A Review of Quantum Mechanics

Chun Wa Wong
Department of Physics and Astronomy, University of California, Los Angeles, CA 90095-1547∗
(Dated: September 21, 2006)
Review notes prepared for students of an undergraduate course in quantum mechanics at UCLA,
c
Fall, 2005 – Spring, 2006. Copyright 2006 Chun Wa Wong

I. The Rise of Quantum Mechanics (e2G = e2 /4πǫ0 in Gaussian units). Hence

1.1 Physics & quantum physics: U

E = T +U = = −T. (5)
2
Physics = objective description of recurrent physi-
cal phenomena Bohr’s two quantum postulates (L quantization and
Quantum physics = “quantized” systems with both quantum jump on photon emission):
wave and particle properties
L = mva = n~
~, n = 1, 2, ...,
1.2 Scientific discoveries: hν = Eni − Enf . (6)
Facts → Theory/Postulates → Facts For ground state:
1.3 Planck’s black body radiation: U1 e2
E1 = = − G = −13.6 eV
Failure of classical physics: 2 2a
~2
1 = −T1 = − ,
Emissivity = uc, λmax T = const, 2ma2
4 ~2
energy density = u = ρν hEi, a = = 0.53Å (Bohr radius), (7)
me2G
hEiexp =
6 kB T, (1)
For excited states:
(ρν = density of states in frequency space).
E1
Planck hypothesis of energy quantization (En = rn = n2 a, En = . (8)
n2
nhν) explains observation:
1.6 Compton scattering:
hν ∞ −nx
P
n=0 ne hν Energy-momentum conservation for a photon of lab
hEi = P ∞ −nx
= hν/k T , (2)
n=0 e e B −1 energy E = hc/λ scattered from an electron (of
mass m) initially at rest in the lab (before=LHS,
where x = hν/(kB T ). after=RHS; 1=photon, 2= recoiling e):
1.4 Einstein’s photoelectric effect: p = p1 + p2 ,
q
Photoelectrons ejected from a metal by absorbed E + mc2 = E1 + m2 c4 + p22 c2 . (9)
light of frequency ν have maximum energy:
Photon scattered into lab angle θ has longer wave-
E = hν − W (work function) (3) length λ1 :
that depends on the frequency ν of light and not λ1 − λ = λe (1 − cos θ), where
its intensity. h hc
λe = = = 2.4 pm (10)
1.5 Bohr’s atomic model: mc mc2
Electron of mass m in a classically stable circular is the electron Compton wavelength.
orbit of radius r = a and velocity v around the 1.7 de Broglie’s matter wave:
atomic nucleus:
Massive matter and massless light satisfy the same
mv 2 e2 U energy-momentum and momentum-wavelength re-
= G , or T = − , (4) lations
r r2 2
E 2 = p 2 c2 + m 2 c4 ,
h
∗ Electronic address: cwong@physics.ucla.edu p = ~k = . (11)
λ
 2

Thus moving matter is postulated to be a wave with LDEs (linear DEs) with constant coefficients
a motional de Broglie wavelength λ = h/p, leading have exponential solutions:
to the diffraction maxima for both matter and light
of d ikx
e = ikeikx ,
dx
2-sided formula: 2d sin θ = nλ,
d −iωt
1-sided formula: D sin θ = nλ. e = −iωe−iωt . (14)
dt
These DE are called eigenvalue equations be-
II. Waves and quantum waves cause the constant ik is called an eigenvalue of the
Classical and quantum waves have identical mathemat- differential operator d/dx, while the solution eikx is
ical properties. called its eigenfunction belonging to its eigenvalue
ik.
2.1 Waves in physics: The 1DWEq (12) has four distinct (i.e., linearly
Waves rise and fall, travel, have coherence in space independent) eigenfunctions:
and in time, interfere, and diffract.
Ψ(x, t) = e±i(kx±ωt) . (15)
Two classes of waves:

(a) Inertial waves in massive media: ocean waves, Energy in both classical and quantum waves is car-
sound waves, vibrations of violin strings; ried by the intensity or energy flux
(b) Noninertial waves in vacuum: EM waves, I ∝ |Ψ|2 . (16)
matter waves.
Its quadratic dependence on Ψ is responsible for the
2.2 Mathematical description of waves: interference between two traveling waves Ψi =
The 1-dimensional wave equation (1DWEq) Ai eiθi with real Ai :

∂2 1 ∂2 |Ψ1 + Ψ2 |2 = |Ψ1 |2 + |Ψ2 |2 + 2Re(Ψ∗1 Ψ2 )

 
− 2 2 Ψ(x, t) = 0 (12)
∂x2 v ∂t = A21 + A22 + 2A1 A2 cos(∆s),
   
∆ω
is an equation of state that gives a wave function where ∆s = ∆k x − t . (17)
Ψ(x, t) describing an unfolding event in spacetime. ∆k
The wave equation is a partial differential equation
In the limit ∆k → 0,
(PDE) because it depends on more than one vari-
able, here the two variables x and t. dω
= vg = group velocity
In contrast, the solution x(t) of an equation of mo- dk
tion such as a Newton equation describes how a ω
6= v = = wave velocity. (18)
single point x(t) on an object evolves in time. Sim- k
ilarly, a snapshot of an object is not necessarily a
wave. A wave function Ψ(x, t) is needed to describe Examples: Certain waves satisfy the relation ω =
a wave’s coherent structure in both space and time. Ak p . Then vg /v = p. Specific examples are p = 2
for transverse vibrations of a thick bar, p = 3/2
However, a wave (any wave) can have particle prop-
for short ripples moving under surface tension, and
erties if a point x can be defined on it such that one
p = 1/2 for long waves on deep sea.
can describe how its position changes in time.
The result (17) is similar in structure to the squared
The wave equation (12) describes linear waves sat-
length of a sum of two vectors in space that inter-
isfying the superposition principle that a sum
sect at an angle θ:
of solutions is also a solution :
|A1 + A2 |2 = A21 + A22 + 2A1 A2 cos θ, (19)
Ψ = a1 Ψ1 + a2 Ψ2 . (13)

Two waves of the same amplitude traveling in op-

Traveling waves:
posite directions interfere form a standing wave:
(a) f (x − vt)= a wave traveling to the right (+x),
Ψ = A[cos(kx − ωt) + cos(kx + ωt)]
(b) g(x + vt)= a wave traveling to the left (-x),
= 2A cos(kx) cos(ωt) (20)
(c) the composite wave a1 f (x − vt) + a2 g(x + vt)
is also a wave. with factorized space and time dependences.
 3

2.3 Matter-wave quantization: is the normalized probablity density of finding the

Light wave of momentum p and energy E = pc is wave in “k space”. (We shall not consider more
described by the wave function: complicated properties that also depend on the
time t or the energy E.)
ei(kx−ωt) = ei[(p/~)x−(E/~)t] . (21) Uncertainty relations: A wave of any kind sat-
de Broglie and Schrödinger suggested that the RHS isfies the uncertainty relation
expression holds also for matter waves so that: 1
∆x∆k ≥ , where

~ ∂
 2
ei(p/~)x = p̂ ei(p/~)x = p ei(p/~)x , (∆x)2 = h(x − x̄)2 i, x̄ = hxi,
i ∂x

∂
 (∆k)2 = h(k − k̄)2 i, k̄ = hki. (28)
i~ e−i(E/~)t = Ĥ e−i(E/~)t = E e−i(E/~)t . (22)
∂t The expression for matter waves is usually written
The differential operators that appear are called as the Heisenberg uncertainty principle:
the momentum operator p̂ and the Hamiltonian (or ~
energy) operator Ĥ, respectively. ∆x∆p ≥ . (29)
2
Commutation relations: The rules of differen-
tial calculus dictates that a differential operator Spreading wave packets: A Gaussian wave
does not commute with its own variable: packet has the spectral amplitude

∂ ∂

∂
 2 1
x = 1 + x , or , x = 1; g(k) = e−αk , with ∆k = √ . (30)
∂x ∂x ∂x 2 α
 
∂ ∂ ∂ It gives rise to a wave function at time t = 0 of
t = 1 + t , or , t = 1. (23)
∂t ∂t ∂t 2 √
f (x, t = 0) ∝ e−x /(4α) , with ∆x = α. (31)
These results give the fundamental commuta-
tors of wave mechanics Hence the wave packet at time t = 0 has the mini-
mal uncertainty product of
~
[p̂, x] = , [Ĥ, t] = i~. (24)
i 1
∆x∆k = . (32)
2
2.4 Wave packet and the uncertainty principle:
A wave packet is a superposition of waves of the For t 6= 0, however, the uncertainty product could
type (21) be greater than the minimal value of 1/2. This hap-
Z ∞ pens if the system is dispersive, meaning a nonzero
1 second Taylor coefficient β in the following expan-
f (x, t) = √ g(k)ei[kx−ω(k)t] dk, (25)
2π −∞ sion:
1 d2 ω 
 
with a linear superposition function g(k) sometimes dω 
ω(k) ≈ ω(k = 0) + k  + k 2 2 
called a spectral amplitude. dk 0 2 dk 0
Expectation value (= mean value) of a property 1
= ω(0) + kvg + k 2 β. (33)
A(x, t) of the wave: 2
Z ∞ The resulting approximate wave function at any
hA(x, t)i ≡ ρP (x, t)A(x, t)dx, where time t can then be evaluated to the closed form
−∞
2
|f (x, t)|2 f (x, t) ∝ e−xt /(4αt ) , with
ρP (x, t) = R ∞ (26)
−∞
|f (x, t)|2 dx iβ
xt = x − vg t, αt = α + t. (34)
2
is the normalized probablity density of finding the
wave at the spacetime point (x, t). Thus the position x of the wave moves in time with
the group velocity
The expectation value of a time-independent prop-
erty B(k) of the wave can be calculated by using

dω 
the spectral amplitude g(k): vg = . (35)
dk 0
Z ∞
hB(k)i ≡ ρ̃P (k)B(k)dk, where Its position uncertainty increases in time to
−∞ "  2 #1/2
|g(k)|2 |αt | √ βt
ρ̃P (k) = R ∞ (27) ∆x(t) = √ = α 1 + . (36)
2
−∞ |g(k)| dx
α 2α
 4

For a massive particle in free space E 2 = p2 c2 + in the sense that the probability for finding the
m2 c4 . Hence state somewhere in space at time t is
Z ∞
d2 E m 2 c6
β=~ 2
= ~ 3 6= 0. (37) P (t) = ρP (x, t)dx
dp E −∞
Z ∞
Thus the spatial width of the Gaussian packet of = |Ψ(x, t)|2 dx = 1. (44)
a massive particle spreads out in both positive and −∞
negative times. By choosing P (t) = 1, the wave function is said to
be normalized.
III. Wave mechanics 3.3 Conservation of |Ψ|2 :
Wave mechanics describes physics by using the The probability density can be moved around, but
Schrödinger wave eq. and its wave functions. cannot be created or destroyed in the absence of
explicit creative or destructive physical processes:
3.1 Schrödinger wave equation (SchWEq) :  
First quantization: The classical E-p relation d ∂ dx ∂
0 = ρP (x, t) = + ρP (x, t)
E 2 = p2 c2 +m2 c4 is quantized into a wave equation dt ∂t dt ∂x
by replacing E, p by the differential operators ∂ ∂
= ρP (x, t) + jP (x, t). (45)
∂t ∂x
∂
E → Ĥ = i~ This is called the continuity equation for the con-
∂t
~ ∂ served probability current density
p → p̂ = : (38)
i ∂x jP (x, t) = v ρP (x, t), (46)

 2 " 2 # where v = dx/dt.

∂ ~ ∂ 2 2 4
i~ Ψ(x, t) = c + m c Ψ, (39)
∂t i ∂x
IV. Wave mechanics in one spatial di-
where the wave function Ψ(x, t) has to be added for mension
the operators to operate on. In this way, a classical Wave mechanics is applied to simple one-dimensional
state of motion “spreads out” into a wave structure. problems to illustrate how the SchWEq is solved and how
Nonrelativistic limit: Likewise, the NR energy its wave character affects the properties of a system.
E = T + V = p2 /2m + V (not containing the rest 4.1 Solving the Schrödinger wave equation
energy mc2 ) is quantized into the SchWEq: (SchWEq) :
Separation of varables: Since solutions are
  "  2 #
∂ 1 ~ ∂
i~ Ψ(x, t) = + V (x) Ψ, (40) unique up to a superposition, we may look for a
∂t 2m i ∂x solution of the factorized form

Commutation relations: A differential operator Ψ(x, t) = ψ(x)φ(t). (47)

does not commute with its variable. Hence
Then
~ 
p̂2

[p̂, x] = , [Ĥ, t] = i~. (41) φ(t) + V ψ(x)
i ψ(x)Ĥφ(t) 2m
= = E. (48)
ψ(x) φ(t) φ(t) ψ(x)
Group velocity: E 2 = p2 c2 + m2 c4 ⇒ EdE =
c2 pdp. Hence On simplification, the first expression is clearly a
function of t only, while the middle expression is a
function of x only. Hence the separation constant
 
E dE
vvg = = c2 . (42) E must be a constant independent of t or x.
p dp
The PDE can thus be broken up into two ordinary
3.2 Born’s interpretation of |Ψ| : 2 DEs:
 
The squared wave function gives the probabililty ∂
Ĥφ(t) = i~ φ(t) = Eφ(t),
density of finding system at spacetime location ∂t
(x, t): "  2 #
1 ~ ∂
+ V (x) ψ(x) = Eψ(x), (49)
|Ψ(x, t)|2 = ρP (x, t) (43) 2m i ∂x
 5

E
q2 E
V0 q2
2
k q2 V0
x=0 k2 Κ2
0
x=0
BCs: Ψ€ = ؐ 0
Ψ '€ = Ψ ' - V0 x=0 x=L

FIG. 2: Wave numbers for a rectangular potential barrier.

