Professional Documents
Culture Documents
Quantum
Mechanics
Second Edition
l July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Quantum
Mechanics
Second Edition
Fayyazuddin
National Centre for Physics
Quaid-i-Azam University Campus
Islamabad 45320
Pakistan
Riazuddin
National Centre for Physics
Quaid-i-Azam University Campus
Islamabad 45320
Pakistan
World Scientific
Singapore. New Jersey. London. Hong Kong
I July 9, 2012 14:29 World Scientifíc Book - 9in x 6in RF-QuantumMechanics I
Publishers' page
¡¡¡
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Publishers' page
iv
¡ July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Publishers' page
V
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Publishers' page
vi
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
vii
l July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Preface
ix
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
X Quantum Mechanícs
Preface xi
Contents
Prejace ix
xiii
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
6. ANGULAR MOMENTUM 87
6.1 Introduction . . . . . 87
6.2 Properties of Angular Momentum . . . . . . . . . . . . . 87
6.3 Eigenfunctions and Eigenvalues of Angular Momentum . 92
6.4 Properties of Associated Legendre Polynomials 97
6.5 Parity of a State 98
6.6 Problems . . . . . . . . . . . . . . . . . . . . 99
Contents XV
9. OPERATORS 175
9.1 State Vectors 175
9.2 Schrodinger Representation . . . . . . . . . . . . . . . . . 181
9.3 Relation between the Momentum and Schródinger Repre-
sentations . . . . . . . . . . . 185
9.4 Matrix Mechanics . . . . . . . 187
9.5 Simple Harmonic Oscillator . 190
9.6 Supersymmetric Oscillator . 194
9.7 Problems . . . . . . . . . . . 195
Contents xvii
Contents xix
Index 541
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 1
BREAKDOWN OF CLASSICAL
CONCEPTS
1.1 Introduction
1
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
2 Quantum Mechanics
Here p is the charge density and j is the electric current and they satisfy
the continuity equation
ªP + di
{)t IV J• = o. (1.2)
( e1 fJt2 -
[)2 )
v'2 E, B = O. (1.3)
On the other hand, the maximum of the amplitude moves with velocity
6w 6v
Vg = 6k = 6(1/ .\)"
(1.8)
4 Quantum M echanics
- - - - 1/ZA.m - - - - _,
Fig. 1.1 The superposition of two simple harmonic waves of slightly different frequencies
and wave nurnbers.
or
t::..kt::..x = 2rr
or
t::..kt::..x > 1. (1.10)
In general, we can represent a wave packet (which is a superposition of
monochromatic waves with wave numbers centered around the mean value
ko) as:
'ljJ(x t) = l A(k)ei(kx-w(k)t)dk L (1.lla) )
A(k)
w = w(k) , (1.12)
in general there will be dispersion in the wave, the wave packet spreading
out as it moves along. The range over which A(k) is appreciably different
from zero is
1 1
ko - -6.k
2 - k -< ko + -6.k
< 2 . (1.13)
6 Quantum M echanics
is appreciably different from zero only in the range .6.x. Now using
Eq. (1.14) in Eq. (1.lla)
1/J(x, t) ~ f~oo A(k)eikox+i(k-ko)xe-i(wot+v9(k-ko)t)dk
= ei(kox-w0t) f~oo A(k)ei(k-ko)(x-v9t)dk
= ei(kox-wot) X(x _ Vgt) . (1.18)
The wave packet is composed of two factors: the first factor represents a
wave of frequency w0/21r and wave length 21r/k0; the last factor describes
a modulation of the amplitude of this wave. This modulation moves with
velocity v9, so that Eq. (1.18) describes a group of waves (wave packet)
which moves without a change in shape with group velocity v9.
In deriving Eq. (1.18), we have neglected the term
1. (d w)2
2it dk2 ko (k - ko)
2
(1.19)
t(~:~) ko (.6.k)2 « 1
or
or
(1.20)
Can the classical concepts outlined above, viz, matter consisting of point
particles and radiation consisting of waves provide a framework for phys-
ical phenomena on a microscopic scale, i.e. when point particles are elec-
trons and protons, each carrying mass and electric charge and interacting
through the fundamental electromagnetic forcEU As we shall see, the classi-
?
cal concepts were found to be quite inadequate in describing the motion of
electrons and their interaction with radiation or other particles. Also the
emission and absorption of radiation could not be explained by the classical
concept of the wave nature of radiation.
E= fo 00 E(w)rku ,
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
8 Quantum Mechanics
or
p=lik. (1.25)
These equations clearly show the relation between the particle parameters
(E, p) of the photon, and the parameters (w, k) of the corresponding wave.
This quantum idea was later used by Einstein to explain the photoelec-
trie effect. If a beam of light of frequency w is incident on the surface of
a metal, electrons may be emitted. Note that when electrons are emitted,
their energy T does not depend on the intensity of radiation, but only on
its frequency. The more intense the beam the more electrons are emitted,
but their energy will depend on the frequency. This cannot be explained
in the classical picture of continuous exchange of energy. However, it is
easily explained by Planck's hypothesis. When the incident energy of the
photon ñw > W, where W is the work required to free the electron from
the attractive potential of the metal, electrons are emitted, and carry the
energy T given by
ñu: = W +T. (1.26)
Otherwise, i.e. for ñw < W, no electron is emitted. Sincebumber of pho-
tons are proportional to intensity of light,~re intense the light, ~ore
n;
.fÁ.t.
..,% L
electrons are emitted if Iua > W.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= 2~
271'
sin2 !l.2 (1.27)
where
10 Quantum M echanics
Later it was found that this dual behaviour was not confined to radia-
tion, but holds for particle as well. This was experimentally discovered by
Davisson and Germer and by G.P. Thomson. They showed that when a ho-
mogeneous beam of electrons passes through a crystal, the emergent beam
exhibits a pattern of alternate maxima and minima of intensity, wholly
similar to~iffraction pattern observed in the diffraction of electromagnetic
radiation. The beam of electrons thus behaves as though it were a train of
waves of wave length
>. = h/p (1.29)
or
p = h/>. = hk ,
where pis the momentum of each electron.
This was predicted, however, by de Broglie before it was discovered
experimentally. Thus an electron of given energy E and momentum p was
associated with a de Broglie wave such that
E= lu/ = nw
p=~=ñk. (1.30)
Now
mc2 mv
E = ---,:;=== p= ---¡:;=== (1.31)
JI -v2/c2 ' JI-v2/c2
Therefore, phase velocity for a de Brogile wave is given by
w E c2
U=-=-=->C. (1.32)
k p V
Now
or
2E dE = 2 e2p.
dp
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore
dE pc2
Vg= - = - =V. (1.34)
dp E
The relationship thus brought out is very attractive; we associate a wave
packet
'f/;(x, t) = j A(k)i(kx-w(k)t)dk
with a particle. For a free particle
w(k) =
2~
[k5
12 Quantum M echanics
where
X(x-vt- ~ Llk) ~ 1: A(k)e'(kx-w(k)<)[x-•<--,!.ak<)dkL(l.37)
~ - R(2_ -
.X - nr 2-)
n~ ' (1.39)
continuously and thus lose energy. As a result its orbit would become
smaller and smaller and it would ultimately be drawn into the nucleus.
The classical theory has two important consequences:
(1) The atom should be very unstable.
(2) It should radiate energy over a continuous range of frequencies.
Both of these consequences completely contradict experiment. The atom is
stable and as we have seen above, the spectrum of the hydrogen atom has
discrete frequencies.
In order to explain the hydrogen spectrum and the stability of the atom,
Bohr postulated the followingrules:
(i) The magnitude of the angular momentum L of the electron is an integral
multiple of !i
L = nñ , n = 1,2, ....
e2 /a2 = maw2 .
Now
L = ma2w = nñ .
(1.41)
where
- ñ,2 -11
ao = --2 = 5.29 X 10 m (1.42)
me
is called the first Bohr atom.
Now
14 Quantum Mechanics
e2 ( 1 1 ) (1.44)
Wnn' = 2fíao n2 - n'2 .
1.3 Problems
1.1 Find the energy and momentum of an X-ray photon whose fre-
quency is 5 x 1018 cycles sec"".
1.2 (a) The velocity of ripples on a liquid surface is J21rS/>.p, where
S is the surface tension and p the density of the liquid. Find the
group velocity of these waves.
(b) Velocity of ocean waves is J g>./21r, where g is the acceleration
due to gravity. Find the group velocity of ocean waves.
1
1.3 Given that A(k) = (2 k)
2, JkJ ~ K, A(k) = O, JkJ > K. Find 1/J(x).
Make a sketch of A(k) and 1/J(x) and hence show that fikfix > l.
1.4 For a Guassian wave packet
1/J(x, t) = N e-x2 /2152 ei(k0x-w0t)
Show that
A(k) = _!!_óe_152(k-ko)2/2
v'21r
Make a sketch of 1/J(x) = 1/J(x, O) and A(k) and hence show that
fixfik = 1 (1.45)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
1.5 Light of wave length >. = 10 meters falls on the surface of a metal.
If the work function of metal is 3.64 eV, is an electron emitted or
not?
1.6 Find the velocity of a proton whose de Broglie wave length is equal
to that of a 1 keV X-ray.
1.7 A photon whose initial frequency was l.5x1019 sec-1 emerges from
a collision with an electron with a frequency l.2x 1019 sec-1. How
much kinetic energy was imparted to electron?
1.8 Derive a formula expressing de Broglie wave length (in A) of an
electron in terms of potential difference V (in volts) through which
it is accelerated.
1.9 Show that velocity v of the electron in the first Bohr orbit of hy-
drogen atom is given by
v/c = e2 /ne= 1/137 .
1.10 A beam of 100 MeV electrons travels a distance of 10 meters. If the
width ~x of initial packet is 10-2 cm, calculate the spread in the
wave packet in traveling this distance and show that this spread is
much less than ~x.
1.11 Find the de Broglie wave length of an electron accelerated through
a 10 Volts potential difference. What is its velocity?
1.12 A light of wave length 2500.A. falls on a surface of metal whose work
function is 3.64 eV. What is the kinetic energy of the electrons
emitted from the surface? If the intensity of light is 4.0 W/m2,
find the average number of photons per unit time per unit area
that strike the surface.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
16 Quantum Mechanics
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 2
QUANTUM MECHANICAL
CONCEPTS
17
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
18 Quantum M echanics
the smaller the wave length, the smaller the uncertainty in position. Hence
if we use light of short wave length to increase the accuracy of position
measurement, there will be a corresponding decrease in the accuracy of
momentum measurement, while light of long wave length will decrease the
accuracy of position measurement with the corresponding increase in mo-
mentum measurement. Combining the above two estimates, we get
(2.1)
which is one form of the uncertainty principie.
Thus if we make a measurement on any object, and we determine the
x-component of its momentum with an uncertainty 6.px, we cannot at the
same time know its x position more accurately than 6.x = h/ 6.Px· The
uncertainties 6.x, 6.px in the position and momentum of a particle at any
instant must satisfy the relation
(2.2)
ñ is so small that the uncertainty principie is important only on the atomic
scale where laws of motion have to be altered to include it.
Hence quantum mechanics cannot make completely definite predictions
concerning the future behaviour of, for instance, electrons. For a given
initial state of the electron, a subsequent measurement can give various
results. The typical problem in quanturn mechanics consists of determining
the probability of obtaining various results on performing this measurement.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Thus only probability relations exist between present and future. How-
ever, there is a casual relation between probabilities. Quantum mechanics
enables us to formulate these relations in a general set of equations that
will supply us with a description of atomic phenomena.
To sum up we no longer have determinacy as is the case on the macro-
scopic scale, where classical mechanics is applicable and we have to replace
determinacy by probability. Quantum mechanics enables us to calculate
the probability.
L. I
I
I
I
, '
f ¡ \
1
I
I
1
1
1
',
\
\
\
\
\
<..">
\
\ \
'
1
I
1 I
I
/
Q
1 /
\ 1 I
\ ,e' I ¡
\ 1 I
· • • ooa ~ I
20 Quantum Mechanics
' represents the motion of the system which '1/J represents. We <leal first with
(ii) and come to (i) later. In analogy with other fields ofphysics, we look for
a differential equation for '1/J(x, t). For this purpose, we describe a particle
with a wave packet,
where w is given by Eq. (2.9). Hence the wave packet associated with the
particle1 is
For the wave function given by Eq. (2.11), the following relation holds [c.f.
Eq. (Llld)J
l::,.kl::,.x ~ 1, (2.12)
which on using relation (2.6) gives
l::,.pl::,.x~ ñ. (2.13)
Thus our description of a particle by wave packet satisfies the basic tenant
of quantum mechanics.
We now proceed to find a differential equation having (2.10) as its so-
lution. We have from Eq. (2.11)
ino'l/J(x, t)
ot = foo
-00
A(k) ñ2k2 ei[kx-(ñk2/2m)tldk
2m
-(n2/2m)°21P =
ax2
foo
-00
A(k)ñ2k2 ei[kx-(ñk2/2m)tldk
2m
22 Quantum M echanics
atomic scale. The above equation is for a free particle, i.e. a particle on
which no force is acting and for which
2
Total energy = Kinetic energy = ;m. (2.15)
It may be seen that Eq. (2.14) can be obtained from Eq. (2.15) by regarding
energy and momentum as differential operators
p-----; ·te
-in-
a (2.16)
OX
that act on the wave function 'ljJ to give
iño'lj; = H'lj;
at
where
ñ2 82
H=---- (2.17)
2m8x2
is the Hamiltonian operator for a free particle.
This is an example of the correspondence principle:
"The essentially definitive relations between physical variables in clas-
sical mechanics, which do not involve derivatives, are also satisfied by the
corresponding quantum operators."
For a particle moving in a field of force given by the potential V(x), the
total energy is given by
p2
E=-+V(x). (2.18)
2m
The rule given in Eq (2.16), gives the differential equation:
o'l/;(x t) ñ2 82
in 0/ 2m
ox2
= ( - + V(x) ) 'l/;(x, t) = H'l/;(x, t) (2.19)
because the particle must be somewhere between -oo and +oo. When it
is satisfied, the wave function is said to be normalised. Thus, in order to
normalise the wave function, all we have to do is to adjust the coefficient
of '1/J(x, t) in such a way that the integral over all space is one.
Eq (2.20) implies that J!';::
P(x, t)dx is independent of time i.e
d
dt 1-oc
00 P(x, t)dx = O. (2.23)
!1-: 1-:
But
P(x, t)dx = (ª'1/J*~;, t) '1/J(x, t) + 'ljJ*(x, t) a'ljJ1:, t))dx.
- ....!L
- 2mi
_§_ (·'· 8,/J* - .!,*
8x '+' ax '+' ax
/}j,_) · (2.26)
Hence
-d
dt
100 P(x, t)dx
-oc
= -.
2mi
n loo -('1/J-
-oo
a
ax
ª'1/J*
ax
- 'ljJ *ª'1/J
-)dx
ax
!i ( a'ljJ* * a'ljJ) +oc
=- 'I/J--'I/J-
2mi ax ax -00
=Ü (2.27)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
24 Quantum M echanics
S(x, t) = ~
2mi
ª 1*).
(1* ax1 - 1 ax ª (2.28)
provided that
1(x, t) ---+ O as x ---+ oo.
This is the boundary condition satisfied when the wave function 1(x, t)
represents a bound state, for example, an electron bound in the hydrogen
atom, or a particle in a potential well. For a bound state, the particle can
never go to infinity, i.e. 1(x, t) = O at x--+ oo. Therefore J l'f/!(x, t)l2dx is
convergent and can be normalised.
There is another kind of wave function encountered in collision problems
in which case a particle arrives from infinity (great distance), is scattered
by a field of force and goes off to infinity (i.e. to a great distance from the
centre of force) again. For such cases the wave function remains finite even
as x --+ oo and J 1112dx, diverges. Here 1112 does not directly determine
the probability and must be regarded as a quantity proportional to this
probability. For example, for a wave function of type
the normalisation integral J::'00 l'I/Jl2dx when taken over the interval diverges
but in this case 'lj; can be defined in a finite interval so that it vanishes or
has a periodic structure at the boundary.
2. In the Schrodinger picture we have considered, x is an algebraic
variable but momentum p is a differential operator
~
X= X,
~
p = ·t;
-in0X.
a
In general the two operators A and B do not commute, i.e.
AB =J BA (2.34)
(cf ordinary numbers which commute). As an example of illustrating this
property of operators, we take the operators p. Then in Schrodinger x,
picture
X]) =X( - iñ fx)
px= ( -iñJx)x. (2.35)
Therefore
xp"'lj;(x,t) = x( -iñ:x)'I/J(x,t)
26 Quantum M echanics
This is known as the quantum condition. It simply means that x and p can-
not be measured precisely and simultaneously. In fact it can be shown that
this quantum condition is completely equivalent to Heisenberg's uncertainty
x
principle. Note that as ñ=» O, and p commute as in classical mechanics
and in this limit quantum mechanics reduces to classical mechanics.
3. The foregoing treatment has been for a one-dimensional case. It may
be readily extended to three dimensions. Then
p2
p=nk; E=2m+V(r). (2.40)
The corresponding operators are
p-+ -ínV (2.41)
E-----, in~at (2.42)
h2
H-----, --V2
2m
+ V(r) (2.43)
so that the Schródinger equation
ina'!/J¡:, t) = H'!fJ(r, t) (2.44)
can now be written
ína'!/J¡:, t) = ( - ;: V2 + V(r) )'!/J(r, t). (2.45)
The probability density
P(r, t) = l'!/J(r, t)l2 (2.46)
and probability current density is
S(r, t) = . ['!/J*V'!/J - '!/JV'!/J*] (2.47)
2nmi
and the equation of continuity is
aP¡:, t) +V· S(r, t) = O. (2.48)
Now
= Pi í = 1, 2, 3 (2.49)
r == (x,y,z) == (x1,x2,x3)
o
= Xi Í = 1, 2, 3. (2.50)
In this case the quantum conditions are given by
[xi,PJ] = iMij L (2.51)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
(
H x, -iñ 8x
a) = -
ñ2 a2
2m 8x2 + V(x). (2.55)
Note that it does not contain t, so that
Hu(x)f(t) = J(t)Hu(x). (2.56)
Therefore dividing both sides of Eq. (2.53) by u(x)f(t), we have
.,. fJ(t) _ H(x, -iñfx)u(x)
i,i7'[i) - u(x) . (2.57)
Since left-hand side depends only on t, and right-hand side only on x, and
these variables can be varied independently, each side must in fact be equal
to a constant. We call this constant E. Thus,
iñÍftf(t) - •
(2.58)
f(t) -
28 Quantum Mechanics
With the expression of Has given in Eq. (2.55), Eq. (2.60) becomes
ñ,2 {)2 )
( -
2m
ax2 + V(x) uE(x) = EuE(x) (2.61)
Solution of Eq. (2.60) or Eq. (2.61) will give us energy eigenvalues or energy
levels and corresponding eigenfunctions. Note that for a stationary state
P(x,t) = l'I/JE(x,t)l2
= luE(x)l2, (2.64)
(1)
Another example is
~ ~( {)) 8
A = A x, ox = Dx x
which gives
u(O) = O
u(7r) = O (Boundary conditions).
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
30 Quantum Mechanics
2.7 Problems
2.1 If the excited level in an atom lasts for about 10-10 sec., what is
the order of magnitude of electron energy spread measured in eV?
2.2 What is the approximate momentum imparted to a proton initially
at rest by a measurement which locates its position within 10-11
meters?
2.3 A particle has an uncertainty in its position lix = 2a. The magni-
tude of the momentum must be at least as large as the uncertainty
in the momentum. Hence find an estimate of its energy.
2.4 Show that
l 2
and u(x) = 2xe-~x
are eigenfunctions of the operator
~( {)) _ d2 2
A x, ax = - dx2 +X •
(
b~
- where mis the mass of the particle.
¡..__ 2.6 ln /3-r\d cay, electrons are mitted from the nuclei with energy of
-( k ~w MeV. Take the energy)l, foV and size of the nucleus 10-13
cm. U e )uncertainty principie to how that electrons cannot be
containecf'in the nucleus before the decay.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
32 Quantum Mechanics
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 3
BASIC POSTULATES OF
QUANTUM MECHANICS
33
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
34 Quantum M echanics
'lf;(x, t) is given by
(3.5)
n
then we have
(3.6)
n
i.e.
where
(3.18)
m=n.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
36 Quantum Mechanics
ómn is called the Kronecker delta. The eigenfunctions are then said to be
orthonormal.
(d) The eigenfunctions Un of a hermitian operator (observable) form a
complete orthonormal set so that any arbitrary state function 'lj; can be
expanded in terms of them, i.e.
(3.19)
n
This is what we mean by a complete set. The Eq. (3.19) is called to the
superposition principle, a basic ingredient of quantum mechanics.
Now in analogy with vector analysis where we express a vector in terms
of basis vectors i,j, k, the Un 's are called basis vectors and C¿ the corre-
sponding coordinates. It followsthat
j u:n'lj;dr = (uml'l/J)
= ¿ Cn(umlun)
n
n
(3.20)
where Cm is related to the probability of finding the system described by
state 'l/; in an eigenstate Um· Thus
(3.21)
gives the probability of operator A having the eigenvalue am, when the
system is described by a state 'lj;. To see this we note that the average value
of the operator A in state 'lj; is given
a= j 'lj;*A'lj;dr
= ¿ ¿
m n
ci.c; j u:nAundr
m n
(3.22)
m
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
The weighing factor ICml2 above gives the probability of finding the Eigen-
value ªm·
(e) The operators do not necessarily obey a commutative law, i.e. two
operators A and E need not give AE = EA. If AE equals EA, the operators
are said to commute, i.e.
[A,EJ =AE-EA=O (3.23)
[A, E] is called the "commutator" of two operators.
(f) If two observables commute, then it is possible to find a set of func-
tions which are simultaneously eigenfunctions of A and B. If they do not
commute, i.e. [A, E] =/= O, this cannot be done except for a state '!/; which
has [A, BJ'I/J = o.
We now show that if Un is a simultaneous eigenfunction of A and E
corresponding to eigenvalues an and bn, then
[A,E]un =O. (3.24)
Now
Aun= llnUn (3.25a)
Bun = bnUn · (3.25b)
Further
ABun = AbnUn
= bnAUn
(3.26a)
E.Aun= EanUn
= anBUn
= anbnUn , (3.26b)
Therefore
i.e.
[A,E]un =O. (3.27)
The implication of this result is as follows:
Since Un form a complete set so that an arbitrary function '!/; can be
expanded in terms of them, it follows that
[A,BJ'I/J=O.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
38 Quantum M echanics
with
(3.33)
Thus ( Ua' lua) has the peculiar property: it is zero everywhere except at
a= a' [see íg, 3.1] nd at a= a' it is infinitely large such that its integral
is l.
o(a-a)
Thus we write
(ua lua) =
1 J u~,(x)ua(x)dx
= ó(a - a') , (3.35)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
40 Quantum Mechanics
Here
o(a - a') = ó(ax - a~)o(ay - a~)o(az - a:) .
~
(b) Closure Relations indexClosure relation · We have
'f/;(r) = ¿ Cnun(r) (for discrete set) (3.43a)
n
Then
C; = Ju~ (r')'f/;(r')dr' (3.44a)
C(a) = J u;(r')'f/;(r')dr' . (3.44b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
1/J(r) = ¿ ( J u~(r')1/J(r')dr')un(r)
n
= J (¿ n
un(r)u~(r') )1/J(r')dr' .
= J 1/J(r') ¿
n
Un(r)u~(r')dr
= ¿ ( J u~(r')1/J(r')dr')un(r)
n
(3.46)
n
J eia(x-x') da .
Now eia(x-x') is an oscillating function and the above integral is not defined.
It is a question of agreeing to give a value to this integral. The prescription
is
f 00
-oo
eia(x-x')da = lim (
E-0
jº
-oo
eia(x-x')+wda + {
lo
00 eia(x-x')-éada)
1 1
=lim( - )
E-o i(x - x') + t: i(x - x') - t:
. (
- lim 2t: )
- E-o E:2 + (x - x')2 ·
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
42 Quantum M echanics
If x-¡. x', (x - x')2 is a fixed number, no matter how small it may be. Now
when e --t O, c2 can be neglected in comparison with (x - x')2 so that the
limit = O. If x = x'
. (
l im 2c ) =
2
1.lffi-=00.
é---+O c2 + (x - x')2 ,:-+O e
The behaviour is that of a 8-function, but we have to verify that its integral
is l. Thus we calculate
00
(
/00. eia(x-x ) da) dx
I
= lim 100 ~ dx
/-oo -oo ,:-+O _00 c2 + (x - x')2
00
2c
= lim r¡ = (x - x')
e---+O /_00 e2 + r¡ 2dr¡,
. [2t an -1 -r¡]oo
= l 1m
e---+O e -oo
= 2 lim ( ~ - ( - ~))
é->Ü 2 2
= 211".
Thus
2_
-oo
21!"
f
00 eia(x-x')da = ó(x - x') . (3.47a)
-1
(21r)3
-f eia·(r-r')da = ó(r - r')
.
(3.47b)
J(x)ó(x - b) = J(b)ó(x - b)
xó(x) = O
8(-x) = 8(x)
1
8(bx) = ¡¡;¡8(x)
1
8(x2 - b2) =
2b
(8(x - b) + 8(x + b)), b >O
These equations have meaning only in the sense of integration; for example,
the first one means
J ó(bx)dx = m.
(e) Fourier Transform
f(x) = rn=
1 /00 C(a)eiªxda (3.48)
y 271" -00
--l
v'27r
/00 f(x)e-'ª. , "dx
-oo
= --1 /00 daC(a)
(27í) -00
!00 e•x(a-a
-oo
. , ldx
= j daC(a)ó(a - a')
= C(a').
Therefore
1 /00 f(x)e-'ª"'dx
C(a) = . rn= . . (3.49)
y 271" -oo
C(a) and f(x) are called the Fourier transforms of each other.
The generalisation to three dimensions is
1
C(a) = (27r)J/2. ¡· f(r)e-ia·rdr.. (3.50b)
@ ,
~genvalue equation is
&
pup(x) = pup(x)
-ili! up(x) = pup(x). (3.51)
A solution of this equation is
44 Quantum M echanics
p = -iñV
and
pup(r) = pup(r)
Now
(up,jup) = J u;,(r)up(r)dr
= IBl2 J e(i/n)(p-p')·rdr
= IBl2ñ3(211-)3J(p - p") .
Thus if we select
1
B = (21rñ)3/2 '
then
(3.54)
º''( )- 1
'// r - (21rñ)3/2
J C( ) (i/n)p·rd
Pe p , (3.56a)
where
1
C(p) = Je-(i/n)p·r'l/;(r)dr (3.56b)
(21rñ)3/2
p= (p) = J 'tf;*(r)p'I/J(r)dr
= -1-JJJdp'dpC*(p')C(p)e-(i/ñ.)p'·r
(21rn)3
X (-in v')e(i/ll.)p·r dr
46 Quantum M echanics
It is clear from Eq. (3.59) that a and /3 are hermitian operators so that we
have
ata = ª2 , /Jt /3 = 132 .
The mean square deviations of the measured values, of A and B about the
mean are given by
Now
= j(at'l/J)*a'l/)dx
= j (a'l/J)*a'l/)dx
= j la'l/Jlzdx.
Then
Hence
(6.a)2(6.b)2 ~ 1 j(a'l/J)*/3'1/)dxl2
= 1 j(at'I/J)*/3'1/Jdxl2
= j 'l/)*a/3'1/)dxl2
1
= JP - iQr (3.62)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
p = J 'l/¡* a{3 ; {3a 'l/¡dx
Since ,.¡3~¡3,. and i af3-:/°' are hermitian operators and the average value of
a large number of measurements of a hermitian operator is real, therefore
P and Q are real numbers.
