Summary of last week.

The problems that classical physics (mechanics, thermodynamics,

electrodynamics, statistical mechanics) failed to solve gave rise
to these requirements for the new quantum mechanics:
(1) Physical variables can take on discrete values
(2) Particles can behave like waves and vice versa (duality)
In addition:
(3) Each particle has an intrinsic magnetic moment (spin)
(4) Two electrons cannot be in the same state
To satisfy (1) and (2), it was assumed that each particle is also
a wave, described by a wave function ψ (x , y , z ,t ); and each physical
variable (called “observable”) A is replaced by an operator ^ A:
Energy

^ =iℏ ∂
E→E
∂t
Momentum
p → ^p =−iℏ ∇
Conservation of energy becomes the Schrodinger equation
∂ψ ^
E=H → ^ ^ ψ → iℏ
E ψ= H =H ψ
∂t
that takes the place of equation of motion F=m a.
The physical variable A can only take on values that are
eigenvalues of ^
A
^
A ua =a ua

When A takes the value a , the system is in the state ua .

Then, as we saw, these naturally led to the probability
interpretation, uncertainty principle, and indistinguishability.
Indistinguishability required that, if we have a system of
identical particles (or atoms), their collective wave-function
has to be either symmetric (bosons) or anti-symmetric (fermions).
This handles the requirement (4) above.
We also saw that the average (expected) value of an observable
when the system is at state ψ is

⟨ A ⟩ =( ψ , ^
Aψ)

*************************
Using Schrodinger equation, it is found that
d ⟨ p ⟩ −d ⟨ V ⟩
=
dt dx
showing that the classical equation of motion (dp /dt=−dV /dx ¿ is
valid for average values.
Energy eigenvalues which are usually of the most interest are
found by solving the eigenvalue problem
^
H u=Eu
2
−ℏ 2
∇ u+ Vu=Eu
2m
with suitable boundary conditions.
One of the few problems that can be solved exactly in qm is the
central force problem or hydrogen atom. This is the qm
counterpart of Kepler problem. Using spherical coordinates, the
electrostatic potential between nucleus and electron is
e
V =−k e
r
assuming the nucleus is so large that it does not move, only
produces V ; e is the charge of proton (−e electron).
When the solution is carried out it gives exactly the same energy
levels as Bohr model
−K
En = 2
n
The eigenfunctions (wave functions) unlm (r , θ , φ) depend on three
integers n , l ,m (called quantum numbers) which take on values as
follows

l=0 (s) m=0 n=1 ,2 , 3 , …

l=1 (p) m=−1 , 0 , 1 n=2 ,3 , 4 , …
l=2 (d) m=−2 ,−1 ,0 , 1 ,2 n=3 , 4 , 5 , …
In the minimum energy case the electron would be in the state
l=0 , m=0 ,n=1. If it gains energy, by em radiation for example, it
would jump to one of the higher levels by absorbing a photon of
frequency ∆ E /ℏ. If it were at a higher level at the beginning,
it could jump into lower levels either by itself (spontaneous
transition) or due to effects of external em (stimulated
transition) by releasing a photon of suitable frequency.
But there is nothing in Schrodinger equation that says how these
energy levels are to be filled when more than one electron is
present (assuming interaction between electrons is much weaker
than the interaction with nucleus; this works surprisingly well
although there are exceptions). That is mainly determined by
Pauli exclusion principle. Schrodinger equation together with
Pauli exclusion principle explains the electron shell structure
(together with couple of observational rules).
Quantum mechanical angular momentum, like the classical angular
momentum, remains constant (is conserved) in the central field
problem.
***********************************
Schrodinger equation can easily be generalized to N particles (as
in classical mechanics). For example, consider two particle case
in one dimension. Let the coordinates of particle 1 and particle
2 be x 1 and x 2. Now the wave function is

ψ (x 1 , x 2 ,t )
Classical Hamiltonian is
2 2
p1 p2
H= + + V (|x 1−x 2|)
2 m1 2m2
Where V is the interaction potential between the two particles
(only depends on the absolute distance). Quantization
∂ ∂
p1 →−ih , p2 →−ih
∂ x1 ∂ x2
And Schrodinger equation becomes
2 2 2 2
∂ ψ −ℏ ∂ ψ ℏ ∂ ψ
iℏ = − +V (|x 1−x 2|)
∂ t 2 m1 ∂ x 21 2 m2 ∂ x 22