FIG. 1: Wave numbers and wave functions at a potential step.

Its probability current density is conserved across

where E appears in both ODEs as the energy
the boundary:
eigenvalue. The first ODE gives the time-
dependent factor ~k
1 − |B|2


j< =
φE (t) = e−i(E/~)t , (50) m
~q
= j> = |AT |2 . (55)
but the second equation is a second-order ODE m
whose solution requires the imposition of two Its reflection and transmission coefficients at the
Boundary conditions at a suitable boundary x = boundary are:
x0 :
R = |B|2 , T = 1 − R. (56)
ψ< (x0 ) = ψ> (x0 ),

′ d  ′ 4.4 Potential barrier:
ψ< (x0 ) ≡ ψ< (x) = ψ> (x0 ). (51)
dx x0 A particle of incident energy E higher than the
height V0 of a rectangular potential barrier is par-
Note that if the first BC is violated, ψ ′ (x0 ) becomes tially reflected and partially transmitted by a potet-
infinite and the DE (differential eq.) cannot be sat- nial barrier. The wave functions for x < 0 before
isfied. If the second BC is violated, ψ ′′ (x0 ) becomes the barrier and for x > L after the barrier remain
infinite. The DE also cannot be satisfied, except for the same as Eq.(54), but the presence of the sec-
the special case where V (x0 ) has a compensating ond boundary at x = L causes reflected waves to
infinity at x0 . be present in the barrier region
4.2 Particle in a box of length L: 0<x<L: ψ< (x) = Ceikx + De−ikx . (57)
With the two BCs ψ(0) = ψ(L) = 0, the solutions
turn out to be the same as those describing classical The two boundary conditions at the second bound-
vibrations of a violin string of length L: ary x = L are enough to determine the two extra
r coefficients C and D. The reflection coefficient re-
2  nπ  mains R = |B|2 , but its value is changed. The
ψn (x) = sin x (52) transmission coefficient is
L L
for a normalized wave function. 1
T =1−R=  2 . (58)
k2 −q2
4.3 Potential step: 1+ 2kq sin2 (qL)
Figure 1 shows a particle of mass m incident from
the left with energy E. The wave numbers k, q are Quantum tunneling: The formula for T works
defined as follows: even when the energy E falls below the height V0 of
~2 k 2 ~2 q 2 the potential barrier, as shown in Fig. 2. Then q 2 =
E= , E−V = ; (53) −κ2 < 0, and q = iκ becomes purely imaginary.
2m 2m
The function [sin(qL)/q]2 in the formula is analyti-
or k 2 = q 2 − k02 , where k02 = 2mV0 /~2 . The parti- cally continued to the function [sinh(κL)/κ]2 . The
cle’s matter wave is partially reflected at the bound- resulting transmission coefficient T across the finite
ary x = 0, and partially transmitted across it (with barrier is shown in Fig. 3. The fact that T does not
no further reflection downstream): vanish for E < E0 , as it does in classical mechanics,
is called quantum tunneling. Note that tunneling
x<0: ψ< (x) = eikx + Be−ikx , can occur only when the barrier thickness is finite.
x>0: ψ> (x) = AT eiqx . (54)
 6

T HaL
1
-W
0.8
External
0.6 Sample field

0.4 HbL

0.2 -W

EV0 Scanning
1 2 3 4 5 Sample tip

FIG. 3: Transmission coefficient across a rectangular potential FIG. 5: Tunneling through a finite potential barrier (a) in
barrier. cold emission, and (b) in a scanning tunneling microscope.

V the electron neutrino) powers the conversion of

1
Alpha mass into solar energy in the interior of the Sun.
emission Fusion The reaction rate depends on the temperature T
0.5 as T 4 .
The luminosity of a star usually increases as M 3 of
rR its mass M , or T 24 of its central temperature. Such
1 2 3
an unusually strong dependence led Bethe in 1938
-0.5
to look for another source of hydrogen burning. He
discovered that this alternative route is a certain
CNO cycle that involves certain carbon, nitrogen
and oxgen isotopes in intermediate steps.
FIG. 4: Tunneling across the Coulomb barrier in α decay and
in nuclear fusion. Cold emission: Electrons are bound in metals by
their work function W . However, they can be in-
duced to tunnel out of the metal by applying an
4.5 Applications of quantum tunneling: external electric field to change the potential to
Alpha decay: An α particle inside a nucleus feels a triangular penetrable potential, as illustrated in
an attractive nuclear potential. As the α particle Fig. 5(a).
leaves the nucleus in alpha decay, it sees the repul- In scanning tunneling microscopes, the tunnel-
sive Coulomb potential outside the nucleus. These ing is induced by the proximity of the scanning tip,
two potentials combine to form a potential barrier as shown in Fig. 5(b). Since the tunneling current
that the α particle must tunnel through on its way increases greatly as the thickness of the potential
out, as illustrated in Fig. 4. The decay probablility barrier decreases, one has a very sensitive device for
is thus f T , the product of the frequency f of the mapping the geometrical shape of a surface down
α particle’s hitting the inside wall of the potential to atomic dimensions.
barrier and the transmission coefficient T through
the potential barrier (that gives the tunneling prob- 4.6 Bound states in a potential well:
ability per attempt).
If V (x) → constant as x → ±∞, the lower of these
Nuclear fusion: is just the time reversed situation two contant values is usually taken to be the zero
of emission. Here a charged particle, say a proton, of the energy scale. Two situations can then be
is trying to enter a nucleus from the outside by tun- realized:
neling through its very high and thick Coulomb po-
tential barrier. The transmission coefficient is typi- (a) A system with E > 0 has continuous energy
cally very small, but increases very rapidly with the spectrum with wave functions that are not lo-
energy E. If a number of atomic nuclei are exposed calized or square-integrable, while
to a gas of protons of increasing temperature, the
(b) a system with E < 0 has a discrete energy
high-energy tail of the Boltzmann energy distribu-
spectrum with localized wave functions that
tion of the proton gas will allow more and more
are square-integrable. The resulting physical
protons to get into atomic nuclei to initiate nuclear
states are called bound states in a potential
fusion reactions. For example, the pp reaction
well.
p + p → d + e+ + ν (59)
The wave numbers associated with a bound state in
(where e+ is the positron and ν is a particle called an attractive one-dimensional square-well potential
 7

This is how a wave packet is constructed, namely

by taking a linear superposition of waves. Also
V =0
Κ2
 
−1 d
E F g(k) = −ixf (x). (64)
dk
q2
Thus a position operator x̂ = i∂/∂k gives the value
-V0 of the position x associated with the wave function
x = -a x=a f (x).
The FT shows that waves can be described either
by wave functions f (x) in space, or by their spectral
FIG. 6: Wave numbers associated with a bound state in an
attractive square-well potential.
amplitudes g(k) in k-space. Functions of one space
can be constructed from those in the other space.
Hence wave properties are completely specified ei-
are summarized in Fig. 6, where ther in ordinary space or in k-space, a “simplifica-
tion” that follows from a wave’s coherent structure.
~2 κ2 ~2 q 2
|E| = −E = , E−V = . (60) Table I gives some simple examples of FTs.
2m 2m
Dirac δ-function: Since
Such an attractive one-dimensional potential has
at least one bound state, the ground state of F −1 {F {f (x)}} = f (x), (65)
even parity whose wave function is proportional to
cos qx. This is because such a wave function al- we find (by writing out the two integral transforms
ready “bends” over at the boundaries, and can al- on the LHS in two steps) that
ways be matched to a decreasing exponential func- Z ∞
tion outside the potential well. δ(x − x′ )f (x′ )dx′ = f (x), (66)
−∞

V. The mathematical structure of quan- where

tum mechanics ∞
1
Z
′
Physical states in quantum mechanics are linear waves. δ(x − x′ ) = eik(x−x ) dk (67)
2π −∞
Both quantum states and classical linear waves can be
superposed to form other waves, in the same way that
ordinary vectors in space can be added to form other
vectors. TABLE I: Some Fourier transforms.
Property If f (x) is then F{f (x)} is
5.1 Fourier transform (FT):
The superposition principle for linear wave func- f(x) g(k)
tions (solutions of linear wave equations) is formal-
k2
ized in the methods of Fourier series and Fourier e−αx
2
√1 e− 4α
2α
transforms (FT). Both methods and their general-
izations are used in quantum mechanics. unit step functiona − √12π k2 sin ka
The FT of a wave function f (x) in one spatial vari-
d n
able x is defined by the integral Derivatives ( dx ) f (x) (ik)n g(k)
Z ∞
1 Derivatives (−ix)n f (x) d n
( dk ) g(k)
F {f (x)} ≡ g(k) = √ f (x)e−ikx dx. (61)
2π −∞
Translation f (x − a) e−ika g(k)
For example,
    Attenuation f (x)e−αx g(k − iα)
d ip
F f (x) = ikg(k) → g(k). (62)
dx ~ CCb Real g ∗ (k) = g(−k)
Thus the momentum operator p̂ = (~/i)∂/∂x yields Real and even g ∗ (k) = g(−k) = g(k)
the momentum value p = ~k associated with the
spectral amplitude g(k). Real and odd g ∗ (k) = g(−k) = −g(k)
Fourier inversion formula:
Z ∞
1 a from −a to a
F −1 {g(k)} ≡ f (x) = √ g(k)eikx dk. (63) b Complex conjugation
2π −∞
 8

is called a Dirac δ-function. transposition with the complex conjugation of all

Dirac has given a simple intuitive construction of “matrix elements”.
his δ-function as a rectangular function of width ǫ The Dirac bracket hφ|ψi is a scalar product of
and height 1/ǫ with a unit area under it: vectors or an inner product of functions
  
1/ǫ, when |x − x′ | ≤ ǫ/2, ψ1
δ(x − x′ ) = lim (68) hφ|ψi = (φ∗1 , φ∗2 ) = φ∗1 ψ1 + φ∗2 ψ2 , (71)
ǫ→0 0, otherwise. ψ2

5.2 Operators and their eigenfunctions: where the complex conjugation on the bra com-
ponents is needed to keep the self scalar product
When a differential operator operates on its eigen- hψ|ψi nonnegative. For two orthonormal vectors
function, it gives a function that is proportional or states:
to the eigenfunction itself:
  hi|ji = δij . (72)
~ ∂
p̂ eikx = eikx = ~k eikx ,
i ∂x Eigenbras and eigenkets: One goes beyond sim-
ple vector spaces by using the eigenfunctions of op-
 
∂
x̂ e−i(x/~)p = i~ e−ixp/~ = x e−ixp/~ . (69) erators such as position x̂ and momentum p̂. In
∂p
fact, we immediately associate their eigenfunctions
The proportionality constant is called its eigen- with abstract eigenstates in the following eigen-
value. bra or eigenket form:
Two functions or eigenfunctions can be combined x̂|xi = x|xi, and hx|x̂ = x∗ hx| = xhx|,
into an inner product:
p̂|pi = p|pi, and hp|p̂ = p∗ hp| = php|, (73)
Z ∞
(φ, ψ) ≡ φ∗ (x)ψ(x)dx where the eigenvalues x and p are real, as is true for
−∞
Z ∞ all physical observables, i.e., properties that can be
= gφ∗ (k)gψ (k)dk. measured experimentally. The eigenvalues x and p
−∞ normally range continuously from −∞ to ∞.
Dirac denotes this inner product by the bracket Projection and quantum destructive mea-
symbol hφ|ψi. surement: A simple 2-dimensional vector |Ai can
be decomposed into its components Ai = hi|Ai with
5.3 Dirac notation: respect to the basis |ii:
Functions in function space and states of a physi-
cal system in quantum mechanics are mathemati- |Ai = |1iA1 + |2iA2
N
!
cally similar to vectors in space. Dirac emphasizes X
this similarity using a mathematical notation that = |iihi| |Ai. (74)
works for vectors, functions and quantum states. i=1

A vector (function, state) can be expanded in terms In a similar way, an N -dimensional vector space is
of basis vectors (functions, states). Dirac distin- made up of the N one-dimensional subspaces pro-
guishes between two kinds of vectors (functions, jected out by the projection operators
states) – column vectors (functions, ket states)
and row vectors (complex-conjugated functions, Pi = Mi = |iihi| (75)
bra states), where the name bra and ket comes from
the two parts of the word “bracket”: in the sense that

|ψi = ket = ψ1 |1i + ψ2 |2i Pi |Ai = Ai |ii (76)

 
ψ1 leaves only that part of the vector that is in this
= = column vector
ψ2 ith subspace. Pi is also called a measurement sym-
hφ| = bra = φ∗1 h1| + φ∗2 h2| bol (hence Mi ). This measurement is not a classi-
= (φ∗1 , φ∗2 ) = row vector cal nondestructive measurement but a quantum de-
structive measurement that destroys all other com-
= |φi† , (70)
ponents of the original vector |Ai. The projection
For simple vectors, the basis states |1i and |2i are collapses the state |Ai into just the part Ai |ii in
just the unit basis vectors usually denoted e1 and the ith subspace.
e2 , respectively. The † symbol is called an adjoint Completeness relations: The original vector |Ai
operation. It combines the matrix operation of can be restored by adding up all its projections
 9

