths of the electron
𝜎𝑒𝑙𝑎𝑠𝑡𝑖𝑐 → 0.
Problem. Liboff CH7.48

The scattering cross section for the scattering of electrons by a rare gas of krypton

atoms exhibits a low-energy minimum at 𝐸 ≅ 0.9 𝑒𝑉. Assuming that the diameter

of the atomic well seen by the electrons is 1 Bohr radius (= 5.29 × 10−11 𝑚),

calculate its depth.

From

ℏ2 𝜋 2 2
ℏ2 𝜋 2
𝐸𝑛 + |𝑉0 | = 2
𝑛 => |𝑉0 | = 2
𝑛2 − 𝐸𝑛
2𝑚(2𝑎) 2𝑚(2𝑎)

(ℏ𝑐)2 𝜋 2 2
(197 × 10−9 𝑒𝑉. 𝑚)2 𝜋 2
|𝑉0 | = 2 2
𝑛 − 𝐸𝑛 = 6 −11 2
12 − 0.9 𝑒𝑉
2(𝑚𝑐 )(2 𝑎) 2(0.5 × 10 𝑒𝑉)(5.29 × 10 𝑚)

|𝑉0 | = 136 𝑒𝑉
 33

Problem 2.23: Evaluate the following integrals

a) Since
−3 < −2 < 1

+1
∫ 𝛿 (𝑥 − (−2) ⏟ 3 − 3𝑥 2 + 2𝑥 − 1) 𝑑𝑥 = 𝑓(𝑥0 ) = 𝑓(−2)
⏟ ) (𝑥
−3 𝑥0 𝑓(𝑥)

𝑓(−2) = (−2)3 − 3(−2)2 + 2(−2) − 1 = −25

b) Since
0<𝜋<∞

+∞
∫ 𝛿 (𝑥 − 𝜋 (𝐶𝑜𝑠3𝑥 + 2) 𝑑𝑥 = 𝑓(𝑥0 ) = 𝑓(𝜋) = 𝐶𝑜𝑠3𝜋 + 2 = −1 + 2
⏟) ⏟
0 𝑥0 𝑓(𝑥)
=1
c) the domain of integration (-1,+1) does not include the point 𝑥0 = 2

+1
∫ 𝛿 (𝑥 − ⏟ (𝑒 |𝑥|+3 ) 𝑑𝑥 = 0
2)⏟
−1 𝑥0 𝑓(𝑥)
 1

Problem 2.33: Rectangular Barrier -Scattering:

E>0
𝑉(𝑥) = 0 𝑥 < −𝑎 (I)
𝑉(𝑥) = 𝑉0 −𝑎 ≤ 𝑥 ≤ 𝑎 (II)
𝑉(𝑥) = 0 𝑥 > 𝑎 (III)

Solution to this problem can be directly written from the solution to the problem of
scattering from the finite potential well.
2
 3

Since the Schrödinger equation for the finite barrier shown above is
identical in form to that for a finite well, the solutions must be identical in
form too.
The only difference in the two solutions is that in region II, 𝓵 → 𝒒;
otherwise the solutions must be identical.

In rectangular Well -Scattering, we had

We obtained the transmition coefficient for the rectangular well scattering

(RWS) given by
2 −1
2
1 ℓ2 − 𝑘 2 2𝑚 2𝑚
𝑇𝑅𝑊𝑆 = {1 + sin (2ℓ𝑎) ( ) } (1) 𝑘 2 ≡ 𝐸 , ℓ 2
≡ (𝐸 + |𝑉0 |)
4 𝑘ℓ ℏ2 ℏ2

Now to find the transition coefficient for the rectangular barrier scattering
(RBS) we make the replacement :
2 −1
2
𝑞2 − 𝑘 2
In eq (1) replace ℓ → 𝑞 => 𝑇𝑅𝐵𝑆 = {1 + sin (2𝑞𝑎) ( ) } (2)
2𝑘𝑞
 4

Now,
2𝑚
𝑞2 = (𝐸 − 𝑉0 )
ℏ2
−1
2 2 2
𝑞 − 𝑘
𝑇𝑅𝐵𝑆 = {1 + sin2 (2𝑞𝑎) ( ) } (2)
2𝑘𝑞

2 2 )2
2𝑚 2𝑚𝐸 2 2𝑚𝑉0 2
(𝑞 − 𝑘 = ( 2 (𝐸 − 𝑉0 ) − 2 ) = ( 2 ) (𝑎)
ℏ ℏ ℏ
2𝑚𝐸 2𝑚
(𝑘𝑞 )2 = ( 2 2 (𝐸 − 𝑉0 )) (𝑏)
ℏ ℏ

(𝑎) 𝑉02 𝑉0
= =
(𝑏) 𝐸(𝐸 − 𝑉0 ) 𝐸(𝐸/𝑉0 − 1)
Therefore
−1
2
1 𝑉0
𝑇𝑅𝐵𝑆 = {1 + sin (2𝑞𝑎) }
4 𝐸(𝐸/𝑉0 − 1)

1
𝑇𝑅𝐵𝑆 = ,𝑅 = 1 − 𝑇
1 1 2𝑚
[1 + sin2 (2𝑎√ 2 (𝐸 − 𝑉0 ))]
4 (𝐸/𝑉0 )(𝐸/𝑉0 − 1) ℏ

Note that we have total transmission

 At very high energies and/or at weak potential barrier:

𝐸 ≫ 𝑉0 ⟹ 𝑇 → 1 𝑎𝑛𝑑 𝑅 → 0
 When
2𝑚 2𝑚
sin2 (2𝑎√ (𝐸 − 𝑉0 )) = 0 => 2𝑎 √ (𝐸 − 𝑉0 ) = 𝑛𝜋 , 𝑛 = 0,1,2, …
ℏ2 ℏ2
They are known as resonances which do not occur in classical physic.
OR
2𝜋
2𝑞𝑎 ≡ 2𝑎 = 𝑛𝜋 => 4𝑎 = 𝑛𝜆𝑞
𝜆𝑞
 5

This means that the transmission is boosted whenever the distance 4a

back and forth that wave traverses inside the well due to reflection is
equal to an integer number of its de Broglie wave length.
If we write the resonance condition in terms of energy, we get

2𝑚
2𝑎√ (𝐸 − 𝑉0 ) = 𝑛𝜋 =>
ℏ2

ℏ2 𝜋 2
𝐸 = 𝑉0 + 2
𝑛2 , 𝑛 = 0,1,2, … : requirement for the perfect transition
2𝑚(2𝑎)
 6

Rectangular Barrier Bound state: Tunnel Effect

E<V0

𝑑2 𝜓 2𝑚
Schrödinger eq: + (𝐸 − 𝑉 (𝑥 ))𝜓 = 0
𝑑𝑥 2 ℏ2

In the region I: x<-a: V(x)=0

𝑑 2 𝜓𝐼 2𝑚
2
+ 𝑘 2 𝜓𝐼 = 0, 𝑘2 ≡ 𝐸>0
𝑑𝑥 ℏ2
The general solution
𝜓𝐼 (𝑥 ) = 𝐴𝑒 𝑖𝑘𝑥 + 𝐵𝑒 −𝑖𝑘𝑥 , 𝑓𝑜𝑟 𝑥 < −𝑎

In the region III: x>a: V(x)=0

𝑑2 𝜓𝐼𝐼𝐼
2
+ 𝑘 2 𝜓𝐼𝐼𝐼 = 0
𝑑𝑥
The general solution
𝜓𝐼𝐼𝐼 (𝑥 ) = 𝐹𝑒 𝑖𝑘𝑥
 7

In the region II: −𝑎 < 𝑥 < 𝑎, 𝑉 = 𝑉0 > 𝐸

𝑑2 𝜓𝐼𝐼 2𝑚
+ 2 (𝐸 − 𝑉0 ) 𝜓𝐼𝐼 = 0
𝑑𝑥 2 ⏟
ℏ
<0

Rewrite it
𝑑2 𝜓𝐼𝐼 2𝑚 2𝑚
− 2 (𝑉0 − 𝐸 ) 𝜓𝐼𝐼 = 0 𝑤𝑖𝑡ℎ 𝜅 2 ≡ (𝑉 − 𝐸 ) > 0
𝑑𝑥 2 ⏟
ℏ ℏ2 0
𝜅2 >0

𝜓𝐼𝐼 (𝑥 ) = 𝑀𝑒 𝜅𝑥 + 𝑁𝑒 −𝜅𝑥 𝑓𝑜𝑟 − 𝑎 < 𝑥 < 𝑎

We can convert a combination of Sin and Cos functions by a simple

replacement into a combination of positive and negative real exponential
functions:

𝑒 𝑖𝑞𝑥 − 𝑒 −𝑖𝑞𝑥 𝑒 𝑖𝑞𝑥 + 𝑒 −𝑖𝑞𝑥

𝐶 sin 𝑞𝑥 + 𝐷 cos 𝑞𝑥 = 𝐶 ( )+( )
2𝑖 2
𝑖𝑞𝑥
−𝐶𝑖 + 𝐷 −𝑖𝑞𝑥
𝐶𝑖 + 𝐷
=𝑒 ( )+𝑒 ( )
⏟ 2 ⏟ 2
𝑀 𝑁
OR

𝐶𝑠𝑖𝑛𝑞𝑥 + 𝐷 cos 𝑞𝑥 = 𝑀𝑒 𝑖𝑞𝑥 + 𝑁𝑒 −𝑖𝑞𝑥 (∗)

Now, in (*) replace

𝑞 → −𝑖𝜅
So that it becomes
𝐶𝑠𝑖𝑛𝑞𝑥 + 𝐷 cos 𝑞𝑥 → 𝑀𝑒 𝑖(−𝑖𝜅)𝑥 + 𝑁𝑒 −𝑖(−𝑖𝜅)𝑥 = 𝑀𝑒 𝜅𝑥 + 𝑁𝑒 −𝜅𝑥
 8

−1
2 2 2
𝑞 −𝑘
𝑇𝑅𝐵𝑆 = {1 + sin2 (2𝑞𝑎) ( ) } (2)
2𝑘𝑞
 9

Now, we can use this result to find the transition coefficient of the
rectangular barrier penetration (RBP) from the transition coefficient of
the rectangular barrier scattering (RBS) in eq. (2) by making the
replacement
𝑞 → −𝑖𝜅

Therefore make the replacement 𝑞 → −𝑖𝜅 in eq(2):

−1
2 2 2
(−𝑖𝜅) − 𝑘
𝑇𝑅𝐵𝑆 = {1 + sin2 (2(−𝑖𝜅)𝑎) ( ) } → 𝑇𝑅𝐵𝑃
2𝑘(−𝑖𝜅)

Using
sin(𝑖𝑧) = 𝑖 sinh(𝑧)
We get the transition coefficient of the rectangular barrier penetration
−1
2 2 2
𝜅 +𝑘
𝑇𝑅𝐵𝑃 = {1 + sinh2 (2𝜅𝑎) ( ) } (3)
2𝑘𝜅
 10

Special cases:
1. 𝐸 ≪ 𝑉0 which implies that

2𝑚𝑎2
𝜅𝑎 ≡ √ 2 (𝑉0 − 𝐸 ) ≫ 1,
ℏ

Then

1 2𝜅𝑎 1
sinh(2𝜅𝑎) = (𝑒 − 𝑒 −2𝜅𝑎 ) ≅ 𝑒 2𝜅𝑎
2 2
So that
−1
2 2
1 2
𝜅 +𝑘 2
4𝑘𝜅 2 −4𝜅𝑎
𝑇𝑅𝐵𝑃 ≅ {( 𝑒 2𝜅𝑎 ) ( ) } => 𝑇𝑅𝐵𝑃 ≅( 2 ) 𝑒
2 2𝑘𝜅 𝜅 + 𝑘2

𝑇𝑅𝐵𝑃 ≠ 0
Classically it was zero. We see that even though the energy is very much
below the top of the barrier there is transmission. This is a wave
phenomenon and in quantum mechanics it is also exhibited by particles
2. When 𝐸 ≈ 𝑉0 => 𝜅𝑎 → 0. Then,
sinh(2𝜅𝑎) ≅ 2𝜅𝑎
Then,
2 −1
2
𝜅2 + 𝑘2
𝑇𝑅𝐵𝑃 ≅ {1 + sinh (2𝜅𝑎) ( ) }
2𝑘𝜅

2 2 )2 −1 2 2 )2 −1
(𝜅 + 𝑘 (𝜅 + 𝑘
𝑇𝑅𝐵𝑃 ≅ [1 + (2𝜅 )2 𝑎2 ] ≅ [1 + 𝑎2 ]
(2𝜅 )2 𝑘 2 𝑘2
 11

2 2 )2
2𝑚 2𝑚 2 2𝑚 2
(𝜅 + 𝑘 = ( 2 (𝑉0 − 𝐸 ) + 2 𝐸) = ( 2 𝑉0 ) (𝑎)
ℏ ℏ ℏ
2𝑚
𝑘2 = 2 𝐸 (𝑏)
ℏ
𝐸≈𝑉
(𝑎) 2𝑚 𝑉02 0
2𝑚
=> 2 ⏞
≅ 𝑉
(𝑏) ℏ 𝐸 ℏ2 0
Thus,
𝐸≈𝑉0 −1
2𝑚𝑎2 𝑉0
𝑇𝑅𝐵𝑃 ⏞ [1 +
≅ ]
ℏ2

Note that 𝑇 < 1, which means that there is no total transmission.

When we take the classical limit, ℏ → 0, then 𝑇 → 0, as in the classical
case.
 12

Let us consider a potential

barrier as shown in the next
Fig. Suppose that an incident
particle (say an electron )with
energy E coming from the left
collides the potential barrier.
The energy E is assumed to be
lower than the top of the
barrier. In the classical
mechanics, the particle is
completely reflected by the
potential barrier. In quantum
mechanics, there is some
probability that it will be
reflected, but the other
penetrates the barrier and pass
through into the right-hand side
region to proceed to the far
right. This is the tunnel
effect.
 13

Let us analyze an apparent difficulty here: The wave function does not vanish
inside the barrier and thus there appears to be probability of finding a particle
with negative kinetic energy . How can this make sense?
We look to the uncertainty relation to remove an apparent paradox that arises
from “too classical” a description of the process.
An experiment to study the particle inside the potential barrier must be able to
localize it within an accuracy
Δ𝑥 ≪ 2𝑎
This measurement will transfer to the particle momentum, with an uncertainty
ℏ
Δ𝑝 ≫
2𝑎
which corresponds to a transfer of energy
Δ𝑝2 1 ℏ 2
Δ𝐸 = ≫ ( ) (𝑎)
2𝑚 2𝑚 2𝑎
In order to observe the negative kinetic energy 𝐾𝐸 = 𝐸 − 𝑉0 , the uncertainty
Δ𝐸 must be much less than |𝐾𝐸 |
ℏ2 𝜅 2
Δ𝐸 ≪ |𝐾𝐸 | = |𝐸 − 𝑉0 | = (𝑏)
2𝑚
From (a) and (b) we get
ℏ2 𝜅 2 1 ℏ 2
≫ Δ𝐸 ≫ ( ) => 2𝜅𝑎 ≫ 1
2𝑚 2𝑚 2𝑎
However, under this circumstances, the quantity to be measured 𝑇𝑅𝐵𝑃
4𝑘𝜅 2 −4𝜅𝑎
𝑇𝑅𝐵𝑃 ≅( 2 ) 𝑒
𝜅 + 𝑘2
is vanishingly small. For example, for 𝜅𝑎 = 10, 𝑇𝑅𝐵𝑃 ≈ 𝑒 −4𝜅𝑎 ≈ 10−18 ,
which means that 𝑇𝑅𝐵𝑃 is immeasurable and we have a total reflection.
Therefore there is no particle in region II to measure the KE.
 14

𝜶 − 𝑫𝑬𝑪𝑨𝒀

The Rutherford’s experiment that led to the discovery of the atomic nucleus:

α-particles were used to hit thin foils (atoms in a crystal), most of them went straight
through the foil as atoms are pretty empty, a few of them got scattered right back (by
electrostatic repulsion) when they were heading for the core, which could be
explained by an atom with a very small, dense, positively-charged nucleus at its
center

Where did these α-particles come from?

α-particle is a disintegration product of the radioactive decay of unstable nucleus:

𝐴 (𝐴−4) 4
⏟
𝑍 𝑋 → (𝑍−2)
⏟ 𝑌 + 𝐻𝑒
2⏟
𝑃𝑎𝑟𝑒𝑛𝑡 𝐷𝑎𝑢𝑔ℎ𝑡𝑒𝑟 𝛼−𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒
𝑁𝑢𝑐𝑙𝑒𝑢𝑠
𝑁𝑢𝑐𝑙𝑒𝑢𝑠

A=(# of proton)+(#of neutron) Z=(# of proton)

EX:
𝟐𝟑𝟎 𝟐𝟐𝟔 𝟒
𝟗𝟎𝑻𝒉 → 𝟖𝟖𝑹𝒂 + 𝟐𝑯𝒆
⏟
𝜶−𝒑𝒂𝒓𝒕𝒊𝒄𝒍𝒆
 15

The diagram below describes the potential energy U(r) of an 𝛼-particle as a

function of its separation from the center of the nucleus. The horizontal line
corresponds to the attractive nuclear force and the exponential curve corresponds
to the repulsive electrical force due to other protons.

The nuclear force is very strong and extremely short range,≈ 10−15 𝑚. α-particle is
confined in the parent nucleus by the nuclear force (when 0<r<R), and it is
insignificant outside the nucleus (when r>R). Beyond the nuclear force range, r>R,
the α-particle feels only the Coulomb potential U(r) which increases closer to the
nucleus.

We can imagine that the alpha particle “moves” about the small space in the
nucleus with some energy (𝐸𝛼 ~9 𝑀𝑒𝑉 ) that is less than the energy required to
overcome the Coulomb barrier (~30 𝑀𝑒𝑉.) Therefore the alpha particle is trapped
in a finite potential well.

Classically, such an α-particle with 𝐸𝛼 ~9 𝑀𝑒𝑉 initially bound to the nucleus can
not escape. However, the explanation of emission of α-particles is possible in
terms of the tunnel effect. According to quantum mechanics, the α-particle, with
its wave attributes, may tunnel through the barrier to appear on the outside. We
shall not go through the detailed arguments used to estimate T in this case, but we
shall note that they involve the approximation

2𝑚(𝑉(𝑥)−𝐸)
−2 ∫𝑏𝑎𝑟𝑟𝑖𝑒𝑟 𝑑𝑥 √
𝑇≅ 𝑒 ℏ2

After a lengthy calculation, including some approximations, the final result is of

the form

𝑇 ≅ 𝐴 𝑒 −𝐵(𝑍−2)/√𝐸𝛼

A and B are constants.

 16

The diagram below describes the potential energy U(r) of an 𝛼-particle as a

function of its separation from the center of the nucleus. The horizontal line
corresponds to the attractive nuclear force and the exponential curve corresponds
to the repulsive electrical force due to other protons.

The nuclear force is very strong and extremely short range,≈ 10−15 𝑚. α-particle is
confined in the parent nucleus by the nuclear force (when 0<r<R), and it is
insignificant outside the nucleus (when r>R). Beyond the nuclear force range, r>R,
the α-particle feels only the Coulomb potential U(r) which increases closer to the
nucleus.

We can imagine that the alpha particle “moves” about the small space in the
nucleus with some energy (𝐸𝛼 ~9 𝑀𝑒𝑉 ) that is less than the energy required to
overcome the Coulomb barrier (~30 𝑀𝑒𝑉.) Therefore the alpha particle is trapped
in a finite potential well.

Classically, such an α-particle with 𝐸𝛼 ~9 𝑀𝑒𝑉 initially bound to the nucleus can
not escape. However, the explanation of emission of α-particles is possible in
terms of the tunnel effect. According to quantum mechanics, the α-particle, with
its wave attributes, may tunnel through the barrier to appear on the outside. We
shall not go through the detailed arguments used to estimate T in this case, but we
shall note that they involve the approximation

2𝑚(𝑉(𝑥)−𝐸)
−2 ∫𝑏𝑎𝑟𝑟𝑖𝑒𝑟 𝑑𝑥 √
𝑇≅ 𝑒 ℏ2

After a lengthy calculation, including some approximations, the final result is of

the form

𝑇 ≅ 𝐴 𝑒 −𝐵(𝑍−2)/√𝐸𝛼

A and B are constants.

17
18
 1

CH0
HISTORICAL REVIEW: EXPERIMENTS AND THEORIES

ORIGIN OF QUANTUM THEORY

The nature and behavior of matter and energy at the atomic and subatomic level is referred to as

Quantum Physics and Quantum Mechanics.

Quantum mechanics was initially invented (at the beginning of the 20th century) because the classical
theories provided no means to explain the properties of atoms, electrons, and electromagnetic radiation..
The physics of the 19th century and before is called the classical physics. It consisted
essentially of

 Classical Mechanics: Newton’s theory of mechanics, Newton’s universal theory of gravitation.

Classical mechanics was used to predict the dynamics of material bodies, and its present form is largely due
to Lagrange, and Hamilton.

 Electricity and Magnetism: Maxwell’s theory of electromagnetic phenomena (which masterfully

dealt with problems involving electric and magnetic fields and their unification)
 Thermodynamics that described heat and Statistical Mechanics which uses probability theory to describe
systems with large number of particles.
By 1900, it seemed that all known physical phenomena could be explained within the framework of these
general theories of matter and radiation.

At the turn of the 20th century, with the development of new experimental techniques to probe atomic and
subatomic structures, it appeared that classical physics fails miserably in providing the proper explanation
for several newly discovered phenomena.

Some of the phenomena that could not be explained by classical physics at the beginning of the 20th
century are

 Black body radiation

 Photoelectric effect
 Compton scattering
 Emission and absorption spectra of atomic systems (the hydrogen spectra )
 Particle diffraction, etc.

It became evident that the validity of classical physics breaks off at the microscopic level and new
concepts radically different from the concepts of classical physics had to be invoked to describe the
structure of atoms and molecules and how light interacts with them.

Some of the new concepts :

 The quantization of physical concepts, such as energy and momentum
 The particle properties of radiation
 Wave properties of matter

All these developments led to the emerge of quantum theory.

 2

BLACK BODY RADIATION

The problem is just about how heated bodies radiate. All objects emit EM radiation to their surroundings
and also absorb radiation from them continuously, whatever their temperatures are.
This is the Thermal radiation which is generated by the thermal motion of charged particles in matter.
It represents a conversion of thermal energy into electromagnetic energy.
 If a gas of particular element is heated to very high temperature it glows, i.e., sends electromagnetic (EM) radiation.
If this EM radiation is examined in a spectroscope, it is seen that only a discrete set of wavelengths –a line spectrum
– is emitted.

 On the other hand, solids (such as the tungsten filament of an electric light bulb) when heated up to similarly high
T, emit a continuous range of wavelengths.

How to investigate the thermal radiation from heated bodies?

When radiation falls on an object, some of it might be absorbed and some reflected.
To begin analyzing heat radiation, we need to be specific about the body doing the radiating:
The simplest possible case is an idealized body which is a perfect absorber. For obvious reasons, this is
called a “black body”.

An object in thermal equilibrium with its surroundings radiates as much energy as it absorbs. It thus
follows that a blackbody is a perfect absorber as well as a perfect emitter of radiation.

The spectrum of the blackbody radiation is the ideal spectrum which is a characteristic of the
temperature T but completely independent of the nature and structure of the body producing it.

How do we construct a perfect black body in the laboratory?

Although an ideal blackbody does not really exist, to a reasonable approximation, all matter in thermal
equilibrium behaves like a blackbody.
To check this ideal (black body) radiation experimentally in 1859 Kirchhoff had a good idea to approximate
black-body radiation:

A cavity whose internal walls perfectly reflect electromagnetic

radiation has a small hole. It is an excellent absorber, since any light
entering the hole would be reflected back and forth by the walls of the
cavity and it is eventually be absorbed (this technique leads to the
alternative term cavity radiation).If this cavity is heated, as the
temperature increases, the hole will eventually begin to glow and the
radiation that leaves the hole is blackbody radiation .
 3

To understand the radiation inside the cavity, one needs simply to analyze the spectral distribution of the
radiation coming out of the hole.

The first quantitative guess based on experimental observation of hole radiation was:
Stefan’s Law (1879):
Josef Stefan found experimentally that total emitted power
P=E/t, from a unit area of a hot body at temperature T at all
frequencies is given by
𝑃 = 𝜎𝑇 4
where
2𝜋 5 𝑘 4
𝜎= = 5.67 × 10−8 𝑊𝑚−2 𝐾 −4
15ℎ3 𝑐 2
∶ 𝑆𝑡𝑒𝑓𝑎𝑛 − 𝐵𝑜𝑙𝑡𝑧𝑚𝑎𝑛𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡

This experimental result was confirmed theoretically by Boltzmann five years later using a combination of
the laws of thermodynamics and Maxwell’s eqs
Until the end of the 19.th century a great volume of experimental data on this subject was cumulated. They
found a radiation intensity/frequency curve close to this:

 𝑢(𝜈, 𝑇)𝑑𝜈 : the radiant energy emitted per time in frequency range (𝜈, 𝜈 + 𝑑𝜈) from the unit area of the
surface at temperature T.

 𝑢(𝜈, 𝑇) is called the radiant energy density.

 The radiation emitted has a continuous energy distribution:

 𝑢(𝜈, 𝑇) increases as T increases.
 The energy density shows a pronounced maximum at a given frequency, which increases with
temperature: 𝜈𝑚𝑎𝑥 ∝ 𝑇 . That is the reason why the color of heated object changes as its
temperature increases from red to yellow to white.
𝑹𝑬𝑫 → 𝑶𝑹𝑨𝑵𝑮𝑬 → 𝒀𝑬𝑳𝑳𝑶𝑾 → 𝑩𝑳𝑼𝑬
𝜈𝑟𝑒𝑑 < 𝜈𝑜𝑟𝑎𝑛𝑔𝑒 < 𝜈𝑦𝑒𝑙𝑙𝑜𝑤 < 𝜈𝑏𝑙𝑢𝑒 , (𝜈𝑟𝑒𝑑 ~400 𝑇𝐻𝑧 , 𝜈𝑏𝑙𝑢𝑒 ~600 𝑇𝐻𝑧)
 4

 From these curves, you can tell what color or wavelength of light is being emitted by the blackbody. This
is done by looking for the peak wavelength using Wien's displacement law:

𝜆𝑝𝑒𝑎𝑘 𝑇 = 2.89 × 10−3 𝑚 𝐾 ∶ 𝐓𝐡𝐞 𝐖𝐢𝐞𝐧′ 𝐬 𝐝𝐢𝐬𝐩𝐥𝐚𝐜𝐞𝐦𝐞𝐧𝐭 𝐥𝐚𝐰

 U(T) : the total radiant energy per unit volume in the cavity at temperature T
∞
𝑈(𝑇) = ∫ 𝑢(𝜈, 𝑇)𝑑𝜈
0

It only depends on temperature T. The actual value can be determined from the Stefan-Boltzmann law

The relation between the emitted power, 𝑷, which Stefan-Boltmann law gives, and
the total stored energy in the cavity, U(T), is given by (see Liboff page 33-34 or
explanation below)
𝑐
𝑃 = 𝑈(𝑇)
4
 5

EX: The sun has 𝑇~5000 𝐾. Power radiated from unit surface of sun is

𝑃 = (5.67 × 10−8 𝑊𝑚−2 𝐾 −4 )(5000 𝐾)4 => 𝑃 ≅ 3.5 × 107 𝑊/𝑚2

From Fig.1: Area under the curve for T=5000 K (Approximating the curve by a triangle)
1
𝑈(𝑇) ≈ (10 × 10−16 𝐽𝑚−3 𝐻𝑧 −1 )(10 × 1014 𝐻𝑧) = 0.5 𝐽𝑚−3
2
𝑐 (3 × 108 𝑚/𝑠)
𝑃 = 𝑈(𝑇) = 0.5 𝐽𝑚−3 => 𝑃 ≅ 3.8 × 107 𝑊/𝑚2
4 4
A pretty good agreement!

Why does the blackbody spectrum have the shape shown in the figure?

The Rayleigh-Jeans Approximation (1900)

J. Rayleigh (1900) and J. Jeans (1905) derived the spectral distribution of thermal radiation on the basis of the
classical electromagnetic theory.

They considered the thermal radiation to be a collection of standing waves having

a temperature T in a cubical enclosure with nodes at the surface.