********************************
We have not talked about spin until now in the development of the
theory. Schrodinger equation does not take spin into account.
There are two main effects of spin:
It adds small terms to total energy (Hamiltonian) and that
results in additional terms in the Schrodinger equation. With
spin terms added, it is sometimes called the Pauli equation.
This brings minute changes to Hydrogen spectra that was also
experimentally verified.
The other effect of spin actually has to do with
indistinguishability:
Consider a qm system consisting of two indistinguishable
particles, their positions x 1 , x 2. Indistinguishability means when
particles are exchanged, the state should be the same. The
physically observable quantity is not the wave function ψ itself
but |ψ|2 . Thus ψ and e iα ψ show the same state (since
2 2
|eiα ψ| =|eiα| |ψ|2=|ψ|2). The wave function for the two-particle system
is ψ ( x 1 , x 2) (suppressing t ); since it should be the same state when
particles are exchanged, this means
iα
ψ ( x 2 , x 1) =e ψ ( x 1 , x 2)
Exchanging the particles again
iα 2iα
ψ ( x 1 , x 2) =e ψ ( x2 , x1 ) =e ψ (x 1 , x2 )
Thus
2 iα iα
e =1 → e =∓ 1
So the wave function for two identical particles is either
symmetric
ψ ( x 2 , x 1) =ψ ( x 1 , x 2)

or antisymmetric
ψ ( x 2 , x 1) =−ψ (x 1 , x 2 )

This can be generalized to any number of identical particles.

It turns out that:
The wave function of a system of identical bosons (integer spin)
is symmetric.
The wave function of a system of identical fermions (half-integer
spin) is antisymmetric.
We see that two fermions in the same state makes the state zero;
two fermions cannot be in the same state. Pauli exclusion
principle is a special case of this (for electrons).
How do we know bosons have symmetric wave functions, and fermions
have anti-symmetric wave functions? This comes out of
relativistic quantum mechanics; spin turns out to be a
relativistic effect (depending on point of view).

We have not taken relativity into account until now. The non-
relativistic quantum mechanics has to be amended with wave-
function symmetry-anti-symmetry and the minute effects of spin
has to be added by hand.
Relativistic qm, as we will see, includes spin naturally (but not
symmetry-anti-symmetry).
As we have seen, the actual space we live in is a Minkowski
space. The equations describing particles should be Lorentz-
invariant; Schrodinger equation is not, since it was deduced by
starting with the non-relativistic energy-momentum relation
2
p
E=
2m
Schrodinger equation should be the non-relativistic limit (i.e.,
making c=∞) of a Lorentz-invariant equation.
As we have seen before (in mechanics), the relativistic energy-
momentum relation is
2 2 2 2 4
E =c p +m c
If we make the substitutions
∂
E →−iℏ , p→−iℏ ∇
∂t
as before, we find
2
2 ∂ψ 2 2 2 2 4
−ℏ 2
=−ℏ c ∇ ψ +m c ψ
∂t
This is called Klein-Gordon equation, and is Lorentz-invariant.
But when a potential is added as in Schrodinger equation, and
solved, it gives wrong energy levels (spectrum) for the electron
in a hydrogen atom (actually, Schrodinger first did that). The
interpretation is, Klein-Gordon equation does not govern the
electron; it is for some other particle (turns out to be
particles of spin zero; electron has spin 1/2. For example, the
Higgs particle you may have heard of has spin zero – Higgs
boson).
So, what is the Lorentz-invariant wave equation for the electron?
Dirac tried to find a first order equation in both space and time
(Lorentz invariant equations must have space and time derivatives
at the same order [2 for Klein-Gordon]; remember, space and time
are on equal footing). In other words, write the energy-momentum
relation as

E=c √ p2 +m 2 c2
Then, making the substitutions we get
∂ψ
=c √−ℏ ∇ +m c ψ
2 2 2 2
−iℏ
∂t
The problem here is the square-root of an operator does not make
sense (think of square-root of d /dx ). We need to get rid of the
square-root.
I will describe the procedure on a simpler example; but it is
basically the same thing.
We want to be able to write

√ x 2+ y2 =ax +by
If these are numbers, this equality cannot be satisfied (except
at special values). The idea is this: let a and b be matrices
and x , y be column vectors. Then write
2 2
x + y =( ax+ by)( ax+ by)
Note that x 2 means x T x (but not xx T , be careful), and a 2 means
T T
a a=a a . Therefore, the above actually means
T T T
x x+ y y=(ax +by ) (ax +by )
In expanding the parentheses, we should be careful that matrices
do not commute when multiplied (actually, this is the whole
point),
T T T T T T
x x+ y y=( ax ) ( ax )+ ( by ) ( by )+ ( ax ) (by)+ ( by ) (ax)
Using the properties of matrix multiplication,