Ai |ii. Hence an identity operator can be con- Dirac δ-function: The equivalence
structed for this linear vector space Z ∞
ψ(x) = δ(x − x′ )ψ(x′ )dx′
N
X −∞
1= |iihi| (77) Z ∞
i=1 = hx|ψi = hx|x′ ihx′ |ψidx′ (82)
−∞
for discrete basis states that satisfy the orthonor-
shows that
mality relation hi|ji = δij .
The idea of completeness or closure holds also for hx|x′ i = δ(x − x′ ). (83)
eigenstates of operators such as x̂ and p̂ that have
It immediately follows that the Dirac δ-function
continuous eigenvalues:
can be represented in infinitely many ways. For
Z ∞ example,
1 = |xihx|dx
−∞ N
X N
X
hx|x′ i = hx|iihi|x′ i = φi (x)φ∗i (x′ ).
Z ∞
(84)
= |pihp|dp. (78) i=1 i=1
−∞

The resulting basis states |xi or |pi are said to be Quantum measurement and wave-function
continuous basis states. collapse: Electrons emitted from a source pass
through a single slit and fall on a distant observing
Representations: A discrete representation is a screen on the other side of the slit: As a coherent
specification of a vector (or state) |Ai by its com- matter wave, each electron forms a diffraction pat-
ponents Ai in a discrete basis. A continuous rep- tern on the observing screen in the direction y of
resentation uses the continuous basis states |xi to the diffraction pattern. (Diffraction = spreading of
give the components A(x) = hx|Ai in the expansion a wave into the geometrical shadow regions.)
Z ∞ A detector of acceptance ∆y placed at position
|Ai = dx |xiA(x). (79) y = d on the observing screen detects electrons
−∞ from signals of their passage through the detec-
tor. If ψ(y ′ ) = hy ′ |ψi is the wave function of self-
Wave functions: The (position) wave function of diffracting electrons normalized in some way, the
a state or wave |ψi is just its component ψ(x) = part that enters the detector is
hx|ψi in the position representation. Its compo- Z d+∆y/2
nent ψ(p) = hp|ψi in the momentum representa-
∆ψ(d) = dy ′ |y ′ ihy ′ |ψi. (85)
tion is called its momentum wave function. A wave d−∆y/2
function can be expanded into an arbitrary discrete
representation of basis states |ii The probability of an electron entering the detector
is
N
X Z d+∆y/2
hx|ψi ≡ ψ(x) = hx|iihi|ψi ∆P (d) = dy ′ |ψ(y ′ )|2 . (86)
i=1 d−∆y/2
N
Note that the electron described by the wave func-
X
= φi (x)ψi , (80)
i=1
tion ψ(y) can be anywhere on the screen before the
detection. However, once it is known to have gone
where φi (x) = hx|ii is the wave function of the basis into the detector, e.g., by the detector signal it gen-
state |ii. Incidentally, N must be infinite to match erates, its wave function has collapsed into ∆ψ(d).
the number of points on the straight line. This is how a quantum measurement can modify
the state itself.
A wave function can also be expanded in another
continuous representation, say the momentum rep- 5.4 Operators and their matrix elements:
resentation: Matrix mechanics
Z ∞ Operators as matrices: Operators can be visu-
hx|ψi = ψ(x) = dp hx|pihp|ψi alized as matrices because they satisfy matrix op-
−∞
Z ∞ erations. Hence quantum mechanics is also called
1 dp i(p/~)x matrix mechanics. For example, an operator Â
= √ e ψ(p). (81)
2π −∞ ~ acting on a state always gives another state:

This is just the inverse Fourier transform relation. Â|ψi = c|φi (87)
 10

in much the same way that a square matrix multi- Note how the operator p̂ becomes a diagonal differ-
plying a column vector always gives another column ential operator in the x-representation, and how it
vector. It can happen occasionally that the result comes out to the right of the Dirac bracket to avoid
is actually proportional to the original state |ψi operating on it. The Hamiltonian operator too has
interesting diagonal representations:
Â|ai = a|ai. (88)
hE|E ′ iE ′ = δ(E − E ′ )E ′ , if continuous

′
These special states are called the eigenstates of hE|Ĥ|E i =
hn|n′ iEn′ = δnn′ En′ , if discrete.
Â. The proportionality constants a are called their
(95)
eigenvalues. Note how the eigenvalue a is used to
label the eigenstate |ai itself. This operator equa-
tion is called an eigenvalue equation. Eigenvalue eq. as wave eq.: Matrix mechanics
Noncommuting operators: Operators, like ma- can be shown to be equivalent to wave mechanics by
trices, do not generally commute. For example, writing the operator eigenvalue equation Ĥ|ΨE i =
E|ΨE i as a Schrödinger wave equation:
~
[p̂, x̂] = . (89) Ehx|ΨE i = hx|Ĥ|ΨE i
i
Z ∞  2 
p̂
In matrix mechanics, commutators like this specify = dx′ hx| + V (x̂) x′ iΨE (x′ )
the fundamental wave nature of the matter waves −∞ 2m
" #
described by these operators. It is the starting Z ∞
1

~ d
2
′ ′ ′
point from which other results can be derived. In = dx δ(x − x ) + V (x )
−∞ 2m i dx′
contrast, the wave property of matter is described
in wave mechanics by postulating the quantization × ΨE (x′ )
rule "  2 #
1 ~ d
= + V (x) ΨE (x). (96)
~ ∂ 2m i dx
p̂ = , (90)
i ∂x
and the wave function ψ(x) on which p̂ operates. 5.5 Hermitian operators and physical observ-
Operator matrix elements: Operators like ma- ables:
trices have matrix elements. The matrix elements Hermitian operator: is one for which
of a matrix or operator M̂ can be expressed conve-
niently in the Dirac notation as Ĥ † = Ĥ T ∗ = Ĥ (itself). (97)

Mij = hi|M̂ |ji, (91) It has the important properties that

a two-sided representation in terms of a suitable (a) Eigenvalues are real.
basis |ii. A 2 × 2 matrix or operator can thus be
written in operator form as (b) Eigenstates can be orthonormalized.

M̂ = M11 |1ih1| + M12 |1ih2| Physical observables can have only real values.
+ M21 |2ih1| + M22 |2ih2|. (92) Hence they are described by Hermitian operators
in quantum mechanics.
The ket-bra combinations on the RHS are them- Dirac notation: For any operator Â:
self operators or square matrices. The off-diagonal
operators are the transition operators Z ∞
hφ|Âψi = φ† (x)[Âψ](x)dx
−∞
ψ1 φ∗1 ψ1 φ∗2
   
ψ1
|ψihφ| = (φ∗1 φ∗2 ) = , (93) Z ∞
ψ2 ψ2 φ∗1 ψ2 φ∗2 = {[† φ](x)}† ψ(x)dx
−∞
while the diagonal operators (with φ = ψ) are pro- †
= h φ|ψi. (98)
jection operators.
An operator in the representation of its own eigen- The Hermiticity of an operator Ĥ is confirmed by
states is a diagonal matrix. Thus the equality hφ|Ĥ|ψi = hĤφ|ψi of the two distinct
integrals involved. Note that if φ(x) is a column
hx|x̂|x′ i = hx|x′ ix′ = x′ δ(x − x′ ) wave function (called a spinor), the transposition
   
′ ′ ~ d ′ ~ d part of the adjoint operator will come into play to
hx|p̂|x i = hx|x i = δ(x − x ) .(94)
i dx′ i dx′ change it into a row spinor.
 11

5.6 One-dimensional harmonic oscillator: Oscillator quanta: The quantum number n (qu.
The 1DHO problem can be solved elegantly in ma- no. = any number used to label a quantum state)
trix mechanics. First write its Hamiltonian gives the number of oscillator energy quanta, each
of energy ~ω, that can be emitted from that state.
p̂2 1 Thus ↠increases the number of energy quanta in
Ĥ = + mω 2 x̂2 (99)
2m 2 the state by 1, and is called a creation operator,
while the destruction operator â decreases n by
in the compact form
1. Physical processes that causes the absorption
Ĥ 1 2 or emission of energy quanta can be expressed in
q̂ + ŷ 2 ,


ĥ = = terms of these operators.
~ω 2
x̂ The ground state which by definition cannot emit
where ŷ = ,
x0 any energy has the qu. no. n = 0. It has a
x0 1 d zero-point energy ~ω/2 that comes from the wave
q̂ = p̂ = . (100) spreading of its wave function from the classical
~ i dy
point x = 0 of stable equilibrium.
p
Here the oscillator length x0 = ~/mω is defined Wave functions: The impossibility of energy
by the equation emission from the ground state is stated by the op-
erator equation
~ω = mω 2 x20 . (101)

The dimensionless position and momentum opera- â|0i = 0. (109)

tors do not commute:
Its y-representaion gives the first-order differential
[q̂, ŷ] = 1/i. (102) equation for the ground-state (GS) wave function:

The dimensionless Hamiltonian ĥ can be factorized: 

d

y+ φ0 (y) = 0. (110)
ĥ = ↠â + 21 , dy

where â = √1 (ŷ + iq̂)

2 The resulting normalized GS wave function is
â †
= √1 (ŷ − iq̂) (103)
2 2
/2 1
φ0 (y) = α0 e−y , where α0 = . (111)
are called the step-down and step-up operators, re- π 1/4
spectively. They do not commute
 † Excited states can be created by adding energy
â, â = 1. (104) quanta to the GS:

Number operator and its eigenstates: The (↠)n

 
number operator N̂ = ↠â is Hermitian. It has real |ni = √ |0i. (112)
n!
eigenvalues n that can be used to label its eigen-
states Consequently, its wave function can be constructed
N̂ |ni = n|ni. (105) by successive differentiations of the GS wave func-
tion:
It then follows that |ni is also an energy eigenstate:  † n
(â )
φn (y) = hy| √ |0i
Ĥ|ni = (n + 21 )~ω|ni. (106) n!
 n
α0 d 2
The eigenvalue n of the number operator is a non- = √ y− e−y /2 . (113)
n
2 n! dy
negative number or integer 0, 1, 2, ... This im-
portant quantization property follows from non-
negative nature of its Hamiltonian (meaning that Operators in the number representation: In
hĤi ≥ 0) and from the nonzero commutators the number representation |ni, the only nonzero
matrix elements of â and ↠lie along the diagonal
[N̂ , ↠] = ↠, [N̂ , â] = −â. (107) one line away from the main diagonal of the matrix:
√
The commutators show that ↠and â change the √0 0 0 .
   
0 1 √0 .
state |ni as follows: 0 0 2 . † 1 √0 0 .
â =  , â =  . (114)

√ √ 0 0 0 . 0 2 0 .
↠|ni = n + 1|n + 1i, â|ni = n|n − 1i. (108) . . . . . . . .
 12

Hence the dimensionless position and momentum Ehrenfest’s theorem: The expectation value of
operators are the square matrices a physical observable satisfies a classical equation
 √  of motion. The expectation value of an operator
√0 1 √0 . Â(t) in the time-dependent state
1 1  1 0 2 .
ŷ = √ (â + ↠) = √  √ ,
2 2 0 2 0 .
.
. . . |Ψ(t)i = e−iĤt/~ |Ψ(0)i is (121)
 √ 
0
√ 1 √0 .
1 † 1 − 1 0 2 . hÂ(t)it ≡ hΨ(t)|Â(t)|Ψ(t)i
q̂ = √ (â − â ) = √  √ .
i 2 i 2 0 − 2 0 . = hΨ(0)|eiĤt/~ Â(t)e−iĤt/~ |Ψ(0)i. (122)
. . . .
(115) A time differentiation of the three time factors gen-
erates three terms:
5.7 Uncertainty principle:
The quintessential wave property described by the d ∂ Â(t) i
hÂ(t)it = h it + h [Ĥ, Â(t)] it . (123)
Heisenberg uncertainty principle takes on a sur- dt ∂t ~
prisingly elegant form in matrix mechanics: If the
Examples:
square uncertainty of the expectation value of a
physical observable  in the quantum state |ψi is d i 1
defined as hx̂it = h [Ĥ, x̂] it = hp̂it ;
dt ~ m
(∆A)2 = h(Â − Ā)2 iψ , (116) d i dV̂
hp̂it = h [Ĥ, p̂] it = −h it . (124)
dt ~ dx̂
then the product of uncertainties of two physical
observables  and B̂ in the same quantum state
satisfies the inequality VI. Quantum mechanics in three spatial
  dimensions
(∆A)(∆B) ≥ 12 h [Â, B̂] iψ  . (117)
 
Different directions in space are independent of one
another in quantum mechanics in the sense that
Simultaneous eigenstates: |a, bi exist for com-
muting operators  and B̂: [x̂i , x̂j ] = [p̂i , p̂j ] = [x̂i , p̂j ] = 0, if i 6= j. (125)
Â|a, bi = a|a, bi, B̂|a, bi = b|a, bi. (118) QM problems in three-dimensional space can be solved
Then ∆A = 0 = ∆B. by using methods of both wave and matrix mechanics.