Since each standing wave originates from an oscillating electric charge, they
(each standing wave ) are considered to be equivalent to a one-dimensional
oscillator (SHO). When the cavity is in thermal equilibrium, the electromagnetic
energy density inside the cavity is equal to the energy density of the charged
particles in the walls of the cavity and can be calculated by finding

𝑇ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑚𝑜𝑑𝑒𝑠

( 𝑠𝑡𝑎𝑛𝑑𝑖𝑛𝑔 𝑤𝑎𝑣𝑒𝑠 ≡ 𝑆𝐻𝑂)𝑜𝑓 𝑡ℎ𝑒 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑒𝑛𝑒𝑟𝑔𝑦
𝑢(𝜈) = 𝑟𝑎𝑑𝑖𝑎𝑡𝑖𝑜𝑛 𝑤𝑖𝑡ℎ ×( 𝑝𝑒𝑟 )
𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝜈 𝑖𝑛 𝑡ℎ𝑒 𝑟𝑎𝑛𝑔𝑒 ⏟𝑠𝑡𝑎𝑛𝑑𝑖𝑛𝑔 𝑤𝑎𝑣𝑒(≡ 𝑆𝐻𝑂)
(
⏟ (𝜈, 𝑣 + 𝑑𝜈) ) 𝜀̅
[𝑁(𝜈)𝑑𝜈]

𝑢(𝜈) = [𝑁(𝜈)𝑑𝜈] × [𝜀̅] (1)

 Assuming that the body is in the form of a cube with edge length L , by solving 3-dimensional wave equation , they
calculated

4𝜋𝜈 2
𝑁(𝜈) = 2 (2)
𝑐3
 6

 To calculate average energy per standing wave (𝑎 𝑆𝐻𝑂),  , classical kinetic theory is called on. According to its
law of equipartition of energy, for a system in thermal equilibrium, each degree of freedom has average energy
1
𝑘𝑇, where 𝑘 = 1.38 × 10−23 J/K : Boltzman constant. Thus a one-dimensional simple harmonic oscillator (a
2
standing wave) has total energy

1 1
𝑘𝑇 = 𝑘𝑇 + 𝑘𝑇
2⏟ 2⏟
𝐾𝑖𝑛𝑒𝑡𝑖𝑐 𝑃𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙
𝑒𝑛𝑒𝑟𝑔𝑦 𝑒𝑛𝑒𝑟𝑔𝑦

The origin of the equipartition law arises, basically , from a more complete result of classical statistical mechanics
called the Boltzmann distribution:
𝑒 −𝜀/𝑘𝑇
𝑃(𝜀) =
𝑘𝑇
such that

𝑷(𝜺)𝒅𝜺: The probability of finding a given entity (standing wave ) with energy  in the interval  ,   d .

The average energy per standing wave, , can be obtained from P ( ) by using
∞
∫0 𝜀 𝑃(𝜀)𝑑𝜀
𝜀̅ = ∞ (3)
∫0 𝑃(𝜀)𝑑𝜀

The denominator of Eq (3) is equal to unity, since it is the probability of finding the standing wave with any
energy.
Using the integral

n  ax n!
 x e dx 
0 a n1
One can easily show that
𝜀̅ = 𝑘𝑇 (4)
Now, substituting the parts, we find
8𝜋𝜈 2
𝑢𝑅𝐽 (𝜈) = ( 3 ) 𝑘𝑇 ∶ 𝑅𝑎𝑦𝑙𝑒𝑖𝑔ℎ − 𝐽𝑒𝑎𝑛𝑠 𝐹𝑜𝑟𝑚𝑢𝑙𝑎
𝑐
Rayleigh-Jeans formula contains everything that classical physics can say about the spectrum of blackbody
radiation. However, it can not possibly be correct !

This unrealistic behavior of the prediction of classical

theory at high frequencies is known as ultraviolet
catastrophe .

The conclusion : the need of revising our basic

assumptions about the behavior of the interaction of E.M
radiation with matter.
 7

PLANC’S RADIATION FORMULA

German physicist Max Planck (1900) solved the problem. He realized (by interpolating between Wien’s rule and
the Rayleigh–Jeans rule) that he could account for the observed curve if

the oscillators in the Rayleigh-Jeans classical approach not to radiate energy continuously, as the classical
theory would demand, but they could only lose or gain energy in integer multiples of 𝒉𝝂, for an oscillator of
frequency 𝝂.

The constant ℎ is now called Planck’s constant,

ℎ = 6.626 × 1034 𝑗. 𝑠
Thus, according to Planck’s Postulate, the energy of each oscillator is given by
𝜀𝑛 = 𝑛ℎ𝜈 , 𝑛 = 0,1,2, …

So, assuming that the energy of an oscillator is quantized, the correct thermodynamic relation for the average energy
be obtained by replacing the integration of (3)—that corresponds to an energy continuum—by a discrete summation
corresponding to the discreteness of the oscillators’ energies
∞ 𝑛ℎ𝜈
∫0 𝜀 𝑒 −𝜀/𝑘𝑇 𝑑𝜀 ∑∞ −
0 𝑛ℎ𝜈𝑒 𝑘𝑇
𝜀̅ = ∞ ⟼ 𝑛ℎ𝜈
∫0 𝑒 −𝜀/𝑘𝑇 𝑑𝜀 ∑∞ − 𝑘𝑇
0 𝑒
⇊ . ⇊
ℎ𝜈
(𝑘𝑇) ⟼ ( ℎ𝜈 )
𝑒 −1
𝑘𝑇

which gives

h
 
e h / kT  1
Then the energy density takes the form

8𝜋ℎ 𝜈3
𝑢(𝜈) = 3 ℎ𝜈 (6)
𝑐
𝑒 𝑘𝑇 − 1
This is known as Planck’s distribution. It gives an exact fit to the various experimental radiation distributions, as
displayed in Figure . The numerical value of h is obtained by fitting (6) with the experimental data .
 At high frequencies ,

ℎ𝜈 ≫ 𝑘𝑇: 𝑒 ℎ𝜈/𝑘𝑇 → ∞ => 𝑢(𝜈) → 0

As observed: No more ultraviolet catastrophe !

 8

 At low frequencies, ℎ𝜈 ≪ 𝑘𝑇, where RJ formula is a good approximation to the data, use

x2
ex  1  x   ... and when x  1 , e x  1  x..
2!

h 1 1 kT
Since  1 then h / kT   .
kT e 1 h h
1 1
kT

Thus , at low freq. Planck’s formula becomes

8 h 3 kT 8 kT 2
u( )d   ( ) d   d
c3 h c3
which is RJ formula!

We can find the total energy density in terms of Stefan-Boltzmann’s total power per unit surface area
∞
8𝜋ℎ ∞ 𝜈3
𝑈(𝑇) = ∫ 𝑢(𝜈, 𝑇)𝑑𝜈 = ∫ ℎ𝜈 𝑑𝜈
0 𝑐3 0
𝑒 𝑘𝑇 −1
Use a change of variable

ℎ𝜈
𝑥=
𝑘𝑇
=𝜋4 /15

8𝜋(𝑘𝑇)4 ⏞∞ 𝑥 3 8𝜋 5 𝑘 4 4 4 4
𝑈(𝑇) = ∫ 𝑑𝑥 = 𝑇 = 𝜎𝑇
(ℎ𝑐)3 0 𝑒 𝑥 − 1 15(ℎ𝑐)3 𝑐

where

2𝜋 5 𝑘 4
𝜎= = 5.67 × 10−8 𝑊𝑚−2 𝐾 −4 ∶ 𝑆𝑡𝑒𝑓𝑎𝑛 − 𝐵𝑜𝑙𝑡𝑧𝑚𝑎𝑛𝑛 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
15(ℎ𝑐)3

Therefore,

4
𝑈(𝑇) = 𝑃
𝑐

In this way, Planck’s relation (6) leads to a finite total energy density of the radiation emitted from a blackbody, and
hence avoids the ultraviolet catastrophe.
 9

To establish the Wien’s relation, note that

1 −ℎ𝜈
lim ( ℎ𝜈 ) → 𝑒 𝑘𝑇
ℎ𝜈
≫1 𝑒 𝑘𝑇 − 1
𝑘𝑇

Then,
8𝜋ℎ𝜈 3 −ℎ𝜈
𝑢(𝜈)𝑑𝜈 ≈ 𝑒 𝑘𝑇 𝑑𝜈
𝑐3

Problem 2.1 To see that this relation is in the form of Wien’s, write
𝑐 1
𝑑𝜈 = |𝑑 ( )| = 𝑐 2 𝑑𝜆
𝜆 𝜆
Thus,
8𝜋ℎ𝑐 3 −ℎ𝑐 1 8𝜋ℎ𝑐 −ℎ𝑐
𝑢(𝜈)𝑑𝜈 → 3 3 𝑒 𝑘𝑇𝜆 (𝑐 2 𝑑𝜆) => 𝑢(𝜆)𝑑𝜆 = 5 𝑒 𝑘𝑇𝜆 𝑑𝜆
𝑐 𝜆 𝜆 𝜆
Or with
−ℎ𝑐
𝑊(𝜆𝑇) ≡ 8𝜋ℎ𝑐 𝑒 𝑘𝑇𝜆
We get
𝑊(𝜆𝑇)
𝑢(𝜆)𝑑𝜆 𝑑𝜆 ∶ 𝑊𝑖𝑒𝑛′ 𝑠𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛
𝜆5
 10

THE PHOTOELECTRIC EFFECT

The photoelectric effect provides a direct confirmation for the energy

quantization of light.

When light is incident on a certain metallic surface, under some

conditions, it ejects electrons from the surface. This phenomenon is called
the photoelectric effect and electrons ejected from the surface in this way
are called photoelectrons

The experimental setup that exhibits the photoelectric effect is as follows:

 An evacuated tube contains two electrodes connected to a

source of variable voltage V.
 Surface of the metal plate is irradiated as the cathode.
 Electrons reaching anode constitute a current called
photoelectric current measured by ammeter A

Experimental results

I) Curves show how the current i varies as V and light intensity I for a fixed value of frequency.

 The photoelectric current 𝑖 depends on the anode

potential.
 The photoelectric current 𝑖 is proportional to the
intensity of light I.
 When all the photoelectrons reach the anode, a
further increase in V does not cause any further
increase in current i,( This is why the graph lines
become horizontal on the right side of the figure.)
This maximum value of the photoelectric current is
known as saturation current .
 Even if V=0, 𝑖 ≠ 0: some of the photoelectrons are emitted with enough energy to reach anod.
 The 𝑖 steadily decreases as V becomes increasingly negative until at the V0, called stopping potential.

1
K max  mv 2  eVStop. : Energy of the fastest electrons
2
 11

II) KE as a function of the frequency 𝜈.

 KE is proportional to the frequency 𝜈, but NOT intensity

I.
 Photoelectric effect is not observed below a certain cut-
off frequency  0

The value of 𝜈0 depends on the type of metal from which the

cathode is made . F.Ex.

𝜈0 ≅ 0.38 × 1015 𝐻𝑧 ∶ 𝑓𝑜𝑟 𝑝𝑜𝑡𝑎𝑠𝑠𝑖𝑢𝑚

𝜈0 ≅ 1.75 × 1015 𝐻𝑧 ∶ 𝑓𝑜𝑟 𝑝𝑙𝑎𝑡𝑖𝑛𝑢𝑚

Classical expectation is in sharp disagreement with this.

 The existence of a cut-off frequency 𝜈0 below which no photoelectric effect is observed can not be
explained by the classical EM theory of light, since according to this theory photoelectric effect
should occur at any 𝜈 of the incident light provided that light is intensive enough.

 According to classical EM theory;

Incident EM radiation (light) is considered as wave of

oscillation frequency  and electric field amplitude E. In this
picture, electrons are ejected from the surface as a result of an
interaction between the oscillating electric field E of the EM
waves and the electric charge of electrons. Thus;

Kinetic Energy ∝ |Amplitude of oscillation|2 = |Amplitude of electric field|2 = |𝐸|2 = 𝐼

,which gives the light intensity I.

 12

THE QUANTUM THEORY OF LIGHT

In 1905, A. Einstein resolved the paradox through the formulation of the photon theory and he received
the Nobel Prize in 1921 for this work.
He generalized the Planck’s notation that it was the absorption and emission of radiation that occurred in
quanta. Instead, Einstein proposed that radiation itself consisted of quanta of energy.

Basic Postulates of the Einsteins Photon Theory:

 EM radiation consists of zero rest mass particles called photons that travel at the speed of light c, with
energy 𝜀 = ℎ𝜈.
 Energy content of EM radiation of frequency 𝜈 in a radiant source can only be an integral multiple of
the quantity ℎ𝜈: 𝐸𝑛 = 𝑛ℎ𝜈 .

The photoelectric effect by the Photon Theory

When striking the metal surface, a photon interacts with

a single electron and ejects it from the surface.

Keep the frequency of light constant, but Increase the intensity of light:
=> increase the number of photons => number of photoelectrons will increase too
=> photoelectric current 𝒊 increases

Keep the intensity of light constant, but increase the frequency of light:
=> the energy of each photon 𝜀 = ℎ𝜈 will increase
=> the photoelectrons will have more energy

The Explanation of the existence of a critical frequency 𝜈0 in Photon Theory:

According to the Einstein’s formulation, the photoelectric effect in a given material should obey the
equation:
𝐾𝑚𝑎𝑥 = ℎ𝜈 − Φ ,
Φ ≡ ℎ𝜈0 : The work function, which is the minimum energy for an electron to escape from a surface (or
else electrons would pour out all the time)

Through the photon concept, the interaction between surface electrons and incident EM
radiation is a particle-particle interaction.
 If ℎ𝜈 > Φ then electron absorb the photon and partially use the photon energy to overcome the work
function. Then rest of the energy appears as the kinetic energy for the electron.
 If ℎ𝜈 < Φ , implies that 𝐾𝑚𝑎𝑥 < 0 , electrons will never be ejected from the surface: Photoelectric
effect is not observed below a certain cut-off frequency 𝜈0 .

The Photoelectric effect Experiment provides an important confirmation of the particle nature of radiation
 13

ATOMIC SPECTRA

When an atomic gas is suitably “exited” it gives off light. The emitted light is then passed through a prism
which separates it into its constituent wavelengths. The result is a spectrum which contains certain specific
wavelengths only. A spectrum of this sort is called a line spectrum.

Spectral Series
In 1885, J.J. Balmer discovered that the wavelengths in the visible part of the hydrogen spectrum showed
regularity.
These regular sets are now called spectral series. The figure below shows the Balmer series.

The line with the longest 𝜆 = 656.2 𝑛𝑚(= 656.2 × 10−9 𝑚) is specified H𝛼 , the next with 𝜆 =
486.1 𝑛𝑚 (= 486.1 × 10−9 𝑚) is specified H𝛽 , and so on.

Balmer’s formula for the wavelengths of this (which were obtained by trial and error) is

1 1 1
= 𝑅 ( 2 − 2) 𝑛 = 3,4,5, …
𝜆 2 𝑛
where
𝑅 = 1.097 × 107 𝑚−1 ∶ Rydberg constant

The H𝛼 line corresponds to n=3, the H𝛽 line to n=4, and so on. The series limit corresponds to n=∞, so that
it occurs at a 𝜆 = 4/𝑅, in agreement with experiment.
Balmer series contains 𝜆′s in the visible portion of the H-spectrum. The spectral lines of H in the ultraviolet
and infrared regions fall into several other series.

In the ultraviolet the Lyman series contains the 𝜆′𝑠 given by the formula
1 1 1
= 𝑅 ( 2 − 2 ) 𝑛 = 2,3,4, … Lyman
𝜆 1 𝑛
In the infrared,
1 1 1
= 𝑅 ( 2 − 2 ) 𝑛 = 4,5,6 … Paschen
𝜆 3 𝑛
1 1 1
= 𝑅 ( 2 − 2 ) 𝑛 = 5,6,7, … Brackett
𝜆 4 𝑛
 14

The Bohr Model: Niels Bohr (1913) developed a model of atomic structure to explain all these features of atomic
spectra of hydrogen atom by combining
Rutherford’s planetary model + Planck’s quantum hypothesis + Einstein’s photon concept
+ some postulates

 Electrons move about the nucleus only in circular,

discrete orbits (known as stationary states) under the
influence of the Coulomb attraction between the
electrons and the nucleus, obeying the laws of
classical mechanics. Atoms can exist only in certain
stable states with definite energies: E1, E2, E3, etc.
 The allowed (stationary) orbits correspond to those
for which the orbital angular momentum of the
electron is an integer multiple of
ℏ ≡ ℎ/2𝜋:
𝐿𝑛 = 𝑚𝑣𝑟𝑛 = 𝑛ℏ , 𝑛 = 1,2,3, …

 As long as an electron remains in a stationary orbit (despite the fact that it is constantly
accelerating,) it does not radiate electromagnetic energy, thus, its total energy e remains
constant. Emission or absorption of electromagnetic radiation can take place only when an
electron, initially moving in an orbit of total energy Ei, jumps to another orbit with total energy
Ef. .The frequency of the emitted or absorbed radiation 𝜈 is equal to

𝐸𝑓 – 𝐸𝑖
𝜈=
ℎ
 15

ENERGY LEVELS AND SPECTRA

H-atom from Bohr’s postulate:

Electron orbits in circular orbit under Coulomb attraction

e2
F  me aR ,
4 0r 2
𝑎𝑅 = 𝑣 2 /𝑟: orbital acceleration
e : magnitude of the charge of the electron

Then the electron velocity is related to its orbit radius r:

𝑒
𝑣= ∶ 𝐞𝐥𝐞𝐜𝐭𝐫𝐨𝐧 𝐯𝐞𝐥𝐨𝐜𝐢𝐭𝐲
√4𝜋𝜖0 𝑚𝑒 𝑟
From the electron’s angular momentum quantization
ℎ
𝐿𝑛 = 𝑛 = 𝑚𝑣𝑟 , 𝑛 = 1,2,3, …
2𝜋
we also find
ℎ
𝑣=𝑛
2𝜋𝑚𝑟
Therefore
𝑠𝑜𝑙𝑣𝑖𝑛𝑔 𝑓𝑜𝑟 𝑟 𝑔𝑖𝑣𝑒𝑠
ℎ 𝑒 𝑛 2 ℎ 2 𝜖0
𝑛 = ⏞
=> 𝑟𝑛 = , 𝑛 = 1,2,3, …
2𝜋𝑚𝑟 √4𝜋𝜖0 𝑚𝑒 𝑟 𝜋𝑚𝑒 2

This is a quantized expression for the radius. Here 𝑟𝑛 represents the radius of the nth orbit. Integer n
is called the quantum number of the orbit.
The radius of the innermost orbit is customarily called the Bohr radius of the H-atom and is denoted
by the symbol 𝑎0
ℎ 2 𝜖0
𝑎0 = 𝑟1 ≡ = 5.292 × 10−11 𝑚
𝜋𝑚𝑒 2
Then, the other radii can be written as
𝑟𝑛 = 𝑛2 𝑎0
As for the total energy of the electron in H-atom;
1 2
𝑒2
𝐸 = 𝐾 + 𝑉 = 𝑚𝑒 𝑣 −
2 4𝜋𝜖0 𝑟𝑛
In deriving this relation, we have assumed that the nucleus, i.e., the proton, is infinitely heavy
compared with the electron and hence it can be considered at rest; that is, the energy of the electron–
proton system consists of the kinetic energy of the electron plus the electrostatic potential energy.

Substituting 𝑣 gives
1 𝑒2
𝐸𝑛 = −
2 4𝜋𝜖0 𝑟𝑛
Substituting 𝑟𝑛 , we find
𝑚𝑒 4 1 𝐸1
𝐸𝑛 = − 2 2
( 2) = 2 , 𝑛 = 1,2,3, …
8𝜖0 ℎ 𝑛 𝑛
Where
 16

𝑚𝑒 4
𝐸1 = −
8𝜖0 2 ℎ2

Note that
 Since the radius is quantized so does energy.
 E<0. This holds for every atomic electron and reflects the fact that it is bound to nucleus.

The energy 𝐸𝑛 of each state of the atom is determined by the value of the quantum number n.
The energies specified by this equation are called the energy levels of the H-atom and are plotted in
figure below, (where, by convention, the energy levels are shown as horizontal lines).

The energies specified by this equation are called the energy levels of the H-atom and are plotted in figure
below, (where, by convention, the energy levels are shown as horizontal lines).

 Energy levels are all negative, which indicates

that the electrons are bounded to the H-atom.
 When n=1,
𝑚𝑒 4
𝐸1 ≡ − = −2.18 × 10−18 𝐽
8𝜖0 2 ℎ2
= −13.6 𝑒𝑉
The lowest energy level 𝐸1 is called the
ground state of the atom, and the higher
levels 𝐸2 , 𝐸3 , 𝐸4 , … are called excited states.
 As 𝑛 increases, the energy level separation
decreases rapidly. Since 𝑛 can take all integral
values from 𝑛 = 1 to 𝑛 = ∞, the energy
spectrum of the atom contains an infinite
number of discrete energy levels.
 As n increases 𝐸𝑛 approaches to zero. In the limit
of 𝑛 = ∞ , 𝐸∞ = 0. Then, the electron is no
longer bound to the nucleus to form an atom.
When E>0, electron is free and its energy is no
longer quantized. ( Such a combination does not
constitute an atom, of course). This situation is
represented in Figure 1.12 by the continuum of
positive energy states. The work needed to
remove an electron from an atom in its ground
state is called its ionization energy. It is equal to
−𝐸1 . For H-atom, it is 13.6 eV.
 17

Origin of Line Spectra

Is there any experimental test of the Bohr model? H-atom exhibits line spectra in both absorption and
emission. Does it follow from this model?
In Bohr model, there are discrete energy levels in H-atom.
When an electron in an excited state drops to a lower state, the
lost energy is emitted as a single photon. If the quantum
number of the initial state is 𝑛𝑖 and the final one is 𝑛𝑓 , then the
photon will have the energy given by
𝐸 = ℎ𝜈 = 𝐸𝑖 − 𝐸𝑓
where 𝜈 is the frequency of the emitted photon. Writing
energies explicitly as,
𝐸1 𝐸1 1 1
𝐸𝑖 − 𝐸𝑓 = ( 2 ) − ( 2 ) = 𝐸1 ( 2 − 2 )
𝑛𝑖 𝑛𝑓 𝑛𝑖 𝑛𝑓
𝑚𝑒 4
(Note that 𝐸1 = − 8𝜖 2 ℎ2
< 0) we find
0
𝐸𝑖 − 𝐸𝑓 𝐸1 1 1
𝜈= = ( 2 − 2)
ℎ ℎ 𝑛𝑖 𝑛𝑓
Since 𝜆 = 𝑐/𝜈
1 𝐸1 1 1
= − ( 2 − 2 ) ∶ Hyrogen spectrum
𝜆 𝑐ℎ 𝑛𝑓 𝑛𝑖
These is the Balmer’s formula provided that
−𝐸1 ?
=
⏞ 𝑅
𝑐ℎ
𝑚𝑒 4
𝐸1 =
8𝜖0 2 ℎ2
4 (9.1 × 10 𝑘𝑔)(1.6 × 10−19 𝐶)4
−31
−𝐸1 𝑚𝑒 1
= 2 2
= −12 2 2 2 8 −34 3
= 1.097 × 107 𝑚−1 = 𝑅,
𝑐ℎ 8𝜖0 ℎ 𝑐ℎ 8(8.85 × 10 𝐶 /𝑁𝑚 ) (3 × 10 𝑚/𝑠)(6.63 × 10 𝐽. 𝑠)

which is indeed the same as the Rydberg constant R. Therefore we can rewrite the hydrogen spectrum that
follows from the Bohr’s postulate
1 1 1
= 𝑅 ( 2 − 2)
𝜆 𝑛𝑓 𝑛𝑖
To produce the Balmer series, we take 𝑛𝑓 = 2, and
1 1 1
𝑛𝑓 = 2 ∶ = 𝑅 ( 2 − 2 ) 𝑛𝑖 = 3,4,5 … Balmer
𝜆 2 𝑛𝑖

Bohr model of H-atom is therefore in accord with the spectral data .

The drawbacks of Bohr's model :
1. It can only explains the origin of spectrum in one electron systems ( hydrogen, He+, Li+, Be+,etc) but not the
origin of spectrum in multi-electron systems like He , Li, Be, etc.
2. It violates the Heisenberg Uncertainty Principle. The Bohr atomic model theory considers electrons to have
both a known radius and orbit i.e. known position and momentum at the same time, which is impossible
according to Heisenberg.
3. Bohr's model did not reproduce the fine or hyperfine structure of electron levels. (Zeeman effect (spectral
lines are split due to external magnetic field) and Stark Effect (spectral lines are split due to external electric
field) couldn’t be described by the model)
4. Bohr model was successful at explaining why atoms emit line spectra, but it was largely of an ad hoc nature.
Assumptions were made so that the theory would agree with the experiment. However, Bohr could give no
reason why the orbits (and the angular momentum ) were quantized.
Quantum mechanics has completely replaced Bohr's model, and is in principle exact for all atoms.
 18

WAVE PROPERTIES OF PARTICLES

Particle properties of light Wave properties of light.

(photoelectric effect experiment) (Young’s double slit experiment)

Wave-Particle Duality
of electromagnetic radiation (light).

The De Broglie Hypothesis

Luis De Broglie suggested that this wave–particle duality was not restricted to light, but must be universal:
All material particles should also display a dual wave–particle behavior.
That is, the wave–particle duality present in light must also occur in matter:
Just as a photon has a light wave associated with it that governs its motion, so a material particle (e.g. an
electron) has an associated matter wave that govern its motion.

From Relativity
𝑟𝑒𝑠𝑡
𝑚𝑎𝑠𝑠=0
𝐸
𝐸 = 𝛾𝑚𝑐 2 = (𝑝𝑐)2 + ⏟
⏟ (𝑚𝑐 2 )2 ⏞
⟹ 𝑝= ∶ momentum of a photon (*)
√𝐾𝑖𝑛𝑒𝑡𝑖𝑐 𝑅𝑒𝑠𝑡 𝑚𝑎𝑠𝑠
𝑐
𝑒𝑛𝑒𝑟𝑔𝑦 𝑒𝑛𝑒𝑟𝑔𝑦
𝑡𝑒𝑟𝑚 𝑡𝑒𝑟𝑚

From Photoelectric Effect

ℎ𝑐 𝐸 ℎ
𝐸 = ℎ𝜈 = ⟹ = ∶ wavelength-energy relation (**)
𝜆 𝑐 𝜆

ℎ
From (∗)and (∗∗) ⟹ we get 𝜆=
∶ for a photon
𝑝
? 𝒉 𝒉
𝐓𝐡𝐞 𝐝𝐞 𝐁𝐫𝐨𝐠𝐥𝐢𝐞 𝐡𝐲𝐩𝐨𝐭𝐡𝐞𝐬𝐢𝐬 ∶ 𝝀 =
⏞ = ∶ for an electron
𝒑 𝒎𝒗
 19

Davisson-Germer Experiment: Experimental confirmation of de Broglie hypothesis.

If electrons propagate as waves, they should exhibit interference.

In order to observe the interference and diffraction effects, which are characteristics of the wave properties,
some geometric parameters of the instruments such as slits must have dimensions comparable to the
wavelength of the electrons.

Let us estimate the wavelength of nonrelativistic electron with m=9.1x10-31 kg.

Its kinetic energy given by
p2
𝐾= ⇒ 𝑝 = √2𝑚𝐾
2𝑚
so that we can express its wavelength (from De Broglie relation) as
ℎ ℎ𝑐
𝜆= =
𝑝 √2(𝑚𝑒 𝑐 2 )𝐾
we find
2
2 −31
108 𝑚 −15
𝑚 2
𝑚𝑒 𝑐 ≡ 𝐸 = (9.1 × 10 𝑘𝑔) (3 × ) = 81.9 × 10 [𝑘𝑔 ( ) ]
𝑠 ⏟ 𝑠
𝐽
Let us convert it into eV:
2
81.9 × 10−15 𝐽
𝐸 = 𝑚𝑒 𝑐 = = 0.511 × 106 𝑒𝑉
1.6 × 10−19 𝐽/𝑒𝑉

Then
𝑚
ℎ𝑐 (6.626 × 10−34 𝐽. 𝑠)/(1.6 × 10−19 𝐽/𝑒𝑉)(3 × 108 𝑠 )
𝜆= =
√2(𝑚𝑒 𝑐 2 )𝐾 √2(0.511 × 106 𝑒𝑉)𝐾

≡1 Å
−7
12.3 × 10 𝑚 𝑒𝑉 12.3 × ⏞
10−10 𝑚 12.3 Å
𝜆= = √𝑒𝑉 => 𝜆 = √𝑒𝑉
√106 𝑒𝑉 𝐾 √𝐾 √𝐾
K(eV) 1 10 100 1000
𝜆(𝐴° ) 12.3 3.89 1.23 0.39

The values of 𝜆’s are comparable to the spacing of atoms in a crystal lattice. Thus, a
crystal can be used as a grating. Experiments of this type were performed in 1927 by
Davisson and Germer by using electron beams
 20

The Experiment

What Davisson and Germer found was that, although the electrons are scattered in all directions from the
crystal, the intensity reaches a maximum for electrons with an energy of 54 eV at an angle of 𝜃 =50o .
This can only result from a constructive interference of the scattered electrons.
Davisson and Germer concluded that set of regularly spaced atomic
planes within the crystal work as a grating so that the intensity peak is
due constructive interference of the scattered electrons from this
grating.
In fact, the intensity maximum of the scattered electrons in the
Davisson–Germer experiment corresponds to the first maximum (n = 1)
of the Bragg formula, follows from

The Bragg condition: If the path difference is equal to an integral

number of 𝜆, then constructive interference occurs.

The path difference is 2ℓ , where

2ℓ = 2𝑑 cos(90 − 𝜑) = 2𝑑 sin 𝜑
Thus, The Bragg condition gives
𝑛𝜆 = 2𝑑 sin 𝜑
where 𝑑 is the spacing between the planes involved in the
diffraction, 𝜑 is the angle between the incident ray and the
crystal’s reflecting planes, while the angle 𝜃 is measured in the
experiment.

The spacing between the planes involved in the diffraction is

d=0.91 Å.

Let’s calculate the experimental wavelength 𝜆𝑒𝑥𝑝 :

The angle 𝜃 is measured in the experiment. The angle of incidence is
𝜃 50
= 90 − 𝜑 => 𝜑 = 90 − => 𝜑 = 650
2 2

For the first order reinforcement, where n=1, Bragg condition is

𝜆𝑒𝑥𝑝 = 2𝑑 sin 𝜑 = 2(0.91 Å) sin 65 ⇒ 𝜆𝑒𝑥𝑝 = 1.65Å

 21

Now, calculate the wavelength from the de Broglie relation, 𝜆𝑑𝑏 :

12.3 Å 12.3 Å
𝜆𝑑𝑏 = √𝑒𝑉 = √𝑒𝑉 ⇒ 𝜆𝑑𝑏 = 1.65Å
√𝐾 √54 𝑒𝑉

Close agreement with the experimental result !