( ax )T (ax )=(x T a T ) ( ax )=x T ( aT a ) x

Doing this for all four terms

x T x+ y T y=x T ( a T a ) x + y T ( bT b ) y + x T ( aT b ) y+ y T ( bT a ) x
The last term can be written as
T
y ( b a ) x=[ ( b a ) x ] y
T T T

since y T u=uT y , for column vectors only, not matrices. Thus

T
y ( b a ) x=[ ( b a ) x ] y=x ( b a ) y
T T T T T

And we have

x T x+ y T y=x T ( a T a ) x + y T ( bT b ) y + x T ( aT b +bT a ) y
Therefore, if we can find matrices a and b satisfying
T T
a a=b b=I
T T
a b+b a=0
(where I is the unit matrix) then we would have
T T T
x x+ y y=(ax +by ) (ax +by )
or
2 2 2
x + y =( ax +by ) ( ax +by )= ( ax+ by )
or,

√ x 2+ y2 =ax +by
There may be an infinite number of matrices satisfying the above
equations (I don’t know); but we need only one of them:

a= ( 10 −10 ) , b=(01 10)

HOMEWORK. Check that the above equations are satisfied by these
a and b .
Going back to
∂ψ
=c √−ℏ ∇ +m c ψ
2 2 2 2
−iℏ
∂t
Since there are four terms inside the square-root, the matrices
in this case will be 4 × 4 and ψ will be 4 ×1 (column matrix or
vector). Otherwise, the procedure is the same. The result is

iℏ
∂ψ
∂t (
=−iℏc α 1
∂ψ
∂x
+α 2
∂ψ
∂y
+ α3
∂ψ
∂z )
+m c 2 α 4 ψ

where α 1 ,2 ,3 , 4 are 4 × 4 matrices satisfying

2 2 2 2
α 1=α 2=α 3=α 4 =I
α i α j+ α j α i=0 ,i ≠ j
(There are an infinite number of solutions to these).
Dirac eqn. (the one above) was made up in the hope that it would
be the correct Lorentz-invariant wave equation for the electron.
It turned out to be very successful:
1. It gives the correct energy levels for the Hydrogen atom.
2. It also gives some of the finer details in the hydrogen
spectrum that was experimentally known but was not predicted
by the Schrodinger eqn.
3. The non-relativistic limit of Dirac equation is the
Schrodinger equation with the Pauli terms added.
4. It gives the spin angular momentum of the electron as 1/2,
and shows that spin plus orbital angular momentum is
conserved, orbital angular momentum by itself is not
conserved.
The non-relativistic limit of Klein-Gordon eqn. is the
Schrodinger eqn. without the spin terms.
HOMEWORK. 4 × 4 version of the example I showed above; we want to
write

√ x 2+ y2 + z 2 +w 2=ax +by +cz + dw

Find the relations that the matrices a , b , c , d must satisfy.
[If you want to see the explicit form of these matrices, look up
Dirac matrices.]
So, Dirac equation gave much more than expected. It includes
spin naturally (Schrodinger equation does not). It looks like
spin is actually a relativistic property.
Another prediction of Dirac equation is the existence of the
antiparticle of electron, called positron. It has all the same
properties of electron except that its charge is positive,
opposite of electron.
Now it is believed that every particle has an anti-particle.
Photon is its own anti-particle.
How do we know half-integer spin particles are fermions, and
integer spin particles are bosons? This is also a result of
relativistic quantum mechanics (called spin-statistics theorem).
**********************************
Up to now we did not say anything about photons, just that the
passage of an electron between the energy levels of an atom is
accompanied by the emission or absorption of a photon of definite
frequency, but we do not have any equations describing photon at
this point. It should be somehow related to Maxwell equations.
Remember that one of the main goals is to describe how elements
emit and/or absorb em radiation (light).
If we consider the em field as a classical (non-quantized) field,
i.e., as a given known function (no photons), this explains some
of the light-matter interaction problems (jumping of electrons
between energy levels in an atom). But it cannot model the
phenomena of spontaneous emission: this is the passage of an
electron to a lower state (thus emitting a photon) without any
external effects. Turns out we need to quantize em field to
explain and compute spontaneous emission.
At any rate, the next logical step is to generalize the
“quantization” concept to electromagnetic field.
The procedure is not really difficult; it relies heavily on
Fourier series and integrals. This subject, “Fourier series and
integrals” or “Fourier analysis” lies at the very heart of
applied mathematics, including so many applications in mechanical
engineering, and you should be familiar with it. It is normally
a part of our Math III course.
The first step is to quantize free em field, naturally. Since em
field is a field, this procedure is called “field quantization”.
Field quantization incorporates symmetry-anti-symmetry into the
equations.
Recall Hamiltonian formulation from mechanics. Hamiltonian was
the total energy. Same thing applies here; Hamiltonian of a free
em field is the total energy, which is
 H=
1
2
∭ (2 1 2
ε 0 E + B dV
μ0 )
The difference here is, it is an integral. As was done in cavity
radiation, the field is considered to be in a cubic box of volume
3
L and resolved into its components (each component is called a
harmonic oscillator). The end result is, after quantization, the
Hamiltonian operator can be written as
❑ ❑
^
H= ∑ ∑ ℏ ωk ¿ ¿ ¿
σ=1 ,2 k

Here, σ =1 ,2 denote the two polarization directions (see note 6),

2π
k= (l ,m , n)
L
ω k =|k|c

l , m, n are integers (so, sum over k is actually a triple sum).