5.8 Classical correspondence: 6.1 Separation of variables in rectangular coor-

Bohr’s correspondence principle: Quantum dinates:
mechanics reproduces classical mechanics in the If a Hamiltonian Ĥ = Ĥ1 + Ĥ2 + Ĥ3 is separable
limit of large quantum numbers. Example: The into terms referring to different dimensions in rect-
probability density of quantum state |ni of the angular coordinates, its energy eigenstates can be
1DHO agrees with the classical value: factorized into the product states |n = n1 n2 n3 i =
2 |n1 i|n2 i|n3 i, one in each spatial dimension:
lim ψnQM (x) → ρCM (x).

(119)
n→∞
Ĥ|n1 n2 n3 i = (E1 + E2 + E3 )|n1 n2 n3 i. (126)
For a periodic system of period T , the classical
probability density can be defined in terms of the The method works for other operators too:
probablity dP (x) of finding the classical point mass
within a width dx of the position x during a mo- p̂2 |p1 p2 p3 i = (p21 + p22 + p23 )|p1 p2 p3 i. (127)
ment dt in time t:
2 Examples:
dP (x) = ρCM (x)dx = dt, or
T
2 (a) 3-dim HO: En = ~ω(n+3/2), n = n1 +n2 +n3 .
ρCM (x) = , (120)
v(x)T (b) Free particle: p2 = p21 + p22 + p23 in the state
|p1 p2 p3 i.
where v(x) = dx/dt is the velocity at position x.
See Fig. 4-18, 19 of Bransden/Joachain for a com- (c) Particle in a box of sides Li : The discrete mo-
parison between quantum and classical ρs for the mentum spectrum is defined by the quantum
1DHO. numbers ni = ki Li /π = 1, 2, ....
 13

6.2 Separation in spherical coordinates: The kinetic energy operator in spherical coordi-
Kinetic energy: In the spherical coordinates r = nates is
(r, θ, φ), p̂2 separates into a radial part and an an- !
gular part: 1 2 L̂2
T̂ = T̂r + T̂L = p̂r + 2 , (136)
2m r
L̂2
p̂2 = p̂2r +,
r2   Central potential: Central potentials are spheri-
 ex ey ez  cally symmetric and depend only on the radial dis-
where L̂ = r̂ × p̂ =  x̂ ŷ ẑ 
 
 p̂x p̂y p̂z  tance from the center of potential (or force):

= L̂x ex + L̂y ey + L̂z ez (128) V = V (r) 6= f (θ, φ). (137)

is the orbital angular momentum operator. Note For central potentials, the Hamiltonian separates
that the position and momentum operators in a in the radial and angular coordinates
typical term of L̂ such as ex ŷp̂z commute because
they refer to different directions in space. However, Ĥ = (T̂r + V̂ ) + T̂L . (138)
L̂ is quite a different type of operator from r̂ or p̂,
As a result, the wave function of definite energy E
because its components L̂i do not commute with
factorizes as follows:
one another:

[L̂i , L̂j ] = i~L̂k , (i, j, k in RH order). (129) ψEℓm (r) = hr|Eℓmi = REℓ (r)Yℓm (θ, φ). (139)

On the other hand, The radial factor satisfies the Schrödinger equation
   
[L̂2 , L̂i ] = 0, for any component i. (130) 1 2 1 2
p̂ + ~ ℓ(ℓ + 1) + V (r) REℓ (r)
2m r r2
The two commuting operators L̂2 , L̂z can be used = EREℓ (r), (140)
to define the simultaneous eigenstate |ℓmi:
where
L̂2 |ℓmi = ~2 ℓ(ℓ + 1)|ℓmi,  
1 d 2 d
L̂z |ℓmi = ~m|ℓmi, (131) p̂2r = −~ 2
∇2r = −~ 2
r
r2 dr dr
 2 
where ℓ and m are called the orbital (angular mo- 1 d
= −~2 r. (141)
mentum) and magnetic quantum numbers, respec- r dr2
tively. This eigenstate describes completely an an-
gular state in spherical coordinates. The resulting Note that the wave function REℓ (r) depends on
angular wave function ℓ as well as the energy eigenvalue E. An en-
ergy eigenstate in three-dimensional space has only
hθφ|ℓmi = Yℓm (θ, φ) (132) three state labels.
Laplace equation: of electrostatics is also im-
is called a spherical harmonic.
portant in QM because its solutions describe the
In the position represention, the L̂z equation can quantum states of a static or resting free particle
be written as the differential equation (V = 0, p = 0 and E = 0):
 
~ ∂ ∇2 ψ(r) = 0. (142)
Yℓm (θ, φ) = mYℓm (θ, φ). (133)
i ∂φ
The solutions regular at the origin r = 0 are the
The solutions in the variable φ are solid spherical harmonics (or harmonic polynomi-
als)
Yℓm (θ, φ) = f (θ)eimφ . (134)
Yℓm (r) = rℓ Yℓm (θ, φ). (143)
These solutions are periodic in φ with a period of
2π: Solutions of the Laplace equations are sometimes
im(φ+2π) imφ called harmonic functions.
e =e . (135)
The solutions in rectangular coordinates are sim-
Hence m must be an integer (0, ±1, ±2, ...). It will pler. They are all rectangular polynomials with
turn out that ℓ is a nonnegative integer bounded terms of the form xn1 y n2 z n3 having the same de-
from below by |m|. gree n = n1 + n2 + n3 . All rectangular polynomials
 14

Spherical coordinates: The energy eigenstates

TABLE II: Rectangular polynomials and solid spherical har-
are
monics.
n Rect. Poly. Harmonic? Number Yℓm ? |nℓmi ∝ (↠)2ν Yℓ=n−2ν,m (↠)|000i, ν = 0, 1, ...(150)

0 1 Y 1 Y00
All the states shown in Table II with x replaced by
the operator â†1 , etc., are valid 3DHO states. The
1 z Y 3 Y1m=0 degeneracy of these states is thus
x ± iy Y m = ±1 n
1 X
D(n) = (n + 1)(n + 2) = (2ℓ + 1). (151)
2 2z 2 − x2 − y 2 Y 5 Y2m=0 2
ℓ=even or odd
z(x ± iy) Y m = ±1
(x ± iy)2 Y m = ±2 6.4 Plane wave in free space:
x2 + y 2 + z 2 N 1 —
Three-dimensional plane waves have the product
wave functions

can be written in the spherical form with definite eik·r = ei(kx x+ky y+kz z)
∞
ℓ, m: X
= Aℓ0 REℓ (r)Yℓ0 (θ, φ)
ℓ=0
 
harmonic, if ℓ = n
rn Yℓm (θ, φ) = , (144) ∞
non − harmonic, if ℓ < n X
= iℓ (2ℓ + 1)jℓ (kr)Pℓ (cos θ) (152)
ℓ=0
but only those where the power n of r is equal to
the degree ℓ of Yℓm are solutions of the Laplace where the choice k = ke3 makes explicit the axial
equation. symmetry of the wave function about the propaga-
Because ∇2 is simple in rectangular coordinates, tion direction ek .
the result The spherical Bessel function jℓ (ρ = kr) that ap-
pears is the solution of the radial equation
d2 n1
∇2 xn1 = x = n1 (n1 − 1)xn1 −2 (145) 
1 d

d

ℓ(ℓ + 1)

dx2 − 2 ρ2 + − 1 jℓ (ρ) = 0. (153)
ρ dρ dρ ρ2
can be used to determine if any rectangular poly-
nomial is a harmonic function. Examples are given It has simple properties when ρ → 0 and ∞. The
in Table II. Functions with definite m include: Legendre polynomials Pℓ (cos θ) are those used in
electrostatics.
m = ±1 : x ± iy = r sin θe±iφ 6.5 Hydrogen atom:
m = 0: z, x2 + y 2 . (146)
Similarity with the Bohr model: In wave me-
chanics, the energy expectation values will turn out
All solid spherical harmonics can be built from
to satisfy the Bohr-model relations
these two classes of functions.
U
6.3 Three-dimensional harmonic oscillator: E = T +U = = −T,
2
Rectangular coordinates: The Hamiltonian of E1
the 3DHO is separable in rectangular coordinates: En = ,
n2
e2 ~2
p̂2 1 E1 = − G, T1 = , (154)
Ĥ = + mω 2 r̂2 = Ĥ1 + Ĥ2 + Ĥ3 . (147) 2a 2ma2
2m 2
where n is the principal (or Bohr) quantum num-
Consequently, its energy eigenstates are factorable: ber, and a is the Bohr radius.
" #" #" # Differences from the Bohr model: The single
(â†1 )n1 (â†2 )n2 (â†3 )n3 circular orbit of radius an = n2 a of the Bohr model
|n1 n2 n3 i = √ √ √ |000i. (148)
n1 ! n2 ! n3 ! spreads out into states of quantum numbers n, ℓ, m,
with 0 ≤ ℓ < n. This result is consistent with the
All possible rectangular polynomials made up of spherical symmetry of the electrostatic attraction
the components â†i of the vector operator ↠can between the electron and the proton. The degen-
create a 3DHO state. Their energy eigenvalues are eracy of states is then
n−1
En = (n + 32 )~ω, where X
d(n) = (2ℓ + 1) = n2 . (155)
n = n1 + n2 + n3 = 0, 1, 2... (149) ℓ=0
 15

Wave functions: are

TABLE III: States of the hydrogen atom. The radial wave
functions are expressed in terms of the dimensionsless distance
ψnℓm (r) = Rnℓ (r)Yℓm (θ, φ). (156)
r̃ = r/a.
The radial functions Rnℓ are functions of n En ℓ m Spect.Symbol Rnℓ (r̃)

~2 κ2 1 E1 0 0 1s e−r̃
ρ = κr, where = −E. (157)
2m
2 E1 /4 1 0, ±1 2p r̃e−r̃/2
Then 0 0 2s (1 − r̃/2)e−r̃/2
unℓ (ρ)
Rnℓ (ρ) = = ρℓ e−ρ v(ρ) (158) n E1 /n2 ℓmax = n − 1 n[ℓmax ] r̃ ℓmax e−r̃/n
ρ ℓ n[ℓ]
can be written in two other forms with each func-
tion satisfying a different differential equation:
7.1 Angular momentum:
2m e2G
   
2 ℓ(ℓ + 1) 2
∇r Rnℓ (r) = − + κ Rnℓ , The orbital angular momentum operator
r2 ~2 r  
   ex ey ez 
′′ ℓ(ℓ + 1) ρ0 L̂ = r̂ × p̂ =  x̂ ŷ ẑ 
 
unℓ (ρ) = − + 1 unℓ ,
ρ2 ρ  p̂x p̂y p̂z 

= L̂x ex + L̂y ey + L̂z ez (163)

ρv ′′ (ρ) + 2(ℓ + 1 − ρ)v ′ + (ρ0 − 2ℓ − 2)v = 0. (159)
is very different from x̂ and p̂ because its compo-
where ρ0 = 2/κa and v ′ (ρ) = dv(ρ)/dρ. All solu- nents do not commute with one another:
tions have
E1 [L̂x , L̂y ] = i~L̂z , or L̂ × L̂ = i~L̂. (164)
ρ0 = 2n, giving En = , (160)
n2 However, every component L̂i commutes with
in agreement with the Bohr model.
L̂2 = L̂2x + L̂2y + L̂2z . (165)
Polynomials vν (ρ): The factor v(ρ) in Eq.(158)
is a polynomial of ρ: The angular momentum states |ℓmi are the simul-
taneous eigenstate of the two commuting operators
(a) v0 (ρ) = 1 gives solutions with ℓ = n − 1, L̂2 , L̂z :
(b) v1 (ρ) = ρ − c(a constant) gives solutions with
ℓ = n − 2, L̂2 |ℓmi = ~2 ℓ(ℓ + 1)|ℓmi,
(c) polynomials of degree ν, namely L̂z |ℓmi = ~m|ℓmi, (166)

ν
X The nature of these states can be studied by using
vν (ρ) = cj ρ j , (161) the
j=0 Ladder operators: for angular momentum states
give solutions with ℓ = n − 1 − ν. L̂± = L̂x ± iL̂y . (167)
Hydrogen states: The properties of the states of These operators have the structure of
the hydrogen atom are summarized in Table III.
The spectroscopy notation used is: x ± iy = r sin θe±iφ (168)

[ℓ] = s, p, d, f, ... for ℓ = 0, 1, 2, 3, ... (162) that in wave functions have the magnetic quan-
tum numbers m = ±1. Hence L̂± can be called
The spectroscopic symbol for a quantum state is the m = ±1 components of L̂ in spherical coor-
2s+1
[ℓ]j , giving the spin degeneracy 2s + 1, the dinates. Operators and states are both matrices,
orbital angular momentum [ℓ] and the total spin though states are not square matrices. They can
quantum number j. be multiplied together. In spherical coordinates,
the m value of a product of functions or matrices
is the sum of the m values of the individual func-
VII. Spin and statistics tions or matrices. Hence we expect that L̂± |ℓmi
The study of orbital angular momentum states leads to be a state of magnetic quantum number m ± 1,
us to intrinsic spins and quantum statistics. respectively.
 16

