Quantum Mechanics PyEd 342

Quantum Mechanics (PyEd 342)
Chapter 1. The Motivation for Quantum Mechanics
1.1. What is Quantum Mechanics?

Quantum Mechanics is a theory that deals with and predicts the behavior of
microscopic objects like molecules, atoms, electrons and photons. Contrary to
Classical Mechanics, Quantum Mechanics is based on radical ideas that the
physical properties of a system are essentially discontinuous (quantized) and
probabilistic (not deterministic).
The development of quantum mechanics was initially motivated by several

observations which demonstrated the inadequacy of classical physics.
1.2. The Breakdown of Classical Physics

Let’s first consider Classical Mechanics which has three formalisms:
F = ma, Newtonian Mechanics
d ∂L ∂L
− = 0, Lagrangian Mechanics, L = T − U
dt ∂q& ∂q
∂H ∂H
= − p& , = q& Hamiltonian Mechanics, H = T + U
∂q ∂p
All three formalisms have the same general goal: to find the equation of motion of
point-like objects and then determine the position and momentum as functions of
time.
In the case of Newtonian formalism of mechanics, the motion will be determined

if the forces are known. The Lagrangian and Hamiltonian approaches use energy
concepts to arrive at the equations of motion and make solving problems simpler.
When Classical Mechanics is applied, in one of its formalisms, to study visually
big objects or system of objects such as ballistics, simple or coupled pendula,
harmonic oscillators, it works fine. But what if we try to apply Classical
Mechanics (N or L or H formalism) to an electron in an atom, to a proton in a
nucleus, to a quark in a proton, to a photon in a laser beam? It breaks down. This
failure of classical mechanics to describe microscopic objects was discovered by
physicists around the turn of the twentieth century. Classical Mechanics seemed to
have a limit!
Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 1

Think:
Is the failure of Classical Mechanics due to incorrect equations? Is it possible to
make the correct modifications to the equations so it is applicable to all objects
irrespective of their size, like it was done in relativity at high speeds?
At this point, it is useful to look at some of the failures Classical Mechanics to

appreciate how physicists of the time concoct new ideas to solve the puzzles and
failures piecemeal and develop Quantum Mechanics over a period of two decades.
1. The ultraviolet catastrophe:
One of the failures of Classical Mechanics is its inadequacy to solve the

problem of the Blackbody Radiation. A blackbody is an idealized object which
absorbs and emits all frequencies. Imagine the blackbody as a cavity whose
inner walls are at a temperature T. The total intensity of the radiation emitted
by the walls at the temperature T is given empirically by the Stefan-Boltzmann
law:
I = σT 4 (1.1)
However, the intensity is not uniformly distributed over all frequencies of the
radiation. So, we can ask: “What is the energy within a unit volume of the
cavity at some given frequencyν, at absolute temperature T?”
In 1900, Rayleigh used the laws of classical thermodynamics and classical

electromagnetism and got the energy density as a function of frequency:
 8πν 2 
u (ν , T ) =  3  kT (1.2)
 c 
Where k is Boltzmann constant and c is the speed of light. Equation (1.2) is

called Rayleigh-Jeans formula.
A graph of the energy density u versus frequency ν is shown below both for the
Rayleigh-Jeans formula and the experimental data. The area under the
experimental curve is finite indicating that the total energy within the cavity is
finite at finite temperature. But the area under the R-J curve is infinite. That is,
when the R-J formula is integrated over all frequencies, it blows up; meaning
the cavity at finite temperature has an infinite total energy inside! The problem
lies in the UV region (high-frequency region) where the R-J curve does not fit
the experimental curve. This disaster was called the Ultraviolet Catastrophe.

So, according to classical physics (R-J

u (ν , T ) R-J
formula), the energy density of the
radiation within the cavity of the
blackbody is infinite due to a divergence
of energy carried by high-frequency Expt
modes. This is in complete contradiction
with experimental results which show no
such divergence, and the total energy
density is finite. The predictions of ν
classical physics were clearly dead wrong!
In the same year (1900), Max Planck obtained a formula that fits the
experimental data for all frequencies and solved the UV catastrophe:
8π hν 3
u (ν , T ) = 3 hν kT (1.3)
c e −1
Where h = 6.626 ×10−34 J .s is a constant, now called Planck's constant.
Max Planck arrived at this result by combining statistical mechanics with a

completely new idea, namely, the quantization or discretization of energy. He
assumed that the oscillating electrons of the cavity walls release energy into the
radiation field not continuously but in lumps, each lump of energy being
proportional to integral multiples of the frequency, i.e.,
E = nhν (1.4)
In classical physics, given a frequency, the energy is determined by the

amplitude of oscillation and any energy is possible. Though Planck’s new idea
was in conflict with the classical view, it was a necessity to move ahead and
solve persisting problems.
Think:
Show that equation (1.3) matches the Rayleigh-Jeans result at small frequencies.
Use e = 1 + x for small x.
x
Integrate equation (1.3) to obtain equation (1.1).
2. The stability of atoms:
According to the Rutherford model of the atom, electrons orbit about the
nucleus of an atom. This model was not flawless, however. Classical
electrodynamics tells us that accelerating charges lose energy by radiating.

Electrons orbiting a nucleus experience a centripetal acceleration and,

therefore, lose energy by emission of radiation and gradually spiral into the
nucleus. That is, a stable electronic orbit is classically forbidden and the atom
should collapse. Classical calculations indicate that the collapse time is about
10–10 sec. But yet this has never been observed to happen.
In 1913, Niels Bohr came up with a semi-quantum explanation for the structure
of atoms, particularly hydrogen. He postulated that:
• Electronic orbits are stable.

Electrons move only in some allowed orbits and in these orbits they do
not radiate energy.
• Electronic angular momentum is quantized.

The magnitude of the angular momentum of the electrons in the stable
orbits has discrete values, that is,
L = r × p = nh (1.5)
• Radiation is quantized.
Electrons release or absorb a quantum of radiation (a photon) only during
transitions between allowed orbits. The energy released or absorbed is
given by
E = E f − Ei = hν (1.6)
Using these postulates and Newtonian Mechanics, Bohr was able to

describe the mystifying experimental facts, especially, the observed
atomic spectra of hydrogen. The Bohr model gives
1  1 1 
= −1.096 × 107 m −1 z 2  2 − 2  (1.7)
λ n m 
for the wavelength of the radiation emitted when the electron jumps from
the nth stable orbit to the mth stable orbit in a one-electron atom. This
result is the same as the empirical formulas of the spectral lines
discovered before the turn of the 20th century.

3. The low specific heats of crystals:
Classically, the molar heat capacity at constant volume, Cv of a crystal is 3R ,

where R the universal gas constant is. This agrees with empirical results
obtained by Dulong and Petit. However, Cv falls sharply for all solids at low
temperatures approaching zero as T approaches absolute zero. This behavior of
solids at low temperatures was inexplicable within the framework of classical
physics.
The problem was solved by Einstein in 1907. He explained that the vibrational
energies of atoms in a crystal are quantized according to E = nhν , which is
Planck's quantization condition for electronic oscillators. According to
Einstein’s theory the molar heat capacity is
2
 hν  e hν kT
Cv = 3 R   2
(1.8)
 kT  ( e hν kT − 1)
Think:
Can you justify that equation (1.8) gives the classical result Cv = 3R for hν kT ?
Verify that equation (1.8) agrees with experimental data for kT hν .
4. Wave-particle duality:
It has been observed that particles and waves behave in a way that cannot be
explained by classical theory. There were ample experimental evidences to
show quite clearly that particles like electrons sometimes act as if they were
waves and waves sometimes act as if they were streams of particles. This
behavior is called wave-particle duality. We will now briefly see some of these
phenomena.
The photoelectric effect
Heinrich Hertz (1886 and 1887) discovered that electrons come out from a
metal surface when the metal is illuminated with ultraviolet light. Later on,
experiments revealed the following facts:
• No electrons are ejected from a metal surface if the frequency of the light
used is lower than a certain minimum threshold no matter how intense the
light is.

• For frequencies above the threshold, the number of electrons emitted is

proportional to the intensity of the light.
• The energy of the emitted electrons is independent of the intensity of
light; instead, the energy of electrons is proportional to the frequency of
the light.
• Above threshold, light emits electrons immediately with no time lag.
Contrary to what is observed in the lab, classical electromagnetic theory

predicts that the energy of ejected electrons should be proportional to the
intensity of the incident light. Also, classically, it requires longer time to eject
electrons from a metal surface as the light intensity goes down.
Einstein solved the problem of the photoelectric effect in 1905. Borrowing

Planck's idea of quantization, he argued that the incident light is coming in
packets of energy, E = hν . These energy packets are now called photons. With
this idea of energy packets and the principle of conservation of energy,
Einstein successfully explained the photoelectric effect:
KEe = hν − φ (1.9)
Where φ , called the work function of the metal, is the energy binding the
electron to the metal?
Note that Einstein’s photoelectric theory predicts that electron energy and light
frequency are related through the same constant as found in Planck's blackbody
curves. This indicates that h has a universal nature.
Think:
Discuss how Einstein’s equation for the photoelectric effect (equation (1.9))
explains all the photoelectric observations listed above: existence of threshold
frequency, linear connection between energy and frequency, independence
of energy on light intensity, and absence of time lag.
Compton scattering
The particle nature of light exhibited in the photoelectric effect was further
justified by the Compton Effect. In 1922, Compton argued that if photons of
light actually behave like particles, then scattering of a photon from an electron
must be observed when they “collide” like two particles. He applied relativistic
mechanics and conservation principles and obtained

h
λ′ − λ = (1 − cos θ ) (1.10)
mc
Where λ ′ and λ are the wavelengths of the scattered and incident photons,
respectively, m is the mass of the electron and θ is the angle of scattering.
The wavelength shift in equation (1.10) was actually observed in the lab!
The Compton Effect marked a radically essential departure from classical

electromagnetism. It is a clear confirmation of the particle-like nature of
waves.
De Broglie waves
Being cognizant of the particle-like properties of waves, de Broglie boldly

proposed in 1923(4) that matter, too, might have wavelike nature. His
proposition was experimentally confirmed in 1927 by Davisson and Germer.
They observed diffraction patterns by bombarding metals with electrons,
implying that the electrons (particles) had behaved as waves!
According to de Broglie, the wavelength of a particle moving with momentum

p is
h
λ= (1.11)
p
Think:
De Broglie's equation offers a justification for Bohr's assumption of quantized
angular momentum. Explain.
From the discussions so far we can conclude that there were two historic tracks
leading to the development of Quantum Mechanics:
1) Energy and angular momentum are discrete

and
2) There is a wave-particle duality.
Next we see how the wave-particle duality leads to the uncertainty principle.

1.3. The Uncertainty Principle

Classically, concepts like position and momentum are independent. In Quantum
Mechanics, however, position and momentum are complementary aspects of a
system due to the wave-particle duality. According to Classical Mechanics, one
can measure both the position and momentum of a particle at the same time to the
desired precision. In the microscopic world, this is impossible. True, one can
measure the position of, say, an electron or its momentum, but not both at the
same time! A simultaneous measurement of position and momentum always
generates intrinsic uncertainties. This is what is known as the uncertainty
principle, developed by W. Heisenberg as a consequence of the wave-particle
duality of microscopic objects. The uncertainty principle states that accurate
knowledge of complementary pairs is impossible. Formally,
∆x∆p ≥ h (1.12)
In the wave-like microscopic world, simultaneous measurements of position and

momentum yield fuzzy position and fuzzy momentum! That is, we get a range of
values for position and momentum instead of single precise values. The
uncertainty principle tells us that complementary quantities, such as position and
momentum, are unknown, except by probabilities.
notice that the uncertainty principle is inconsequential to macroscopic objects

since Planck's constant, h, is so small (~10-34). Suppose the momentum of a
moving ball is measured to an accuracy of 10kg.m/s. The uncertainty in the
position of the ball is 10-35 meters.
The uncertainty principle puts a limit to our knowledge of reality because it clearly
shows that Nature has a built-in indeterminacy, unpredictability.
1.4. Postulates of Quantum Theory

The framework of Quantum Mechanics is built on a few theoretical principles
derived from experimental observations. These principles, called the postulates of
Quantum Mechanics, are discussed below:
Postulate 1:
The state of a quantum mechanical system is completely specified by the wave
function Ψ ( r,t ) , where r represents the space coordinates ( x, y, z ) of the system
and t is the time.
The wave function must be continuous, single valued and square integrable.

The probability of finding the system within a volume dv of space at time t is:
P ( r, t ) = Ψ* ( r, t ) Ψ ( r, t ) dv
Square integrable means

2
∫
all − space
Ψ ( r, t ) dr = finite
2
where Ψ ( r , t ) = Ψ* ( r , t ) Ψ ( r , t )
Postulate 2:
In Quantum Mechanics, every observable (i.e., any measurable property of the
system), is described by a linear, Hermitian operator.
For example, for a particle in one dimensional motion, the momentum operator
corresponding to the classical momentum px is pˆ x = −ih ∂ ∂x .
Postulate 3:
The only possible results of any measurement of an observable are the eigenvalues
of the operator that corresponds to such observable.
That is, if Â is the operator associated with the observable A being measured,
then the only values that will ever be observed are the eigenvalues a which satisfy
ÂΨ = aΨ
Although measurements must always yield an eigenvalue, the state of the system
is not necessarily an eigenstate of Â . Generally, the state of the system is any
arbitrary state that can be expanded in a complete set of eigenvectors as
Ψ = ∑ i ciψ i ,
where the sum can run to infinity in principle. The probability of observing
eigenvalue ai is given by ci*ci .
Postulate 4:
The average value of many measurements of an observable A , corresponding to
operator Â , is given by

*
=∫ *
Ψ ÂΨdv
A
∫ Ψ Ψdv
This average value is called the expectation value.
Postulate 5:
The evolution of the wavefunction Ψ ( r,t ) in time is described by the time
dependent Schrodinger equation:
∂Ψ ( r, t )
ĤΨ ( r, t ) = ih
∂t
ˆ h2 2 ˆ
where H = − ∇ + V ( r ) is the operator associated with the total energy of the
2m
p2
system, E = + V (r ) .
2m

Chapter 2. Mathematical Foundation
2.1. Operators
At the end of Chapter 1, we have briefly seen the postulates of Quantum
Mechanics. Postulate 2 tells us that any observable in quantum mechanics is
described by a linear, Hermitian operator. Now it’s time to see in detail about the
properties of quantum mechanical operators.
What is an operator?
An operator Â is a mathematical entity that acts on a given function f ( x ) and

transforms into another function g ( x ) . Symbolically,
Â f ( x ) = g ( x ) (2.1)
d
Example: The differentiation operator transforms a differentiable function
dx
into another function; the integration operator ∫ transforms an
integrable function into another function.
In 1D Quantum Mechanics, the coordinate operator is x̂ = x and the momentum

operator is p̂ = −ih ∂ ∂ x . Generally, an operator Â that represents an
observable A is obtained by first writing the classical expression of the observable
in terms of Cartesian coordinates, A = A ( x, p ) , and then substituting the
coordinate x and the momentum p by their operators. As an example, consider
the total energy
p2 1 2
E ( x, p ) = K + V = + kx (2.2)
2m 2
of the 1D harmonic oscillator. The energy operator then becomes (see postulate 5):
h2 ∂ 2 1 ˆ 2
Hˆ = − + kx (2.3)
2m ∂ x 2 2
Think:
What is the operator corresponding to angular momentum?

2.2. Basic Properties of Operators

Sum and difference of two operators
For any two operators, Â and B̂ , and a function f :
( Aˆ + Bˆ ) f = Afˆ + Bfˆ ( Aˆ − Bˆ ) f = Afˆ − Bfˆ (2.4)
Product of two operators
For any two operators, Â and B̂ , and a function f :
ˆ ˆ f =A
AB ˆ Bˆ f (2.5)
 
 
Equality
Two operators, Â and B̂ , are equal if, for all functions f :
Â f = B̂ f (2.6)
The identity operator
The identity operator, Î , does not change the operand it acts on:
Î f = f (2.7)
Associativity
The associative law holds for operators, that is,
A ( )( )
ˆ BC
ˆ ˆ = AB
ˆ ˆ Cˆ (2.8)
Commutativity
The commutative law is not generally valid for operators. That is,
ˆ ˆ ≠ BA
AB ˆ ˆ , in general (2.9)

The commutator of two operators
The commutator of two operators Â and B̂ , is defined by:
ˆ B
 A, ˆ ˆ − BA
ˆ  ≡ AB ˆˆ (2.10)
 
Here order matters!
ˆ B
 A, ˆ  = −  B,
ˆ Aˆ  (2.11)
   
ˆ B
If Â and B̂ commute, then  A, ˆ  =0. (2.12)
 
Example
The operators Â = x and B̂ = x 2 commute because for any function ψ ( x )

we have:
ˆ B
 A, ˆ ψ = xx 2ψ − x 2 xψ = 0
 
Think:
Do the operators Â = x and B̂ = ∂ ∂ x commute?
Some rules of commutators
The following are useful rules for evaluating commutators:
ˆ B
 A, ˆ  +  B,
ˆ Aˆ =0 ( see equation 2.11) (2.13)
   
ˆ A
 A, ˆ =0 (2.14)
 
ˆ B+C
 A, ˆ ˆ  =  A,
ˆ B ˆ C
ˆ  +  A, ˆ (2.15)
     
Aˆ + B,C
ˆ ˆ  =  A,C
ˆ ˆ  +  B,
ˆ Cˆ (2.16)
     

ˆ BC
 A, ˆ ˆ  =  A,
ˆ Bˆ Cˆ +B ˆ C
ˆ  A, ˆ (2.17)
     
ˆ ˆ Cˆ  =  A,
 AB, ˆ Cˆ B ˆ  B,
ˆ +A ˆ Cˆ (2.18)
     
ˆ  B,C
 A, ˆ ˆ   + C,
ˆ  A,
ˆ Bˆ   + B ˆ A
ˆ C, ˆ  = 0 (2.19)
           
If the operators Â and B̂ commute with their commutator, then
ˆ B
 A, ˆn =nB ˆ B
ˆ n-1  A, ˆ (2.20)
   
Aˆ n, B ˆ n-1  A,
ˆ  = nA ˆ Bˆ (2.21)
   
ˆ B
Finally, if  A, ˆ then it is possible to show that
ˆ  = iC
 
∆A ∆B ≥ 12 C (2.22)
Where ∆A and ∆B are the uncertainties in the observables A and B, given

by:
2 2 2 2
( ∆A ) = A2 − A and ( ∆B) = B2 − B
Equation (2.22) is the famous Heisenberg uncertainty principle.
Think:
Derive all the rules (2.14) – (2.22).
Obtain the well-known relation ∆x ∆px ≥ h2 from the general rule (2.22).
2.3. Linear Operators

Most operators in Quantum Mechanics are linear operators. An operator Â is said
to be a linear operator if and only if it satisfies the following two conditions:
ˆ ( c f ) = cA
A ˆ f and Â ( f + g ) = Aˆ f + Aˆ g (2.23)

where c is a constant and f and g are functions. Combining the two conditions we
can write
Â ( af + bg ) = aAˆ f + bAˆ g (2.24)
where a and b are constants.
d 2
Example: Consider the two operators and ( ) . Which one is linear?
dx
It is easy to see that the first operator is a linear operator because it satisfies the
conditions for linearity:
d d d
 a f ( x ) + b g ( x )  = a f ( x) + b g ( x)
dx dx dx
The second operator, however, is not a linear operator because

2 2 2
 a f ( x ) + b g ( x )  ≠ a  f ( x )  + b  g ( x ) 
Other operators relevant to Quantum Mechanics are the antilinear operators

defined by
Â (α f + β g ) = α *Aˆ f + β *Aˆ g (2.25)
Where α and β are complex constants and the asterisk represents their complex
conjugate.
2.4. Hermitian Operators

An operator Â is Hermitian if and only if
*
*
( )
∫ψ 1 Aˆ ψ 2dx = ∫ Aˆ ψ 1 ψ 2dx (2.26)
where ψ 1 and ψ 2 are any two state functions.
Example
Consider the momentum operator p̂ = −ih ∂ ∂ x . Now, by equation (2.26):

*
* ∂   
∞ ∞
∂
∫−∞ψ 1  −ih ∂ x ψ 2 dx = −∞∫  −ih ∂ x ψ 1  ψ 2dx
∂ψ 1*
∞ ∞
∂ψ 2
*
−i h ∫ ψ 1 dx = ih ∫ ψ 2 dx
−∞
∂x −∞
∂x
∞
 ∂ψ 2 ∂ψ 1*  ∞
−ih ∫ ψ 1* + ψ 2  dx = −ih ψ 1*ψ 2  = 0
−∞ 
∂x ∂x 
−∞
Therefore, the momentum operator is Hermitian.
Note: By Postulate 2 quantum mechanical operators are linear and Hermitian.
2.5. Eigenstates and Eigenvalues

An eigenfunction of an operator Â is a function f such that the application of Â on
f gives a constant multiple of f , that is,
Â f = k f (2.27)
The constant k is called the eigenvalue of Â .
In Quantum Mechanics, the eigenfunction of the operator Â is the eigenstate of

the observable A and the eigenvalue is the expectation value of the observable A
(see postulate 4 in section 4, Chapter 1). Thus if
Âψ ( r ) = aψ ( r ) , (2.28)
then, by Postulate 4,
* *
ψ (r ) Âψ ( r ) dr ∫ψ (r ) aψ (r ) dr
A =∫ * = *
∫ψ ( r )ψ (r ) dr ∫ψ ( r )ψ ( r ) dr
a ∫ψ * ( r )ψ ( r ) dr
= *
=a
∫ψ ( r ψ
) ( r ) d r
Hermitian operators have the following two important properties:
Property 1: The eigenvalues of a Hermitian operator are real.