Thus, the Davisson and Germer experiment directly verifies de Broglie’s hypothesis of the wave nature
of moving bodies.

Inspired by de Broglie’s hypothesis, verified by the Davisson and Germer experiment, Schrödinger
constructed the theory of wave mechanics which deals with the dynamics of microscopic particles.
He described the motion of particles by means of a wave function (𝒓 ⃗ , 𝒕) , which corresponds to the de
Broglie wave of the particle.
We will deal with the physical interpretation of 𝜓(𝑟, 𝑡) in the chapter 1.

These lecture notes are not original; they contain lots of material from several different sources
 1

CH 3

FORMALISM

In CH 1 and CH 2, we have seen a number of interesting properties of simple

quantum systems, like the distinct energy states of the harmonic oscillator that are
evenly spaced or, a more general one; the uncertainty principle satisfied between
the position and the momentum of a free particle.

The purpose of this chapter is to cast the quantum theory in a more powerful and
general form. There is not much here that is truly new, the idea rather is to
generalize what we have already discovered in particular cases.

Quantum theory is based on two constructs: wave functions and operators.

1. Operators: In quantum mechanics, observables are represented by operators.

In general, an operator is a rule for turning one function into another. If 𝐴̂ is an
operator and 𝑓(𝑥) is a function then
𝐴̂𝑓(𝑥 ) = 𝑔(𝑥)
is also a function.
The operators of quantum theory are linear operators
If 𝐴̂ is a linear operator, it satisfies the following rules for two functions 𝑓(𝑥 )
and ℎ(𝑥 ):
𝐴̂(𝑓 (𝑥 ) + ℎ(𝑥 )) = 𝐴̂𝑓 (𝑥 ) + 𝐴̂ℎ(𝑥)
𝐴̂(𝑐𝑓(𝑥 )) = 𝑐𝐴̂𝑓(𝑥 )
 2

EX:

𝜕 𝜕𝑓(𝑥 ) 𝜕ℎ(𝑥)
̂=
𝐷 such that ̂ 𝑓 (𝑥 ) =
𝐷 ̂ ℎ(𝑥) =
and 𝐷
𝜕𝑥 𝜕𝑥 𝜕𝑥
is a linear operator such that
𝜕𝑓(𝑥 ) 𝜕ℎ(𝑥 )
̂ (𝑓(𝑥 ) + ℎ(𝑥) =
𝐷 + =𝐷̂ 𝑓(𝑥) + 𝐷
̂ ℎ(𝑥)
𝜕𝑥 𝜕𝑥
̂ (𝑐𝑓(𝑥 )) = 𝑐𝐷
𝐷 ̂ 𝑓(𝑥)

But the operator 𝑂̂ ≡ √ such that

𝑂̂𝑓(𝑥) = √𝑓(𝑥)
is not:

𝑂̂ (𝑓(𝑥 ) + ℎ(𝑥 )) = √𝑓(𝑥) + ℎ(𝑥 ) ≠ √𝑓(𝑥) + √ℎ(𝑥)

2. Wave functions In quantum mechanics, the state of a system is represented by its
wave function. Mathematically, wave functions satisfy the defining conditions for
abstract vectors. But, the “vectors” in quantum mechanics are functions and they live
in an infinite dimensional abstract linear space of square integrable functions.
Physicists call this space Hilbert space.

The formalism of quantum mechanics deals with operators that are linear and
wave functions that belong to an abstract Hilbert space.

Thus, the structure and mathematical properties of Hilbert spaces and the linear
operators are essential for a proper understanding of the formalism of quantum
mechanics and the natural language of quantum mechanics is therefore the linear
algebra.

For this, we are going to review briefly the properties of Hilbert spaces and those of
linear operators. Then we will introduce a new notation: Dirac’s bra-ket notation
 3

Hilbert Space
A Hilbert space ℋ consists of a set of vectors 𝜙1 , 𝜙2 , … and a set of scalars 𝑎, 𝑏, 𝑐, …
which satisfy the following properties:
(a) 𝓗 is a linear space:
If 𝜙1 , 𝜙2 ∈ ℋ, and 𝑎 is scalar

 𝑎 𝜙1 ∈ ℋ

 (𝜙1 + 𝜙2 ) ∈ ℋ

 Commutativity: 𝜙1 + 𝜙2 = 𝜙2 + 𝜙1

 Associativity : (𝜙1 + 𝜙2 ) + 𝜙3 = 𝜙1 + (𝜙2 + 𝜙3 )

(b) 𝓗 has a defined scalar product.

The scalar product of an element 𝜙1 with another element 𝜙2 is in general a
complex number, denoted by
⟨𝜙1 |𝜙2 ⟩ = 𝑎 = 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑛𝑢𝑚𝑏𝑒𝑟
The scalar product obtained by exchanging the role of 𝜙1 and 𝜙2 results in the
complex conjugate of that number
⟨𝜙1 |𝜙2 ⟩ = ⟨𝜙2 |𝜙1 ⟩∗ = 𝑎∗ = 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑐𝑜𝑛𝑗𝑢𝑔𝑎𝑡𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑛𝑢𝑚𝑏𝑒𝑟
(c) 𝓗 is separable and complete.

Every Cauchy sequence {𝜙𝑛 } of functions in it converges to an element 𝜙𝑚 of

ℋ=> it has no hole in it such that

‖𝜙𝑛 − 𝜙𝑚 ‖ ⟶ 0 𝑎𝑠 𝑛, 𝑚 ⟶ ∞
 4

The dimension of a vector space is given by the maximum number of linearly

independent vectors the space can have. For instance, if the maximum number of
linearly independent vectors that a space has is N (i.e., 𝜙1 , 𝜙2 , … , 𝜙𝑁 ), this space is
said to be N-dimensional.

Linearly independent Functions:

EX:

(a) Consider the set#1: {𝑓1 , 𝑓2 } = {𝑆𝑖𝑛𝑘𝑥, 𝑒 𝑖𝑘𝑥 }

𝑖𝑘𝑥
𝑒 𝑖𝑘𝑥 − 𝑒 −𝑖𝑘𝑥 𝑎1 𝑎1
𝑎1 𝑆𝑖𝑛𝑘𝑥 + 𝑎2 𝑒 = 0 = 𝑎1 ( ) + 𝑎2 𝑒 𝑖𝑘𝑥 = 𝑒 𝑖𝑘𝑥 ( + 𝑎2 ) + 𝑖𝑒 −𝑖𝑘𝑥 = 0
2𝑖 2𝑖 2

=> 𝑎1 , 𝑎2 = 0

(b)Consider the set#2: {𝑓1 , 𝑓2 } = {3𝑒 𝑖𝑘𝑥 , 𝑒 𝑖𝑘𝑥 }

𝑎1 3𝑒 𝑖𝑘𝑥 + 𝑎2 𝑒 𝑖𝑘𝑥 = 0 = 𝑒 𝑖𝑘𝑥 (𝑎1 3 + 𝑎2 ) => 𝑎2 = −3𝑎1

 5

Basis of a vector space:

It consists of a set of the maximum possible number of linearly independent vectors
belonging to that space. If the dimension of space is N, =the maximum number of
linearly independent vectors in this space. Then, the set of N vectors, 𝜙1 , 𝜙2 , … , 𝜙𝑁 ,
to be denoted in short by {𝜙𝑖 }, is called the basis of the vector space, while the
vectors 𝜙1 , 𝜙𝑁 , … , 𝜙𝑁 are called the base vectors. It is convenient to choose them
orthonormal; that is, their scalar products satisfy the relation
⟨𝜙𝑖 |𝜙𝑗 ⟩ = 𝛿𝑖𝑗
Moreover, the basis is said to be complete if it spans the entire space; that is, there is
no need to introduce any additional base vector. In this N-dimensional vector space,
an arbitrary vector 𝜌 can be expanded as a linear combination
𝑁

𝜌 = ∑ 𝑎𝑖 𝜙𝑖 (∗)
𝑖=1

The expansion coefficients 𝑎𝑖 are called the components of the vector 𝜌 in the basis
{𝜙𝑖 }. Each component is given by the scalar product of 𝜌 with the corresponding
base vector:

𝑎𝑖 = ⟨𝜙𝑖 |𝜌⟩
 6

Examples of Hilbert space

(1) One of the most familiar example of a Hilbert space is the Euclidean vector
space consisting of three-dimensional vectors, denoted by R3, and equipped with
the scalar product. (A finite dimensional
Euclidean space is in many ways like an infinite
dimensional Hilbert space but there are some ways in
which the infinite dimensionality leads to subtle
differences we need to be aware of.)

Every point in three-dimensional

Euclidean space is determined by
three coordinates
(𝑥, 𝑦, 𝑧).
There a vector 𝐴⃗ is a set of three
numbers, called components
(𝐴𝑥 , 𝐴𝑦 , 𝐴𝑧 ).

Any vector in this space can be expanded in terms of the three unit vectors 𝑒̂𝑥 ,𝑒̂𝑦 , 𝑒̂𝑧 ,
(or 𝑖̂, 𝑗̂, 𝑘̂ ) which is called a basis with the property:

(𝑒̂𝑥 , 𝑒̂𝑦 , 𝑒̂𝑧 ): basis vectors of space

0 𝑖𝑓 𝑖 ≠ 𝑗
𝑒̂𝑖 ∙ 𝑒̂𝑗 = 𝛿𝑖𝑗 = {
1 𝑖𝑓 𝑖 = 𝑗
The vectors 𝑒̂𝑥 ,𝑒̂𝑦 , 𝑒̂𝑧 are said to span the vector space.

Alternatively, in spherical coordinate system

(𝑒̂𝑟 , 𝑒̂𝜃 , 𝑒̂𝜙 ): basis vectors of space

 7

Any vector 𝐴⃗ of the Euclidean space can be written in terms of the base vectors as
𝐴⃗ = 𝐴𝑥 𝑒̂𝑥 + 𝐴𝑦 𝑒̂𝑦 + 𝐴𝑧 𝑒̂𝑧 = ∑ 𝐴𝑖 𝑒̂𝑖
𝑖=𝑥,𝑦,𝑧

where 𝐴𝑥 , 𝐴𝑦 , 𝐴𝑧 are the components of 𝐴⃗ in the basis (𝑒̂𝑥 , 𝑒̂𝑦 , 𝑒̂𝑧 ). They are
determined as

𝐴𝑖 = 𝐴⃗ ∙ 𝑒̂𝑖

The scalar product of two vectors 𝐴⃗ and 𝐵

⃗⃗ is defined as

𝐴⃗ ∙ 𝐵
⃗⃗ = 𝐴𝑥 𝐵𝑥 + 𝐴𝑦 𝐵𝑦 + 𝐴𝑧 𝐵𝑧 = ∑ 𝐴𝑖 𝐵𝑖
𝑖=𝑥,𝑦,𝑧

Note that the scalar product in Euclidian space is real and symmetric.
The length of the vector is

|𝐴⃗| = √𝐴⃗ ∙ 𝐴⃗

In 3-dimensinal vector space

if 𝐴⃗ and 𝐵
⃗⃗ are orthogonal then 𝐴⃗ ∙ 𝐵
⃗⃗ = |𝐴⃗||𝐵
⃗⃗ | cos(𝐴⃗, 𝐵
⃗⃗ ) = 0
 8

(2) The second example of a Hilbert space is the space of the entire square
integrable complex functions 𝑓(𝑥 ).
It is easy to verify that the space of square-integrable functions possesses the
properties of a Hilbert space. For instance, any linear combination of square-
integrable functions is also a square-integrable.
A good example of square-integrable functions is the wave function of quantum
mechanics, 𝜓(𝑥, 𝑡). We have seen that, according to Born’s probabilistic
interpretation of 𝜓(𝑥, 𝑡) the quantity |𝜓(𝑥, 𝑡) |2 𝑑𝑥 represents the probability of
finding, at time t, the particle in dx centered around the point x . The probability of
finding the particle somewhere in space must then be equal to 1:
∞

⟨𝜓|𝜓⟩ = ∫ |𝜓(𝑥, 𝑡) |2 𝑑𝑥 = 1
−∞
hence the wave functions of quantum mechanics are square-integrable. Wave
functions satisfying above eq. are said to be normalized or square-integrable. As
wave mechanics deals with square-integrable functions, any wave function which is
not square-integrable has no physical meaning in quantum mechanics.

Quantum wave functions belong to a linear vector space, a Hilbert space.

The dimension of this space is infinite since each wave function can be expanded in
terms of an infinite number of linearly independent functions.
--------------------------------------------------------------------------------------
EX: Particle in an infinite potential well.
2 𝑛𝜋 ℏ2 𝜋 2 2
𝜓𝑛 (𝑥 ) = √ sin ( 𝑥), 𝐸𝑛 = 𝑛 , 𝑛 = 1,2,3, …
𝐿 𝐿 2𝑚𝐿2
There are infinitely many 𝑛. Thus the orthonormal basis { 𝜓𝑛 } is of infinite
dimension.
---------------------------------------------------------------------------------------
 9

In this prescription; the physical state of a system is represented by elements of

Hilbert space; these elements are called state vectors.
We can represent the state vectors in different bases by means of function expansions.
Let, 𝑓(𝑥) ∈ ℋ and let {𝜙𝑖 } be a basis of it. Then
∞

𝑓 (𝑥 ) = ∑ 𝑎𝑖 𝜙𝑖 (∗)
𝑖=1

This is analogous to specifying an ordinary (Euclidean) vector by its components in

various coordinate systems. For instance, we can represent equivalently a vector by its
components in a Cartesian coordinate system,
𝐴⃗ = 𝐴𝑥 𝑒̂𝑥 + 𝐴𝑦 𝑒̂𝑦 + 𝐴𝑧 𝑒̂𝑧

Or in a spherical coordinate system,

𝐴⃗ = 𝐴𝑟 𝑒̂𝑟 + 𝐴𝜃 𝑒̂𝜃 + 𝐴𝜙 𝑒̂𝜙

The meaning of a vector is, of course, independent of the coordinate system chosen
to represent its components.
Similarly, the state of a microscopic system (*) has a meaning independent of the basis
in which it is expanded.
 10

DIRAC NOTATION
To free state vectors from coordinate meaning, Dirac introduced an economical and
powerful notation in quantum mechanics; it allows one to manipulate the formalism
of quantum mechanics with ease and clarity. He introduced the concepts of kets,
bras, and bra-kets.
Each state function 𝜓(𝑥) is associated with a ket |𝜓⟩:

𝜓
⏟ (𝑥 ) ⟶ |𝜓
⏟⟩
𝑎 𝑠𝑡𝑎𝑡𝑒 𝑎 𝑘𝑒𝑡
𝑓𝑢𝑛𝑐𝑖𝑜𝑛

Each complex conjugate state vector 𝜓 ∗ (𝑥) is associated with a bra ⟨𝜓|.

⏟∗ (𝑥 ) ⟶ ⟨⏟
𝜓 𝜓|
𝑎 𝑐𝑜𝑚𝑝𝑙𝑒𝑥 𝑎 𝑏𝑟𝑎
𝑐𝑜𝑛𝑗𝑢𝑔𝑎𝑡𝑒
𝑠𝑡𝑎𝑡𝑒
𝑣𝑒𝑐𝑡𝑜𝑟

Dirac denoted the scalar (inner) product by the symbol, ⟨ | ⟩, which he called a
bra-ket.

⟨⏟ | ⟩ ∶ 𝑎 𝑏𝑟𝑎-𝑘𝑒𝑡
𝑠𝑐𝑎𝑙𝑎𝑟
𝑝𝑟𝑜𝑑𝑢𝑐𝑡

For instance, the scalar product of two states vectors, 𝜓(𝑥) and 𝜙(𝑥) is denoted by
the bra-ket ⟨𝜓|𝜙⟩.

⟨𝜓|𝜙⟩: the scalar product of two states vectors, 𝜓(𝑥) and 𝜙(𝑥)

which is given by the integral

∞
⟨𝜓|𝜙⟩ = ∫ 𝜓 ∗ (𝑥 )𝜙(𝑥 )𝑑𝑥 (∗)
−∞
 11

Some properties of this notation: If 𝑎 is any complex number,

Ordinary notation Dirac Notation

𝑎 𝜓(𝑥 ) |𝑎𝜓⟩ 𝑜𝑟 𝑎|𝜓⟩
𝑎 ∗ 𝜓 ∗ (𝑥 ) ⟨𝑎𝜓| or 𝑎∗ ⟨𝜓|
𝐴̂ 𝜓(𝑥 ) |𝐴̂𝜓⟩ 𝑜𝑟 𝐴̂|𝜓⟩
𝜓+𝜙 |𝜓 + 𝜙⟩ = |𝜓⟩ + |𝜙⟩
In addition:

 ⟨𝜙|𝜓⟩∗ = ⟨𝜓|𝜙⟩
∞ ∗ ∞
⟨𝜙|𝜓 ⟩∗ = (∫ 𝜙 𝑥 )𝜓(𝑥 )𝑑𝑥 ) = ∫ 𝜓 ∗ (𝑥 )𝜙(𝑥 )𝑑𝑥 = ⟨𝜓|𝜙⟩
∗(
−∞ −∞

 𝑎⟨𝜓|𝜙⟩ = ⟨𝑎∗ 𝜓|𝜙⟩ = ⟨𝜓|𝑎𝜙⟩

∞ ∞
𝑎⟨𝜓|𝜙⟩ = ∫ 𝜓 𝑥 )𝑎𝜙(𝑥 )𝑑𝑥 = ∫ (𝑎∗ 𝜓)∗ (𝑥 )𝜙(𝑥 )𝑑𝑥
∗(
−∞ −∞

 𝑎∗ ⟨𝜓|𝜙⟩ = ⟨𝑎𝜓|𝜙⟩ = ⟨𝜓|𝑎∗ 𝜙⟩

 ⟨𝜓|𝑎𝜙1 + 𝑏𝜙2 ⟩ = 𝑎⟨𝜓|𝜙1 ⟩ + 𝑏⟨𝜓|𝜙2 ⟩

 ⟨𝑎𝜙1 + 𝑏𝜙2 |𝜓⟩ = 𝑎∗ ⟨𝜙1 |𝜓⟩ + 𝑏 ∗ ⟨𝜙2 |𝜓⟩

∞
 ⟨𝑎1 𝜓1 + 𝑎2 𝜓2 |𝑏1 𝜙1 + 𝑏2 𝜙2 ⟩ = ∫−∞(𝑎1 𝜓1 + 𝑎2 𝜓2 )∗ (𝑏1 𝜙1 + 𝑏2 𝜙2 )𝑑𝑥
∞ ∞ ∞ ∞
= 𝑎1 𝑏1 ∫ 𝜓1 ∗ 𝜙1 𝑑𝑥
∗
+ 𝑎2 𝑏2 ∫ 𝜓2 ∗ 𝜙2 𝑑𝑥
∗
+ 𝑎1 𝑏2 ∫ 𝜓1 ∗ 𝜙2 𝑑𝑥
∗
+ 𝑎2 𝑏1 ∫ 𝜓2 ∗ 𝜙1 𝑑𝑥
∗
−∞ −∞ −∞ −∞

Thus,
⟨𝑎1 𝜓1 + 𝑎2 𝜓2 |𝑏1 𝜙1 + 𝑏2 𝜙2 ⟩
= 𝑎1∗ 𝑏1 ⟨𝜓1 |𝜙1 ⟩ + 𝑎2∗ 𝑏2 ⟨𝜓2 |𝜙2 ⟩ + 𝑎1∗ 𝑏2 ⟨𝜓1 |𝜙2 ⟩ + 𝑎2∗ 𝑏1 ⟨𝜓2 |𝜙1 ⟩
 12

 A state vector 𝜙𝑖 (𝑥) is said to be normalized if its scalar product with itseif is
equal to unity:
∞ ∞
⟨ 𝜙𝑖 | 𝜙𝑖 ⟩ = ∫ 𝜙𝑖∗ (𝑥 )𝜙𝑖 (𝑥 )𝑑𝑥 = ∫ |𝜙𝑖 (𝑥 )|2 𝑑𝑥 = 1
−∞ −∞

 Two state vectors 𝜙𝑖 (𝑥) and 𝜙𝑗 (𝑥) are said to be orthogonal if their scalar
product is zero:
∞
⟨𝜙𝑖 |𝜙𝑗 ⟩ = ∫ 𝜙𝑖∗ (𝑥 )𝜙𝑗 (𝑥 )𝑑𝑥 = 0
−∞

 A set of state vectors {𝜙𝑛 } are said to be orthonormal if they are normalized
and mutually orthogonal

⟨𝜙𝑚 |𝜙𝑛 ⟩ = 𝛿𝑚𝑛

 Moreover, a set of state vectors {𝜙𝑛 } is said to be complete if it spans the entire
space; that is, there is no need to introduce any additional base vector, such that
any other function in this ∞-dimensional Hilbert space, 𝜒(𝑥) can be expanded
as a linear combination:

𝜒(𝑥) = ∑ 𝑐𝑛 𝜙𝑛 (𝑥)
𝑛=1

The expansion coefficients 𝒄𝒏 above are called the components of the vector 𝜒 in the
basis {𝜙𝑛 } and given by the scalar product of 𝜒 with the corresponding base
vector, 𝑐𝑖 = ⟨𝜙𝑖 |𝜒⟩.

𝑐𝑖 = ⟨𝜙𝑖 |𝜒⟩: 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡𝑠 of the vector 𝜒 in the basis {𝜙𝑖 }

 13

Euclidean space Space of wave functions

of quantum mechanics
Elements ⃗⃗1 , 𝑉
𝑉 ⃗⃗2 , 𝑉
⃗⃗3 , … : 3. dim 𝑣𝑒𝑐𝑡𝑜𝑟𝑠 𝜓1 , 𝜓2 , 𝜓3 , … 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛𝑠
Scalar ⃗⃗1 ∙ 𝑉
𝑉 ⃗⃗2 = ∑ 𝑉1𝑖 𝑉2𝑖 = a scalar ⟨𝜓1 |𝜓2 ⟩ = ⟨𝜓2 |𝜓1 ⟩∗
𝑖=𝑥,𝑦,𝑧
Product
= ∫ 𝜓1 ∗ 𝜓2 𝑑 3 𝑟⃗ = a scalar

Basis set 0 𝑖𝑓 𝑖≠𝑗 If {𝜑𝑛 } 𝑐𝑜𝑚𝑝𝑟𝑖𝑠𝑒 𝑎 𝑏𝑎𝑠𝑖𝑠 for

𝑒̂𝑖 ∙ 𝑒̂𝑗 = 𝛿𝑖𝑗 = {
1 𝑖𝑓 𝑖=𝑗 a Hilbert space ℋ, then
⟨𝜑𝑖 |𝜑𝑗 ⟩ ≡ ⟨𝑖|𝑗⟩ = 𝛿𝑖𝑗
The functions {𝜑𝑛 } are said to
The vectors 𝑒̂𝑥 ,𝑒̂𝑦 , 𝑒̂𝑧 are said span the Hilbert space ℋ.
to span the vector space.
Dimension(d) d=3 2≤ 𝒅 < ∞
⃗⃗ = 𝑉𝑥 𝑒̂𝑥 + 𝑉𝑦 𝑒̂𝑦 + 𝑉𝑧 𝑒̂𝑧
Completeness 𝑉 Any function |𝜓⟩ in ℋ
= ∑ 𝑉𝑖 𝑒̂𝑖
𝑖=𝑥,𝑦,𝑧 |𝜓⟩ = ∑ 𝑐𝑖 |𝜑𝑖 ⟩
𝑖

Projection ⃗⃗ ∙ 𝑒̂𝑖
𝑉𝑖 = 𝑉 𝑐𝑖 = ⟨𝜑𝑖 |𝜓⟩
Norm 1/2 1/2
√𝑉 ⃗⃗ = (∑ 𝑉𝑖2 )
⃗⃗ ∙ 𝑉 ⟨𝜓|𝜓⟩1/2 = (∑|𝑐𝑖 |2 )
𝑖 𝑖
 14

HERMITIAN OPERATORS

Consider an observable 𝑄 represented by an operator 𝑄̂ . The expectation value of

this observable is

〈𝑄 〉 = ∫ 𝜓 ∗ 𝑄̂ 𝜓 𝑑𝑥 = ⟨𝜓|𝑄̂ 𝜓⟩

Observables are represented by operators whose expectation values are real, since the
outcome of a measurement has got to be real:
!
⏞ 〈𝑄 〉∗ (∗)
〈𝑄 〉 =
But the complex conjugate of a scalar product reverses the order:
∗
〈𝑄 〉∗ = ⟨𝜓|𝑄̂ 𝜓⟩ = ⟨𝑄̂ 𝜓|𝜓⟩

From (*) we have that if 𝑄̂ represents an observable then it satisfies the condition

⟨𝜓|𝑄̂ 𝜓⟩ = ⟨𝑄̂ 𝜓|𝜓⟩ for all 𝜓

We call such operators Hermitian.

Most books require a stronger condition:

⟨𝑓|𝑄̂ 𝑔⟩ = ⟨𝑄̂ 𝑓|𝑔⟩ for all 𝑓(𝑥) 𝑎𝑛𝑑 𝑔(𝑥)

But, it is possible to show that these two definitions are equivalent (See Probl.3.3)

The Hermitian conjugate of an operator 𝑄̂ is an operator 𝑄̂ † such that

⟨𝑓|𝑄̂ 𝑔⟩ = ⟨𝑄̂ † 𝑓|𝑔⟩ 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑓 𝑎𝑛𝑑 𝑔

Therefore, if 𝑄̂ represents an observable then

𝑄̂ † = 𝑄̂
 15

EX.(1) Consider a complex number 𝑐. Find Hermitian conjugate 𝑐 † .

We should find the object 𝑐̂ † that fits eq.(*) for all 𝑓 and 𝑔 in ℋ.

1.st way:

⟨𝑐 † 𝑓|𝑔⟩ = ⟨𝑓|𝑐 𝑔⟩ (∗)

𝑅𝐻𝑆 = ⟨𝑓|𝑐 𝑔⟩ = 𝑐⟨𝑓 | 𝑔⟩ = ⟨𝑐 ∗ 𝑓|𝑔⟩

𝑅𝐻𝑆 = 𝐿𝐻𝑆 => ⟨𝑐 ∗ 𝑓|𝑔⟩ = ⟨𝑐 † 𝑓|𝑔⟩ => 𝑐 † = 𝑐 ∗

2.nd way:

⟨𝑓|𝑐 𝑔⟩ = ∫ 𝑓 ∗ (𝑐)𝑔 𝑑𝑥 = ∫(𝑐 ∗ 𝑓)∗ 𝑔 𝑑𝑥 = ⟨𝑐 ∗ 𝑓|𝑔⟩

Compare with (*)=> 𝑐 † = 𝑐 ∗

𝜕
̂=
EX.(2) Consider the operator 𝐷 ̂ † =?
defined in a ℋ. 𝐷
𝜕𝑥

̂ † 𝑓|𝑔⟩ = ⟨𝑓|𝐷
⟨𝐷 ̂ 𝑔⟩
∞
𝜕
̂ 𝑔⟩ = ∫ ⏟
𝑅𝐻𝑆 = ⟨𝑓|𝐷 𝑓∗ 𝑔 𝑑𝑥
⏟
𝜕𝑥
−∞ 𝑢 𝑑𝑣
∞ ∞
∗
𝜕 𝜕 𝜕
⏟∗ 𝑔|∞
=𝑓 −∞ − ∫ (
̂ 𝑓|𝑔⟩
𝑓 ∗ ) 𝑔 𝑑𝑥 = ∫ (− 𝑓) 𝑔 𝑑𝑥 = ⟨− 𝑓|𝑔⟩ = ⟨−𝐷
𝜕𝑥 𝜕𝑥 𝜕𝑥
=0 −∞ −∞

̂ † 𝑓|𝑔⟩ = ⟨−𝐷
𝐿𝐻𝑆 = 𝑅𝐻𝑆 => ⟨𝐷 ̂ 𝑓|𝑔⟩ => 𝐷
̂ † = −𝐷
̂
 16

†
EX: (3) Given a product operator 𝐶̂ ≡ 𝐴̂𝐵̂ . What is 𝐶̂ † = (𝐴̂𝐵̂) ?
†
We should find an object (𝐴̂𝐵̂) that fits in
†
⟨(𝐴̂𝐵̂) 𝑓|𝑔⟩ = ⟨𝑓|(𝐴̂𝐵̂)𝑔⟩

𝑅𝐻𝑆 = ⟨𝑓|(𝐴̂𝐵̂ )𝑔⟩ = ⟨𝐴̂† 𝑓|𝐵̂ 𝑔⟩ = ⟨𝐵̂† 𝐴̂† 𝑓|𝑔⟩

†
𝑅𝐻𝑆 = 𝐿𝐻𝑆 => (𝐴̂𝐵̂) = 𝐵̂† 𝐴̂†

EX: (4) Is the square of a Hermitian operator Hermitian?

† †
(𝐴̂𝐵̂ ) = 𝐵̂† 𝐴̂† => (𝐴̂𝐴̂) = 𝐴̂† 𝐴̂† = 𝐴̂𝐴̂
EX: (5) Momentum Operator.
𝜕
𝑝̂ = −𝑖ℏ
𝜕𝑥
What is 𝑝̂ † ?