Thus Eq. (3.62) gives
(~a)2(~b)2 ~ p2 + Q2 ~ Q2
= 11 'l/¡*iª{3;{3a'l/¡dxl2
(~p)2(~x)2 ~ ~ 1 j 'l/¡*'l/¡dxJ2 = ~
The root mean square deviation is often called the uncertainty (standard
deviation), i.e.
Example
1
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
r
48 Quantum M echanics
~ ~ ·i. 8
A =p--->
-in-
OX
~ p2 f¡,2 82
B=H=----- (3.66)
2m 2m8x2
so that
8 f¡,2 82
[p,H]v,(x) = [-ilt- - --]vi(x)
ax 2m8x 2
=0
[p,H] =0.
The energy and momentum of a free particle can be known exactly, simul-
taneously. In other words, it is possible to find a wave function which is a
simultaneous eigenfunction of both momentum and energy.
3.5 Problems
3.1 If (A) denotes the average value for a large number of measurements
of an operator A for an arbitrary state function v,, show that (A)
is real if A is hermitian.
3.2 The state function for a free particle moving in x-direction is given
by
Normalise this wave function. Find the state function </J(p) in mo-
mentum space.
(i) Show that for the state v,(x) given above
'
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(p) = J pJ</>(p)J2dp
(p2) = J p2J</>(p)J2dp
show that
[H, r] = -ili.!!__
m
(p, H] = [p, V(r)] = -ili"vV
Using the above results and the fact that H is hermitian, show that
50 Quantum Mechanics
b) If it is in a state
Chapter 4
SOLUTION OF PROBLEMS IN
QUANTUM MECHANICS
= V for x > O
V ( x)
V(x) = O for x < O.
L ( 4.1)
T1=E-V=-. P?
2m
51
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
52 Quantum Mechanics
E0 - - - - - - - - - - - - - - - - - - - - - - (i)
E0- - - - - - - - - - - - - - - - - - - - - - (ii)
The particle will, therefore, move with reduced momentum. Further, since
the particles have suffi.cient energy to go over the barrier, there is total
transmission. (ii) E - V(x) < O , for x > O. In this case the particles
hitting the potential barrier from the left are unable to go across it and are
completely reflected back. There is, therefore, no transmission through it.
Having discussed the classical aspect of the problem, we now take up
the quantum mechanical treatment.
The quantum mechanical motion is determined by the eigenvalue equa-
tion
Hu=Eu (4.3a)
or
¡¡,2 82
( -
2m
ox2 + V(x) )u(x) = Eu(x). (4.3b)
Now
V(x) =O x<O
= V X> o.
k2 _ 2mE _ 2m(E- V)
O - ñ,2 '
k 21 - ñ,2 . (4.4)
l July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
\~ uo(x)~-
~
incident )
+~
(reflected)
L (4.7a) i
q U1 (x) D eik1x
'-v--"
. (4.7b)
( transmi tted)
From Eq. ( 4.8) we can find B and D in terms of A for any value E (i.e. no
restriction on energy values). These are given by
B _ (ko - k1)A. D- 2koA (4.10)
- (ko + k1) ' - (ko + k1)"
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
54 Quantum Mechanics
Si(x) = ~
2mi
(u'!' OUí
i 8x
- Ui out)
8x
(4.11)
- - - - - - - - - - - - - - - - - -Eo
X
56 Quantum Mechanics
Therefore
iko(A - B) = -KG. (4.20b)
'
B = (ko - iK)A . C= 2koA (4.21)
(ko + iK) ' (ko + iK)
for any value of E, so that again there is no restriction on possible energy
values. Here
R = IBl2 /IAl2 = 1 ,
i.e, we get total reflection as in the classical case.
There is, however, a difference from the classical result in that the prob-
ability of finding the particle in the classically forbidden region, i.e. x > O
is not zero. This probability is given by
P(x) ex lu1(x)l2 = ICl2e-2Kx X> Ü.
Thus
(4.22)
This is appreciable near the barrier edge and falls exponentially to negligible
value at distances large compared with 1/ K. The wave function will thus
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
58 Quantum M echanics
V(x)
.-a-- ....
Define
k2 = 2;,;E , K2 = 2; (V - E) .
[::2 + k2]u2(x) = O. x ~ a
The solutions of these equations are
uo(x) = Aeikx +Be-ikx L (4.29a) )
u1(x)
u2(x)
=
=
c e r>
A' eikx.
+DeKx L (4.29b) )
(4.29c)
There is no reflected wave in region (2)í we have only a transmitted
wave as given in Eq. (4.29c). The continuity conditions give
and
D = ~1 (l + tr:: (4.32a)
(4.32b)
From Eqs.
(4.33a) J
(4.33b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
60 Quantum Mechanics
and
Fig. 4.5 The incident and transmitted waves for a rectangular potential barrier.
We expect this shape for the wave function beca use when Ka » 1 then,
as is clear from Eq. (4.38b), D ~ O; thus
uo(x)=A(eikx+!-ikx)l (4.43a) }
U1(x) ~ C e-Kx [_ (4.43b) )
u2(x) ~ A' e=. (4.43c)
4.3 Parity
62 Quantum M echanics
f¡,2 {)2 )
( - 2m 8x2 + V(-x) UE(-x) = E UE(-x) (4.46)
f¡,2 32 )
( -
2m
ox2 + V(x) ue(-x) = EuE(-x). (4.47)
Thus we see that uE( -x) is also a solution of Eq. ( 4.45) for a given value
of E. If we assume that there is only one linearly independent solution for
a given eigenvalue E, then the two solutions can differ only by a constant,
therefore
Changing x-+ -x
(4.48b)
Therefore
r¡ = ±1 .
The constant r¡ is called the parity of the state. For positive parity states,
(r¡ = +1), uE(x) is an even function of x; for negative parity state (r¡ =
-1), uE(x) is an odd function of x.
If there is more than one solution, and a particular solution <loes not
have definite parity, then from uE(x) and uE(-x) one can always construct
solutions
(4.49a)
(4.49b)
+ 00 + 00
V(x)
Fig. 4.6 The energy díagram of a one-dimensional infinite well where a particle is con-
fined to the region
64 Quantum M echanics
A sin ka= O
or
Bcoska =O.
, e do not want both A and B = O since his ¡voul9 give a phy ically
\ \.\ uninteresting solutio.n tt(x) = O everywhere. Also ~ sin ka and cos ka}--
~ ~{~ cannot] be zero for a given value of k, Ther foro there are two possibl
classes 'o( solutions. For the fírst class
A = O, cos ka = O,
so that
ka=mr/2,
where nis an odd integer.
For the second class, B = O,sinka = O, so that ka= (mr)/2, where
n is an even integer. Thus two classes of eigenfunctions and their energy
eigenvalues are
;a
,,.(x) = Ecos mrx, k,, = nm s; = ft2 ( 7r2 ) n2 ij
2a 2m 4a2
J fA
n
tJ odd~.55a)
. n1rx n2 ( 1r2 )
Un ( X ) = A sin--, E,,= n2; neven. (4.55b)
~ 2a 2m 4ª2
Combining the above values of E,, we see that the possible energy levels
are
n2 ( 1r2 )
E,, = 2m 4a2 n2 , n= 1,2, .... (4.56)
- -- - l
f
1
- - - a X
T Vo
1
1
1
1
1
l
1
1
1
66 Quantum M echanics
L
6
(::2 - K2)u1,a(x) = O (4.61a)
i.e.
D=±A.
Thus we have
u3(x) = Ae-Kx L (4.65a) )
u1(x) = ±Ae-Klxl. (4.65b)
Thus there are two possible types of solutions of definite parity, the ones
with even parity
u3(x) = Ae-Kx, u1(x) = Ae-Klxl
u2(x) = B' coskx L (4.66)
and the ones with odd parity
u3(x) = Ae-Kx, u1(x) = -Ae-Klxl
'
u2(x)=C'sinkx. (4.67)
For each solution we have two continuity conditions
u2(a) = u3(a)l-_ )
u;(a) = u;(a). (4.68)
Since an overall constant factor remains arbitrary unless determined
by the normalisation condition, this means that there are two conditions
to determine one constant. Thus the energy E must be regarded as an
adjustable parameter. The continuity conditions can only be satisfied for
certain discrete values, En, of the energy. The conditions ( 4.68) give for
even parity solution
Ae-Ka = B' coskal (4.69a) J
AK e-Ka= B'ksinka. (4.69b)
This can be satisfied only if
k tan ka = K ( even parity). (4.70)
Similarly Eq. ( 4.68) give for odd parity solution
A e-Ka = C' sin ka (4.71a)
-AKe-Kª = C'kcoska (4.71b)
which can be satisfied if
kcotka = -K (odd parity). (4.72)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
68 Quantum Mechanics
If we substitute Eqs. (4.58) and (4.59) for K and k, Eqs. (4.70) and (4.72)
respectively determine the energy eigenvalues En corresponding to even and
odd parity solutions, The equations can be solved graphically for En. In
fact we notice from Eq (4.58) and (4.59), that Ka and ka can be combined
to give
2 2 k2 2 2mVa2
K a+ a=--- (4.73)
f¿
which is the equation of a circle in the (Ka, ka) plane with a radius ~.
Thus K and k of the energy eigenfunctions or eigenvalues are determined by
the intersections of the plots of Eq (4.70)[for even modes] and Eq (4.72)[for
odd modes] with the circles given in Eq. (4.73). Hence classically bound
states when treated quantum mechanically yield discrete energy levels.
4.6 Problems
lt
4.4 Consider a rectangular potential barrier 4 volts high and 10-9 me-
ters wide. Calculate a rough value for the probability for an elec-
tron of kinetic energy 3 voltf to peneti-;,te the barrier.
4.5 A particle of mass m, moves in a potential (with energy E< V)
V(x)=oo x::;O
V(x) = O O< x < a
V(x) = V a::; x::; b
V(x) = O x >b
the energy has a value as obtained in Problem 4.1, find the relative
intensity at x = b and x = a. For the region x > b, show that there
is equal intensity in the beams travelling to left and right.
4.6 Find the transmitivity and reflectivity for a particle of energy E(>
O) scattered by the square well potential of Fig. 4. 7. Show that (i)
for K -=I=- k, T = 1 and R = O for energies E given by
n21r2 lt2
E=-V+---
8ma2
(ii) T has a minimum value for energies
E=-V+----
(2n + 1)21r21t2
32ma2
(iii) as E -+ oo, K -+ k, T -+ l.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
70 Quantum M echanics
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 5
SIMPLE HARMONIC
OSCILLATOR
5.1 Introduction
F = BV = -mw2x (5.1)
ax
so that
1
V(x) = -mw2x2 (5.2)
2
The equation of motion is
d2x 2
m- =-mw x (5.3)
dt2
with the solution
x = acos(wt + 8) (5.4)
representing an oscillator motion of angular frequency w, time period T =
2; and amplitude a. The energy E of the oscillation is the potential energy
at the extreme position x = ±a so that
71
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
72 Quantum Mechanics
p=
•ts
-z,iax·
a (5.10)
It is convenient to change the variable x to y:
fmw
y= y----¡¡:-x. (5.11)
Now
(5.15)
(5.22)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
74 Quantum M echanics
and
n m
(5.24)
n
(5.30)
Thus
(5.35)
and
(xlx) = ICnl2(Un-1lun-1)
= ICnl2· (5.36)
Hence from Eqs. (5.21) and (5.36)
ICnl2 = n. (5.37)
Selecting the phase, so that Cn is real
e; = vn,. (5.38)
Hence
(5.39)
Similarly, we get
(5.40)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
76 Quantum Mechanics
we get
1
Nn = .JnIº (5.42)
Hence
(5.43)
(5.47)
l_
o
2
which has solution e=F. Therefore we write
1 2
un(Y) = e-,y Hn(Y) (5.48)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
~ H~-2yHb=O
Thus either Ho(Y) = a constant or
r--------"
H"
H~ = 2y
o
implying ln Hb(y) = y2 or
Hb(y) = eY2
But this solutio~i· not possible as it blows up for y--------, ±oo, thus Ho(Y) =
i.,..)L constant which take as A0• Thus
1 2
uo(Y) = Aoe-2Y • (5.50)
u3(y) = v'6(at)3uo(Y)
1 1 {) {) {) 1 2
= v'6/8(-0Y +y)(-0y +y)(-0Y +y)A0e-2Y
Ao , 2
v'fJy(-3 + 2y2 )e-'fY . (5.53)
78 Quantum M echanics
Hn(-y) = -Hn(Y)
Hn(-y) = Hn(Y) (5.55)
n = O, 1, 2 ... (5.56)
The lowest energy state n = O, or ground state is not zero and has the value
!ñw; it is called zero point energy. After this energy, the energy levels are
equally spaced. The normalized energy eigenfunctions are given by
Un
1
(x ) = (2ny'nn! )!(mw)! -(~)x2H
ti e n
(
y¡mw)
----¡¡:x . (5.57)
The eigenvalues and the eigenfunctions for n=0,1,2 are shown in Fig. 5.1.
Eq. (5.57) follows from Eq. (5.54) by putting y= ~x and the relation
f
00
y2 1C.
i.e. e-2 H¿ and e- 2 Hi¿ are orthogonal. This is not surprising be-
2
cause e-~ H¿ being energy eigenfunctions and eigenfunctions of an ob-
servable form a complete orthonormal set. The normalization constant for
1C. -1
e- 2 Hn(Y) is (y'n2nn!)2.
Thus
(5.60)
From Eq. (5.60):
Ho(y) = 1
H1(Y) = 2y
H2(y) = 2(-1 + 2y2)
+ 2y)e-s2+2sy = ~
LO
(-2s Hn(Y) nsn-1
L n!
n=l
which gives
_2 ~ Hn(Y) 8n+1 + 2Y ~ Hn(Y) 8n _ ~ n Hn(Y) 8n-1 =O
L n! L n! L n!
Q0l
n n n
Comparing s" term:
Hn+l - 2yHn + 2nHn-l = O. (5.61)
We now derive Eq. (5.58), by using the generating function, mm Eq. V
(5.59) .
.1
-oc
S(y, s)S(y, t)e-Y2 dy = ¿ ¿ :~~:
n k -oc
.1 Hn(y)Hk(y)e-Y2 dy. (5.62)
Joc
e-(s+t-y)2 dy
-oc
= e2st,/ii
= '-~ (2str (5.63)
v1r L n.1 ·
n
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
80 Quantum M echanics
(5.64)
-(X)
u,
Fig. 5.1 Wave functions of stationary states of the harmonic oscillator far n = 1 and
n= 2.
iu/
average of
lu n12
x=
V{,i:y=
~
p =-in-= o
A "f-
V¡--ft(a+at)
-;¡:;;;;;;
• ~ o = -i ,~fUJ.J(
-ivmfUJ.J- -- a- a t) .
L (5.66)
(5.65) )
ox 2 {)y
The energy eigenfunctions un(x) of a simple harmonic oscillator form a
complete orthonormal set, so that any arbitrary function '1.i(x, t) can be
expanded in terms of them
'1.i(x, O) = w(x) = ¿
n
Cnun(x)
L (5.67)
(5.68)
)
n
If '1.i(x) is normalized:
J w*(x)w(x)dx = l.
Thus
J¿ ¿ n k
C~Cku~(x)uk(x)dx = 1
or
or
n
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
82 Quantum Mechanics
Thus if a system is in state W(x, t), the probability of obtaining the result
En is equal to ICnl2, The average value of x in the state W(x, t):
= L L C~Ckei(n-k)wtXnk
n k
where
auk
atuk =
= vkuk-1
vk + luk+l·
L )
Hence, we have
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and
5.5 Problems
OL
5.2 Show that
(P)n = J u~(x)(-iñ:x)un(x)dx = (un[fiun) =
}
o
5.3 Show that for a simple harmonic oscillator
1 1
(V)n = 2(n + 2)!u..J
1 1
(T)n = 2(n + 2)fu..J
(T)n = (V)n Virial r-i:
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
84 Quantum Mechanics
L
b.p = [((p- (p))2)]!
5.8 Show that the Schrodínger eqution for a simple harmonic oscillator
in momentum space is given by:
82 2 p2
ap2<P(P) + mñ2w2 (E- 2m)cp(p) = O.
5.9 Obtain the ground state and first excited state wave functions </Jo(P) - °"'\
and cp1 ([) in momentum spaceL L_)
Answer
</J1(P) = (i 1r
l
(mñw)3
) ~ e-.!f e-2!:w.
86 Quantum Mechanics
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 6
ANGULAR MOMENTUM
6.1 Introduction
87
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
88 Quantum Mechanics
Now
[Lx, Ly]'lf; = (LxLy - LyLx)'l/J
= -li2 ((y!_ - z!_ )(z!_ - x!_) - (z!_ - x!_)
f)z oy OX f)z ox f)z
f) e )
x(y02 -z0Y) 1/J
2 81/J 821/J 2 821/J 821/J fJ2'!f;
= -li (y ox + yz 8z8x - z 8y8x - yx f)z2 + zx f)yf)z
021/J 021/J 2 021/J o'lf; 021/J
-zy OXOZ + xy oz2 + Z OXOy - X f)y - XZ ozf)y)
(6.6c)
'
The relations (6.6) can be written symbolically
LX L = iliL. (6.7a)
Relations (6.6) can also be written as
[Li, Li] = in.éijkLk, i,j, k = 1, 2, 3. (6.7b)
Since Lx, Ly, L; do not commute with each other, we cannot determine
them simultaneously. In other words it is not possible to assign a fixed
direction to angular momentum.
Proof. Suppose there exists a simultaneous eigenfunction 1/J of Lx, Ly,
t., with eigenvalues Ax, Ay, Az. Then
Lx'l/J = Ax1P
Ly'lf; = Ay'l/!
Lz'l/J = Az'lp
are satisfied simultaneously. We have
Lx(Ly'lf;) = Ay(Lx'l/J)
= AyAx1P·
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ANGULAR MOMENTUM 89
Likewise
Ly(Lx'l/J) = Ax(Ly'l/;)
= AxAy'l/;#1
?
L o
so that
(L,L, - L,L,),J, - o
But
therefore
Hence
Similarly
o
Ax= Ay= O L
Hence we cannot determine simultaneously the eigenvalues Ax, Ay and Az,
o
the exception being the eigenvalue zero for each component of angular mo-
mentum.
(2) We now show that the magnitude of the angular momentum L2 =
L; + L~ + L~ commutes with t.; Ly, Lz.
Proof. [L2, Lz] = [L~, Lz] + (x --+ y)
= Lx[Lx, Lz] + [Lx, Lz]Lx + Ly[Ly, Lz] + [Ly, Lz]Ly
= Lx(-inLy) + (-inLy)Lx + Ly(inLx) + (inLx)Ly
=Ü.
Similarly
(6.8)
This means we can determine L2 and one of the components of L, say L;
simultaneously, that is to say we can simultaneously determine the eigen-
values of L2 and Lz.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
90 Quantum M echanics
(X, y, z)
Then we have
t; = -iñ:</> (6.9a)
t; = :e
iñ (sin</> + cot cos </> :</>) e (6.9b)
ANGULAR MOMENTUM 91
so that
í\ = sin(} cos epi + sin sin </Jj é + cos (}k
l )
dr = drr1 + r sin O( - sin epi+ cos </Jj)d</J + r cos (}d(}( cos epi+ sin </Jj) + r( - sin B)d(}k.
Compare it with
dr = drr1 + rdB01 + r sin (}d</J<P1.
Thus
~1 = cos B(cos epi+ sin </Jj) - sinBkL
92 Quantum Mechanics
giving
L
2
= =ñ
2 ( 1 a .
sin() a/sm () 8())
a + sin12 () 8cp2
a2 )
·
So Eq. (6.9d) is proved.
or
e2immr = l.
ANGULAR MOMENTUM 93
(6.15)
We write
'1/J>.m = R(r)Y>.m(8, </;), (6.17)
where we have separated the dependence on r, since neither L¿ nor L2
depend on r. We can then write the eigenvalue equations (6.12) as
LzY>.m(8, </;) = mriY>.m(8, cp)L (6.18a) )
L2Y>.m(8, cp) = >..n2Y>.m(8, cp). (6.18b)
Since L¿ depends only on </; and its eigenfunctions are characterised by
if>m(<P) we can write
(6.19)
Substituting Eq. (6.19) in Eq. (6.18b) with L2 given in Eq. (6.9d), we have
1 8 ( . n 8) 1 82 p (/)) imcf,
---7- ( sin 8 88 sm 17 88 + sin2 8 8cp2
)
>.m 17 e
La'v\ = -AP>,m(8)eimcf,
s óJV'1l., / º N.. or
1 ·,Nl, 1, .._
(6.20)
The dependence on <p has been eliminated. This is an eigenvalue equation
for x, which determines the possible eigenvalues of L2• Since eigenfunctionl _5
are properly normalised
j IY>-ml dn 2 = 1. (6.21a)
Now
dO = sin 8d8d</;.
This leads to (using Eq. (6.15))
94 Quantum Mechanics
or
IC~1.l2 = .! Y;'m(L:¡:)(L±)Y>-.mdD
= .! Y.x"m(L2 - L; =f nLz)Y>-.mdD
ANGULAR MOMENTUM 95
Hence
(6.26b)
First taking the positive sign, we have
,\-m2 +m 2: O
or
1
.\+42:(m-1/2)2. (6.27a)
(6.27b)
Jc?ii· :ºL
/C>-m2 - O.
(6.29a)
(6.29b)
}
L
Hence [c.f. Eq. (6.26a)J
Therefore
,\ - m? - m1 = O, x - m~ + m2 = O
o
(6.30)
so that
Therefore
[The other solution m1 = m2 - 1
o
is not possible as m1 > m2]
1
=v\+1/4-2. (6.31)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
96 Quantum Mechanics
-l <m $_ l l
o'
Thus l must be a positive integer or zero. Then from Eq. (6.30)
>.=l(l+l), l=0,1,2,···
so that eigenvalues of L2 are l(l + l)ñ and we write eigenvalue Eqs. (6.18)
LzYim(O, <P) = mñYim(O, <P) h.. (6.32a) )
L2Yim(O, <P) = l(l + l)ñ2Yim(O, <P), (6.32b)
where
m = O, ±1, ±2, · · · , ±l
= -l, -l + 1, · · · , o, 1, 2, · · · , (l - 1), l.
That is (2l+l) different values in all. From Eq, (6.26a) choosing the phases
of e¡;,
to be zero, we get )
e¡;,= Jl(l + 1) - m(m ± 1) ñ 7
= J(l =f m)(l ± m + 1) ñ. }---/ (6.33)
Now Eq. (6.25) becomes
L±Yim(O, <P) = J(l =f m)(l ± m + l)ll.Yi,m±l (O, <P) (6.34) "'
With >. = l(l + 1) and writing
Pr satisfies
Px,« as P1m, we'ti. see from Eq. (6.20) that
V
v
~() dd()(sin()dP~(())) + (l(l + 1) - ~: )P1m(O) = O (6.35)
sm sm ()
which is known as Legendre's equation and P1m(O) are called associated
Legendre's polynomials.
Hence from Eqs, (6.15) and (6.19) normalised simultaneous eigenfunc-
tion of L2 and L¿ are given by
Yim((), e/>)= N1mP¡m(O)<I?m(<P)
= (-1r((2l4 :(~)]:)7)!) 112Pt(cosO)eim</>. (6.36)
ANGULAR MOMENTUM 97
or
(6.37)
where w = cosO, Pt(w) are known as Associated Legendre Polynomials.
It may be noted that for m = O, we get
o ) 1 d1 2 )
(6.38)
P¡ (w = 21z! dwl (w - 1) = P1(w '
which are called Legendre Polynomials.
98 Quantum M echanics
One final remark: Classically angular momentum vector may have any
orientation1, In quantum mechanics it is restricted to values for which its z-
component'Ís an integral multiple of This fact is sometimes illustrated on
ñ.
a vector diagram. Consider, for example, l = 2 for which l(l + l)li2 = 6n2.
Possible values of m are -2 - O 1 2 and possible orientations of L are
then illustrated as in h figure helo (the radius of the semi-circle is v6n) 'f'
2ñ
Oñ
-2ñ ------
Fig. 6.2
ANGULAR MOMENTUM 99
6.6 Problems
[L+,V+J =0 l l
[z; V+J = liV+.
Lz(V+YíL) = (l + l)h(V+Yii)
L2(V+Yi1) = (l + l)(l + 2)ñ2(V+Yi1).
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
6.2 Show that when the system is described by a state function 'lj;(r),
n
f:::.Lxf:::.Ly ~ 21 (Lz) 1
where
(f:::.Lx)2 = ((Lx - (Lx) )2)
(!:::.Ly)2 = ((Ly - (Ly))2)
and (Lx) and (Ly) denote the average values of L¿ and Ly in the
state 'lj;(r). If 'lj;(r) is normalised eigenfunction of Lz, with eigen-
value mñ, what is f:::.Lx f:::.Ly?
6.3 Find the commutator
to show that
[Li, Xj] = +ihEijtXt ~ )
[L2,x¡J = (-ih) LCijdLixj +xjLi]
ij
which gives
= __!_((2l + 1)!)1/2
ª 21l! 2
so that
pt-l(IJ) = 2~l! ((2l; l)!)1¡2sinl(IJ). (6.48)
Now from Eq. (6.47), taking the upper sign and changing m ---t
m-1
p1m(IJ) = ((l - m)(l + m + 1))-112(-1) sinm+l g d(c!slJ)
X (sin-m+l(IJ)P¡m-l(IJ)).
Using this relation twice, we obtain
-1/2 d
Pt(IJ) = ( (l - m + l)(l + m) ) (-1) sin'" IJ d(cosB)
x sin
. ma d(coslJ)2
d2 (. -m+2apm-2(8))
sm l •
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 7
MOTION IN CENTRALLY
SYMMETRIC FIELD
7.1 Introduction
H = --'v
r,,2
2µ
2 + V(r). (7.2)
iño'lj) = H'lj)
8t
= (-~:'v +V(r))'l/J.
2 (7.3)
103
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(7.6b)
L is the magnitude of the orbital angular momentum of the particle. Sub-
stituting Eq. (7.6) in Eq. (7.5), we have
n2 _!_~(r2~) +
( - 2µ L2((), </>) + V(r)) u= Eu. (7.7)
r2 ar ar 2µr2
Thus for a central field of force, the Hamiltonian operator can be written
in the form
H = - ri2 _!_~(r2~) + L2((), </>) + V(r). (7.8)
2µ r2 ar ar 2µr2
Since L2 and L¿ do not depend on r and since
[L2,L2] = [L2,Lz] = O
it follows that
[H,L2] = O, [H,Lz] =
"
O
Thus it is possible to find simultaneous eigen functions of H, L2 and L¿ .