Proof: Using equation (2.26) with eigenfunctions ψ n , we obtain
*
( )
∫ψ nAˆ ψ n dx = ∫ Aˆ ψ n ψ ndx
*
Substituting equation (2.28):
* *
∫ψ nan ψ n dx = ∫ ( anψ n ) ψ ndx
an ∫ψ n*ψ n dx = an* ∫ψ n*ψ n dx
(an − an* ) ∫ψ n*ψ n dx = 0
Since the eigenfunctions ψ n are square integrable (see Postulate 1), we conclude:
an = an* (2.29)
Property 2: The eigenfunctions of a Hermitian operator are orthogonal. That is, if

ψ n and ψ m are two different eigenfunctions of a Hermitian operator
Â such that Âψ n = anψ n and Âψ m = amψ n , then
*
∫ψ nψ mdx = 0
Proof: using equation (2.26) together with equation (2.28):
* *
∫ψ nam ψ m dx = ∫ ( an ψ n ) ψ m dx
( am − an ) ∫ψ n*ψ m dx = 0 Equation (2.29) has been used.
Since am ≠ an for two different eigenfunctions, we must have
*
∫ψ n ψ m dx = 0 (2.30)
The eigenfunctions are said to be normalized to 1 if they satisfy the condition
*
∫ψ n ψ n dx = 1 , (2.31)

2.6. Hilbert Space

We have seen in the previous section that the eigenfunctions ψ i of a linear and
Hermitian operator Â satisfy equations (2.30) and (2.31) which can be combined
into one:
*
∫ψ n ψ m dx = δ nm (2.32)
where δ nm is the Kronecker delta. Such functions are said to be orthonormal

functions. These orthonormal eigenfunctions form a complete basis set just like
the unit base vectors in 3D Euclidean space. That is, any wavefunction Ψ can be
expanded as a linear combination of the orthonormal eigenfunctions:
Ψ ( r ) = ∑i ciψ i ( r ) (2.33)
where the expansion coefficients ci are generally complex numbers.
The representation of the wavefunction Ψ in

ψ 2 ( x) terms of the basis functions (eigenfunctions) is
similar to the representation of an ordinary
vector in Euclidean space as shown in the
Ψ ( x) diagram for 2D case. The eigenfunctions form
c2
a function space with infinite dimensions
called Hilbert-Space.
The constants ci are the “components” of the

wavefunction Ψ and can be determined as
c1 ψ1 ( x) follows:
Ψ ( x ) = c1 ψ 1 ( x ) + c2 ψ 2 ( x ) Multiply both sides of equation (2.33) by the

complex conjugate of a basis eigenfunction
ψ *j and integrate:
∞ ∞
∫ ∑ i ciψ j (r )ψ i ( r )dr
*
∫ ψ j ( r ) Ψ ( r ) dr =
*
−∞ −∞
∞
= ∑ i ci ∫ ψ *j ( r )ψ i ( r ) dr = ∑ i ciδ ji
−∞
∞
⇒ c j = ∫ ψ *j ( r ) Ψ ( r ) dr (2.34)
−∞

Think:
What is the expectation value of the observable A when a quantum mechanical
system is in a state given by equation (2.33)?
Hint:
*
∫ Ψ Aˆ Ψdr  ∑ ciψ i ( r )  A ˆ
 ∑ j c jψ j ( r )dr
*
A = =
 i ∫
*
∫ Ψ *Ψ d r  ∑ i ciψ i ( r )   ∑ j c jψ j ( r )  dr
 ∫   
You will see that the average value of A is a weighted average of eigenvalues.
Continuous Representation
Some operators have a continuous spectrum of eigenvalues. In such cases the

eigenfunctions form a continuous basis set. The sum in equation (2.33) is then
replaced by an integration of the form
∞
Ψ ( x ) = ∫ φ ( x′)ψ ( x′, x ) dx′ ,
−∞
with
∞
φ ( x′ ) = ∫ ψ * ( x′, x ) Ψ ( x ) dx
−∞
2.7. Ket and Bra Vectors

The concept of ket and bra vectors was introduced by Paul Dirac to represent
states as vectors. The advantage is that the coordinate dependence need not be
specified when working in the ket and bra vector space. According to such
representation, called the Dirac notation or bra-ket notation, the state function
ψ ( x ) is written as a ket ψ and its complex conjugate ψ * ( x ) as a bra ψ .
Therefore, for any function Ψ that can be expressed in terms of basis functions,
Ψ ( x ) = ∑ i ciψ i ( x ) , (2.35)
the ket-vector representation becomes
Ψ = ∑i ci ψ i (2.36)
where ψ i are basis ket vectors.

Taking the complex conjugate of equation (2.36) and remembering that a bra is the
complex conjugate of a corresponding ket, we obtain
* * *
Ψ =  ∑ i ci ψ i  = ∑ i ci* ψ i
Ψ = ∑ i ci* ψ i (2.37)
The bra vector ψ defined by equation (2.37) is said to be dual to the ket vector
ψ . The inner (scalar) product of a bra and a ket is defined by
∞
*
ΨΦ = ∫ Ψ Φ dx ,
−∞
(2.38)
Using equations (2.32), (2.36) and (2.37) into equation (2.38)
Ψ Φ = ∑ n ∑ m cn* d m ψ n φm = ∑ n ∑ m cn* d mδ nm
Ψ Φ = ∑ n cn* d n (2.39)
The bra-ket notation for the expectation value of an observable is

∞
ˆ ψ dx = ψ A
A = ∫ ψ *A ˆψ (2.38)
−∞
It is easy to show that

*
ΨΦ = ΦΨ (2.39)
Ψ Ψ = real number (2.40)
Ψ Ψ ≥0 (2.41)
The bra-ket notation can be further simplified for enumerated (discrete)

eigenfunctions:
ψn ≡ n and ψn ≡ n (2.42)

Think:
1. Verify equations (2.39) – (2.41)
2. Write the definition of Hermiticity of an operator, equation (2.26), in Dirac’s
notation
3. Show that ψ j ψ i is an operator that projects any vector Ψ into the direction
of ψ j .
2.8. Wave Functions in Position and Momentum Space

In many cases momentum appears to be a better variable than position and,
therefore, it might be convenient to have all wave functions represented as a
function of momentum rather than position. The wave function in momentum
space is obtained from its representation in position space by the Fourier transform
∞
1
φ ( p) = ∫
2π h −∞
e−ipx hψ ( x ) dx (2.43)
If the wave function in momentum space is known, its counterpart in coordinate

space is calculated by the inverse Fourier transform:
∞
1
ψ ( x) = ∫
2π h −∞
eipx hφ ( p ) dp (2.44)
The correspondence between φ ( p ) and ψ ( x ) is unique, that is, for a given wave
function ψ ( x ) in position space there is only one representation in momentum
space φ ( p ) . An important property of the Fourier transform is that if ψ ( x ) is
normalized, then φ ( p ) is normalized as well. The normalization condition holds
both in position and momentum spaces:
∞ ∞
* *
∫ ψ ( x )ψ ( x ) dx = ∫ φ ( p )φ ( p ) dp = 1
−∞ −∞
(2.45)

Chapter 3. The Schrödinger and Heisenberg Equations
3.1. Schrödinger’s Equation

In Newtonian Mechanics an object is pictured as particle wit definite mass,
momentum, acceleration, etc.
z
r r
r Liner Velocity: v = d r dt
v
m r r
Liner Momentum: p = mv
r
r r
nd dp r
Newton’s 2 Law: = −∇V
dt
x Etcetera
y
Figure 3.1: Newtonian Picture of the dynamics of a particle
According to quantum mechanics, the Newtonian picture of nature is wrong if

examined very carefully. The problem is that a definite numerical position or a
definite numerical linear momentum for the particle simply does not exist.
Quantum mechanics describes the particle in terms of the wave function discussed
in Chapter 1.
We will see in this Chapter that knowledge of the wave function Ψ ( x, y, z; t )

provides all information available about the particle. Figure 3.2 shows a mental
2
picture of the square magnitude of the wave function, Ψ . The physical meaning
attached to the square magnitude is known as Born's statistical interpretation. The
darkest part of the gray region is the location where the particle is most likely to be
z found. More precisely, the probability of
finding the particle within a small volume of
r
size d 3r = dxdydz around a given location
m r
r = ( x, y, z ) at the time t is, if such a
measurement is attempted,
x r 2 r
y Ψ ( r ; t ) d 3r (3.1)
Figure 3.2: Visualization of a
wave function representing a Such a position measurement, if actually
particle done, affects the wave function. After the
measurement, the wave function collapses,

that is, it will be restricted to the volume where the particle is “caught”. If a second
measurement is made immediately after the first, the same result for the position is
obtained indicating that the wave function is still in its collapsed state. If, however,
the measurement is repeated after sufficiently long time, a different value is
obtained for the position of the particle. This shows that the wave function will
spread out again in time if allowed to do so.
If equation (3.1) is integrated over all possible locations of the particle, we must
obtain 1. That is, the total probability of finding the particle must be 100%. In
quantum mechanics, this is expressed by
r 2 r
ΨΨ = ∫ Ψ ( r ; t ) d 3r = 1 (3.2)
r
All r
Wave functions satisfying equation (3.2) are said to be normalized.
The wave function for a quantum “particle” is obtained by solving the equation
h2 2 ∂Ψ
− ∇ Ψ + V Ψ = ih (3.3)
2m ∂t
∂2 ∂2 ∂2
where ∇ ≡ 2 + 2 + 2 is a differential operator called the Laplacian and V is
2
∂x ∂y ∂z
the potential energy function. Equation (3.3) is the time-dependent Schrödinger
equation.
Separable Solutions
r
In this Chapter, we consider 1D quantum problems so Ψ ( r ; t ) ≡ Ψ ( x; t ) . The time-
dependent Schrödinger equation, equation (3.3), now becomes
h2 ∂ 2 Ψ ( x; t ) ∂Ψ ( x; t )
− + V Ψ = ih (3.4)
2m ∂x 2
∂t
For cases where the potential energy is independent of time, the solution of the
time-dependent Schrödinger equation is separable, that is, we can find a function
of position and a function of time such that the product
Ψ ( x; t ) = ψ ( x ) f ( t ) (3.5)
is a perfect solution of equation (3.4).

Substituting equation (3.5) into equation (3.4) and then dividing by equation (3.5),
we have:
2
h2 1 d ψ ( x ) 1 df ( t )
−
2m ψ ( x ) dx 2
+ V ( x ) = i h
f ( t ) dt
(3.6)
The left-hand side of equation (3.6) is a function that depends only on x and the
right-hand side is a function only of t . The equality holds if and only if both
functions are equal to the same constant, say, E . Then,
2
h2 1 d ψ ( x )
− +V ( x) = E (3.7)
2m ψ ( x ) dx 2
1 df ( t )
and ih =E (3.8)
f ( t ) dt
Let’s first consider equation (3.8) which can be rearranged as
d f (t ) i
= − E dt
f (t ) h
Integration of this equation gives
i
ln f ( t ) = − Et + const
h
Or f ( t ) = C e−(iE h )t (3.9)
where C is a constant of integration.
Using equation (3.9) into equation (3.5) the separable solution to the time-
dependent Schrödinger equation becomes:
Ψ ( x; t ) = ψ ( x ) e−(iE h)t (3.10)
The constant C is absorbed into the position-dependent function ψ ( x ) .
Equation (3.7) can be rearranged in the form

 h2 d 2 
− 2
+ V ( x ) ψ ( x ) = Eψ ( x ) (3.11)
 2m dx 
The expression in square brackets is the Hamiltonian operator Ĥ corresponding to

the total energy (see Postulate 5, Chapter 1 and the first section of Chapter 2).
Ĥψ ( x ) = Eψ ( x ) (3.12)
It should be clear now that ψ ( x ) is the eigenfunction and E , the eigenvalue of

the Hamiltonian operator, which is Hermitian. Therefore, the constant E is the
expectation value of the total energy as can be easily verified:
ˆ Ψdx = ψ *H
H = ∫ Ψ *H ∫ ˆ ψ dx = E ∫ψ ψ dx = E
*
We can see that the time-dependent Schrödinger equation (3.4) is reduced to an

eigenvalue problem (equation 3.12) which is called the time-independent
Schrödinger equation. Solving the time-independent Schrödinger equation requires
a knowledge of the potential energy as a function of position.
Think:
1) Show that the probability density is independent of time, that is,
2 2
Ψ ( x; t ) = ψ ( x ) . In this case, the wave functions are said to be stationary
states.
2) Show that every measurement of the total energy always returns the same value
E for a stationary state: σ H = 0
3.2. One-Dimensional Infinite Potential Well

In this section we will see the general procedure for analyzing quantum systems
that will be useful in understanding more advanced problems of quantum
mechanics discussed in later chapters. The system considered here is a single
particle, say, an electron, confined inside an infinitely deep well.
The physical system

The particle inside the infinite well is a rough idealization of a physical system
such as a valence electron trapped in a copper surface. The copper surface
corresponds to the walls of the infinite well.

Here the conduction electron is not totally free as there are other electrons and
nuclei within the copper surface and these should be ignored in the analysis. But
we can refine our imagination and say the particle is a lone helium atom stuck
inside a carbon nanotube.
A classical picture of the particle inside the infinite well is shown in Figure 3.3
while the actual quantum system we need to study is pictured as in Figure 3.4.
V ( x) V ( x)
ψ ( x)
x (t )
x x
a a
Figure 3.3. The classical picture Figure 3.4. The quantum picture
The classical particle, imagined as a little sphere, is bouncing back and forth
between the walls of the infinite well. If friction is not assumed, the particle will
keep bouncing forever.
A quantum “particle” like an electron does not have specific shape or size, but has
a wave property. So in quantum mechanics, we have a wave “reflecting” between
the walls of the infinite well instead of a particle bouncing around.
Conditions to be met
The particle inside the infinite well is assumed to be free, that is, it experiences no
forces except when it hits the sides of the well. Therefore, the potential energy
must be constant inside the well. Remember: force is the negative derivative of
potential energy. Further, since the value of the constant does not have physical
significance, we can simplify calculations by assuming that the potential energy is
zero inside the well. So, for the infinite potential well we are considering,
0, if 0 ≤ x ≤ a
V ( x ) =  (3.13)
∞, otherwise
where a is the length of the one-dimensional infinite well.

Since the particle is confined within the potential well, the probability of finding it
outside the well is zero, that is, ψ ( x ) = 0 for x < 0 and x > a . Moreover, the wave
function is required to be single-valued and continuous (see postulate 1, Chapter 1)
and, therefore, the following boundary conditions must be satisfied:
ψ ( 0) = 0 at x = 0 (3.14)
ψ (a) = 0 at x = a (3.15)
The Hamiltonian
Next we find the Hamiltonian, that is, the total energy operator. Using equation
(3.13), the 1D Hamiltonian becomes:
h2 d 2 h2 d 2
Ĥ = − + V ( x ) = − (3.16)
2m dx 2 2m dx 2
The Hamiltonian eigenvalue problem

Having determined the Hamiltonian, we then proceed to the Hamiltonian
eigenvalue problem, that is, the time-independent Schrödinger equation.
Substituting equation (3.16) into equation (3.12):
h2 d 2ψ
− = Eψ (3.17)
2m dx 2
Solutions of the eigenvalue problem

Now we are ready to solve the Hamiltonian eigenvalue problem, equation (3.17),
which is a second order ordinary differential equation. Rewriting (3.17):
d 2ψ 2mE
2
=− 2 ψ (3.18)
dx h
The best way here is to look up the solution in the form of an exponential function.
Then suppose the solution is
ψ ( x ) = Ceα x (3.19)
where C and α are constants.

Substituting equation (3.19) into equation (3.18), we obtain:
α 2ψ = − 2mE
2
ψ
h
−2mE
Or α =±
h2
We see that equation (3.19) actually contains two solutions:
ψ + ( x ) → e+ −2mE x h and ψ − ( x ) → e − −2mE x h
A more general solution is a linear combination of the two:
−2 mE x h −2 mE x h
ψ ( x ) = C1e+ + C2e− (3.20)
Let’s now apply the first boundary condition, equation (3.14):
ψ ( 0) = C1 + C2 = 0 ⇒ C2 = −C1
Equation (3.20) now becomes
(
ψ ( x ) = C1 e+ −2 mE x h
− e− −2 mE x h
) (3.21)
Applying the second boundary condition, equation (3.15):
(
ψ ( x ) = C1 e+ −2 mE a h
− e− −2 mE a h
)=0 (3.22)
Let’s assume that the energy E is negative so the square root −2mE and hence
the exponentials become real numbers.
Since the exponentials in equation (3.22) are not equal, the expression in
parentheses cannot be zero. Then the constant C1 will have to be zero producing
the trivial solution ψ ( x ) = 0 . A zero wave function is physically unacceptable
because inside the potential well, where there is a nonzero probability of finding
the particle, the wave function should not vanish. What is the problem then?
The problem must be the assumption that the energy is negative. Actually, the
energy cannot be negative! Remember that the potential energy is zero and, for the

particle in the infinite well, the total energy is all kinetic. According to Classical
Mechanics the kinetic energy, being proportional to the square of the velocity, can
never be negative. Quantum Mechanics agrees with Classical Mechanics on the
fact that the kinetic energy cannot be negative, but not on the reason. Quantum
Mechanics tells us that the particle’s kinetic energy is nonnegative because the
wave function must satisfy the boundary conditions.
What if the energy E is zero? In this case equation (3.18) becomes
d 2ψ
=0
dx 2
with the solution
ψ ( x ) = C1 + C2 x
The first boundary condition requires that ψ ( 0 ) = C1 + C2 × 0 = 0 , producing C1 = 0 .