⟨𝑓|𝑝̂ 𝑔⟩ = ⟨𝑝̂ † 𝑓|𝑔⟩

∞
𝜕
𝑓∗
𝐿𝐻𝑆 = ⟨𝑓|𝑝̂ 𝑔⟩ = −𝑖ℏ ∫ ⏟ 𝑔 𝑑𝑥
⏟
𝜕𝑥
−∞ 𝑢 𝑑𝑣
∞ ∞
∗
∗ |∞
𝜕 ∗ 𝜕 𝜕
= −𝑖ℏ 𝑓
⏟ 𝑔 −∞ + 𝑖ℏ ∫ 𝑔 ( 𝑓 ) 𝑑𝑥 = ∫ 𝑔 (−𝑖ℏ 𝑓) 𝑑𝑥 = ⟨−𝑖ℏ 𝜕𝑥 𝑓|𝑔⟩
𝜕𝑥 𝜕𝑥
=0 −∞ −∞

= ⟨𝑝̂ 𝑓|𝑔⟩

𝐿𝐻𝑆 = 𝑅𝐻𝑆 => ⟨𝑝̂ 𝑓|𝑔⟩ = ⟨𝑝̂ † 𝑓|𝑔⟩ => 𝑝̂ † = 𝑝̂

 17

Problem 3.4
(a) Consider two Hermitian operators
𝐴̂ † =𝐴̂
𝐴̂+ = 𝐴̂ => ⟨𝑓|𝐴̂𝑔⟩ = ⟨𝐴̂† 𝑓|𝑔⟩ =
⏞ ⟨𝐴̂𝑓|𝑔⟩
𝐵̂+ = 𝐵̂ => ⟨𝑓|𝐵̂𝑔⟩ = ⟨𝐵̂† 𝑓|𝑔⟩ =
⏟ ⟨𝐵̂𝑓|𝑔⟩
𝐵̂ † =𝐵̂

Their sum
⟨𝑓|(𝐴̂ + 𝐵̂)𝑔⟩ = ⟨𝑓|𝐴̂𝑔⟩ + ⟨𝑓|𝐵̂𝑔⟩ = ⟨𝐴̂𝑓|𝑔⟩ + ⟨𝐵̂ 𝑓|𝑔⟩ = ⟨(𝐴̂ + 𝐵̂)𝑓|𝑔⟩
+
(𝐴̂ + 𝐵̂) = (𝐴̂ + 𝐵̂)

(b) 𝑄̂ † = 𝑄̂ ∶ 𝐴 ℎ𝑒𝑟𝑚𝑖𝑡𝑖𝑎𝑛 𝑜𝑝𝑒𝑟𝑎𝑡𝑜𝑟. Under what condition 𝛼𝑄̂ is Hermitian?

(𝛼 is a complex number)

̂ is hermitian => ⟨𝑓|𝑄̂ 𝑔⟩ = ⟨𝑄̂ 𝑓|𝑔⟩

Since Q
Now,
!
If (𝛼𝑄̂ ) is hermitian => ⟨𝑓|(𝛼𝑄̂ )𝑔⟩ =
⏞ ⟨(𝛼𝑄̂ )𝑓|𝑔⟩

𝐿𝐻𝑆 = ⟨𝑓|(𝛼𝑄̂ )𝑔⟩ = 𝛼⟨𝑓|𝑄̂ 𝑔⟩

𝑅𝐻𝑆 = ⟨𝛼𝑄̂ 𝑓|𝑔⟩ = 𝛼 ∗ ⟨𝑄̂ 𝑓|𝑔⟩

𝐿𝐻𝑆 = 𝑅𝐻𝑆 if 𝛼 = 𝛼 ∗ . Then 𝛼𝑄̂ is hermitian.

 18

(c) Consider two Hermitian operators 𝐴̂ and 𝐵̂. Under what condition their product
𝐴̂𝐵̂ is Hermitian?
𝐴̂+ = 𝐴̂ => ⟨𝑓|𝐴̂𝑔⟩ = ⟨𝐴̂† 𝑓|𝑔⟩ = ⟨𝐴̂𝑓|𝑔⟩
𝐵̂ + = 𝐵̂ => ⟨𝑓|𝐵̂𝑔⟩ = ⟨𝐵̂† 𝑓|𝑔⟩ = ⟨𝐵̂𝑓|𝑔⟩
If their product is Hermitian we have
!
(𝐴̂𝐵̂) = (𝐴̂𝐵̂)† => ⟨𝑓|(𝐴̂𝐵̂)𝑔⟩ =
⏞ ⟨(𝐴̂𝐵̂)𝑓|𝑔⟩
!
⟨𝑓|(𝐴̂𝐵̂ )𝑔⟩ = ⟨𝐴̂𝑓|𝐵̂𝑔⟩ = ⟨𝐵̂𝐴̂𝑓|𝑔⟩ =
⏞ ⟨(𝐴̂𝐵̂)𝑓|𝑔⟩ =>
Therefore, if
𝐵̂𝐴̂ = 𝐴̂𝐵̂ => [𝐴̂, 𝐵̂] = 0
Then 𝐴̂𝐵̂ is Hermitian.
̂ corresponding to energy 𝐸 with
(d) Energy operator Hamiltonian operator 𝐻
̂ 𝜓ℓ = 𝐸ℓ 𝜓ℓ
𝐻
Thus the Hamiltonian operator should be a Hermitian operator
! 𝑝̂ 2
̂†
𝐻 = ̂ with
⏞ 𝐻 ̂=
𝐻 + 𝑉(𝑥)
2𝑚
From EX(3) we have
(𝑝̂ 2 )† = 𝑝̂ † 𝑝̂ † = 𝑝̂ 2
For V(x), which is a real function 𝑉̂ † =?
⟨𝑉̂ † 𝜓ℓ |𝜓𝑛 ⟩ = ⟨𝜓ℓ |𝑉̂ 𝜓𝑛 ⟩
∞ ∞

𝑅𝐻𝑆 = ∫ 𝜓ℓ∗ 𝑉 (𝑥 )𝜓𝑛 𝑑𝑥 = ∫ (𝑉 (𝑥 )𝜓ℓ )∗ 𝜓𝑛 𝑑𝑥 = ⟨𝑉̂ 𝜓ℓ |𝜓𝑛 ⟩

−∞ −∞

𝑅𝐻𝑆 = 𝐿𝐻𝑆 = ⟨𝑉̂ 𝜓ℓ |𝜓𝑛 ⟩ = ⟨𝑉̂ † 𝜓ℓ |𝜓𝑛 ⟩ => 𝑉̂ † = 𝑉̂

̂† = 𝐻
𝐻 ̂
 19

Properties of Hermitian Operators

Let {𝜓𝑛 } represent eigenfunctions a Hermitian operator 𝐴̂ in a Hilbert space ℋ with
the corresponding eigenvalues {𝑎𝑛 }
𝐴̂𝜓𝑛 = 𝑎𝑛 𝜓𝑛
𝐴̂† = 𝐴̂
OR in Dirac notation
𝐴̂|𝜓𝑛 ⟩ = 𝑎𝑛 |𝜓𝑛 ⟩ 𝑂𝑅 |𝐴̂𝜓𝑛 ⟩ = |𝑎𝑛 𝜓𝑛 ⟩ (∗)
Theorem 1: Eigenvalues of a Hermitian operator are real.i.e.,
Show that 𝑎𝑛∗ = 𝑎𝑛
Proof:
Multiplying (*) from the left with ⟨𝜓𝑛 | =>
⟨𝜓𝑛 |𝐴̂𝜓𝑛 ⟩ = ⟨𝜓𝑛 |𝑎𝑛 𝜓𝑛 ⟩ = 𝑎𝑛 ⟨𝜓𝑛 |𝜓𝑛 ⟩ (∗)
Also,
𝐿𝐻𝑆 = ⟨𝜓𝑛 |𝐴̂𝜓𝑛 ⟩ = ⟨𝐴̂† 𝜓𝑛 |𝜓𝑛 ⟩ = ⟨𝐴̂𝜓𝑛 |𝜓𝑛 ⟩ = ⟨𝑎𝑛 𝜓𝑛 |𝜓𝑛 ⟩ = 𝑎𝑛∗ ⟨𝜓𝑛 |𝜓𝑛 ⟩ (∗∗)
(∗) = (∗∗) => 𝑎𝑛∗ = 𝑎𝑛
Theorem 2: Eigenfunction of a Hermitian operator are orthogonal.i,e.,
Show that ⟨𝜓𝑙 |𝜓𝑛 ⟩ = 0 𝑤ℎ𝑒𝑛 𝑛 ≠ 𝑙
Proof:
Multiplying (*) from the left with ⟨𝜓𝑙 | =>
⟨𝜓𝑙 |𝐴̂𝜓𝑛 ⟩ = ⟨𝜓𝑙 |𝑎𝑛 𝜓𝑛 ⟩ = 𝑎𝑛 ⟨𝜓𝑙 |𝜓𝑛 ⟩ (∗∗∗)
𝐿𝐻𝑆 = ⟨𝜓𝑙 |𝐴̂𝜓𝑛 ⟩ = ⟨𝐴̂† 𝜓𝑙 |𝜓𝑛 ⟩ = ⟨𝐴̂𝜓𝑙 |𝜓𝑛 ⟩ = ⟨𝑎𝑙 𝜓𝑙 |𝜓𝑛 ⟩ =
(𝑇ℎ1)
𝑎𝑙∗ ⟨𝜓𝑙 |𝜓𝑛 ⟩ ⏞ 𝑎𝑙 ⟨𝜓𝑙 |𝜓𝑛 ⟩(4*)
=
(∗∗∗) = (4 ∗) => 𝑎𝑙 ⟨𝜓𝑙 |𝜓𝑛 ⟩ = 𝑎𝑛 ⟨𝜓𝑙 |𝜓𝑛 ⟩ => (𝑎𝑙 − 𝑎𝑛 )⟨𝜓𝑙 |𝜓𝑛 ⟩ = 0
Since 𝑛 ≠ 𝑙 => (𝑎𝑙 − 𝑎𝑛 ) ≠ 0 => ⟨𝜓𝑙 |𝜓𝑛 ⟩ = 0
 20

Determinate States

Suppose that an observable Q which is represented by a Hermitian operator 𝑄̂ will

be measured in a specific experiment. One considers an ensemble of identically
prepared systems, all in the same state Ψ(𝑥, 0) .

In each of the experiments, the particles are described by the same function
Ψ(𝑥, 0) at some initial time t=0.

There are two possible cases for the choice of Ψ(𝑥, 0):
1. Wave function Ψ(𝑥, 0) may be a general superposition of the possible
solutions of the Sch.eq:
∞

Ψ(𝑥, 0) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 )
𝑛=1

When a measurement of the observable Q is made on each experiment then, you do

not get the same result and the set of the results must be analyzed statistically.

EX: Suppose that a SHO is in a superposition state

Ψ(𝑥, 0) = 𝑎𝜓0 (𝑥 ) + 𝑏𝜓1 (𝑥 )
where 𝜓0 and 𝜓1 are stationary states of the Hamiltonian:
What is its energy? It may be 𝐸0 with probability |𝑎|2 , 𝐸1 with probability |𝑏|2 . But
it is not certain.
 21

2. In each of the experiments, the particles may be described by the function given
by
Ψ(𝑥, 0) = 𝜓𝑢 (𝑥 ) (∗)
such that every measurement of the observable 𝑄 is certain to return the same
value, say 𝑞𝑢 . Such a state is called a determinate state.

Actually we have seen such a state, stationary states of the Hamiltonian:

𝑖𝐸 𝑡
− 𝑛
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 )𝑒 ℏ ∶ stationary states of the Hamiltonian
A measurement of energy on such a state Ψ𝑛 is certain to yield the
corresponding “allowed” energy E𝑛 .

In the determinate state (*), the expectation value is

〈𝑄 〉 = 𝑞𝑢
Then the standard deviation of 𝑄 in such a state would be zero:

2
2
𝑄 〉) 〉 = 〈(𝑄 − 𝑞𝑢 )2 〉 = ⟨𝜓𝑢 |(𝑄̂ − 𝑞𝑢 ) 𝜓𝑢 ⟩
𝝈𝟐 = 𝟎 => 〈(𝑄 − 〈⏟
=𝑞𝑢

= ⟨(𝑄̂ − 𝑞𝑢 )𝜓𝑢 |(𝑄̂ − 𝑞𝑢 )𝜓𝑢 ⟩ = 0

The only function whose inner product with itself is zero is 0:

(𝑄̂ − 𝑞𝑢 )|𝜓𝑢 ⟩ = 0 => 𝑄̂ |𝜓𝑢 ⟩ = 𝑞𝑢 |𝜓𝑢 ⟩

This is an eigenvalue equation for the operator 𝑄̂ , 𝜓𝑢 is the eigenfunction of 𝑄̂ , 𝑞𝑢

is the corresponding eigenvalue. Measurement of Q on such a state is certain to yield
the eigenvalue 𝑞𝑢 .

̂.
Determinate states are eigenfunctions of 𝑸
 22

For example, determinate states of the total energy are eigenfunctions of the
Hamiltonian operator

̂ 𝜓𝑛 = 𝐸𝑛 𝜓𝑛
𝐻

which is precisely the time-independent Schrödinger equation.

Continuous Spectra

So far, we have excluded continuous spectra from the discussion. From a physical
point of view, a continuous spectrum (e.g. the energy of a free quantum object) makes
perfect sense. But we have the problem that the corresponding eigenfunctions are not
square integrable, and therefore we cannot properly define a scalar product.
Since these unphysical states are not square integrable, the
previously-developed probability concept of quantum mechanics cannot work with
them (or at least not readily)—that is the essential difficulty.

If the spectrum of a hermitian operator is continuous, the eigenfunctions are not

normalizable and the proof of Th1 and 2 fail, because the scalar product may not
exists:
∞

⟨𝜓𝑙 |𝜓𝑛 ⟩ = ∫ 𝜓𝑙∗ (𝑥 )𝜓𝑛 (𝑥 )𝑑𝑥

−∞

Nevertheless, there is a sense in which the three essential properties (reality,

orthogonality and completeness) still hold. It is best to approach this subtle case
through specific examples.
 23

Example 3.2
Find the eigenfunctions and eigenvalues of the momentum operator.
Let 𝜙𝑝 (𝑥 ) be the eigenfunction and 𝑝 the eigenvalue:
𝑑
𝑝̂𝑥 𝜙𝑝 (𝑥 ) = 𝑝 𝜙𝑝 (𝑥 ) ⟹ −𝑖ℏ 𝜙𝑝 (𝑥 ) = 𝑝 𝜙𝑝 (𝑥 )
𝑑𝑥
OR
𝑑𝜙𝑝 (𝑥 ) 𝑝
− 𝑖 𝜙𝑝 (𝑥 ) = 0
𝑑𝑥 ℏ
The general solution is

𝑝𝑥
𝑖
𝜙𝑝 (𝑥 ) = 𝐴 𝑒 ℏ 𝑜𝑟 with 𝑝 = 𝑘ℏ
This is not square integrable,
∞ ∞

⟨𝜙𝑝 |𝜙𝑝 ⟩ = ∫ 𝜙𝑝∗ (𝑥) 𝜙𝑝 (𝑥 )𝑑𝑥 = |𝐴|2 ∫ 𝑑𝑥 → ∞

−∞ −∞
However, when we restrict ourselves to real eigenvalues, we do recover a kind of
“orthonormality”:
∞ ∞
(𝑝−𝑝′ )𝑥
𝑖
⟨𝜙𝑝′ |𝜙𝑝 ⟩ = ∫ 𝜙𝑝∗ ′ (𝑥) 𝜙𝑝 (𝑥 )𝑑𝑥 = |𝐴 |2 ∫𝑒 ℏ 𝑑𝑥
−∞ −∞
|2
= |𝐴 2𝜋ℏ 𝛿 (𝑝 − 𝑝 ′)
≠ 0 𝑜𝑛𝑙𝑦 𝑖𝑓 𝑝 = 𝑝′

If we pick 𝐴 = 1/√2𝜋ℏ, so that

1 𝑖
𝑝𝑥
𝜙𝑝 (𝑥 ) = 𝑒 ℏ
√2𝜋ℏ
Then
⟨𝜙𝑝′ |𝜙𝑝 ⟩ = 𝛿 (𝑝 − 𝑝′ ) (∗)

Remember orthonormalization condition we had for a discrete set of the functions

{𝑓𝑛 (𝑥)}
⟨𝑓𝑚 |𝑓𝑛 ⟩ = 𝛿𝑛𝑚 (∗∗)
 24

with the completeness

Ψ(𝑥, 0) = ∑ 𝑐𝑛 𝑓𝑛 (𝑥 ) (∗∗∗)
𝑛

The eq.(*) is similar to (**) but the indices (𝑝, 𝑝′) are now continuous variables, and
the Kronecker delta has become a Dirac Delta , but otherwise it looks just the same.

We call eq. (*) Dirac orthonormality.

Note that the eigenfunctions {𝝓𝒑 (𝒙)} are complete and since the indices are now
continuous variables, the sum in (***) is replaced by an integral to write any
function 𝑔(𝑥 )
∞

𝑔(𝑥 ) = ∫ 𝑐(𝑝) 𝜙𝑝 (𝑥 )𝑑𝑝

−∞

The expansion coefficient, now a function, 𝑐(𝑝) is obtained by Fourier’s trick

∞ ∞ ∞ ∞

∫ 𝑔(𝑥) 𝜙𝑝∗ ′ (𝑥 )𝑑𝑥 = ∫ 𝑐(𝑝) 𝑑𝑝 ∫ 𝜙𝑝∗ ′ (𝑥 )𝜙𝑝 (𝑥 )𝑑𝑥 = ∫ 𝑐(𝑝) 𝛿 (𝑝 − 𝑝′ )𝑑𝑝 = 𝑐(𝑝′ )
−∞ −∞ ⏟
−∞ −∞
⟨𝜙𝑝′ |𝜙𝑝 ⟩

𝑐 (𝑝′ ) = ⟨𝜙𝑝′ |𝑔⟩ = ∫ 𝑔(𝑥) 𝜙𝑝∗ ′ (𝑥 )𝑑𝑥

−∞

(For its interpretation see page 30)

Conclusion:
 Eigenfunctions of momentum operator are not square integrable, they don’t
represent possible physical states, so they don’t live in Hilbert space.
 However, for real eigenvalues of the momentum operator, we do recover a kind of
“orthonormality”. They don’t represent possible physical states, but they are still
very useful, as we have seen in the scattering examples.
 25

Example . Eigenfunctions and eigenvalues of position operator 𝑥.

Suppose that a particle is found at a position 𝒚. What is the wave function
immediately after this measurement ?
It must be an eigenfunction of the position operator 𝑥:

𝑥 𝜓𝑦 (𝑥 ) = 𝑦 𝜓𝑦 (𝑥)

Here 𝑦 is a fixed number, but 𝑥 is a continuous variable.Note that:

𝜓𝑦 (𝑥 ) = 0 𝑤ℎ𝑒𝑛 𝑥 ≠ 𝑦

𝜓𝑦 (𝑥 ) must have the property that multiplying it by 𝑥 is the same as multiplying it

by 𝑦. This implies that it is Dirac delta function
𝜓𝑦 (𝑥 ) = 𝐴 𝛿 (𝑥 − 𝑦 )

As with momentum eigenfunctions, position eigenfunctions are not normalizable; the

eigenfunctions are not square integrable, but again from Dirac orthonormality
∞ ∞

⟨𝜓𝑦 |𝜓𝑦′ ⟩ = |𝐴|2 ∫ 𝛿 (𝑥 − 𝑦) 𝛿

⏟(𝑥 − 𝑦′) 𝑑𝑥 = |𝐴|2 ∫ 𝛿 (𝑥 − 𝑦)𝑓(𝑥) 𝑑𝑥
−∞ 𝑓(𝑥) −∞
2 |2
= |𝐴 𝑓(𝑦) = |𝐴 𝛿(𝑦 − 𝑦′)
|
If we pick 𝐴 = 1, so
⟨𝜓𝑦 |𝜓𝑦′ ⟩ = 𝛿 (𝑦 − 𝑦′)
Thus
𝜓𝑦 (𝑥 ) = 𝛿 (𝑥 − 𝑦)
These eigenfunctions {𝜓𝑦 (𝑥 )} are also complete, for and function 𝑓(𝑥)
∞ ∞

𝑓(𝑥 ) = ∫ 𝑐 (𝑦′) 𝜓𝑦′ (𝑥 ) 𝑑𝑦′ => 𝑐 (𝑦) = ⟨𝜓𝑦 |𝑓⟩ = ∫ 𝑓(𝑥 ) 𝜓𝑦∗ (𝑥 ) 𝑑𝑥
−∞ −∞
∞
𝑐 (𝑦) = ∫−∞ 𝑓 (𝑥 ) 𝛿 (𝑥 − 𝑦 ) 𝑑𝑥 => 𝑐 (𝑦) = 𝑓(𝑦) (For its interpretation see page 29)
 26

GENERALIZED STATISTICAL INTERPRETATION

In Ch1, we have seen how to calculate
 the probability that a particle would be found at time t in a particular location
x in (x,x+dx)
|Ψ(𝑥, 𝑡)|2 ≡ 𝑃(𝑥, 𝑡) ∶ Probability Density
 the expectation value of any quantity, 𝑄(𝑥, 𝑝),
∞
ℏ 𝜕
〈𝑄(𝑥, 𝑝)〉 = ∫ Ψ∗ 𝑄 (𝑥, ) Ψ 𝑑𝑥
−∞ 𝑖 𝜕𝑥

In Ch2, the time dependent Schrödinger equation with a time independent Hamiltonian
𝜕Ψ(𝑥, 𝑡)
𝑖ℏ = 𝐻Ψ(𝑥, 𝑡)
𝜕𝑡
has the most general solution given by
∞ ∞
𝑖𝐸𝑛 𝑡
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡) = ∑ 𝑐𝑛 𝜓𝑛 (𝑥 ) 𝑒 − ℏ
𝑛=1 𝑛=1

Here
𝑖𝐸𝑛 𝑡
Ψ𝑛 (𝑥, 𝑡) = 𝜓𝑛 (𝑥 ) 𝑒 − ℏ

With
𝐻𝜓𝑛 (𝑥 ) = 𝐸𝑛 𝜓𝑛 (𝑥 )

are stationary state solutions, in the sense that all probabilities and expectation
values are independent of time.
Possible outcomes of an energy measurement in state Ψ(𝑥, 𝑡) will be 𝐸𝑛 with the
probability |𝑐𝑛 |2 .
 27

We now state the general statistical interpretation:

Consider an observable 𝑄 represented by an operator 𝑄̂ (𝑥, 𝑝). Let 𝑓𝑛 (𝑥 ) be the

eigenfunctions of this observable operator 𝑄̂ with the corresponding eigenvalues 𝑞𝑛 :
𝑄̂ 𝑓𝑛 (𝑥 ) = 𝑞𝑛 𝑓𝑛 (𝑥 )

 𝑄̂ (𝑥, 𝑝) is hermitian.
 {𝑓𝑛 (𝑥 )} : orthogonal => ⟨𝑓𝑚 |𝑓𝑛 ⟩ = 𝛿𝑛𝑚
and square integrable =>
∞
∫ 𝑓𝑛∗ (𝑥 ) 𝑓𝑛 (𝑥 )𝑑𝑥 < ∞
−∞
=>eigenfunctions are normalizable.
 Eigenvalues 𝑞𝑛 : real.

If {𝑓𝑛 (𝑥 )} is also complete, any wave function 𝑔(𝑥 ) can be written uniquely and
exactly as a linear combination of them:

𝑔(𝑥 ) = ∑ 𝑐𝑛 𝑓𝑛 (𝑥 ) OR in Dirac Notation: |𝑔⟩ = ∑ 𝑐𝑛 |𝑓𝑛 ⟩

𝑛 𝑛
The coefficients are given by
∞ ∞

⟨𝑓𝑚 |𝑔⟩ = ∑ 𝑐𝑛 ⟨𝑓𝑚 |𝑓𝑛 ⟩ = ∑ 𝑐𝑛 𝛿𝑛𝑚 = 𝑐𝑚

𝑛=1 𝑛=1

𝑐𝑛 = ⟨𝑓𝑛 |𝑔⟩ = ∫ 𝑓𝑛∗ (𝑥 ) 𝑔(𝑥 )𝑑𝑥

Then,

|𝑔⟩ = ∑ 𝑐𝑛 |𝑓𝑛 ⟩ => |𝑔⟩ = ∑|𝑓𝑛 ⟩⟨𝑓𝑛 | 𝑔⟩ => ∑|𝑓𝑛 ⟩⟨𝑓𝑛 | = 𝕀 ∶ 𝑡ℎ𝑒 𝑢𝑛𝑖𝑡 𝑜𝑝𝑒𝑟𝑎𝑡𝑜𝑟
𝑛 𝑛 𝑛

Now, if you measure the observable 𝑄 on a state 𝜓(𝑥) you are certain to get one of
the eigenvalues:
 28

|𝑐𝑛 |2 : 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡 𝑜𝑓 𝑄 𝑖𝑛 𝑠𝑡𝑎𝑡𝑒 𝑔(x) 𝑤𝑖𝑙𝑙 𝑔𝑖𝑣𝑒 𝑞𝑛

Of course, the total probability had got to be one:

!
∑|𝑐𝑛 =
⏞1 |2
𝑛

This follows from the normalization of the wave function:

∗
1 = ⟨𝑔|𝑔⟩ = ⟨(∑𝑚 𝑐𝑚 𝑓𝑚 (𝑥 ))|(∑𝑛 𝑐𝑛 𝑓𝑛 (𝑥 ))⟩ = ∑ ∑ 𝑐𝑚 ⟨𝑓𝑚 |𝑓𝑛 ⟩
𝑐𝑛 ⏟
𝑛 𝑚 𝛿𝑛𝑚

1 = ∑|𝑐𝑛 |2
𝑛

Similarly,

𝑞𝑛 𝑓𝑛
⏞
〈𝑄 〉 = ⟨𝑔|𝑄̂ 𝑔⟩ = ⟨(∑𝑚 𝑐𝑚 𝑓𝑚 (𝑥 ))| (∑𝑛 𝑐𝑛 𝑄
̂ 𝑓𝑛 (𝑥 ))⟩ = ∑ ∑ 𝑐𝑚
∗
𝑐𝑛 𝑞𝑛 ⟨⏟𝑓𝑚 |𝑓𝑛 ⟩
𝑛 𝑚 𝛿𝑛𝑚

= ∑|𝑐𝑛 |2 𝑞𝑛 =>
𝑛

〈𝑄 〉 = ∑|𝑐𝑛 |2 𝑞𝑛
𝑛
 29

EX: In Example 3.3 , we found eigenfunctions 𝜓𝑦 (𝑥 ) and eigenvalues 𝑦 of the

position operator

𝑥 𝜓𝑦 (𝑥 ) = 𝑦 𝜓𝑦 (𝑥)

where
𝜓𝑦 (𝑥 ) = 𝛿 (𝑥 − 𝑦 )

with the Dirac orthonormality

⟨𝜓𝑦 |𝜓𝑦′ ⟩ = 𝛿 (𝑦 − 𝑦′)
These eigenfunctions {𝜓𝑦 (𝑥 )} are also complete such that for any function 𝑔(𝑥)
∞ ∞

𝑔(𝑥 ) = ∫ 𝑐(𝑦′) 𝜓𝑦′ (𝑥 ) 𝑑𝑦′ => 𝑐 (𝑦) = ⟨𝜓𝑦 |𝑔⟩ = ∫ 𝜓𝑦∗ (𝑥 )𝑔(𝑥 ) 𝑑𝑥
−∞ −∞
∞

𝑐(𝑦) = ∫ 𝛿 (𝑥 − 𝑦 )𝑔(𝑥 ) 𝑑𝑥 => 𝑐(𝑦) = 𝑔(𝑦)

−∞

From the general statistical interpretation we have

|𝑐(𝑦)|2 𝑑𝑦 = |𝑔(𝑦)|2 𝑑𝑦 =
∶ probability of finding the particle at position 𝑦 in the range
(𝑦, 𝑦 + 𝑑𝑦)
which is precisely the original statistical interpretation.
 30

Discrete spectra Continuous spectra

𝑄̂ 𝑓𝑛 (𝑥 ) = 𝑞𝑛 𝑓𝑛 (𝑥 ) 𝑥 𝜓𝑦 (𝑥 ) = 𝑦 𝜓𝑦 (𝑥)

{𝑓𝑛 (𝑥 )} : eigenfunctions are labelled by {𝜓𝑦 (𝑥)} : eigenfunctions are labelled

a discrete index n. by a continuous index y.
∞
𝑔(𝑥) = ∑ 𝑐𝑛 𝑓𝑛 (𝑥)
𝑔(𝑥 ) = ∫ 𝑐 (𝑦′) 𝜓𝑦′ (𝑥 ) 𝑑𝑦′
𝑛
−∞
|𝑔⟩ = ∑ 𝑐𝑛 |𝑓𝑛 ⟩ ∞
𝑛 |𝑔⟩ = ∫ 𝑐(𝑦′) |𝜓𝑦′ ⟩𝑑𝑦′
−∞
∞
𝑐𝑛 = ⟨𝑓𝑛 |𝑔⟩ = ∫ 𝑓𝑛∗ (𝑥 ) 𝑔(𝑥 )𝑑𝑥
𝑐(𝑦) = ⟨𝜓𝑦 |𝑔⟩ = ∫ 𝜓𝑦∗ (𝑥 )𝑔(𝑥 ) 𝑑𝑥
−∞

|𝑐𝑛 |2 : 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑡ℎ𝑎𝑡 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑚𝑒𝑛𝑡 |𝑐 (𝑦)|2 𝑑𝑦 = |𝑔(𝑦)|2 𝑑𝑦 =

𝑜𝑓 𝑄 𝑖𝑛 𝑠𝑡𝑎𝑡𝑒 𝑔(x) 𝑤𝑖𝑙𝑙 𝑔𝑖𝑣𝑒 𝑞𝑛 ∶ probability of finding the particle

at position 𝑦 in the range
(𝑦, 𝑦 + 𝑑𝑦)
 31

EX: In Example 3.2 we found the eigenfunctions of the momentum operator

1 𝑖
𝑝𝑥
𝜙𝑝 (𝑥 ) = 𝑒 ℏ
√2𝜋ℏ
If we restrict ourselves to real eigenvalues, we do recover a kind of “orthonormality”.
⟨𝜙𝑝′ |𝜙𝑝 ⟩ = 𝛿 (𝑝 − 𝑝′ ) : Dirac orthogonality

Note that the eigenfunctions {𝜙𝑝 (𝑥)} are complete so that any function 𝜓(𝑥) can be
expanded as

∞ ∞
1 𝑝𝑥
𝜓(𝑥 ) = ∫ 𝑐(𝑝) 𝜙𝑝 (𝑥 )𝑑𝑝 => 𝑐 (𝑝) = ⟨𝜙𝑝 |𝜓⟩ = ∫ 𝑒 −𝑖 ℏ 𝜓(𝑥 )𝑑𝑥
√2𝜋ℏ
−∞ −∞

This is an important quantity and we give it a special name and symbol

∞
1 −𝑖
𝑝𝑥
𝑐 (𝑝) → Φ(𝑝) = ∫𝑒 ℏ 𝜓(𝑥 )𝑑𝑥 ∶ momentum space wave function
√2𝜋ℏ
−∞
∞
1 𝑝𝑥
𝜓(𝑥 ) = ∫ 𝑒 𝑖 ℏ Φ(𝑥 )𝑑𝑝 ∶ position space wave function
√2𝜋ℏ
−∞

According to the generalized statistical interpretation

|Φ(𝑝)|2 𝑑𝑝
∶ probability of finding the particle in state 𝜓(𝑥 ) at the momentum p in the range dp
(𝑝, 𝑝 + 𝑑𝑝)
 32

Example 3.4A particle in free space is initially in a wave packet given by

𝛼 1/4 −𝛼𝑥 2 /2
Ψ(𝑥, 0) = ( ) 𝑒
𝜋
What is the probability that its momentum is in the range (𝑝, 𝑝 + 𝑑𝑝)?
 