Before demonstrating this result, let us first note 7.2 Addition of angular momenta:
however that the ladder operators, like their con- Product states: Two angular momentum opera-
stituent rectangular components of L̂x , L̂y , com- tors L̂ and Ŝ are independent if [L̂i , Ŝj ] = 0. The
mute with L̂2 itself. Hence they do not change the simultaneous eigenstates of the commuting opera-
orbital quantum number ℓ when acting on the state tors L̂2 , L̂z , Ŝ2 , Ŝz are the product states
|ℓmi. However, they too do not commute with L̂z :
|ℓmℓ sms i = |ℓmℓ i|sms i. (178)
[L̂z , L̂± ] = ±~L̂± . (169)
These commutators can be used to show explicitly They have the degeneracy
that L̂+ is the step-up operator for m, increasing
its value by 1, while the step-down operator L̂− d(ℓ, s) = (2ℓ + 1)(2s + 1). (179)
decreases m by 1:
|φ± i = L̂± |ℓmi = Aℓm |ℓm ± 1i. (170) Coupled states: Two independent angular mo-
menta L̂ and Ŝ can be added to the total angular
Aℓm can be found by calculating the normalization momentum:
hφ± |φ± i using
Ĵ = L̂ + Ŝ. (180)
L̂2 − L̂2z = L̂2x + L̂2y = L̂∓ L̂± ± ~L̂z . (171)
The result is The four operators L̂2 , Ŝ2 , Ĵ2 , Jˆz can be shown to
p commute among themselves. Hence their simulta-
L̂± |ℓmi = ~ ℓ(ℓ + 1) − m(m ± 1) |ℓm ± 1i. (172)
neous eigenstates |ℓsjmi exist:
Possible ℓm values: m2 is bounded because
L̂2 |ℓsjmi = ~2 ℓ(ℓ + 1)|ℓsjmi,
hL̂2z i 2
< hL̂ i, 2
or m < ℓ(ℓ + 1). (173)
Ŝ2 |ℓsjmi = ~2 s(s + 1)|ℓsjmi,
Hence mmin ≤ m ≤ mmax . From Eq.(172): Ĵ2 |ℓsjmi = ~2 j(j + 1)|ℓsjmi,
L̂+ |ℓmmax i = 0 ⇒ mmax = ℓ, Jˆz |ℓsjmi = ~m|ℓsjmi. (181)
L̂− |ℓmmin i = 0 ⇒ mmin = −ℓ. (174)
So the total number of steps N on the ladder is Clebsch-Gordan coefficients: The two distinct
ways (product or coupled) of specifying the states
mmax − mmin = 2ℓ = N − 1. (175) of two angular momenta are related by a change of
representation:
Assuming that ladders with any number of steps
exist in nature, one finds !
X
N −1 |ℓsjmi = |ℓmℓ sms ihℓmℓ sms | |ℓsjmi,
Odd N = 2ℓ + 1 ⇒ ℓ= = integer, mℓ ms
2  
N −1
Even N = 2s + 1 ⇒ s= = half integer. X
2 |ℓmℓ sms i =  |ℓsjmihℓsjm| |ℓmℓ sms i, (182)
(176) j

The orbital quantum numbers ℓs are integers be- where each sum involves only one summation in-
cause they are also the integer degrees of the har- dex because mℓ + ms = m. The transformation
monic polynomials Yℓm . The half-integer s’s are (Clebsch-Gordan) coefficients are chosen real, and
the spin quantum numbers of an intrinsic spin an- therefore
gular momentum Ŝ. The word “spin” is also used
for any angular momentum. hℓsjm|ℓmℓ sms i = hℓmℓ sms |ℓsjmi. (183)
Angular-momentum representations: L̂2 , L̂z
and therefore also Lˆ ± , L̂x , L̂y are N × N matrices
Examples: Two spin 1/2 states can be coupled to
for those states with N steps in their ladder. a total spin of S = 1 or 0:
Example: For spin s = 1/2, the ladder has only
two steps, with spin states “up” and “down”. | 12 21 SM = 11i = |↑↑i, | 12 12 1 − 1i = |↓↓i,
Then the spin operators are 2 × 2 matrices: Ŝ2 = | 12 12 10i = √12 |↑↓+↓↑i;
(3/4)~2 diag(1, 1). The Pauli spin matrices are
| 21 12 00i = √1 |↑↓−↓↑i, (184)
the components of σ̂ = 2Ŝ/~: 2
     
1 0 0 1 0 −i where ↑ (↓) denotes the spin up (down) magnetic
σz = 0 −1 , σx = 1 0 , σy = i 0 . (177)
state of the spin 1/2 system.
 17

Vector model: The length J of the vector sum

TABLE IV: electronic structure of the elements. The quan-
J = L + S satisfies a triangle inequality:
tum numbers of some atomic ground states are also given as
the spectroscopic symbol 2S+1 [L]J .
|L − S| ≤ J ≤ L + S, but
Z Element Spatial Configuration Ground State
jmin = |ℓ − s| ≤ j ≤ jmax = ℓ + s, (185) 1 H 1s 2
S1/2
2 He (1s)2 1
S0
for quantum angular momenta involves the quan-
tum numbers.
2
Construction of total angular momentum 3 Li (He)2s S1/2
states: The jth ladder of states has 2j + 1 steps. 4 Be (He)(2s)2 1
S0
The topmost state m = ℓ + s of the longest ladder
jmax = ℓ + s is simply 5 B (He)(2s)2 2p 2
P1/2
6 C (He)(2s)2 (2p)2
|ℓs ℓ + s ℓ + si = |ℓℓssi. (186) 7 N (He)(2s)2 (2p)3
8 O (He)(2s)2 (2p)4
Then step down to find the remaining states on the 9 F (He)(2s)2 (2p)5 2
P3/2
10 Ne (He)(2s)2 (2p)6 1
S0
ladder:

Jˆ− |ℓsjmi = ~ j(j + 1) − m(m − 1) |ℓsjm − 1i

p
X If β = α is the same s.p. state:
= (L̂− + Ŝ− ) |ℓmℓ sms i
mℓ ms Bosons : Ψαα (1, 2) = ψα (1)ψα (2) 6= 0, but
× hℓmℓ sms |ℓsjmi, (187) Fermions : Ψαα (1, 2) = 0. (191)
where the RHS is a sum over two terms involving These important results are called:
L̂− |ℓmℓ i and Ŝ− |sms i, respectively.
Bose-Einstein condensation: The ground state
The topmost state of the next ladder with j = of a system of identical bosons is one where every
ℓ + s − 1 is the remaining or unconstructed state boson is in the lowest energy single-particle state.
with m = ℓ + s − 1. It is constructed by orthonor-
Pauli exclusion principle: No two identical
malization from |ℓs ℓ + s ℓ + s − 1i, using the prod-
fermions can populate the same single-particle
uct representation. By convention, its overall phase
state.
is taken to be the Wigner (aka Condon-Shortley)
phases. The other states on the ladder are found Many identical particles: The wave function of
by stepping down. This construction method works N identical bosons is symmetric under any per-
for all the remaining j ladders. mutation of the N particle labels. For identical
fermions, the wave function changes sign under an
7.3 Identical particles in quantum mechanics: odd permutation of the N particle labels, but re-
Spin-statistics theorem: Identical particles of mains unchanged under an even permutation.
integer (half integer) spins satisfy the Bose-Einstein
7.4 The Periodic Table:
(Fermi-Dirac) statistics. They are called bosons
(fermions). Electronic structure of the elements: is sum-
marized in Table IV:
Two identical particles are indistinguishable in
their probability densities Helium atom: Its two electrons with spins point-
ing ↑ and ↓ can populate the same 1s spatial state.
ρ(2, 1) = ρ(1, 2), (188) They complete the principal qu. no. n = 1 shell.
The shell n can accomodate 2n2 electrons.
but their wave functions may differ by a negative To get the lowest energy, both electrons in the He
sign ground state (GS) occupy the lowest s.p. spatial
 state nℓm = 100. The total wave function
bosons
Ψ(2, 1) = ±Ψ(1, 2) for . (189)
fermions Ψ0 (1, 2) = ψ100 (r1 )ψ100 (r2 )χSM =00 (1, 2)
= −Ψ0 (2, 1) (192)
Boson (Fermion) wave functions are symmetric (an-
tisymmetric) in the particle labels. must be antisymmetric in the two fermion labels 1
Examples: If α, β are single-particle (s.p.) states: and 2. Hence the spin “function” χ (not a “state”)
for the total intrinsic spin must be the antisymmet-
Ψαβ (1, 2) ∝ [ψα (1)ψβ (2) ± ψα (2)ψβ (1)] . (190) ric function with S = 0.
 18

The atomic Hamiltonian The corresponding states |nx i can be labeled by the
qu. no. nx . They form a discrete basis for wave
Ze2G Ze2G e2G
Ĥ(1, 2) = − − + (193) functions in the representative 1D box.
r1 r2 r12
Three-dimensional space: The normalized 3D
gives the GS energy (with Z = 2) wave functions in the representative cube are
E0 (He) ≈ −Z 3 ER + E12 ≈ −109 eV + 30 eV. (194) hr|ni = 1
eik·r , where ki = 2π
L ni , (198)
L3/2

where ER = 13.6 eV is the Rydberg energy. and n = n1 n2 n3 .

Closed subshells: The ms value of the spin Fermi gas: For identical fermions, each spatial
state |sms i changes when the quantization axiz ez state can be populated at most singly (the spin
changes direction. However, if all ms states are degeneracy being treated separately). Then n is
filled by 2s + 1 particles, they will remain filled called a (vector) occupation number.
when the arbitrarily chosen direction ez changes. p A free parti-
cle of mass m has energy Ek = (~k)2 c2 + m2 c4 .
Hence the total MS is 0, and this unique state of So the lowest energy state of a free Fermi gas of
2s + 1 particles has S = 0. Similarly, the complete N fermions in the cube is one where the states of
filling of the 2ℓ + 1 states of different mℓ gives a lowest k or occupation number n = |n| are popu-
unique spherical symmetric spatial system of 2ℓ + 1 lated. With spatial isotropy (spherical symmetry),
particles that has L = ML = 0. This is why the the occupation pattern is that of a sphere in n or
ground states of He, Be and Ne are all 1 S0 states. k space of radius k = kF :
Spin-orbit interaction: The atomic Hamilto- Z Z Z
nian contains a relativistic term ALS L̂ · Ŝ, ALS > 3 L 3
d3 k


N = dN = d n = 2π
0. It gives a negative (attractive) energy of
−(ALS /2)(ℓ+1) for the s.p. state with the total an-
Z kF
dN L 3 4π 3



gular momentum j = ℓ − 1/2. For the j = ℓ + 1/2 = dk dk = 2π 3 kF . (199)
0
state on the other hand, the energy is repulsive,
namely (ALS /2)ℓ. This explains why the B ground Here
state in the spatial 2p configuration has the quan-
L 3
dN
4πk 2


tum numbers 2 P1/2 of the lowest-energy single elec- dk = 2π (200)
tron state outside the closed n = 1 shell and the
closed 2s subshell. is called the density of states in k-space. The radius
of the Fermi sphere in k-space
The F ground state is one electron short of the
closed n = 2 shell. The lowest energy is obtained by N
kF = (6π 2 ρ# )1/3 , where ρ# = , (201)
putting the missing electron or hole in the higher L3
j = 3/2 state to give a 2 P3/2 ground state, i.e.,
is called the Fermi momentum. It does not depend
leaving a higher state unpopulated by populating
on the size of the cube. For electrons with their spin
all the lower-energy states in the 2p subshell.
degeneracy d(s) = 2, the total electron number is
7.5 Free Fermi gas: Ne = 2N .
Box normalization: For a particle extending over The energy per particle in a free Fermi gas is
all space, the plane-wave wave functions eik·r are R dN

not square-integrable. This feature causes awk- E Ek dk
= hEk igas = R dkdN

wardness including the appearance of Dirac δ- N dk dk
functions. The problem can be circumvented by 
(3/5)ǫF , if nonrelativistic;
subdividing space into identical cubes of sides L = (202)
(3/4)~ckF , if very relativistic.
satisfying a periodic boundary condition (BC). The
wave functions in each identical cube can then be The nonrelativistic energy
normalized.
One-dimensional space: The periodic BC ~2 kF2
ǫF = (203)
2m
ikx (x+L) ikx x
φkx (x + L) = e = φkx (x) = e (195)
at the surface k = kF of the Fermi sphere is called
gives the discrete spectrum the (NR) Fermi energy.
Chandrasekhar limit: A white dwarf star con-
kx L = 2πnx , with nx = 0, ±1, ±2, ... (196)
tains Z protons, Z electrons and A − Z neutrons.
and the normalized wave functions The main contributions to its energy are the repul-
sive kinetic energy of the highly relativistic elec-
φkx (x) = √1 eikx x . (197) tron gas and the attractive gravitational potential
L
 19