With this result, the wave function corresponding to zero energy will be
ψ ( x ) = C2 x
The second boundary condition gives: ψ ( a ) = C2 a = 0 , that is, C2 = 0 .
Once again we have the trivial solution ψ ( x ) = 0 . So, the assumption that the
energy E can be zero must be wrong too.
There is no objection for zero energy in Classical Mechanics because that would
imply a particle sitting motionless in the infinite potential well. Quantum
Mechanics predicts that the particle cannot have zero kinetic energy because of
Heisenberg's uncertainty principle. Since the particle’s position cannot be
determined with certainty, its linear momentum must be uncertain as well.
Measurements will show a range of values for the momentum of the particle. It is
in motion and therefore has kinetic energy.
Note how the boundary conditions restrict the value of the energy the particle can
have. We are now left with one possibility – a positive energy. In that case,
equation (2.21) becomes
(
ψ ( x ) = C1 e+ik x − e−ik x ) (3.23)

where k is positive defined by
2mE
k= (3.24)
h
Using Euler’s formula eikx = cos kx + i sin kx we can write equation (3.23) as
ψ ( x ) = C sin kx (3.25)
Applying the second boundary condition to (3.25) we obtain
ψ ( a ) = C sin ka = 0
Since C cannot be zero (because, if it is, it would lead to the trivial solution),
equation (3.25) is satisfied only if the sine factor is zero. That is,
sin ka = 0
Or ka = 0, ± π , ± 2π , ± 3π ,L
The zero and negative values are not acceptable because k > 0 (E is positive in
equation (3.24)). We conclude: there is a nonzero solution for the time-
independent Schrödinger equation corresponding to each of the following values
of the positive constant k :
nπ
k =
n
, for n = 1,2,3,L (3.26)
a
The solutions are given by equation (3.25) for the different values of k :
ψ 1 = C1 sin  πa x  ,ψ 2 = C2 sin  2aπ x  ,ψ 3 = C3 sin  3aπ x  ,L

     
Or in a generic form,
ψ n = Cn sin  naπ x  , for n = 1,2,3,L (3.27)

 
Check the Solutions

The solutions given by equation (3.27) all satisfy the boundary conditions, that is,

ψ n ( 0) = Cn sin ( 0) = 0
ψ n ( a ) = Cn sin ( nπ ) = 0
Also, substitution of each solution into the ordinary differential equation (equation
3.18)
d 2ψ 2mE
2
=− 2 ψ
dx h
shows that they all satisfy it, provided that their energy values are:
n 2h 2π 2
En = for n = 1, 2,3,L (3.28)
2ma 2
Note that the energy values [equation (3.28)] could be obtained by combining
equations (2.24) and (2.26).
What about the normalization condition? Are the solutions normalized? Let’s
check.
a
2 2  nπ 
1= ψn ψn = ∫x=0 sin  a
Cn x  dx

After integration we have
a 2
Cn = 1
2
Choosing Cn to be a positive real number, the normalization condition requires

that
2
Cn =
a
We finally have the normalized solutions that satisfy the time-independent

Schrödinger equation:
ψ n = a2 sin  naπ x  , for n = 1,2,3,L (3.29)

 

Summarizing: there are infinitely many energy eigenfunctions and discrete energy
eigenvalues.
Think:
Show that the normalized wave functions are also orthogonal.
Discussion of the Energy Values

We have seen that the only energy values the particle in the infinite potential well
can take are the eigenvalues of the Hamiltonian operator. It cannot have negative
energy, nor zero energy.
The first few of the allowed energy values are shown in Figure 3.5. In this “energy
spectrum”, the energy levels are plotted vertically and the corresponding quantum
number n is indicated to the right of each energy level.
Note the difference between the

quantum energy values and the
E5 = 25 π 2 h 2 2ma 2 n=5
classical one. Classically, the total
energy of the particle can have
any nonnegative value. Quantum
mechanics, on the other hand, tells
us that the possible energy levels
E4 = 16π 2 h 2 2ma 2 n=4 are discrete, that is, the total
energy must be one of the levels
shown in the energy spectrum.
This difference, however, is
E3 = 9π 2 h 2 2ma 2 n=3 insignificant for a macroscopic
particle due to the smallness of
Planck's constant h . At quantum
E2 = 4 π 2 h 2 2ma 2 n=2 scales, the discreteness of the
energy values can make a major
E1 = π 2 h 2 2ma 2 n =1 difference. See the “Think”
problem below.
Figure 3.5: A portion of the energy spectrum The lowest possible energy level
for the particle in the 1D infinite potential wellE1 is called the ground state. The
particle will be found in the
ground state when the temperature is lowered to absolute zero. We can see that
quantum mechanics does not allow zero energy even at zero Kelvin. In classical
physics, the particle has no kinetic energy (so zero total energy) at absolute zero.

Think:
1) Calculate the ground state energy for m = 9.10938 ×10−31 kg (the mass of an
electron) and a = 2 ×10−10 m . How big is it? Note that h = 1.05457 ×10−34 J .s .
Express the result in units of eV, where 1eV = 1.60218 ×10−19 J .
2) Now use macroscopic values, m = 1 kg and a = 1 m , into the expression for the
ground state energy and see how big it is.
3) What is the quantum number that produces a macroscopic amount of energy,
say, 1 J, for macroscopic values m = 1 kg and a = 1 m ? With that many energy
levels involved, would you see the difference between successive ones?
Discussion of the eigenfunctions

The eigenfunctions given by equation (2.29) are the sate functions of the quantum
system we analyzed, that is, the particle in the 1D infinite potential well. The first
three state functions and their magnitude squares are plotted in Figure 3.6.
2
ψ1 ψ1
2
ψ2
ψ2
2
ψ3
ψ3
The ground state eigenfunction and its square magnitude are plotted at the top of
Figure 3.6. Since the square magnitude gives the probability distribution of the
particle’s position, it is seen that in the ground state, the particle is much more
likely to be found close to the middle of the potential well than close to the ends.

The middle graphs in Figure 3.6 show the first excited state wave function ψ 2 and
the corresponding probability density curve. Here there are two regions, to the left
and right of the center, where the particle is more likely to be found. But unlike to
the ground state, the probability of finding the particle close to the center is zero.
The second excited state ψ 3 and its square amplitude are shown at the bottom of
Figure 3.6. We can see that the probability curve picks up maximum values in
three regions indicating that the likelihood of locating the particle in these regions
is high.
Note the following:
• In these eigenfunctions, the particle is not localized to within any particular

small region of the infinite potential well.
• In general, there are regions where the particle may be found separated by
regions where there is little chance to find the particle in.
• The higher the energy, the more such regions there are.
• The number of nodes (zero crossings) in the nth state is n − 1 , excluding the
end points.
• The energy eigenstates are stationary states: the probabilities shown in
figure 3.6 do not change with time.
In order to endow a particle with localized wave function bouncing around,

stationary states of different energy must be combined. That is, the most general
solution to the time dependent Schrödinger equation is a linear combination of the
stationary states. Using equation (3.28) and (3.29) in equation (3.10), we obtain
the stationary states for the particle in the 1D infinite potential well:
 
2  nπ  −i n2π 2h 2ma2  t
Ψ n ( x; t ) = sin  xe (3.30)
a  a 
A linear sum of these solutions gives the most general solution:
 
∞ ∞
2  nπ  −i n2π 2h 2ma2  t
Ψ ( x; t ) = ∑ Cn Ψ n ( x; t ) = ∑ Cn sin  xe (3.31)
n =1 n =1 a  a 
At some initial moment (t=0), the wave function takes the form
∞
∑ Cn sin  naπ x 
2  
Ψ ( x;0) = (3.32)
a  
n=1

Think:
1) For the nth eigenfunction ψ n , how many distinct regions are there where the
particle may be found?
2) There are no forces inside the pipe, so the particle has to keep moving until it
hits the side of the infinite well, and then reflects backward until it hits the
other side and so on. So, it has to cross the center of the pipe regularly. But in
the energy eigenstate ψ 2 , the particle has zero chance of ever being found at
the center of the pipe. What goes wrong?
3) Show that the expansion coefficients Cn are given by
a
2  nπ 
Cn = ∫ sin  x  Ψ ( x;0 ) dx ,
a 0  a 
where Ψ ( x;0 ) is the initial wave function given by equation (3.32).
3.3. Three-Dimensional Solution

In three dimensions, equation (3.17) becomes
h2  ∂ 2 ∂ 2 ∂ 2 
ψ x, y, z ) = Eψ ( x, y, z )
2m  ∂x 2 ∂y 2 ∂z 2  (
−  + + (3.33)
We have seen in the previous section that the wave function for the infinite
potential well is separable. That means the three-dimensional solution is a
combination of the one-dimensional solutions for all three coordinates:
ψ ( x, y, z ) = ψ 1 ( x )ψ 2 ( y )ψ 3 ( z ) (3.34)
Substituting (3.34) into (3.33) and dividing by the full wave function we obtain:
1 d 2ψ 1 ( x ) 2
1 d ψ 2 ( y)
2
1 d ψ3 ( z)
+ + = −k 2
ψ 1 ( x ) dx 2
ψ 2 ( y ) dy 2
ψ 3 ( z ) dz 2
(3.35)
where k is defined by equation (3.24).
Equation (3.35) is valid for functions of independent variables ( x, y, z ) only if each

term on the left hand side is a constant, that is,

1 d 2ψ 1 ( x )
= − k x2
ψ 1 ( x ) dx 2
1 d 2ψ 2 ( y )
= − k y2 with k 2 = k x2 + k y2 + k z2 (3.36)
ψ 2 ( y ) dy 2
1 d 2ψ 3 ( z )
= −k z2
ψ 3 ( z ) dz 2
Each of the equations above is similar to equation (3.18) and, therefore, produces
the same solutions. After all, there is no fundamental difference between the three
coordinate directions:
 nπ  nx2π 2
(ψ 1 )n = 2 sin  1 x  , k x2 = for n1 = 1,2,3,L
1 ax  ax  ax2
2 2
nπ  nπ
(ψ 2 )n = 2 sin  2 y  , k y2 = y 2 for n2 = 1,2,3,L (3.37)
2 ay a
 y  ay
nπ  nz2π 2
(ψ 3 )n = 2 sin  3 z  , k z2 = for n3 = 1,2,3,L
3 az  az  az2
Now, the three-dimensional eigenfunctions are simply products of the one

dimensional ones:
 
ψ n1n2n3 ( x, y, z ) = a a8 a sin  na1π x  sin  na2π y  sin  na3π z 
   
(3.38)
x y z  x   y   z 
There is one such three-dimensional eigenfunction for each set of three quantum
numbers {n1 , n2 , n3} .
The energy eigenvalues of the three-dimensional problem are:
h2 2 h 2 2 2 2 h 2π 2  n12 n22 n32 

En1n2n3 =
2m
k = (
k + k y + kz =
2m x
+ + )
2m  ax2 a 2y az2 
(3.39)
 
For example, when all three quantum numbers are equal to 1, we get the lowest
energy state – the ground state – with
     
ψ 111 ( x, y, z ) = a a8 a sin  aπ x  sin  aπ y  sin  aπ z 
x y z  x   y   z 

and
h2π 2  1 1 1 
E111 = + +
2m  ax2 a y2 az2 
 
Think:
1) If the dimensions a y and az are one tenth of the dimension ax , by what
percentage is the one-dimensional ground state energy E1 in error as an
approximation to the three-dimensional one E111 ?
2) If ax = a y = az = a , how many states are there with the same total energy of
 h 2π 2 
E = 14  2 
 2ma 
Check yourself!
The initial wave function for the particle in the infinite square well is found to be a
linear combination of the first two stationary states:
Ψ ( x;0 ) = A ψ 1 ( x ) + ψ 2 ( x ) 
a) Normalize Ψ ( x;0 ) .
2
b) Find Ψ ( x; t ) and Ψ ( x; t )
c) Compute x . What is the frequency and amplitude of the oscillation?
d) Compute p .
e) Find the expectation value of H . How does it compare with E1 and E2 ?
f) A classical particle in this well would bounce back and forth between the
walls. If its energy is equal to the expectation value you found in (e), what
is the frequency of the classical motion? How does it compare with the
quantum frequency you found in (c)?

3.4. The Finite Potential Well

In this section we consider a particle of mass m and total energy E interacting with
a finite potential well. Equation (3.13) now becomes:
0, if 0 ≤ x ≤ a
V ( x ) =  (3.40)
V0 , otherwise
where V0 > 0 .
Now, depending on the total energy E of the particle, we will obtain two types of
solutions to the time-independent Schrödinger equation. If E < V ( x ) , the solutions
are discrete and normalizable, like the eigenstates of the infinite potential well. On
the other hand, if E > V ( x ) , the solutions are non-normalizable, and labeled by a
continuous index. A particle represented by discrete, normalizable solutions is
said to be in a bound state. In the latter case, where the total energy exceeds the
potential energy, the particle is unbounded and we have a scattering state.
We will first look the bound states, E < V ( x ) .
Classically, a particle in the potential well simply moves between x = 0 and x = a ,

which are called turning points. The situation is quite different in Quantum
Mechanics. Schrodinger’s equation in the three regions is
d 2ψ 2m (V0 − E )
= ψ; x < 0 and x > a
dx 2 h2
V ( x)
d 2ψ 2mE
2
=− 2 ψ; 0< x<a
dx h
Adopting the notation E

V0
2 2mE
k = 2 ; k >0 (3.41)
h a x
and Figure 3.7. The finite potential well

with V0 > E
2m (V0 − E )
q2 = ; q>0 (3.42)
h2

the Schrödinger equation becomes
d 2ψ
2
= q 2ψ ; x<0 (3.43)
dx
d 2ψ
2
= −k 2ψ ; 0< x<a (3.44)
dx
d 2ψ
2
= q 2ψ ; x>a (3.45)
dx
The general solution in the region x < 0 is
ψ ( x ) = A1eqx + A2e−qx
The second term is not admissible as it shoots to infinity for x → −∞ . Therefore,
ψ left ( x ) = Aeqx for x < 0 (3.46)
Similarly, for the region x > a :
ψ ( x ) = D1eqx + D2e−qx
We must choose D1 = 0 for the wave function becomes unphysical when x → ∞ .

Then,
ψ right ( x ) = De− qx for x > a (3.47)
Now, for the middle region, the general solution is
ψ middle ( x ) = B cos kx + C sin kx (3.48)
Since the wave function and its first derivative must be continuous, the solutions
(3.46), (3.47) and (3.48) must satisfy the boundary conditions at x = 0 and x = a .
The boundary conditions fix the relations between the constants A, B, C and D.
At x = 0 :
ψ left ( 0 ) =ψ middle ( 0) or A = B (3.49)

and
dψ left dψ middle
= or qA = kC (3.50)
dx dx
x =0 x =0
At x = a :
ψ middle ( a ) =ψ right ( a ) or B cos ka + C sin ka = De−qa (3.51)
and
dψ middle dψ right
= or − kB sin ka + kC cos ka = −qDe−qa (3.52)
dx dx
x =a x =a
From equations (3.49) – (3.52) we obtain:
q
B = A; C= A
k
 
(3.53)
q+k
D = − sin ka + cos ka  Aeqa
 q−k 
Using equations (3.53) back into equation (3.51), or (3.52) for that matter,
produces the transcendental equation:
2qk
= − tan ka (3.54)
q2 − k 2
Let θ = ka . Then
2m (V0 − E ) 2mV0 2mE 2mV

q2 = 2
= 2 − 2 = k02 − k 2 where k02 = 2 0
h h h h
In terms of θ,
2 θ02 −θ 2 θ02 − θ 2
q = and q= ; where θ0 = k0a
a2 a
Substituting this result into (3.54) and using θ = ka :

θ θ02 − θ 2
= − tan θ (3.55)
0.5θ02 − θ 2
A graphical solution for the transcendental equation (3.55) is shown below for
θ0 = 4π .
7π/2 4π
π/2 π 3π/2 2π 5π/2 3π θ
Figure 3.8: Graphical solution to equation (3.55) for θ0 = 4π
Now, the solutions to equation (3.55) correspond to the intersection of the curve
θ θ 02 − θ 2 ( 0.5θ02 − θ 2 ) with the curve [ − tan θ ] . Figure 3.8 shows that these curves
intersect three times to the left of θ = 3π and only once to the right of θ = 3π for
the particular value θ0 = 4π .
Since, for the bound state under consideration, 0 < E < V0 , the range of values of
θ must be between zero and θ0 = k0 a = a 2mV0 h . We can now consider two
extreme cases:
I. θ0 = a 2mV0 h 1 (Deep, wide well)
As θ0 increases (i.e., as the well becomes deeper), there are more and more bound
states. In this limit (i.e., the limit in which the well becomes very deep), the
solutions to equation (3.55) correspond to [ − tan θ ] → 0 . This gives
θ ≅ nπ , where n is a positive integer. (3.56)

For θ0 = 4π , the maximum value of θ , we have n = 4 bound states as shown in the

figure. Now, using θ = ka and equation (3.41) into equation (3.56), we obtain:
En ≅
n2π 2h2 , n≤
θ0
(3.57)
2ma2 π
The energies of the finite well approach the energies of the infinite potential well
as V0 → ∞ ! However, for finite V0 , no matter how large, there are a finite number
of bound states.
II. θ0 = a 2mV0 h < 1 (Shallow, narrow well)
As θ0 decreases, the number of bound states decreases as well. But, no matter how
small θ0 becomes (i.e., no matter how shallow the well becomes), there is always
at least one bound state remaining.
Think:
Normalize the wave function for the finite square well, that is, determine the
constant A in equations (3.53).
Analyze the intersection points in Figure 3.8.
Let’s now examine the scattering states. In the three regions we have
d 2ψ 2m ( E −V0 )
2
= − 2
ψ = −q 2ψ ; for x < 0 and x > a
dx h
d 2ψ 2mE
2
= − 2 ψ = −k 2ψ ; 0< x<a
dx h
The general solutions are:
 A eiqx + A e−iqx
 1 2

ψ ( x ) =  B1e + B2e−ikx
ikx
 iqx −iqx
C1e + C2e

We assume the particle is going from left to right in figure 3.7. When it encounters
the well it is either reflected or transmitted. Thus, A1 is the amplitude of the
incident wave, A2 is the amplitude of the wave reflected at x = 0 and C1 is the
amplitude of the wave transmitted at x = a . Since there is no particle (wave)
coming from the right side of the well in figure 3.7, C2 must be zero.
Note that when the wave functions above are multiplied by exp ( −iEt h ) , they
represent progressive waves:
C1e ( )
i qx − Et h
represents a wave traveling from left to right
C2e (
−i qx + Et h )
represents a wave traveling from right to left
The boundary conditions at x = 0 give
Continuity of ψ : A1 + A2 = B1 + B2 (3.58)
Continuity of dψ dx : q ( A1 − A2 ) = k ( B1 − B2 ) (3.59)
The boundary conditions at x = a give
Continuity of ψ : B1eika + B2e−ika = C1eiqa (3.60)
Continuity of dψ dx : kB1eika − kB2e−ika = qC1eiqa (3.61)
Working on equations (3.60) and (3.61), we obtain:
 k + q  i( q −k )a
B1 = C1   e (3.62)
 2k 
 k − q  i ( q + k )a
B2 = C1   e (3.63)
 2k 
Inserting equations (3.62) and (3.63) into equations (3.58) and (3.59) generates:
C1 iqa
A1 + A2 = e {k cos ka − iq sin ka} (3.64)
k

C1 iqa
A1 − A2 = e {q cos ka − ik sin ka} (3.65)
q
We can now express the amplitudes A2 and C1 in terms of the amplitude A1 :
A2 =
(
i k 2 − q 2 sin ka ) A1 (3.66)
(
2kq cos ka − i k 2 + q 2 sin ka )
2kqe−iqa
C1 = A1 (3.67)
(
2kq cos ka − i k 2 + q 2 sin ka )
The reflection and transmission probabilities are given by, respectively,
R=
A2
2
=
( )
k 2 − q 2 sin 2 ka
(3.68)
2 2
A1 2 2
4k q + k − q ( 2 2
) sin 2
ka
4k 2 q 2
2
C1
T= 2
= 2
(3.69)
A1 2 2
4k q + k − q ( 2 2
) sin 2
ka
For total transmission, T = 1 , equation (3.69) gives sin ka = 0
2mE a
which is valid if ka = nπ or = nπ
h
Consequently, for total transmission, the energy values will be the same as those
for the infinite potential well.
Think:
1) Verify all the equations (3.62) – (3.69).
2) Show that the sum of the reflection and transmission probabilities is 1.
3) An electron is trapped in a well 1Å wide, and V0 = 100eV deep. How many
possible bound states can this electron be in? (Explain clearly your reasoning)
(You may use me = 500 keV , hc = 2000 eV Å to keep the numbers simple)

3.5. The Finite Potential Barrier

The potential barrier is described by the V ( x)
potential energy function
V0
V for 0 < x < a
V ( x ) =  0
0 otherwise
where V0 > 0 . The wave function to the x=0 x=a x

left and to the right of the barrier must Figure 3.9: The potential barrier
satisfy the time independent Schrödinger
equation:
d 2ψ 2mE
2
= − 2 ψ = −k 2ψ (3.70)
dx h
where k is the usual constant k = 2mE h .
Inside the barrier ( 0 < x < a ) , the wave function must satisfy
d 2ψ 2m ( E − V0 )
2
=− 2
ψ = −q 2ψ (3.71)
dx h
where q = 2m ( E − V0 ) h , as before.
Let’s first consider a particle with E > V0 . The solutions in the different regions
for this case are:
 Aeikx + Be−ikx for x < 0


ψ ( x ) = Ceiqx + De−iqx for 0 < x < a
 ikx −ikx
 Fe + Ge for x > a
The solution for x < 0 consists of a plane-wave of amplitude A traveling to the

right [note that the time-dependent wave-function has the factor exp ( −iEt h ) ], and
a plane-wave of complex amplitude B traveling to the left. The first wave is
interpreted as an incoming particle, and the second as a particle reflected by the
potential barrier.