1/4
 1
 ( p )   dx   e  x /2 e  ipx /
2


  2

Complete the square

   1 
1/4 1/2

 dxe  ( xip / /2
   )2
e p
2 2


    2  


1/4
 1   p2 /2

2

2 
e
  

From this we find that the probability the momentum is in the range (p, p + dp) is

1/2
 1   p2 /
| ( p) |2 dp  
2

2 
e
  
 33

1
𝑚𝜔 4 −𝑚𝜔𝑥 2
𝜓0 (𝑥 ) = ( ) 𝑒 2ℏ
𝜋ℏ
1
𝑚𝜔 4 −𝑚𝜔𝑥 2 −𝑖𝐸0 𝑡
Ψ0 (𝑥, 𝑡) = ( ) 𝑒 2ℏ 𝑒 2 𝑤𝑖𝑡ℎ 𝐸0 = ℏ𝜔/2
𝜋ℏ
∞
1 −𝑖
𝑝𝑥
Φ(𝑝, 𝑡) = ∫𝑒 ℏ Ψ(𝑥, 𝑡)𝑑𝑥 ∶ momentum space wave function
√2𝜋ℏ
−∞

1 ∞
1𝑚𝜔 4 −𝑖𝐸0 𝑡 −𝑖
𝑝𝑥 𝑚𝜔 2
Φ(𝑝, 𝑡) = ( ) 𝑒 2 ∫𝑒 ℏ 𝑒 2ℏ 𝑥 𝑑𝑥
−
√2𝜋ℏ 𝜋ℏ
−∞

𝑝𝑥 𝑚𝜔 2 𝑚𝜔 𝑝𝑥
𝑒𝑥𝑝 [−𝑖 − 𝑥 ] = 𝑒𝑥𝑝 [− (2𝑖 + 𝑥 2 )]
ℏ 2ℏ 2ℏ 𝑚𝜔
𝑚𝜔 𝑝 2 𝑝2
= 𝑒𝑥𝑝 [− (𝑥 + 𝑖 ) ] 𝑒𝑥𝑝 [− ]
2ℏ 𝑚𝜔 2ℏ𝑚𝜔
1 ∞
1𝑚𝜔 4 −𝑖𝐸0 𝑡 − 𝑝2 𝑚𝜔 𝑝 2
Φ(𝑝, 𝑡) = ( ) 𝑒 2 𝑒 2ℏ𝑚𝜔 ∫ 𝑒𝑥𝑝 [− (𝑥 + 𝑖 ) ] 𝑑𝑥
√2𝜋ℏ 𝜋ℏ ⏟
2ℏ 𝑚𝜔
−∞
𝜋
√𝑚𝜔/2ℏ

1
1 4 − 𝑝2 𝑖𝐸0 𝑡
Φ(𝑝, 𝑡) = ( ) 𝑒 2ℏ𝑚𝜔 𝑒 − 2
𝜋ℏ𝑚𝜔
 34

Classical momentum is
𝑝2
𝐸= => 𝑝 = ±√2𝑚𝐸.
2𝑚
In the ground state:𝐸 = ℏ𝜔/2. Then,
𝑝 = ±√𝑚ℏ𝜔
Thus the classical range of the momentum is
−√𝑚ℏ𝜔 ≤ 𝑝 ≤ √𝑚ℏ𝜔
The normalization=>
∞

1 = ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 =
−∞
𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒
𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒
⏞
−√𝑚ℏ𝜔 0 √𝑚ℏ𝜔 ⏞ ∞

= ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 + ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 + ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 + ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝

−∞ ⏟
−√𝑚ℏ𝜔 0 −√𝑚ℏ𝜔
𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒

Thus, the probability that a measurement of p on a particle in the ground state would yield a value
outside the classical range is given by
𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒
𝑜𝑢𝑡𝑠𝑖𝑑𝑒 𝑡ℎ𝑒 𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒
⏞
−√𝑚ℏ𝜔 ⏞ ∞

∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 + ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 ≡𝑃

−∞ −√𝑚ℏ𝜔

0 √𝑚ℏ𝜔

=1− ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 + ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝

⏟
−√𝑚ℏ𝜔 0
( 𝑐𝑙𝑎𝑠𝑠𝑖𝑐𝑎𝑙 𝑟𝑎𝑛𝑔𝑒 )
√𝑚ℏ𝜔

=> 𝑃 = 1 − 2 ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝

0
 35

Now,
1 2
√𝑚ℏ𝜔 √𝑚ℏ𝜔
1 4 𝑝2 𝑖𝐸 𝑡
− − 0
∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 = ∫ |( ) 𝑒 2ℏ𝑚𝜔 𝑒 2 | 𝑑𝑝
𝜋ℏ𝑚𝜔
0 0

1 √𝑚ℏ𝜔
1 2 𝑝2
−
=( ) ∫ 𝑒 ℏ𝑚𝜔 𝑑𝑝
𝜋ℏ𝑚𝜔
0

Chance of variable:

2 ℏ𝑚𝜔
𝑧≡√ 𝑝 => 𝑑𝑝 = √ 𝑑𝑧
ℏ𝑚𝜔 2
√𝑚ℏ𝜔 1 √2
1 ℏ𝑚𝜔 𝑧22
−
2
∫ |Φ(𝑝, 𝑡)| 𝑑𝑝 = ( ) √ ∫ 𝑒 2 𝑑𝑧
⏟𝜋ℏ𝑚𝜔 2
0 0
√1
2𝜋
√2
𝑧2 𝜋 𝜋
−
∫𝑒 2 𝑑𝑧 = √ 𝐸𝑟𝑓[1] = √ (0.843) = 1.056
2 2
0

√𝑚ℏ𝜔
1 𝜋 (0.843)
∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 = √ ∗ √ (0.843) =
2𝜋 2 2
0

√𝑚ℏ𝜔
(0.843)
𝑃 = 1 − 2 ∫ |Φ(𝑝, 𝑡)|2 𝑑𝑝 = 1 − 2 = 0.157 ≅ 0.16
2
0
 1

THE UNCERTAINTY PRINCIPLE

We stated the uncertainty principle in the form

ℏ
𝜎𝑥 𝜎𝑝 ≥
2

The more precise a wave’s position is, the less precise is its wavelength
(=momentum) , and vice versa.
In Ch1 and we have checked it several times, in the problems.

Now, we will prove it’s a more general version.

The starting point will the Schwarz inequality. In the real Euclidian space, with
three-dimensional vectors 𝑎⃗ and 𝑏⃗⃗, we have

𝑎⃗ ∙ 𝑏⃗⃗ = |𝑎⃗||𝑏⃗⃗| cos(𝑎⃗, 𝑏⃗⃗) ≤ |𝑎⃗||𝑏⃗⃗|

For any two state vectors |𝜓⟩ and |𝜙⟩ of the Hilbert space, we can show that

|⟨𝜓|𝜙⟩|2 ≤ ⟨𝜓|𝜓⟩⟨𝜙|𝜙⟩

Consider two Hermitian operators 𝑨 ̂ and 𝑩

̂ corresponding to two observables 𝐴 and
𝐵. By definition, their standard deviations with respect to a normalized state vector
|𝜓⟩ are given by

2
𝜎𝐴2 = 〈(𝐴 − 〈𝐴〉)2 〉 = ⟨𝜓|(𝐴̂ − 〈𝐴〉) 𝜓⟩ = ⟨𝜓|(𝐴̂ − 〈𝐴〉)(𝐴̂ − 〈𝐴〉)𝜓⟩
= ⟨(𝐴̂ − 〈𝐴〉)𝜓|(𝐴̂ − 〈𝐴〉)𝜓⟩

𝜎𝐴2 ≡ ⟨𝑓|𝑓⟩ (𝑎)

where

|𝑓⟩ ≡ |(𝐴̂ − 〈𝐴〉)𝜓⟩

 2

Likewise

2
𝜎𝐵2 = 〈(𝐵 − 〈𝐵〉)2 〉 = ⟨𝜓|(𝐵̂ − 〈𝐵〉) 𝜓⟩ = ⟨𝜓|(𝐵̂ − 〈𝐵〉)(𝐵̂ − 〈𝐵〉)𝜓⟩

= ⟨(𝐵̂ − 〈𝐵〉)𝜓|(𝐵̂ − 〈𝐵〉)𝜓⟩

𝜎𝐵2 ≡ ⟨𝑔|𝑔⟩ (𝑏)

where

|𝑔⟩ ≡ |(𝐵̂ − 〈𝐵〉)𝜓⟩

Note that from (a) and (b):

𝜎𝐴2 𝜎𝐵2 = ⟨𝑓 |𝑓 ⟩⟨𝑔|𝑔⟩

Now, from Schwarz inequality

𝜎𝐴2 𝜎𝐵2 = ⟨𝑓|𝑓⟩⟨𝑔|𝑔⟩ ≥ |⟨𝒇|𝒈⟩|𝟐 (∗)

where

|⟨𝑓|𝑔⟩|2 = ⟨𝑓|𝑔⟩⟨𝑓|𝑔⟩∗

For any complex number z

2
1
|𝑧 |2 = [𝑅𝑒(𝑧) ]2 + [𝐼𝑚(𝑧) ]2 ≥ [𝐼𝑚(𝑧 )]2 ∗
= [ (𝑧 − 𝑧 )]
2𝑖

Letting

2 2
1 1
⟨𝑓|𝑔⟩ ≡ 𝑧 => |⟨𝒇|𝒈⟩|𝟐 = |𝑧|2 = [ (𝑧 − 𝑧 ∗ )] = [ (⟨𝑓|𝑔⟩ − ⟨𝑔|𝑓⟩]
2𝑖 2𝑖

Thus, from (∗) we have

 3

2
1
𝜎𝐴2 𝜎𝐵2 ≥ | (⟨𝑓|𝑔⟩ − ⟨𝑔|𝑓⟩| (∗∗)
2𝑖

Now,

⟨𝑓 |𝑔⟩ = ⟨(𝐴̂ − 〈𝐴〉)𝜓|(𝐵̂ − 〈𝐵〉)𝜓⟩ = ⟨𝜓|(𝐴̂ − 〈𝐴〉)(𝐵̂ − 〈𝐵〉)𝜓⟩

= ⟨𝜓|(𝐴̂𝐵̂ − 𝐴̂〈𝐵〉 − 〈𝐴〉𝐵̂ + 〈𝐴〉〈𝐵〉)𝜓⟩

= ⟨𝜓|𝐴̂𝐵̂𝜓⟩ − 〈𝐵〉⟨𝜓|𝐴̂𝜓⟩ − 〈𝐴〉⟨𝜓|𝐵̂𝜓⟩ + 〈𝐴〉〈𝐵〉⟨𝜓|𝜓⟩

= 〈𝐴𝐵〉 − 〈𝐵〉〈𝐴〉 − 〈𝐴〉〈𝐵〉 + 〈𝐴〉〈𝐵〉 =>

⟨𝑓|𝑔⟩ = 〈𝐴𝐵〉 − 〈𝐵〉〈𝐴〉

Similarly,

⟨𝑔|𝑓⟩ = ⟨(𝐵̂ − 〈𝐵〉)𝜓|(𝐴̂ − 〈𝐴〉)𝜓⟩ = ⟨𝜓|(𝐵̂ − 〈𝐵〉)(𝐴̂ − 〈𝐴〉)𝜓⟩

= ⟨𝜓|(𝐵̂𝐴̂ − 〈𝐵〉𝐴̂ − 𝐵̂〈𝐴〉 + 〈𝐵〉〈𝐴〉)𝜓⟩

= ⟨𝜓|𝐵̂ 𝐴̂𝜓⟩ − 〈𝐵〉⟨𝜓|𝐴̂𝜓⟩ − 〈𝐴〉⟨𝜓|𝐵̂𝜓⟩ + 〈𝐵〉〈𝐴〉⟨𝜓|𝜓⟩

= 〈𝐵𝐴〉 − 〈𝐵〉〈𝐴〉 − 〈𝐴〉〈𝐵〉 + 〈𝐴〉〈𝐵〉 =>

⟨𝑔|𝑓 ⟩ = 〈𝐵𝐴〉 − 〈𝐵〉〈𝐴〉

So that

⟨𝑓 |𝑔⟩ − ⟨𝑔|𝑓 ⟩ = 〈𝐴𝐵 〉 − 〈𝐵𝐴〉 = ⟨𝜓|𝐴̂𝐵̂𝜓⟩ − ⟨𝜓|𝐵̂𝐴̂𝜓⟩ = ⟨𝜓|(𝐴̂𝐵̂ − 𝐵̂𝐴̂)𝜓⟩

= 〈[𝐴̂, 𝐵̂]〉

where

[𝐴̂, 𝐵̂] ≡ 𝐴̂𝐵̂ − 𝐵̂𝐴̂

 4

Conclusion: From (∗∗)

2
1
𝜎𝐴2 𝜎𝐵2 ≥ | 〈[𝐴̂, 𝐵̂]〉| (3 ∗)
2𝑖

This is the generalized uncertainty principle

𝑑
EX: Suppose 𝐴̂ = 𝑥̂ and 𝐵̂ = 𝑝̂ = −𝑖ℏ . Their commutator is
𝑑𝑥

[𝑥̂, 𝑝̂ ] = 𝑖ℏ

So,

1 2
ℏ2
𝜎𝑥2 𝜎𝑝2 ≥ | 𝑖ℏ| = [ ]
2𝑖 2

ℏ
𝜎𝑥 𝜎𝑝 ≥
2

OR with 𝜎𝑥 ≡ Δ𝑥 , 𝜎𝑝 ≡ Δ𝑝, it is the original Heisenberg uncertainty principle:

ℏ
Δ𝑥Δ𝑝 ≥
2
--------------------------------------------------------------------------------------------
 5

-------------------------------------------------------------------------------------------------
Note that the uncertainty principle is not an extra assumption in quantum
theory, but rather a consequence of the statistical interpretation.
When you measure the position of the particle the act of the
measurement collapses the wave function to a narrow spike.
(An eigenstate of the position operator!) To produce a matter
wave that was spatially localized that much, we put several
matter waves of particle with different momenta together.
This wave function therefore necessarily carries a broad
range of wavelengths (and hence momenta) in its Fourier
decomposition.
Poisiton =precise
Momentum=unknown (∞′ 𝑙𝑦 𝑚𝑎𝑛𝑦)

If we now measure the momentum, the state will

collapse to a long sine wave (an eigenstate of the
momentum operator !) with now a well-defined
wavelength. But, the particle no longer has the position you got in the first
measurement.
Momentum: precise
Position: unknown (∞′ 𝑙𝑦 𝑚𝑎𝑛𝑦 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑝𝑜𝑠𝑖𝑡𝑖𝑜𝑛𝑠)

The second measurement renders the outcome of the first measurement

obsolete. Only if the wave function were simultaneously an eigenstate of both
observables (x and p here) it would be possible to make the 2.nd measurement
without disturbing the state of the particle (the 2.nd collapse would not change
anything in that case). This is possible if both observables, say A and B are
compatible: [𝐴̂, 𝐵̂ ] = 0.

In fact there is an “uncertainty relation” for every pair of observables whose

operators do not commute- we call them incompatible observables. Incompatible
observables do not have common eigenfunctions; at least they cannot have a complete
set of common eigenfunctions. By contrast, compatible observables do admit
complete sets of simultaneous eigenfunctions.

-------------------------------------------------------------------------------------------
 6

Suppose that 𝑃̂ and 𝑄̂ are two operators that have a complete set of common
eigenfunctions {𝑓𝑛 }:
𝑃̂𝑓𝑛 = 𝜆𝑛 𝑓𝑛
𝑄̂ 𝑓𝑛 = 𝑞𝑛 𝑓𝑛
That is 𝑓𝑛 (𝑥) is an eigenfunction of both of 𝑃̂ and 𝑄̂ . Since the set {𝑓𝑛 } is complete
=> any function 𝑔(𝑥) in Hilbert space can be expressed as a linear combination
𝑔(𝑥 ) = ∑ 𝑐𝑛 𝑓𝑛
𝑛
Now,
[𝑃̂, 𝑄̂ ]𝑔(𝑥 ) = (𝑃̂𝑄̂ − 𝑄̂ 𝑃̂)𝑔(𝑥 ) = (𝑃̂𝑄̂ − 𝑄̂ 𝑃̂) ∑ 𝑐𝑛 𝑓𝑛
𝑛
= 𝑃̂ ∑ 𝑐𝑛 𝑄̂ 𝑓𝑛 − 𝑄̂ ∑ 𝑐𝑛 𝑃̂𝑓𝑛 = ∑ 𝑐𝑛 𝑞𝑛 𝑃̂𝑓𝑛 − ∑ 𝑐𝑛 𝜆𝑛 𝑄̂ 𝑓𝑛
𝑛 𝑛 𝑛 𝑛
= ∑ 𝑐𝑛 𝑞𝑛 𝜆𝑛 𝑓𝑛 − ∑ 𝑐𝑛 𝜆𝑛 𝑞𝑛 𝑓𝑛 = 0
𝑛 𝑛
Since this is true for any function g=>[𝑃̂, 𝑄̂ ] = 0
̂ , 𝐿̂2 , 𝐿̂𝑧 } are mutually compatible observables=>
For example, in the H-atom, {𝐻
̂ , 𝐿̂2 ] = [𝐻
[𝐻 ̂ , 𝐿̂𝑧 ] = [𝐿̂2 , 𝐿̂𝑧 ] = 0 =>

We construct simultaneous eigenfunctions of all three:

𝜓𝑛𝑙𝑚 (𝑟, 𝜃, 𝜙) = 𝑅𝑛𝑙 (𝑟)𝑌𝑙𝑚 (𝜃, 𝜙)
̂ 𝜓𝑛𝑙𝑚 (𝑟, 𝜃, 𝜙) = 𝐸𝑛 𝜓𝑛𝑙𝑚 (𝑟, 𝜃, 𝜙)
𝐻
𝐿2 𝑌𝑙𝑚 = ℏ2 𝑙(𝑙 + 1) 𝑌𝑙𝑚
𝐿̂𝑧 𝑌𝑙𝑚 = ℏ 𝑚 𝑌𝑙𝑚 , 𝑚 = −𝑙, … , 𝑙
But we cannot find a common eigenfunction of both the position and the
momentum since they are not compatible.
 7

2
1
𝜎𝐴2 𝜎𝐵2 ≥ | 〈[𝐴̂, 𝐵̂]〉|
2𝑖
𝑝̂2
𝐴̂ = 𝑥̂ and 𝐵̂ = 𝐻 = + 𝑉(𝑥)
2𝑚
2
1
𝜎𝑥2 𝜎𝐻2 ̂ ] 〉|
≥ | 〈[𝑥̂, 𝐻
2𝑖
𝑝̂ 2 1
̂ ] = [𝑥̂,
[𝑥̂, 𝐻 + 𝑉(𝑥)] = [𝑥̂, 𝑝̂ 2 ] + [⏟𝑉(𝑥), 𝑥̂ ] =
2𝑚 2𝑚
=0

1 1 𝑖ℏ𝑝̂
([⏟
𝑥̂, 𝑝̂ ]𝑝̂ + 𝑝̂ [𝑥̂, 𝑝̂ ]) = ̂] =
(2𝑖ℏ𝑝̂ ) => [𝑥̂, 𝐻
2𝑚 2𝑚 𝑚
𝑖ℏ

̂ ] ≠ 0, position and energy observables are not compatible and satisfy an

Since [𝑥̂, 𝐻
uncertainty relation
2
1 𝑖ℏ𝑝 ℏ
𝜎𝑥2 𝜎𝐻2 ≥ | 〈 〉| => 𝜎𝑥 𝜎𝐻 ≥ |〈𝑝〉|
2𝑖 𝑚 2𝑚

For stationary states, 𝜎𝐻 ≡ ΔE = 0.

𝑑〈𝑥〉
Also 〈𝑝〉 = 0 =>Ehrenfest’s theorem:〈𝑝〉 = 𝑚 . (Expectation values obey
𝑑𝑡

classical laws)

𝑑〈𝑥〉
In a stationary state, expectations do not change in time, and so .
𝑑𝑡
 8

Time Development of the Expectation Values

As a measure of how fast the system is changing let us compute the time derivative
of the expectation value of some operator 𝑄̂ :
𝑑 𝑑 𝜕𝛹 ̂ 𝜕𝑄̂ 𝜕𝛹
〈𝑄 〉 = ⟨𝛹|𝑄̂ 𝛹⟩ = ⟨ |𝑄 𝛹⟩ + ⟨𝛹| 𝛹⟩ + ⟨𝛹|𝑄̂ ⟩
𝑑𝑡 𝑑𝑡 𝜕𝑡 𝜕𝑡 𝜕𝑡

Now, the Schrödinger eq

𝜕𝛹 𝜕𝛹 1
𝑖ℏ = 𝐻Ψ => = 𝐻Ψ
𝜕𝑡 𝜕𝑡 𝑖ℏ

𝑑 1 𝜕𝑄̂ 1
〈𝑄 〉 = ⟨ 𝐻Ψ|𝑄̂ 𝛹⟩ + ⟨𝛹| 𝛹⟩ + ⟨𝛹|𝑄̂ 𝐻Ψ⟩
𝑑𝑡 𝑖ℏ 𝜕𝑡 𝑖ℏ

1 1 𝜕𝑄̂
=− ⟨𝐻Ψ|𝑄̂ 𝛹⟩ + ⟨𝛹|𝑄̂ 𝐻Ψ⟩ + ⟨𝛹| 𝛹⟩
𝑖ℏ 𝑖ℏ 𝜕𝑡

1
̂
1
̂ 𝜕𝑄̂
= − ⟨Ψ|𝐻𝑄 𝛹⟩ + ⟨𝛹|𝑄 𝐻Ψ⟩ + ⟨𝛹| 𝛹⟩
𝑖ℏ 𝑖ℏ 𝜕𝑡

1 𝜕𝑄̂
=− ⟨Ψ|(𝐻𝑄̂ − 𝑄̂ 𝐻)𝛹⟩ + ⟨𝛹| 𝛹⟩
𝑖ℏ 𝜕𝑡

𝑑 𝑖 𝜕𝑄̂
〈𝑄 〉 = 〈[𝐻, 𝑄̂ ]〉 + 〈 〉
𝑑𝑡 ℏ 𝜕𝑡

𝜕𝑄̂ 𝑑 𝑖
= 0 => 〈𝑄 〉 = 〈[𝐻, 𝑄̂ ]〉
𝜕𝑡 𝑑𝑡 ℏ

In the event that 𝑄̂ commutes with 𝐻, the quantity 〈𝑄 〉 is constant in time

 9

Problem 3.17. Apply the eq. below to the special cases:

𝑑 𝑖 𝜕𝑄̂
̂
〈𝑄 〉 = 〈[𝐻, 𝑄 ]〉 + 〈 〉
𝑑𝑡 ℏ 𝜕𝑡
a) 𝑄 = 1 . [𝐻, 1] = 0
𝑑 𝑑
〈𝑄 〉 = 0 = ⟨𝛹 |𝛹 ⟩
𝑑𝑡 𝑑𝑡
(this is the conservation of normalization, which we originally proved in Eq. 1.27).
b)
𝑄̂ = 𝐻
𝜕𝐻 𝑑 𝑑
= 0 , [𝐻, 𝐻 ] = 0 => 〈𝐻 〉 = 〈𝐸 〉 = 0
𝜕𝑡 𝑑𝑡 𝑑𝑡
=> 〈𝐸 〉 𝑖𝑠 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑜𝑓 𝑚𝑜𝑡𝑖𝑜𝑛

This is conservation of energy in the sense that

〈𝐻 〉 = ∑|𝑐𝑛 |2 𝐸𝑛
𝑛

Probability of getting a particular energy is independent of time.

c) 𝑄 = 𝑥
𝑑 𝑖 𝜕𝑥̂
〈𝑥 〉 = 〈[𝐻, 𝑥̂ ]〉 + 〈 〉
𝑑𝑡 ℏ 𝜕𝑡

𝑝̂ 2 1
[𝐻, 𝑥̂ ] = [ + 𝑉(𝑥), 𝑥̂] = [𝑝̂ 2 , 𝑥̂ ] + [⏟𝑉(𝑥), 𝑥̂ ] =
2𝑚 2𝑚
=0

1 1 𝑖ℏ𝑝̂
[𝐻, 𝑥̂ ] = (𝑝̂ [⏟𝑝̂ , 𝑥̂ ] + [𝑝̂ , 𝑥̂ ]𝑝̂ ) = (−2𝑖ℏ𝑝̂ ) => [𝐻, 𝑥̂ ] = −
2𝑚 2𝑚 𝑚
−𝑖ℏ

𝑑 𝑖 𝑖ℏ𝑝̂ 𝑑 〈𝑝̂ 〉
〈𝑥 〉 = 〈− 〉 => 〈𝑥 〉 =
𝑑𝑡 ℏ 𝑚 𝑑𝑡 𝑚
 10

d) 𝑄 = 𝑝

𝑑 𝑖 𝜕𝑝̂
〈𝑝〉 = 〈[𝐻, 𝑝̂ ]〉 + 〈 〉
𝑑𝑡 ℏ 𝜕𝑡

𝑝̂ 2 1
[𝐻, 𝑝̂ ] = [ + 𝑉(𝑥), 𝑝̂ ] = [𝑝̂ 2 , 𝑝̂ ] + [𝑉(𝑥), 𝑝̂ ]
2𝑚 2𝑚

𝑑𝑓 𝑑 𝑑𝑉
[𝑉 (𝑥 ), 𝑝̂ ]𝑓 (𝑥 ) = −𝑖ℏ (𝑉(𝑥 ) − (𝑉𝑓)) = 𝑖ℏ 𝑓
𝑑𝑥 𝑑𝑥 𝑑𝑥

𝑑𝑉
[𝐻, 𝑝̂ ] = 𝑖ℏ
𝑑𝑥

𝑑 𝑖 𝑑𝑉 𝑑 𝑑𝑉
〈𝑝〉 = 〈𝑖ℏ 〉 => 〈𝑝 〉 = − 〈 〉
𝑑𝑡 ℏ 𝑑𝑥 𝑑𝑡 𝑑𝑥
This is in accord with the Ehrenfest’s theorem.
 11

The Energy-Time Uncertainty Principle

There is also an uncertainty relation between energy and time,
ℏ
Δ𝐸 Δ𝑡 ≥ (∗)
2
which is not as strict as the position-momentum uncertainty principle
ℏ
Δ𝑥 Δ𝑝 ≥ (∗∗)
2
In the non-relativistic quantum mechanics, 𝑡 and 𝑥 are treated on a very unequal
footing and eq.(*) is not implied by eq.(**)
The relation involves the minimal time window Δ𝑡 which is required to observe a
system with energy uncertainty Δ𝐸.
Note that position, momentum and energy are all dynamical variables, they are
measurable characteristics of the system, at any given time. But the time itself is not
a dynamical variable (not an observable): you do not go out and measure the “time”
of a particle as you might do its position or its energy. Time is an independent
variable, of which the dynamical quantities are functions. It is not an operator in
QM.=> In particular, the Δ𝑡 in (*) is NOT the standard deviation of a collection of
time measurements: roughly speaking

𝚫𝒕 is the time it takes the system to change substantially.