energy of its mass M . Both energies are inversely The equation is solved by equating terms of the
proportional to R, the radius of the star, but they same power of λ on the two sides of the equation:
have different M dependences:
 λ0 : Ĥ0 ψn(0) = En(0) ψn(0) ⇒ ψn(0) = φn ,
4/3 2  > 0, unbound,
M M λ1 : En(1) = Hnn
′
,
ETot = b −c = 0, critical,
R R  < 0, collapse. X ′
Hkn
ψn(1) = − (0) (0)
φk ,
(204) k6=n Ek − En
′
The white dwarf will collapse gravitationally if its where Hkn = hk|Ĥ ′ |ni. (210)
mass exceeds the critical mass
 3/2 (1)
b 8.2 Calculation of En :
Mc = . (205)
c Many perturbations are made up of sums of fac-
torable terms of the type
VIII. Approximation methods for time-
Ĥ ′ = Q(r)Â (211)
independent problems
Few problems in quantum mechanics are exactly solv- involving a radial factor Q(r) and an angle-spin fac-
able. We first consider some approximation methods for tor  for a system with intrinsic spin S. Then
Hamiltonians that are time-independent.
8.1 Time-independent perturbation theory: En(1) = Qnℓ hÂiα ,
Z ∞
2 × 2 matrices: can be diagonalized exactly. The where Qnℓ = |Rnℓ (r)|2 Q(r)r2 dr,
Hermitian matrix 0
  α = ℓmℓ sms or ℓsjm. (212)
0 b
H= b a , (206)
Examples:
where both a and b are real, has the eigenvalues
2
and orthonormal eigenvectors hL̂ · Ŝi = 1
2 hĴ − L̂2 − Ŝ2 iℓsjm ,
E1,2 = 1 hL̂z + 2Ŝz iℓmℓ sms = (mℓ + 2ms )~,
2 (a ∓ ∆E),
b2 ~m
hŜz iℓsjm = 2 hĴ · Ŝiℓsjm . (213)
p
∆E = a2 + 4b2 ≈ a + 2 ; ~ j(j + 1)
  a 
cos θ − sin θ
ψ1 = sin θ
, ψ2 = cos θ
, 8.3 Variational method:
E1 b Variational principle: For any wave function ψ:
tan θ = ≈− . (207)
b a
hψ|Ĥ|ψi
Many problems of physical interest involving only Eψ = hĤiψ ≡ ≥ E0 (GS). (214)
hψ|ψi
two dominant states can be solved approximately
by this method. The approximate wave function ψ gives a higher
Perturbation theory: Let the Hamiltonian Ĥ = energy when it contains excited-state components
Ĥ0 + Ĥ ′ contains a main part Ĥ0 whose energy of higher energies.
(0)
eigenvalues En , eigenfunctions φn or eigenstates Variational wave function: is any wave func-
|ni are already known. If a complicated but weak tion carrying a number of parameters b that can be
perturbation Ĥ ′ is now added to the system, the varied in order to minimize the energy
eigenvalue equation
hψ(b)|Ĥ|ψ(b)i
(Ĥ0 + λĤ ′ )ψn = En ψn (208) E(b) = . (215)
hψ(b)|ψ(b)i
can be solved by a systematic expansion in powers
of λ that can be set back to its numerical value of The energy minimum E(b0 ) where dE(b)/db = 0 is
1 at the end of the expansion. the best variational estimate because it is closest to
This λ expansion is done in both En and ψn : the true ground-state (GS) energy E0 .
  He atom: The ee repulsion between the two elec-
(Ĥ0 + λĤ ′ ) ψn(0) + λψn(1) + ... trons in the He atom can be eliminated approxi-
   mately by reducing the actual nuclear charge Z = 2
= En(0) + λEn(1) + ... ψn(0) + λψn(1) + ... . (209) to the effective charge z ≈ 27/16.
 20

Hartree mean-field screening: In an atom with where |ψi i are the time-independent eigenstates of
Z electrons, the mutual ee interactions can be re- Ĥ. Then it evolves at time t into
placed by a mean field experienced by each electron
characterized by an effective nuclear charge of |Ψ(t)i = e−iĤt/~ |ψi = d1 |Ψ1 (t)i + d2 |Ψ2 (t)i
 = d1 e−iE1 t/~ |ψ1 i + d2 e−iE2 t/~ |ψ2 i. (221)
Z, for r → 0
z(r) = (216)
1, for r → ∞
Unperturbed basis: It is often useful to express
|Ψ(t)i in terms of the time-dependent unperturbed
8.4 WKB approximation: basis states
Since any complex number can be written in the
 
(0)
−iE1 t/~ 1
amplitude-phase form Aeiφ , one can look for a so- |Φ1 (t)i = e 0
,
lution of the 1D Schrödinger wave equation in this  
(0) 0
amplitude-phase form. The WKB approximation |Φ2 (t)i = e−iE2 t/~ 1 (222)
is a semi-classical result used between the two clas-
sical turning points x1 and x2 of the potential V (x) (0) (0)
where the local wave number of energies E1 = 0 and E2 = a:
|Ψ(t)i = c1 (t)|Φ1 (t)i + c2 (t)|Φ2 (t)i. (223)
r
2m
k(x) = [E − V (x)] (217)
~2 The probabilities |ci (t)|2 can be found by using the
is real. It is obtained by ignoring the A′′ (x) = energy eigenstates
d2 A(x)/dx2 term. Then 
cos θ
 
− sin θ

ψ1 = sin θ , ψ2 = cos θ
(224)
ψ(x) = A(x)eiφ(x)
C ±i x k(x′ )dx′
R
in Eq.(221):
≈ ψW KB (x) = p e x1 . (218)
k(x) 2
|c1 (t)|2 = d1 cos θ − d2 e−i∆ωt sin θ ,


WKB quantization: At a turning point where 2

|c2 (t)|2 = d1 sin θ + d2 e−i∆ωt cos θ ,

(225)
the potential has a finite slope, the wave func-
tion can penetrate into the classically forbidden re- where ∆ω = (E2 − E1 )/~. These probabilities are
gion where the local wave number becomes purely in general functions of t. Indeed, their time depen-
imaginary k → iκ. Under this analytic continua- dences betray the fact that the φi are not energy
√ the amplitude
tion, √ factor becomes complex, A ∝ eigenstates, for otherwise they will be stationary
1/ k → e−iπ/4 / κ, thereby contributing an addi- states with time-indepedent probabilites.
tional phase of −π/4. The resulting WKB-modified
The expansion (223) is particularly interesting
Bohr-Sommerfeld quantization condition is
when the basis states |φi i are themselves physically
Z x2  m observable, as happens in
k(x)dx = n − π, n = 1, 2, ..., (219)
x1 4 Neutrino oscillations: Neutrinos come in three
distinct varieties or “flavors”: νe , νµ , ντ , associated
where m is the number of turning points where the with the electron e, muon µ and tau τ , respec-
potential is not an ∞ wall. The resulting WKB tively. Although they are physically distinct from
energies are often quite good. They even agree one another, they are not mass eigenstates (energy
with the exact energies in the case of the one- eigenstates at zero momentum). The mass eigen-
dimensional harmonic oscillator potentials. states are linear combinations of these neutrino ba-
sis states of different flavors. For this reason, the
IX. Approximation methods for time- flavor probabilities are not constants of motion, but
dependent problems instead they oscillate in time.
Time-dependent problems are more complicated partly Solar neutrino oscillations: The neutrinos ini-
because one has to keep track of the time evolution of a tially produced in the solar interior where masses
state, but mostly because the Hamiltonian itself might are converted to energy are νe = φ1 , and not
be time-dependent. νµ = φ2 , i.e., c2 (t = 0) = d1 sin θ + d2 cos θ = 0,
or d1 = cos θ, d2 = − sin θ. Allowing only oscil-
9.1 The time-dependent two-state problem: lation into νµ for simplicity, the probabilities for
is exactly solvable for the time-independent Hamil- finding νe and νµ at a later time t are
tonian (206). Let a state be prepared initially as
|c1 (t)|2 = 1 − sin2 2θ sin2 (∆ωt/2),
|ψi ≡ |Ψ(t = 0)i = d1 |ψ1 i + d2 |ψ2 i, (220) |c2 (t)|2 = sin2 2θ sin2 (∆ωt/2). (226)
 21

The neutrinos may have very small masses, so that (TD) perturbation Ĥ ′ . Unfortunately, the TD ex-
pansion coefficients cn (t) must be obtained more
m2i c3 laboriously by actually solving the TDSchEq
Ei ≈ pi c + . (227)
2pi
∂
i~ |Ψ(t)i = (Ĥ0 + λĤ ′ )|Ψ(t)i. (234)
The the probability for finding νe of energy E at a ∂t
distance L = ct from the Sun is
The resulting differential equations (DEs) of mo-
2 2 2 2 4 tion for ck (t),
|c1 (t)| = 1 − sin 2θ sin [A(∆m c )L/E];
∆m2 = |m21 − m22 |,
X
′ iωbk t
i~ċb (t) = λ Hbk e ck (t), (235)
1 k
A = = 1.27 GeV(eV)−2 km−1 . (228)
4~c
are a set of coupled first-order DEs that can be
(0) (0)
Experimental data for solar neutrinos of energies solved easily in a PC. Here ωbk = (Eb − Ek )/~.
broadly distributed around a mean value of E =
Systematic TD perturbation theory: is con-
3 MeV give
cerned with the Taylor expansion in λ (eventually
set to 1):
(∆m2 )c4 = 7.3 × 10−5 (eV)2 , tan2 θ = 0.41. (229)
(0) (1)
cb (t) = cb (t) + λcb + ... (236)
9.2 Time-dependent perturbation theory:
Eq.(235) can then be separated term by term into
Expansion in exact eigenstates: Suppose Ĥ 6=
separate equations for cb(n) :
Ĥ(t), and its energy eigenvalues Ej and eigenstates
|ψj i are known. Then any arbitrary quantum state (0) (0)
at time t can be expanded in terms of these energy λ0 terms : i~ċb (t) = 0 ⇒ cb (t) = const,
(1) ′ iωbk t (0)
X
eigenstates: λ1 terms : i~ċb (t) = Hbk e ck , ... (237)
k
N
X X
|Ψ(t)i = dj |Ψj (t)i = dj e−iEj t/~ |ψj i. (230) These first-order DEs in time can be integrated di-
j=1 j rectly, making it easy to calculate first-order wave
functions in TD perturbation theory.
Expansion in basis states: This same state can Sinusoidal perturbation: Let the sinusoidal per-
also be expanded in terms of a known set of basis turbation
(0)
states |Φn (t)i = e−iEn |φn i that are eigenstates of
(0) Ĥ ′ = V̂ cos ωt, 0 ≤ t ≤ t0 only, (238)
a known Hamiltonian Ĥ0 of energies En :
X X (0)
be turned on for a finite period of time. After the
t/~
|Ψ(t)i = cn (t)|Φn (t)i = cn (t)e−iEn |φn i.(231) perturbation has ceased, a state initially in |φa i
n n will find itself in the state |φb i, b 6= a, with the
probability
It happens occasionally that both |ψi and |φi states
are physically observable. Then it is of interest to  Vba 2 2 2
2  
 j0 (z) t0 ,

 (1)
cb (t ≥ t0 ) ≈  (239)

express one observational probabilities in terms of

2  ~2
the other expansion coefficients. For example,
 2 where z = 12 (ωba ± ω)t0 ,
N 
X j0 (z) = sin(z)/z,
|cn (t)|2 Unj dj e−iEj t/~  ,

=  (232) (0)
j=1  ωba = [Eb − Ea(0) ]/~. (240)

is a generalization of what is done for neutrino os- Thus most of the excitations are close to the two
(0) (0)
cillations to higher dimensions. Here U , with states |φb i with energies Eb = Ea ± ~ω.
Golden rule: The transition rate during the in-
Unj = hφn |ψj i, (233) teraction lasting a time t0 from a state a to a group
of states b is thus
is an N ×N unitary matrix if both sets of states are
1 ∞ (1) 2
Z
normalized. U is then made up of the eigenstates Wba ≡ |c | ρb d(~ωba )
|ψj i stored columnwise. t0 −∞ b
2
TD perturbation: The expansion (231) can also

2π  Vba 
≈ ρb , (241)
be used if Ĥ = Ĥ0 + Ĥ ′ contains a time-dependent ~  2 
 22

where ρb = dNb /dEb is the density of final states. The new features include the appearance of a ge-
Example: The Bohr transition of an atomic elec- ometrical phase ϕgeo
n and of transitions to other
tron from a higher atomic state b to a lower state states k 6= n.
a with the emission of a photon of energy ~ω = Impulsive perturbation: An impulsive pertur-
(0) (0)
Eb − Ea is caused by the electric dipole interac- bation contains a δ-function in time:
tion
Ĥ ′ (x, t) = Â(x)δ(t). (249)
Ĥ ′ = V̂ cos(ωt). (242)
The TDSchEQ can be solved exactly to show that
Here V̂ = −ǫ· D̂E0 is the electric-dipole interaction the δ-function makes the state discontinuous at t =
between the electric dipole operator D̂ = qr̂ of the 0:
atomic electron and the electric field E = ǫE0 of
the emitted photon. The latter can be obtained !−1 !
 Â
from the photon energy density in a cube of side L: |Ψ(0+ )i = 1 − 1+ |Ψ(0− )i, (250)
2i~ 2i~
~ω ǫ0 E02
3
= , (243)
L 2
where ǫ0 is the permittivity of free space. When
X. Applications
Ĥ ′ is used in the Golden Rule, the resulting transi- In this chapter, we see how quantum problems of great
tion rate is the Einstein coefficient for spontaneous physical interest have been solved exactly or approxi-
photon emission by the atom: mately.