For the region x > 0 , the solution consists of a plane-wave of complex amplitude
F traveling to the right and a plane-wave of complex amplitude G traveling to the
left. The second wave is unphysical as there is no particle coming from the right
side of the barrier and, therefore, we set G = 0 . The wave traveling to the right
interpreted as a particle transmitted through the barrier.
Now, the boundary conditions require that ψ and dψ dx be continuous at x = 0

and x = a . At x = 0 we have
A+ B = C + D (3.72)
kA − kB = qC − qD (3.73)
Likewise, at x = a we have
Ceiqa + De−iqa = Feika (3.74)
qCeiqa − qDe−iqa = kFeika (3.75)
You are welcome to show that the above four equations yield
R=
B
2
=
( )
k 2 − q 2 sin 2 qa
(3.76)
A 2
2 2
(
4k q + k − q 2 2
) sin 2
qa
2
F 4k 2q 2
T= = 2
(3.77)
A 2 2
(
4k q + k − q 2 2
) sin 2
qa
Note that the reflection and transmission probabilities satisfy the condition
R +T =1.
Classically, a particle of energy E > V0 slows down when it moves across the
potential barrier and remains unaffected otherwise. That is to say, according to
classical physics, reflection is impossible when E > V0 and, thus, we must have
R = 0 and T = 1 . This classical result is obtained from equations (3.76) and (3.77)
in the limit of relatively small potential barriers (i.e., V0 E ). However, when V0
is of order E , there is a significant probability that the incident particle will be
reflected by the barrier. The reflection and transmission probabilities are plotted in
Figures 3.10 and 3.11.

We can see from equation (3.76) or Figure (3.11) that the probability of reflection
is zero at certain barrier widths irrespective of the energy of the incident particle.
These special barrier widths correspond to
qa = nπ ; n = 1,2,3,L (3.78)
The momenta of the particle outside and inside the barrier are, respectively,
poutside = hk and pinside = hq . Using the de Broglie relation p = h λ , we have,
inside the potential barrier:
2π (3.79)
q=
λinside
Combining equations (3.78) and (3.79) gives
a = n 
λinside 
 (3.80)
 2 
In words, the special barrier widths are integer multiples of half the de Broglie
wavelength of the particle inside the barrier. The absence of a net reflection at the
special barrier widths is accounted for by the destructive interference of the waves
reflected from left and right edges of the barrier.
1
0.9
T
0.8
0.7
0.6
R, T
0.5
0.4
0.3
0.2 R
0.1
0
0 0.2 0.4 0.6 0.8 1
Vo/E
Figure 3.10: Transmission and reflection probabilities for a potential barrier as a function
of the ratio V0 E . Here, a potential barrier of width a = 1.25λ is assumed. λ is the free-
space de Broglie wave-length, λ = 2π k .

1
0.9
0.8 T
0.7
0.6
R, T
0.5
0.4
0.3
0.2 R
0.1
0
0 1 2 3 4 5
/λ
a/λ
Figure 3.11: Transmission and reflection probabilities as a function of the ratio a λ ,

where a is the width of the barrier and λ is the free-space de Broglie wave-length. Here,
an incident particle of energy E = 4V0 3 is assumed.
Let us, now, consider the case E < V0 . In this case, we define q = −2m ( E − V0 ) h
which is real and positive.
The general solution to the time-independent Schrödinger equation then takes the
form
 Aeikx + Be−ikx for x < 0


ψ ( x ) = Ceqx + De−qx for 0 < x < a
 ikx
 Fe for x > a
The boundary conditions generate the following four equations:
At x = 0 : A+ B = C + D (3.81)
ik ( A − B ) = q ( C − D ) (3.82)
At x = a : Ceqa + De− qa = Feika (3.83)
q ( Ceqa − De− qa ) = ikFeika (3.84)

After considerable algebra, the above four equations yield
R=
B
2
=
( )
k 2 + q 2 sinh 2 qa
(3.85)
A 2
2 2
4k q + k + q ( 2 2
) sinh 2
qa
2
F 4k 2 q 2
T= = 2
(3.86)
A 2 2
4k q + k + q ( 2 2
) sinh 2
qa
Note again that R + T = 1 .
These results could have been easily obtained from Equations (3.76) and (3.77) by
simply replacing q by −iq .
Now, according to classical physics, a particle with a total energy less than the
height a potential barrier, E < V0 , cannot be reflected if incident on such a potential
barrier. That is, for such a particle, the classical probability of reflection is unity,
and the classical probability of transmission is zero.
0.9 R
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1 T
0
1 1.5 2 2.5 3 3.5 4 4.5 5
V0/E
Figure 3.12: Transmission and reflection probabilities as functions of the ratio V0 E . The
width of the potential barrier is taken to be a = 0.5λ , where λ is the free-space de
Broglie wave-length.

1
0.9
0.8 R
0.7
0.6
0.5
0.4
0.3
T
0.2
0.1
0
0 0.5 1 1.5 2
a/λ
Figure 3.13: Transmission and reflection probabilities as functions of the ratio a λ ,

where λ is the free-space de Broglie wave-length. The incident particle is assumed to
have energy E = 3V0 4 .
The reflection and transmission probabilities [Equations (3.85) and (3.86)] are
plotted in Figures (3.12) and (3.13). Figure (3.12) shows that the classical result,
2 2
R = 1 and T = 0 , is obtained for relatively thin barriers (i.e., qa 1 ) in the limit
where the height of the barrier is relatively large (i.e., V0 E ). However, when V0
is of order E , there is a substantial probability that the incident particle will be
transmitted by the barrier. According to classical physics, transmission is
impossible when V0 > E . Figure 3.13 shows that the transmission probability
decays exponentially as the width of the barrier increases. Nevertheless, even for
very wide barriers (i.e., qa 1 ), there is a small but finite probability that a
particle incident on the barrier will be transmitted. This phenomenon, which is
inexplicable within the context of classical physics, is called tunneling.
Think:
Using the notations ε = Vo E and λ = 2π h 2mE , show that equations (3.76) and
(3.77) can be written as
ε 2 sin 2 2π 1 − ε a λ
( ) 4 (1 − ε )
R= T=
4 (1 − ε ) + ε 2 sin 2 ( 2π 1 − ε a λ ) 4 (1 − ε ) + ε 2 sin 2 ( 2π 1 − ε a λ )

Think:
In Figure 3.9 we get a potential step if a → ∞ . Suppose a stream of particles of
mass m and energy E > 0 encounter a potential step of height V0 ( < E ) , that is,
V ( x ) = 0 for x < 0 and V ( x ) = V0 for x > 0 with the particles incident from − ∞ .
Show that the fraction reflected is
2
 k −q
R= 
k +q
where k 2 = 2mE h 2 and q 2 = 2m ( E − V0 ) h 2
3.6. Probability Current Density

At the beginning of this Chapter we have seen Max Born’s interpretation of the
wave function:
Ψ ( x, t ) is a function associated with the presence of a particle.

The probability of finding the particle somewhere between ( x, x + dx ) at
time t is given by
2
Probability = P ( x, t ) dx = Ψ ( x, t ) dx (3.87)
Since the particle is somewhere, the probability must be normalized to unity:

∞ ∞
2
∫ P ( x, t ) dx =
−∞
∫
−∞
Ψ ( x, t ) dx = 1 (3.88)
In this section, we will see that if equation (3.88) is satisfied at time t = 0 , then it
will stay satisfied at any time afterward. We begin by taking the time derivative of
the probability density function defined by equation (3.87).
2
P ( x, t ) = Ψ ( x, t ) = Ψ* ( x, t ) Ψ ( x, t ) (3.89)
∂P ( x, t ) ∂Ψ* ( x, t ) ∂Ψ ( x, t )
= Ψ ( x, t ) + Ψ * ( x, t ) (3.90)
∂t ∂t ∂t
From the time-dependent Schrödinger equation:

∂Ψ ( x, t ) ih ∂ 2 Ψ ( x, t ) i
= − V ( x ) Ψ ( x, t ) (3.91)
∂t 2m ∂x2 h
∂Ψ* ( x, t ) 2 *
ih ∂ Ψ ( x, t ) i *
=− + V ( x ) Ψ* ( x, t ) (3.92)
∂t 2m ∂x 2 h
Substituting equations (3.91) and (3.92) into equation (3.90) and assuming a real
potential energy function, we obtain:
∂P ( x, t ) i h  ∂ Ψ ( x, t )
2 *
∂ 2 Ψ ( x, t ) 
=−  *
Ψ ( x, t ) − Ψ ( x, t )  (3.93)
∂t 2m  ∂x 2 ∂x 2


Or
ih ∂  ∂Ψ ( x, t ) ∂Ψ ( x, t ) 
*
∂P ( x, t )
=− Ψ ( x, t ) − Ψ * ( x, t )  (3.94)
∂t 2m ∂x  ∂x ∂x 
Now, we define a new function J ( x, t ) by
ih  ∂Ψ ( x, t ) ∂Ψ ( x, t ) 
*
J= Ψ ( x, t ) − Ψ * ( x, t )  (3.95)
2m  ∂x ∂x 
Equation (3.94) now becomes
∂P ( x, t ) ∂J ( x, t )
=− (3.96)
∂t ∂x
Integrating equation (3.96) we find

∞
∂P ( x, t ) ∞
∂J ( x, t )
∫−∞ ∂t dx = − ∫−∞ ∂x dx = − J ( ∞, t ) + J ( −∞, t ) = 0 (3.97)
Note that J ( x, t ) must go to zero at infinity because the square integrable wave
function and its first derivative vanish at infinity.
∞
d
Thus, ∫
dt −∞
P ( x, t ) dx = 0 (3.98)

That is what we set off to prove: The probability integral is independent of time; if
it equals 1 at some time, t = 0 , it equals 1 for all other times!
Now, what exactly is the function J ( x, t ) ? To see the physical significance of this
function, let’s redo the integration in equation (3.97) from x = a to x = b , instead
of integrating over from −∞ to ∞ :
d
b b
∂J ( x, t )
∫
dt a
P ( x, t ) dx = −∫ ∂x
dx = − J (b, t ) + J ( a, t ) = 0 (3.99)
a
The left hand side of equation (3.99) is the time variation of the probability of
finding the particle somewhere between x = a and x = b . If we think of a stream of
particles, this equation then tells us about the flux of particles entering the region
( a, b ) per unit time. So, J ( x, t ) must represent the flux, or probability current at
the point x = a at the time t .
Equation (3.99) can be expressed in differential form as
∂P ( x, t ) ∂J ( x, t )
=− (3.100)
∂t ∂x
which in 3D becomes:
∂P ( x, y, z; t )
= −∇⋅ J ( x, y, z; t ) (3.101)
∂t
3.7. Schrödinger’s and Heisenberg’s Pictures

The time evolution of a quantum system has different mathematical formulations.
So far we have followed the mathematical representation in which the state of a
quantum system is described by a time-dependent wavefunction which determines
probability densities. This is known as the Schrödinger representation of quantum
mechanics. In this representation, we have seen that quantum observables are
expressed in terms of operators that remain the same at all instants of time. For
example, the momentum operator pˆ x is mathematically represented as −ih ∂ ∂x for
any time t .
The time evolution of the wave function in the Schrödinger picture can be written
as

Ψ ( x; t ) = U ( t ) Ψ ( x;0) (3.102)
where U ( t ) is the time evolution operator that turns into an identity operator at
t = 0 , that is,
U ( 0) = 1 (3.103)
Moreover, U ( t ) must be unitary to satisfy the normalization condition imposed on

the wavefunction:
∞ ∞
*
∫−∞ Ψ * ( x; t ) Ψ ( x; t ) dx = −∞∫ Ψ ( x;0) Ψ ( x;0) dx = 1 for all t
∞ ∞ *
U ( t ) Ψ
∫ Ψ * ( x; t ) Ψ ( x; t ) dx = ∫  ( x;0) U (t ) Ψ ( x;0) dx
  
−∞ −∞
∞
∫−∞ Ψ ( x;0)U (t )U (t ) Ψ ( x;0) dx

* †
=
Obviously,
U † ( t ) U ( t ) = U ( t )U † ( t ) = 1 (3.104)
Let’s now use equation (3.102) in the time-dependent Schrödinger equation:
∂
ih Ψ ( x; t ) = Hˆ Ψ ( x; t )
∂t
∂
ih U ( t ) Ψ ( x;0)  = HU
ˆ ( t ) Ψ ( x;0)
∂t  
∂U ( t ) ˆ ( t ) Ψ ( x;0 )
ih Ψ ( x;0) = HU
∂t
Here, the last equation generates an important relation, namely,
∂U ( t ) i
= − Hˆ U ( t ) (3.105)
∂t h
The Hermitian conjugate of equation (3.105) produces:

∂U † ( t ) i †
= U ( t ) Hˆ (3.106)
∂t h
Since the wavefunction has no direct physical meaning, we will eventually be
interested in observables. So instead of concentrating on a time-evolving
wavefunction, we can look at a time-evolving operator. In this case, the
expectation value of the corresponding observable is, in the bra-ket notation:
A ( t ) = Ψ ( t ) Â Ψ ( t )
ˆ U (t ) Ψ ( 0)
= U (t ) Ψ (0) A
ˆ U (t ) Ψ ( 0)
= Ψ ( 0) U † (t ) A
The last expression here suggests a time-varying operator:
ˆ (t ) = U † (t ) A
A ˆ U (t ) (3.107)
Equation (3.107) tells us that the dynamics of a quantum system can be analyzed
using time-dependent operators. This is an alternative approach to the study of
quantum mechanics and is known as the Heisenberg representation of quantum
mechanics.
So, we obtained two different physical interpretations:
1) Schrödinger Picture
Leave operators unchanged but transform the eigenvectors:
Ψ (t ) = U (t ) Ψ ( 0)
2) Heisenberg Picture
Leave eigenvectors unchanged but transform the operators:
ˆ (t ) = U † (t ) A
A ˆ U (t ) .
This second picture is physically appealing because in reality particles move, that
is, position and momentum evolve with time.

Let’s look at the time-evolution of a quantum system in the two representations.

First, the Schrödinger picture:
The Schrödinger Picture
The time-dependent Schrödinger equation is
∂
ih Ψ = Ĥ Ψ (3.108)
∂t
In the bra notation this becomes:
∂ ˆ†= ˆ
−ih Ψ = Ψ H Ψ H (3.109)
∂t
Note that the Hamiltonian operator is Hermitian: Hˆ † = Hˆ
The expectation value of an observable A that is represented by an operator Â is
ˆ Ψ (t )
A = Ψ (t ) A (3.110)
Taking the time derivative of equation (3.109) and multiplying by ih :
∂Ψ ( t ) ˆ ˆ
ih
d A
= ih A Ψ ( t ) + ih Ψ ( t )
∂A ˆ ∂Ψ ( t ) (3.111)
Ψ ( t ) + ih Ψ ( t ) A
dt ∂t ∂t ∂t
Since operators in the Schrödinger picture do not have explicit time dependence,
we must have
∂Â
=0 (3.112)
∂t
Using equations (3.108), (3.109) and (3.112) in equation (3.110), we get:
d A (t )
ih ˆ ˆ Ψ ( t ) + Ψ ( t ) AH
= −Ψ ( t ) HA ˆ ˆ Ψ (t )
dt
ˆ ˆ − HA
= Ψ ( t ) AH ˆ ˆ Ψ (t )
= Ψ (t ) ˆ ˆ
 A,H Ψ (t )
 

d A (t )
ih ˆ ˆ
= A,H (3.113)
dt  
In the Schrödinger representation, therefore, an observable will be a constant of

motion if it commutes with the Hamiltonian.
Heisenberg Picture
In this picture of quantum mechanics, the Schrödinger operators are transformed

into Heisenberg operators by equation (3.107):
ˆ (t ) = U † (t ) A
A ˆ U (t ) (3.114)
H S
with the condition
ˆ ( 0) = A
A ˆ (3.115)
H S
The time-evolution of the operators in the Heisenberg picture is:
ˆ ( t ) ∂U † ( t )
dA ˆ
H
= A ˆ ∂U ( t )
ˆ U (t ) + U † ( t ) ∂A S U ( t ) + U † ( t ) A
dt ∂t S ∂t S ∂t
The middle term vanishes because Schrödinger operators are time independent.
So,
dÂ H ( t ) ∂U † ( t ) ˆ ˆ ∂U ( t )
= A SU ( t ) + U † ( t ) A S (3.116)
dt ∂t ∂t
Using equation (3.104) in equation (3.116):
dÂ H ( t ) ∂U † ( t ) ˆ U (t )U † ( t ) ∂U ( t )
ˆ U (t ) + U † (t ) A
= U ( t )U † ( t ) A S S
dt ∂t ∂t
But we have, from the time derivative of equation (3.104),
∂U † ∂U
U = −U †
∂t ∂t

Using this result and equation (3.114) yields
dÂ H ( t ) ∂U † ( t ) ˆ ˆ ∂U † ( t )
= U (t ) A H − A H U (t ) (3.117)
dt ∂t ∂t
At this point, we go back to the Schrödinger picture and borrow equation (3.106)
and plug it into equation (3.117) using the subscript S for Schrödinger.
Here is what we have:
dÂ H ( t ) i † ˆ
dt { ˆ −A
= U (t ) HS U (t ) A
h H
ˆ U † (t ) H
H
ˆ U (t )
S }
But, we know that,
ˆ = U † (t ) H
H ˆ U (t )
H S
Therefore,
dÂ H ( t ) ˆ ˆ ˆ A ˆ = Aˆ ,H
ˆ 
ih = AH HH − H H H  H H (3.118)
dt  
Equation (3.118) is known as the Heisenberg equation of motion. We have shown

that, if Â does not explicitly depend on time then
dÂ H ( t ) 1  ˆ ˆ 
= A H ,H H  (3.119)
dt ih  
describes the time evolution of operators in the Heisenberg picture.
Think:
Given the unitary operator
U ( t ) = e−iHt h
for a time-independent Hamiltonian, show that U and H commute, and for a time-
dependent Hamiltonian, U and H need not commute.

Since the wavefunction in the Heisenberg picture is time-independent,
∂
Ψ =0
∂t H
we can relate the Schrödinger and Heisenberg wavefunctions as
Ψ S (t ) = U (t ) Ψ H (3.120)
So,
Ψ H = U † (t ) Ψ S (t ) = Ψ S ( 0)
The predictions of quantum mechanics are independent of the representation:
In both pictures, the expectation values are the same:
ˆ Ψ = Ψ UA
AH = Ψ H A ˆ U† Ψ = Ψ Aˆ Ψ = A
H H S H S S S S S
In either picture the eigenvalues are preserved:
ÂS ψ n S
= an ψ n S
ÂSU ψ n H
= anU ψ n H
by (3.120)
ˆ U ψn
U† A = anU †U ψ n
S H H
ÂH ψ n H
= an ψ n H
by (3.104) and (3.107)

Chapter 4. The Harmonic Oscillator
In this Chapter a basic quantum system will be discussed in detail. The system is a
particle constrained by forces to remain close to its equilibrium position. If the
forces pulling the particle back to its equilibrium position are negatively
proportional to the distance of the particle from this position, you have what is
called the harmonic oscillator. A real example of such a system is an atom in a
solid or in a molecule.
A model of a 3D harmonic oscillator is shown in Figure 4.1. The constraining

forces are represented by springs and the particle’s displacement from its
equilibrium position will be indicated by ( x, y, z ) . It will be assumed that all the
z three springs are equally stiff, that is, they
have the same spring constant k .
y
The classical frequency of vibration for
k the harmonic oscillator is
m x
k
k ω= k (4.1)
m
in radians per second. This frequency is
used often in quantum calculations.
Figure 4.1: The 3D harmonic oscillator
4.1. The Hamiltonian

As in Chapter 3, the energy spectrum of the oscillating particle can be found by
first identifying the total energy operator – the Hamiltonian.
For the 3D oscillator shown in Figure 4.1, the total potential energy is the sum of
the potential energies due to each spring. That is,
1
(
V ( x, y, z ) = k x2 + y 2 + z 2
2 ) (4.2)
Adding the kinetic energy operator to this, we get the Hamiltonian:
h2  ∂ 2 ∂ 2 ∂ 2  1
Ĥ = − + +
2m  ∂x2 ∂y 2 ∂z 2  2 (
+ k x2 + y 2 + z 2 ) (4.3)

4.2. Separable Solutions of the Schrödinger Equation

The energy eigenfunctions ψ and energy eigenvalues E of the harmonic oscillator
must satisfy the time-independent Schrödinger equation:
 2  2 
− h  ∂ ∂ 2 ∂ 2  1
 2m  ∂x 2

+ + +
∂y 2 ∂z 2  2
k x 2
(
+ y 2
+ z)2 

ψ = Eψ (4.4)
 
The boundary condition is that ψ becomes zero at large distance from the
equilibrium position. This must be true because at large distances we have a
system with infinite potential energy which is unphysical. Therefore, the chances
of finding the particle far away from the equilibrium position should be
vanishingly small.
Like the potential well eigenfunctions, the eigenfunctions of the harmonic

oscillator are assumed to be separable. That is to say, each eigenfunction ψ is a
product of one-dimensional eigenfunctions, one in each direction:
ψ ( x, y, z ) =ψ x ( x )ψ y ( y )ψ z ( z ) (4.5)
Substituting equation (4.5) into equation (4.4) and dividing the resulting equation
by ψ x ( x )ψ y ( y )ψ z ( z ) we obtain:
 h 2 1 ∂ 2ψ x 1 2   h 2 1 ∂ 2ψ y 1 2   h 2 1 ∂ 2ψ z 1 2 
− 2 + kx  +  − 2 + ky  +  − + kz  = E
 2m ψ x ∂x 2   2m ψ y ∂y 2   2m ψ z ∂z 2
 2 
This equation reveals that E consists of three parts corresponding to each

eigenfunction. Labeling these energies by Ex , E y and Ez we have:
E = Ex + E y + Ez (4.6)
h2 1 d 2ψ x 1 2 (4.7)
− + kx = Ex
2m ψ x dx2 2
2
h2 1 d ψ y 1 2 (4.8)
− + ky = E y
2m ψ y dy 2 2
h2 1 d 2ψ z 1 2 (4.9)
− + kz = Ez
2m ψ z dz 2 2

Note that Ex , E y and Ez are constants independent of all three variables x, y, z .
Each of the equations above represents a one-dimensional eigenvalue problem.