Now, take the relation
2
1
𝜎𝐴2 𝜎𝐵2 ≥ | 〈[𝐴̂, 𝐵̂]〉|
2𝑖
and there pick 𝐴̂ = 𝐻 and 𝐵̂ = 𝑄̂ :
2 2 2
1 1 1ℏ𝑑
𝜎𝐴2 𝜎𝐵2 ≥ | 〈[𝐴̂, 𝐵̂]〉| → 𝜎𝐻 𝜎𝑄 ≥ | 〈[𝐻
2 2 ̂ , 𝑄̂ ]〉| = | 〈𝑄 〉|
2𝑖 2𝑖 2𝑖 𝑖 𝑑𝑡
Which follows from the relation
 12

𝑑 𝑖
〈𝑄 〉 = 〈[𝐻, 𝑄̂ ]〉
𝑑𝑡 ℏ
Then,
2
2 2
ℏ2 𝑑〈𝑄 〉 ℏ 𝑑 〈𝑄 〉
𝜎𝐻 𝜎𝑄 ≥ | | 𝑂𝑅 , 𝜎
⏟𝐻 𝜎𝑄 ≥ | |
4 𝑑𝑡 2 𝑑𝑡
∆𝐸

OR
𝜎𝑄 ℏ
∆𝐸 ≥
𝑑 〈𝑄 〉 2
| |
𝑑𝑡
where 𝜎𝐻 ≡ ∆𝐸. We now identify
𝜎𝑄
∆𝑡 ≡
𝑑 〈𝑄 〉
| |
𝑑𝑡
Then
ℏ
Δ𝐸Δ𝑡 ≥
2
It remains to interpret the quantity 𝚫𝒕.
𝑑 〈𝑄 〉
| | : 𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑠 ℎ𝑜𝑤 𝑓𝑎𝑠𝑡 〈𝑄 〉 𝑐ℎ𝑎𝑛𝑔𝑒𝑠 𝑜𝑣𝑒𝑟 𝑡𝑖𝑚𝑒
𝑑𝑡

∆𝑡 then tells you the approximate amount of time it takes for the expectation value
of an observable to change by a standard deviation.
𝜎𝑄 ← 𝑎𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑐ℎ𝑎𝑛𝑔𝑒
∆𝑡 ≡
𝑑 〈𝑄 〉 ← 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦 𝑜𝑓 𝑐ℎ𝑎𝑛𝑔𝑒
| |
𝑑𝑡
 13

Problem 3.18 Test the energy-time uncertainty principle for the wave function of
𝑑〈𝑥〉
Problem 2.5 and observable x, by calculating 𝜎𝐻 , 𝜎𝑥 and exactly.
𝑑𝑡

In a stationary state for which the energy is uniquely determined all expectation
values are constant in time:
𝑑 〈𝑄 〉
=0
𝑑𝑡
And ∆𝐸 = 0 => ∆𝑡 = ∞. To make something happen you must take a time
dependent linear combination state.
1
Ψ(𝑥, 𝑡) = (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )
√2
Energy-time uncertainty relation says
ℏ 𝜎𝑄
Δ𝐸Δ𝑡 ≥ , 𝜎𝐻 ≡ ∆𝐸 𝑤𝑖𝑡ℎ ∆𝑡 ≡
2 𝑑 〈𝑄 〉
| |
𝑑𝑡
For the position observable 𝑄 = 𝑥
𝜎𝑥
∆𝑡 ≡
𝑑 〈𝑥 〉
| |
𝑑𝑡

Now
2
ℏ 2 2
ℏ2 𝑑〈𝑥 〉
Δ𝐸Δ𝑡 ≥ => 𝜎𝐻 𝜎𝑥 ≥ [ ]
2 4 𝑑𝑡

1 1
(Δ𝐸)2 ≡ 𝜎𝐻2 = 〈𝐻 2 〉 − 〈𝐻 〉2 = (𝐸1 − 𝐸2 )2 = (3ℏ𝜔)2
4 4
𝑎2 1 5 32 2
𝜎𝑥2 = 〈𝑥 2〉
− 〈𝑥 〉2 = [ − 2
− ( 2 ) cos2 3𝜔]
4 3 4(𝜋) 9𝜋
 14

𝑑 〈𝑥 〉 𝑑 𝑎 32 𝑎 32 𝑑 〈𝑥 〉 8ℏ
= [1 − 2 cos 3𝜔] = 3𝜔 sin 3𝜔 => = sin 3𝜔
𝑑𝑡 𝑑𝑡 2 9𝜋 2 9𝜋 2 𝑑𝑡 3𝑚𝑎
2
2 2
ℏ2 𝑑 〈𝑥 〉
𝜎𝐻 𝜎𝑥 ≥ [ ]
4 𝑑𝑡

2 2
1 𝑎 1 5 32
𝐿𝐻𝑆 ≡ 𝜎𝐻2 𝜎𝑥2 = (3ℏ𝜔)2 [ − 2
− ( 2 ) cos 2 3𝜔]
4 4 3 4(𝜋) 9𝜋
3 2 1 5 32 2
= ( ) (𝑎ℏ𝜔)2 [ − 2
− ( 2 ) cos2 3𝜔]
4 3 4(𝜋) 9𝜋
2 2
ℏ2 𝑑 〈𝑥 〉 ℏ2 8ℏ ℏ 8ℏ 2 2
𝑅𝐻𝑆 ≡ [ ] = [ sin 3𝜔] = ( ) sin 3𝜔
4 𝑑𝑡 4 3𝑚𝑎 2 3𝑚𝑎
2
ℏ2 𝑑〈𝑥 〉 8 2
𝑅𝐻𝑆 ≡ [ ] = ( 2 ) (ℏ𝜔𝑎)2 sin2 3𝜔
4 𝑑𝑡 3𝜋
?
3 2 1 5 32 2
8 2
2
( ) (𝑎ℏ𝜔) [ − 2 ⏞(
− ( 2 ) cos 3𝜔] ≥ ) (ℏ𝜔𝑎 )2 sin2 3𝜔
4 3 4(𝜋) 2 9𝜋 3𝜋 2

?
1 5 32 2 4 × 8 2
[ − ⏞(
− ( 2 ) cos2 3𝜔] ≥ ) sin2 3𝜔
3 4(𝜋) 2 9𝜋 3 × 3𝜋 2

?
1 5 4×8 2 2 32 2 32 2
[ − ⏞(
]≥ 2
) sin 3𝜔 + ( 2 ) cos 3𝜔 = ( 2 )
3 4(𝜋)2 3 × 3𝜋 2 9𝜋 9𝜋
𝑌𝐸𝑆
1 5 32 2
[ − ⏞ (
] ≥ )
3 4 (𝜋 )2 ⏟9𝜋 2
⏟
0.20668 0.12978
 15

Details of the calculation:

Δ𝐸 ≡ 𝜎𝐻 = √〈𝐻 2 〉 − 〈𝐻 〉2
〈𝐻 〉 = ⟨Ψ|𝐻 |Ψ⟩
1 1
=⟨ (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|𝐻 (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
√2 √2
1
= ⟨(𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|(𝐸1 𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝐸2 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
2
1
〈𝐻 〉 = (𝐸 + 𝐸2 )
2 1
Note that all the cross terms vanish due to orthogonality of energy eigenfunctions:

⟨𝜓𝑖 |𝜓𝑗 ⟩ = 𝛿𝑖𝑗

〈𝐻 2 〉
1 1
= ⟨ (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|𝐻 2 (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
√2 √2
1
= ⟨(𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|𝐻(𝐸1 𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝐸2 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
2
1
= ⟨(𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|(𝐸12 𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝐸22 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
2
Here

𝐻 2 𝜓1 = 𝐻𝐻𝜓1 = 𝐸12 𝜓1 , 𝑒𝑐𝑡.

1 2
〈𝐻 2 〉 = (𝐸1 + 𝐸22 )
2
1 1
𝜎𝐻2 = 〈𝐻 2 〉 − 〈𝐻 〉2 = (𝐸12 + 𝐸22 ) − {〈 (𝐸1 + 𝐸2 )〉2 }
2 2
 16

1 2 1 1
(𝐸1 + 𝐸22 ) − (𝐸12 + 𝐸22 ) + 2𝐸1 𝐸2 => 𝜎𝐻2 = (𝐸12 + 𝐸22 − 2𝐸1 𝐸2 )
2 4 4
1
= (𝐸1 − 𝐸2 )2
4
1
(Δ𝐸)2 ≡ 𝜎𝐻2 = (𝐸1 − 𝐸2 )2
4
2 2 2 2 2 2
ℏ 𝜋 ℏ 𝜋
(𝐸1 − 𝐸2 )2 = [ 2
(1 − 4 )] = [− 2
3] ≡ (3ℏ𝜔)2
2𝑚𝑎 2𝑚𝑎

1 1
(Δ𝐸)2 ≡ 𝜎𝐻2 = (𝐸1 − 𝐸2 )2 = (3ℏ𝜔)2 (𝑎)
4 4
∞ ∞
∗
〈𝑥 〉 = ∫ Ψ 𝑥̂ Ψ 𝑑𝑥 = ∫ 𝑥̂ |Ψ(x, t)|2 𝑑𝑥
−∞ −∞
From problem 2.5(c)
 17

𝑎 32
〈𝑥 〉 = [1 − 2 cos 3𝜔]
2 9𝜋
𝑑〈𝑥〉 𝑑 𝑎 32 𝑎 32 𝑑〈𝑥〉 8ℏ
= [1 − cos 3𝜔] = 3𝜔 sin 3𝜔 => = sin 3𝜔 (b)
𝑑𝑡 𝑑𝑡 2 9𝜋2 2 9𝜋2 𝑑𝑡 3𝑚𝑎

Now calculate 𝜎𝑥2 = 〈𝑥 2 〉 − 〈𝑥 〉2 :

〈𝑥 2 〉
1 1
= ⟨ (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )|𝑥 2 (𝜓1 (𝑥)𝑒 −𝑖𝐸1 𝑡/ℏ + 𝜓2 (𝑥)𝑒 −𝑖𝐸2 𝑡/ℏ )⟩
√2 √2
1
= [⟨𝜓1 |𝑥 2 |𝜓1 ⟩ + ⟨𝜓2 |𝑥 2 |𝜓2 ⟩ + 𝑒 𝑖(𝐸1 −𝐸2 )𝑡/ℏ ⟨𝜓1 |𝑥 2 |𝜓2 ⟩
2
+ 𝑒 −𝑖(𝐸1 −𝐸2 )𝑡/ℏ ⟨𝜓2 |𝑥 2 |𝜓1 ⟩]
 18

From problem (2.4)

𝑎
2 𝑛𝜋𝑥 1 1
⟨𝜓𝑛 |𝑥 2 |𝜓𝑛 ⟩ = ∫ 𝑥 2 sin2 ( ) 𝑑𝑥 = 𝑎2 [ − ]
𝑎 𝑎 3 2(𝑛𝜋)2
0

1 1 1 1
⟨𝜓1 |𝑥 2 |𝜓1 ⟩ = 𝑎2 [ − ] , ⟨ 𝜓 2 | 𝑥 2|
𝜓 2 ⟩ = 𝑎 2
[ − ]
3 2(𝜋)2 3 2(2𝜋)2

𝑎
2 𝑛𝜋𝑥 𝑚𝜋𝑥
⟨𝜓𝑛 |𝑥 2 |𝜓𝑚 ⟩ = ∫ 𝑥 2 sin ( ) sin ( ) 𝑑𝑥 =
𝑎 𝑎 𝑎
0
𝑎
1 (𝑛 − 𝑚)𝜋𝑥 (𝑛 + 𝑚)𝜋𝑥
= ∫ 𝑥 2 [cos ( ) − cos ( )] 𝑑𝑥
𝑎 𝑎 𝑎
0
𝑎 𝑎
2
𝑢 2𝑎2 𝑥 𝑢 𝑎 3 𝑢𝜋𝑥 2 𝑢
∫ 𝑥 [cos ( 𝜋𝑥)] 𝑑𝑥 = { 2 2 cos ( 𝜋𝑥) + ( ) [( ) − 2] sin ( 𝜋𝑥)}
𝑎 𝑢 𝜋 𝑎 𝑢𝜋 𝑎 𝑎 0
0
𝑎
(𝑛 − 𝑚)𝜋𝑥
∫ 𝑥 2 [cos ( )]
𝑎
0
𝑎
2𝑎2 𝑥 𝑢 𝑎 3 𝑢𝜋𝑥 2 𝑢
= { 2 2 cos ( 𝜋𝑥) + ( ) [( ) − 2] sin ( 𝜋𝑥)}
𝑢 𝜋 𝑎 𝑢𝜋 𝑎 𝑎 0
 19

2𝑎3 𝑎 3
= { 2 2 cos(𝑢𝜋) + ( ) [(𝑢𝜋)2 − 2] ⏟
sin(𝑢𝜋)} , 𝑢 = 𝑛 − 𝑚
𝑢 𝜋 𝑢𝜋 =0

𝑎
2
(𝑛 − 𝑚)𝜋𝑥 2𝑎3
∫ 𝑥 [cos ( )] = 2 2
(−1)𝑛−𝑚
𝑎 (𝑛 − 𝑚) 𝜋
0
𝑎
2
(𝑛 + 𝑚)𝜋𝑥 2𝑎3
∫ 𝑥 [cos ( )] = 2 2
(−1)𝑛+𝑚
𝑎 (𝑛 + 𝑚) 𝜋
0

2|
1 2𝑎3 (−1)𝑛−𝑚 (−1)𝑛+𝑚
⟨𝜓𝑛 |𝑥 𝜓𝑚 ⟩ = [ − ]
𝑎 𝜋 2 (𝑛 − 𝑚)2 (𝑛 + 𝑚)2
1 2𝑎3 (𝑛2 + 𝑚2 + 2𝑛𝑚)(−1)𝑛−𝑚 (𝑛2 + 𝑚2 − 2𝑛𝑚)(−1)𝑛+𝑚
= [ − ]
𝑎 𝜋2 (𝑛 − 𝑚)2 (𝑛 + 𝑚)2
=0 =2(−1)𝑛+𝑚

1 2𝑎3 (𝑛2 + 𝑚2 ) ⏞
[(−1)𝑛−𝑚 − (−1)𝑛+𝑚 ] + (2𝑛𝑚) ⏞
[(−1)𝑛−𝑚 + (−1)𝑛+𝑚 ]
=
𝑎 𝜋2 (𝑛 − 𝑚)2 (𝑛 + 𝑚)2
[ ]

2|
2𝑎2 (4𝑛𝑚)(−1)𝑛+𝑚
⟨𝜓𝑛 |𝑥 𝜓𝑚 ⟩ = 2 [ ]
𝜋 (𝑛2 − 𝑚2 )2

2| 2|
2𝑎2 (8)(−1)3 16𝑎2
⟨𝜓1 |𝑥 𝜓2 ⟩ = ⟨𝜓2 |𝑥 𝜓1 ⟩ = 2 [ ]=−
𝜋 (1 − 4)2 9𝜋 2

Thus,

1 2 1 1 1 1
〈𝑥 2 〉 = [𝑎 [ − ] + 𝑎 2
[ − ]
2 3 2(𝜋)2 3 2(2𝜋)2

16𝑎2 𝑖(𝐸 −𝐸 )𝑡/ℏ

− 2
[𝑒
⏟ 1 2 + 𝑒 −𝑖(𝐸1 −𝐸2 )𝑡/ℏ ]]
9𝜋
2 cos(𝐸1 −𝐸2 )𝑡/ℏ
 20

(𝐸1 − 𝐸2 ) ℏ2 𝜋 2 ℏ𝜋 2
= (1 − 4) = −3 ≡ −3𝜔
ℏ 2𝑚𝑎2 ℏ 2𝑚𝑎2

2〉
𝑎2 2 5 32
〈𝑥 = [ − − cos 3𝜔]
2 3 8(𝜋)2 9𝜋 2

Therefore
2
𝑎2 2 5 32 𝑎 32
𝜎𝑥2 = 〈𝑥 2〉
− 〈𝑥 〉2 = [ − − cos 3𝜔] − ( [1 − 2 cos 3𝜔])
2 3 8(𝜋)2 9𝜋 2 2 9𝜋

𝑎2 2 5 32 𝑎2 32 2 2
32
= [ − − cos 3𝜔] − (1 + ( ) cos 3𝜔 − 2 cos 3𝜔)
2 3 8(𝜋)2 9𝜋 2 4 9𝜋 2 9𝜋 2

𝑎2 1 5 32 2
𝜎𝑥2 = [ − 2
− ( 2 ) cos2 3𝜔] (𝑐)
4 3 4(𝜋) 9𝜋

From (a), (b) and (c), write

2
2 2
ℏ2 𝑑 〈𝑥 〉
𝜎𝐻 𝜎𝑥 ≥ [ ]
4 𝑑𝑡

2 2
1 𝑎 1 5 32
𝐿𝐻𝑆 ≡ 𝜎𝐻2 𝜎𝑥2 = (3ℏ𝜔)2 [ − 2
− ( 2 ) cos 2 3𝜔]
4 4 3 4(𝜋) 9𝜋
3 2 2
1 5 32 2
= ( ) (𝑎ℏ𝜔) [ − 2
− ( 2 ) cos2 3𝜔]
4 3 4(𝜋) 9𝜋
22
ℏ2 𝑑 〈𝑥 〉 ℏ2 8ℏ ℏ 8ℏ 2 2
𝑅𝐻𝑆 ≡ [ ] = [ sin 3𝜔] = ( ) sin 3𝜔
4 𝑑𝑡 4 3𝑚𝑎 2 3𝑚𝑎
2
ℏ2 𝑑〈𝑥 〉 8 2
𝑅𝐻𝑆 ≡ [ ] = ( 2 ) (ℏ𝜔𝑎)2 sin2 3𝜔
4 𝑑𝑡 3𝜋
?
3 2 1 5 32 2
8 2
( ) (𝑎ℏ𝜔)2 [ − ⏞(
− ( 2 ) cos 2 3𝜔] ≥ ) (ℏ𝜔𝑎 )2 sin2 3𝜔
4 3 4(𝜋) 2 9𝜋 3𝜋 2
 21

?
1 5 32 2 4×8 2 2
[ − 2 ⏞(
− ( 2 ) cos 3𝜔] ≥ ) sin 3𝜔
3 4(𝜋)2 9𝜋 3 × 3𝜋 2
?
1 5 4×8 2 2 32 2 32 2
[ − ⏞(
]≥ 2
) sin 3𝜔 + ( 2 ) cos 3𝜔 = ( 2 )
3 4(𝜋)2 3 × 3𝜋 2 9𝜋 9𝜋
𝑌𝐸𝑆
1 5 32 2
[ − ⏞ (
] ≥ )
⏟3 4(𝜋)2 ⏟9𝜋 2
0.20668 0.12978
 1

Matrix Representations of Operators and Wave Functions

During 1925 and 1926 two different mathematical descriptions of quantum
mechanics were proposed.
I. The first model became known as matrix mechanics, discovered by Heisenberg
in 1925. There he associated physical quantities like position and momentum of
the particle with essentially, ∞ dimensional square matrices. The formulation in
terms of matrices is very useful in that it allows us to deal with quantities that have
no classical counterpart, like spin.
II. In 1926, Erwin Schrödinger proposed a wave theory of quantum mechanics,
based on the Schrödinger equation along the lines of the de Broglie hypothesis.

At the end of 1926, there was two different mathematical descriptions of quantum
mechanics. It was subsequently shown that the two different mathematical models
are equivalent because they may be transformed into one another.

We have described the mathematical structure of quantum mechanics in essentially the

form proposed by Schrödinger in 1926.
Now, we will consider the matrix mechanics.
The matrix formulation was built on the premise/assumption that all physical
observables must be represented by matrices.
 2

We introduced the scalar product:

∞
⟨𝜓|𝜙⟩ = ∫ 𝜓 ∗ (𝑥 )𝜙(𝑥 )𝑑𝑥
−∞

We now give some additional motivation for this assumption.

We start with a Hamiltonian H with a discrete energy spectrum
𝐻 𝜑𝑛 = 𝐸𝑛 𝜑𝑛
Its eigenfunctions {𝜑𝑛 } form a complete orthonormal set. Therefore the most general
solution to the TDSE can be written as
∞ ∞
𝑖𝐸 𝑡
− 𝑛
Ψ(𝑥, 𝑡) = ∑ 𝑐𝑛 Ψ𝑛 (𝑥, 𝑡) = ∑ 𝑐𝑛 𝜑𝑛 (𝑥 ) 𝑒 ℏ
𝑛=1 𝑛=1

Here the expansion coefficient is determined from the initial linear combination state
∞

Ψ(𝑥, 0) = ∑ 𝑐𝑛 𝜑𝑛 (𝑥 ) (∗)
𝑛=1

given by
∞

𝑐𝑛 = ∫ 𝜑𝑛∗ (𝑥 )Ψ(𝑥, 0)𝑑𝑥 (∗∗)

−∞

We can understand this situation a little differently: We take into account the fact that
the eigenfunctions {𝜑𝑛 } represent an orthonormal basis of the vector space ℋ (a
Hilbert space), like the 3 unit vectors {𝑒̂𝑖 } in the space of 3 dimensional vectors (ℝ3 ).
 3

Euclidean 3- space Hilbert space of wave

(the set ℝ3 )
functions of quantum
mechanics
Elements ⃗𝑽𝟏 , ⃗𝑽𝟐 , ⃗𝑽𝟑 , … : 3. dim 𝑣𝑒𝑐𝑡𝑜𝑟𝑠 𝝍𝟏 , 𝝍𝟐 , 𝝍𝟑 , … 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛𝑠
Basis set {𝒆̂𝒊 } {𝝋𝒏 }
0 𝑖𝑓 𝑖 ≠ 𝑗
𝑒̂𝑖 ∙ 𝑒̂𝑗 = 𝛿𝑖𝑗 = {
1 𝑖𝑓 𝑖 = 𝑗 ⟨𝜑𝑖 |𝜑𝑗 ⟩ ≡ ⟨𝑖|𝑗⟩ = 𝛿𝑖𝑗
The vectors 𝑒̂𝑥 ,𝑒̂𝑦 , 𝑒̂𝑧 are said to
span the vector space. The functions {𝜑𝑛 } are said to
span the Hilbert space ℋ.
Scalar ⃗ 𝟏∙ 𝑽
𝑽 ⃗ 𝟐 = ∑ 𝑉1𝑖 𝑉2𝑖 = a scalar ⟨𝝍𝟏 |𝝍𝟐 ⟩ = ⟨𝜓2 |𝜓1 ⟩∗
𝑖=𝑥,𝑦,𝑧
Product
= ∫ 𝜓1 ∗ 𝜓2 𝑑 3 𝑟 = a scalar

⃗ = 𝑉𝑥 𝑒̂𝑥 + 𝑉𝑦 𝑒̂𝑦 + 𝑉𝑧 𝑒̂𝑧

Completeness 𝑉 Any function Ψ in ℋ

⃗ = ∑ 𝑽𝒊 𝑒̂𝑖
𝑉 Ψ = ∑ 𝒄𝒊 𝜑𝑖 (𝑥)
𝑖=𝑥,𝑦,𝑧 𝑖

Projection
∞
⃗ ∙ 𝑒̂𝑖
𝑽𝒊 = 𝑉 𝒄𝒊 = ∫ 𝜑𝑖∗ (𝑥 )Ψ(𝑥 )𝑑𝑥
−∞

𝒄𝒊 = ⟨𝜑𝑖 |Ψ⟩
 4

Note that from

𝑽𝒙 = ⃗𝑽 ∙ 𝑒̂𝑥
⃗ ∙ 𝑒̂𝑦
𝑽𝒚 = 𝑽

𝑽𝒛 = ⃗𝑽 ∙ 𝑒̂𝑧
⃗ or 𝑽𝒙 , 𝑽𝒚 , 𝑽𝒛 , we can calculate 𝑉
it makes no difference whether we specify 𝑽 ⃗

uniquely from 𝑉𝑥 , 𝑉𝑦 , 𝑉𝑧 and vice versa if the unit vectors are known.
The situation described in (*) and (**) is quite analogous by simple replacements
⃗ → 𝜳(𝒙)
𝑉
{𝑒̂𝑖 } → {𝜑𝑛 }
𝑉𝑖 → 𝒄𝒊
For example, when {𝜑𝑖 } is known, 𝒄𝒊 can be determined uniquely, if 𝜳(𝒙) is given,
and vice versa:
∞

𝒄𝒊 = ⟨𝜑𝑖 |𝚿⟩ = ∫ 𝜑𝑖∗ (𝑥 )𝚿(𝒙)𝑑𝑥

−∞

We can thus denote the expansion coefficients 𝑐𝑖 as coordinates of a vector 𝒄, in the

form of a column matrix
𝑐1 ⟨𝜑1 |𝚿⟩
𝒄 = (𝑐2 ) = (⟨𝜑2 |𝚿⟩)
⋮ ⋮
Now consider the scalar product of two wave functions ⟨𝜓|𝜙⟩ where

𝜓(𝑥 ) = ∑ 𝑐𝑖 𝜑𝑖 (𝑥 ) and 𝜙(𝑥) = ∑ 𝑑𝑖 𝜑𝑖 (𝑥)

𝑖 𝑖
𝛿𝑖𝑗
∞ ⏞∞
∗
∗(
⟨𝜓|𝜙⟩ = ∫ 𝜓 𝑥 )𝜙(𝑥)𝑑𝑥 = ∑ 𝑐𝑖 𝑑𝑗 ∫ 𝜑𝑖∗ (𝑥 )𝜑𝑗 (𝑥)𝑑𝑥
−∞ 𝑖,𝑗 −∞
 5

⟨𝜓|𝜙⟩ = ∑ 𝑐𝑖∗ 𝑑𝑖 (3 ∗)
𝑖,𝑗

We now show that this is exactly the same result when we take the scalar
product of the two coordinate vectors 𝒄 and 𝒅.
However when we try to multiply two vectors when they are both viewed as column
matrices
𝑐1 𝑑1
𝒄 ∙ 𝒅 = (𝑐2 ) ∙ (𝑑2 ) (4 ∗)
⋮ ⋮
𝒄 ∙ 𝒅 = (𝑛 × 1) ∙ (𝑛 × 1): 𝑁𝑜𝑡 𝑑𝑒𝑓𝑖𝑛𝑒𝑑
we cannot multiply them.
To rectify this problem, we can take the transpose of the first vector, turning it into
a 1 × 𝑛 row matrix. With this change, the product is well defined. Therefore, instead
of (4*), we now have

𝑑1
†
𝑐 ∙𝑑 = (𝑐1∗ 𝑐2∗ …) ∙ (𝑑2 ) = 𝑐1∗ 𝑑1 + 𝑐2∗ 𝑑2 + ⋯ = ∑ 𝑐𝑖∗ 𝑑𝑖
⏟ ⋮ 𝑖,𝑗
(1×𝑛)∙(𝑛×1)

𝑐 † ∙ 𝑑 = ∑ 𝑐𝑖∗ 𝑑𝑖
𝑖,𝑗

Comparing this with (3*),

∞
⟨𝜓|𝜙⟩ = ∫ 𝜓 ∗ (𝑥 )𝜙(𝑥 )𝑑𝑥 = ∑ 𝑐𝑖∗ 𝑑𝑖
−∞ 𝑖,𝑗

we see that
∞
The expression ⟨𝝍|𝝓⟩ = ∫−∞ 𝝍∗ (𝒙)𝝓(𝒙)𝒅𝒙 is clearly a scalar product.
 6

We conclude that
 Any function Ψ in ℋ

Ψ = ∑ 𝒄𝒊 𝜑𝑖 (𝑥)
𝑖

is uniquely represented by the expansion coefficient 𝒄𝒊 in the given basis {𝜑𝑛 }.

 The expansion coefficient 𝒄𝒊 is viewed as a column matrix

𝑐1 ⟨𝜑1 |𝚿⟩
𝒄 = (𝑐2 ) = (⟨𝜑2 |𝚿⟩)
⋮ ⋮
Matrix Equation for the Eigenvalue Equation:
We now apply the same prescription to eigenvalue equation. We start with an
eigenvalue problem of wave mechanics, considering a wave function 𝜓(𝑥) and an
operator 𝑄̂ with a discrete spectrum:
𝑄̂ 𝜓(𝑥 ) = 𝑞 𝜓(𝑥)
OR in Dirac notation
𝑄̂ |𝜓⟩ = 𝑞|𝜓⟩
This equation can be written as a matrix equation.