But we know that
.,/
l o
/')
J
L2Ytm(B,</>) = l(l +l)lfY¡.,,,(6,11) /_
L,Xtm(B, </J) = mhYtm(O, </i) L
Therefore we can write the solution of Eq. (7.7) in the form
u(r, (), </>) = R(r)Yzm((), </>), (7.9)
Hence using Eq. (7.9), we get from Eq. (7.7) the differential equation for
the radial part
In this form in H = T +V
n2 a2
T=---+----
t(l + 1)ñ2 (7.13)
2µ 8r2 2µr2
Thus one has a simple physical interpretation
ñ2
---
82
---+-
fJ; (7.14a)
2µ 8r2 2µ'
where Pr is the radial momentum and then the transverse momentum Pt is
given by
p¡ l(l+l)ñ2
(7.14b)
2µ 2µr2
so that
(7.14c)~
Remarks:
(1) Element of volume in polar coordinates is r2drd0. with dO. =
sin ()d()dcp. But Yim ( (), </)) are normalised over dO.. Hence normali-
sation condition on R(r) is determined by
(7.15a)
j lx(r)l2dr = l. (7.15b)
(2) The condition on the wave function R(r) is that it should be finite
everywhere. Since R(r) = x~), we require that x(r) - O at least
as fast as r - O. In particular x(O) = O.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
x V::r,
= K = zt y"2µ, E= -€2. (7.17)
Equation (7.16) now becomes
d2x
dx2
+ (K _ x
l(l+l)
x2
-c2) x=O. (7.18)
(7.19)
Therefore x --+ e±"x for large x. We must distinguish between two cases
E > O and E < O. We consider here only the case E < O. The solution
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
08 2cs + 2d - K 2c
Cs-1 = s2 + (2l + l)s ---+ ----¡
l
relationship between the coefficients of successive powers of x is
o
Let us compare with the known series
1
1¿
s.(2cx)ª
(7.27)
8
First we note that energy is quantized and that energy levels do not
depend on l and m. nis called principal quantum number (n = 1, 2, 3, · · ·).
r,-;
¿@
l
z2a2(µc2)
o
En == - , n = 1, 2, · · ·
2n2
o =~
2
i callad fine structure constant a ~ 1/137 lJ
is dimensionless.
In particular far hydrogen atom Z = , and ~nceµw.cleus i very heavy
1..i......J
as compared to lectron, 1.-::__
l""uu-vuu9 .!. ¡l mo + ffiN
= .L -1- ~ _!__
m. (µ ~ me)· Henc
for hydrog n atorn ¡1, can be replaced by me· ote mec2 ~ 0.5llMeV the
rest energy of electron.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(7.31)
J R';1r2dr = l. (7.34)
R nl = N nl (Kx)l e -c:xL21+1
n+l
(Kx) , (7.38)
n n
where
Nnl
-- Bnl
C o (- 1 ) [(n
(n/K)l
+ l)!]2 ( n - l - 1 )'(. 2 l + 1 )'. (7.39)
Ío
F" e -y y
2! (L21+1( ))2 2d = 2n[(n + l)!]3
n+l y y y (n - l - 1)! .
(7.44)
Using this relation, the normalisation constant Nnl can now be found and
is given by
2Z 3(n-l-1)!)112
( (nao) 2n[(n + l)!]3 (7.45)
Nnl = - '
where
ñ,2
mee2
ªº = -- (7.46)
and is called the Bohr radius. (Here we have put µ = me, the mass of
electron as we are taking the nucleus to be at rest.) Since
Ze2 2Zr K
Kx= 2me~r= --, €= -2
,. n ªº
we can write the normalised energy eigenfunctions for hydrogen-like atom
as
Unlm(r, e, </Y)= Rn1(r)Ytm(B, </Y)
Rn1 = - (( 2Z )3 (n - l - 1)!) 1/2 e-(Z/nao)r (2Zr) l L2l+l (2Zr) .
nao 2n[(n + l)!]3 nao n+l nao
(7.47)
The energy levels are given by
z2e2
E----
n - 2aon2 · (7.48)
Radial functions for n = 1 and n = 2 are given below
n=l: l=O
R10 = (Z/ao)3f22e-Zr/ao
1
Yoo = --
J4:;;:
ls level: u100 = (z/ao)312 11} 72e-Zr/ao
corresponds to
z2e2
E1 =---
2ao
n = 2: l = O, 1
R20 = (Z/2ao)3f2(2 - Zr )e-Zr/2ao
ªº
R21 = (Z/2ao)3/2 a::t-Zr/2ao
l = O, m = O
l=l, m=-1,0,-1.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
2s level:
z3/2
U200 = (2 - Zr/ao)e-Zr/2a0
4v2rra~12
corresponds to
J~ 1:1:\~tm(r,(),q>)untm(r,(),4>)r2drsin()d()dq>
Also eigenfunctíons are orthogonal
= ll
00!1rf21r u~'l'm'(r,(),4>)untm(r,(),4>)r drsin()d()dq>=O
! 0 0 0
2
7.4 Problems
7.8 For
p2 A Ze2
H = - + V(r) = T- -
2m r
Show that
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
Chapter 8
COLLISION THEORY
Thus far, we have been considering the Schréidingerequation for the motion
of a single particle in an external field of force. In many problems, such
as the hydrogen atom, we have to consider the problem of the motion of
two interacting particles (nucleus and electron in case of hydrogen atom).
Previously we considered the nucleus as a fixed centre about which the
electron moves. We shall discard this assumption now. We now show that
the problem of the motion of two interacting particles can be reduced to that
of one particle. Consider two particles with coordinates and masses r1 =
(x1,Y1,z1), m1, and r2 = (x2,Y2,z2), m2 moving in a potential V(r1,r2).
The wave function of the system is denoted by 'ljJ ( r1, r1, t). The Hamiltonian
is
H = -21-P1A2 + -21-P2
A2
+ V( r1,A r2A ) · (8.1)
m1 m2
The potential V(i1, rA2) includes the interaction energy between the two
particles as well as other sources of potential energy.
In the Schréidingerrepresentation
f¡,2 f¡,2
H = --2-'vi - -2-'v~ + V(r1,r2) (8.2)
m1 m2
where
a2 a2 a2
'vi =
8X1 2 + 8Y12 + 8Z12
and 'ljJ(r1, r2, t) now satisfies the Schródinger equation
in81P = H'ljJ
at
f¡,2 f¡,2 )
( --'vi
2m1
- -'v~
2m2
+V '1/J. (8.3)
117
I July 9, 2012 14:29 ~ World Scientific Book - 9in x 6in RF-QuantumMechanics I
The interpretation of '!/; is that l'l/Jl2dT1dT2 (dT1 = d3r1, dT2 = d3r2) gives
the probability that the first particle is at r1 in volume element dT1 while
at the same time the second particle is in volume element dT2 at r2.
We now assume that V(r1, r2) depends only on r = r1 - r2, so that
V = V(r1 - r2). In many physical problems, the potential is of this form.
Introduce the relative coordinates
r = (r1 - r2), r = (x, y, z) (8.4a)
and the coordinates of the centre of mass
= m1r1 + m2r2
R ----, R=(X,Y,Z). (8.4b)
m1 +m2
Then using relations like
a ax a ax 8
-=--+--
OX1 OX1 ax ax1 {)X
=-+
8 m1 a
ax m1 +m2 ax , (8.5)
we have
(8.6a)
Similarly we have
(8.6b)
This can have a solution for arbitrary r and R only if the part involving R
and the part involving r are separately and identically constant. Thus we
obtain
or
(8.11)
and
r,,2 1 2 ,
---V
2µ u(r)
u(r) + V(r) = Er - E. (8.12)
Equation (8.11) is the eigenvalue equation for the motion of a free particle
of mass M. This means that centre of mass of the system moves like a free
particle with kinetic energy É and mass M. Let us denote the difference
between the total energy Er and the energy associated with the centre of
mass É by the symbol E (relative energy) so that E= Er - É. Thus Eq.
(8.12) for u(r) becomes
r,,2
--'v2u(r)
2µ
+ V(r)u(r) = Eu(r). (8.13)
This equation is the eigenvalue equation for the energy of the relative mo-
tion of the two particles and is the same as the equation for a particle that
has reduced mass µ in an external field V.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Usually, in solving problems like that of the hydrogen atoms, we are not
interested in the energy of the motion of the atom as a whole, but merely
the energy resulting from the relative motion of the electron and nucleus,
i.e. we are interested in the energy levels E associated with the relative
motion. In practice we shall, therefore, usually solve only the equation for
u(r) and obtain the possible values of E.
For hydrogen-like atorrl., in our previous result of energy levels if we
allow for the motion of nucleus,we have to replace me by the reduced mass
memp
me ----t µ = '
me+mp
where mp is the proton mass. Since me << mp, this introduces a small
correction which however is within experimental errors,
The analysis into centre of mass and relative ~n is essential for the
collision of two particles which is our next topic.
Remarks: (1) µ = mass of one of the particles, say m1 when the other
particle mass m2 --+ oo. (2) Fixed source approximation is a good one,
whenever one particle (the target particle in two-particle collision) is very
massive compared to the other (the projectile).
We have considered two types of stationary states which involve (i) discrete
spectrum of energy eigenvalues and correspond to bound states in which
the par lcl i · restraíned by the potential to a particular región oí spac<}{ )
(íí) f°ntin.uous spectrum of energy eig nvalues. The latter sítuation arises
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
M = m1 +m2
1 1 1
-=-+-
µ m1 m2
r = r1 - r2
R = m1r1 + m2r2
M
( as shown previously), the Schrodinger equation is separable and we can
write the wave function
'l/;(r, R, t) = u(r)U(R)e-i(E+É)t/ñ. (8.14)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
@
122 Quantum Mechanics
-:: V2
--VR
2M
u+t
u(r) and U(R) then satisfy the equations
n2 2
V=
=EU.
•
u L (8.15)
(8.16)
}
Equation (8.16) implies that centre of mass moves like a free particle. Equa-
tion (8.15) describes relative motion of two particles. The results obtained
from Eq. (8.15) therefore describe the collision with reference to a system of
coordina tes moving with the centre of mass ( called centre of mass system)
and must be transformed appropriately to give the motion of particles in
system of coordinates fixed in the laboratory in which the observations are
made. The laboratory frame is defined in which one of the particles ( called
the target particle) is initially at rest. Take m2 to be target particle and@)
m1 to be the bombarding particle. rl
We define the appropriate quantities in the two frames as follows: 2' 1 ( 1>
Let q1 and q1 be initial and final momenta of the bombarding particle in
the laboratory system. Let q2 be recoil momentum of the target particle
in this frame. Conservation of momentum gives
(8.18)
or
P1 = -P2 = P, P~ = -p; = P 1 · (8.20a)
Relative velocity v is given by
V=---=-,
P1 P2 P (8.20b)
m1 m2 µ
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
/
/
p =p
• 1. - --f--') 0_: ---1-----;e•m_L
m1 I P2=-p
I
I
I
(a) (b)
Fig. 8.2 (a)- Scattering of two particles in the laboratory system. (b)- Scattering of
two particles in the centre of mass system.
Also
L (8.22)0
or
(8.23) D
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Take q1 along z-axis. Then from Eq. (8.21) we see that p is also along
z-axis
ql = (0,0,q),p = (0,0,p).
From Fig. 8.la,µnd Eq. (8.22) we have
q1, cos fh M q
= m1 + p , cos ºeL )
q~ sin é'¿ = p' sin Be.
We now confine ourselves to elastic scattering IPI = IP'I = p. Then from
the above equations, we have
sin Be
tanOL = e m !l.
cos e+7tf-P
sin Be
(8.24a)
- cosOe + ~ ·
m2
The azimuthal angle </J about the bombarding particle is the same in both
frames
(8.25)
Equations (8.24) and (8.25) give the relations between the two frames for
the elastic scattering. Ll
Remarks: (1) In the lirnit m2 , aboratory frame is equivalent to <:]'1,J.. ¡ 1.
\:9ntre of mass frame and ~ll-at..aegl . the recoil of the target is a very /\e.,ilJ(..11"-q
reasonable approximation íf mi/m2 << l. (2) When m1 = m2, as in the O
case of protonjneutron scattering; the distinction between the two frames
is e sential, Then
1
Te=
2TL
sin Be
tan é'¿ = O = tan(Oe/2) (8.26)
cos e+l
that is
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
In the centre of mass frame, the Schródinger equation for stationary states:
ñ,2
[--V2
2µ
+ V(r)]u(r,O,</J) = Eu(r,O,<p). (8.27)
2
In a scattering process, an incoming beam of particles of energy T = ~µ is
moving in z-dírection. As it enters the scatt ring región, it is deviat d from
its original path {scattered) and is detected in a detector ,b'ig. (8.2 Th
incoming beam is repre ented by aplane wave eikz, p = lik. The outgoing
ll<r
scattered wave i r presented by an outgoing spherical wave !,:--. Therefore
in the region V = O:
1 ikr
u(r,O,</J)-----tA[ eikz +f(O,</J) -e ]. (8.28)
r--> oo '-.,,,-"
incident wave
..._,_,
r
outgoing scattered wave
The above quantity has dimension of area; f(B, </>) has dimension of length
called scattering amplitude.
Now
!~ = 1 f ( (), </>) 2
1 : differential cross section (8.34)
a= j(!~)dD j lf(B,</>)l
= dD
2 : total cross section. (8.35)
<°&~
~
,_____ • z-axis
Fig. 8.3 The incident plane wave and the scattered wave moving radially outwards.
( ddu)
H
n
L
= (~ + cos Bc)2[1 + 2~ cosBc + (~ )2]!
( m
1 + m~ cosBc
)
(du)
dn e.
H
(8.38)
b S. R
Pñ <R
p -
or
e S. P: = kR.
If the energy of the beam is sufficiently small
kR < 1, l < l.
e at low energies, P = O (s-wave) is relevant.
(8.39)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where ag = (2f + l)il and it(kr) is the spherical Bessel function which
is regular at r = O and has asymptotic behavior r -----+ oo : sin(kr - k\
R. e-'!f (e'"r ( l)te--,k~)
; )= 2~; . Further we hav replaced Legendre polynomial
Pi ( cos B) bycherical hannoníc:
Yeo = ·
7f
P,(cos 8). (8.41)
J ikrcosOy0~d0 = ~L R.
de---+radt(kr)f
2f + 1
Yo~YeodO
o
1 .
= ~L e JuTIªOt(kr)ólD
2f + 1
= ~aojo(kr)~ (8.42)
Now
J eikrcosOy;*
00
dO = _1_
v,i7r ¡ eikrcos(} sineded,I,
'I'
=
1
--21r-[e' 1 ºk
e-•ºk r].
r - (8.43)
v,17r ikr
Therefore
a J.
0 0
(kr) = _l_[eikr - e-ikr] (8.44)
2ikr ·
Hence for s-wave
eikrcos(} = _l_[eikr - e-ikr]. (8.45)
2ikr
The same result can also be obtained by the following simple argument. ¡J _
Since _s-wavescattering is isotropic
wave is
Hf
= O component in the incident ) TIJ,-
The effect of scattering potential can only be to alter the outgoing wave
ékr. Hence for s-wave:
Seikr _ e-ikr
u(r) rv 2ikr (8.47)
1í
sio2 ó =
k2 7í
(2sin8)2
k
L (8.52) o
Hence (2 t 8)
is effective radius of target. Since sin ó ~ 1,
41r
cr<
- -k2 (8.53)
i.e. the upper limit for s-wave scattering at an energy corresponding to k, d
is purely geometric and i · independent oflli!j;eraction potential. '1 NL,
Finally we note from q . (8.4.8)and (8~50):
L ó
\J\."-K"
~~
Then from the prutu4 of S-mattix vizl
s = 1 + 2ikf
SS* = 1
(8.54)
(8.55)
o
I July 9, 2012 14:29 World Scientific Book - 9i n x 6in RF-QuantumMechanics
we have
1 + 2ik(f - f*) + 4k2lfl2 = 1
e'·k reos e ~
E -.
u +-1[e'·k r - • .kr]Pe(cosB).
(-Iye-• (8.57)
r->oo 2ikr
e
After scattering, one expects that only outgoing spherical wave would
change its phase. Thus
2e + 1
( e Le -.2ikr-[e 2'ºte'
i s: "k •
r -(-Iye-• "k
r]Pe(cosB)
®Now
k
(T = J lf(B)l2dn = J IJ(e)l221rsinOde
Hence from Eqs. (8.62) and (8.63), we obtain the optical theorem:
tion shows that the the contributions of different angular momenta are not
additive in each direction but may interfere with the each other.
We note that the dimension of cross section is that of area. The cross
section is measured in barn;
lbarn = 10-24cm2 = 10-28meter2. (8.66)
For a given V(r) and given energy, the asymptotic form of u(r, O) in E
(8.58) determines 8e in much the same way that the boundary conditions
determine the energy levels of a bound system.
Finally we derive an expression for 8e in terms of the potential. For
a spherica~mmetric potential, the simultaneous eigenfunctions of Hamil-
tonian and angular momentum L2 and L; are Re(r)Yem(B, </>). Hence the
general solution of Eq. (8.27) can be written in the form
00
u(r,B) = ¿Ae(k)Re(r)Pe(cosB).
i=O
Now the wave function u(r, B) for each partial wave is
Ge
Re(r)Pe(cos(}) = -Pe(cos(})
r
where the radial wave functions Re(r) and Ge(r) satisfy the differential
equations.
1 d(2dRe) ( 2µ f(f+l)) ) ,,
k2r2 dr r dr + 1- fi2k2 V(r) - k2r2 Re= O~ .67)
)
1 d2Gt, ( 2µ f(f + 1))
k2 dr2 + 1- fi2k2 V(r) - k2r2 G¿ = O. (8.68)
For the free particle, the differential equation for R1 and G1 are:
~j_
x2 dx
(x dR1) + (l -
2
dx
f(f+
x2
1)) Rºe O = (8.69)
(8.70)
where x = kr Fil· ·t, we note that Eq. ( .69~ is the Bessel equation,
J
l l@ S tOi whose solution. vhich is regular at r = ~ is the spherical Bessel function
je(x) = jt,(kr): •••
(kr)e
R1 ( r) = u ( kr) -----+
kr_,o
-----'---"--
l.3.5 ..... (2f + 1) (8.71)
dG1 - G0R.dG1¡
G e---¡¡;:- dr r o _
- - f r
cº(')2µV(r)G
e r ñ,2 e (')d'
r r. (8. 73)
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(8.74)
Ggº( r) -
-1 sin(kr - br
-) (8.77)
k 2
up to and overall phase factors ei(óe+~) and e'~", which cancel out, on Í.Í'\
both sides of Eq. (8.74). Using Eqs. (8.76) and (8.77), we get from Eq.
(8.74)
®
.
sm82 = - 2µ
li2 k J
00
o
G2(r)V(r)Gg(r)dr
o
=
o
~z; k J
00
where Ímax "' ka. Only the terms for R: O :=::; f < ka will contribute. Hence
the method is easier to apply for small ka. ~
Now the solution of the Bessel Eq. (8.69), which is regular at r = O, }_ i$ Ü
• R. 1 R ( 1 d ) R. sin kr
J1.(kr) = (-1) kR.+l r -; dr -r-. (8.81)
The second solution of the Bessel equation which is irregular at r = O is
k ) = (- 1 )e-i±__
n¡, ( r kR.+1
(!~)R.
r dr
coskr
r ·
(8_82)
For s-wave scattering in the Born approximation, we have from Eqs. (8.79)
and (8.81) for f = o,
as = 2µ
r,,2 J V(r)r 2
dr
. (kr) ---
Je
..,_(k_r_)t,--
kr--+0 (2€ + l)!!
L )
nt(kr) ~
. (2e - 1)!!
- kH2rt+l L )
j¡,(kr) -
kr->oo
~ sin(kr -
kr
1
br)
2
br
L )
n¡,(kr) - -- cos(kr - - ). (8.85)
kr=roo kr 2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
~ 84:ütotl (8.87) C)
We have so far discussed elastic scattering for which
Se= e2ióe (8.88)
ISel2 = 1 (8.89)
and phase shifts 8e are real. In many scattering experiments, there can be
inelastic scattering viz. the target may also get excited or change its state,
or another particle may emerge. This means that there is an absorption
of the incident beam in such experiments and if we describe the scattering
process as a one channel picture, then 8 is complex. This is because
(Flux)in -(Flux)out = Absorption
(Flux)¡n -=/:- (Flux)out (8.90)
In order to satisfy the above condition, we write
8e -------, 8e + if3e ( 8e real)
Se(k) = e-2/3e(k)e2ióe(k)
= r¡e(k)e2ióe(k)
o
(8.91)
ISe(k)l2 =
ISel2 -=/:-1
ISel2 s1 (8.93)
or
o < r¡¡ < l. (8.94)
The partial wave scattering amplitude ft(k) is now given by
k) = Se(k) - 1
fe ( 2ik
we2ióe - 1
=
2ik
r¡e sin 28e . 1 - r7P. cos 28e
= 2k +i 2k . (8.95)
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics l
aabs =
47!" '°'
k2 1 L..,(2f + 1)(1 - ISel 2 )
4 e
= ; ¿(2e + 1)(1- rJ}). (8.97)
e
Thus the total cross section is
But
ªtot = ael
=;
+ aabs
~(2e+1)[2-2r¡ecos28e] ~
/
(8.98) o
Im ft(k) = 1- cos 28e
r]t
2k
Imf(O) = ¿(2e+ l)Imf¡(k)
e
1
= ~(U+ 1)(1 - 'f/e cos 28e)
2k
k
= (8.99)
471" ªtot•
Thus the optical theorem is indeed satisfied and is a consequence of con-
servation of flux or probability.
Now aeu is maximum, when
(8.lOOa)
(8.lOOb)
(8.lOOc)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
a-el,i = k2
tt
(2í! + 1) L (8.lüla) J
a-abs,R
7f
= k2 (21! + 1) L
The incoming and outgoing waves are coherent in elastic scattering (because
Se enters in a-e1 as opposed to ISel2 in a-abs) and therefore can interfere
(8.lülb)
o
constructively or destructively.
We see that when Se = O or tte = O, we have total absorption. Never-
theless there is still elastic scattering in that partial wave. This becomes
clear in scattering by a black disc Fig. .4 . Let a be the radius of the disc.
Now only those partial waves are important for which
l :S L = ka
and we shall take here ka > > l. Write í! = pk, so that p :S: a. Thus for the
black disc limit í! < L (p < a), there is complete absorption:
tte = O, e-2f3t = O,
Se= O.
a-el = ; t,(21!
e=O
+ 1) ~ ;L2 = 1ra2 L )
s, = O e < L, p <a
St = 1 e > L, p > a, (8.103)
hence
dael 1 L 2
dD = (8.104)
4k2[¿(2R+l)Pt(cosB)[.
l=O
Now for high energy ka > > 1, L = ka and we replace the summation by
the integration
1
X= kp--+l+-
2
1
2
dael 1 ka
dD = k2 xdxPx ( cos B) (8.105)
1 1
Pn(cosB) = Jo(nsinB)
= Jo(nB), (8.106)
1
2
do"! 1 ka
dD = k2 xdxJa(xB) (8.107)
1
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
1 2
= k204 JkaBJ1 (kaB)J
k2a4JJ1(kaB)/kaBl2.
= (8.109)
A plot of d:;~1 as a function of () is shown in ig. .5. ow
duel ,
dO. is maximum
da"'
dQ
Fig. 8.5 The incident plane wave and the scattered wave moving radially outwards.
'ljJ(t) = o, t < o
'1/J(t) = '1/J(O)e(-iEot-~t)/", t~O (8.110)
(8.111)
T is the mean life time, t1;2, the half life of resonance. Now
(8.112)
f
00
'ljJ(E) = eifl-'ljJ(t)dt
o
= 'ljJ(O) f
00
i(E-Eo+i~Hdt
=
o
ih'l/;(O)
(E- Eo) +i2
r
L )
l'I/J(E)l2 = l'!/J(O)l2ñ2 r2. (8.113)
(E- Eo)2 +4
The probability density for finding the particle with energy E:
00
= l7/J(O)l2n2 iJ -oo
1ró(E - Ea)dE
(8.117)
From Eq. (8.115):
2 r (8.118)
17/J(O) 1 = 21rn2.
Hence
P(E) = l·'·(E)l2 =-.!:._ l (8.119)
'P 21r (E - Ea)2 + ~2 •
Eq. (8.119) gives the natural line shape of decaying particle. Now P(E~
has its maximum value ~r at E = E0. At E = Ea±~' P(E) = ir.
P(E) as a function of E given in Eq. (8.119) is shown in the ig, 'ffe.6ft
r is the width of resonance.
A striking feature of a scattering process is~earance of high, narrow
peaks in the scattering cross section expressed as a function of energy. The
peaks are called resonances.
Partial wave scattering amplitude is
Put
cot' óe(E)I
(E=Eo)
= -i
so that near resonance
(8.122)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and
1
fe(k) = k[-~(E - Eo) - i]
r
- -2 (8.123)
- r·
E-Eo+i2
At resonance
P(E)
-r2 _
1t
¡- - - -1- - - -,
, r, 1
1 1
1
1
Fíg. 8.6 Natural líne shape of a decayíng state. r is the full wídth at half maxímum.
e
If the resonance occurs in partial wave (i.e. angular momentum of res-
onance is €), then near resonance onlyl_ih partial wave in expansion of
scattering amplitude given in Eq. (8.59) is important, so that
The partial wave amplitude ft(k), near resonance has the form in Eq.
(8.123). Hence near the resonance
(8.126)
and then
(2P + 1)(-f) (2f + 1)(-~)[(E - Ea) - if]
(8.127)
f(O) = k[(E - Ea) + i~] k[(E - Ea)2 + r4 2]
This is called Breit.{Wigner formula. If both the projectile and the target
has spin J1 and Jz respectively, then
(8.130)
The solution ar
uo(x) = Aeikx + Be-ikx L (8.136) ,)
u1 (x) = Ce.il<a: + De-il<a: L (8.137) )
u2(x) = Feikx. (8.138)
T(E) = 1 ~ 12 = IS(E)l2
4
- K (8.141)
4cos2 2ka + (--¡;; + Kk )2 sin 2 2ka ·
The S-matrix S(E), regarded asan analytic function of ~ has pole~when ,) / )-
. K k .
2cos2ka - i( k + K) sm2ka = O. (8.142)
The solutions (i) and (ii) give the bound states of even and odd parity
respectively on negative E-axis.
In the neighborhood of a resonance E; > O, the S-matrix:
1
S(E) = . =} ±1 (8.147)
cos 2ka[l - !( f + f) tan 2ka]
so that
tan 2ka --+ O
cos 2ka --+ ±1 (8.148)
as E--+ Er, By Taylor expansion near resonance a function
(8.149)
In our case
f(E) = K
(k + K)tan2ka
k L )
f(Er) = O. (8.150)
Put
4
(8.151)
r
so that near resonance
S( ) ±1
E= [1-!i(E-Er)]
±r ±·r
22
= 2 = (8.152)
f-i(E-Er) (E-Er)+if'
Thus S ( E) appears to have a pole at
r
E= E; -i- (8.153)
2
9.-,X'.. ~ shown in the Fig. M8.6~ From Eq. (8.147)
/ cos 2K a + ! ( 1 + f) sin 2K
S ( E ) = ~~~~~~~~~~~~
[cos2 2Ka + i( f + f) sin2 2Ka]
= IS(E)leio(EJ (8.154)
where
1 K k
tan 8 ( E) =
2 ( k + K) tan 2K a. (8.155)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
V(x) = Vo8(x)
Schródinger Equation is
ñ,2 d2 ]
[ - 2m dx2 + V(x) u(x) = Eu(x) (8.158)
or
d2 2m
[ dx2 + k2 ] u(x) =-----¡; V(x)u(x)
2m
=-----¡; Vo8(x)u(x) (8.159)
where
k2 = 2mE
ñ,2 , E> O.
First we determine the boundary conditions for the 8-function potential.
Integrate Eq. (8.159) between the limits O - E and O+ E:
2~Vo
Je d~2 u(x)dx + 7ek2u(x)dx = 7e8(x)u(x)dx, (8.160)
O-e O-e O-e
which gives
d O+e IO+e 2 V,
....!:!'.. + k2u(x) = ~u(O) (8.161)
dx I O-e O-e ñ
or in the limit E--+ O:
du)
( dx O+é
_ (du)
dx O-e
= 2mV0 u(O).
ñ.