Since each have the same mathematical form, their solutions must also have the
same mathematical form. Therefore, we first obtain the eigenfunction ψ x ( x ) and
the corresponding eigenvalue Ex .
4.3. Eigenvalues and Eigenfunctions for the 1D Harmonic Oscillator

Equation (4.7) can be written as
d 2ψ x 2m  1 2 
(4.10)
=  kx − E ψx
dx2 h2  2
x

Using equation (4.1), the classical angular frequency of the harmonic oscillator,
equation (4.10) becomes:
d 2ψ x  m2ω 2 2 2m  (4.11)
= x − 2 Ex ψ x
dx2  h2 h 
Now, for convenience, we change the variable from x to ξ so that
ξ = mω x (4.12)
h
Accordingly, we change the derivative with respect to x to derivative with respect
to ξ :
dψ x dψ x dξ mω dψ x
= =
dx dξ dx h dξ
Differentiating once more:
d 2ψ x mω d 2ψ x
= (4.13)
dx2 h dξ 2
Using equations (4.12) and (4.13) into equation (4.11) gives
d 2ψ x
dξ 2 ( )
− ξ 2 −ε ψ x = 0 (4.14)

where ε = 2 Ex (4.15)
hω
The solution to equation (4.14) above must satisfy the boundary condition ψ x → 0
as ξ → ∞ . In the limit ξ → ∞ , equation (4.14) simplifies to
d 2ψ ∞
2
− ξ 2ψ ∞ 0 (4.16)
dξ
2 2
with asymptotic solution ψ ∞ = Ce−ξ (4.17)
Think:
2 2
Check that ψ ∞ = Ce−ξ 2 + De+ξ 2 a more general approximate solution to
equation (4.16) above. Here we must set D = 0 . Why?
Now, the solution to equation (4.14) must have the form:

2 2
ψ x (ξ ) = h (ξ ) e − ξ (4.18)
2
where h (ξ ) is a function yet to be determined such that ψ x = he− ξ 2
→ 0 as
ξ → ∞ . Substituting equation (4.18) into equation (4.14), we obtain
d 2h dh
dξ (
2
− 2ξ + ε −1) h = 0 (4.19)
dξ
Let us attempt a power-series solution of the form

∞
h (ξ ) = ∑ a jξ j (4.20)
j =0
The coefficients a j are independent of ξ and can be determined easily.
Substituting equation (4.20) into equation (4.19) yields:

∞ ∞ ∞
∑
j =2
j ( j −1) a jξ j −2
− 2ξ ∑ ja jξ
j =1
j −1
+ (ε −1) ∑ a jξ j = 0
j =0
After adjusting the index j so all the summations start from j = 0 , we get
∞ ∞ ∞
∑
j =0
( j + 2)( j +1) a j+2ξ − ∑ 2 ja jξ
j =0
j j
+ ∑ (ε −1) a jξ j = 0
j =0
Or
∞
∑
j =0
 j + 2 j +1 a
(

)( ) j+2 − ( 2 j +1− ε ) a j ξ j = 0 (4.21)
Equation (4.21) clearly shows that for the power series to vanish the coefficients
of ξ j must be zero for all ξ and for each j . Therefore,
a j +2 =
( 2 j +1− ε ) a (4.22)
( j + 2)( j +1) j
Equation (4.22) is the recursion relation between the coefficients. The recursion
relation shows that only the first two coefficients, a0 and a1 , are completely
arbitrary constants and the rest can be determined in terms of these two. If a0 is
known, we can use the recursion relation to generate a2 , a4 , a6 , a2 ,L and if a1 is
known we can generate a3 , a5 , a7 , a9 ,L . So,
Even j Odd j
a0 a1
1− ε 3−ε
a2 = a a3 = a
2! 0 3! 1
a4 =
(5 − ε ) (1 − ε ) a a5 =
( 7 − ε ) (3 − ε ) a
0 1
4! 5!
a6 =
(9 − ε ) (5 − ε ) (1 − ε ) a a7 =
(11 − ε ) ( 7 − ε ) (3 − ε ) a
0 1
6! 7!
M M

Equation (4.20) now becomes the sum of even and odd power series:
 1 − ε 2 ( 5 − ε ) (1 − ε ) 4 ( 9 − ε ) ( 5 − ε ) (1 − ε ) 6 
h (ξ ) = a0 1 + ξ + ξ + ξ + L
 2! 4! 6! 
 
(4.23)
 3 − ε 3 ( 7 − ε ) ( 3 − ε ) 5 (11 − ε ) ( 7 − ε ) ( 3 − ε ) 7 
+ a1 ξ + ξ + ξ + ξ + L
 3! 5! 7! 
 
Both these series converge for all ξ since the ratio of two consecutive coefficients
approaches zero as j → ∞ :
a j +2 ( 2 j + 1 − ε ) 2
= → → 0 for j → ∞ (4.24)
a j ( j + 2 )( j + 1) j
But this convergence is not rapid enough to satisfy the requirement that
2
ψ x = he −ξ 2 → 0 in the limit ξ → ∞ . To see this, let’s examine the asymptotic
behavior of both the even and odd power series in equation (4.23) above.
For j 1 ,
∞
1 2
h (ξ ) → C ∑ ξ 2 j → Ce ξ j 1 (4.25)
j j!
Inserting this into equation (4.18), we can see that in the limit ξ → ∞ and for
dominant powers of h (ξ ) , the wavefunction ψ x varies as
2 2 2 2 2 2 2
ψ x = he−ξ e ξ e−ξ = eξ →∞
This behavior of ψ x for large values of ξ is unacceptable, since it does not meet
the boundary condition ψ x → 0 as ξ → ∞ . The wavefunction ψ blows up because
we let j → ∞ in the power series expansion of h (ξ ) [equation (4.20)]. The
wavefunction remains finite as ξ → ∞ if the power series is cut off at some value
of j , say j = n , so that an+ 2 = 0. Hence, from the recursion relation (equation
4.22) we get:
an+2 =
( 2n +1− ε ) a = 0
( n + 2)( n +1) n
Or ε = 2n + 1 (4.26)
Combining equations (4.15) and (4.26) we obtain the energy eigenvalues for the
1D harmonic oscillator:
Exn = ( n + 12 ) hω , n = 0,1,2,3,L (4.27)
This result tells us that a harmonic oscillator has equally spaced quantized energy
levels, the spacing between successive energy levels being hω , where ω is the
classical frequency of oscillation. When n = 0 , the 1D oscillator possesses the
lowest finite energy which is sometimes called the zero-point energy:
Ex 0 = 12 hω zero-point energy (4.28)
V(x)
E 6 =13ħω/2
E 5 =11ħω/2
E 4 =9ħω/2
E 3 =7ħω/2
E 2 =5ħω/2
E 1 =3ħω/2
E 0 = ħω/2
x
Figure 4.2: Energy levels of the quantum oscillator
Having determined the proper energies for the simple harmonic oscillator in one
dimension, the next step is to obtain the corresponding eigenfunctions. Using
equation (4.26) in equation (4.23) produces:
For even n ( a0 ≠ 0 but a1 = 0 ) ,
 2n 2 2n ( 2n − 4 ) 4 2n ( 2n − 4 )L8 × 4 n 
hn (ξ ) = a0 1 − ξ + ξ −L + ξ , (4.29)
 2! 4! n! 
 

and for odd n ( a0 = 0 but a1 ≠ 0 ) ,
 2n − 2 3 ( 2n − 2 )( 2n − 6) 5 ( 2n − 2 )( 2n − 6 )L8 × 4 n 
hn (ξ ) = a1 ξ −
 ξ + ξ −L + ξ ,
 3! 5! n! 
 
(4.30)
The constants a0 and a1 are so defined that the coefficient of ξ n is 2n . Therefore,

for the even and odd series we have
a0 = (−1)n 2
n! , even n (4.31)
( n 2 )!
n!
a1 = 2(−1)( )
n −1 2
, odd n (4.32)
( n − 1) 2  !
 
With this choice a0 and a1 , the functions given by equations (4.29) and (4.30)
become Hermite polynomials, H n (ξ ) . If we write these polynomials in descending
powers of ξ , we get a single expression valid for both even and odd n :
n n ( n − 1) n−2 n ( n − 1)( n − 2 )( n − 3) n−4

H n ( ξ ) = ( 2ξ ) − ( 2ξ ) + ( 2ξ ) − L (4.33)
1! 2!
The Hermite polynomials can be generated by
n 2 d n −ξ 2
H n (ξ ) = ( −1) eξ e (4.34)
dξ n
The first seven Hermite polynomials are listed below.
H0 = 1
H1 = 2ξ
H 2 = 4ξ 2 − 2
(4.35)
H 3 = 8ξ 3 −12ξ
H 4 = 16ξ 4 − 48ξ 2 + 12
H 5 = 32ξ 5 −160ξ 3 + 120ξ
H 6 = 64ξ 6 − 480ξ 4 + 720ξ 2 −120

Think:
Obtain equations (4.31) and (4.32) from equations (4.29) and (4.30).
Obtain H 7 and H 8 first using equation (4.33) and then using equation (4.34).
Now, the wavefunctions for the 1D harmonic oscillator [equation (4.18)] could be
written as
2 2
ψ n (ξ ) = Cn H n (ξ ) e−ξ (4.36)
The constant Cn is determined from the normalization condition:
∞ ∞
2
∫ ψ n2 (ξ ) d ξ = Cn2 ∫ e−ξ H n2 (ξ ) dξ = 1 (4.37)
−∞ −∞
Using equation (4.34) and carrying out the integration it is easy to show that
1
Cn = (4.38)
π 1 4 2 n n!
Combining equations (4.36) and (4.38) the eigenfunctions become:
1 2
ψ n (ξ ) = 14
H n (ξ ) e−ξ 2 (4.39)
π 2 n!
n
The lowest energy state ( n = 0 ) is

2 2 2 2
ψ 0 ( ξ ) = π −1 4 H 0 e − ξ = π −1 4 e − ξ (4.40)
Or, using equation (4.12) this becomes:
14
ψ 0 ( x ) =  mω 
  2 2h
e− mω x (4.41)
 πh 
As a simple check, it is easy to show that:

2 2
 h2 ∂ 2 1 2   mω  − mω x2 2h  hω   mω  −mω x2 2h
ˆ
Hψ 0 =  − + 2 kx    e =   e = E0ψ 0
2
 2m ∂x  π h   2  π h 

Think:
Verify that the eigenfunctions given by equation (4.39) satisfy
∞
∫−∞ψ nψ mdξ = δ nm
Show how to obtain equation (4.41) from equation (4.40).
Write the wavefunctions ψ 1 , ψ 2 , ψ 3 , ψ 4 , ψ 5 , ψ 6 and discuss their nature.
ψ0 ψ1 ψ2
ψ3 ψ4 ψ5
Figure 4.3: The first 6 wavefunctions of the quantum oscillator
4.4. The Raising and Lowering Operators

Equations (4.14) and (4.26) can be combined to give:
 d2 
− + ξ 2 ψ xn = ( 2n + 1)ψ xn (4.42)
 dξ 2 
 
This can be rewritten as
1  d 2 
hω − 2 + ξ 2 ψ xn = ( n + 12 ) hωψ xn (4.43)
2  dξ 

Obviously, the operator acting on the left side of this equation is the Hamiltonian:

1  d2 
Hˆ = hω  − 2 + ξ 2  for 1D oscillator (4.44)
2  dξ 

Consider now the differential operators
1  d  1  d 
aˆ+ =  − + ξ  and aˆ− =  + ξ  (4.45)
2  dξ  2  dξ 
Combining the two:
1 d  d 
aˆ+ aˆ−ψ xn =  − + ξ 
 dξ
+ ξ ψ
 xn
2  dξ  
(4.46)
1  d 2ψ xn 
= −
2  dξ 2
+ ξ 2ψ xn −ψ xn  =
 ( n + 12 )ψ xn − 12ψ xn

Similarly,
1 d  d 
aˆ− aˆ+ψ xn =  + ξ  −
 dξ
+ ξ ψ
 xn
2  dξ  
(4.47)
1  d 2ψ xn 
xn  = ( n + 2 )ψ xn + 2ψ xn
2 1 1
= − + ξ ψ +ψ 
2  dξ 2 xn

Simplifying equations (4.46) and (4.47) further we obtain:
aˆ+ aˆ−ψ xn = nψ xn (4.48)
aˆ− aˆ+ψ xn = ( n + 1)ψ xn (4.49)
Subtracting equation (4.49) from equation (4.48) we obtain the commutation

relation:
 aˆ+ , aˆ−  = aˆ+ aˆ− − aˆ− aˆ+ = −1 (4.50)

 
Multiplying both sides of equation (4.46) by hω and rearranging we obtain:

hω ( aˆ+ aˆ− + 12 )ψ xn = ( n + 12 ) hωψ xn (4.51)
We can see from this equation that the Hamiltonian for the harmonic oscillator can
be written in the form
Hˆ = hω ( aˆ+ aˆ− + 12 ) (4.52)
But what is the effect of the operator â+ or â− acting on the wavefunction ψ n ?
Let’s see:
Hˆ ( aˆ+ψ n ) = hω ( aˆ+ aˆ− + 12 ) aˆ+ψ n = hω ( aˆ+ aˆ− aˆ+ + 12 aˆ+ )ψ n = hω aˆ+ ( aˆ− aˆ+ + 12 )ψ n
Using equation (4.49) we have:
Hˆ ( aˆ+ψ n ) = hω ( n + 1 + 12 ) aˆ+ψ n = En+1 ( aˆ+ψ n ) (4.53)
Similarly,
Hˆ ( aˆ−ψ n ) = hω ( n −1 + 12 ) aˆ−ψ n = En−1 ( aˆ−ψ n ) (4.54)
Equation (4.53) shows that ( aˆ+ψ n ) is an eigenfunction of the Hamiltonian with

eigenvalue En +1 . Therefore,
aˆ+ψ n = αψ n+1, where α depends on n
Likewise, aˆ−ψ n = βψ n−1, where β depends on n
α and β cab be determined by inspection of equations (4.48) and (4.49):
aˆ+ψ n = n +1ψ n+1 (4.55)
aˆ−ψ n = nψ n−1 (4.56)
We conclude that â+ and â− are raising and lowering operators, respectively, for
the harmonic oscillator, that is, operating on the wavefunction with â+ causes the
quantum number n to increase by unity, and vice versa.

4.5. The 3D Harmonic Oscillator

The 3D eigenvalue problem of the simple harmonic oscillator has now been
solved, because the equations for the y and z directions are mathematically the
same as the equation for the x direction and must therefore have identical
solutions:
E yn = ( ny + 12 ) hω, ny = 0,1,2,3,L (4.57)
Ezn = ( nz + 12 ) hω, nz = 0,1,2,3,L (4.58)
The total energy E of the 3D harmonic oscillator is given by equation (4.6):
E = ( nx + ny + nz + 32 ) hω , nx , ny , nz = 0,1, 2,3,L (4.59)
The corresponding eigenfunctions are:
ψ nx ny nz (ξ ,ζ ,η ) =ψ x (ξ )ψ y (ζ )ψ z (η )
1 − 12 ξ 2 +ζ 2 +η 2  (4.60)
= 32
H nx (ξ ) H ny (ζ ) H nz (η ) e  
π 3 4 ( 2n n!)
where
ξ = mω h x, ζ = mω h y, η = mω h z
Think:
1. Write out the ground state energy and the energy E100 .
3. Write out the ground state wavefunction and write the eigenstate ψ 100 fully.
4.6. Discussion of the Eigenvalues

We have seen that every set of three nonnegative integers nx , ny , nz , determines
one unique eigenstate, equation (4.60), and a corresponding energy eigenvalue,
equation (4.59). In this section we will look at the energy eigenvalues in detail.
Discrete energy levels: The energy eigenvalues are of great physical importance,
because according to Postulate 3 in Chapter 1, they are the only measurable values

of the total energy; the oscillator cannot have any other total energy than these
eigenvalues. Figure 4.4 shows the energy spectrum of the quantum oscillator and
the sets of quantum numbers nx , ny , nz generating each energy level.
nx = 1 2 1 2 1 0 0 3 0 0 Ten states with

E = 92 hω ny = 1 1 2 0 0 2 1 0 3 0 the same energy
nz = 1 0 0 1 2 1 2 0 0 3
nx = 1 1 0 2 0 0 Six states with

7
E = hω
2
ny = 1 0 1 0 2 0 the same energy
nz = 0 1 1 0 0 2
nx = 1 0 0 Three states with the same energy

5
E = hω
2
ny = 0 1 0 E100 = E010 = E001
nz = 0 0 1
nx = 0
E000 = 32 hω ny = 0 Lowest-energy state
nz = 0
Figure 4.4: Energy levels (spectrum) of the simple harmonic oscillator
The energy spectrum clearly shows that the energy of the quantum oscillator takes
only certain discrete values which evenly spaced. Note that each energy value is
an odd multiple of 12 hω . However, h is so small that on a macroscopic scale the
different energy levels are indistinguishably close together as if they form a
continuum. The Newtonian theory which assigns continuous energy values for the
harmonic oscillator is therefore an excellent macroscopic approximation though it
is strictly speaking incorrect.
No escape: Note that the energy levels have no largest value. Meaning, however
high the energy of the oscillator may be, it will never escape: the further it goes
from the equilibrium position, the larger the forces that pull it back.
Zero-point energy: A striking feature of the energy spectrum is the existence of a

nonzero minimum energy which occurs for nx = 0, ny = 0, nz = 0 . This energy
which is the 3D version of equation (4.28) is:
E000 = 32 hω (4.61)