The first step is to expand the wave function in terms of the eigenfunctions {𝜑𝑖 } of
the Hamiltonian (or any other basis system in ℋ):

|𝜓⟩ = ∑ 𝑐𝑛 |𝜑𝑛 ⟩ => 𝑐𝑛 = ⟨𝜑𝑛 |𝜓⟩

𝑛

Then, rewrite the eigenvalue equation

𝑄̂ |𝜓⟩ = 𝑞|𝜓⟩ => ∑ 𝑐𝑛 𝑄̂ |𝜑𝑛 ⟩ = 𝑞 ∑ 𝑐𝑛 |𝜑𝑛 ⟩

𝑛 𝑛
 7

∗ ( )
Multiply by 𝜑𝑚 𝑥 ≡ ⟨𝜑𝑚 |

̂ |𝝋𝒏 ⟩ = 𝑞 ∑ 𝑐𝑛 ⏟
∑ 𝑐𝑛 ⟨𝝋𝒎 |𝑸 ⟨𝜑𝑚 |𝜑𝑛 ⟩
𝑛 𝑛 𝛿𝑛𝑚

̂ |𝝋𝒏 ⟩ on the LHS is a number depends on n and m. We call it 𝑸𝒎𝒏

⟨𝝋𝒎 |𝑸

𝑄𝑚𝑛 ≡ ⟨𝜑𝑚 |𝑄̂ |𝜑𝑛 ⟩ = ∫ 𝜑𝑚

∗ ( )̂
𝑥 𝑄 𝜑𝑛 (𝑥 ) 𝑑𝑥

Thus

∑ 𝑸𝒎𝒏 𝑐𝑛 = 𝑞𝑐𝑚 𝐎𝐑 ℚ𝒄=𝑞𝒄

𝑛

This is a matrix equation: LHS is the product of a matrix ℚ with a column vector 𝒄
and RHS is also a column vector 𝑞𝒄.
Now, this is a matrix representation, in the basis {𝜑𝑛 }, of the eigenvalue equation we
started with
𝑄̂ |𝜓⟩ = 𝑞|𝜓⟩ →
⏟ ℚ𝒄=𝑞𝒄
𝑖𝑠
𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑒𝑑
𝑏𝑦

In basis {𝜑𝑛 }; 𝑄̂ is represented by the matrix ℚ

⟨𝜑1 |𝑄̂|𝜑1 ⟩ ⟨𝜑1 |𝑄̂|𝜑2 ⟩ ⟨𝜑1 |𝑄̂|𝜑3 ⟩ ⋯ 𝑄11 𝑄12 𝑄13 ⋯
̂ ̂ ̂ 𝑄23 ⋯
𝑄̂ ⟶
⏟ ℚ = ⟨𝜑2 |𝑄 |𝜑1 ⟩ ⟨𝜑2 |𝑄|𝜑2 ⟩ ⟨𝜑2 |𝑄 |𝜑3 ⟩ ⋯ = (𝑄21 𝑄22 )
𝑖𝑠 ⟨𝜑3 |𝑄̂ |𝜑1 ⟩ ⟨𝜑3 |𝑄̂|𝜑2 ⟩ ⟨𝜑3 |𝑄̂ |𝜑3 ⟩ ⋯ 𝑄31 𝑄32 𝑄33 ⋯
𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑒𝑑
⋮ ⋮ ⋮ ⋱ ⋮ ⋮ ⋮ ⋱
𝑏𝑦 ( )
|𝜓⟩ is represented by a column vector 𝒄

⟨𝜑1 |𝜓⟩
⟨𝜑 |𝜓⟩
|𝜓⟩ ⟶
⏟ 𝒄=( 2 )
𝑖𝑠 ⟨𝜑3 |𝜓⟩
𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑒𝑑 ⋮
𝑏𝑦
 8

• The set of numbers 𝑐𝑛 = ⟨𝜑𝑛 |𝜓⟩ constitute a complete description of the state |𝜓⟩

wrt the basis {𝜑𝑖 }.

• The set of numbers 𝑄𝑚𝑛 ≡ ⟨𝜑𝑚 |𝑄̂ |𝜑𝑛 ⟩ constitute a complete description of

operator 𝑄̂ wrt basis {𝜑𝑖 }

Now consider a Hermitian operator 𝐺̂ and suppose that {𝜑𝑛 } is the eigenfunctions of 𝐺̂
with the eigenvalues 𝑔𝑛 :
𝐺̂ |𝜑𝑛 ⟩ = 𝑔𝑛 |𝜑𝑛 ⟩
In basis {𝜑𝑛 }, the matrix representation of 𝐺̂ is obtained by
𝐺𝑚𝑛 ≡ ⟨𝜑𝑚 |𝐺̂ |𝜑𝑛 ⟩ = 𝑔𝑛 ⟨𝜑𝑚 |𝜑𝑛 ⟩ = 𝑔𝑛 𝛿𝑚𝑛
Then
𝐺11 𝐺12 𝐺13 ⋯ 𝑔1 0 0 ⋯
𝐺23 ⋯ 0 𝑔2
𝐺̂ → 𝔾 = (𝐺21 𝐺22 )=( 0 ⋯)
𝐺31 𝐺32 𝐺33 ⋯ 0 0 𝑔3 ⋯
⋮ ⋮ ⋮ ⋱ ⋮ ⋮ ⋮ ⋱
Thus, the matrix of a Hermitian operator in a basis of the eigenfunctions of that operator
is diagonal.
How to represent eigenfunctions of a diagonal matrix?
In basis {𝜑𝑛 }; the column vector representation of the eigenfunctions are
(𝑛)
|𝜑𝑛 ⟩ = 𝕀|𝜑𝑛 ⟩ = ∑ |𝜑𝑚 ⟩ ⏟
⟨𝜑𝑚 |𝜑𝑛 ⟩ => |𝜑𝑛 ⟩ = ∑ 𝑐𝑚 |𝜑𝑚 ⟩
𝑚 (𝑛) 𝑚
𝑐𝑚

where
(𝑛)
𝑐𝑚 ≡ ⟨𝜑𝑚 |𝜑𝑛 ⟩ = 𝛿𝑚𝑛 ≠ 0 𝑖𝑓 𝑚 = 𝑛
 9

(1) (2)
𝑐1 ⟨𝜑1 |𝜑1 ⟩ 1 𝑐1 ⟨𝜑1 |𝜑2 ⟩ 0
(1) (2)
⟨𝜑 |𝜑 ⟩ ⟨𝜑 |𝜑 ⟩
|𝜑1 ⟩ ⟶ 𝑐2 ≡ ( 2 1 ) = (0) , |𝜑2 ⟩ ⟶ 𝑐2 ≡ ( 2 2 ) = (1) ,
(1) ⟨𝜑3 |𝜑1 ⟩ 0 (2) ⟨𝜑3 |𝜑2 ⟩ 0
𝑐3 ⋮ 𝑐3 ⋮
⋮ ⋮
( ⋮ ) ( ⋮ )
𝑒𝑡𝑐
The matrix representation of the eigenvector |𝜑𝑛 ⟩ is a column vector with a single
nonzero unit entry in the n.th slot.
The Energy Representation
We learned that the matrix representing an operator is always diagonal in its own

basis, and eigenvectors are unit vectors in their own basis:

Thus, in energy representation (basis), Hamiltonian operator is diagonal.

In general,
𝐻|𝜓𝑛 〉 = 𝐸𝑛 |𝜓𝑛 〉

𝐻𝑘𝑛 ≡ 〈𝜓𝑘 |𝐻|𝜓𝑛 〉 = 𝐸𝑛 〈𝜓𝑘 |𝜓𝑛 〉 = 𝐸𝑛 𝛿𝑛𝑘 ⟹ 𝐻𝑘𝑛 = 𝐸𝑛 𝛿𝑛𝑘

𝐸1 0 0 ⋯ 1 0
̂ →ℍ=(0
𝐻
𝐸2 0 ⋯ ) , |𝜓 〉 → (0) , |𝜓 〉 → (1) , 𝑒𝑡𝑐.
1 2
0 0 𝐸3 ⋯ 0 0
⋮ ⋮ ⋮ ⋱ ⋮ ⋮
 10

ELEMENTARY MATRIX PROPERTIES

 Product of Two Matrices 𝐴 and 𝐵:
(𝐴 𝐵)𝑛𝑞 = ∑ 𝐴𝑛𝑝 𝐵𝑝𝑞
𝑝
 Product of Two Wavefunctions |𝜓⟩ and |𝜙⟩:
𝑏1
⟨𝜓|𝜙⟩ = ∑⟨𝜓|𝜑𝑛 ⟩⟨𝜑𝑛 |𝜓⟩ = ∑ 𝑎𝑛∗ 𝑏𝑛 = (𝑎1∗ , 𝑎2∗ , 𝑎3∗ , ⋯ ) (𝑏2 )
𝑏3
𝑛 𝑛
⋮
∗ ∗ ∗
= 𝑎1 𝑏1 + 𝑎2 𝑏2 + 𝑎3 𝑏3 + ⋯
 𝐴 : inverse of 𝐴 such that 𝐴−1 𝐴 = 𝐴𝐴−1 = 𝐼
−1

 𝐴̃ : transpose of 𝐴. The matrix elements of 𝐴̃ are obtained by reflecting the elements

𝐴𝑛𝑞 through the major diagonal of the matrix of 𝐴:
(𝐴̃) = 𝐴𝑞𝑛
𝑛𝑞
 If 𝐴̃ = 𝐴, then 𝐴 is symmetric
If 𝐴̃ = −𝐴, then 𝐴 is antisymmetric
 Trace of 𝐴 : is the sum over its diagonal elements:
𝑇𝑟(𝐴) = ∑ 𝐴𝑞𝑞
𝑞
 Hermitian adjoint of A: 𝐴 = 𝐴 † ̃∗
Matrix elements of 𝐴† are given by : ( 𝐴† )𝑞𝑛 = (𝐴̃∗ )𝑞𝑛 = (𝐴∗ )𝑛𝑞
∗ ∗
To see this: ⟨𝜑𝑞 | 𝐴† 𝜑𝑛 ⟩ = ⟨ 𝐴† 𝜑𝑛 |𝜑𝑞 ⟩ = ⟨𝜑𝑛 |𝐴𝜑𝑞 ⟩ = (𝐴∗ )𝑛𝑞
 If 𝐴† = 𝐴, 𝐴 is hermitian. OR,
( 𝐴† )𝑛𝑞 = (𝐴)𝑛𝑞 ⟹ (𝐴)𝑛𝑞 = (𝐴∗ )𝑞𝑛 ⟹ 𝐴̃∗ = 𝐴
 Unitary Matrices: 𝑈 † = 𝑈 −1 . Matrix Elements satisfy
( 𝑈† )𝑛𝑞 = ( 𝑈−1 )𝑛𝑞
⟹ 𝑈 ̃ ∗ = 𝑈 −1
∗ −1
(𝑈 )𝑞𝑛 = ( 𝑈 )𝑛𝑞

If 𝑈 is unitary , 𝑈 † 𝑈 = 𝑈 𝑈 † = 𝐼
( 𝑈† 1𝑈)𝑛𝑞 = 𝛿𝑛𝑞 = ⟨𝜑𝑛 | 𝑈 † 1𝑈|𝜑𝑞 ⟩ = ∑⟨𝜑𝑛 | 𝑈 † |𝜑𝑝 ⟩⟨𝜑𝑝 |𝑈|𝜑𝑞 ⟩
𝑝

= ∑( 𝑈 † )𝑛𝑝 (𝑈)𝑝𝑞 = ∑( 𝑈 ∗ )𝑝𝑛 (𝑈)𝑝𝑞 = 𝛿𝑛𝑞

𝑝 𝑝
 Singular Matrix A: det(A)=0.
 11

Position and Momentum Operators:

1 1
𝑥̂ = (𝑎̂ − + 𝑎̂ + ) , 𝑝̂ = (𝑎̂ − − 𝑎̂ + )
√2𝛽 𝑖√2𝛽

We find the matrices representing operators 𝑥̂ and 𝑝̂ in the so-called energy

representations. Energy representation {𝜓𝑛 } is such that Hamiltonian is represented by a
diagonal matrix
𝐻|𝜓𝑛 〉 = 𝐸𝑛 |𝜓𝑛 〉

𝐸1 0 0 ⋯ 1/2 0 0 ⋯
̂ → ℍ = ( 0 𝐸2 0 ⋯ ) = ℏ𝜔 ( 0
𝐻
3/2 0 ⋯ )
0 0 𝐸3 ⋯ 0 0 5/2 ⋯
⋮ ⋮ ⋮ ⋱ ⋮ ⋮ ⋮ ⋱
To simplify the notation we instead write
𝐻|𝑛〉 = 𝐸𝑛 |𝑛〉
Then, the matrix elements of the position operator are
1 1
𝑥𝑛𝑘 = 〈𝑛|𝑥̂|𝑘〉 = 〈𝑛|(𝑎̂ − + 𝑎̂ + )|𝑘〉 = [〈𝑛|𝑎̂ − |𝑘〉 + 〈𝑛|𝑎̂+ |𝑘〉]
√2𝛽 √2𝛽
1 − + ]
𝑥𝑛𝑘 = [𝑎̂𝑛𝑘 + 𝑎𝑛𝑘
√2𝛽

To find the matrices representing operators 𝑥̂ and 𝑝̂ in the so-called energy

representations we need to find the matrices representing operators 𝑎̂− and 𝑎̂+ in the
same representation
 12

Creation and Annihilation Operators:

The matrix elements are 〈𝑛|𝑎̂ − |𝑘〉 such that we have
⟨0|𝑎̂− |0⟩ ⟨0|𝑎̂− |1⟩ ⟨0|𝑎̂− |2⟩ ⋯ ̂−
𝑎 00 ̂−
𝑎 01 ̂−
𝑎 02 ⋯
−
− − −
⟨1|𝑎̂ |0⟩ ⟨1|𝑎̂ |1⟩ ⟨1|𝑎̂ |2⟩ ⋯ ̂−
𝑎 ̂−
𝑎 ̂−
𝑎 12 ⋯)
𝑎̂ → ( ) = ( 10 11
⟨2|𝑎̂− |0⟩ ⟨2|𝑎̂− |1⟩ ⟨2|𝑎̂− |2⟩ ⋯ ̂−
𝑎 20 𝑎−
̂ 21 ̂−
𝑎 22
⋯
⋮ ⋮ ⋮ ⋱ ⋮ ⋮ ⋮ ⋱
We know how 𝑎̂− and 𝑎̂+ operators act upon the energy eigenstates:
𝑎̂− |𝑘〉 = √𝑘|𝑘 − 1〉

𝑎̂+ |𝑘〉 = √𝑘 + 1|𝑘 + 1〉

Thus,
−
𝑎̂𝑛𝑘 = 〈𝑛|𝑎̂ − |𝑘〉 = √𝑘〈𝑛|𝑘 − 1〉 ⟹
− +
n k 𝑎̂𝑛𝑘 𝑎𝑛𝑘
−
𝑎̂𝑛𝑘 = √𝑘𝛿𝑛,𝑘−1 ≠ 0 if 𝑛 = 𝑘 − 1
0 0 0 0
+ +
𝑎𝑛𝑘 = 〈𝑛|𝑎̂ |𝑘〉 = √𝑘 + 1〈𝑛|𝑘 + 1〉 ⟹ 0 1 √1 0
0 2 0 0
⋮ ⋮ ⋮ ⋮
+
𝑎𝑛𝑘 = √𝑘 + 1𝛿𝑛,𝑘+1 ≠ 0 if 𝑛 = 𝑘 + 1 1 0 0 √1
1 1 0 0
0 √1 0 0 0 ⋯ 1 2 0
√2
0 0 √2 0 0 ⋯
−
𝑎̂ → 0 0 0 √3 0 ⋮ ⋮ ⋮ ⋮
⋯ 2 0 0 0
0 0 0 0 √4 ⋱
(⋮ ⋮ ⋮ ) 2 1 0 √2
2 2 0 0
0 0 0 0 0 ⋯ 2 3 0
√3
√1 0 0 0 0 ⋯
+
𝑎̂ → 0 √2 0 0 0 ⋯
⋮ ⋮ ⋮ ⋮
0 0 √3 0 0 ⋱ 3 0 0 0
( ⋮ ⋮ ⋮ ) 3 1 0 0
3 2 0 √3
Note that 𝑎̂− and 𝑎̂† are
3 3 0 0
non-diagonal matrices and they are 3 4 √4 0
hermitian conjugate of each other. ⋮ ⋮ ⋮ ⋮
 13

Position and Momentum Operators:

1
𝑥̂ = (𝑎̂ − + 𝑎̂+ ) ,
√2𝛽
1 1
𝑥𝑛𝑘 = 〈𝑛|𝑥̂|𝑘〉 = 〈𝑛|(𝑎̂ − + 𝑎̂+ )|𝑘〉 = [〈𝑛|𝑎̂|𝑘〉 + 〈𝑛|𝑎̂† |𝑘〉]
√2𝛽 √2𝛽

1 − + ]
𝑥𝑛𝑘 = [𝑎̂𝑛𝑘 + 𝑎𝑛𝑘
√2𝛽
This gives the matrix
0 √1
0 0 0 ⋯ 0 0 0 0 0 ⋯
1 0 0
√2 0 0 ⋯ √1 0 0 0 0 ⋯
𝑥̂ ⟶ 0 0 √3 0 ⋯ + 0
0 √2 0 0 0 ⋯
√2𝛽
0 0
0 0 √4 ⋱ 0 0 √3 0 0 ⋱
[( ⋮ ⋮ ⋮ ) ( ⋮ ⋮ ⋮ )]
0 √1 0 0 0 ⋯
√1 0 √2 0 0 ⋯
1
𝑥̂ ⟶ 0 √2 0 √3 0
√2𝛽 0 0 √3 0 ⋯
√4
0 0 0 √4 0 ⋱
( ⋮ ⋮ ⋮ )
Similarly, for the momentum operator,

1
𝑝̂ = (𝑎̂ − − 𝑎̂+ ) ,
𝑖√2𝛽
1 1
𝑝𝑛𝑘 = 〈𝑛|𝑝̂ |𝑘〉 = 〈𝑛|(𝑎̂ − − 𝑎̂+ )|𝑘〉 = [〈𝑛|𝑎̂|𝑘〉 − 〈𝑛|𝑎̂† |𝑘〉]
𝑖√2𝛽 𝑖√2𝛽

1 − + ]
𝑝𝑛𝑘 = [𝑎̂𝑛𝑘 − 𝑎𝑛𝑘
𝑖√2𝛽
 14

This gives the matrix

0 √1 0 0 0 ⋯ 0 0 0 0 0 ⋯
1 0 0 √2 0 0 ⋯ √1 0 0 0 0 ⋯
𝑝̂ ⟶ 0 √3 0 ⋯ − 0 0 0 0 ⋯
𝑖√2𝛽 0 0 √2
0 0 0 0 √4 ⋱ 0 0 √3 0 0 ⋱
[( ⋮ ⋮ ⋮ ) ( ⋮ ⋮ ⋮ )]

0 −𝑖√1 0 ⋯
0 0
𝑖√1 0 −𝑖√2 0 0 ⋯
1
𝑝̂ ⟶ 0 𝑖√2 0 −𝑖√3 0
√2𝛽 0 0 −𝑖√3 0 −𝑖√4 ⋯
0 0 0 −𝑖√4 0 ⋱
( ⋮ ⋮ ⋮ )

These matrices are hermitian and commutes with Hamiltonian.

𝑝̂ 2 1
𝐻= + 𝑚𝜔2 𝑥 2
2𝑚 2

One can show that

2 2
0 −𝑖√1 0 0 0 ⋯ 0 √1 0 0 0 ⋯
𝑖√1 0 −𝑖√2 0 0 ⋯ √1 0 √2 0 0 ⋯
1 1 1 1
0 𝑖√2 0 −𝑖√3 0 + 𝑚𝜔 2 0 √2 0 √3 0 =
2𝑚 2𝛽 2 ⋯ 2 2𝛽 2
0 0 −𝑖√3 0 −𝑖√4 0 0 √3 0 √4 ⋯
⋱
0 0 0 −𝑖√4 0 0 0 0 √4 0 ⋱
( ⋮ ⋮ ⋮ ) ( ⋮ ⋮ ⋮ )

1
0 0 ⋯
2
3
= ℏ𝜔 0 2
0 ⋯ ≡ℍ
5 ⋯
0 0
2 ⋱
(⋮ ⋮
⋮ )
 15

Example: 3.8 Suppose that the Hamiltonian of a system is represented by a Hermitian

matrix
ℎ 𝑔
ℍ=( )
𝑔 ℎ
where 𝑔 and ℎ are real constants. If the system starts out (at t=0) in an state
1
|𝜓1 〉 ≡ |1〉 → ( )
0
What is its state at time t?
The answer is the solution to the TDSE:
𝑑
𝑖ℏ |𝒢⟩ = ℍ|𝒢⟩
𝑑𝑡
As always we begin by solving the TISE
𝐻|ℊ⟩ = 𝐸|ℊ⟩ (∗)
Here
ℎ 𝑔 𝛼
𝐻→ ℍ=( ) and |ℊ⟩ → (𝛽 )
𝑔 ℎ
It is given that
1
|ℊ(0)⟩ → ( )
0

The characteristic equation determines the eigenvalues:

(𝐻 − 𝐸𝕀̂)|ℊ⟩ = 0 => 𝑑𝑒𝑡(𝐻 − 𝐸𝕀̂) = 0 =>

ℎ 𝑔 𝐸 0 ℎ−𝐸 𝑔
𝑑𝑒𝑡 [( )−( )] = 𝑑𝑒𝑡 ( ) = 0 =>
𝑔 ℎ 0 𝐸 𝑔 ℎ−𝐸

(ℎ − 𝐸 )2 − 𝑔2 = 0 => ℎ − 𝐸 = ±𝑔 => 𝐸± = ℎ ± 𝑔
 16

To determine the eigenvectors

𝛼
|ℊ⟩ = (𝛽 )

We write the eq. (*) in matrix form

ℎ 𝑔 𝛼 𝛼
( ) (𝛽 ) = 𝐸± (𝛽 )
𝑔 ℎ
𝐸 = 𝐸+ = ℎ + 𝑔

ℎ𝛼 + 𝑔𝛽 = (ℎ + 𝑔)𝛼 => 𝛼 = 𝛽
𝑔𝛼 + ℎ𝛽 = (ℎ + 𝑔)𝛽 => 𝛼 = 𝛽
1
|ℊ+ ⟩ = 𝛼 ( )
1
Normalization=>
1
1 = ⟨ℊ+ |ℊ+ ⟩ = 𝛼 2 (1 1) ( ) = 2𝛼 2 => 𝛼 = 1/√2
1
1 1
|ℊ+ ⟩ = ( ) : Normalized eigenvector of 𝐻 with 𝐸 = 𝐸+ = ℎ + 𝑔
√2 1

𝐸 = 𝐸− = ℎ − 𝑔
ℎ𝛼 + 𝑔𝛽 = (ℎ − 𝑔)𝛼 => 𝛼 = −𝛽
𝑔𝛼 + ℎ𝛽 = (ℎ − 𝑔)𝛽 => 𝛼 = −𝛽
1
|ℊ− ⟩ = 𝛼 ( )
−1
Normalization=>
1
1 = ⟨ℊ− |ℊ− ⟩ = 𝛼 2 (1 −1) ( ) = 2𝛼 2 => 𝛼 = 1/√2
−1
1 1
|ℊ− ⟩ = ( ) : Normalized eigenvector of 𝐻 with 𝐸 = 𝐸− = ℎ − 𝑔
√2 −1
 17

The set {ℊ+ , ℊ− } is orthogonal and complete:

1 1 1 1
|ℊ+ ⟩ = ( ) , |ℊ− ⟩ = ( )
√2 1 √2 −1

⟨ℊ− |ℊ+ ⟩ = ⟨ℊ+ |ℊ− ⟩ = 0

Therefore, we expand the initial state |ℊ(0)⟩ as a combination of |ℊ+ ⟩ and |ℊ− ⟩

1
|ℊ(0)⟩ = ( ) = 𝑐+ |ℊ+ ⟩ + 𝑐− |ℊ− ⟩
0
1 1 1
𝑐+ = ⟨ℊ+ |ℊ(0)⟩ = (1 1) ( ) => 𝑐+ =
√2 0 √2

1 1 1
𝑐− = ⟨ℊ− |ℊ(0)⟩ = (1 −1) ( ) => 𝑐− =
√2 0 √2

1 1
|ℊ(0)⟩ = ( ) = (|ℊ+ ⟩ + |ℊ− ⟩)
0 √2

1
|𝒢(t)⟩ = ∑ 𝑐𝑛 |ℊ𝑛 (0)⟩𝑒 −𝑖𝐸𝑛 𝑡/ℏ = (|ℊ+ ⟩𝑒 −𝑖𝐸+𝑡/ℏ + |ℊ− ⟩𝑒 −𝑖𝐸−𝑡/ℏ )
𝑛=+,−
√2

1 1 1 −𝑖(ℎ+𝑔)𝑡/ℏ 1 1
= (
( )𝑒 + ( ) 𝑒 −𝑖(ℎ−𝑔)𝑡/ℏ )
√2 √2 1 √2 −1

1 1 −𝑖(ℎ+𝑔)𝑡/ℏ 1
= (( ) 𝑒 + ( ) 𝑒 −𝑖(ℎ−𝑔)𝑡/ℏ )
2 1 −1

1 −𝑖ℎ𝑡/ℏ 1 −𝑖𝑔𝑡/ℏ 1
= 𝑒 (( ) 𝑒 + ( ) 𝑒 𝑖𝑔𝑡/ℏ )
2 1 −1

1 −𝑖ℎ𝑡/ℏ 𝑒 −𝑖𝑔𝑡/ℏ + 𝑒 𝑖𝑔𝑡/ℏ −𝑖ℎ𝑡/ℏ cos(𝑔𝑡/ℏ)

= 𝑒 ( −𝑖𝑔𝑡/ℏ ) => |𝒢(t) ⟩ = 𝑒 ( )
2 𝑒 − 𝑒 𝑖𝑔𝑡/ℏ −𝑖 sin(𝑔𝑡/ℏ)
 18

We can easily show that

I ) |𝒢(t) matches the initial state when t=0:
cos(𝑔0/ℏ) 1
|𝒢(t)⟩ = 𝑒 −𝑖ℎ0/ℏ ( )=( )
−𝑖 sin(𝑔0/ℏ) 0
II) |𝒢(t) satisfies the TDSE
𝑑
𝑖ℏ |𝒢⟩ = ℍ|𝒢⟩
𝑑𝑡
ℎ 𝑔
−𝑖 cos(𝑔𝑡/ℏ) − sin(𝑔𝑡/ℏ) ?
ℎ 𝑔 −𝑖ℎ𝑡/ℏ cos(𝑔𝑡/ℏ)
𝑖ℏ𝑒 −𝑖ℎ𝑡/ℏ
( ℏ ℏ )=
⏞( )𝑒 ( )
ℎ 𝑔 𝑔 ℎ −𝑖 sin(𝑔𝑡/ℏ)
−𝑖 (−𝑖) sin(𝑔𝑡/ℏ) − 𝑖 cos(𝑔𝑡/ℏ)
ℏ ℏ

ℎ cos(𝑔𝑡/ℏ) − 𝑖𝑔 sin(𝑔𝑡/ℏ) ⋎
−𝑖ℎ𝑡/ℏ ℎ 𝑔 −𝑖ℎ𝑡/ℏ cos(𝑔𝑡/ℏ)
𝑒 ( 𝑔𝑡 )=
⏞( )𝑒 ( )
−𝑖ℎ sin ( ) + 𝑔 cos(𝑔𝑡/ℏ) 𝑔 ℎ −𝑖 sin(𝑔𝑡/ℏ)
ℏ
 19

Projection Operators

Consider the following combinations of vectors and operators:

𝐴̂ 𝐵̂|𝜓⟩⟨𝜙|𝐶̂ |χ⟩

This expression can be interpreted in different ways:

 The operators 𝐵̂ and 𝐴̂ act successively on the vector |𝜓⟩ multiplied by the scalar

⟨𝜙|𝐶̂ |χ⟩

𝐴̂ 𝐵̂|𝜓⟩ ⟨𝜙|𝐶
⏟ ⏟ ̂ |χ⟩ = 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟
𝑎 𝑣𝑒𝑐𝑡𝑜𝑟 𝑎 𝑠𝑐𝑎𝑙𝑎𝑟

 The vector |𝜒⟩ is acted upon by the operator 𝐴̂ 𝐵̂|𝜓⟩⟨𝜙|𝐶̂ .