The operators a^ +¿¿
kσ and a^ kσ are called photon creation and
annihilation (destruction) operators, respectively.
+¿¿
a^ kσ creates a photon with frequency ω k polarization σ , momentum
ℏ k and energy ℏ ω k .
a^ kσ annihilates a photon with frequency ω k polarization σ
momentum ℏ k and energy ℏ ω k .
Note that this does not say anything about where photon is; since
its momentum is known precisely (momentum is ℏ k ), photon can be
anywhere.
The interpretation is as follows: The form of ^
H suggest that em
field is equivalent to an infinite number of independent (not
interacting with each other) particles (photons) of various
frequencies.
Photons are bosons with spin 1.
There are two specific equations named Schrodinger: The general
form
∂Ψ ^
iℏ =H Ψ
∂t
is valid in any qm problem since it expresses ^
EΨ= ^
H Ψ , energy
conservation. For example, the photons satisfy this:
 ❑ ❑
∂Ψ
iℏ = ∑ ∑ ℏ ω k ¿¿ ¿
∂ t σ =1 ,2 k
The other one is when we substituted the non-relativistic
Hamiltonian in the general equation and obtained
2
∂ ψ −ℏ 2
iℏ = ∇ ψ +Vψ
∂t 2m
only valid for a non-relativistic, spinless particle moving in a
potential V (x , y , z ).
The interaction of em field with a Hydrogen atom in the
nonrelativistic case is modeled as follows. In classical
electrodynamics, Hydrogen atom (central potential plus electron)
plus em field have the Hamiltonian

H=
1
2m
1 1
(
( p−e A )2 +V + ∭ ε 0 E 2+ B2 dV
2 μ0 )
where, V =−k e e/r is the central potential of nucleus, p is the
momentum of the electron; and A is called the magnetic vector
potential, it is related to magnetic field
B=∇ × A
Interaction of H atom and em radiation comes from p ∙ A term when
the square is expanded. When this system is quantized last term
(triple integral) becomes as written before, p becomes ^p in 3-D
motion (generalizing the before ^p x =−iℏ∂ /∂ x )

^p=−iℏ∇
And A is obtained during the quantization of em field,
❑ ❑
A= ∑
^ ∑ Γ k ω k u kσ ¿ ¿ ¿
σ =1 , 2 k

Γ k is a constant, ukσ is the polarization vector. This

formulation does handle spontaneous transition mentioned before.
One important conclusion from field quantization is this: the em
force is due to exchange of “virtual photons” between charge
carriers. This idea later proved to be very useful in developing
theories of other forces.
******************************************
After the quantization of em field, the next logical question is
this: Quantizing em field gave photon particles. Electrons are
particles; which field equation should be quantized to obtain
electron particles?
The obvious choice is the Dirac equation. Em field equations are
Lorentz-invariant, and so is Dirac equation.
This is done by following the same steps, i.e., first find the
Dirac Hamiltonian

H=∭ ¿ ¿

where ψ +¿¿ denotes transpose followed by complex conjugate (called

Hermitian conjugate). Put electrons in a box and separate into
components. Then introduce creation-destruction operators for
particles, etc. EM-Dirac coupled theory is called quantum
electrodynamics (qed). QED is said to be the most successful
physical theory ever developed, looking at its impressive
agreement with experiments.
Since Dirac equation itself was obtained by quantizing
relativistic equation of motion, quantizing the Dirac equation is
sometimes called “second quantization”. Second quantization or
field quantization includes symmetry-anti-symmetry naturally. In
the case of Dirac equation, anti-symmetry property has to be
used, symmetric functions give meaningless results.
*************************************
This same field quantization (or second quantization) method can
also be applied to Schrodinger equation. It was found that
particles resulting from Schrodinger eqn. as a field eqn. can be
fermions or bosons; both work. Since Schrodinger equation is the
limiting case of both the Klein-Gordon equation (boson, spin
zero) and the Dirac equation (fermion, spin 1/2), this is to be
expected (logical). Hamiltonian for the Schrodinger equation is

( )
2
ℏ
H=∭
¿ ¿
∇ ψ ∙ ∇ ψ +V ψ ψ dV
2m
Similar field quantization procedures can be applied to other
field equations. For example, quantizing the acoustic field,
“phonons”, quanta of acoustic waves are obtained; quantizing the
plasma oscillation waves gives “plasmons” (plasma is ionized gas
– considered to be the fourth state of matter).