The physical meaning of equation (4.61) is that even at absolute zero temperature,
the particle is not completely at rest at its equilibrium position. It still has 3hω 2
worth of kinetic and potential energy left. This lowest-energy state is the ground
state of the quantum oscillator.
But why is the energy not zero at absolute zero? The answer comes from
Heisenberg’s uncertainty principle. If the potential energy is zero, the particle must
certainly be at its equilibrium position for. But the uncertainty principle does not
allow a certain position. Conversely, if the kinetic energy is zero, the linear
momentum would have to be zero for certain which is again not allowed by the
uncertainty principle.
Degeneracy: As seen in the energy spectrum in Figure 4.4, except the ground
state, three or more eigenstates produces the same energy. When multiple states
produce the same energy degeneracy is said to occur and such states are said to be
degenerate states.
Note that the number of degenerate sates increases rapidly when we go up the
energy level. For example, there are three degenerate states ψ 100 , ψ 010 , and ψ 001
with the energy E = 52 hω and ten degenerate states with the energy E = 92 hω .
Think:
1. Verify that the sets of quantum numbers shown in the spectrum Figure (4.4) do
indeed produce the indicated energy levels.
2. Verify that there are no sets of quantum numbers missing in Figure (4.4). The
listed ones are the only ones that produce those energy levels.
3. Write the energy level that should come next in Figure (4.4) with all sets of
quantum numbers nx , ny , nz producing it.
4. Write out the ground state wavefunction and show that it is spherically
symmetric.
5. Show that the ground state wave function is maximal at the origin and, like all
the other energy eigenfunctions, becomes zero at large distances from the
origin.
6. Show that the 3D harmonic oscillator eigenfunctions are orthonormal, that is,
ψ 000 ψ 000 = 1, ψ 100 ψ 000 = 0, ψ 100 ψ 100 = 1, ψ 010 ψ 100 = 0, etc

Chapter 5. Angular Momentum
5.1. Definition of Angular Momentum

This Chapter looks at the properties of angular momentum which will serve as the
basis to understand the structure of the hydrogen atom discussed in Chapter 6. We
will explore the quantum properties of angular momentum starting with its
classical definition.
ur
Newtonian physics defines the angular momentum vector L as the cross product
uur uur uur
L = r× p (5.1)
r ur
where r is the position of the particle in question and p is its linear momentum.
Or in terms of Cartesian components:
Lx = ypz − zp y
Ly = zpx − xpz (5.2)
Lz = xp y − ypx
Following this rdefinition,

ur
quantum mechanics substitutes the operator
representations of r and p , to express the angular momentum operator as:
uur
Lˆ = rˆ × −ih∇
( ) (5.3)
ur  ∂ ∂ ∂ 
where ∇ ≡  , ,  is the gradient operator.
 ∂x ∂y ∂z 
In terms of components, equation (5.3) becomes:
 ∂ ∂
Lˆx = ypˆ z − zpˆ y = −ih  y − z 
 ∂z ∂y 
 ∂ ∂
Lˆ y = zpˆ x − xpˆ z = −ih  z − x  (5.4)
 ∂x ∂z 
 ∂ ∂
Lˆz = xpˆ y − ypˆ x = −ih  x − y 
 ∂y ∂x 

Unlike the Hamiltonian, the angular momentum operator is not specific to a given
system. All observations about angular momentum will apply regardless of the
physical system being studied. Therefore, in the following we investigate the
commutation relations of the angular momentum operators.
5.2. Commutation of Angular Momentum Operators

We know that
 xˆ , p 
 i ˆ j  = −  pˆ j , xˆi  = ihδi j 
 

 xˆ , xˆ  = 0  where i, j = 1, 2, 3 (5.5)
 i j  
pˆ , pˆ  = 0 
 i j  
Using these relations and equations (5.4), we can now obtain the commutation
relations between the components of angular momentum:
 Lˆx , Lˆ y  =  ypˆ z − zpˆ y , zpˆ x − xpˆ z  = ypˆ z − zpˆ y ( zpˆ x − xpˆ z ) − ( zpˆ x − xpˆ z ) ypˆ z − zpˆ y
( ) ( )
   
= ypˆ x pˆ z z − xypˆ z pˆ z − zzpˆ x pˆ y + xpˆ y zpˆ z − ypˆ x zpˆ z + zzpˆ x pˆ y + xypˆ z pˆ z − xpˆ y pˆ z z
= ypˆ x ( pˆ z z − zpˆ z ) + xpˆ y ( zpˆ z − pˆ z z )
= [ z, pˆ z ] ( xpˆ y − ypˆ x ) = ihLz
Instead of repeating the whole procedure for  Lˆ y , Lˆz  and  Lˆz , Lˆx  , we can the
   
results for these commutators by cyclic permutation of the indices, i.e., by
replacing x by y , replacing y by z , and replacing z by x . Hence, we have
 Lˆ , Lˆ  = ihL  Lˆ , Lˆ  = ihL  Lˆ , Lˆ  = ihL (5.6)

 x y  z  y z  x  z x  y
Let’s now define an operator for the square of the angular momentum by
Lˆ2 = Lˆ Lˆ = Lˆ2x + Lˆ2y + Lˆ2z (5.7)
To evaluate the commutators of L̂2 with each component we use the relations
 A2 , B  = A  A, B  +  A, B  A and  An , A = 0
   

Hence,
 Lˆ2 , L  =  Lˆ2 + Lˆ2 + Lˆ2 , L  =  Lˆ2 , L  +  Lˆ2 , L 

 x
  x y z x
  y x   z x 
= Lˆ y  Lˆ y , Lx  +  Lˆ y , Lx  Lˆ y + Lˆz  Lˆz , Lx  +  Lˆz , Lx  Lˆz (5.8)
       
= −ihLˆ y Lˆz − ihLˆz Lˆ y + ihLˆz Lˆ y + ihLˆ y Lˆz = 0
In words, L̂2 commutes with Lˆx . Likewise, we can easily verify that the square
magnitude of the angular momentum commutes with the other two components.
Thus,
 Lˆ2 , Lˆ  =  Lˆ2 , Lˆ  =  Lˆ2 , Lˆ  = 0 (5.9)

 x 
 
y 
 
z

These results show that we can simultaneously determine the square magnitude of
the angular momentum and one of its components. On the other hand, equations
(5.6) show that no two components of commute and, therefore, we cannot in
general specify more than one component at the same time. That means we cannot
usually determine the direction of angular momentum in quantum mechanics! But
we can choose an arbitrary direction and take it as the z-direction of our coordinate
system. Then we find the eigenvalues and common eigenfunctions of the two
observables L̂2 and Lˆz .
The x and y-components of the angular momentum operator L̂ can be combined to

generate two more useful operators:
Lˆ± = Lˆx ± iLˆ y (5.10)
It is easy to show that
Lˆ+ Lˆ− = Lˆ2 − Lˆ2z + hLˆz (5.11)
Lˆ− Lˆ+ = Lˆ2 − Lˆ2z − hLˆz (5.12)
Subtracting we obtain:
 Lˆ , Lˆ  = 2hLˆ (5.13)
 + −  z

We also have
 Lˆ , Lˆ  = −hLˆ (5.14)
 + z  +
 Lˆ , Lˆ  = hLˆ (5.15)
 − z  −
Think:
Verify that L̂+ and L̂− are not hermitian operators, but are hermitian conjugates
of one another, that is,
† †
Lˆ+ = Lˆ− and Lˆ− = Lˆ+
( ) ( )
Verify equations (5.11)-(5.15)
5.3. Representation of Angular Momentum in Spherical Coordinates

The transformation from Cartesian to spherical coordinates is achieved by the
equations
x2 + y2
2 2 2
r = x + y + z , θ = arctan , ϕ = arctan y (5.16)
z x
The inverse transformations are
x = r sin θ cosϕ , y = r sin θ sin ϕ , z = r cosθ (5.17)
Making use of the transformation equations we can obtain the conversion

equations for ∂ ∂x , ∂ ∂y and ∂ ∂z :
∂ ∂ cosθ cos ϕ ∂ sin ϕ ∂

= sin θ cosθ + − (5.18)
∂x ∂r r ∂θ r sin θ ∂ϕ
∂ ∂ cosθ sin ϕ ∂ cos ϕ ∂

= sin θ sin ϕ + + (5.19)
∂y ∂r r ∂θ r sin θ ∂ϕ
∂ ∂ sin θ ∂
= cosθ − (5.20)
∂z ∂r r ∂θ

Now using equations (5.4) and the results above, the components of the angular
momentum become:
 ∂ ∂  ∂ ∂ 
Lˆx = −ih  y − z  = ih  sin ϕ + cot θ cosϕ  (5.21)
 ∂z ∂y   ∂θ ∂ϕ 
 ∂ ∂  ∂ ∂ 
Lˆ y = −ih  z − x  = −ih  cosϕ − cot θ sin ϕ  (5.22)
 ∂x ∂z   ∂θ ∂ϕ 
 ∂ ∂ ∂
Lˆz = −ih  x − y  = −ih (5.23)
 ∂y ∂x  ∂ϕ
Also,
 1 ∂  ∂  1 ∂ 2 
Lˆ2 = −h2   sin θ  + (5.24)

sin θ ∂θ  ∂θ  sin 2 θ ∂ϕ 2 
and
 ∂ ∂ 
Lˆ+ = he+iϕ  + + i cot θ  (5.25)
 ∂θ ∂ϕ 
 ∂ ∂ 
Lˆ− = he−iϕ  − + i cot θ  (5.26)
 ∂θ ∂ϕ 
We can see that all the angular momentum operators involve only the angular
spherical coordinates, θ and ϕ , but not the radial coordinate, r . This reduces the
number of independent variables from three (x, y, z) to two (θ, φ). Furthermore,
equation (5.23) reduces one of the eigenvalue problems to one variable, ϕ .
5.4. Angular Momentum Eigenstates and Eigenvalues

Suppose Φ (ϕ ) is an eigenfunction of the operator Lˆz [equation (5.23) above]. The
eigenvalue of this operator is, naturally, the z-component of the angular
momentum. Hence:
∂Φ
Lˆz Φ = −ih = Lz Φ (5.27)
∂ϕ

Integration gives Φ (ϕ ) = AeiLzϕ h (5.28)
Since the eigenfunction Φ must be single valued, we can write
Φ (ϕ ) = Φ (ϕ + 2π ) (5.29)
The reason is that if we increase the angle ϕ by 2π , we make a complete circle

around the z-axis and return to the same point. Then the eigenfunction (5.28) must
again be the same. The condition (5.29) means
ei 2π Lz h = 1
Or cos 2πhLz + i sin 2πhLz = 1 (5.30)
Equation (5.30) is satisfied only if
Lz = m h, where m = 0, ±1, ± 2, ± 3, L (5.31)
So the angular momentum in an arbitrary direction, labeled as the z-direction,

cannot just take on any value: it must be an integral multiple of Planck's
constant h . The integer m is called the magnetic quantum number.
Comparing Lz with the linear momentum component pz , which can take on any
value within the accuracy permitted by the uncertainty principle, we see that Lz
can only take discrete but precise values.
The eigenfunctions of Lˆz [equation (5.28)] now become:
Φ m (ϕ ) = Aeimϕ (5.32)
The normalization condition fixes the constant A:
2π 2π
* 2
∫ Φ Φ dϕ = A ∫0 e
−imϕ imϕ
e dϕ = 2π A2 = 1
0
1
A= (5.33)
2π
Inserting this into (5.32) we have the properly normalized eigenfunctions of Lˆz :

1 imϕ
Φ m (ϕ ) = e , m = 0, ± 1, ± 2, ± 3, L (5.34)
2π
Think:
1. What is the magnetic quantum number of a macroscopic, 1 kg, particle that is
encircling the z-axis at a distance of 1 m at a speed of 1 m/s? Write out as an
integer, and show digits you are not sure about as a question mark.
2. Based on the derived eigenfunction, equation (5.32), would any macroscopic
particle ever be at a single magnetic quantum number in the first place? In
particular, what can you say about where the particle can be found in an
eigenstate?
Now let’s determine the eigenfunctions and eigenvalues of the operator L̂2 ,
equation (5.24). It is reasonable to assume that the eigenstates of the square of the
angular momentum operator are functions of the angular variables only: Y (θ , ϕ ) .
The eigenvalues are, definitely, L2 , the magnitude squared of L . Therefore,
ˆ2 h 2 ∂  ∂Y (θ ,ϕ )  h 2 ∂ 2Y (θ ,ϕ )
L Y (θ ,ϕ ) = − sin θ −
2 2
= L2Y (θ ,ϕ ) (5.35)
sin θ ∂θ  ∂θ  sin θ ∂ϕ
 
Writing Y (θ , ϕ ) in the separable form
Y (θ ,ϕ ) = Θ (θ ) Φ m (ϕ ) (5.36)
and putting it back into equation (5.33) we obtain:
h 2Φ ∂  ∂Θ  h2Θ ∂ 2Φ 2
−  sin θ − = L ΘΦ (5.37)
sin θ ∂θ  ∂θ  sin 2 θ ∂ϕ 2
Note that the function Φ (ϕ ) is the eigenstate of Lˆz described by equation (5.34).
Using this function in equation (5.37) we come to an eigenvalue problem in the
variable θ. That is,
h2 ∂  ∂Θ  h2m2
 sin θ  − Θ = − L2Θ (5.38)
sin θ ∂θ  2
∂θ  sin θ

Instead of solving this eigenvalue problem directly, we use the properties of the
angular momentum operators discussed in the previous section. Let’s see first the
effect of the operator L+ on Φ m (ϕ ) . Using equations (5.27) and (5.32):
 
   
{ }
Lˆz  Lˆ+Φ m  =   Lˆz , Lˆ+  + Lˆ+ Lˆz  Φ m = hLˆ+ + Lˆ+ Lˆz Φ m

(5.39)
=  hLˆ+ + mhLˆ+  Φ m = ( m + 1) h  Lˆ+Φ m 
   
Hence, L+ raises the z-component of angular momentum by one unit of h .

Similarly, L− lowers the z-component of angular momentum by one unit of h ,
that is,
Lˆz  Lˆ−Φ m  = ( m −1) h  Lˆ−Φ m  (5.40)

   
Let’s now turn to the eigenfunctions Y (θ , ϕ ) of L̂2 . It can be easily verified that
these are also eigenfunctions of Lˆz :
LˆzY = Lˆz ( ΘΦ m ) = ΘLˆz Φ m = mhΘΦ m = mhY (5.41)
Consider then the inner product Lˆ+Y Lˆ+Y :
Lˆ+Y Lˆ+Y = Y Lˆ†+ Lˆ+Y = Y Lˆ− Lˆ+Y (5.42)
where Lˆ†± = Lˆm has been used. [See the “Think” problem on page 78.] Using
equation (5.12) equation (5.42) becomes
Lˆ+Y Lˆ+Y = Y  Lˆ2 − Lˆ2z − hLˆz  Y

 
Applying equation (5.41) to this equation we obtain
Lˆ+Y Lˆ+Y = Y  Lˆ2 − Lˆ2z − hLˆz  Y = L2 Y Y − m2h2 Y Y − mh2 Y Y

 
{
= L2 − m ( m +1) h2 Y Y }
We know that the inner product of any eigenfunction with itself is a positive-
definite real quantity, that is,
Lˆ+Y Lˆ+Y ≥ 0
Therefore, we find that
Lˆ+Y Lˆ+Y = L2 − m ( m +1) h2 Y Y

{ }
= {L − m ( m +1) h } ≥ 0
2 2
Or L2 ≥ m ( m + 1) h2 (5.43)
Equation (5.43) tells us that the magnetic quantum number m has restricted
values. For example, if the magnitude of the angular momentum vector L is
6 2 h , then by the condition (5.43) we have
72 ≥ m ( m + 1)
which gives m ≤ 8 . So, the possible vaules of m in this case are:
m = 0, ± 1, ± 2, ± 3, ± 4, ± 5, ± 6, ± 7, ± 8.
A second restriction on the magnetic quantum number is obtained if we start with

the inequality
Lˆ−Y Lˆ−Y ≥ 0 .
The result is
L2 ≥ m ( m −1) h2 (5.44)
In the restrictive conditions (5.43) and (5.44), the equality holds for the maximum
or minimum values of m . Suppose the maximum value of m is some positive
integer l , that is,
mmax = l

Using the condition given by equation (5.43) we obtain
L2 = l ( l + 1) h2
We thus obtained the magnitude of the total angular momentum:
L = l ( l +1) h (5.45)
If we choose the minimum value of m to be the negative of l , that is, if
mmin = −l
then we have from the restriction (5.44)
L = −l ( −l −1) h = l ( l +1) h
Therefore, inequalities (5.43) and (5.44) are equivalent to the following constraint:
−l ≤ m ≤ l (5.46)
where l is a non-negative integer. We, thus, conclude that the quantum number m
can only take a restricted range of integer values. The integer l is known as the
orbital (azimuthal) quantum number.
Using the two quantum numbers l and m , equations (5.35) and (5.41) now read
Lˆ2Yl,m (θ ,ϕ ) = l ( l + 1) h2Yl,m (θ ,ϕ ) (5.47)
LˆzYl,m (θ ,ϕ ) = mhYl,m (θ ,ϕ ) (5.48)
We can now set the rules for the allowed values of the quantum numbers l and
m:
• l takes the non-negative integer values 0, 1, 2, 3, L .

• Once l is given, the quantum number m can take any integer value in the
range
−l, − l + 1, − l + 2, L, 0, L, l − 2, l − 1, l (5.46*)

Now, do we get the same the result as in equations (5.39) and (5.40) if we use the
raising and lowering operators to act the eigenstates Yl,m (θ , ϕ ) ? Let us see:
 
Lˆz  Lˆ+Yl,m  =   Lˆz , Lˆ+  + Lˆ+ Lˆz  Yl,m =  hLˆ+ + Lˆ+ Lˆz  Yl,m
      
Using equation (5.41) we obtain
Lˆz  Lˆ+Yl,m  =  hLˆ+ + mhLˆ+  Yl,m = ( m + 1) h  Lˆ+Yl,m  (5.49)

     
Similarly,
Lˆz  Lˆ−Yl,m  = ( m −1) h  Lˆ−Yl,m  (5.50)

   
Thus, we conclude, as in equations (5.39) and (5.40), that the effect of the raising
operator L̂+ acting on an eigenstate of Lˆz corresponding to the eigenvalue mh is to
convert this eigenstate to an eigenstate corresponding to the eigenvalue ( m + 1) h .
Likewise, the lowering operator L̂− changes an eigenstate of Lˆz corresponding to
the eigenvalue mh into an eigenstate corresponding to the eigenvalue ( m − 1) h .
So, let α l , m and β l , m be parameters such that
Lˆ+Yl,m = α l,mYl,m+1 (5.51)
Lˆ−Yl,m = βl,mYl,m−1 (5.52)
Now, (5.51) and (5.52):
Lˆ− Lˆ+Yl,m = Lˆ− α l,mYl,m+1 = α l,m Lˆ−Yl,m+1 = α l,m βl,m+1Yl,m

( ) (5.53)
Also, using equations (5.12), (5.47) and (5.48),
Lˆ− Lˆ+Yl,m =  Lˆ2 − Lˆ2z − hLˆz  Yl,m = l ( l + 1) h2 − m2h2 − mh2 Yl,m

 
( )
Lˆ− Lˆ+Yl,m = l ( l + 1) − m ( m +1) h2Yl,m
( ) (5.54)

Comparing equations (5.53) and (5.54), we have
(
α l,m βl,m+1 = l ( l + 1) − m ( m + 1) h2 ) (5.55)
In the same way, we can show that
)
Lˆ+ Lˆ−Yl,m = βl,mα l,m−1Yl,m = l ( l +1) − m ( m −1) h2Yl,m
(
Or (
βl,mα l,m−1 = l ( l + 1) − m ( m −1) h2 ) (5.56)
It is easy to verufy that equations (5.55) and (5.56) are satisfied if
α l,m = l ( l + 1) − m ( m + 1) h (5.57)
β l ,m = l ( l + 1) − m ( m − 1) h (5.58)
Hence, we can now write equations (5.51) and (5.52) as
Lˆ+Yl,m = l ( l +1) − m ( m +1) hYl,m+1 (5.59)
Lˆ−Yl,m = l ( l + 1) − m ( m −1) hYl,m−1 (5.60)
Think:
1. Prove that the condition given by (5.44) is correct.
2. Verify eqautions (5.50) and (5.56).
3. Check the validity of equations (5.57) and (5.58) by substituting into equations
(5.55) and (5.56), respectively.
4. Suppose m = ±l in equations (5.59) and (5.60). What would be the result?
Discuss.
5. using equations (5.25) and (5.26) write out the differential equations for
L̂+Yl,l and L̂−Yl, −l

5.5. Spherical Harmonics

We have seen that the functions Yl ,m (θ , ϕ ) are simultaneous eigenstates of L̂2 and
Lˆz . These functions are known as spherical harmonics and we now investigate
their functional form. We begin by choosing the eigenstate corresponding to the
lowest possible value of m (i.e., m = −l ) and act on this eigenstate with the
lowering operator:
Lˆz Lˆ−Yl,−l = Lˆ− LˆzYl,−l +  Lˆz , Lˆ−  Yl,−l