4 ω3
  
1 10.1 Solving special matrix eigenvalue problems:
A = Wba = |Dba |2 , (244)
3 c3 ~ 4πǫ0
Special matrices of arbitrarily large dimensions of
9.3 Sudden, adiabatic and impulsive perturba- great physical interest can sometimes be solved an-
tions: alytically by symmetry considerations.
In sudden perturbation, the Hamiltonian To begin, it is easy to verify by inspection the eigen-
changes suddenly by a finite amount to the final values and eigenvectors of the Pauli matrix σx :
value Ĥa . If the change is so sudden that the state     
|ψi at t = 0 has not changed, the state at times 0 1 1 1
= ,
t > 0 is just 1 0 1 1
    
0 1 1 1
|Ψ(t)i = e−iĤa t/~ |ψi = − . (251)
1 0 −1 −1
X (a)
= e−iEn t/~ |φn ihφn |ψi
n
Indeed, three simple rules often go a long way in
X finding the eigenvectors and eigenvalues of certain
= |Φn (t)ihφn |ψi. (245) Hermitian matrices:
n
(a) The vector v1 = (1, 1, ..., 1), where all compo-
Adiabatic perturbation: When a time-
nents are 1, is an eigenvector of a matrix where
dependent Hamiltonian Ĥ(t) changes very slowly
the matrix elements of each row is a permu-
in time, we expect the instantaneous eigenvalues
(t) (t) tation of those of the first row. Its eigenvalue
En (t) and eigenstates |ψn i defined by the eigen- E1 is just the sum of the matrix elements of
value equation any row.
Ĥ(t)|ψn(t) i = En(t) (t)|ψn(t) i (246) (b) Eigenvectors of a Hermitian matrix can be
made othogonal to one another.
to give a good approximation. That is
(c) The sum of eigenvalues is equal to the trace of
En (t) ≈ En(t) (t), (247) the matrix.
a result known as the adiabatic theorem. Example 1: The vector v1 = (1, 1, 1) is an eigen-
If one watches the system for a sufficiently long vector of the 3 × 3 matrix ((0,1,1),(1,0,1),(1,1,0)),
time, however, sooner or later additional features with eigenvalue E1 = 2. The three vectors v2 =
will appear. They can be described conveniently (2, −1, −1), v3 = (−1, 2, −1), v4 = (−1, −1, 2) are
(t)
by using the time-dependent basis |ψk i: all orthogonal to v1 , but not linearly independent
R t (t′ ) ′
of one another. Each has the eigenvalue -1. The
(t)
X ′
|Ψ(t)i = ck (t)e−(i/~) 0 Ek (t )dt |ψk i. (248) orthogonal eigenvectors can be taken to be v1 , v2 ,
k and v5 = (v3 − v4 )/3 = (0, 1, −1).
 23

Example 2: The vectors v1 , v2 and v5 are also Hence Ĥ ′ has nonzero matrix elements only be-
the orthogonal eigenvectors of the Hermitian ma- tween two opposite-parity states i, j with the same
trix ((1,1,1),(1,1,1),(1,1,1)), with the eigenvalues 3, magnetic quantum number: mi = mj .
0, and 0, respectively. This is a special case of the Example: In the n = 2 shell of the hydrogen atom,
general result that two matrices M1 and M2 have there are four states: nℓm = 200; 210, 21 ± 1. Ĥ ′
common eigenvectors if connects only the two opposite-parity m = 0 states.
For these two states alone, it takes the form of a
M2 = M1 + aI, (252) 2 × 2 matrix b((01), (1, 0)) whose eigenstates are
where I is the identity matrix. Since the parame- E1,2 = ∓b : ψ1,2 = √1 (φ200 ± φ210 ) . (256)
2
ter a can be complex, we see that non-Hermitian
matrices can also have orthogonal eigenvectors. Their energies E1,2 thus move away from the undis-
Example 3: The arbitrarily large N × N matrix turbed energy of the two states with m = ±1 by an
whose matrix elements are all 1’s has one “collec- amount proportional to b, which is proportional to
tive” eigenvector v1 = (1, 1, ..., 1) with eigenvalue E. [The state |ℓmi has parity (−1)ℓ .]
N . The remaining eigenvalues are all 0. The or- 10.3 Aharonov-Bohm effect:
thogonal eigenvectors can be chosen to be v2 =
(N − 1, −1, −1, ..., −1), v3 = (0, N − 2, −1, ..., −1), Gauge transformation: In classical electromag-
..., vN = (0, 0, ..., 0, 1, −1). netism (EM) the electric and magnetic fields are
uniquely defined, but the EM potential (the scalar
BCS theory of superconductivity: can be un- potential ϕ and vector potential A) are not unique.
derstood conceptually by using a large dimensional They can change by a gauge transformation
N × N matrix Hamiltonian with the same small
negative matrix element −ǫ everywhere. All eigen- A → A′ = A + ∇Λ,
states have energy 0, except the collective eigen- ∂Λ
state v1 = (1, 1, ..., 1) that has energy ∆ = −ǫN . ϕ → ϕ′ = ϕ − , (257)
∂t
The energy gap ∆ that separates the ground state
from the excited states makes it hard for the sys- where Λ is any scalar field, without changing the
tem to be excited. If the ground state contains EM fields
current-carrying electrons, then at sufficiently low
B = ∇ × A,
temperatures (below a certain critical temperature
Tc ), these electrons cannot lose energy by inelastic ∂A
E = −∇ϕ − . (258)
collisions with the crystal lattice in the conductor. ∂t
When this happens, both dissipation and resistiv-
Relativistic electrodynamics: The basic dy-
ity vanish, and the medium becomes superconduct-
namical variable in classical mechanics is the 4-
ing. (BCS = Bardeen, Cooper and Schrieffer who
momentum (p, iE/c). To add EM, one needs an
discovered this fundamental theory of superconduc-
EM 4-vector. There is a unique qualifying candi-
tivity.)
date: the EM 4-potential (A, iϕ/c). Hence electro-
10.2 Stark effect in an external electric field: dynamics for a particle of mass m and charge q can
be built up from the sum (p − qA, i(E − qϕ)/c).
A charge q moved a distance z against a constant This combination transforms like a 4-vector un-
external electric field E = Eez has the (dipole) in- der Lorentz transformations, thus guaranteeing the
teraction energy correct result in different inertial frames. As a re-
Z sult, the EM 4-potential plays a more fundamental
′
E = −q E.dr = −qEz. (253) role in quantum mechanics than the EM fields.
The energy-momentum relation for relativistic elec-
The resulting perturbing Hamiltonian is trodynamics is then

Ĥ ′ = −E D̂z , where D̂ = qr̂ (254) (E − qϕ)2 = (p − qA)2 c2 + m2 c4 . (259)

is called the electric dipole operator of the charge. One can find from this the nonrelativistic kinetic
This operator is odd in parity, changing sign when energy TN R :
z → −z. Hence all first-order energies vanish: T ≡ E − mc2
1
En(1) = hn|Ĥ ′ |ni = 0. (255) ≈ TN R = qϕ + (p − qA)2 . (260)
2m

Since z = r cos θ is independent of the azimuthal Quantum electrodynamics: We are now in a

angle φ, D̂z has magnetic quantum number m = 0. position to quantize the mechanical terms in NR
 24

electrodynamics, leaving the EM potential unquan- after one complete circuit. Hence its Aharonov-
tized for simplicity: Bohm phase can only be an integral multiple of 2π:

1 qΦ
Ĥ = (p̂ − qA)2 + V̂ + qϕ. (261) g = 2πn = , or Φ = nΦ0 . (268)
2m ~
As a result, the magnetic flux enclosed by c can
Aharonov-Bohm phase: The EM interaction re- only exist in integral multiples of a quantized unit
sides primarily in the phase (or gauge) of quantum flux Φ0 = h/q. Experimental measurement deter-
wave functions. For the special case where ϕ = 0, mines that q = −2e in superconductors, thus show-
this important result can be demonstrated readily ing that the super current is carried by pairs of
by showing that the wave function ψ in the pres- electrons, now called Cooper pairs.
ence of the vector potential A differs from the wave
function ψ ′ without A by an A-dependent phase g: Gauge interactions: Other fundamental inter-
actions also work through the phase of quantum
ψ(r) = eig ψ ′ (r). (262) wave functions. Hence they are called gauge inter-
actions. Indeed, the information contained in the
If true, the phase factor eig serves the purpose of phase of a wave function is usually more important
removing the A dependence from the Schrödinger than that contained in its amplitude. A picture of
equation when moved to the left of a “gauged” fac- Snow White can be made to appear in our eye as
tor: a coherent wave. One can take the phase informa-
    tion from Snow White’s wave function and use it
~ ig ′ ig ~ with the amplitude part of the wave function from
∇ − qA e ψ (r) = e ∇ ψ ′ (r). (263)
i i a picture of Happy, one of the seven dwarfs. A
composite wave like this has been constructed the-
Since oretically. Believe it or not, one sees Snow White
    in the resulting hybrid picture, and not the dwarf.
~ ~
∇ eig ψ ′ (r) = eig ~(∇g) + ∇ ψ ′ (r), (264) 10.4 Magnetic resonance:
i i
Rabi measured the magnetic moments of atomic
g has to satisfy the differential equation states by finding the Zeeman splitting of the en-
ergies of different magnetic substates by resonance
~∇g = qA, or ~dg = qA · dr. (265) matching. Atomic magnetic moments are tradition-
ally expressed as gµB in units of the Bohr magneton
Hence µB = e~/2m, the theoretically value g being called
q
Z r the Landè g factor. The Hamiltonian involved is
g= A(r′ ) · dr′ , (266)
~ r0 Ĥ = Ĥ0 + Ĥ ′ , (269)
where the line integral is in general path-dependent. where Ĥ0 = Ĥ1 + ĤLS + Ĥ2 contains the usual
The integral is then said to be non-integrable. This NR Hamiltonian Ĥ1 , the relativistic spin-orbit term
non-integrability can be made explicit by integrat- ĤLS , and the Zeeman Hamiltonian
ing around a closed circuit c once and simplifying
the result with the help of Stokes’s theorem: Ĥ2 = −M̂ · Bz
gµB Bz ˆ
q Jz = ω0 Jˆz
I
= (270)
g = A(r′ ) · dr′ ~
~ c
q
Z
q that gives the splitting between the magnetic sub-
= (∇ × A) · dS = Φ, (267) states of the same j. Finally a small oscillatory
~ S ~
magnetic field is applied in the x-direction to give
where the total magnetic flux Φ across the surface a perturbing Hamiltonian
S enclosed by the circuit c depends on the magnetic
field ∇ × A on and inside c. Such a dependence on Ĥ ′ = ωx Jˆx cos(ωt), where ~ωx = gµB Bx . (271)
properties of the system outside the path c shows This perturbation causes transitions between the
why quantum mechanics in particular, and waves magnetic substates. Any quantum state
more generally, describe nonlocal phenomena that
are spread out in space.
X
|Ψ(t)i = cm (t)|Φm (t)i
Magnetic flux quantization: Suppose the path m
c is located in a region free of any EM field, then
X
= cm (t)e−iωm t |φm i, (272)
the wave function must return to its original value m
 25

(0)
where ωm = Em /~, can be expressed in term of the is the differential solid angle in a suitable inertial
unperturbed states |Φm (t)i. The time-dependent frame.
expansion coefficient cm (t) then satisfies the TD- Example: Two beams of the same cross sectional
SchEq area A collide head-on where their paths cross each
X other. Each beam is made up of bunches of Ni (i =
i~ċm (t) = ′
Hmk eiωmk t ck (t), (273) 1, 2) particles per bunch. The target can be taken
k to be one bunch of N2 particles in beam 2. Then
where ωmk = ωm − ωk . the effective incident flux is I = f N1 /A, where f is
the frequency of collision between bunches.
Example: The problem can be solved approxi-
mately for cm (t) for the atomic n[ℓ]j = 2p1/2 con- Rutherford cross section: For the scattering of
figuration that has only two magnetic substates φ± α-particles of charge Z1 = 2 from an atomic nu-
of energies E±(0)
= ±~ω0 /2. For example, if the sys- cleus of charge Z2 , Rutherford found from classical
tem is initially in the upper state (c+ = 1, c− = 0), mechanics that in the CM frame
then the probability of finding the system in the dσ
 2
d 1
upper and lower states at a later time t are respec- = 4 ,
dΩ 4 sin (θ/2)
tively:
Z1 Z2 e2G
P++ = |c+ |2 = cos2 (ωR t/2) + sin2 χ sin2 (ωR t/2), where d = , (277)
TCM
P−+ = |c− |2 = cos2 χ sin2 (ωR t/2), where (274)
is the distance of closest approach (or turning
p point) for head-on collision at the NR kinetic en-
ωR = (∆ω)2 + (ωx /2)2 , ∆ω = ω0 − ω, ergy TCM in the CM frame. Note that for backscat-
∆ω ωx /2 tering θ = π, the Rutherford differential cross sec-
sin χ = , cos χ = . (275) tion d2 /16 gives a direct measurement of the posi-
ωR ωR
tion d of the classical turning point.
Magnetic resonance: As the applied frequency Total cross section: is proportional to the total
ω passes the Zeeman frequency ω0 , the minimum reaction rate:
population of the φ+ state falls sharply to 0 at ω0 Z Z
dσ
Z
dR R
and then rises sharply back up to ≈ 1 again, thus σ = dσ = dΩ = = . (278)
giving a clear signal for the measurement of ω0 . dΩ L L
The third frequency ωx controls the period τ of
oscillation of the population between the two mag-
netic substates. At full resonance (ω = ω0 ), one 11.2 Quantum theory of scattering in a nutshell:
finds τ = 4π/ωx . Partial-wave expansion: of a plane wave into
spherical waves of good angular momentum ℓ
around the origin of coordinates has the Rayleigh
XI. Scattering theory form:
Information about the dynamical properties of micro- ∞
X
scopic systems like atoms, nuclei and particles can be ψk (r) = eik·r = iℓ (2ℓ + 1)Pℓ (cos θ)jℓ (kr),
obtained by scattering beams of projectiles from targets ℓ=0
containing them, as first demonstrated by Rutherford, ∞
X i
Geiger and Marsden who elucidated atomic structure by ∼ iℓ (2ℓ + 1)Pℓ (cos θ)
scattering α particles from a gold foil.
r→∞
2k
ℓ=0
e−i(kr−ℓπ/2) ei(kr−ℓπ/2)
 