(8.162)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
. . 2mVo
C(ik) - ik(A- B) = ~u(O)
= 2mVoc
ñ,2
L (8.166) ()
From Eq. (8.165) and (8.166):
e
A
ikñ2
ikñ2 - mVo
L (8.167) )
B = C _ 1 = mVo (8.168)
A A ikñ2-mV0
Hence from Eqs. (8.163) and (8.164):
T = 1º12 = E (8.171)
A E+ mV02
21i2
mV02
R = 1-B12 = 21t2 2 (8.172)
A E+ mVo
21i2
R+T=l.
The S-matrix
S(E) = C = Coefficient of transmitted wave
A Coefficient of incident wave
ikñ2
=---- (8.173)
ikñ2 - mVo ·
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ikñ? - mVo = O
or
For
E<O, E=-IEI
k = iK, K = J ":i~EI.
2 (8.175)
Boundary conditions, namely u continuous and the one given in Eq. (8.162)
give
2m
-KA-KA= -V
ñ,2
0A (8.178)
so that
Hence
where Ep = :m.
2
It is convenient to write the olution of Eq. (8.181) as
The subscript on '1/J indicates that we are using boundary condition for t <
ti. Substituting Eq. (8.183) in Eq. (8.181) and using that Haup(r)=Epup(rl,
we obtain
~;ñ(ó.)pp.(t)up(r)e -•:•, ~ ~app,(t)Vup(r)e=':"• L o
By multiplying on the right by u;, ( r) and integrating over d3r and using
the orthogonality relation
we obtain
iñapp;(t) = L ª~PJt)eT(Ev-E,,)t(p'IVI-+ p) (8.185a)
p
where
gives
t
ª~'p, (t) = OiJp, +~ j dt ¿ (pjV)p)e ,,_' (Ep-E,,)f¡ app, (t). (8.186)
t« p
For t < ti, V = O and for t = ti, the integral vanishes and we recover the
boundary condition. This is an exact result. We now assume that V is
(\ . ,,..
small perturbation . Then to the zeroth order
.._¡ V
a'PP; (t) = Opp,
app, () t ;::::;J
Á
Upp, + Ín1 Ít-t, t '"'(
L.., p
'IVI p ) eT
-i(E
P
-E·)tÁ
P Upp,
t, p
=o·PP; + -in1 ¡t t;
dt(iílVlv-)eT
i
-i(E
P,-
E) t
" K. (8.187)
Taking t--+ oo, p--+ P¡, we have [since V vanishes for t < ti]
(8.188)
Now the probability that the interaction V causes a transition from the
state IPi) to a state lv') is given by
Ppp, = 1 j d3ru;,'lf'i(r, t{
which on using Eq. (8.183) and the orthogonality relation (8.184) gives
P.¡;,P, = la~p/· (8.189)
Thus for p = P¡ =/= Pi, we get from Eq. (8.188) transition probability
amplitude:
ªPiP, = -21rio(Ep, - Ep1)(v11Vlvi)
= 21rio(Ep, - Ep1 )T¡i (8.190)
so that the scattering matrix in Born approximation is
Ts = -(P¡IVIPi)
= - j u; (r)V(r)up,(r)d r
1 3 (8.191)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
TB = ---
1
(2rrñ)3
J ipf.r
e- " V(r)e_n_d3r
ipi.r
where
q= (Pi -p¡) = ñ(ki -k¡) = ñk (8.194)
is the momentum transfer.
Now TB has dimension 3, E~:
where as scattering amplitude J((}, </J) has
Now
q2 = PT + PJ - 2Pi·P¡ = 2p2(1 - cos (})
Thus, we have
__ i:
- ñ,2
J [ékr _ e-ikr]
ikr
2
r V(r)dr
00
On the other hand for large velocities or high energies, from Eqs. (8.194)
and (8.196), k = ñs~;
ª-, and in Eq. (8.199) for fB(B) it appears in the
2
denominator. Thus the scattering is mainly through the small angles:
sin J~J
and
io = --2µ 1 00
V(r)
sin(~rB)
r2dr
0
h
ñ2 o (~)Br· -
= <l>(p B), say ~ (8.201)
Thu · the scattering cross-s tion
Note that for small B, sin(),....,() and for large B, sin é] < 1 and cI>(p B),....,
!. Thus 0Jtegral over () converges so rapidly that the integration can be
extended to oo without any great error. Thus putting ~() = x,
1
®
i.e.
a= -2-
p =:
o
xdxl<I>(x)l2
ñ,2 ñ,2
(J"(X-=--
p2 2µE
or
lim (c,E) = constant, (8.203)
E->oo
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(8.205)
For electron scattering (ignoring its spin) with a point nucleus of charge
Ze:
(mec2)2 z2e4
(8.211)
4(cp)4sin4 ~ ·
Taking into account the spin of electron, the cross section was calculated by
Mott. The Mott scattering cross section for scattering of electron of spin
1/2, on a spinless point target of charge Ze is given by:
v2 . 2
2e ) .
da ) 2 2 E2 ( (8.212)
( dfl Mott = 4(Ze ) (cq)4 1 - c2 sin
For ~ << 1, E------> mec2, it reduces to (~~)R. For high energy electron
scattering i.e. ~ ------> 1: E------> cp
d ) E2 2 o 2o
_.!!___ = 4(Ze2)2 cos 2 = (ze2)2 cos 2 . (8.213)
( dfl 4E4 sin4 fl.2 4E4 sin" fl.2
Mott
! [cpV'!/J - '1/JVcpifS = O.
~-
-----
(8.217)
(Y
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
For
iq.r
efJ =e--¡¡;-, and 1/J = V(r),
we have from Eqs. (8.216) and (8.217):
= -
q2
n,2
J ,q.r
e----¡¡;-V(r)d r.
3
(8.218)
T = - (2 :n) J e'1'V(r)d r
3 3
1
= ---(--)
(21fn)3
n,2
q2
J iq.r 2
e----¡¡;-'v V(r)d r
3
Ze2 1
= ~2
21f ¡¿ q
J ,q.r
e " p(r)d r,
3
(8.219)
(8.220)
and
(8.221)
where
p(r) = (_..!_ )3
21f
Je-'':;' F(q )d q. 2 3 (8.223)
The form factorbq2) denotes the structure of the target. p(r) gives the
charge ofthe target. For low momentum transfer (small q2), we can expand
(8.224)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(8.225)
Now
.,..
j q.r dD. = O.
o
Thus, we have
Hence we have
(8.227)
(r2) is called the mean square radius. Electron spin can be taken into
account, by replacing (1~)R in Eq. (8.221) by (1~)Mott"
For most of the potentials, no exact solution is available. As such one has J
to resort to approximation techniques. The Wentzely¡[ ramersñsrillioum or \J.. W L-,
WKB approximation is specially useful when one is dealing with slowly
varying potentialsg what we mean by such potentials will become clear as
) we develop the WKB approximation.
The Schri:idingerequation for stationary states for a particle of mass m
moving in a potential V(r) is ~ ~
and
i.e.
A dS
= (8.238)
w2 dx
and using the condition (20.192), Eq. (20.191) then gives
A dS
2 = -d = ±nk(x),
W X
(8.239)
i.e.
S(x) = -sn {'1' k(x')dx' (8.240)
Íxo
and
A ) 1;2
w(x) = ( nk(x) , (8.241)
(nk1x))
112
{ C1 exp(i 1: k(x')dx')
+ C2 exp(-i 1: k(x')dx')}
u(x) =
C'
e -f"'=o lk(x')ldx1
E< V(x). (8.243)
~ ,
The values of Xi for which E = V(xi) are called classical turning
points. They correspond to those points in space where the classical par-
ticle comes to a halt [k(xi) = O] and turns back. The wave function given
in Eq. (20.196) or (20.197) become infinite at these points and the WKB
approximation breaks down. The problem then is to connect the solutions
(20.196) and (20.197) on two sides of a turning point. Such connection
formulae can be derived as follows.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
V(x)
a - - - - - - - - - b- - E
~---------------x
Fig. 8.7 The classical turning points a and b. The particle is confined in the region
a :S X s
b.
Suppose the potential V(x) has the form shown in Fig 8.7. There are
two turning points a and b. Away from the turning points, the solution is
given by the WKB approximation. Since the particle is confined between
a and b, the wave function is real and in the WKB approximation is given
in this region by [c.f. Eq. (20.196)]
u(x) = ~ sin(r¡(x) + a), (8.244a)
where
r¡(x) = ¡x k(x)dx, (8.244b)
and a is an arbitrary phase factor.
In the region x > b, the wave function is given by Eq. (20.197)
u(x) =
V
C'
!Ti:Texp(-
lkl
1"'
b
lk(x)ldx), (8.245)
In order to find the solution near the turning points, we solve the
Schródinger equation exactly by using the following approximation for the
potential V(x). Near the turning point x = a, we expand V(x) in powers
of (x - a) and retain only the first power, Jl&e'l.~e--solutimc near
~1:.ning..p.o.mts, we oJve the..Schr"c-lingm·--equatmñ exactl b using the
~g approximation for the pe nttitt Vtx). ear the turning point )
9r= et, we expand~)-tn--pew , ·~ t:arrrnrrty-'cl'iefirst power,-
8V
V(x) = V(a) + ax ix=a(x - a)
= V(a) - F(x - a), (8.247)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
(8.248a)
(8.248b)
where y = x - a.
Using the above approximation, the Schródinger equation near x = a
can be written as
d2u
dy2 + 2mFyu = O. (8.249)
By writing
2
u(y) = ylf2cp(V2mF3y3f2) = y1f2cp(r¡), (8.250)
Using Eqs. (20.208) and (20.209), w see that for large r¡, the solution
(20.206) takes the form of Eq. (8.244ae4b), i.e.
t
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
u= IYl1;2</J(lr¡I) - -~ A2 (2mFlyl)-lf4e-lTTI
2 V'ff
-- 3 A2
----e -ITTI (8.257)
2 y0rlkf
provided A2 = -B2. Now for large r¡, the u= y112<jJ(r¡) and u= IYl1/2</J(lr¡I) 4
should go over to the WKB solution given in Eqs, (8.244al8.244b) and if
(8.246) respectively. We see that this is so if [._,.
3~ = c,« = n/4,C" = -~ ~-
Now the wave function should be continuous at y = O. This condition gives
-1A2 = A1. Hence we have the final result that the WKB solution to
the right of the turning point a is given by
u(x) = e
míi:T
v2lkl
exp(- 1ª
X
lk(x)ldx) . (8.259)
In the region a :=::; x :=::; b, the particle is confined in a potential and the
energy spectrum is discrete. The WKB solution to the left of the turning
point b is given by
u(x) =
C'
~ sin(
yk
¡b
X
k(x)dx
tt
+ -)
4
. (8.260)
Now the wave functions given in Eqs. (8.258) and (8.260) must be the same
in the whole region a to b. This requires that
{b
l
x 7r 7r
a k(x)dx + 4 + lx k(x)dx +4= (n + l)1r
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
or
l a
k(x)dx = (n + l)rr - -
2
b 1
or
l ª p(x)dx = (n +
2)rrñ
J:
where § p(x)dx = 2 p(x)dx is the integral taken over the whole period of
quasi-classical motion of the particle.
We now consider the radial part of the Schródinger equation
d2 X
dr2
2m
+ ,if (
E - V(r) - 2m
n l (l + 1) )
r2 x(r) = O. (8.262)
This equation is identical with the Schródinger's equation for one dimen-
sional motion in a field of effective potential
V,(
l r) - r +2
+2 1) .
_ V( ) !f_ l(l (8.263)
r m
Hence the solution of this equation in the WKB approximation can easily
be written in the same way as discussed earlier.
We now apply the WKB approximation to the problem of tunneling
through a barrier. Let us consider a particular example for which we take
V(r)=-Vo r «a
= O r > a, (8.264)
so that
Vi(r) = -Vo + !f_ l(l + l)
2m r2
ñ2 l(l + 1)
r > a. (8.265)
2m r2
loside the potential well we ignore the a. guiar momentum. Now if we
r plot Vi(1·) ver us 1·, we have a situation r - ·. mblin_ g hat shown in Fig. 8, . In
the región a< x < bf/b is the second turning point at whieh E- lf¡(r) = O),
the WI<B wave function is given by
exp (- J lk(r)ldr) ,
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
V(x)
where
lk(r)I = ( l(l ~ 1) - 2; E) 1/2 (8.266)
The transmission coefficient T is given by
T = IFluxloutsidebarrier
IFluxlinsidethe well
X exp (-2 lb lk(r)ldr)
a
!
transmission is given by Eq. (8.267). The number of times per secondg the
particle hits the wall of the potential well is given by
1/2
(E_)_!._= E+Vo .!__ (8.268)
m 2a ( 2m ) a
Hence the total probability of transmission is given by
.!_=(E+ Vo)1/2 T
T 2ma2
= (__!!___)
2ma2
1/2 exp (-2 Jb (l(l + 1) -
a r2
2m E)1/2 dr)
n2
(8.269)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
b = li l(l + 1) _ (8.270)
2mE
The integral occurring in Eq. (20.221) can be expressed as
la
b (l(l + l)
r2
-
2m
f¿2
E) dr = )l(l + 1) f'Y
1( 2_ - 1)112dx,
x2
(8.271)
where
X=
l(l
2mE
2mE a
+ l)n2 r L (8.272) )
'Y= (8.273)
l(l + 1) Ti:
For a special case when 'Y<< 1, the integral
(8.274)
Hence
2) 1/2
T = 2~a "(-2\f'!(l+l)
(
2x
ddr2 + [2µn, (E - V(r)) ] X= O.
2
In the WKB approximation, the wave function x(r) for s-wave bound states
is given by
where r1 is the first turning point (k(r1) = O). The above solution is to the
left of the turning point. Now
1/2
k(r) =
[
~ (E - V(r)) ]
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or for
1 [2µ(E - V(r)]1/2 drQ(n- i:)1rñ, (8.277b)
lo
r k(r)dr
1 = /2µ, (E)312
n e lo
¡1 (1 - x)112dx
= /2µ, .'!:___(E)3/2
ñ. 3(3 .
Hence from Eq. ( 8. 277a, 277b), we have the energy levels for s-states
(8.279)
(8.280)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
Forme= 1850 MeV and f3 = 1/(b2nc) with b = (1950 MeV)-1, one gets
from Eq. (20.230)
In order to evaluate the integral, we assume that the average value of the
sin2 term is !.We have
1
N2 -
2
¡r10
-dr=
k(r)
1 N2
-
2
¡r1 [ 2µ
2(E-f3r)
ñ
]-1/2
dr
0
=1 (8.282)
This gives
N2 = ,/2µ,~(E)-1/2. (8.283)
= N2 E3 r,¡:;- r(1/2)r(3)
2(33 y 2µE r(7 /2)
8 E2
- 15 (32. (8.284)
(8.285)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
8.13 Problems
eikr
u(r, (), cp)------, A(eikz + f((), cp)-)
r--+oo r
represents the incident plane wave and the scattered wave radially
moving outward. Let S denotes the probability current density.
Show that
S0ut = v3JAJ2lf((), cp)l2r + higher terms which can be neglected for large r.
r
Calculate S0ut ·d~ where Jd~J = dY:, is the surface element r2dD..
Hence show that
SouJ_° d~ = lf((),cp)J2dn = da,
in
8.2 Consider the scattering of a beam of particles by a hard sphere of
radius a:
V:: {~~·L
Show that the phase shiít Ót is given by
jp_(ka)
tan Óp_ = np_(ka),
where ie and np_ have the usual meaning and k is the wave number.
Find the cross section when a<< 1/k. For a>> 1/k, show that
Óp_ ~-ka+ br/2.
8.3 Consider the scattering of a particle of mass m by a hard sphe~re
of
radius a
a~óo
- - tan(ka)
·a- tan ka
tanó1
· L + ka tan ka ,
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
80 ~ -ka,
Show further that
da
drl "'a2
( 1 - -3-
(ka)2 + 2(ka)2 cos () + O(ka)3 ) .
Ze2
V(r) = --exp(-r/a).
r
Find also the total cross section .
. 6 Consider the pot ntial
V=Vo Jfor
= () ¡ for r > a.
Find the scattering amplitude f(()) in the Born approximation.
8.7 A solution of the equation
('92 + k2)Gk(r) = 83(r)
is given by
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
C+ +C- = l.
Verify that
'1/J(r) = eik·r +~ J Gk(r - r')'I/J(r')V(r')d3r'
is solution of
ñ2
(-
2µ
v'2 + V(r))'I/J(r) = Ek'l/J(r),
where
Ek
nk2
= -2.
2µ
Show that in the limit of zero energy the asymptotic form of 'ljJ is
'lp rv 1-~
T-HX) rl
where
1 00
f(O) =
2ik
¿(2l + l)(e2icl"e - l)Pg(cosO).
l=O
The Born formula for the scattering amplitude f (O) is
fB(O) = - 2µ
r,,2 lo
r= sinqr
qr
V(r)r2dr.
n2 k lot" V(r)j¡(kr)r2dr.
Jf ~ 2µ
Since for f > fmax, the phase shifts are small, we can use the Born
amplitude as a device for summing up the partial wave series for
f > fmax· Using the above statement show that
f(O) -
2:k
'}~'t2t + l)((e"'' - lf (e'"f - l))P,(eosO)+ fn(O).
1 (I_All ~h <l-wv t ~
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
l)
COLLISION THEORY 171
~Hint: V=O
=00
?
/8.1 and P = -g~'t[V here is volume].
Show that for s-wave bound states, the wave function at the origin
is given by
d2x 2µ
dr2 +[,i2CE-V(r))]x=O.
.!!__( "/ ) = 2µ dV
dr X X ñ2 dr
For a bound state, the wave function x(r) is real. From the above
equation, we have
f 00 [ d ( "/ )] 2µ f 00 dV
lo X dr X X xdr = ,i2 lo X dr xdr.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Now
we have
Hence we have
= (2µ/n,2)(R(r)ldV/drlR(r))
ar
l1/Js(O)l2 = (2µ/41rri?)(dV/dr).
This result is useful far the interpretation of the newly discov-
ered resonance J/1/J at 3100 MeV. This resonance is regarded as
a s-wave bound state of charmed quark and antiquark (ce called
charmonium). It is possible to determine 11/Js(O) 12 from the decay
rate J/ 'ljJ ----, e- e+. Then using the above relation, we can get infar-
1\ • mation about the potential V(r), which keeps ce bound together.
~ Since the mass ofkharmed quark is high (1500 MeV-2000 MeV), it
is a good approximation to treat the ce system non-relativistically
with µ = 1/2mc,
8.12 Using the WKB approximation, show that far the s-wave bound
states far the logarithmic potential V(r) = Clnr/r0, the energy
eigenvaluesare given by
1
En = C ln[(2n -
2 )vn] + C ln[(l/ vm:c)l/rol,
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
COLLISION THEORY 173
Chapter 9
OPERATORS
o
We now define the scalar product as
L
This being a scalar product is a number, in general complex. The com-
pl x conjugate of < c/Jl'l/J > is defined as
Operators
:
OPERATORS 177
But, by definition
(v,tF)* = (Ftv,).
Therefore we can write
(Ft = (<tit.Jt. (9.10)
One can easily prove that
c.Jt)t = A
(A+.B)t=.Jt+_¿t
(AA)t = A* .Jt •
where A is a number, and
(A.B)t = _¿t_Jt L
If .Jt = A, A is called hermitian. A is called unitary if .Jt A = .J.Jt = l.
Recall that an observable in quantum mechanics corresponds to a hermitian
a
operator. Thus for a hermitian operator A,
(<tit.Jtv,) = (1/JtAt<ti)*. (9.11)
Definitions:
Two state vectors t<ti) and tv,) are said to be orthogonal if
(<titv,) = o (9.12a)
or
(1/Jt<ti) = o. (9.12b)
A state vector is said to be normalised if
(v,tv,) = l. (9.13)
Eigenvectors:
Theorem:
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Proof:
Eigenvalue equation is
so that
Therefore
or
Hence for
(i) m = n, an = an *, (anlan)-=/- O
(ii) m -=1- n: We must have (amlan) = O
i.e, they are orthogonal. If the eigenstates are normalised:
In this case
m=n
m-=f-0.
r4-. Completeness:
OPERATORS 179
(9.16)
Therefore
(9.17)
Substituting in Eq.(9.15a)
(9.18)
n
(a'l'l/J) = j C(a)(a'la)da
= j C(a)(a' - a)da
= C(a')
or
C(a) = (al'l/J). (9.20)
Substituting in Eq. (9.15b)
j la) (a Ida = i.
Thus for any set of eigenvectors la), the completeness condition can be
formally expressed as
Sala)(al = i, (9.22a)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
S¿ = ¿ if a= an, discrete
and
(9.25)
n m n m
and similarly for the continuous case. If & and {3 are two linear operators
then
L(nl&lk)(kl/-Jln)
k
n m
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
OPERATORS 181
(9.26c)
n
Thus the weighting factor, which appears in the calculation of the average,
namely,
(9.27)
gives the probability of a resultan in a measurement of A on Jvi).
By the use of the completeness relations, we can easily see that the
quantum mechanics of the previous chapters followsas a particular case of
í\ the present formalism.
\ Finally for a Unitary operator Ü with eigenvector Ju) and corresponding
eigenvalue >.:
ÜJu) = >.Ju)
implies
(uJüt = .X*(uJ
and since
ütü = 1
(uJu) = (uJütüJu)
= (ul.X* >.Ju)
= I.XJ2(ulu)
i.e.
1.x12 = l.
Thus a unitary operator has eigenvalues which are of the type eílJ.
which we denote as
'lj;(x) = (xl'l/J) (9.29)
and call this the state function in x or the Schrodínger representation. Then
'lj;*(x) = ('l/Jlx). (9.30)
The expression (xl'l/J) may be thought of as the component of a state vector
l'l/J); when x is a continuous variable, we have an infinity of components
which run together to form the state function
'lj;(x) = (xl'l/J).
Note that the average value of repeated measurements of position is given
by
x,¡, = ('l/Jlxl'l/J)
= jj ('l/Jlx')(x'lxlx)(xl'l/J)dxdx'
= JJ ('l/Jlx')xó(x' - x)(xl'l/J)dx' dx
= J l(xl'l/J)l2xdx. (9.31)
This shows that l'l/J(x)l2 = 1 (xl'l/J)l2, determines the probability density P(x)
of the particle at x.
Let us now express a,¡, in the Schrodinger or x-representation
a,¡,= ('l/JIAl'l/J)
if we write
(xlAlx')
A
= ó(x' - x)A A ( /8 ) .
x, ox' (9.33)
Example:
The momentum operator p in the Schródinger Representation. We have
the quantum condition
x'fJ-fix = in. (9.34)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
OPERATORS 183
Then
(xl(xp - fix)lx') = iñ(xlx'). (9.35)
Now
(xlfixlx') = x' (xlfilx') (9.36)
(xlxfilx') = (x'l(i:fi)tlx)*
= (x'lfixlx)*
= x(x'lfilx)*
= x(xlfilx'). (9.37)
Therefore Eq. (9.35) becomes
(x - x')(xlfilx')~li.(xlx')
Consider now
v+> (9.38)
, J~
/ J \,~
~ .!fa The first term on the ~
9 f(x) Jx8(x - x')dx,
= f(x)8(x - x')l1!_9 - f~9 f'(x)8(x -
-g < x' < g
x')dx.
gives zero, since 8(±g-x')
(9.39)
= O. Thus
[9 f(x)~8(x - x')dx = - [9 J'(x)(x - x')dx
l.¿ OX L; .
= -f'(x'). (9.40)
Therefore
Similarly
f~9 f(x') 8~, ( 8(x - x') )dx', -g <X< g
@ Therefore
of(x')
ox'
of(x')
ax'
8(x - x')dx'
1
x'=x
(9.41b)
= -in :x
(xl'l/J). (9.44)
It is clear from Eqs. (9.43a, 9.43b) and (9.44) that in the Schródinger rep-
resentation
"'t. {)
(9.45)
A
p = -inax.
Eigenfunctions:
The eigenvalue equation
(9.46)
becomes in the Schródinger representation (xlA.lan) = an(xlan)· Using
completeness conditions, we can write
or
J dx' (xlAlx') (x' lan) = an (xlan) (9.47)
J dx' (x - x')Á.(x',
·
~ª ) (x' lan) =
ux'
an (xlan) , (9.48)
i.e.
(9.49)
We write
Uan (x) = (xlan)· (9.50)
The above equation shows that Uan(x) is an eigenfunction of Á.(x, ffx) be-
longing to the eigenvalue an.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
OPERATORS 185
(xl'I/J) = j dp(xlp)(pl1/J)
= -1-/ dpe(ip/li.)x(Pl1/J). (9.56)
V27r1i
On the other hand, expanding 11/J) in terms of eigenvectors of the position
operator
(pl'I/J) = j dx(plx)(xl'I/J)
= -1
~
-f dxe-i(p/ñ)x(xl1/J). (9.58)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore (xl1f1) and (Pl7P) are Fourier transforms of each other. Previously
we labelled (Pl7P) by C(p) and (xl7P) by 1fJ(x). The probability ofthe system
having momentum pf\ when in state 17P) is 1 (Pl7P) 12•
J Examples:
(i) Simultaneous eigenfunctions of angular momentum L2 and z-
component of L, t.;
Using polar coordinates, select the basis vectors to be ¡e, </J). Since L2,
L¿ depend on () and <p only, the simultaneous eigenfunctions of L2 and L¿
are given by
L21Zm)r:ln2Z(Z + l)IZm)
Lzllm) '-G)hmllm), (9.60)
(</Jlm) 1 . "'
-eim.,,_
= (9.61)
J27r
(ii) Discrete eigenfunctions of the hydrogen atom. Select the basis vec-
tors to be Ir, e, </J), since the Hamiltonian depends on all three coordinates
r, e, </J. Then the eigenfunctions can be written as
(arnlan) = Ómn
L lan)(anl = .i (9.64)
n
beco me
or
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
OPERATORS 187
J u~,Jx)uan(x)dx = s.;
¿(xlan)(anlx') = (xlx') (9.65)
n
= 8(x - x')
(9.66)
or
¿ Uan (x )u~n (x') = 8(x - x'). (9.67)
n
H = LL IEn)(EnlHIEm)(Eml
n m
= LL IEn)Emómn(Eml
n m
(9.71)
In this basis
H =¿ ¿ laj)(ailHlai)(ail
i j
= ¿ ¿ laj)Hij(ail (9.74)
j
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and since basis vectors IEn) and iai) form a complete set, one can write
(9.75)
where
u: = (ailEn)
Un/= (ailEn)* = (Enlai)· (9.76)
Now
n
(9.77)
(EmlEn) = L Uni(Emlai)
i
(9. 78)
Hence, we have
Thus one can go from one set of basis vectors to another set by a unitary ~
transformation. The unitary matrix Uni is overlap (anlEn) between two set /~J
\.: .
of basis vector.
Now
and
(9.80)
n
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
OPERATORS 189
Therefore
(EnlHIEm) = ¿¿(EmlaJ)(aJIHlai)(ailEm)
i j
or
a= U,6 (9.87)
where a and ,6 are column matrices
ª1
a2
a= ,6=
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ñ.waat = H + ~ñ.wL )
1
ñ.wat a -- H - -ñ.w
2 . (9.92)
OPERATORS 191
Therefore
(9.95b)
Let In) denote the normalized eigenstate of H belonging to eigenvalue
En:
Hin)= Enln). (9.96a)
Let
IF) = aln)
then
(FI = (nlat (9.96b)
Now,
(FIF) = (nlat aln)
= ¿(nlatlm)(mlaln)
m
Therefore
(nlat aln) 2:: O. (9.98a)
Similarly we can show that
(nlaatln) 2:: O (9.98b)
Then from Eq.(9.98a), using Eq. (9.92), we have
1
(nl(H - 2:: O
2ñw)ln)
or
1
(nlHln) 2::
2ñw(nln)
or
i.e.