 
=  Lˆ− Lˆz − hLˆ−  Yl,−l
 
= −l −1 h  Lˆ−Yl,−l 
( )  
The effect is clearly to convert the eigenstate into that of a state with a lower value
of m , i.e., m = −l − 1 . However, no such state exists because of the constraint
(5.46) or (5.46*) . We must, therefore, have
Lˆ−Yl,−l = 0 for m = −l ( minimum value ) (5.61)
Similarly,
Lˆ+Yl,+ l = 0 for m = +l ( maximum value ) (5.62)
These results could be directly obtained from equations (5.59) and (5.60).
We continue our investigation by picking one of the equations above, say equation
(5.62) which tells us that there is no state for which m has a larger value than + l .
The eigenstate corresponding to m = + l is, using equation (5.34),
1
Yl,l (θ ,ϕ ) = Θl,l (θ ) Φl (ϕ ) = Θl,l (θ ) ei lϕ (5.63)
2π
Inserting this into equation (5.62) and making use of equation (5.25) we obtain
 ∂ ∂ 
heiϕ   Θl,l (θ ) e
i lϕ
+ i cot θ =0 (5.64)
 ∂θ ∂ϕ 

Carrying out the derivative we obtain the differential equation
d Θl,l
− l cot θ Θl,l = 0 (5.65)
dθ
Integrating (5.65) gives
l
Θl,l = Cl,l ( sin θ )
where Cl,l is a constant of integration. Equation (5.63) now becomes:
Cl,l l i lϕ
Yl,l (θ ,ϕ ) =
2π
(sinθ ) e (5.66)
Think:
In a similar way demonstrate that
Cl,l l
Yl,−l (θ ,ϕ ) =
2π
(sinθ ) e − i lϕ (5.67)
Normalization requires that
2
∫ Yl,l (θ ,ϕ ) d Ω = 1 where d Ω = sin θ dθ dϕ (5.68)
Now,
π 2π
1 2 2l
Cl,l ∫ ∫ ( sin θ ) sin θ dθ dϕ = 1
2π θ = 0 ϕ =0
2 π l
∫ (1 − cos θ ) sin θ dθ = 1
2
Cl,l
θ=0
Let u = cosθ . Then
2 +1 l
∫ (1 − u ) du = 1
2
Cl,l (5.69)
−1

The integral in equation (5.69) cab be easily evaluated:
2
+1 l 2 ( ) 2l l!
∫
−1
(1 − u )2
du =
2l + 1 ( 2l )!
(5.70)
Inseting this into equation (5.69) yields an expression for the constant Cl,l :
2l + 1 1
Cl,l =
2 ( 2l )! l
2 l!
(5.71)
The normalized spherical harmonics for m = + l are, therefore,
2l + 1 1 l
Yl,l (θ ,ϕ ) = ( 2l )! l ( sin θ ) ei lϕ (5.72)
4π 2 l!
Having determined the form of Yl,l , we can obtain the lower state eigenfunction
Yl ,l −1 by using equations (5.26) and (5.60) for m = l , that is, we use the lowering
operator on Yl,l :
 ∂ ∂ 
Lˆ−Yl,l = he−iϕ  − + i cot θ  Y = l ( l + 1) − l ( l − 1) hYl,l −1
 ∂θ ∂ϕ  l,l
This yields
1 −iϕ  ∂ ∂ 
Yl,l−1 = e  − + i cot θ  Yl,l (5.73)
2l  ∂θ ∂ϕ 
Using (5.72) into (5.73), we obtain
Yl,l−1 =
( −1) Cl,l i l −1 ϕ  d 
e ( )  + l cot θ  ( sin θ )
l
(5.74)
2π 2l  dθ 
Now, for any function of θ, we can easily show that
 d  1 
d
(sin θ ) f (θ )
l
 + l cot θ  f (θ ) = l  (5.75)
 dθ 
  (sin θ ) dθ 

Hence, the eigenstate corresponding to m = l − 1 becomes
i l −1)ϕ  
( −1) Cl,l e (  1  1 d  2l 
Yl,l−1 =  
l −1 


(sinθ )  (5.76)
2π 2l 

(sinθ )  sin θ dθ  

By operating on Yl ,l −1 with the lowering operator (5.26) and using equation (5.60)
with m = l − 1 , we get:
 ∂ ∂ 
Lˆ−Yl,l−1 = he−iϕ  − + i cot θ Y = l ( l + 1) − ( l −1)( l − 2 ) hYl,l−2
 ∂θ ∂ϕ  l,l−1
Substituting equation (5.76) and using (5.75) we obtain the eigenstate Yl,l − 2 for
which m = l − 2 :
2 i l − 2)ϕ  
( −1) Cl,l e (  1  1 d 
2
2l 
Yl,l−2 =  l − 2  sin θ dθ
 (sinθ )  (5.77)
2π 2!( 2l )( 2l − 1) ( sin θ )    
 
Proceeding in the same way, we can show that the eigenstate with m = l − 3 is
 
( −1) Cl,l e ( )
3 i l −3 ϕ 3
 1  1 d  2l 
l −3  sin θ dθ  (
Yl,l−3 =  sin θ )  5.78)
2π 3!( 2l )( 2l − 1)( 2l − 2 )  ( sin θ )   
 
Inspection of equations (5.77) and (5.78) reveals a general functional form of the
spherical harmonics. So, for m = l − k , where 0 ≤ k ≤ 2l , we can write:
 
( −1) Cl,l e ( )
k i l −k ϕ k
 1  1 d  2l 
Yl,l−k =  
l−k  
 ( sin θ ) 
2π k !( 2l )( 2l − 1)L( 2l − k + 1)  ( sin θ )  sin θ dθ  
 
Or, in terms of the magnetic quantum number m :

l −m  
( −1) Cl,l eimϕ  1  1 d 
l −m
2l 
Yl,m =  m   (sin θ ) 
2π ( l − m )!( 2l )( 2l − 1)L( l + m + 1)  ( sin θ )  sin θ dθ  


Now,
2l ( 2l −1)( 2l − 2 )L( l + m + 1)( l + m )!

2l ( 2l −1)( 2l − 2 )L( l + m + 1) =
( l + m )!
=
( 2l )!
( l + m )!
Therefore,
Cl,l ( l + m )! eimϕ  1  −1 d 
l−m
2l


Yl,m =  m   (sin θ ) 
2π ( l − m )!( 2l )!  ( sin θ )  sin θ dθ  
 
Or, substituting the expression for Cl,l , i.e., equation (5.71), we have
 
( 2l + 1)( l + m)! eimϕ  1  −1 d 
l−m
2l 
Yl,m =
4π ( l − m )!
  
2l l!  ( sin θ )m  sin θ dθ 
(sinθ )  (5.79)

 
The first few spherical harmonics obtained from equation (5.79) are listed below:
l m Yl,m l m Yl,m
1 7
0 0 Y0,0 =
4π
3 0 Y3,0 =
16π
( 5cos3 θ − 3cos θ )
3 21
1 0 Y1,0 = cos θ ±1 Y3,±1 = m sin θ ( 5cos 2 θ − 1) e± iϕ
4π 64π
±1 Y1, ±1 = m
3
sin θ e± iϕ ±2 Y3,±2 = 105 sin 2 θ cos θ e ±2iϕ
8π 32π
5 35
2 0 Y2,0 = ( 3cos 2 θ − 1) ±3 Y3, ±3 = m sin 3 θ e±3iϕ
16π 64π
±1 Y2,±1 = m 15 sin θ cos θ e ± iϕ

8π
15
±2 Y2,±2 = sin 2 θ e ±2iϕ
32π

The spherical harmonics are conventionally written as
Yl,m = ( −1)
m ( 2l + 1)(l − m )! eimϕ P cosθ ,
l,m ( ) for m ≥ 0 (5.80)
4π ( l + m )!
where the Pl ,m are the associated Legendre polynomials given by
−m 2
(
l + m ( l + m )! 1 − u
2
)  d 
l−m l
Pl,m ( u ) = ( −1)
(l − m)! 2l l!
 
 du 
(
1− u 2 ) (5.81)
for m ≥ 0 . Alternatively,
m2
(
l 1− u 2 )  d 
l+m l
Pl,m (u ) = ( −1) 2l l!
 
 du 
(1− u ) 2
(5.82)
for m ≥ 0 . The spherical harmonics for m < 0 can be easily obtained from the
relation
m
Yl,−m = ( −1) Yl*,m (5.83)
We have seen that these spherical harmonics are simultaneous eigenstates of Lˆz
(eigenvalues mh ) and L̂2 (with eigenvalues l ( l + 1) h 2 ). The spherical harmonics
given by equatios (5.79) or (5.80) are orthonormal:
*
∫ Yl′,m′Yl,md Ω = δ ll′ δ mm′ (5.84)
They also form a complete set. That is, any function of θ and φ can be represented
as a superposition of spherical harmonics. Hence,
∞ l
ψ (θ ,ϕ ) = ∑ ∑ cl,mYl,m (θ ,ϕ ) (5.85)
l=0 m=− l

5.6. Spin Angular Momentum

The general motion of an extended classical system is a combination of two types
of motin: motion of the center of mass of the system due to a net external force or
torque and motion of the system relative to its own center of mass. The system,
therefore, has two types of angular momentum – external and internal. A good
example of such a sytem is the Earth. The external angular momentum of the Earth
is its orbital angular momentum due to its revolution around the Sun. The internal
angular momentum of the Earth is its spin angular momentum due to its rotation
(spinning) about an axis passing through its center of mass. Quantum particles
such as electrons do possess both orbital and spin angular momentum. The orbital
angular momentum of an elecrtron in an atom, for instance, is due to its motion
around the nuclus of the atom. However, there is no perfect classical analogy to
the electron’s spin angular momentum since electrons are considered as
structureless point particles. In quantum mechanics, the spin angular momentum is
thought of as an intrinsic property possessed by elementary particles just as they
possess a characteristic charge and mass.
Spin Operators
Spin angular momentum, just like orbital angular momentum, obeys the general
commutation relation
Sˆ × Sˆ = ihSˆ (5.86)
If we define three operaqtors S x , S y , S z such that S ≡ ( S x , S y , S z ) , then equation

(5.86) is equivalent to equqtions (5.6). Meaning,
 Sˆ , Sˆ  = ihSˆ  Sˆ , Sˆ  = ihSˆ  Sˆ , Sˆ  = ihSˆ (5.87)

 x y  z  y z  x  z x  y
We can also define the magnitude squared of the spin angular momentum vector
by the operator
Sˆ 2 = Sˆx2 + Sˆ y2 + Sˆz2 (5.88)
Following the algebra of orbital angular momentum in section 5.2, we can show
that
 Sˆ 2 , Sˆ  =  Sˆ 2 , Sˆ  =  Sˆ 2 , Sˆ  = 0 (5.89)
 x 
 
y 
 
z


That is, we can simultaneously measure the magnitude squared of the spin angular
momentum and, at most, one cartesian component which is, by convention, the z-
component ( Sˆz ).
Analogous to equations (5.10), we can define raising and lowering operators for
spin angular momentum:
S± = S x ± iS y (5.90)
It can be easily demonstrated that the raising and lowering operators defind by
equation (5.90) satisfy the relations:
†
 Sˆ  = Sˆm (5.91)
 ±
 
Sˆ+ Sˆ− = Sˆ 2 − Sˆz2 + hSˆz (5.92)
Sˆ− Sˆ+ = Sˆ 2 − Sˆz2 − hSˆz (5.93)
 Sˆ , Sˆ  = m hS (5.94)
 ± z  ±
Eigenvectors of the Spin Operators
The spin angular momentum of a quantum particle has nothing to do with motion
in real space; it is just a fundamental property of the quantum particle. Therefore,
unlike the orbital angular momentum, the spin angular momentum cannot be
expressed in terms of space coordinates x, y, z or r , θ , ϕ . This means, the spin
operators cannot be represented by differential operators as was done in section
5.3 for the case of orbital angular momentum operators. Now the question is: what
could the wavefunctions upon which the spin operators act possibly be? Definitely
we cannot have spin wavefunctions that depend on real space coordinates like the
spherical harmonics. It may be inappropriate to take them as regular functions in
the first place. We can think of spin eigenvectors in an abstract vector space, say
the spin space, on which the spin operators act. The eigenvectors in this space
correspond to the internal state of the particle under investigation.
Suppose a quantum particle has N linearly independent internal states which are
characterized by linearly independent spin vectors χ n . These N independent spin
vectors determine the dimensionality of the spin space. Now, it is a fundamental
assumption of quantum mechanics that the internal states of a particle are

superposable. Hence, a general spin state χ is obtained by a linear supeposition of

the independent spin vectors, that is,
N
χ = ∑ cn χ n (5.95)
n =1
where the cn are complex numbers.
The norm of a spin vector (a spinor) is a real positive-definite number given by
2
χ = χ †χ ≥ 0 (5.96)
Two spin states χ a and χ b are said to be mutually orthogonal if
χ a† χb = χb† χ a = 0 (5.97)
Since the independent spin vectors discussed above span the spin space, they must
be orthonormal, that is,
χ n† χ m = δ nm (5.98)
These are the eigenvectors or eigenstates of the spin operators we are looking for!
If Ô represents a general spin operator, then
Oˆ χ n = α n χ n (5.99)
where α n is the eigenvalue of Ô . A measurement of O will then yield the result

2
α n with probability cn according to equation (5.95).
Eigenstates of Sˆz and Ŝ 2
The commuting spin operators Sˆz and Ŝ 2 have simultaneous eigenstates. By

analogy with equations (5.47) and (5.48) we can write:
Sˆz χ s,ms = ms h χ s,ms (5.100)
Sˆ 2 χ s,ms = s ( s + 1) h2 χ s,ms (5.101)

The eigenstates of Sˆz and Ŝ 2 are assumed to be orthonormal:
χ †s,ms χ s′,ms′ = δ ss′δ msms′ (5.102)
We can also verify that
Sˆ± χ s,ms = s ( s + 1) − ms ( ms ±1) hχ s ,ms ±1 (5.103)
Now, using equations (5.96) and (5.102) we obtain two inequalities
s ( s + 1) − ms ( ms ± 1) ≥ 0 (5.104)
If s ≥ 0 , these two inequalities imply that
−s < ms < s (5.105)
So, for some given value of s , the possible values of ms are restricted within the
range between the minimum value − s and the maximum value + s . Therefore,
spin states like χ s , s +1 or χ s ,− s −1 do not exist. Using the raising and lowering
operators we then have
Sˆ+ χ s,s = Sˆ− χ s,− s = 0 (5.106)
Given the eigenstate χ s ,− s , how many times should the raising operator be
employed to arrive at the eigenstate χ s , s ? Let’s see:
2 3
Sˆ+ χ s,− s → χ s ,− s +1 ,  Sˆ 
 + χ s ,− s → χ s ,− s + 2 ,  Sˆ 
 + χ s , − s → χ s , − s +3
   
4 k
 Sˆ  χ s ,− s → χ s ,− s + 4 , ------------  Sˆ  χ s ,− s → χ s ,− s + k = χ s , s
 +  +
   
We see from the pattern above that, after employing Ŝ+ k times, − s + k = s . Or
k = 2s (5.107)
where k is a positive integer. Hence, the values of the spin quantum number s are
s = k = 0, 1 , 1, 3 , 2, 5 , 3, K (5.108)
2 2 2 2

Unlike the orbital quantum number l , the spin quantum number s can take
positive half-integer values in addition to positive integer values. In our discussion
of orbital angular momentum, the quantum number m , which is analogous to ms ,
is restricted to take integer values due to the requirement that the wavefunctions be
singlevaued [see equation (5.29)]. Consequently, the quantum number l , which is
analogous to s , is also restricted to take integer values. We have seen that these
restrictions are end results of the representation of the orbital angular momentum
operators as differential operators in real space. There is no such representation
of the corresponding spin angular momentum operators. Hence, there is no reason
to restrict the quantum number s to integer values only.
By using equations (5.108) and (5.105) in equations (5.100) and (5.101), we can
determine the eigenvalues of the spin operators Sˆz and Ŝ 2 as shown below:
eigenvalues of Ŝ 2 eigenvalues of Sˆz

s=0 0 0
s = 12 34 h2 − 12 h , 12 h
s =1 2h 2 −h , 0, h
s = 32 15 h 2
4
− 32 h , − 12 h , 12 h , 32 h
M M M
In quantum mechanics, all fermions (electrons, protons, neutrinos, etc) possess

half-integer spin whereas all bosons (pi mesons, photons, gravitons, etc) possess
integer spin.
Spin 1 2 and Pauli Matrices
We have discussed that there are no coordinates θ and ϕ associated with internal
angular momentum. So the only thing we can do now to develop a spinor
representation for spin 1 2 .
The total spin for electrons is s = 1 2 . This means, there are only two possible spin
states for an electron. The usual basis states are the eigenstates of sz : ± 12 h .
Let us denote these two independent spin eigenstates of an electron as
χ + = χ 1 , 1 = 12 , 12 Spin up (5.109)
2 2

χ − = χ 1 ,− 1 = 12 , − 12 Spin down (5.110)

2 2
It then follows that [see equations (5.100) and (5.101)]
Sˆz χ ± = ± 12 hχ ± (5.111)
Ŝ 2 χ ± = 34 h2 χ ± (5.112)
These two eigenstates satisfy the orthonormality conditions
χ +† χ + = χ −† χ − = 1 (5.113)
and χ +† χ − = χ −† χ + = 0 (5.114)
We simply represent the + 12 h eigenstate as the first component of a 2D vector and

the − 12 h eigenstate as the second component. So the basis eigenstates are:
1  0
χ + =  

and χ − =   (5.115)
0
  1 
As you might expect, these forms satisfy the orthonormality conditions. An

arbitrary spin-½ state can then be represented as a linear combination of these two:
a
χ =   = a χ + + bχ − (5.116)
b 
2 2
with the normalization condition that a + b = 1 .
Now we derive the matrix operators for spin. A general spin operator which
operates on a spinor like the one given by equation (5.116) above must be
represented as a 2 × 2 matrix. Thus, using equations (5.111) and (5.115) for spin-
½ particles we can write
 c11 c12  1  1  1 
 = h for spin up
c
 21
c22 
  0  2  0 
    
 c11 c12   0  1  0
    = − h  for spin down
c
 21
c22   1  2 1

Solving for the constants we obtain
c11 = 12 h, c12 = 0, c21 = 0, c22 = − 12 h
Hence, the matrix representation of the spin operator Sˆz is
h 2 0  1 1 0 
Sˆz =   = 2 h   (5.117)
 0 − h 2  0 −1
Similarly, using equations (5.112) and (5.115) we obtain
1 0
Sˆ 2 = 34 h2   (5.118)
0 1
Now, using equation (5.103) for spin-½ particles yields
Ŝ± χ m = hχ ±
From this we obtain the matrix representation
0 1
Sˆ+ = h   (5.119)
 0 0 
 0 0
Sˆ− = h   (5.120)
1 0
Using the definition (5.90) we obtain
S +S 0 1
Sˆx = + − = 2h   (5.121)
2 1 0
S −S  0 −i 
Sˆ y = + − = h2   (5.122)
2i i 0 
We can see from equations (5.117), (5.121) and (5.122) that the spin operator Ŝ
can be conveniebtly expressed in vector form by
Sˆ = h2 σˆ

where the components of σ

σ̂ are the 2 × 2 matrices generally known as the Pauli
spin matrices:
0 1 0 −i  1 0
σˆ x =   σˆ y =   σˆ z =  
1 0 i 0   0 −1
These are Hermitian, Traceless matrices.
Think:
1. Show that the Pauli matrices satisfy the usual commutation relations from
which we derived the properties of angular momentum operators.
2. Find the Pauli representations of Sˆx , Sˆ y , and Sˆz for a spin-1 particle.
3. Find the Pauli representations of the normalized eigenstates of Sˆx and Sˆ y for a
spin-½ particle.
4. Suppose that a spin-½ particle has a spin vector which lies in the xz plane,
making an angle θ with the z axis. Demonstrate that a measurement of S z
yields h 2 with probability cos 2 (θ 2 ) , and −h 2 with probability sin 2 (θ 2 ) .
5. An electron is in the spin-state
1 − 2i 
χ = A   Pauli representation
 2 
Determine the constant A by normalizing χ . If a measurement of Sˆz is made,
what values will be obtained, and with what probabilities? What is the
expectation value of Sˆz ? Repeat the above calculations for Sˆx and Sˆ y .
6. Consider a spin-½ system represented by the normalized spinor
 cosα 
χ =  

 e sin α
iβ

in the Pauli representation, where α and β are real. What is the probability
that a measurement of Sˆ y yields −h 2 ?