⏟̂ 𝐵̂|𝜓⟩⟨𝜙|𝐶̂ )
(𝐴 ⏟⟩
|𝜒 = 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟
𝑎𝑛 𝑜𝑝𝑒𝑟𝑎𝑡𝑜𝑟 𝑎 𝑣𝑒𝑐𝑡𝑜𝑟

In this case, we see that we can construct operators with the combination of a ket and

a bra, called the exterior product

̂ ≡ |𝜓⟩⟨𝜙| ,
operator ∶ 𝐷 ̂ † ≡ |𝜙⟩⟨𝜓|
adjoint operator: 𝐷

If the vectors are now chosen to be their conjugate,

⟨𝜙| = | 𝜓⟩† ≡ ⟨𝜓|

we get an important class of operators, the projection operators

𝑃̂ ≡ |𝜓⟩⟨𝜓|
 20

With the property that it picks out the portion of any other vector |𝛽 ⟩ that lies along|𝜓⟩:

𝑃̂ |𝛽 ⟩ = |𝜓⟩⟨𝜓| 𝛽 ⟩ ≡ ⟨𝜓|𝛽 ⟩|𝜓⟩

If {|𝜑𝑛 ⟩} is a discrete orthonormal basis then

⟨𝜑𝑛 |𝜑𝑘 ⟩ = 𝛿𝑛𝑘 𝑎𝑛𝑑 ∑|𝜑𝑛 ⟩⟨𝜑𝑛 | = 𝕀 ∶ identity operator

𝑛

since if we let this operator act on any function |𝜓⟩ we recover the expansion of |𝜓⟩

in the {|𝜑𝑛 ⟩} basis :

|𝝍⟩ = 𝕀|𝝍⟩ = ∑ |𝜑𝑛 ⟩ ⟨⏟𝜑𝑛 | 𝝍⟩

𝑛 𝑐𝑛

=>

|𝝍⟩⟩ = ∑ 𝑐𝑛 |𝜑𝑛 ⟩
𝑛
 21

Consider the projection operator

𝑃̂ ≡ |𝜓⟩⟨𝜓|
Let it acts on the state |𝛽 ⟩
𝑃̂ |𝛽 ⟩ = |𝜓⟩⟨𝜓| 𝛽 ⟩

One more times

|𝜓⟩⟨𝜓|
⏞̂
𝑃̂2 |𝛽 ⟩ = |𝜓⟩⟨𝜓| 𝑃 𝜓|𝜓⟩ ⟨𝜓| 𝛽 ⟩ = 𝑃̂ |𝛽 ⟩ =>
𝛽 ⟩ = |𝜓⟩ ⟨⏟
1
𝑃 |𝛽 ⟩ = 𝑃̂|𝛽 ⟩ => 𝑃̂2 = 𝑃̂
̂2

The eigenvalue equation

𝑃̂ |𝜂 ⟩ = 𝜆|𝜂 ⟩

𝑃̂2 |𝜂 ⟩ = 𝜆 𝑃̂|𝜂 ⟩ = 𝜆2 |𝜂 ⟩ => 𝜆 = 𝜆2 => 𝜆 = 0,1

𝜆 = 1: =>
𝑃̂ |𝜂 ⟩1 = |𝜂 ⟩1

Any complex multiple of |𝜂 ⟩1 is an eigenvector of 𝑃̂ with the eigenvalue 𝜆 = 1

𝜆 = 0: =>
𝑃̂ |𝜂 ⟩0 = 0
Any vector orthogonal to |𝜂 ⟩1 is an eigenvector of 𝑃̂ with the eigenvalue 𝜆 = 0.
 22

Hermitian conjugate (or adjoint) of an operator:

Hermitian adjoint of an operator 𝐴̂ is written 𝐴̂† . For an 𝑛 × 𝑛 square matrix, to
construct 𝐴̂† , one first forms the complex conjugate of 𝐴̂ and then transposes:
𝐴̂† = 𝐴̃∗
In other words, the matrix representing the operator 𝐴̂ , 𝔸, and its hermitian adjoint
matrix 𝔸† are given by
𝐴11 𝐴12 𝐴13 ⋯ 𝐴1𝑛 ∗
𝐴11 𝐴∗21 𝐴∗31 ⋯ 𝐴∗𝑛1
𝐴21 𝐴22 𝐴23 ⋯ 𝐴2𝑛 ∗
𝐴12 𝐴∗22 𝐴∗32 ⋯ 𝐴∗𝑛2
𝔸 = 𝐴31 𝐴32 𝐴33 ⋯ 𝐴3𝑛 , 𝔸† = 𝐴13
∗
𝐴∗23 𝐴∗33 ⋯ 𝐴∗𝑛3
⋮ ⋮ ⋮ ⋱ ⋮ ⋮ ⋮ ⋮ ⋱ ⋮
∗
(𝐴𝑛1 𝐴𝑛2 𝐴𝑛3 … 𝐴𝑛𝑛 ) (𝐴1𝑛 𝐴∗2𝑛 𝐴∗3𝑛 … ∗
𝐴𝑛𝑛 )
Matrix elements of 𝐴̂† in a basis system {𝜑𝑛 } are given by :
∗ ∗
( 𝐴† )𝑚𝑛 ≡ ⟨𝜑𝑚 | 𝐴̂† 𝜑𝑛 ⟩ = ⟨ 𝐴̂† 𝜑𝑛 |𝜑𝑚 ⟩ = ⟨𝜑𝑛 |𝐴̂𝜑𝑚 ⟩ = (𝐴∗ )𝑛𝑚
i.e.,
( 𝐴† )𝑚𝑛 = (𝐴̃∗ )𝑚𝑛 = (𝐴∗ )𝑛𝑚
If 𝐴̂† = 𝐴̂, 𝐴̂ is a hermitian operator. OR,
( 𝐴† )𝑛𝑚 = (𝐴)𝑛𝑚 ⟹ (𝐴)𝑛𝑚 = (𝐴∗ )𝑚𝑛 ⟹ 𝐴̃∗ = 𝐴
Unitary Operators
If the Hermitian adjoint 𝑈 ̂ † of an operator 𝑈
̂ is equal to 𝑈̂ −1 , the inverse of 𝑈
̂, i.e.,
̂† = 𝑈
𝑈 ̂ −1
then 𝑈 ̂ is said to be unitary. The matrix elements of 𝑈 ̂ satisfy the relations
( 𝑈 † )𝑛𝑚 = ( 𝑈 −1 )𝑛𝑚
⟹ 𝑈 ̃ ∗ = 𝑈 −1
∗ −1
(𝑈 )𝑚𝑛 = ( 𝑈 )𝑛𝑚

̂ is unitary ,
If 𝑈
̂†𝑈
𝑈 ̂=𝑈
̂𝑈̂ † = 𝕀̂
Matrix elements:
̂ † 𝕀̂𝑈
( 𝑈 † 𝑈)𝑛𝑚 = 𝛿𝑛𝑚 = ⟨𝜑𝑛 | 𝑈 ̂|𝜑𝑚 ⟩ = ∑⟨𝜑𝑛 | 𝑈
̂ † |𝜑𝑝 ⟩⟨𝜑𝑝 |𝑈
̂|𝜑𝑚 ⟩
𝑝

= ∑( 𝑈 † )𝑛𝑝 (𝑈)𝑝𝑚 = ∑( 𝑈 ∗ )𝑝𝑛 (𝑈)𝑝𝑚 = 𝛿𝑛𝑚

𝑝 𝑝
 23

Unitary and Similarity Transformations

With unitary operators we can define unitary transformations of states and operators.
Common notations are
̂ |𝜓⟩ and
|𝜓⟩ → |𝜓 ′ ⟩ = 𝑈 𝐴̂ → 𝐴̂′ = 𝑈
̂ 𝐴̂ 𝑈
̂†

The interesting thing about the unitary transformations is that they leave important
properties and quantities unchanged. In this respect, a unitary transformation is an
analogue of the rotation in elementary vector calculus. One can obtain a new
orthogonal basis in 3-space through a rotation of axes about the origin. In this new
basis,

 The lengths of vectors

 The angle between the vectors
 The scalar products
 The matrix elements
 The eigenvalues
are all unchanged.

We already know that a given Hilbert space has many basis. The transformation from
one basis to another is a unitary transformation. It has all the properties listed above
that a rigid rotation in 3-space has.
 24

Transformation of Basis:
Let {𝜑𝑛 } and {𝑓𝑛 } are two basis :
⟨𝜑𝑛 |𝜑𝑘 ⟩ = 𝛿𝑛𝑘 and ∑|𝜑𝑛 ⟩⟨𝜑𝑛 | = 𝕀̂
𝑛
and
⟨𝑓𝑛 |𝑓𝑘 ⟩ = 𝛿𝑛𝑘 and ∑|𝑓𝑛 ⟩⟨𝑓𝑛 | = 𝕀̂
𝑛
Now, write
|𝑓𝑛 ⟩ = 𝕀̂|𝑓𝑛 ⟩ = ∑|𝜑𝑝 ⟩ ⟨𝜑
⏟ 𝑝 |𝑓𝑛 ⟩
𝑝 ∗
⟨𝑓𝑛 |𝜑𝑝 ⟩

̂
with the matrix elements of a operator 𝑈
∗
𝑈𝑛𝑝 ≡ ⟨𝑓𝑛 |𝜑𝑝 ⟩ and then 𝑈𝑛𝑝 ≡ ⟨𝜑𝑝 |𝑓𝑛 ⟩
Then

∗
|𝑓𝑛 ⟩ = ∑ 𝑈𝑛𝑝 |𝜑𝑝 ⟩ ∶ Transformation of basis
𝑝

̂ is unitary, i.e., 𝑈
Let us prove that 𝑈 ̂†𝑈
̂=𝑈
̂𝑈̂ † = 𝕀̂

∗
⟨𝑓𝑖 |𝑓𝑗 ⟩ = 𝛿𝑖𝑗 ⟹ ⟨∑ 𝑈𝑖𝑘 𝜑𝑘 | ∑ 𝑈𝑗𝑙∗ 𝜑𝑙 ⟩ = 𝛿𝑖𝑗 ⟹ ∑ 𝑈𝑖𝑘 𝑈𝑗𝑙∗ ⟨𝜑𝑘 |𝜑𝑙 ⟩ = 𝛿𝑖𝑗
𝑘 𝑙 𝑘

∑ 𝑈𝑖𝑘 𝑈𝑗𝑙∗ 𝛿𝑘𝑙 = 𝛿𝑖𝑗 ⟹ ∑ 𝑈𝑖𝑘 𝑈𝑗𝑘

∗
= 𝛿𝑖𝑗
𝑘 𝑘

∗
̃𝑘𝑗 ̂𝑈̂ † = 𝕀̂ . Therefore U is a unitary matrix.
⟹ ∑ 𝑈𝑖𝑘 𝑈 = 𝛿𝑖𝑗 𝑂𝑅 𝑈
𝑘

̂.
{𝝋𝒏 } and {𝒇𝒏 } are related through the unitary transformation 𝑼
 25

Transformation of State Vectors:

|𝜓⟩: An arbitrary state vector

In {𝜑𝑛 } basis:
|𝜓⟩ = ∑ 𝜓𝑛 |𝜑𝑛 ⟩ ⟹ 𝜓𝑛 = ⟨𝜑𝑛 |𝜓⟩
𝑛
In {𝑓𝑛 } basis:
|𝜓⟩ = ∑ 𝜓𝑛′ |𝑓𝑛 ⟩ ⟹ 𝜓𝑛′ = ⟨𝑓𝑛 |𝜓⟩
𝑛
We can show that
̂ |𝜓⟩
|𝜓 ′ ⟩ = 𝑈
Proof:

Two basis are transformed with unitary transformation

∗
|𝑓𝑛 ⟩ = ∑ 𝑈𝑛𝑝 |𝜑𝑝 ⟩
𝑝
Then

𝜓𝑛′ = ⟨𝑓𝑛 |𝜓⟩ = ⟨∑𝑝 𝑈𝑛𝑝

∗
𝜑𝑝 |𝜓⟩ = ∑ 𝑈𝑛𝑝 ⟨𝜑𝑝 |𝜓⟩ = ∑ 𝑈𝑛𝑝 𝜓𝑝 =>
𝑝 𝑝

𝜓𝑛′ = ∑ 𝑈𝑛𝑝 𝜓𝑝
𝑝
This is the matrix representation of the equation

̂ |𝜓⟩
|𝜓 ′ ⟩ = 𝑈
That is,
⏟ ̂ |𝜓⟩ ⟶
|𝜓 ′ ⟩ = 𝑈 𝜓𝑛′ = ∑ 𝑈𝑛𝑝 𝜓𝑝
𝑇ℎ𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 ⏟ 𝑝
𝑖𝑡𝑠 𝑚𝑎𝑡𝑟𝑖𝑥 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑎𝑡𝑖𝑜𝑛
with 𝑈𝑛𝑝 ≡ ⟨𝑓𝑛 |𝜑𝑝 ⟩

This eq. says how an arbitrary state vector transforms under a change of basis.
 26

Problem 11.12 (Liboff): Show that the scalar product ⟨𝜓|𝜙⟩ is preserved under a
unitary transformation.

Let |𝜓⟩ and |𝜙⟩ be two arbitrary vectors in Hilbert space. Under the transformation
̂ of basis
𝑈
̂
𝑈
{𝜑𝑛 } ⟶
⏞ {𝑓𝑛 }
State vectors transform
̂ |𝜓⟩
|𝜓⟩ ⟶ |𝜓 ′ ⟩ = 𝑈
̂ |𝜙⟩
|𝜙⟩ ⟶ |𝜙 ′ ⟩ = 𝑈
How the scalar product ⟨𝜓|𝜙⟩ transform?
̂ 𝜓|𝑈
⟨𝜓|𝜙⟩ ⟶ ⟨𝜓 ′ |𝜙 ′ ⟩ = ⟨𝑈 ̂ 𝜙⟩ = ⟨𝜓| 𝑈
̂†𝑈
̂ 𝜙⟩ = ⟨𝜓|𝕀̂𝜙⟩ = ⟨𝜓|𝜙⟩

The scalar product is invariant under unitary transformation.

 27

Transformation of Operators: Unitary-Similarity Transformation.

A typical quantum mechanical eq. appears as
|𝜓⟩ = 𝐹̂ |𝜙⟩ (∗)
Consider two basis {𝜑𝑛 } and {𝑓𝑛 } and they are connected by a unitary
̂:
transformation 𝑈
∗
|𝑓𝑛 ⟩ = ∑ 𝑈𝑛𝑝 |𝜑𝑝 ⟩
𝑝

Let in {𝜑𝑛 } basis:

|𝜓⟩ = 𝐹̂ |𝜙⟩
In {𝑓𝑛 } basis:
|𝜓 ′ ⟩ = 𝐹̂ ′ |𝜙 ′ ⟩
where
̂ |𝜓⟩ and |𝜙 ′ ⟩ = 𝑈
|𝜓 ′ ⟩ = 𝑈 ̂ |𝜙⟩

𝐹̂ ′ =?
From
|𝜓 ′ ⟩ = 𝐹̂ ′ |𝜙 ′ ⟩ ⟹ 𝑈
̂ |𝜓⟩ = 𝐹̂ ′ 𝑈
̂ |𝜙⟩
̂ −1 :
Multiply from the left by 𝑈
̂ −1 𝑈
⟹ ⏟
𝑈 ̂ |𝜓⟩ = 𝑈
̂ −1 𝐹̂ ′ 𝑈 |𝜙⟩ ̂ −1 𝐹̂ ′ 𝑈
⟹ |𝜓⟩ = 𝑈 ̂ |𝜙⟩
𝕀̂

Compare this with

|𝜓⟩ = 𝐹̂ |𝜙⟩
which implies that

𝐹̂ = 𝑈
̂ −1 𝐹̂ ′ 𝑈
̂ 𝑂𝑅 𝐹̂ ′ = 𝑈
̂ 𝐹̂ 𝑈
̂ −1 ∶ Unitary-similarity Transformation
The more general case of transformation:
𝐴̂ ⟶ 𝐴̂′ = 𝑆̂𝐴̂ 𝑆̂ −1 𝑠𝑢𝑐ℎ 𝑡ℎ𝑎𝑡 𝑆̂ −1 ≠ 𝑆̂ † .Then 𝑆̂ is similarity transformation
 28

----------------------------------------------------------------------------------------
A Hermitian matrix is always diagonalizable by a unitary-similarity transformation.
Let 𝐴̂ be a Hermitian 𝑛 × 𝑛 matrix. Let the column vectors of the 𝑛 × 𝑛 matrix 𝑆̂ be
comprised of the orthonormalized eigenvectors of 𝐴̂. Then
 𝑆̂ is unitary.
 𝑆̂ −1 𝐴̂𝑆̂ is a diagonal matrix comprised of the eigenvalues of 𝐴̂.
EX: Consider the matrix
0 0 1
𝐴 = (0 0 0)
1 0 0
Find the eigenvalues and corresponding eigenvectors.
(𝐴 − 𝜆𝕀̂)𝜓 = 0 => 𝑑𝑒𝑡(𝐴 − 𝜆𝕀̂) = 0 => 𝜆1 = 0, 𝜆2 = 1, 𝜆3 = −1
𝑎
𝜆1 = 0: 𝜓1 = (𝑏 )
𝑐
0 0 1 𝑎
(0 0 0) (𝑏 ) = 0 => 𝑐 = 0, 𝑎 = 0
1 0 0 𝑐
0
𝜓1 = (1)
0
𝑎
𝜆2 = 1: 𝜓2 = (𝑏 )
𝑐
0 0 1 𝑎 𝑎
(0 0 0) (𝑏 ) = (𝑏 ) => 𝑐 = 𝑎, 𝑏 = 0
1 0 0 𝑐 𝑐
11
𝜓2 = ( 0)
√2 1
 29

𝑎
𝜆3 = −1: 𝜓3 = (𝑏 )
𝑐

0 0 1 𝑎 𝑎
(0 0 0) (𝑏 ) = − (𝑏 ) => 𝑐 = −𝑎, 𝑏 = 0
1 0 0 𝑐 𝑐
−11
𝜓3 = (0)
√2 1
Then, construct the matrix whose columns are comprised of orthonormalized
eigenvectors of 𝐴̂:

0 1/√2 −1/√2 0 1/√2 −1/√2

𝕊= ( 1) ( 0 ) ( 0 ) = (1 0 0 )
⏟ 0 ⏟1/√2 ⏟ 1/√2 0 1/√2 1/√2
( 𝜓1 𝜓2 𝜓3 )
0 1 0
𝕊† = ( 1/√2 0 1/√2)
−1/√2 0 1/√2
0 1 0 0 0 1 0 1/√2 −1/√2
𝕊† 𝐴𝕊 = ( 1/√2 0 1/√2) (0 0 0) (1 0 0 )
−1/√2 0 1/√2 1 0 0 0 1/√2 1/√2

0 1 0 0 1/√2 1/√2 0 0 0
= ( 1/√2 0 1/√2) (0 0 0 ) = (0 1 0)
−1/√2 0 1/√2 0 1/√2 −1/√2 0 0 −1
It is a diagonal matrix whose diagonal elements are eigenvalues (0,1,-1) of A. Thus, the
unitary matrix 𝕊 diagonalizes the matrix A.
 30

Invariance of Eigenvalues
Since the eigenvalues of an operator corresponding to an observable are physically
measurable quantities, one doesn’t expect these values to be affected by a
transformation of basis in Hilbert space. For example, the eigenenergies of SHO are
𝐸𝑛 = ℏ𝜔(𝑛 + 1/2)
in all representations. In a similar vein, the eigenvalues of a Hermitian operator must
be real. It follows that
(1) The eigenvalues of a Hermitian operator are preserved under a unitary-
similarity transformation.
(2) The Hermiticity of an operator is preserved under a unitary-similarity
transformation.

Proof of (1): Consider an eigenvalue equation

𝐴̂|𝜓⟩ = 𝑎 |𝜓⟩
Consider two basis {𝜑𝑛 } and {𝑓𝑛 } and they are connected by a unitary
̂:
transformation 𝑈
∗
|𝑓𝑛 ⟩ = ∑ 𝑈𝑛𝑝 |𝜑𝑝 ⟩
𝑝

Let
In {𝜑𝑛 } basis:
𝐴̂|𝜓⟩ = 𝑎 |𝜓⟩
In {𝑓𝑛 } basis:

̂ |𝜓⟩ => |𝜓⟩ = 𝑈

|𝜓 ′ ⟩ = 𝑈 ̂ −1 |𝜓 ′ ⟩
Then
𝐴̂ 𝑈
̂ −1 |𝜓 ′ ⟩ = 𝑎 𝑈 ̂ 𝐴̂𝑈
̂ −1 |𝜓 ′ ⟩ => 𝑈 ̂ −1 |𝜓 ′ ⟩ = 𝑎 |𝜓 ′ ⟩
 31

Which implies that both 𝐴̂ and its similarity transformation 𝑈

̂ 𝐴̂𝑈
̂ −1 have the same
̂ 𝐴̂𝑈
eigenvalue 𝑎. The eigenvectors of 𝑈 ̂ −1 are
̂ |𝜓⟩
|𝜓 ′ ⟩ = 𝑈
̂ to be unitary for this statement to be true. For any
Note that it is not necessary for 𝑈
operator 𝐹̂ , ….

Proof of (2): Let 𝐴̂ be hermitian, 𝐴̂† = 𝐴̂. Under unitary transformation,

𝐴̂ → 𝐴̂′ = 𝑈
̂ 𝐴̂𝑈
̂ −1
Then

𝐴̂† → 𝐴̂†′ = (𝑈 ̂ −1 )† = (𝑈
̂ 𝐴̂𝑈 ̂ −1 )† 𝐴̂† 𝑈
̂†

̂ is unitary, 𝑈
Since 𝑈 ̂ −1 = 𝑈 ̂ † )† = 𝑈
̂ † and ( 𝑈 ̂. Thus,

𝐴̂† → 𝐴̂†′ = 𝑈
̂ 𝐴̂𝑈
̂ −1 = 𝐴̂
 32

Example 1: The Hamiltonian 𝐻 for a certain three-level system is represented by the

matrix

0 0 0
ℍ = ℏ𝜔 (0 1 0)
0 0 −1

What values will we obtain when measuring the energy and with what probabilities?

------------------------------------------------------------------------------------------------

The results of the energy measurement are given by the eigenvalues 𝐸𝑛 of ℍ.

𝐻 |𝜓𝑛 ⟩ = 𝐸𝑛 |𝜓𝑛 ⟩ , 𝑛 = 1,2,3.

It is a diagonal matrix and its eigenvalues are the diagonal elements.

𝐸1 = 0, 𝐸2 = ℏ𝜔, 𝐸3 = −ℏ𝜔

Which of these possible results of the energy values will appear after the

measurements depend on the state in which we make the measurements.

Case1: Measure the energy in one of the eigenstates of H, |𝜓1 ⟩, |𝜓2 ⟩, |𝜓3 ⟩.

1 0 0 0 1 1
𝐸1 = 0 in state |𝜓1 ⟩ = (0) with 𝑃 = 100%, since ℏ𝜔 (0 1 0 ) (0) = 0 (0)
0 0 0 −1 0 0

0 0 0 0 0 0
𝐸2 = ℏ𝜔 in state |𝜓2 ⟩ = (1) 𝑤𝑖𝑡ℎ 𝑃 = 100%, since ℏ𝜔 (0 1 0 ) (1) = ℏ𝜔 (1)
0 0 0 −1 0 0

0 0 0 0 0 0
𝐸3 = −ℏ𝜔 in state |𝜓3 ⟩ = (0) 𝑤𝑖𝑡ℎ 𝑃 = 100%, since ℏ𝜔 (0 1 0 ) (0) = −ℏ𝜔 (0)
1 0 0 −1 1 1
 33

Case 2: Measure the energy in an arbitrary state, say

1 1−𝑖
|𝜙⟩ = (1 − 𝑖 )
√5 1
Since the set {|𝜓1 ⟩, |𝜓2 ⟩, |𝜓3 ⟩} forms an orthonormal system, we write |𝜙⟩ in terms
of them,

1−𝑖 1 1−𝑖 0 1 0
|𝜙⟩ = 𝑐1 |𝜓1 ⟩ + 𝑐2 |𝜓2 ⟩ + 𝑐3 |𝜓3 ⟩ = (0 ) + ( 1) + (0 )
√5 0 √5 0 √5 1

2
1−𝑖 2
𝑃(𝐸 = 𝐸1 ) = |𝑐1 |2 = | | =
√5 5
2
1−𝑖 2
𝑃(𝐸 = 𝐸2 ) = |𝑐2 |2 =| | =
√5 5
2
1 1
𝑃(𝐸 = 𝐸3 ) = |𝑐3 |2 =| | =
√5 5
 34

Example 2: Consider a system whose Hamiltonian 𝐻 is represented by the matrix

1 −1 0
ℍ = 𝜀 (−1 1 0)
0 0 −1

where 𝜀 has the dimension of energy.

a) What values will we obtain when measuring the energy and with what

probabilities?

The results of the energy measurement are given by the eigenvalues 𝐸𝑛 of ℍ.

𝜀−𝐸 −𝜀 0
(ℍ − 𝐸𝕀̂)𝜓 = 0 => 𝑑𝑒𝑡(ℍ − 𝐸𝕀̂) = 0 => 𝑑𝑒𝑡 ( −𝜀 𝜀−𝐸 0 )=0
0 0 −(𝜀 + 𝐸)
−(𝜀 − 𝐸 )(𝜀 − 𝐸 )(𝜀 + 𝐸 ) + 𝜀 2 (𝜀 + 𝐸 ) = 0 => (𝜀 + 𝐸 )[𝜀 2 − (𝜀 − 𝐸 )2 ] = 0
𝐸1 = 0 , 𝐸2 = −𝜀 , 𝐸3 = 2𝜀
The corresponding energy eigenstates:
𝑎
𝐸1 = 0: 𝜓1 = (𝑏 )
𝑐
1 −1 0 𝑎
𝜀 (−1 1 0 ) (𝑏 ) = 0 => 𝑎 − 𝑏 = 0, 𝑏 − 𝑎 = 0, 𝑐 = 0
0 0 −1 𝑐
11
𝜓1 = (1 )
√2 0
𝑎
𝐸2 = −𝜀: 𝜓2 = (𝑏 )
𝑐
1 −1 0 𝑎 𝑎
𝜀 (−1 1 0 ) (𝑏 ) = −𝜀 (𝑏 ) => 𝑎 − 𝑏 = −𝑎, 𝑏 − 𝑎 = −𝑏, 𝑐 = 𝑐
0 0 −1 𝑐 𝑐
 35

0
𝜓 2 = (0 )
1
𝑎
𝐸3 = 2𝜀: 𝜓3 = (𝑏 )
𝑐
1 −1 0 𝑎 𝑎
𝜀 (−1 1 0 ) (𝑏 ) = 2𝜀 (𝑏 ) => 𝑎 − 𝑏 = 2𝑎, 𝑏 − 𝑎 = 2𝑏, 𝑐 = −2𝜀𝑐
0 0 −1 𝑐 𝑐
1 −1
𝜓3 = (1)
√2 0

the set {|𝜓1 ⟩, |𝜓2 ⟩, |𝜓3 ⟩} forms an orthonormal system,

If we measure the energy in one of the eigenstates of H, |𝜓1 ⟩, |𝜓2 ⟩, |𝜓3 ⟩.

1 1 1 −1 0 1 1 1 1
𝐸1 = 0 in state |𝜓1 ⟩ = (1) with 𝑃 = 100%, since 𝜀 (−1 1 0) (1 ) = 0 (1)
√2 0 0 0 −1 √2 0 √2 0

0 1 −1 0 0 0
𝐸2 = −𝜀 in state |𝜓2 ⟩ = (0) with 𝑃 = 100%, since 𝜀 (−1 1 0 ) (0) = −𝜀 (0)
1 0 0 −1 1 1

1 −1 1 −1 0 1 −1 1 −1
𝐸3 = 2𝜀 in state |𝜓3 ⟩ = ( 1 ) with 𝑃 = 100%, since 𝜀 (−1 1 0) ( 1 ) = 2𝜀 (1)
√2 0 0 0 −1 √2 0 √2 0

b) Now consider another observable of the same system, 𝐴, represented by the matrix
0 4 0
𝔸 = 𝛼 (4 0 1)
0 1 0
Suppose that when we measure the energy, we obtain a value of −𝜀. Immediately

afterwards, we measure 𝐴. What values will we obtain for A and with what

probabilities?
 36

The results of the measurement of A gives one of its eigenvalues

𝐴̂𝜙 = 𝜆𝜙

Therefore, we first find 𝜆.

−𝜆 4𝛼 0
(𝔸 − 𝜆𝕀̂)𝜙 = 0 => 𝑑𝑒𝑡(𝔸 − 𝜆𝕀̂) = 0 => 𝑑𝑒𝑡 ( 4𝛼 −𝜆 𝛼𝛼 ) = 0
0 𝛼 −𝜆

(−𝜆)(𝜆2 − 𝛼 2 ) − 4𝛼 (−4𝛼𝜆) = 0 => 𝜆1 = 0, 𝜆2 = √17𝛼, 𝜆3 = −√17𝛼

The corresponding eigenvectors:

𝑎
𝜆1 = 0: 𝜙1 = (𝑏 )
𝑐

0 4 0 𝑎
𝛼 (4 0 1) (𝑏 ) = 0 => 4𝛼𝑏 = 0, 4𝛼𝑎 + 𝛼𝑐 = 0, 𝛼𝑏 = 0
0 1 0 𝑐

1 1
𝜙1 = (0)
√17 −4

𝑎
𝜆2 = √17𝛼: 𝜙2 = (𝑏 )
𝑐

0 4 0 𝑎 𝑎
𝛼 (4 0 1) (𝑏 ) = √17𝛼 (𝑏 ) => 4𝑏 = √17𝑎, 4𝑎 + 𝑐 = √17𝑏, 𝑏 = √17𝑐
0 1 0 𝑐 𝑐

1 4
𝜙2 = (√17)
√34 1
 37

𝑎
𝜆3 = −√17𝛼: 𝜙3 = (𝑏 )
𝑐
0 4 0 𝑎 𝑎
𝛼 (4 0 1) (𝑏 ) = −√17𝛼 (𝑏 ) => 4𝑏 = −√17𝑎, 4𝑎 + 𝑐 = −√17𝑏, 𝑏 = −√17𝑐
0 1 0 𝑐 𝑐
1 4
𝜙3 = (−√17)
√34 1

The set {|𝜙1 ⟩, |𝜙2 ⟩, |𝜙3 ⟩} forms an orthonormal system. (Show this)They are the
determinate states of A.
If a measurement of the energy yields 𝐸2 = −𝜀, this means that the system is left in
the state 𝜓2 . The state 𝜓2 is not a determinate state for the observable A. However,
we can expand it in terms of the determinate states of A:
|𝜓2 ⟩ = 𝑐1 |𝜙1 ⟩ + 𝑐2 |𝜙2 ⟩ + 𝑐3 |𝜙3 ⟩
1 0 1 −4
𝑐1 = ⟨𝜙1 |𝜓2 ⟩ = (1 0 −4) (0) = (−4) => 𝑐1 =
√17 1 √17 √17

1 0 1 1
𝑐2 = ⟨𝜙2 |𝜓2 ⟩ = (4 √17 1) (0) = (1) => 𝑐2 =
√34 1 √34 √34

1 0 1 1
𝑐3 = ⟨𝜙3 |𝜓2 ⟩ = (4 −√17 1) (0) = (1) => 𝑐3 =
√34 1 √34 √34

−4 2 16
𝑃(𝐴 = 0 in state 𝜓2 ) = |𝑐1 |2 =| | =
√17 17
2
1 1
𝑃(𝐴 = √17𝛼 in state 𝜓2 ) = |𝑐2 |2 =| | =
√34 34
2
1 1
𝑃(𝐴 = −√17𝛼 in state 𝜓2 ) = |𝑐3 |2 =| | =
√34 34