(9.99)
Hence the eigenvalues of H are positive and have a minimum given by
(9.100)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
alO) = O. (9.104)
We now construct a system of vectors
(9.105)
From Eq. (9.95b), we have
Hat¡o) = at HIO) + ñwat¡o)
= (.\ + ñw)atJo)
= ~ñwatlo). (9.106)
2
Hence if at¡o) i= O, atJo) is an eigenstate of H belonging to eigenvalue !nw.
Similarly,
HatatJo) = atHatJo) + nwatatJo)
= at(l + ñw)at¡o) + ñwatatJo)
= (>. + 2ñw)atat¡o)
~nwatat10).
= (9.107)
2
Continuing this process we get
(9.108)
(9.109)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
OPERATORS 193
(9.112)
Then
atln) = _1_(at)n+110)
vn!
= _l_J(n + 1)! 1 (at)n+110)
vn! J(n+ 1)!
= v'n+lln+l), (9.113)
where In+ 1) is a normalised eigenstate of H belonging to the eigenvalue
1
En+l = (n + 1 + =En+ hw. (9.114a)
2hw)
Similarly,
(9.114b)
where In - 1) is a normalised eigenstate H with eigenvalue
1
En-1 = (n - 1 + 2 )ñw = En - lu», (9.115)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Thus, we see that a and at are the operators which respectively annihilate
and create energy in the system in units of lua.
We also see that
(mlatln) = -./n + l(mln + 1)
= -./n + lóm,n+{./ (9.116a) }
(mlaln) = yn(mln - 1)
\fnóm,n-1·
= (9.116b)
The above expressions give the matrix elements of at and a in the basis
defined by eigenvectors of the Hamiltonian of a simple harmonic oscillator.
From Eqs. (9.116a, 9.116b) one can also determine the matrix elements of
x and p by using Eq. (9.89).
ln)B ex (att~O)a\ }
HB = (n + )In~ (9.118)
2
We now introduce the fermionic oscillator, which is defined by the operators
d and dt satisfying anti-commutation relations { d, dt} = 1 [seeproblem 9.6].
The Hamiltonian is HF = dt d-!
and eigenstates of HF are IO) F and dt lü) F
so that
dlü)F = Ü. (9.119)
For the combined system, the total Hamiltonian is
H=HB+HF
=ata+ dtd (9.120)
'sanreA uonaioadxa aAr¡:e'.ilau-uou y'!la1 aA'Rl[ ( q)
'u'8mur1aq cll'R ('R)
swa¡qo.1d L"6
' ( :)!J'.)clUIUIASladns)
s-? - 1
g-d +1
1_(g_c> - 1)(g_c> + 1) =
\ 'l:>] = [H'b]
O~-
~ H = {+b'b}
11 {+b'ib}={b'b}
'tllqa'.ay'R
:)p'.)aUIUIASladns '.)S<>Jdtu¡s at¡'.) ªP!AO.td +b pu'tl b pu'R 'parqo :)!UO!UilclJ lO
:)1uosoq amd '8 1a'.auoy ou s(Ej} iPº = +b 'p +v = b srorarodo aq'.) auyaa
'Al'.)auttUAS 'tl SJ <>iaq'.) '.)'Rl['.) sa'.)'8:)!Pª! Afüaua ut k>'81aua~ap
=u ·.10'.)'RH!JSO J!UOUI.I'Rl[ Jp'.)aurur.As1adns ª! sa'.)'8'.)S JO AJ'81aua~ap aq'.) ª'.lºN
(1Z1'6) ... (Tl+P = (zl '(ül+P = (11
... (Ol¡;(-fv) = (zl '(ol+v = (11 '(ol
961 SYO.LV'l!3d0
A2
= \,
'
2m
has a set of eigenstates In) with energy eigenvalues En. The lowest
eigenstate IO) has the energy E0. Show that
n2
¿(En - Eo)l(nlxlü)l2 =
n
2m.
Hint:
[x,ft] = in
[x, [H,x]J = ~-
Verify the above sum rule for a simple harmonic oscillator.
9.4 In the momentum representation, show that the position operator
x is represented by
(plxlp') = in! (PIP')
(plxl7/J) = in :P (Pl7/J)
that is
A
X=
•t,
Z1i8p.
a
9.5 If In) denotes a normalised eigenstate of a simple harmonic oscil-
lator, belonging to the eigenvalue En= (n + !)ñw, show that
Jn/mw(n!1)112, m=n+l
(nlxlm) = Jn/mw('f!J.)112, m = n- 1 (9.123)
{
O otherwise.
9.6 If an operator A has the following properties
Á.2 = o
Á.At + At A= 1,
show that
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
OPERATORS 197
A.tia) = 1
.A_t11) = o.
9.7 Let ÍJ and 6 be two anticommutating operators
(Olx210) = _n__
2mw
9.9 In the three dimensional vector space with an orthonormal set of
7 basis vector {11), 12), 13)} ~
= ~
L
41re2 Zl'l/im(O)l2
where for the Coulomb potential
v2V = 41re2 Z83(x)
9.11 Show that 3 x 3 matrices
e, = li (~ ~i b)
o o o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
OPERATORS 199
jl) - m. m. m.JO) - 1- 1) -
Show that the eigenvector!:M IY) in the basis in which G3 is not
diagonal are:
JG,--1)- ~ G).
Now the matrix which connect the two bases is
Show that
-1 i O )
U=~
(
O o_J2 .
-1 -i O
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 10
HEISENBERG EQUATION OF
MOTION, INVARIANCE
PRINCIPLE AND PATH
INTEGRAL
10.1 Introduction
(10.1)
Now dimensionally
[x][p] = ( Ac~on).
201
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
= j dx(xlHlx)(xlw(t))
where
(xlHlx) = t5(x - x)H(x, =iñ :x) (10.5)
so that we have in the Schródinger representation
(10.llb)
n
Let us write \J!(O) = 'l/;,
(10.12)
n
so that
a(En) = (Enl'l/J). (10.13)
Thus the time dependent solution, which satisfies Eq. (10.6) is
(xl\J!(t)) = L e-iEntfli(xlEn)(Enl'l/J). (10.14)
n
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
We see that all the time dependence is in the state vector. This is called
the Schródinger picture. In this picture, the operators representing the
dynamical variables are regarded as being independent of time. The state
vectors represent the observed systems and these change with time, as do
the results of observations or measurements. Thus the average result of
repeated measurements of an observable A, made at time t, on a system in
the state [w(t)) is
ªwCtJ = (w(t)IA[w(t))
= j dxd±(w(t)lx)(xlAl±)(±lw(t))
a
J
A
= dx(w(t)lx)A(x, ax)(x[w(t))Yl
a
J
A
Then
l
But
eiHtfñHe-iHt/ñ = H(t) )
eiHt/ñ Ae-iHt/ñ = .A(t). (10.21)
Therefore
thus
H(t) = H(O) = H. (10.24)
i.e H is constant of motion.
Thus we can write
iñd~?) = [.F,H] = O.
This implies that
S-picture H-picture
States represented Moving or time dependent fixed or independent
by vectors of time
iñcftl\[!(t))s iñcftl\[!)H = Ü
The Hamiltonian:
(10.27)
ó
Hamilton's canonical equations of motion are
dq, . aH
dt = Qi = api'
(10.28)
Then (we will confine ourselves to the case when F <loes not depend explic-
itly on time)
and the dynamical variable .A(t) and the Hamiltonian are regarded as op-
erators.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(qi(t),pj(t))PB = ~
L.,
( ~~
oqi ºPí - ~~
oqi opí ) = ~
L., (óirÓjr) = Óij· (10.33)
[qi(t), 1/j(t)]
op; uqr
=o= [pí(t),Pj(t)]
r
l )
(10.35)
First we have to construct the states lq, t) and IP, t) which are '1instanta-
((
neous" eigenstates of the Heisenberg picture operators q(t) and p(t):
qlq) = qlq)
e*Htqlq) = qe-ll:Htlq)
ef,Htqe--ll:Hte-ll:Htlq) = qe*Htlq)
q(t)lq(t)) = qlq(t)), (10.36)
where lq(t)) = e-li:Htlq). Now using the completeness relation,
J dqlq)(ql = 1,
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
we get
j dqef.Htjq)(q¡e-f.Ht = e-kHte--kHt
=1
i.e.
j dqlq(t))(q(t)I = 1. (10.37)
Further
while
Then
Further using
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
we get
=
1
21rñ J dpe-K
'p2
2"'
t '
eKPQ, (10.39)
1
2
=- 1 (21rmñ)
__ .!c.!!!J!.:
eñ2 t:
21rñ it
where we have used the Gaussian Integral
Thus finally
id xe -a"'22 = ~7í -.
a
K(t, q) = ( --
m )! "' e-2,1ttq
2
(10.40)
21rñit
which isJ§:.ee particle propagator, Green's function for the Schródinger op-
~ª)
era t or ( -2.,..na2aq2 - i,i8t .
then
('lj!u¡'lj!u) = ('lj!1utu'lj!)
= ('lj!l'lj!) (10.46)
so that transformed state l'lj!u) is also normalised.
If the lan)'s forma complete set of states so that
(10.47a)
n
Then
(10.47b)
n
i.e.
(10.48)
also form a complete set.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
i.e,
Urlx)
Uflx) = Ui1lx)
lx - a)
x + a)
/;lrd
~10.63)
u« 17/J) - 17/JT) (10.64)
7/Jr(x), xl7J;r) ~
= (x1Url7/J) = (x + ai7/J)
= 7j;(x + a). (10.65)
Now by Taylor expansion, for infinitesimal a:
7/JT(x) = 7j;(x + a)
= 7j;(x)
a
+ a &x 7j;(x) + O(a2)
= (1 + a :x + ... ) 7j;(x)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
i.e
= x
A
h ( -ino")
+ ia = x + a.
A
(10.70)
Hence in general
A(x +a)= e~ A(x) -!!j¡-1?. (L0.71)
We now consider several examples of a closed syste~t acteJ hy externa!
forces). The simplest exarnple i that f a free particle f~i h th
tf ¡ 1/\~ G'l\.
Hamiltonian
A2
H= R._ PA=-iñV (10.72)
2m'
is obviously translational invariance:
[f>,H] = O. (10.73)
..U. ;\ Thus momentum is conserved and it corresponds to Newton's first law of
~)'J.- motion~w of in rtia)b classical mechanics.
Th Hamiltonian for an isolated system of particles mutually interacting
with one another:
A2
p
H = -
2m
+ V(xi -xj) (10.74)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(10.77)
Thus it is possible to find simultaneous eigenstates of H and T(a). Let In)
(>.f\ be~gen tate of H with eigenvalue In), n denotes the nth lattice site:
Hin) = Eoln) L (10.7 ) ,)
HT(a)]n} = T(a)Hln) = EoT(a)ln). (10.79)
Hence
T(a)ln) = In - 1) (10.80)
is also an eigenstate of H with eigenvalue E0. We have n-fold degeneracy.
T(a) is unitary but not hermitian, hence its eigenvalues are oftype ei6. The
simultaneous eigenstates of H and T(a) are constructed as follows. Define
the state
()(]
(10.81)
n=-oo
while
()(]
T(a)lka) = ¿ einkªln - 1)
n=-oo
n=-oo
(10.83)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
The translational invariance does not hold for a particle moving in a har-
monic oscillator potential centered on sorne fixed point, since the potential
defines a natural coordinate origin. For such a system momentum is not
conserved.
3- Rotation operator:
Just as there is a relation between the momentum operator p and dis-
placement operator e'~ª, there exis~ re1ationship between the angular s
momentum operator and rotation operator. onsid r a rotation of coordi-
nates
(10.90)
The length
x¡ = Xi + x~ + x¡ = x2 + y2 + z2 (10.91)
is invariant under the rotation of quardinates:
~(D
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
x/ = R.jxjRikXk
= (R)ij(RT)kiXjXk
RTR= 1
(RT R)kj = Ókj
= RRT.
L- (10.93)
)
y= -xsinB + y cose
z= z. (10.94)
For infinitesimal rotation:
x=x- () x x, () = ()e"z,. (10.95)
For a general infinitesimal transformation
(J~~ ~x-wx
~ Xi~ - EijkWjXk L (10.96)
Corresponding to rotation of coordinates, there is a unitary operator UR:
1/JR)L (10.97) )
UR[x) - +w x x)L
.)
Uk [x) = X - W X x). (10.98)
6
to check that L; satisfy the commutation relation,t
[Li, Lj] = i1íEijkLk l.-
Hence angular momentum is the generator of rotation. Intrinsic spin, a
purely quantum mechanical concept with no classical analogue can be taken
into account by replacing L by J with the same commutation relations as
those for L:
(10.102)
(In quantum mechanics, the commutation relations are fundamental enti-
tles). The commutation relations define the angular momentum Ji, Hence _.¡)
the generators of rotation group S0(3) are Jfi,
ch. 11) 4'"
UR = e-fw.J. (10.103)
Rotational invariance implies
which gives
[J,H] =0 (10.104)
i.e angular momentum is conserved. This holds for an isolated system (i.e.
a system not in an externa! field) or a system in a centrally symmetric
potential for which:
:m + V(r)
·2
H = (10.105)
PH&IIP
i.e
[F,HJ=O (10.116)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
so that
dP
dt =0
i.e parity is conserved. This is a type of conservation law which plays as
important role in quantum mechanics and has no classical analogue.
The law of conservation of parity i not universal. There is class of [
in eractions called weak int raction,s-, (responsible for {3-decay)jfor whích e: )
thi law is violated, /
ii- Time reversal
®
Under time reversal
X
p
{}.
J J. (10.118)
o
H nce, under time reversal, the basic commutation relatíons ofthe quantum
mechanics viz L-
[xi, Pj] = iMij L )
[Ji, Jj] = i~ijkh· J >( )_
are no invariant, in con ras to space r fíectíontx=-« -x --+ -p -> ¿ ~ a
J), in which case they ar invariant, Thus an operator corresponding to
time reversal cannot be a unitary operator as it <loes not preserve the
fundamental laws of quantum mechanics. A way out of this difficulty is ,J
to simultaneously change i -----+ -i. Accordingly we wíll talcebmplex <fi.-t_
conjugate of all complex numbers to preserve the invariance of physical
laws under time reversa!. For time reversal, the corresponding operator T
is not unitary i.e. rt -=J y-1. ~
10.7.1 Introduction
There are two ways to effect a transition between classical and quantum
physics:
(1) Canonical quantization
[q(t),p(t)] = iñ (10.123)
which we discussed in sec. 10.2, and originated from Hamiltonian mechan-
ics. As is clear from Eq. (10.34), which hol~· equal time, time plays a
central role.
(ii) Path integral The path integral approach was formulated by Dirac,
Feynman and is essentially tied to Lagrangian formulation of mechanics
based on Action functional.
t¡
(i) A. Zee. Quantum Field Theory, Princeton University Press 2003, Chapter 1.2.
(ii) P. Raymond, Field Theory, A Modern primer, The Benjamin Cummings Publishing
Company, 1981, Chap. JI $.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
We need to construct the states ¡q, t) and IP, t) which are ii,iinstantaneous"
eigenstates of the Heisenberg picture operators q(t) and p(t). This we have
already done in sec. 10.3:
s q(t)lq(t)) ~ qlq(t))~
)
j dqlq(t))(q(t)I = 1.
x= (qF¡e--i\:HT¡q¡)
()
x= rr J
N-1
i=l
dqi(QF¡e-*Hót¡QN-1)(QN-1le-iH8t¡QN-2) .....
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Consider now
I j = ( Qj+l 1 e _.iHótl
" Qj
)
with
A2
.!!__ + V(q)
H =
2m
Introduce a complete set of one particle states J dplp) (PI = 1 so that
t, = j dp(qH1IP)(ple--kót[~+V(qJ¡qj)-
I·
1
21rñ
= - 1 J
dp eKP(qi+1-q;)e-¡¡ót[2=le-,;8tV(q;)_
i i p2 i
(10.128)
we get
I·1 = -
1 ( 21r )2 1
e".iót(!!!c(q;+1-q;)-V(q))
2 ------r.-- , .
21rñ i8t/mñ
Thus
J( =
-im ) N/2 NIT-1
--
( 21rMt
J dq·e
{ót(i1
,= 0
T(qi+J,-qi )2-V(q;)J
• (10.129)
i=l
'
Now take the continuum limit 8t----+ O, N----+ oo
Q+1
J
- q·J ----+ q.
8t
ót~----+
N
l t
dt (10.130)
= J Dq(t)efi- {[!mri2-V(q))
{'f L(q,q)
=
J Dq(t)e T (10.131)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
f Dq(t) = J~ 00 (~;;
. )N/2N-1
!! f dq¿ (10.132)
and
1
o L(q, <i)
is the LagrangiaI:\,f In terms of the action
S([q], t, T)
=
=
2m<i -
f
t
V(q)H
dt L(q, q)
(10.133) /
(10.134)
T
we obtain
K, = (qF, tFIQI, t¡)
= f Dq(t)eiS([q],t,T). (10.135)
This implies that to obtain K,, we simply have to integrate over all possible
paths q(t) such that q(T) = q1 and q(t) = QF, weighted by the exponential
k
of times the action evaluated for the particular path.
In the classical limit O, the integral is determined by the minimum
ñ --,
of S i.e. by the classical path Qc(t) that extremizes the action and must
obey the Eule4!Lagrange equations:
dt o<i
!!:_
áq
(ºL) _ 8L
= O.
"Thus the path integral formalism clearly brings out the distinction between
the classical and quantum mechanics. In the former the particle takes
one pa h to go from qt to QF i.e. qc(t), while all paths contribute to the
latter -4!'.Raymond]"
Finally the transition amplitude from sorne initial state 14>) at t¡ to a
final state 17/1) at t = tF is given by
K,(v,,cp) = (7/1,tFlcp,t¡)
= f dqFdq1(1PFIQF(tF))(qF(tF)lq1(t1))(q1(ti)lct>1)
= f dqFdq¡V,}(qF)K,(qF,tF,QJ,t¡)cp¡(q¡) (10.136)
where
(10.137)
and is called the propagator, since it contains all the information regarding
the development of the system.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
1
2
= 21riMt)
~ e.:=
1 i-m: < qi+1-qj-1 l2
(
where we have used the Gaussian integral (10.128). Thus the :first integra-
tion [j = 1, q0 = q¡] gives
1
2
= ( 47íÍótri)
--- e=ióm (q -q¡)2
3
3m
where we have again the Gaussian integral (10.128) with b = O. Therefore
after (N - 1) integrations, we get
rr
N-1
.
a-J-
j +1
·
= aN-11.....
(N
2 .... N
-1
)
=
N-1
-ª--
N
J=l
6;/;.
where a = 21ri Thus
(10.138)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ZFI(J)
_· )N/2N:}]:;:1! ·8t
dqie'-¡;-[L-j=O
N-1,.,,(qJ+1-qJ)2
=
( 27í~~t 2 -----;,¡---- -V(qj)+J(kj)qj]
'F
i;; .f
dt[!rn<i2-V(q)+J(q)q]
=
J Dq(t)e t t . (10.142)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
(10.143)
N-1
f I: q;J(t;)8t .
We first look ate i=0 • The exponent multiplying ic5t(q0 - q1] is
qoJo(to) + qiJ(ti) ..... qN-iJ(tN-1)
= qo[Jo(to) + q1J(t1) .... + QN-iJ(tN-1)]
+ (q1 - qo)J(t1) + (q2 - qo)J(t2) ...... (QN-1 - qo)J(tN-1)
N-1
= QJ L J(tj) + {(q1 - q1)(J(t1) + J(t2) + ..... J(tN-1)
j=O
]2
N-1{
= ~{ ; q1J(ti)c5t +2 ~t; N-1 [
(qj+l - Qj) +m
ót N-1
k~l J(tk)c5t
lN-1
- ;;t ;
(Ó )2 [
~
N-1
k~l J(tk)c5t
l 2 }}
. (10.145)
I(tj+1) =
N-1
L
k=J+l
J(tk)c5t = JT
t;+1
J(T)dT (10.146)
N-1
I(O) = L J(tj)ót
j=O
(10.147)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Then
- 2~ X; (.E, J(t,)ót)'
ót
= --
1 ~
~ [J(tj+i)] 2 ót = JT dtI 2 (t). (10.148)
2m
j=O O
.
ZFI ( J) = (____!!!!___ )"~
21rMt
rrNI
i=l
dq·e*q¡
'
I(O) e* 2ót [¿ f=o1[Cv+1 -qj )+f;i;J(tj+1)]]2 e--lt 2;. J¿ dtI2(t)
.
(10.149) U
To <leal with the second term in Eq. (10.149), we proceed as in/ee particle <(}/v.__
case and look at
Thus using the Gaussian integration (10.128), the dqi integration gives
(10.151)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ZFI(J) =
m
C1rdiT)
2
1
exp
[ i·
h,
[ m ( qp - qt )2
2T +
qp - qr
T ¡T
O
dtI(t) + 2mT
1
x (
fo T dtl(t) )2 + q¡J(O) - 2~
fo T dtl2(t)]]
(10.152a)
1
= (21r7nT) 2 exp(F].
where
·
F = !._ [m(qp - q¡) 2 + qp - q¡ { T dtI(t) + -1- (
{T dtI(t) )2
n 2T T lo 2mT lo
+ q¡J(O) - 2~
foT dtI2(t)]].
(10.152b)
We note that [see Eq. (10.146)]
where
(} (r - t) = 1 if T > tl )
= Ü if T < t.
Thus
= fot J ( T) dr.
We now show that ZFI(J) serves as the generating functional for time
products of operator ili(t) of the free field (zero potential).
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
If
f [g(x)] = 1b dx [g(x)f,
then
óf(t) {T
óJ(ti) = la drO(r - t)ó (r - t1) (10.157)
=O(t1-t)
and
ó!(O) (10.158)
8J(ti) = () (t1)
= 1 for t1 > O.
Further, on using Eq. (10.157)
óJ( )
t1
Ó ¡TO
I(t)dt =
¡T
O
O(t1 - t)dt =
¡t1
O
=t t
() (t1 - t) dt + 1'1'
ti-t¡
() (t1 - t) dt
=
r:: dt
lo = t1 - t t- (10.159)
Thus
8Z(J) ( m )~ (10.160)
-di óJ(ti) = +ch. 21rinT exp[F].G(t1)
where
G(t1)=
l)
( ¡¡,
[QF T- q¡ (t1-t1)+2mT2l
1 {T
dtI(t).(ti-t1)+q1 0
- .L ¡t dt2I(t)O(t1
2m la
- t)]
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
Hence
-l
0 Ji 1)
1 J=O = J Dq(t)q(t1) exp [~lar dt~mq2 l
= [q¡ + QF - q¡ (t1 - t)] (~) ! eK m(qi;.;qr)2 (10.161a)
T 21rlñT
u
Next we calculate (-lñ)2 1 2[,
T
j \
-
2~
j dtB(t2 - t)B(t1 - t)]
o
where we have used (10.157), (10.159) and (10.160). Thus
o
¡ T
dte(ti - t)B(t2 - t)
0
=
T
¡t2
(tF - t¡)
, dt = (t2 - t1)
since B(ti - t) and B(t2 - t) contribute only when t < ti as well < t2 and
we are taking t1 > t2.
For t1 < t2, interchanging t2 +-+ ti
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Thus using
B(ti - t2) = 1,
=Ü, [)
J
(qF, tplT (q(t1)q(t2)) lq¡, t¡)J=O
ZFI[J] serves as the generating functional for time-ordered products of the
free (zero potential) theory.
10.8 Problems
10.1 Show that for the simple harmonic oscillator, the Heisenberg equa-
tion of motion gives
ina(t) = [a(t), H] = nJJJa(t)
which has the solution
a(t) = a(O)e-iwt = ae-iwt
at(t) = ateiwt.
Hence show that
q(t) = q(O) coswt + p(O) sinwt
mw
p(t) = [p(O) coswt - mwq(O) sinwt].
Hence show that
o
= XiPJ -x·p¡
= EijkLk [
Thus in quantum mechanics
i~ [ L; Lj] L
6
= Eijki,k
[ti,tjJ = infijktk L
10.3 Show that
A -A 1
= B + [A,B] + + . ..
A A A A A A A
e Be [A,B]]
2[A,
Hint: Consider f (>..) = e>.A iJe->.A,
and show that
df(>..)
d);
[,,i, !(>..)]
d2 J(>..)
d>!2 [A, [A, f(>..)J] etc.
and use the Taylor's series
A x df (>..) 1
A >..2 d2 f I
f (0) + 1! d>..! >.=O+ 2! d).,2 >.=O+ .....
f(>..) =
[
x,exp
A (iafjx)J
T /
where a is a real number. Using the above result, prove that
exp ( T iafix) I x') is. .
an eigenstate o f t h e coordimate operator x re- A [
H = - -efX.
2m
Calculate the operators x(t) and p(t) in the Heisenberg picture.
Suppose at t = O, the particle is in the state l1Po) whose wave
function in the x-representation is 1Po(x) = (xl,¡po) = eikx</>(x),
J
where </>(-x) = </>(x) and </>(x) is real and l</>(x)l2dx = l. Define
the uncertainty (~A)t in the observable A at time t as
1
2,
(~A)t = { ((.A(t) - (.A(t))))2}
(x(O))
(x(O)p(O)
= O, (p(O))
+ p(O)x(O))
= ñk,
= o.
(p(0)2) = 2m(E) l )
(ii)
where
is the Lagrangian.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
with 8 = It
dt [!m<i2 - !mw2q2). Calculate Sc1 for this case. clint:
First
show that for S.H.O, q(t) = sinlwT [<ir sinwt + <Jo sinw(T - t)]
For
A2
H = E__ -/Jq
2m
show that
S = /Jq¡ t - f!__t3 + /Jqi t + m( q¡ - Qi)2.
el 2 24m 2 2t
K(q¡,t;qi,O) = m
( 21rhit
)! exp (iSc1)
T .
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 11
(11.3)
where 1rr2 is the area enclosed by the circular loop and n is a unit vector
normal to the plane of the circular loop. Now the circumference of the
237
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
circular loop = 21rr, therefore the charge at a fixed point on the orbit =
e/21rr. The current j is then given by
. e
J = 21rrVe
ep
= (11.4)
Thus
(11.5)
Now
Ir x PI= rp.
It is conventional to write
e
µ=--rxp (11.6)
2mec
where e is the electric charge, of the particle.
But r x p = L, the angular momentum. Therefore
e
µ=--L. (11.7)
2mec
The interaction energy is then given by
Vs = -µ·B
e
=---B·L. (11.8)
2mec
This is a general result, not restricted to a circular orbit. By the correspon-
dence principle, we also regard Eq. {11.8)"1. as the interaction energy of the ~
hydrogen atom with the magnetic field B m quantum mechanics provided /
we regard L as an operator. Thus we write the total Hamiltonian operator
of the system
e
Hs =H- --B·L. (11.9)
2mec
Take the z-axis in the direction of B, then
e
Hs = H - --BLz. (11.10)
2mec
The extra term - e 2BLz in Hs only contains the operator Lz. Therefore
mee
the simultaneous eigenstates of H, L2, L¿ are also eigenstates of Hs. We
have denoted the eigenfunction of H by
Un!m(r, B, </J) = (r, B, </Jllmn) (11.11)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
But
HBUnlm(r, e, <P) = E;;unim(r, (), <P) L (11.12)
6.E~B. = (11.15)
2mec
This effect is observed and shows why m is known as the magnetic quantum
number.