11.1 The scattering cross sections: × − (279)
r r
The differential cross section
This is the time-independent wave function of a sta-
dσ 1 dR
= (276) tionary scattering state. The time-dependent wave
dΩ L dΩ function carries an additional time factor e−iωt ,
is the angular distribution dR/dΩ of the reaction where ω = ~k 2 /2m. The additional time factor
rate R per unit luminosity L. L itself is the product shows that the wave function e−ikr /r describes an
INT , where I is the flux or current density of the ingoing spherical wave collapsing towards the ori-
incident beam (number of incoming particles per gin, while the wave function eikr /r describes an
second per cross sectional area of the beam, or Ṅi /a outgoing spherical wave expanding out from the ori-
for a beam of uniform cross section a) and NT is gin.
the number of target particles illuminated by this Scattered wave: When a target particle is placed
beam. L depends on both beam and target. dΩ at the origin of coordinates, the radial part jℓ of the
 26

wave function (279) will be replaced by the func- Total cross section: is the result integrated over
tion Rℓ (r) from the SchEq with a potential in it. At the solid angle dΩ:
large distances r, the ingoing waves are still collaps- Z  
dσ
Z
ing towards the origin, and therefore do not know if σtot = dΩ = f ∗ (θ)f (θ)dΩ
a potential is present there. In elastic scatterings, dΩ
an outgoing spherical wave has the same normal-
X
= σtot,ℓ , (286)
ization as before, and can be changed at most by a ℓ
phase, here taken to be 2δℓ , that resides in a phase
factor Sℓ = e2iδℓ called an S-matrix element: where the two sums over ℓ, one from each f , has
been simplified to only one sum by using the or-
∞
X i thogonality relation for Legendre polynomilas
ψ(r) ∼ iℓ (2ℓ + 1)Pℓ (cos θ)
r→∞
2k 1
4π
Z
ℓ=0
 −i(kr−ℓπ/2) i(kr−ℓπ/2)
 Pℓ (cos θ)Pℓ′ (cos θ)dΩ = δℓℓ′ , and (287)
e e −1 2ℓ + 1
× − Sℓ
r r 4π
σtot,ℓ = (2ℓ + 1) sin2 δℓ . (288)
= ψk (r) + ψsc (r) where (280) k2

Optical theorem: If δℓ is real, one finds that

∞
X
ψsc (r) ∼ iℓ (2ℓ + 1)Pℓ (cos θ) X X sin2 δℓ
r→∞ Imf (θ = 0) = (2ℓ + 1)Imfℓ = (2ℓ + 1)
ℓ=0 k
ℓ ℓ
ei(kr−ℓπ/2)
 
Sℓ − 1 k
× = σtot . (289)
2ik r 4π
ikr
e
= f (θ) . (281) This theorem holds even when the phase shift δℓ =
r αℓ + iβℓ becomes complex.
is the scattered wave function, the additional wave
function generated by the potential. Containing
only outgoing spherical waves, it can be written 11.3 Breit-Wigner resonance:
asymptotically (r → ∞) in terms of an angle- The partial-wave total cross section σℓ =
dependent quantity called the scattering amplitude σℓ, max sin2 θℓ reaches a maximum value of σℓ, max =
(4π/k 2 )(2ℓ + 1) whenever δℓ (k) = π/2. This maxi-
∞
X mum value is called the unitarity limit.
f (θ) = (2ℓ + 1)fℓ Pℓ (cos θ),
ℓ=0
If δℓ (E) rises through δℓ (E = ER ) = π/2 sifficiently
Sℓ − 1 1 rapidly, a Taylor expansion about E = ER needs
where fℓ = = eiδℓ sin δℓ (282) to be taken only to the terms
2ik k

π dδℓ 
if the scattering is elastic, meaning that the phase δℓ ≈ + (E − ER ) . (290)
shift δℓ is real. 2 dE ER
Flux: or probability current density is Γ/2
Then tan δℓ ≈ − , where
E − ER
~ Γ 1
j= (ψ ∗ ∇ψ − ψ∇ψ ∗ ) ≈ jk + jsc , (283) = > 0. (291)
2im 2 (dδℓ /dE)ER
where jk = ~k/m is the flux of the plane wave, and Resonance: In the neighborhood of a resonance,
2 the partial-wave total cross section has the simple
~k |f (θ)| E-dependence:
jsc = er (284)
m r2
tan2 δℓ
is the flux of the scattered wave at large distances. σℓ = σℓ, max
1 + tan2 δℓ
Differential cross section: A detector of area (Γ/2)2
dA = r2 dΩ then measures the differential cross sec- = σℓ, max . (292)
(E − ER )2 + (Γ/2)2
tion
This function has a sharp maximum at E = ER and
dσ 1 dR 1 dA
= = (jsc · er ) falls rapidly to half its maximal value at E − ER =
dΩ Lk dΩ ~k/m dΩ ±Γ/2. Hence Γ is called the resonance width (or
= |f (θ)|2 . (285) full width at half maximum, FWHM).
 27

Decaying state: The partial-wave scattering am- The photon in question cannot propagate freely. It
plitude near resonance has the simple form can only exist momentarily around its point source
(the point charge), and is then said to be a virtual
1 iδℓ 1 1
fℓ = e sin δℓ = photon. In Feynman’s diagrammatic language, the
k k cot δℓ − i Coulomb interaction arises from the emission of a
1 −Γ/2 virtual photon by a source (the q here) and its ab-
= . (293)
k (E − ER ) + iΓ/2 sorption by a test charge at distance r from it, or
vice versa.
At the complex energy E = ER − iΓ/2, the time-
The virtual particle in Yukawa’s theory of inter-
dependent probability density decays exponentially
actions is a boson, otherwise it cannot be emitted
|Ψ(r, t)|2 = |ψ(r)e−i(ER −iΓ/2)t/~ |2 because of the conservation of fermion number. A
virtual boson that is massive satisfies the Yukawa
= |ψ|2 e−Γt/~ . (294) equation
(∇2 − µ2 )φ(r) = −4πgδ(r), (299)
11.4 Yukawa’s theory of interactions: where the unit of charge is now denoted g, and µ =
Yukawa showed in 1935 that both the Coulomb mc/~ is the inverse reduced Compton wavelength.
interaction mediated by the exchange of massless The solution of this DE that vanishes at r = ∞ can
photons between charges, and the strong or nuclear be shown to be the Yukawa potential
interaction mediated by the exchange of massive g
φ(r) = e−µr . (300)
bosons called mesons, can be derived from quan- r
tum mechanics.
Nuclear forces: were known in 1937 to have a
Klein-Gordon equation: The Einstein energy- range of 1/µ ≈ 1.4 fm. (fm = 10−15 m.) This
momentum relation E 2 = p2 c2 +m2 c4 can be quan- fact allowed Yukawa to predict that these forces
tized into the Klein-Gordon equation in free space: are caused by the exchange of bosons of mass
Ĥ 2 Φ(r, t) = p̂2 c2 + m2 c4 Φ(r, t),


(295) m = µ~/c ≈ 140 MeV/c2 . Such strongly inter-
acting bosons, now called π mesons or pions, were
where Ĥ = i~∂/∂t and p̂ = (~/i)∇. The result- discovered in 1947.
ing wave functions Φ oscillates in time as e−iEt/~ .
These free-space solutions are said to be on the
energy shell, because the E-p relation is satisfied 11.5 Scattering at low energies:
on a spherical shell in p-space. Impact parameter: In classical mechanics, a par-
Static solutions of the KG equation: The ticle impacting at a transverse distance b (the im-
KG equation does not have time-independent so- pact parameter) from the center of a square-well
lutions in free space, because these solutions have potential of range R will not “see” the potential if
E = 0 and therefore cannot satisfy Einstein’s b > R. If the incident momentum in the center-
E-p relation. Yukawa showed that there are of-mass frame is p, the interaction vanishes when
static (hence energy-nonconserving) solutions near the angular momentum ℓ = r × p exceeds the value
a point charge q located at the origin that satisfy ℓmax = Max(kR/~), where Max(z) is the largest
the inhomogenious DE: integer in z.
The wave spreading in quantum mechanics makes
p̂2 c2 + m2 c4 φ(r) = 4π(~c)2 qδ(r),


(296) the connection between ℓ and b less sharply defined,
but it remains true that the scattering phase shift
For m = 0, the Yukawa equation simplifies to the δℓ vanishes sharply when ℓ > ℓmax . Consequently,
Poisson equation
ℓX
max
2 f (θ) ≈ (2ℓ + 1)fℓ Pℓ (cos θ), (301)
∇ φ(r) = −4πqδ(r). (297)
ℓ=0
One can show with the help of Gauss’s theorem
(or Gauss’s law in electrostatics) that the resulting Hence at sufficiently low energies, the S-wave (ℓ =
static wave function of a massless photon around a 0) term dominates:
point charge that vanishes at r = ∞ is sin2 δ0 4π
σtot ≈ 4π = 2 . (302)
q k2 k + k 2 cot2 δ0
φ(r) = . (298)
r The fact that the leading Taylor term for ℓ = 0 is
This is just the Coulomb potential around the
√ point independent of k comes from the properties of the
charge q (in Gaussian units, or q = qSI / 4πǫ0 ). scattering wave function.
 28

Effective-range expansion: 11.7 Born approximation for two-particle scatter-

ings:
1 1 1
k cot δ0 = − = − + r0 k 2 + ..., (303) The Born (or first Born) approximation for the
a(k) a0 2
scattering of two particles is a first-order TDPT
where a0 is called the scattering length and r0 is that can be obtained from the
called the effective range. Golden rule: A weak two-body interaction

Ĥ ′ = V̂ e−iωt + V̂ † eiωt (311)

11.6 Phase shifts for finite-range potentials:
causes the scattering from an initial two-particle
Radial wave function: The wave function in the
state i to a final two-particle state f within the solid
presence of a potential V is
angle d2 Ωf around the final relative momentum kf .
∞ The transition rate is given, in first-order TDPT,
X uℓ (r)
ψ(r) = iℓ (2ℓ + 1)Pℓ (cos θ) , (304) by the Golden Rule:
kr
ℓ=0
dwf i 2π ρf (E)
where the radial wave function uℓ satisfies the ra- = |hf |V̂ |ii|2
dΩ ~ 4π
dial wave equation µf (~kf )  3 
= 2 4
L |hf |V̂ |ii|2 . (312)
d2 uℓ 2µ ℓ(ℓ + 1) (2π) ~
2
+ 2 (E − V )uℓ − uℓ = 0. (305)
dr ~ r2 Here µj , kj are the reduced mass and relative mo-
where µ is the reduced mass in the center-of-mass mentum, respectively, in the state j (= i, f ) of the
system. two-body system in a cube of side L, and the den-
sity of final state factor is
Attractive S-wave square-well potential: For
ℓ = 0 in an attractive square-well potential of depth ρf (E)

L
3
k 2 dk
V0 , the radial wave equation is = . (313)
4π 2π dE
d2 uin
For r < R : + κ2 uin = 0, The scattering is elastic if kf = ki , and inelastic if
dr2 kf , ki .
uin = sin(κr); (306)
The resulting differential scattering cross sec-
tion
d2 uout
+ k 2 uout = 0,
 
For r > R : dσf i 1 dwf i
dr2 ≡ (314)
dΩ Jinc dΩ
uout = sin(kr + δ0 ); (307)
is just the transition rate per unit incident flux (or
2mE 2mV0 current density), which is
k2 = , κ20 = , κ2 = k 2 + κ20 . (308)
~2 ~2 vi ~ki
Jinc = 3
= (315)
The phase shift δ0 is determined by matching the L µi L3
inside and outside logarithmic derivatives Lin = in a cube of side L. Note that Jinc has the expected
Lout at r = R, where L = (du/dr)/u, to give dimension of m−2 s−1 . The final result for the Born
1 1 approximation can be written compactly as
tan κR = tan(kR + δ0 ), or
κ k dσf i µi µf
 
kf
= |Ṽ (q)|2 , (316)
 
k
δ0 = tan−1 tan κR − kR. (309) dΩf (2π)2 ~4 ki
κ
where q = kf − ki , and
Levinson’s theorem: If there are m bound states Z
inside the potential, Ṽ (q) ≡ L3 hf |V̂ |ii = e−iq·r V (r)d3 r (317)
δℓ (k = 0) = mπ, (310)
is obtained by using the plane-wave wave functions
except for an S-wave bound state at E = 0, which in the relative coordinate r normalized to one par-
contributes only π/2, half of the normal contribu- ticle in a cube of side L:
tion.
1 1
hr|ii = eiki ·r , hf |ri = e−ikf ·r . (318)
L3/2 L3/2
 29

Thus the cross section in the Born approximation

can be calculated directly from the Fourier trans-
form Ṽ (q) of the interaction potential V (r).
Examples of Ṽ (q) are those from the δ-shell and
the Yukawa potentials:
4πR
V (r) = A0 δ(r − R) ⇒ Ṽ (q) = A0 sin(qR);
q
e−αr 4π
V (r) ∝ ⇒ Ṽ (q) ∝ . (319)
r α2 + q 2