Chapter 6. The Hydrogen Atom
This Chapter examines a critically important quantum system – the hydrogen

atom. The hydrogen atom consists of a single proton (the nucleus) and a single
electron encircling the proton. We choose the spherical coordinate system for the
quantum analysis of the hydrogen atom and assume the proton, which is much
heavier than the electron, to be at rest at the origin of the spherical coordinates.
The quantum tereatment determines the energy levels of the electron leading to the
spectrum of hydrogen: the radiations (photons) that the atom can absorb or emit.
The result of this treatment is also very important for understanding the properties
of the other elements and of chemical bonds.
6.1. The Hamiltonan

The electrostatic interaction between the electron and proton of the hydrogen atom
gives rise to potential energy
1 e2
V (r ) = − (6.1)
4πε 0 r
Where r is the distance from the nuclus, e = 1.6 × 10−19 C is the electronic charge
and ε 0 = 8.85 ×10−12 C 2 Jm is the permittivity of space.
Adding to this the kinetic energy operator, we get the Hamiltonian of the hydrogen
atom in spherical coordinates:
h2 2
Ĥ = − ∇ +V (r )
2me
h 2  ∂  2 ∂  1 ∂  ∂  1 ∂ 2  1 e2
=−  r  +  sin θ + −
2me r 2  ∂r  ∂r  sin θ ∂θ  ∂θ  sin 2 θ ∂ϕ 2  4πε 0 r
where me = 9.109 × 10−31 is the mass of the electron.
Note: The small proton motion can be corrected for by using the reduced mass
me m p
µ= = 9.1044 × 10−31 kg . This makes the solution exact, except for extremely
me + m p
small effects due to relativity and spin.

6.2. Solution of Energy Eigenvalue Equation

The method to find the energy eigenfunctions and eigenvalues is the same as the
one for the harmonic oscillator. We assume the eigenfunctions ψ ( r ,θ , ϕ ) are
separable in the form of a product of functions of each of the three coordinates,
that is
ψ ( r ,θ ,ϕ ) = R ( r ) Y (θ ,ϕ ) = R ( r ) Θ (θ ) Φ (ϕ ) (6.2)
Substituting this expression into the Hamiltonian eigenvalue problem Ĥψ = Eψ ,

we get:
h 2  d  2 dR  R ∂  ∂Y  R ∂ 2Y  1 e2
− Y  r  + sin θ +  − RY = ERY
2me r 2  dr  dr  sin θ ∂θ  ∂θ  sin 2 θ ∂ϕ 2  4πε 0 r
Multiplying this equation by 2me r 2 RY throughout and then then splitting the
terms yields:
h 2 d  2 dR  1  h 2 ∂  ∂Y  h 2 ∂ 2Y  2me r 2 e2
−  r  + −  sin θ  − 2 − = 2me r 2 E
R dr  dr  Y  sin θ ∂θ  ∂θ  sin θ ∂ϕ  4πε 0 r
2
We have seen in Chapter 5 that the expression in the square brackets is the square
of the angular momentum operator acting on the spherical harmonics Y, equation
(5.35), which is reproduced here:
h2 ∂  ∂Y  h 2 ∂ 2Y
Lˆ2Y = −  sin θ  − = l ( l + 1) h 2Y
sin θ ∂θ  2
∂θ  sin θ ∂ϕ 2
Using this equation we have:
2 2
h 2 d  2 dR  2 2me r e
−  r  + l ( l + 1) h − = 2me r 2 E (6.3)
R dr  dr  4πε 0 r
This can be further simplified by defining a function u ( r ) so that
u ( r ) = rR ( r ) (6.4)
It can be easily verified that

d  2 dR  d 2u (6.5)
 r  = r
dr  dr  dr 2
Using equations (6.4) and (6.5) in equation (6.3) and dividing by 2me r 2 we obtain:
 
h 2 d 2u  h 2 l ( l + 1) 1 e2  (6.6)
− + − u = Eu
2me dr 2  2me r 2 4πε 0 r 
  
This equation is known as the radial equation and the expression in square
brackets is known as the effective potential energy.
1 e2 h2 l ( l +1)
Veff = − +
4πε 0 r 2me r 2
Now let’s examine the asymptotic behaviour of u ( r ) . There are two cases: r → ∞
and r → 0 .
As r → ∞ , the radial equation reduces to
d 2u 2me E
2
= − 2
u = κ 2u
dr h
where κ = −2me E h as usual. The general solution to this equation is
u ( r ) = Ae−κ r + Be+κ r
The second term is not physically acceptable for it makes the function blow up as
r → ∞ . This implies that B = 0 . So,
u ( r ) e−κ r for large values of r (6.7)
As r → 0 , the radial equation takes the form
d 2u l ( l +1)
= u
dr 2 r2
Here we assume a solution of the form u = Cr s where C is a constant. Putting this
into the equation above yields:

s ( s −1) = l ( l +1)
Solving for s , we obtain two possible values: s = −l and s = l + 1 . The general

solution as r → 0 is thus,
u ( r ) = Cr l+1 + Dr −l
Obviously, we must chose D = 0 to keep the function from blowing up as r → 0 .

Therefore,
u ( r ) r l+1 for r 1 (6.8)
Now a solution to the radial equation which encompasses the asymptotic behaviors
given by equations (6.7) and (6.8) can be written as
u ( r ) = r l+1e−κ r v ( r ) (6.9)
where v ( r ) is a function yet to be determined. Carefully carrying out the

derivative of equation (6.9) upto the second order, we get
   
d 2u  d 2v dv  l ( l +1)
2
=  r 2
+ 2 ( l + 1 − κ r ) + − 2κ ( l + 1) + κ 2r  v  r le−κ r
dr  dr dr r
    
 
Putting this into equation (6.6) and noting that κ = −2me E h , the radial equation
becomes
d 2v dv  2mee2 
r
dr 2
+ 2 ( l + 1 − κ r ) dr  4πε h2 ( ) v = 0
+ − 2κ l + 1 (6.10)
 0 
It is more convenient if we work with the dimensionless variable
ρ =κr (6.11)
dv d ρ dv d 2v 2
2 d v
Then, = =κ and = κ
dr dr dρ dr 2 dρ2
Substituting these into equation (6.10) and using equation (6.11) we obtain

2
ρ d v2 + 2 ( l +1 − ρ ) dv + α − 2 ( l +1)  v = 0 (6.12)
dρ dρ  
2
where α = 2mee2 (6.13)
4πε 0h κ
We now assume a power-series solution to equation (6.12) and then determine the
coefficients in the power series, that is,
∞
v ( ρ ) = ∑ bj ρ j (6.14)
j =0
dv ∞
Differentiating: = ∑ jb j ρ j −1 (6.15)
d ρ j =1
d 2v ∞
2
= ∑ j ( j −1) b j ρ j −2 (6.16)
dρ j =2
using equations (6.14) – (6.16) into equation (6.12):

∞ ∞ ∞
∑
j =2 j =1   j =0
( )
j ( j −1) b j ρ j −1 + ∑ 2  l +1 jb j ρ j −1 + ∑ α − 2l − 2 − 2 j b j ρ j = 0
Adjusting the index j this becomes
∞
   j
∑
j =0 

( j + 1)( j + 2l + 2 ) b j +1 −  2 ( l + 1 + j ) − α  b j ρ = 0
  
(6.17)
This equation is valid for all values of ρ if the coefficient vanishes, that is,
( j +1)( j + 2l + 2) b j +1 − 2 ( l +1+ j ) − α  b j = 0

This is a recursion relation between the coefficients:
2 ( l +1 + j ) − α
b j +1 = bj (6.18)
( j +1)( j + 2l + 2)

For large j , the recursion relation takes the form
2
b j +1 = b
j +1 j
From this we can easily obtain
2j
bj = b for large j
j! 0
Using this in equation (6.14) we have, for large j ,
v ( ρ ) = b0 ∑
∞ ( 2ρ ) = b0e2 ρ
j =0 j!
This shows that the power series blows up as ρ → ∞ which is undesirable.

Fortunately, the way out is simple: cut off the power series at some j so that
b j +1 = 0 . Now the recursion relation gives
2 ( l +1+ j ) −α = 0
Or 2n = α (6.19)
where n = j + l + 1 = 1, 2,3,L
It turns out that the solutions of the radial problem can be numbered using a third
quantum number, n , called the principal quantum number. The principal quantum
number is larger than the azimuthal quantum number l , which in turn must be at
least as large as the absolute value of the magnetic quantum number m :
n>l≥ m
In terms of the quantum numbers n and l , the recursion relation (6.18) becomes
2 ( j + l +1− n )
b j +1 = bj (6.18*)
( j +1)( j + 2l + 2)
Now combining equations (6.13) and (6.19)

2
κ = mee 2 1 = 1 (6.20)
4πε 0h n a0n
where the length a0 is called the Bohr radius and has the value
4πε 0h2
a0 = 2
= 0.529 ×10−10 m (6.21)
mee
Also, we can easily obtain from equations (6.11) and (6.20) the relation
ρ= r (6.22)
na0
Inserting κ = −2me E h in equation (6.20) we get the energy eigenvalues:
2
m  e2  1 h2 1
En = − e2  = − (6.23)
2h  4πε 0  n2 2me a02 n2
This is called the Bohr formula.
The eigenfunctions for the radial equation of the hydrogen atom can now be
obtained by combining equations (6.4), (6.9), (6.11) and (6.14):
n −l −1
Rnl ( ρ ) = ρ l e− ρ vnl ( ρ ) = ρ le− ρ ∑
j =0
bj ρ j (6.24)
For n = 0, l = 0 equation (6.24) gives
R1,0 ( ρ ) = b0e− ρ
Or in terms of r , using equation (6.22),
R1,0 ( r ) = Ce−r a0
where the costant C is determined from the normalization condition:

∞ 2 ∞
∫0 R1,0 r 2dr = C 2 ∫ e−2r a0 r 2dr = 1

0

This gives C = 2 a03 and, therefore,
2 −r a0
R1,0 ( r ) = e (6.25)
a03
1
Let’s try now R2,0 : R2,0 ( ρ ) = e− ρ ∑ b j ρ j = e− ρ ( b0 + b1ρ )
j =0
From the recursion relation, eqution (6.18*), we have b1 = −b0 . Thus,
R2,0 ( ρ ) = b0e− ρ (1 − ρ )
Or, in terms of r , using equation (6.22),
 r 
R2,0 ( r ) = Ce−r 2a0 1 − 
 2a0 
Applying the normalization condition determines the constant C:

∞ 2
C 2
0
−r a
∫ e 0 (1− r 2a0 ) r dr = C 2a0 = 1
2 2 3
( )
1  r  −r 2a0
So, R2,0 ( r ) = 1 −  e (6.26)
3
2a0  2a0 
All the other radial functions can be determined in a similar way. Generally,
however, the radial wave functions Rnl are obtained from the following generic
function:
3 l
Rnl ( r ) =
 2 
 
( n − l −1)! e−r na 
0  2r 
 
L2nl−+l1−1 
2r 
(6.27)
 na0  3  na0   na0 
  2n ( n + l )!    
 
The functions L2nl−+l1−1 are the associated Laguerre polynomials.
There are contradictory definitions of the associated Laguerre polynomials in

different references. Here, the most common definitions are presented. The
associated Laguerre polynomials are given by:

qdq
Lqp ( z ) = ( −1) dz q Lp+q ( z ) (6.28)
where the Laguerre polynomials are
d p −z p
Lp ( z ) = e z
e z (6.29)
dz p
p and q in equations (6.28) and (6.29) are integers.
The first few Laguerre pliynomials obtained from equation (6.29) are listed below:
L0 = 1
L1 = − z + 1
L2 = z 2 − 4 z + 2
L3 = − z 3 + 9 z 2 − 18 z + 6
L4 = z 4 − 16 z 3 + 72 z 2 − 96 z + 24
L5 = − z 5 + 25 z 4 − 200 z 3 + 600 z 2 − 600 z + 120
The first few associated Laguerre polynomials obtained from equation (6.28) are:
L00 = 1 L10 = 1 L20 = 2 L30 = 6

L10 = − z + 1 L11 = −2 z + 4 L12 = −6 z + 18 L31 = −24 z + 96
L02 = z 2 − 4 z + 2 L12 = 3 z 2 − 18 z + 18 L22 = 12 z 2 − 96 z + 144 L32 = 60 z 2 − 600 z + 1200
The first few radial wave functions for hydrogen are.
R10 = 2a0−3 2 e − r a0
1 −3 2  r  − r 2 a0 1 −3 2 r − r 2 a0
R20 = a0 1 − e R21 = a0 e
2  2a0  24 a0
2a0−3 2  2r 2r 2  − r 3a0 8a0−3 2  r   r  − r 3a0

R30 =  1 − + 2 
e R31 = 1 −  e
27  3a0 27a0  27 6  6a0   a 
4a0−3 2 r 2 − r 3a0
R32 = 2
e
81 30 a0

In terms of the three quantum numbers, the final energy eigenfunctions of the
hydrogen atom are:
ψ nlm ( r ,θ ,ϕ ) = Rnl ( r ) Ylm (θ ,ϕ ) (6.30)
where Ylm are the spherical harmonics discussed in Chapter 5. These energy
eigenfunctions of hydrogen are, of course, orthonormal:
* 2
∫ψ nlmψ n'l'm' r sin θ drdθ dϕ = δ nn'δ ll 'δ mm '
Think:
1. Use the tables for the radial wave functions and the spherical harmonics to
write down the wave function
ψ nlm ( r ,θ ,ϕ ) = Rnl ( r ) Ylm (θ ,ϕ )
for the case of the ground state ψ 100 and check that this state is normalized.
2. An electron in the Coulomb field of a proton is in the state described by the
wave function
1
Ψ=
6(4ψ 100 + 3ψ 211 −ψ 210 + 10ψ 21−1 )
Find the expectation value of the Energy, L2 and Lz .
6.3. The Hydrogen Spectrum

The only energy values an electron in the hygrogen atom can have are those
obtaind in the previous section:
h2 1
En = − , n = 1, 2, 3, L (6.31)
2me a02 n2
Evidently, the energy depends only on the principal quantum number n , and not
also on the orbital (azimuthal) quantum number l and the magnetic quantum
number m . Since the lowest possible value of the principal quantum number is
one, the lowest possible energy is

E1 = −13.6eV (6.32)
This corresponds to the ground state of the hydrogen atom given by the
eigenfunction ψ 100 . The ground state is the state in which the hydrogen atom will
be at absolute zero temperature. Even at room temperature, it will still be in the
ground state since the thermal energy at room temperature is unlikely to raise the
energy level of the electron to the next higher one, E2 .
We see from equation (6.31) that the energy eigenvalues are all negative, unlike
those of the harmonic oscillator, which were all positive. This actually results from
defining the potential energy of the harmonic oscillator to be zero at the
equilibrium of the particle, while the hydrogen potential is instead defined to be
zero at large distance from the nucleus. Another difference is that the energy of the
hydrogen atom has a maximum value, namely zero, while there is no upper limit
for the energy values of the harmonic oscillator. This means, physically, that while
the particle can never escape in a harmonic oscillator, in a hydrogen atom, the
electron escapes if its total energy is greater than zero. Such a loss of the electron
is called ionization of the atom. The ionization energy of the hydrogen atom is
13.6 eV.
If the electron is excited (say, by passing a spark through hydrogen gas) from the
ground state to a higher but still bound energy level, it will in time make a
transition back to a lower energy level. The energy lost by the electron during a
transition is emitted as electromagnetic radiation in the form of a photon. The
energy of the photon emitted during electron transition is the difference between
the initial and final energy eigenvalues. That is,
h 2  1 1  (6.33)
E photon = Eni − En f = −
2me a02  n2f ni2 
 
According to Planck’s formula, the natural frequency of the electromagnetic

radiation emitted is simply the photon's energy divided by h :
E photon h  1 1  

1 1 
ν= = − = RH − (6.34)
2π h 4π me a02  n2f ni2   n2 n 2 
i
  f  
Where RH = 109677.6 cm −1 is the Rydberg constant for the Hydrogen atom. Figure
6.1 shows some of the possible transitions which constitute the spectrum of
hydrogen. The most energetic photons, in the ultraviolet range, are emitted by
Lyman transitions. Balmer transitions emit visible light and Paschen ones infrared.

Figure 6.1: Spectrum of the hydrogen atom.
There is a convention to name l = 0 the " s " states, l = 1 the " p " states, l = 2 the
" d " states and l = 3 the " f " states. From there on follow the alphabet with
g , h, i, L . The ground state of Hydrogen has n = 1 and l = 0 . Therefore, according
to the convention, the ground state is called the 1s state. The first excited state of
Hydrogen has n = 2 . Corresponding to this, there are four degenerate states (not
counting different spin states). In terms of ψ nlm , these are ψ 200 , ψ 21−1 , ψ 210 , ψ 211 .
These are called the 2s and 2 p states.
The second excited state has n = 3 with the degenerate states 3s , 3 p and 3d . This
totals 9 states with the different allowed m values.
In general, for a given principal quantum number n , there are n 2 degenerate

states, again not counting the different spin states.
The Hydrogen spectrum was primarily investigated by measuring the energy of

photons emitted in transitions between the states, as depicted in figures 6.1 and
6.2.

Figure 6.2: Possible electron transitions in a Hydrogen atom. Note that transitions
which change l by one unit are strongly preferred.

Quantum Mechanics PyEd 342

Uploaded by

Copyright:

Available Formats

Quantum Mechanics PyEd 342

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quantum Mechanics PyEd 342

Uploaded by

Copyright:

Available Formats

Quantum Mechanics (PyEd 342)

Chapter 1. The Motivation for Quantum Mechanics

1.1. What is Quantum Mechanics?

The development of quantum mechanics was initially motivated by several

1.2. The Breakdown of Classical Physics

F = ma, Newtonian Mechanics

In the case of Newtonian formalism of mechanics, the motion will be determined

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 1

At this point, it is useful to look at some of the failures Classical Mechanics to

1. The ultraviolet catastrophe:

One of the failures of Classical Mechanics is its inadequacy to solve the

In 1900, Rayleigh used the laws of classical thermodynamics and classical

Where k is Boltzmann constant and c is the speed of light. Equation (1.2) is

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 2

So, according to classical physics (R-J

Where h = 6.626 ×10−34 J .s is a constant, now called Planck's constant.

Max Planck arrived at this result by combining statistical mechanics with a

In classical physics, given a frequency, the energy is determined by the

Integrate equation (1.3) to obtain equation (1.1).

2. The stability of atoms:

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 3

Electrons orbiting a nucleus experience a centripetal acceleration and,

• Electronic orbits are stable.

• Electronic angular momentum is quantized.

Using these postulates and Newtonian Mechanics, Bohr was able to

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 4

3. The low specific heats of crystals:

Classically, the molar heat capacity at constant volume, Cv of a crystal is 3R ,

Verify that equation (1.8) agrees with experimental data for kT hν .

The photoelectric effect

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 5

• For frequencies above the threshold, the number of electrons emitted is

Contrary to what is observed in the lab, classical electromagnetic theory

Einstein solved the problem of the photoelectric effect in 1905. Borrowing

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 6

The Compton Effect marked a radically essential departure from classical

Being cognizant of the particle-like properties of waves, de Broglie boldly

According to de Broglie, the wavelength of a particle moving with momentum

1) Energy and angular momentum are discrete

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 7

1.3. The Uncertainty Principle

In the wave-like microscopic world, simultaneous measurements of position and

notice that the uncertainty principle is inconsequential to macroscopic objects

1.4. Postulates of Quantum Theory

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 8

Square integrable means

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 9

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 10

Chapter 2. Mathematical Foundation

An operator Â is a mathematical entity that acts on a given function f ( x ) and

In 1D Quantum Mechanics, the coordinate operator is x̂ = x and the momentum

Prepared by Mesfin Tadesse, CoE, AAU, Megabit 2000 EC 11

2.2. Basic Properties of Operators

For any two operators, Â and B̂ , and a function f :

( Aˆ + Bˆ ) f = Afˆ + Bfˆ ( Aˆ − Bˆ ) f = Afˆ − Bfˆ (2.4)

Product of two operators

For any two operators, Â and B̂ , and a function f :

Two operators, Â and B̂ , are equal if, for all functions f :

The identity operator

The associative law holds for operators, that is,

Â (α f + β g ) = α Aˆ f + β Aˆ g (2.25)