According to Eq. (11.13) or Eq. (11.14), there should be no splitting
of the ground state (n = 1, 1 = O, m = O). However, experimentally the
lowest level is observed to split into two. If the splitting has the same
physical origin as that derived above, it must be associated with the angular
momentum i. which satisfies
2j +1=2
. 1
or J = -
2
Wlith m = -2,1 2.1
Since we have shown quite generally that the orbital angular momen-
tum l can only take integral values, this result indicates the necessity for
generalization of the formulism. The answer is to introduce a generalized
definition of angular momentum and show that half integral eigenvalues are
also possible.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
L= f X f>. (11.16)
Then using the commutation relation
[xi,Pil = iMij
we have shown that L satisfy the commutation relations
LX L =iñL. (11.17)
Also we have shown that for eigenvalues mñ of Lz, mis an integer dueto
the fact that with L defined as in Eq. (11.16) L¿ has the representation
-in!_
8</J
in spherical polar coordinates. We now assume Eq. (11.17) holds whether
or not angular momentum can be expressed as in Eq. (11.16). Thus in
general for the angular momentum operator J, we take the commutation
relations
J X J = iñJ (11.18a)
as the defining property of the angular momentum J. Equation (11.18a) is
equivalent to
[Jx, Jy] = iríJz L )
[Jy, Jz] = iñJx l- )
[Jz,Jx] = inJy, (11.18b)
From Eq. (11.18b) it follows that
[J2, Jz] = O etc. (11.18c)
Thus we can measure J2 and any single component of J which we take to
be Jz,
We now show that in general J2, Jz can have half integral as well as
integral eigenvalues. Only when J = L given in Eq. (11.16) the eigenvalues
of L2 and Lz are restricted to integers.
Let us denote the simultaneous eigenvectors of J2 and Jz by JAm):
J2JAm) = An2JAm) (11.19a)
JzJAm) = mñ!Am) (11.19b)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(11.2lb)
Thus
[J+, L] = 2fiJz. (11.21c)
Also
[Jz, J=f]
(J2,h]
=
=
=r:-ñJ=f
o.
L (11.2ld)
(11.2le)
)
i.e. the eigenvalues form a series in which successive terms díffer by one
unit of ñ. We now show that this series terminates on both sides. Note
that if we write
therefore
1 1 2
(>.+¡) 2:: (m-2). (11.26)
m2=1/2±J-X+l/4. (11.29)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and
.>.=j(j+l) (11.36)
where
j = o, 1/2, 1, 3/2, · · · .
Hence the eigenvalues of J2 and Jz respectively are
j(j + l)ñ2jjm)l
J2jjm) = (11.39a)
)
Jzjjm) = mñjjm). (11.39b)
Thus by taking the commutation relations J xJ = iñJ as the defining prop-
erty of angular momentum, we see that both half integral as well as integral
eigenvalues of angular momentum are allowed.
therefore
O= (jmjJ2hl/m') - (jmlhJ2jj'm').
Using Eq. (11.39a) we have
i.e.
(jmlhlj'm') = 811,(jmlhl/m'). (11.41)
Now
therefore
±ñ(jmlJ±ljm') = (jmlJj±ljm') - (jmlJ±Jzlim')
= ñ(m - m')(jmlhljm').
Hence
(jmlJ±ljm') =/= O, only if
(m-m') = ±1
or
m=m'±l (11.42)
i.e. we can write
(jmlJ±ljm') = cé±')jóm,m'±l· (11.43)
But since (h)t = J-:r-,
(jmlJ+lim') = (jm'IJ_ljm)*
(11.44)
Hence
or
m-1
C (+)j =
cr:
(-)jº
Now we note from Eq. (11.21a) that
(jmlJ+J_ljm') = (jml(J2 - f; + ñJz)ljm')
= n2j(j + l)ómm' - li2m2ómm' + n2mómm'· (11.45)
Now putting a complete set of states:
¿(jmlJ+lin)(jnlJ_ljm') = li2 (j(j + 1) -m(m -1)) Ómm'
n
i.e.
n
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or
IC[~)jl2 = ñ?(j + m)(j - m + 1). (11.46)
This leaves the phase of the matrix elements of J_ arbitrary. This is not
of any physical significance. We therefore choose all phases to be zero and
obtain
C~)i = ñ/(j + m)(j -m + 1)
-
cm-1
(+)j
L (11.47) )
e~ )j = !i-,/~(J-. )-(J-.
+_m_+_l --m-). (11.48)
From Eqs. (11.43) and (11.47)
(jmlJ-ljm')
(jmlJ+ljm')
=
=
fi/(j + m')(j - m' + l)óm,m'-1
nJ(j + m' + l)(j - m')óm,m'+l
L (11.49) )
11.4 Spin
}
(11.54a)
J
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
n (~ ~),
where the rows and columns correspond to (mi and lm') respectively. Sim-
ilarly J+ is represented by
n (~ ~).
Now from J± = Jx ± iJy, we have
1 i
Jx = 2(J+ + J_), Jy = -2(J+ - J_).
Therefore, Jx and Jy are represented respectively by
2n (º o1) , ~ (º -i)
1 2 i O .
~ G ~1),
while from Eq. (11.53), J2 is represented by the matrix
n2(!(!+1) o )=~n2(1º)·
O !(!+1) 4 01
For j = 1/2, it is convenient to write
J = (1/2)nn (11.54b)
so that u satisfies the commutation relations
~w ~~,"'
UxUy - <ly<J{~crz, etc. (11.55)
It follows from above that ax, cry, a¿ are represented by the matrices
d
= Ü ~
Writing
0
(11.54) that the eigenvalues of O"z are given by
(T: = ±1 L (11.58)
Thus there are two eigenvectors of Jz = (1/2)ooz with eigenvalue equations
where
a= (+1/2Jx), b = (-1/2Jx).
Thus the probability of its having m = ±1/2 is given by
Px(+l/2) = 1(+1/2Jx)l2 = Jal2 L )
Px(-1/2) = l(-1/2Jx)l2 = JbJ2. (11.62a)
The normalizing condition is
1 = (xlx) = JaJ2 + lbJ2. (11.62b)
This ensures that the total probability of one or other of the two possible
spin orientations is unity as it should be.
For a particle with spin 1/2, we have to take into account two additional
degrees of freedom due to spin. Thus we may represent the state function
in this case by
'l/J(r, (1/2)a~) or ,¡/J(r)Jx) , where
,¡/J(r)Jx) = ,¡/J(r, +1/2)1 + 1/2) + ,¡/J(r, -1/2)1-1/2) (11.63)
Let us write
,¡/J(r, +1/2) = ,¡/J1 (r)L
,¡/J(r, -1/2) = 'l/Jz(r), (11.64a)
)
where
l'l/J12 (r)l2 = 11/J(r, ±1/2)12 (11.64b)
is the probability of finding the particle at the point r with a~ = ±l.
Since J ± 1/2) are represented by
J2 = n2
(
1(1 + 1)
o
O
1(1 + 1)
O
o
) (1 O
= 2n2 o 1 o ~
º) )
O O 1(1+1) 001
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
i: = 1rnlí
v2
(º º)
1 O1 1
O1O L )
)
1 O O)
Jz = /i O O O .
( O O -1
where Lis the orbital angular momentum and S = !ñn is the spin. Then
i, = e, + 1Mz l (11.67) O
We now show that spin provides a basis for the Zeeman splitting of
the ground state of the hydrogen atom. It is natural that the Hamiltonian
( 11.10) is now replaced by
H = H - --L
eB
- b=o
n (11.68)
a 2mec z 2 z
where we determine the constant b from the consideration that the splitting
of the ground state agrees with the experiment. Since uz commutes with
H (as H does not involve any spin variable) and also with Lz, it follows
that the eigenstates of Ha will be those which are simultaneous eigenstates
of H, L¿ and Uz.
For the ground state E1 (l = O, m¡ = O) the wave functions are
u10ol + 1/2), U10ol - 1/2),
where
Hu10o(r) = E1Ui10o(r) ~ )
LzU10o(r) = O L )
iuzl ± 1/2) = ±il ± 1/2).
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Thus we see that the ground state splits into two levels, with a separation
6.E = bñ.
Experimentally
eñB
( 6.E)exp = -- ,
mee
therefore
(T eñ
an d E1 =E1::¡::--.
2mec
Thus
(11.71a)
Note the curious fact that in arder to get agreement with experiment, H(T is
not obtained from HB by replacing L¿ by J¿ = Lz+!ncrz = z-component of
the total angular momentum. The factor of 1/2 in the spin term is missing
in Eq. (11.71a). This is known as the "magnetic anomaly of the spin". We
may also say that the magnetic moment of the electron is
eñ
µe=--,
2mec
the Bohr magneton, since this is the factor by which a magnetic field B has
to be multiplied to obtain the observed change in the energy. The so called
(( 'Mnagnetic anomaly of the spin" naturat'ises in Dirac theory (see chapter r\/\
20). J
We consider the addition of two spin 1/2 operators S1 and S2. Let S1
and S2 be spin operators corresponding to two spin 1/2 particles. We now
considera system of two spin 1/2 particles and define the total spin operator
(11.71b)
July 9, 2012 14:29 World Scientific Book - 9ín x 6ín RF-QuantumMechanícs
Now
(11.71c)
since S1 and S2 commute with each other, for they refer to two independent
particles. This implies that S obeys the usual commutation relations of
angular momentum.
Now we can express S2 as
S2 = st + s? + S1+S2- + S1-S2+ + 2S1zS2z
Further we note that the eigenstates of Sf, S1z and Sl, S2z
L (11.71d)
are given by
o
sr\x±(l)) = ~li-2\x:Í:(1)) (11.72a)
1
S1z\X±(l)) = ±2ñ\x±(l)) (11.72b)
and similar equations for Sl and S2z· Further we note (see problem (11.4))
S1+\X+(l)) =O (11.73a)
S1-\x-(l)) = O (11.73b)
S1-lx+(l)) = ñ\x-(1)) (11. 73c)
s1+\x-(1)) = ñ\x+(l)) (11.73d)
and similar equations for S2±.
For the system of two spin 1/2 particles, we can form the following
states out of lx±(l)) and lx±(2))
lx++(l, 2)) = lx+(l))\x+(2))l, (11.74a) 1
)
lx+-(1,2)) = lx+(l))lx-(2))lt, (11.74b)
lx-+(1, 2)) = \x-(l))lx+(2)) L. (11.74c) )
\x--(1, 2)) = lx-(l))lx-(2)). (11.74d)
From Eqs. (11.75), we see that the eigenvalues of S; are -n, o,lvhere
the eigenvalue O has multiplicity 2.
To find the eigenstates of 82, we note using Eqs. (11. 74), (11.72) and
(11.73) that
S2Jx++(l, 2)) = (Sf +Si+ 81+82- + 81-82+ + 2B1zB2z)lx++(l, 2))
3 3 1 2
= (4 + +
4 2)n lx++(l, 2))
= 2n2lx++(l, 2)) ¡,_ (11.76a) )
2¡ 3 3 1 2 2 )
s
x+-(1,2))=(4+4-2)n lx+-(1,2))+n lx-+(1,2)
=n2(lx+-(1,2))+1x-+(1,2))) L (11.76b) )
B2lx-+(1,2)) = n2 (lx-+(1,2)) + lx+-(1,2))) 1.,- )
s21x--(1, 2)) = 2n2lx--(1, 2)) L (11.76c)
8
We see that above are not the eigenstates of 82•
From Eqs, (11.75) and (11.76), we see that normalized simultaneous eigen-
states of 82 and S¿ are
s21xt1(1, 2)) = 2n2Jxt1(1, 2)) (11.77a) L )
s21xi(l, 2)) = 2n21xic1, 2)) l- (11.77b) )
s21x¡1(1, 2)) = 2n2lx¡1(1, 2)) L (11.77c) )
s21x~(1, 2)) = o, (11.77d)
where
lxt1(1, 2)) = lx++(l, 2)) = 1 ii) L (11.78a) )
º
lx+Cl,2)) =
1
v'2 (lx+-(1,2)) + lx-+(1,2))) =
r v'2 r)
1 1) + 11 L (11.78b) )
We conclude that the addition of two spin 1/2 angular momenta give the
angular momenta with j = O and l.
j(j + 1)
= j(j + l)ti2 I (11.80a)
j
j-1
(11.80b)
-j
where I is an unit matrix of (2j + 1) x (2j + 1) dimensions. j is called the
highest weight of the representation. The particles with spin angular mo-
mentum j = O, 1, 2 ... are called bosons and those with j = are called !, !, ...
fermions. We saw in chapter 10 that components of angular momentum J
are generators of the rotation group
UR(()) = e-f.O n·J L
Consider rotating a spin j particle with by an angle () around an axis, say
z-axis: The rotation that does this is
Thus
9;(2~) : ~/%i:r ~f
Thu bosons comdba.ck to themselves alter a rotation by 21r but fermions
do not. Pauli's -Fc?u: ion principle holds for fermíon but not for b ons
(s e &hapter 141 ~~cause tbey behav so differently it eems unlikely that a
syrnñietry can fxtst which converts bosons into fermions and fermions into
bosons. This is so if one only talks about conserved charges coming from
symmetries whose generators satisfy commutation relations. The idea of
supersymmetry (which unify fermions and bosons) involve a symmetry in
which sorne generators satisfy commutation relations and sorne anticom-
mutation relations. We alr ady discussed in .;.hapter 9 a simple example o(. ~ ( O'\
supersymmetric system in he form of/supersymmetric oscillator. 1'k.\_
Consider two angular momentum o/5mators J1 and J2, The basis vectors
for J 1 and J 2 are given by
li1m1), m1 = i1, · · · , =i:
lhm2), m2 = h, · · ·, -j2
in a space R1 of (2j1 + 1) dimensions and in a space R2 of (2j2 + 1) di-
mensions respectively. ff, J1z are represented by matrices of (2j1 + 1)
x (2j1 + 1) dimensions, whereas J? and hz are represented by matrices of
(2j2 + 1) x (2j2 + 1) dimensions. The vectors
liii2m1m2) = li1m1)lj2m2) (11.81)
forma basis for the new space of dimension (2j1 + 1)(2j2 + 1).
Consider the operator
J=J1+J2.
This is not a matrix addition, because J1 and J2 are represented by ma-
trices of different dimensions. The addition is symbolic. The operators J 1
and J 2 commute, for they refer to two independent systems; this implies
that the components of J obey the usual commutation relations of angular
momentum.
Now
(ll.82a) )
Jzliii2m1m2) = (J1z + l2z)li1hm1m2)
= (J1z + J2z)lj1m1)1i2m2)
= (m1 +m2)nlj1m1)li2m2)
= (m1 + m2)nlid2m1m2)- (11.82b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
J =L+S,
o 1
2 2
3
1 2
1
2 6 4 EB 2
5
2 2
3 3 1 1 3
2 -2, -2, 2' 2 10 6 EB4
From Eq. (11.86) we have the result that the angular momentum op--
erator J2 has the eigenvalues j(j + l)ñ-2, where j can take the following
values
i, + Í2, . . . , li1 - h l ·
It may be noted that the product space of dimension ( (2j1 + 1) x (2j2 +
1)) is reducible, that is the space R1 x R2 splits into a number of invariant
irreducible subspaces, each corresponding to one of the following allowed
values of j:
i, + n.i, + h - 1, ... 'li1 - hl·
The total number of independent vectors for the irreducible subspaces,
which correspond to pos ible values of .i are given in the following table:
Highest weight of Number of independent
irreducible representation vectors
i, + Í2 2j1 + 2h + 1
i, + Í2 -1 2j1 + 2h -1
i, + i2 - 2 2j1 + 2j2 - 3
Clebscl~Gordan Coefficients
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
l}d2m1m2) =
m1m2
~
m.=Tn¡ +TTI,2
ljmjd2)(jmjd2l}d2m1m2)LJ11.87b) o
We shall choose the phase factors in such a way that the ClebscbJ1}ordan~
coefficients (C.G.), namely, the coefficients in expansion (ll.87a) are real
(ji}2m1m2l}mj1}2) = (jmji}2IJ1J2m1m2)*
= (jmji}2l}d2m1m2) (reality). (11.88)
Orthogonality relations:
(j1}2m~m~l}d2m1m2) = ¿(ji}2m~m~ljmj1}2)(jmj1hl}d2m1m2)
jm
(1 l.89a)
since the left-hand side vanishes unless m~ = m2, m~ = m2. Similarly we
have
(ll.89b)
since the left-hand side vanishes unless j = i', m = m', Now 8(j,j1,}2) = 1
if j has one of the values of the set
}1 + }2,}1 + J2 - 1, ... 'l}1 - }21
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
i.e.
Table I
J m2 = 1/2 m2 = -1/2
ji-m+l/2
J + 1/2 2ji+I
ji +m+I/2
J-1/2 2j1 +1
Table II
j m2 = 1 m2 =0 m2 = -1
.!. l
(11.92)
Denoting the spin wave functions
11 1 1
1--)=x+
22
l---)=x-
' 2 2
(11.93a)
and noting that
({hj>jlm¡) = Yím1 (O,</>), (11.93b)
1 11 1 1 1
ll-m - --) = llm - -)1- + -)
2 22 2 2 2
1
= llm - 2)x+, (11.94)
etc. we have
l-m+.!.
2l + 1 2 Yim+! lx-)
L )
(11.95)
l+m+ !
2l + l Yim+! lx-).
(11.96)
Now
J2 = L2 + 82 + 2L · S, (11.97)
since L and S commute with each other. Thus we see that L · S has the
following eigenvalues
~ (j(j + l)n2 - l(l + l)n2 - s(s + l)n2), (11.98)
where j = l + !, l - !, s = !· Hence the eigenvalues of L · S are (1/2) ln2
and (-1/2) (l+l)n2 respectively. It is then clear from Eqs. (11.95)-(11.97)
that the corresponding eigenstates r\ L
L · S-'l,l,1+1/2,m =
2lh 'I/Jt+l/2,m (11.99) J
L · S'l/;1-1¡2,m = -~(t + l)ñ-2,/;1-1¡2,m- (11.100)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o
Thus
J2Rwlim) = RwJ2ljm) = j(j + l)fi2Rwljm) L (11.103)
Consider in particular the following rotation. Let p0 specify the initial
direction and p the desired final direction, p being obtained from Po by a
rotation in the positive sense about the direction
n = Po x p/lPo x PI
The rotation angle is determined from the relation
L (11.104) o
Po · p = cos (),
o
O ~ () ~ it,
)
L
o
n .
= (-smcp,coscp,0) PoXP
= -. -()-
sm
Now
(Po X p) X Po
n x Po= . ()
sin
p - Pocos()
= sín é
(11.106b)
giving for the infinitesimal rotation () about direction n
P = Po + ()n x Po.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics [
Compare it with
p~ = Po + Po x Bn.
In this case we look at Po in the rotated frame of reference whereas in
the former case we looked at the rotated vector itself. The former can be
obtained from the latter by substituting B --+ -B. We now show that Re
can be written in a rather simple form. We have
Let us look at
(11.108)
and
(11.109)
It follows from Eq. (11.106a) that Re= R(<fa, B, -<fa) acting on a state with
IP, B = O, <fa = O) gives the state
IP, e,</>)= R(</>, e, -</>)IP, e= o, <fa= o) (11.111)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
m'
j
= L ljm')d~)m(w) (11.112)
m'=-j
where
(11.113)
Rw=O = l.
ii) Inverse of Rw exists
)
u; = e-i¡¡!,·J L
ti.; = ei¡¡!,·J L J
RwR-w = l.
(11.114)
m"
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or
(11.115)
rri"
or equivalently
(11.116)
Because of this property the rotation matrix d(íl(w) gives a representation
of the rotation group. For the identity transformation (no rotation), w = O
and
d~?m(O) = (jm'jjm) = Óm'm· (11.117)
The matrices d(j) (w) are unitary, since
= Ómm' -
2\/+Um'jLljm)
i
- 2ñc(jm,IJ+lim)
-*Ez(jm'IJzljm). (11.120)
The use of Eqs. (11.49) and (11.50) then give
where
j = o, 1/2, 1, · · · .
For the finite rotation w = 8n, R(cp, 8, -</>) = ci/:n·J and we have
d~!m(</>, 8, -</>) = (jm'IR(cp, 8, -</J)ljm)
= ( jm 'I e-" J ·e-"JYe"iq, J • I jm 1
ief, iO
= e-im'<t>eim<f,(jm'le-1/:JYljm)
= ei(m-m')q,d~!m(8). (11.122)
Now
R(cp, 8, -</>)18 =o,</>= O)= 18</J),
so that
(jmlR(cp, 8, -</J)IOO) = (jml8</J) (11.125)
or
¿ (jmlR(cp, 8, -</J)ljm')(jm'IOO) = (jml8</J),
m'
i.e.
¿ d~;,,, (</>, 8)(jm'IOO) = (jml8</J). (11.126)
m'
For the particular case j being an integer, we have
¿d~;,,,(cp,8)~~1(0,0) = ~~(8,cp), (11.127)
m'
where
"Yjm(8,cp) = (8,</Jljm).
Now
(11.128)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and we have
v=c:
{2[+T """"" (j)
L., dmm'
m'
*
( (/), () ) Óm',O -- ~m((), (/)) (11.129)
or
(11.130)
or
e-im,t,d(j) (()) =
mO
J 41r Y*
2j + 1 Jm
(() "').
,'!'
(11.131)
(11.132)
since J 1 and J 2 commute with each other. Therefore, we have
e-(i/ñ)w·J lid2m1 m2)
= e-(i/ñ)w·J1 li1m1)e-(i/ñ)w·J2 li2m2)
L., 11·1 m'1 )iíil
= """"" mím1 (w)"""""
L., 11· 2m'2)i12l
m2m2 (w)
Tnt m;
= L., 11· 1 1· 2m'1 m'2)ihl
""""" mím1 (w)im12l2m2 (w) (11.133)
m~mfu
or
(jij2m~ m; 1 e-(i/ñ)w·J lid2m1 m2)
= L., (1· 1 1· 2 m"m"I
""""" 1 2 J. 1 J. 2 m'1 m'2 )imí m2m2 (w)
11l m1 (w)d(h)
m~m;
== L
m~m~
J:
Um"m
1
J:
'1 Um"m
2
'2 d(j 1)
,
m1rn1
(
W )d(j,)
,
m2m2
(W )
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Hence
But
(jij2m~ m;le-(i/ñ)J ·w ljd2m1 m2)
= L Ud2m~m;l/m'j1i2)(j'm'jd2le-(i/ñ)J·wljmjij2)
jj'm'm
x (jmjd2U1i2m1m2). (11.135)
Now since e-(i/1i)J·wljmjd2) is an eigenstate of J2, we can write
[c.f. Eq. (11.113)]
(j'm'jd2le-(i/ñ)J·wljmjij2) = Ójj'd~,!m(w). (11.136)
Hence
Ud2m~ m; le-(i/ñ)J·w lj1hm1 m2)
= L (jij2m~m;ljm'j1i2)(jmjd21id2m1m2)d~;m(w)
jmm'
(11.137)
and on using Eq. (11.134) and the reality of C.G. coefficient, we have
ij!)
m1m1
(w)d(j~) (w)
m2m2
= L Ud2m~m;ljm'jii2)(j¡j2m1m2ljmjd2)d~;m(w).
jmm'
(11.138)
Note that the factors multiplying d~!m(w) are C.G. coefficients. On the
other hand, the inverse relation is given by
"""' """'ij¡)
L.., L.., m~ m1 (W )im;m2
12l ( W )('J1J2m1
· m2 1·JffiJ1)2
· · )('J' m /·J1J2· 1·.J1J2m1 m2 )
(11.139)
Now we use Eq. (11.138) to prove sorne of the important results.
For j1 and h both integers and m1 and m2 both zero, m must also be
zero. In this case, from Eq. (11.138), we have
iiil (w)ij2) (w)
m~O m;o
= ¿U1hm~m;ljm'jd2)(jüjd2U1hOO)d~!0(w). (11.140)
jm'
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
(j) ( )
dmo W = y{4ir
2J+l~m* ( 8,<p,)
we have from Eq. (11.140) removing the prime and taking the complex
conjugate
YJ1 m1 ( (}, <p )YJ2m2 ( (}, <p)
._, (2i1+1)(2i2+1)(.. ... )(·o· .1. ·oo)Y (8,1.,)
= újm
47r(2j + l) J1J2m1m2 1 JffiJ1J2 J J1J2 J1J2 jm , ..,, .
(11.141)
Multiplying both sides by ~*,m, ( (}, cp) and integrating over the angles and
using the orthogonality of spherical harmonics, and replacing j'm' by jm
at the end, we have often used the integral
The set of (2j + 1) states jjm), for fixed j, transform into themselves under
rotation according to the transformation law
j
Rwljm) = L jjm')d~'m(w).
m'=-j
= (1/Jw!RwAR,::;-111/Jw)
therefore (11.143)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
RwSH:;;1 = S.
For an infinitesimal rotation w
i i
(1 - ,i,w · J)S(l + ,i,w · J) = S
or
[hB]=O
i.e. a scalar operator commutes with the angular momenturr[_.
2. Vector operator: A vector operator V should transform in the same
way as J, i.e. a vector operator obeys the followingcommutation relations.
[h V:il = iñEijkvk. (11.144)
For infinitesimal rotation, using Eq. (10.103)
i i
- ,i,w · J)1!¡(1 + ,i,w · J)
1/¡ + ékilWk Vz
= -v¡ + (V X w )i. (11.145)
On the other hand under infinitesimal rotation, we know that a vector
operator transforms as
(11.146)
Noting that
Vi=
Vx+iV11
v'2
L )
Vo=Vz L (11.147) )
_ Vx -iV11
V-1 - ../2
Then we can write the commutation relations of Vq as
(kq'[h[kq) =
(kq'[Jz[kq) = ñq(kq'[kq)
!iy'k(k + 1) -
=
q(q ± l)óq ,q±l
ñq8qq'·
1 L )
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Hence
q'
(11.150)
so that
[h, rJklJ = n../k(k + 1) - q(q ± 1)1~~1
= Eq'Td,k) (kq'[h[kq) (11.151a)
[Jz,TtlJ = nqTtl = Eq,Td~)(kq'[Jz[kq) (11.151b)
or
(11.151c)
Under infinitesimal rotation
rCkl R-W 1
R=:« = rCkl
q
- iw
n, . [J lqrCkl]
= rtl - ~Eq,Td,kl(kq'[w. J[kq)
= Eq,Td,k)(kq'[Rw[kq). (11.152)
The above result also holds for finite rotation since a finite rotation can
be built from infinitesimal rotations. Hence, we have the result that under
rotation an irreducible tensor operator rtl transforms as
RwrtlR;:,1 = ¿rtl(kq'[Rw[kq)
q'
k
= L Td,k)d~~i(w). (11.153)
q'=-k
Conversely we can start from this definition of the tensor operator and can
obtain results about the commutation relation of rJkl with J.
Taking the Hermitian conjugate of Eq. (11.153), we have
t k t /
(RwrtlR:;1) = q'~k (r?l) d~~t(w) "- )
R w r(k)t
q
R-1 =
w
'°'
~
k
T(k)t
q'
ik)*
q' q
(w) .
q'=-k
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
But
ik)* (w) = (-1)2k-q-q' ik) (w)
q'q -q'-q
therefore
R r(k)t R-1 rCkJt
= "\;""' (-1)2k-q-q' ikl (w)
w q w L q -q'-q ·
q' 1
Now q' being a dummy index, we can write q' --+ -q'. Changing q --+ -</-._ )
¡- )attdfmultiplyingby ( -1 )q on both sides of the above equation and noting
that (k + q) is always an integer so that (-1)2k+2q = 1, we have
k
u; ((-l)qT~~t) R;:;1 = ¿ r!_kj,\-1ld~~~(w).
q'=-k
Hence (-l)qT~~t transforms under rotation in the same way as rJk). If
rJk) is to represent a physical quantity, it should return to itself after a
rotation through 21r so that k is an integer.
rJkl lo:j1 m1) transforms under rotation in the same way as the state
lkj1qm1) [cf. Eq.(11.133)]. Hence we can write
rJkllo:j1m1)
= lo:kj1qm1) = ¿ lo:jm)(jmkj1lkj1qm1), (11.156)
jm
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
J2lajm) = j(j + l)ñ2lajm)
Jzlajm) = mñlajm).
Conversely we can write
lajm) = ¿ TJk)laiim1)(kj1qm1ljmkj1). (11.157)
From Eq. (11.156), changing j1m1 --+ jm and replacing Ljm by Lj"m"'
we have
rt)lajm) = j~' laj"m")(j"m"kjlkjqm)L (11.158) '
(a'j'm'ITJk)lajm) = ¿ (a'j'm'laj"m")(j"m"kjlkjqm).
j"m"
(11.159)
Now from the orthogonality relation
( a I J•/ m 'I O:J·// m ") = s s
Uj'j"Um'm" (a I J•/ m 'I O:J•/ m') . (11.160)
Hence
(a'j'm'ITt)lajm) = ((a'j'm'laj'm')) (j'm'kjlkjqm). (11.161)
Now we prove that (a'j'm'laj'm') is independent of m'. To see this,
consider the transformation
lajm) = ¿ lf3jm)(f3jmlajm),
{3
so that
lajm + 1) = ¿ lf3jm + 1)((3jm + llajm + 1)
{3
and
J+lajm) = nJ(.j + m + l)(j - m)lajm + 1)
= liJ(j + m + l)(j - m) ¿
lf3jm + 1)((3jm + llajm + 1).
{3
(11.162a)
But
(11.162b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(j'm'IYkq(B, c/>)jjm)
= j (j'm'IBc/>)Ykq(B, c/>)(Bc/>ljm)dD
= J Y/,m,(Bc/>)Ykq(Bc/>)°Yjm(Bc/>)dD
Far the El transition one comes across (see Sec. 17.6) the matrix ele-
ments of the form (l'm'lrJlm). Now r in the canonical base can be written
as
r = ( r1, ro, r - l) ,
where
X+ iy T • () i"'
r1 = - -- = - - sin e .,,
v'2 J2
r0 = z = r cos () (11.168)
_ X - iy _ T . () -i<f,
r_ 1 - J2 - J2 sm e
ar
"« = y¡¡;
3rY q(O,rp),q1
= 1,0,-1.
= lf ~:-:+-+-~
, (llqm¡z'm'll)(l'OlllllOO). (11.169)
(llqm¡z'm'll) =1- O,
far l' = l + 1, l, l -1 and far m' = m + q. Hence we have the selection rules
= z' - l = ±1,0
!:::,.l L (11.170) ,
!:::.m = m' - m = q = ±1, O.
Furthermore, using the C.G. Table II, and Eq. (11.169) we have
(l + lm/Y10/lm) = a (l + 1)2 - m2
(2l + 3)(2l + 1) l )
(l - lm/Yío/lm) = a z2 -m2
(2l + 1)(2l - 1)
t: )
(l + lm + 1/Y11/lm) = a (! + m+ 1)(! + m + 2)
2(2l + 1)(2l + 3)
L (11.171)
)
(l - lm + 1/Y11/lm) = a 41r
(l -m - l)(l - m)
2(2l + 1)(2l + 3) L )
(l + lm-1/Y1_1/lm) = a 41r
(l-m+l)(l-m+2)
2(2l + 1)(2l + 3) i )
11.10 Problems
in!lx(t)) = Hlx(t)).
Find the probability at t > O, for finding the eJectron in the spin
state (a) Sx = (1/2)n, (b)t = -(1/2)n, and (c)~= (1/2)n. Hint:
As far as spin is concerned, the Hamiltonian H = µoBaz, where µo
is the magnetic moment of the electron. Use matrix representation
for Sz etc.
11.7 Consideran operator ú which obeys the commutation relation
[ú, Jz] = (1/2)M L j
[[ú, J2], J2] = ~(úJ2 + J2ú) + 31:4 ú.
Consider the representation in which the basis vectors are lim), the
simultaneous eigenvectors of J2 and Jz belonging to the eigenvalues
j (j + 1) n2 and mñ respectively. Show that if
(jmlúlj'm') =I- O
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Show that
1 2 2 B
6.E = (2a mee )a e/a2
B
---~9--- x 13.6eV,
(2.4 x 10 Gauss)
2 1 .
where a = ~e 137 is the fine structure constant and E1 =
!mec2a2 = 13.6 eV is the binding energy of electron in the first
Bohr orbit of hydrogen atom. a is the radius of first Bohr orbit.
Draw the energy level diagram for the Zeeman effect for p and d
level. In atomic transitions D..m = m¡ - mi = -1, O, l. Draw 9
possible transitions on the energy level diagram.
11.10 An electron can be regarded as a magnetic dipole of magnetic mo-
ment µ = µ0u, where µ0 = 2 meeefi. is the electron magnetic moment.
The interaction energy between magnetic dipoles is given by
V = ~ { (µ . µ ) - 3 (µ1 . r) (µ2 . r) } .
\ r3 1 2 r2
=
() ..Bl.Il -n
COS - - i
e · U
2 2 '
where () is the angle of rotation about n(n2 = 1). In general a spin
state can be written as
1 1
lx) =Gil+ + C2I -
2) 2)
= (~:)
called a two component spinor. Its transformation law under the
rotation is
lx') = e-(i/2)0n·u1x).
Show that
(i) for O= 21r, lx) = -lx)
(ii) for rotation about X-axis, for which ñ = ex, U· ñ = O'x,
(C~)
C2
= (
i
!
_c~s 6 -i si~!)
sm 2 cos 2
(º1) .
C2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
gives
<T~ = <Txl )
o-~ = + sin ()o-z
cos Buy L- )
o-: = - sin ()o-y + cos ()o-z,
i.e. u transforms like a vector under the rotation.
11.15 Show that
O' . pe-(i/2)9n· u = e-(i/2)9n· u <T . ez = e(if2)(}n,u <T z >
L Ud2m~m;ljm'jd2)d~)m(w).
m'=m~+m;
[Hint: operate Eq. (11.138) on IJ"m"jd2) and then take matrix
elements with (jd2m1m21, use the orthogonality property of C.G.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= .~d;,.,m(w).
where j is an integer; (), </> and ()', </>' are the spherical polar coor-
dina tes of the same physical point in the old and new coordinate
systems while ,6, a are the spherical polar coordinates of z' axis in
the old coordinate system.
[Hint: consider the eigenstates IJm) and IJm)' respectively of Jz
and J; where
IJm)' = RwlJm).
I ii 1) 1 ii O) 1 rr -1)
I i11) 1 i1 O) 1 u -1)
! li 1) ! li O) ! li -1)
I U 1) 1 U O) 1 !! -1)
a. What are the possible eigenvalues of ffatal?
b. For each of the eigenvalues found in part (a) what are possible
eigenvalues of Jtotal,zll
c. Determine (J1,x + h,x)2! rr 1) and J1,xh,yl n 1).
d. Write down the normalized state with total angular momentum
eigenvalue O in terms of the individual spin states given above.
!
11.21 A particle of spin is in a D-state of the orbital angular momen-
tum:
a. What are its possible states of total angular momentum?
b. For the Hamiltonian
H = a+ bL · S + cL2
~ where alb and e are numbers, fínd the values of the energy for eac.h
of the dilÍerent states of total angular momentum (expre s your
'tJ..1/ an wer in terms of at~·
11.22 Show that:
a)
)
b)
)
e) For any vector operators A:
[Jx, (J,, TJ'll] + [.J,, [J,, TJ'l]] + [J,. [J,. TJ'll] - k( k + 1 )ñ'TJ'l Le)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 12
TIME INDEPENDENT
PERTURBATION THEORY
12.1 Introduction
There are very few problems which can be solved exactly. One has to resort
S
to approximate methods, which are of two types:
(i) Perturbative
(ii) Non-perturbative
We first discuss perturbative methoJ. suppose we have som problem
which cannot be solved exactly. We loolt1I!"a similar problem that can be
B Cíff-5
solved exactly. Then see what is the deviation between the problem we
want to solve and the problem we have solved. In perturbation theory this
deviation is small. So we have a main effect which we know and the small
effect which we get as an approximation.
283
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
write
(12.4)
n
We insert Eq. (12.4) in Eq. (12.1) with Has given in Eq. (12.2). Thus
(Ho +AV)¿ CnsUn(x) = E8 ¿ CnsUn(x).
n n
¿
n
Cns(Es - en) J u;,.(x)un(x)d3x
= ¿
n
so; J u~(x)Vun(x)d3x.
This gives
n n
or
Cms(Es - cm) =A¿ e; Vmn, (12.5a)
n
where
Vmn = J u;,.(x)Vun(x)d3x
= (m[Vln). (12.5b)
Note that Vmn are matrix elements of the perturbation with respect to the
eigenfunctions of the unperturbed Hamiltonian. Equations (12.5a) are an
infinite set of linear equations for Cms· When A = O, perturbation is zero,
then
(12.6)
For A=/=- O, we make an expansion of C's
ems -_ Ums
s:
+ /\,ac1)ms + /\, 20c2)
ms + ... (12.7a)
E8 = e + AE(l) + A2 EC2) + ...
S S 8 (12.7b)
'l/Js = Us + A'l/JF) + A2'l/JF) · · · . (12.7c)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
We substitute the above equations back in Eq. (12.5a) and solve these
equations for a certain power of >.. In the first order (keeping terms up to
>. only)
(12.11)
m
Since
it follows
cc1J· + cc1J
SS SS
= o
that is the real part of c!!l = O. There is no loss of generality (because
we can select a phase such that C~!l is real) in making the simple choice
(1)
lm 088 = O, so that
C SS(l) = o•
Thus using Eq, (12.9), we have from Eq. (12.11) up to first order in>.
(12.12)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
(12.13b)
We wríte
Also note
We also note from Eq. (12.12) that the condítíon for the applícabílity of
the perturbation theory is that
(12.14)
Let us calculate Es to second order in >.. Substituting Eq. (12.7) in
Eq. (12.5a) and retaining terms up to order >.2, we have
[8ms + >.cgi + >.2C~l] [es+ >.E?l + >.2 Ei2l - cm]
= A ¿(8ns + >.Ci~l)Vmn·
n
(12.16)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Now
)
s-=f=.m.
e
Th ereiore c(2)
88
+C SS
(2)* _
- -
n
'L..Jncf-s
°""" v,;. Vn.
(e.-enP
_
-
2Re c(2)
SS •
o
We take Im d;l = O, hence
_ ~2 """"' v;:s V.is
L {e - e, )2
n;és 8
~
L 7
w»
«t, ~:.X """8 L._.,¡
mcf-s
C(llu
ms m + .X '°' cC
2 L._,,¡
m#-s
2>u m
ms + .x2cCss2>u s
Us +AL
_,_ Es-Em
VmsUm + _x2 ¿ _,_
({¿
_,_ (Es-En)(Es-Em)
VnsVmn
m-,-s m.,...s n-,-s
_ VmsVss })
Um+"
,2(-~"""
L( E
IVnsl2 )2 ) Us, (12.18)
( E8 - Em) 2 2 En 8 -
ncf-s
12.3 Degeneracy
We now consider the case when Ho has degenerate eigenvalues i.e corre-
sponding to eigenvalue Es, there are more than one eigenfunctions. Let us
consider the case of two fold degeneracy i.e
Hous,I = EsUs,IL '
HoUs,2 = EsUs,2·
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
then we get
, _,- i
J n
We now make the expansions
/ Cms,i Umsdi + .xcms,i
(1) + A 2 cms,
(2) i + ... (12.24a)
Now dic5ms(é8 -em) is zero because when m =/= s, Óms = O and when m = s,
c = cm· Therefore up to terms of order A, we have
8
Hence for m = s
diEi1) = ¿ d1 vij, (12.26a)
j
where
Yij = J u;ivu8jd3x. (12.26b)
Also from Eqs. (12.24a,c) and Eq. (12.19) , up to zero'th order
'l/;~ = ¿¿
ÓnsdjUnj
j n
= ¿ djUBj . (12.27)
j
Now from Eq. (12.26)
d1Ei1) = s. Vu + da V12L
d2EF) = d1 Vz1 + d2 Vz2, (12.28) )
A solution of these hom[;~e~e~¡:¡'nnui1:":~uls ::ations exists onl~::.29)
Vi1 V22 - E8
Now
V12 = V2*1 or V21 = Vi*2,
Then Eq. (12.29) gives
E;~ = ~ ( (Vu + V22) ± /(V11 - Vz2)2 + 4IVi.zl2) . (12.30)
Thus from Eq. (12.24b) to order >.
\ (1)
Es± = t:8 + /\E8±
or
x ).
Es±= e, +
2(Vu + Vz2) ± 2J(Vi.1 - Vz2)2 + 4IVi.212. (12.31)
Thus due to perturbation the level t:s (which is degenerate with respect
to the unperturbed Hamiltonian H0) is split up into two levels given in
Eq. (12.31) when we turn on the perturbation.
Thus we can remove the degeneracy by turning on a small perturbation.
Also we note from Eq. (12.28) that t
has two values corresponding to two
values E~~ given in Eq. (12.30). Thus it follows that the zero'th order
wave functions are 'l/;~~ which are linear combinations of the unperturbed
degenerate wave functions Us1 and U82
(12.32)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
E=-V<p=-~!~r (12.33)
where es denotes the energy in the absence of the electric field. Now
Vsn = (slVln)
=E· (slrln) (12.34)
=E·rsn ~ )
rsn = (slrln)
= J u;rund3r
= (nlrls)*
(12.35)
Thus to second order
(12.36)
The shift of energy levels in the electric field is known as the Stark effect.
The above expansion to order e2 holds to a good accuracy when IEI is small
compared to the internal electric field of the atom.
The energy eigenfunction of the perturbed Hamiltonian H = H0 + eV
to first order in e is given by
(12.37)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
We now calculate the static electric dipole moment of the system in one
of its stationary states. In quantum mechanics, the electric charge density
is given by
Thus to order e2
p(r) = -e'I/J;'I/Js L o
2E ~ (u;rnsUn + u~rsnUs)
p (r ) = -eus* U8 - e · L...., (12.38)
ncfs E:s - E:n
and the dipole moment is given by
D = J p(r)rd3r
= -e J u;ru d3r
8
-e2E · ¿~
n#s Es En
(rns ju;rund3 r + rsn Ju~ru d r)
5 3
_ 2 ~ rnsrsn + rsnrns . E
- -erss - e L....,
n#s Es - En
(12.39)
where
=ii·E,
where
(12.42)
('
V
This can be seen as follow~:
Fu8(r) = ±u8(r)
where
Therefore
P denotes the parity operator. But
Fu8(r) = (-r)l
U8
o
(12.43)
Now
r88 = J u;(r)ru5(r)d3r l
In the integration change
r----+ -r,
so that
rss = J u;(-r)(-r)u8(-r)d3r
= J u;(r)(-r)u8(r)d3r,
(12.44)
Hence r ss = O. For H atom,
Hounlm(r) = €nUnlm(r)
where
me4
€n = - 2ñ2n2, n = 1, 2, · · · , l + 1 ~ n, -l ~ m ~ l
(12.46a)
Therefore
[Lz, V]= O. (12.46b)
Henct = 2 is understood]
Hence
((-1)1 + (-1)1') (lmlVll'm') = O. (12.50)
Thus
(lmlVll'm') = O for l = r. (12.51)
Hence using Eqs. (12.46e) and (12.51), the 4x 4 determinant equation which
determines E~1), the first order correction, is
-E~l)
O -E~1l
o (OOIVl10)
O O
o 1
=Ü. (12.52)
(101VIOO) O -E~l) O
Now
l O O O -E~1l
elEl-1- ~ (- 72a5)
=
16a4 3
= -3elEla. (12.55)
Equations (12.52) and (12.55) give
-E~1) ( -E?) (E?l2) - 3elEla( -3elElaE~1))) = O. (12.56)
The four roots are
E~1) = O, E~1) = O, E~1) = ±3elEla. (12.57)
Thus half of the four fold degeneracy is removed in the first order. The
third and fourth values of E~1) correspond to
1 1
v'2 (u200 - U210) and v'2 (u200 + U210)
and the first two correspond to any linearly independent combination of
u2n and u21_1. Thus a hydrogen atom in its excited state (n = 2) behaves
as though it has a permanent electric dipole moment of magnitude 3ea that
can be oriented in three different ways: one state parallel to the external
field, one state anti-parallel to the field and two states with zero component
along the field.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(1/!IHl1/l) > E
(VJIVJ) - O,
where Ea is the ground state energy and 11/l) is an arbitrary state. This can
be seen as follows: In the basis IEn), En being energy eigenvalues of HM )
H = L L IErn)(ErnlHIEn)(Enl
rn n
rn n
Thus
Now 1(1/!IEn)l2 > O and En 2: E0, E0 being the lowest possible energy,
i.e ground state energy corresponding to 11/!0). Thus
o
Eg=---+---
2ao 2ao
= - z2e2 .. (12.64) •
ªº
The first order change in the energy is the expectation value of the H1 in
the ground state, on using the wave functions in Eq. (12.62) is
!:::,.E= (H1)
= (1(100)2(100)IH111(100)2(100))
- :,(Z/ao)'e' j j d r d"r,1,,
3 1 ~r,le-!:(•,+•,) t2.65) ó
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Now using
(12.66)
2z
o
d3k
J d3
re
-ik·r-
=o
ro
= [k2
= 41rk2dk, and putting a0k/2Z = x, one obtains
167r Zao
+ (2Z/ao)2]2' (12.68)
4 Ze2 I" dx
t::.E = :;;:~ Ío (x2 + 1)4
= i Ze2 (51r)
1r ªº 32
5Ze2 = L (12.69)
8ao ! ... ,
One can determine experimentally the ground tate energy for~o elec-
tron atom by measuring the ionization energy E, i.e the minimum energy
required to remove one of the electrons to oo. Then
E, = -(Ea - Eí0) (12.70)
where Ejo is the ground state energy of the remaining electron and is given
in Eq. (12.63). For He (Z = 2), E, = l.807Ry experimentally so that
EOexp = E'jO - E i
= [-4 - (1.807)]Ry
= -5.807Ry
~ -79eV (12.71)
while the perturbative result, using Eq. (12.64) and Eq. (12.68) gives
(Z = 2)
E¡;er = Eg +t::.E
5 Ze2
= -(Z-8)~
5
= -(Z -
8)Z.2Ry
= -5.5Ry
~ -74.8eV. (12. 72)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
8
= (Ho)+ (H1)
= (Ho)+-- 5Z'e2 (12.74)
ªº
The expectation value (Ho)is with respect to the ground state of a Coulom-
L o
bic potential Z' e2 / r, so that
Z'e2
(K.E of each electron) = -2- (12. 75)
and
ªº
(- Ze2)
r
~~~
~(- Z'e2)
Z' r
Z Z'2e2
z,C-~) (12.76)
o
Thus
E0(Z')
8Eo(Z') = O
Z' Z- ~
gives = (12.78)
8Z' 16
while
82 :~\f2
= 2 > O. Thus E0(Z') has a minimum at Z' = Z - i56• Thus
· expected the effective change is somewhat less than Z. The minimum
,.__,,.,.....
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Eaªr = -(Z - ~ )2 e2
16 a2
= -5.695Ry
:::::! -77.46eV (12.79)
12.6 Problems
If
ua(x) = (xlü) L D
u1(x) = (xll)
is the first excited state of the same oscillator when F = O, show
that to the first order in perturbation the probability of finding it
in the first excited state is
1(116)12 = 1 j ui(x)ua(x)dxl 2
F2 1
2mñw3 º
Also calculate the shift in the ground state energy to second order
in F.
12.2 A hydrogen atom is placed in an externa! constant field F in the
L
z-direction. For the ground state s of the atom, write
e2 F2 L IV.nl2
ncfs€8 -E:n
= -~F2a t>
U
Show that
16
4a3 <a< -a3
3
where a is the Bohr radius ;;2 •
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
eF ~ lXJ R~0(r)Rn1(r)r dr 3 L
l o
Hence show that for the ground state
E1 =E1+-3-~
e2 F2 '°'
n,fl
1 fa°° R10Rn1r3drl2
E1 - En
12.4 Consideran atom which has a nucleus of charge Z and one electron.
Using first order perturbation theory, calculate the energy shift of
o
the ls, 2s and 2p states of 1 H and 235U atoms, assuming that
the nucleus is a uniformly charged sphere of radius R = r0A 1/3,
where ro~ 1.2 x 10-13 cm and A is the atomic number. Note that
R < < a, where a is the Bohr radius.
Hint: The Coulomb potential is given by
3Ze2 2 1 2
V(r) = - 2R3 (r - 3r ) )
Ze2
r >R.
r
12.5 Zeeman Effect:
The Hamiltonian for a hydrogen like atom can be written as
Ha=---+
p2
2me
ze2
r
ze2
2m~c2r3
S·L
L o
The last term represents the interaction of the magnetic moment
of the electron µ = __ mee
e_5 with the magnetic field B = 21e Ex v =
iffr Xv = 2~ec ffr X p which arises due to the motion of an
electron in the electric field E of nucleus of charge Z e.
Consider the atom in an external weak homogeneous magnetic field
B along the z-axis so that the magnetic interaction energy is given
by
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= _.::!!_(njml(Lz
2mec
+ 2Sz)lnjm)
- eBfl m (1 ± -1-) j = l ± 1/2.
- 2mec 2l + 1 '
12.6 If H(>..) depends on a parameter >.., and 11/lk(>..)) is an eigenvector
of H(>..) which is normalized to unity, prove that
\
dE:1>..) = (1/lk(>..)I º~Y) 11/lk(>..))
1) [HelmannJ,FeynmanTheeorem]
12.7 Estimate the ground state energy of the hydrogen atom by using
the three dimensional harmonic oscillator ground state function as
a trial function. How this estimate compares with the exact value,
. h.
Wh lC 18 -~.
me47
(~r
Three dimensional harmonic oscillator ground state function is:
1\-!a2r2, a=~ L Q
Take a as variational parameter. Useful integral
{DO X2nea2x2 da: = __1_ V'ff (2n)! •
Ío 2a2n+l n! 22n
Repeat the same problem with the followingtrial function:
r
c/Ja(r)=l--, r::;a
a
= O, r >a
which one is the better trial function?
12.8 Use the variational principle to estimate the ground state energy
of the anharmonic oscillator
-2
H=L+>..x4.
2m
Compare your result with the exact result
f¡,2 ) 2/3
Ea = 1.060>..1/3 ( 2m
You may use the trial function
Ae-a2x2 /2.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
12.9 Use the variational principle to estimate the ground state energy
of a particle in the potential
V=oo for x<O L )
V=cx for x>O
L
Take Axeiax as the trial function. Why can't one select Ae-ax as
the trial function? Useful integral:
o
{'x, n -bx n!
Ío x e dx = bn+1 .
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
Chapter 13
TIME DEPENDENT
PERTURBATION THEORY
Let us discuss how the perturbation method works for the time dependent
Schródinger equation
l
given in Eq. (13.lb) and make use of
Haun(r) =
303
énUn(r)
.:
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Then we obtain
n n
or
n n
xe'í'--,,),~:~mn(t) L (13.9)
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o
to
~ '1/J(r, t) ~ Us(r)e-i(es+>.Vss) t ~ to
where
(13.17b)
i.e. energy is changed slightly and is given above in the first approximation.
We are interested in Cn(t) for ni- s, since then ICn(t)12 gives the prob-
ability of transition from state s to another state nas given in Eq. (13.8).
Hence using Eq. (13.12) to first order
P(s,n) = ICn(t)l2
(13.18)
For second order, we put Eq. (13.12), replacing m by n, back into Eq.
(13.6a); so that we get
where
Em -En
Wmn= (13.19)
n
Special cases:
(A) Sudden Approximation
t < t0 no perturbation
t 2". t0 there is perturbation
Thus
Vms(t) = O t < to
and assume that
Cm(t) = _i_>.Vms
n ita
r ei(E;m.-E,)t/lidt
ei(Em-Ea)t/li ( . t-t ) ~
= ->.Vms 1- e-i(Em.-E,)? (13.20)
E:m -E:s
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Hence
This shows that if the perturbation is turned on, the transition probability
oscillates between O and sorne number.
From Eq. (13.21)
l (13.21)
o
,2
/\
. 21 t
2s1n 2Wns
P(s,n) = t;,2 IVnsl (wns/2)2, (13.22)
where Wns = En-,;,:• and we have taken t0 = O. We now show that far large
o
t,
Now
P(n,s)-d L
·21 t ·21 t
lim sm 2Wns = lim sm 2Wns t
t-+oo (wn8/2)2t t-+oo (wn8t/2)2
o
Further
L
Thus
.
hm (
t-+oo
Hence the transition probability per unit time (in the limit of t ----> oo) is
given by
(13.27)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
This explicitly exhibits the fact that energy is conserved. If the transition
is to a continuum of states about state n, then p1(e:n)de:n is the number
of final states with energies between E:n andén+ den, where P¡(e:n) is the
density of final states, i.e. the number of states per unit energy interval. In
this case the transition probability per unit time is given by
Wns =
2;
A2 f
IVnsl2ó(cn - és)P1(cn)dcn
2rr 2 2
= --,¡.X IVnsl P¡(cn)· (13.28)
The above formula represents a constant rate of transition. The formula
(13.28) is a fundamental result and has been called the Golden Rule by
Fermi. For an application of the Golden Rule to potential scattering, see
Sec. 18.7.
We end this section by the following remarks: we note from the above
analysis if we plot 4 sin2w • 1"'""t
as a function of Wns for fixed t, we get the
curve shown in ig A13. l .
4sin2co Vco2
"' ""
'°ns
13.2 Problems
For t ;::=: O, take F(t) = F0(l - e-t/T). Show that in this case, the
probability of finding the system in 11) for large times is
F.2
P(O, 1) ~ 2m~3,
if WT » l. If T = O, show that this probability oscillates indefi-
nitely.
13.2 A hydrogen atom in its ground state is placed in an electric field
E(t), which is spatially uniform and has the time dependence
E(t) = O t< ol )
= Eoe-t/T t > O.
What is the first-order probability of finding the atom in the 2p
state after a long time?
I July 9, 2012 14:29 World Scíentific Book - 9in x 6in RF-QuantumMechanics I
Chapter 14
14.1 Introduction
L o
the particles' coordinates implying
l'l/J(l. 2)12 = l'l/J(2. 1)12 (14.3)
Hence
iµ(l, 2) = ±'l/J(2, 1) , (14.4)
311
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
P121/J(l 2) = 1/;(2,
~z
P121/J(l 2) = 1/J(l, 2).
1)L (14.5)
)
Thus Í'i2 has two eigenvalues ±l. Let the corresponding eigenstates be 1/J±
l
..
1/J+ = !(1 + A2)1/J = H1/J(1,2) + 1/J(2, 1)) (14.6) )
1/J- = !(1 -A2)1/J = !(1/J(l, 2) -1/;(2, 1)) L (14. 7)
Now
Hence
(14.9)
Pij is an operator that interchanges the variables of the ith and jth parti-
cles:
Pi2'!/J(l, 2, 3) = '!/J(2, 1, 3) L )
Pi3'!/J(l, 2, 3) = '!/J(3, 2, 1), etc. (14.10)
Then, for example
Ha(l)'l/Ja(l) = Ea'l/Ja(l)L
Ha(2)'l/Jf3(2)= Ep'l/Jf3(2). (14.20) )
We can also write the solution as
(14.21)
Here particle 1 is in state (3 and particle 2 is in state a, but energy E =
E°'+ Ep while 'l/Ja,{3 and 'l/Jp,a are two different wave functions belonging to
the same energy eigenvalue. Therefore, we have degeneracy, called exchange
degeneracy.
We now consider a system of ti particles. We write the Hamiltonian as
with eigenvalue
'fJ'
FermiJ,Oirac statistics:
1
7/JA(l, 2) = J2 ('1f'o.(l)7/J¡3(2) - 7/J¡3(1)7/Jo.(2)) (1 state) (14.27)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Bose-Einstein statistics:
Possible symmetric wave functions are
'l/Ja(l)'l/Ja(2), 1('l/Ja(l)'l/J13(2) + 'l/J13(l)'l/Ja(2)) and 'l/J13(l)'lj;13(2) W{3 states) /
In the classical situation, one can think of the different 1/Jo: as specifying
possible orbits of the particles. In this case, each of the followingrepresents
a classically distinguishable situation
'l/Ja(l)'l/Ja(2), 'l/Ja(l)'l/J13(2), 'l/J13(l)'l/Ja(2) and 'l/J13(1)'l/J13(2) (4 states)
This way of counting the number of different states of the system is
known as classical or Boltzmann statistics. The above considerations for
non-interacting identical particles will be useful where we can treat the
interaction between the identical particles as a perturbation. This is illus-
trated in the next section.
H=Ho+V (14.29a)
!'--where
Ha = Ho(l) + Ho(2)
J3?
=---+---
2e2 p~ 2e2
(14.29b)
2m r1 2m r2
e2
V=-
r12
We treat V as a perturbation. For the unperturbed part the solution can
be written in the form
(14.30a) )
(14.30b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
>.V22 = e2 j u2__.!_u2d3r1d3r2
r12
1
~
(14.33b) )
(14.33c) )
(14.33d)
Now
Vi1 = "'22; Vi2 = "'21; Vi2 = V:t1, (14.34)
and therefore Vi2 and V21 are real. To order >.
x x
E±= E+ 2(Vi1 + V22) ± 2 /(Vi1 - Vi2)2 + 4IVi2l2
=E+ >.Vi1 ± >.Vi2 =E+ >.E11), (14.35)
where
E~)= Vi1 ± Vi2- (14.36)
The zeroth order wave function is then given by
1 ( E~) - Vi1 )
'l/J± = y'2 U1 + V12 U2
1
= v'2(u1 ± u2) (14.37)
1
= v'2 (ua(r1)u13(r2) ± u13(r1)ua(r2)). (14.38)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(14.41a) )
(14.41b) )
(14.41c)
For ortho state the wave function is given by
(14.42)
When the two electrons are in the same state a = {3 then 1/J- vanishe, ):uldi )
only para ·tate is pos íble and M the spins of the two electrons are anti-
parallel. Thus the ground state for the helium atoro is a para state. Also
we note that for the para state electrons can be very close to each other,
whereas for the ortho state two electrons tend to stay away from each other,
since in this case 'l/J_ vanishes as r1 ---+ r2. We have repulsíon when the spíns
of two electrons are parallel to each other and we have attraction when their
spins are anti-parallel. This is how a hydrogen molecule is formed, viz. we
get a chemical bond between two hydrogen atoms when two electrons have
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
opposite spins, because they can then come close to each other. Since no
more than two electrons with opposite spins can be roughly at the same
place, the hydrogen atoms must keep away from each other; the chemical
bond is formed between two hydrogen atoms only.
Consider the scattering of two particles a and b in the centre of mass system.
b
b
n-e
~ 2
14.6 Problems
where
'l/J10o(r) = ~¡33/2e-í3r,
Ea = Ebº) + 6.E
where
Z2e2 Ze2
Ebº) = ---, 6.E = (5/8)-
ao ªº
Hint: Ground state wave function is
where
)
un(r) = (f)213sinyxsin nf'ysinyz.
From this it follows that the number of quantum states between
momentum p and p + dp is given by [p = nk and 27T!;1 = kx etc.]
3 3
2 (!) dkxdkydkz = 2 (2~) d3k
3
= 2 ( 2~) 47rk2dk,
47r 3
V(r1 - r2) = -Vo3a ~(r1 - r2).
Treat this potential as a perturbation. Then using equations similar
U to Eqs. (14.33), (14.34), (14.36), (14.37), (14.39) and (14.41) in the
-1 µ,_ text and the result ofJvrst part of problem (14.4), calculate energy
levels and energy eigenfunctions. For a numerical estímate take
L = 10-8cm, a < 10-10cm and Vo = 10-3eV.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 15
15.1 Introduction
A two state system, i.e. when two states form a complete set, is simple /'
but rich in content. It provides a simple framéJvork for quantum mixing; L..I
a consequence of superposition principle of quantum mechanics. This leads
to quantum mechanical phenomena of interferometry which provides a sen-
sitive method to probe extremely small effects. Consider a system with
orthonormal basis {11), 12)}:
In general this is not a diagonal matrix and therefore the basis state~
and 12) are not eigenstates of H. Its diagonalization
#
det IH - MI = O
(15.3)
325
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Now the set of four matrices 1, rri (i = 1, 2, 3), the Pauli matrices, form
the basis in which one can express Has
H = nol + n· u (15.4)
1 .
= + H22), z(H12 - H21), (Hn - H22))
2((Hn
where
1
no= ~Tr(H) = Hn; H22' n = (15.5)
2Tr(Hu).
inc ·u)2 = n21, the eigenvalues of n-rr are ±[n[. This implies that the
eígeuvalues of H are
E±= no± [n] (15.6)
with the eigenvalue equation
no+ ~3
n1 - in2)
( n1 + in2 no - na
(U±)
V±
=(no± In!) (U±).
V±
(15.7)
U+=
(n n3)112
(2n)1/2 v+ =
(n - na)1l2
(2n)112
n1 + in2
Jn~ + n~
L )
(n - na)1l2 n1 + in2 (n + na)1l2
(15.10)
u.. =- (2n)l/2 Jnr -
n~' tr., = (2n)l/2
We can express the vector n in terms of spherical polar coordinates:
n = n( sin O cos (/J, sin O sin (/J, cos O) :
n= [n]
7
e sm2
() ·cp () I
I+) =cos2ll)+e' sin212) -, )
1-) = -.e " ·cp sin e
211) + cos ()
212)
L (15.12) )
or we can invert them
() ·cp () 1
11) = cos
21+) - e' sin
21-) v-- )
·cp () ()
12) = e-i sin + cos (15.13)
21+) 21-)-
Now the Schródinger equation of motion is
(15.18)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
In other words
The important conclusion one draws from this equation is that although at
t = O the system was in state 11), after time t it develops the component
12). The probability of finding the system in state 12) after time t is
. . !:iEt
Pi-2(t) = l(21</>(t))l2 = sm2 Osm2 -n-. (15.28)
P,_, = 1-},'_,
where e is the charge of the particle and g8 is the spin gyromagnetic ratio
of the electron, Dirac equation gives g8 = 2 (see Ch. (20)). Thus for the
electron (-e, e > O)
(15.33)
where
eñ eñ
Ws = 9s-2-B,Wb = 9s--b. (15.34)
me 2mc
Note that here the Hamiltonian explicitly depends on time. The
Schródinger equation of motion for the spin state lx(t)) is
The spin state lx(t)) can be written in terms of the basis provided by I!)
and 1 !), the eigenstate of S¿ = ~<Tz given in Eq. (15.30). Since H
-
where
e;2t o )
V(t) = -iwt
( O e----Y-
(15.37)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
and
V(t)vt (t) = 1, V(O) = 1 = vt (O)
and
tt = ñ (T ~)
~~
2 2
Ws Wb )
= ñ(
2o-z + 2a-x · (15.38)
Then the Eq. (15.35) takes the form
d
= V t (t)HV(t)lx(t))
-
iñ dtlx(t)) (15.39)
or
iñ :tlx'(t)) = H'lx'(t)) (15.40)
where
lx'(t)) = V(t)lx(t)), lx'(O)) = lx(O)) (15.41)
L )
(15.43) )
_)
(15.44)
Then from Eq. (15.27), using boundary conditions lx(O)) = lx'(O)) = 11) =
1
Then
+ !)
lx'(t)) = (cos ~t - i cos B sin ~t)ll) - i sin Bsin ~tl2) ~5.45)
(15.47)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore the probability that spin has flipped by time t i.e. the particle
has spin down is given by
= w~
(w - w8)2 + w~
. 2 [((w-ws)2+w~)!
sm
2
t l (15.48)
At resonance w = W8
This small deviation was expected and can be explained by radetive cor- O
rections to Dirac theory.
o
15.3 Particle Mixing
15.3.1 Introduction J
In this section, we discuss two state problem/ involving particle mixinfYt ~ ·
{j l:.
Pa:rticles are divided into two classes: leptons ~~'dnadrons. Leptons are pin ,<__~
!
particl~~ere are three generation · of leptons; he lightest one is/~}-ctron .J-4\.!:::)
( -) whi ·h Ís table. The experimental Iimit on the lifetime of electrón ~ ¡~
q:J x)_T-k:J~;5x1026yr. The second and third generation of leptons are muon (µ-)
and -r-lepton (-r-); each type of charged lepton ha k!respondu1g neutral fi\,
/ partner called neutrin<U Ve Vµ vT which are stable, The heavy charged
leptons decay into light r leptons by ,6-decay:
µ
T
--t
--t
Vµ +e- +ve
vT+µ-+vµ
l.-- )
--t VT +e-+ Ve.
Corresponding to each lepton, there is an antilepton with the same mass,
same spin but opposite charge. Each lepton is assigned a quantum number
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Hadrons are specified by intrinsic spin and intrinsic parity JP, J de-
., I\.CA. nofi1 the spin and P intrinsic parity. Hadrons are divided into two classes:
r,\ ~ _j baryon and mesons¡{ ~ow lying baryons have JP = ~ + and are assígned a
V\..,,, · quantum number B ~li,¡ antibaryons have B = -1, but same mass spin
j ~ negatíve intrinsic parity, The lightest baryon Li.Eroton (p), which is
1 . _ '!](.) stable _and ~sitiw,.e,1:t&-gep1),e experimental limit on protón life time:
~~~Jl)!Pe{!031yr. It has a neutral parEñer, called neutrón (n) which is unstable
and decaytnto a proton by ,8-decaywith a life time Tl :::: 4624sec.
conserving both charge, baryon number and lepton number. Each hadron
member is assigned an interna! quantum number, such as isospin I and
hypercharge Yff which are related to electric charge of each hadron:
)
y
Q=h+---;¡·
lfeutron and proton arebospin doublet I = !3 = (protón), h = !, ! -!
(neutron) and have hyper harg Y = l. Heavy baryons d cay to light r
baryons by weak interaction. Low lying mesons with spin and parity JP =
o- are 1r-, 1r0, 1r+ called pions:
fand kaons:
I = .!_
2' !3 =
1 -1
2'2' y= 1
L )
-=-=° 1 1 -1
(K ,K - ) : I= - 13 = 2' 2' y= -1.
2'
They are unstable; for example
Kº -t n+n- t- K° , ~Y=~lt- )
-o
Kº -t 7r+7["- -t K, l~YI =2.
(KºIHIK) f= o
i.e. off diagonal matrix elements are not zero. Thus Kº and K0 cannot be
mass eigenstates, when weak interaction Hamiltonian is included in H.
'lfJ(t) = (;~:D
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
is given by
12).
:22
(15.52)
=
r:a,L
ri2· (15.53)
)
Since xº is btiparticle of xº, xº and xº have same mass and decay
width:
(x°IMlxº) = (xºIMlxº)
(15.54)
Diagonalization of mass matrix M, gives the eigenvalues (c.f Eq. (15.3)]
(15.59)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
where
<J.= {M-;; = f fi2
o
m12 - = (Et. (15.61)
P V¡;¡;; m12 - ir12 q
The mass eigenstates IX±) form a complete set:
11/J(t)) = a_(t)lx-) + a+(t)lx+) L (15.62)
In this basis, the mass matrix M is diagonal so that:
i
. d
dt
(ª-(t)) _ (m- - ~r _
ª+ ( t) - o m+ - ~ r +
o ) (ª-(t))
ª+ ( t) .
(15.63)
The solution is
= a±(O)e-im±t-!r±t.
a±(t) (15.64)
Ifwe start with state lxº), i.e 11/J(O)) = lxº) so that from Eq (15.50) a(O) = 1,
a(O) = O, then from Eqs (15.60) and (15.62):
+exp(-im+t - ~r +t)lx+)l
= 21[exp(-zm_t-
. 1 . 2r+t
2r_t)+exp(-zm+t- 1 )l l x)o
_<.l.[exp(-im_t - ~r _t) - exp(-im+t - ~r +t)]lxº).
p 2 2
(15.66)
where we have used Eq. (15.60). Thus although at time t = O, the system
was in the state lxº), after time t it develops the component lxº). The
probability of finding the particles Jxº) i°r lxº) at time lis represented by/
P(xº ------, xº, t) = 1 (xº 11/J(t) )12 = -[e-rt + e-r +t + 2e- <co . .6.m~1~67)
P(xº ·-> Xº, t)- 1 (x°l,P(t)) I' - ~ 1 ~ I' [e4" -' + -r •' - 2 -r -• cos t.mt~¿
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
/}..m=m+-m-, (15.69)
One may make the following remark here: under the operation ( called
charge conjugation) which takes a particle state into its antiparticle state,
lxº) S lx°)
while under parity
lxº) s. -lxº)
so that under the combined operator C P
(15.70)
-o
15.3.3 Kº - K Complex
We now apply the general formalism developed in Sec. 15.3.2 to Kº - Kº
system. Here we denote K_ and K+ as Ks and KL. The subscripts S
and L stand for short and long lived since the particles Ks and KL which
have definite mass and life time have Ts = (0.8935 ± 0.0008)10-10s and
TL = (5.17 ± 0.14)10-8 s respectively so that 'rs « TL· Suppose that at
t = O, Kº (Y = 1) is produced by the reaction rr-p -----+ Kº A O• The initial
state is then pure Y = l. It is clear from Eq. (15.66) [with xº = Kº] that a
kaon beam which has been produced in a pure Y = 1 state has changed into
one containing both the parts with Y = 1 and Y = -1. Experimentally
Kº can be verified through the observation of hadronic signature such as
K p -----+ rr+A O since rr+A O can only be produc d by K° and not by Kº. The
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
P (K o -:-:O)
K , t = 1( 1 _ rs cos(b.m ))t .
4 + e ""S 2e_!_..L
__L
--+ - 2
(15.72)
(15.74)
where CPIKP,2) = ±IKP,2). In the decay K--+ 21r, (21r) state has CP = -l-L,
Thus the CP non conservation manifests itself by the ratio
where m12 = lm12le2if3. The mass eigenstates B'f', which we now denote as
BL,H are:
o
IBf) = ~[IBº) - e-2if31IJ°)JL
These equations are like those of a damped harmonic oscillator, the angular
frequency of which is given by
6.m 2rrli
W = ----¡¡:, T = 6.,m.
(15.83)
B~
= li d TBD ~
d
0.77
(6.mBo)
Xs = /i • TB~ ~ 26.2.
where
(15.85)
The mismatch between flavor and mass eigenstates gives rise to neutrino
oscillations.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics 1
l'lj!(t)) -iEt
= e----¡¡-
[ ( !::iEt - i cos Wsin-----¡¡:--
=:t: !::iEt) !::iEt lvx) ]
lve) - isin Wsin-----¡¡:--
(15.86)
where E= E+!E_ = Er!E2 and !::iE = Ei;E2. Then from Eq. (15.28) and
(15.29), the probability of conversion from Ve to Vx and survival probability
are respectively
. 2 • 2 !::iEtL
Pve-+Vx = sin 2() sin -ñ- )
Pve-+Ve = 1- Pve-+Vx· (15.87)
Here () is the mixing angle. Now
m2 1
e, = (k2
2
+ m¡)2 ~ (15.88)
lkl + k,
since k > > mi and we take the extreme relativistic limit. Then !::iE =
TkF
mf-m2
= -~Ll.m2 where !::im12
2
= m22-m1,2 it
· ·
is customary to take m2 > m1
and write lkl ~ Ev. Define
>. = 2rr = 4rrEv = 47r~ eV2 1012_1 __ (15.89)
E2 - E1 /::im2 MeV /::im2 MeV
Putting e and Ii back
1 cñ -13
MeV = MeV = L97x 10
m
e; eV2 (15.90)
.47mMeV /::im2'
Now L = et, (e = 1, t = L) is the distance travelled after which Ve is
converted into Vx· Hence
t
1
+
sin2 2()
t
-L- X= et
.tf- where tlm2 >o}< O) correspond to normal (inverted) mass spectrum hier-
archy. We also fuí'nmarize the various sources which provide neutrinos
15.4 Problems
11 .n. .n n ..
Pxil =
2{21 ± 2(sm2
2tsmBcosBcoswt + sm 2tcos 2tsmBsmwt)}. /
Chapter 16
Quantum Computation
16.1 Introduction
/
quantum teleportation and quantum network .
At present the consistent development in tbe subject has led to~~\
sophisticated quantum algorithms fundarnentally newer quantum comput-
ers, quantum clocks and quantum satellites for communication between dis-
tant parties. Secure communication, that is, cryptography is conventionally
based on factorization of large numbers, such as in RSA. It is challenged
by quantum ciphers for factorization such as Shor's Algorithm, and r scued
by the quantum cryptographic techniques, such as BB84 and Ekert92 . n
recent years secure data communication between distant parties basad on
quantum laws has crossed the barrier of 100 kilometers.
The infinite dimension Hilbert pa e simplifies drastically as we require
only two dimensions to develop a protobype of a classical bit. A quantum
mechanical bit ora qubit, therefore, becomes a simplest example of Hilbert
space and provides understanding of the superposition principle in quan-
tum mechanics. The understanding helps us to think of more and more
physically controllable systems to engineer quantum computers.
345
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
16.2 Qubit
IO) z
11}
Fig. 16.1
individual system. For this reason if l1Pi), 11?'2), ... 11fln) express n numbers of
qubits, the combined state of the system becomes
(16.4)
Hence, in a simplified situation the combined state of a two qubit system
becomes
J1P1,2) = J1P) @ l1P2)
J1P1,2) = (lc1IJO) + lc2IJl)) ® (Jd1IJO) + ld2Jll))
l1P1,2) = aJO, O)+ bJO, 1) + cll, O)+ dll, 1) (16.5)
where lü, O) = JO) ® IO) and same ·~ for the other terms. Interestingly,
in conventional computers for a two bit cp.se, we have the same four op-
tions, which are: 00, 01, 10, 11. However, inl,.Q.uantum domaln in addition to
these four options we have their superposition state as well. Furthermore,
the probability amplitudes are complex numbers which relate a phase to
each of the four options. This particular feature makes quantum mechanics
different from statistical mechanics and gives rise to the interference phe-
nomenon. It is important to note that as a special case if any one of the
amplitudes becomes zero, the two systems are not independent. Hence, we
r- canyot write the final state as a tensor product of the two independent
J states as we did in Eq.(16.5). In this situation two states are correlated, or
entangled to each other.
l1P) = ( :: ) .
This implies that the definition of a particular qubit requires the definition
of complex probability amplitudes, c1 and c2. Alternatively we say that
for a particular choice of c1 and c2 we have one qubit. We consider a two
t qubits system), say, J1P1) and 11?'2). The corresponding combined state of
the system shall be a tensor product of these two states. As we define tensor
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
S productJ..:;s a matrix of matrices, the combined state of the two qubits, lx),
becomes
(16.6)
Here, state vectors ['I/J1) and ['I/J2) are themselves matrices which appear as
elements of another matrix. The above expression can further be simplified
as:
(16.7)
11
1>
X
Ü>
Fig. 16.2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(16.9)
It is worth mentioning here that the square matrix defines NOT gate oper-
ation. Generalizing the statement we say that any operator can be defined
as a square matrix.
In contrast to one single bit logic gate in conventional computational,
in quantum dornain we have more. In sorne applications it is n eded to
introduce a particular phase on a qubit. For that trratte~we require a phase /
gate Y. The phase gate has he basis vec ors IO} and 11) a its eigenst¡1s" ,
witb eig nvalues +l and -1, respectively, ~Wi{ as shown in ig. 16.3~
/O •
:0> 1 y 1 ~~ 0
Fig. 16.3
BIO) =
A 1
1210}
+ 1
1211)
L )
, 1 1
Hll) = -10) - -11).
/2 /2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(IO>+ll>)/2
/\
H (IO>+ll>)/2
Fig. 16.4
ICg 2
1
ICe 12
t
Fig. 16.5
Fig. 16.6
Logic circuits are created from conventional logic devices to perform specific
tasks. Here our aim is to bring into light these tasks using quantum logic
gates, as discussed above. For this purpose, we consider two simple appli-
cations, one is the swapping operation, and the other is quantum cloning
in a special situation.
We consider a specific circuit shown below [Fjg. 16. 7]. The above cir-
l 'J!. >
1 l '1'2 > l 'l'out >
Fig. 16.7
l1/J1) = Ucont[l1/Ji)] k. L )
l'lf'1) = Ucont[la) ® lb)]
)
l'lf'1) = la)® la E9 b). (16.13)
Now the control bit for the next stage l1/J2) is la E9 b) and target bit is laL )
thus,
l'lf'2) = Ucont[l1/J1)] L )
l'lf'2) = Ucont [la) ® la E9 b)] ~ )
l'lf'2) = la E9 a E9 b) ® la E9 b) l- J
11/Jz) =lb)® la E9 b). (16.14)
First term in tensor product results state lb). Next considering lb) as control
(JI
1
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
bit and laEBb) as a target bit for the output state, lv,0), we proceed as follows:
l'if.io1tt) = 'Ucont[!1P2) J l- L. )
'ª A
>---,av
1 \jll > u
cnot
----1 O © a>= 1 a
'º ~-----~
Fig. 16.8
the output state of the systern becomes IV'oul:,uch that [st Fig. 16.8]
IV'out) 'Ucont[lv,1)] L
)
=
>
IV'out) = 'Ucont [la) © IO) JL. )
IV'out) = la) © IO EB a) L )
l1Pout) = la)© la).
This generates a clone of the controller qubit state, la), at the target qubit.
A physical example of quantum cloning phenomena is observed in the stim-
ulated emission process as an excited atom, in the presence of a resonant
photon, ernits an identical photon. The interaction, therefore, transforms
a single photon in the system to two identical photons.
A simple modification in the circuit for quantum cloning justifies that quan-
tum cloning in general is not possible. In addition, the outcome reveals
another purely quantum phenomenon, that is, entanglement. We consider
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
that the controller qubit of the above circuit now passes through a single
bit Hadamard gate, before the application of two qubit controlled NOT
gate operation. The corresponding schematic diagram is shown below 'g.
16.9]: Considering la) = IO), then
r: 1\
\.. l./
Fig. 16.9
= Ucont
1
11/Jout = .,/'j,(10) ® IO) + 11) ® 11)).
In the above equation we have controller qubit before tensor notation and
target qubit after it in each of the two terms. At this stage we have an
important and urpri ing ect. The control and target qubit tates cannot ,/ _
b ~ac_to~·ed ou and di cussed ü~depen~ently they are entan.~le~. IfÍfJr~t JH ,/
qubit is in stat zero th s cond 1. • al o in tate zero, wh r füJ.f ·t qubí il\ór j-- [-fµ..
i in stat - one, second shall ah o be in state one, and t he probability of
occurrence of the two events is equal.
From the above discussion we infer that in the process of quantum
cloning if in case the controller qubit is in a superposition state, in the
output state 11/Jout), we shall not have its clone at the target qubit. In
contrast we have engineered an entanglement between the controller qubit
and the target qubit. This implies that in general cloning is not possible,
which is established as no-cloning theorem.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o
_ \O, O) ± 11, 1)
1 ª± ) - v'2 .
Furthermore, as we apply a NOT gate to the target qubit before CNOT
operation, we obtain two other entangled states \13+) and \,6_), viz/_ •
I.B±) = 10, 1) ~11, O).
The four entangled states la±)and I.B±) are orthogonal and normalized to
unity. These are known as EPR states or EPR pairs. Therefore they define
a complete set of the basis vectors for a two qubit system. This implies
that we can write any arbitrary two qubit state in the vector space spanned
by the four EPR entangled states.
Historically, EPR entangled states initiated a hefty philosophical discus-
sion as they were used primarily to question the foundations of quantum
mechanics by Einstein, Podolsky and Rosen. In later years, the discussion
ik
between Einstein, Schrodinger and Max Born, led to comprehend the newly
develop d me hanics at a very fin scale. Interestingly in 196'1 John Bell
with the help olpla.me entangled tates settl d the lon~nding discus ion
on the validity of the classical laws or the quantum laws in understanding
~} [Bhysical world¡ai:eu,rd by calculating upper limit of strength of correlations,
t-1'4! J known as Bell's inequalities. He showed that the upper limit of correlation
between two parties would differ when calculated using classical mechanics
and quantum mechanics. The experimentally verifiable Bell's inequalities
led 0·"'-e} name these state F(Dell's state .
'
16.10 Problems
16.1 Show that the CNOT quantum logic gate is a reversible two&bit v'
gate.
{'\
) l/31:IO, 1) ~11, O) 1
define a complete set of the basic vectors for a two qubit system,
that is, the four entangled states are orthogonal and normalized to
unity.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
16.3 Show that the tensor product of two unit operators is a unit oper-
ator.
16.4 Show that the tensor product of two unitary operators is a unitary
operator.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 17
PERTURBATION INDUCED BY
ELECTROMAGNETIC FIELD
../
L
B
E= -v«: ! aA
= y' X A.
e at
We note that physically observable quantities E and B do not determine cp
(17.3a)
(17.3b) '
and A uniquely. For, if we change
359
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Eqs. (17.3a) and (17.3b) remain unchanged. This is known as gauge trans-
formation.
In terms of potentials A and </J, Eqs. (17.1) can be written as
'v2A- 2_ 82A -
c2 8t2
v (!e 8</J8t -r v A) = - 41rj(r,
e
t)l (17.5)
'v ·A= O.
This is called the Coulomb gauge,¡ fpr this case, we have
'77¿</J = -41rp(r)L (17.7a) }
o/-
1 82 A 41r.
'12 A- --
c2 8t2
= --J(r,t).
C
(17.7b)
For the non-static case, it is more convenient to select the gauge such that
1 8</J
--+'v·A=O. (17.8)
e 8t
This is called the Lorentz gauge. For this gauge Eq. (17.7b) remains unal-
tered, but Eq. (17.7a) becomes
2 1 82</J
'v <P - c2 = p(r, t). (17.9)
8t2
/\ For a source}!:,ee region, Eqs. (17.7b) and (17.9) become
1 82A
'v2 A - c2 8t2 = O, (17.lüa)
82<P
'72</J - 2_ = O. (17.lüa)
c2 8t2
For a free particle, the Hamiltonian is given by
p2
H=-. (17.11)
2m
For a particle of mass m and charge e moving in an electromagnetic field,
the classical Hamiltonian is given by
H
2m
= __]___ ~A(r,
e
(P -
2 + e</J(r). t)) (17.12)
[pi,Ai(r,t)] = -iñºA;(r,t)
Xi
= -iñV ·A. (17.16)
Since cp is independent of time, it follows from the Lorentz gauge that
V·A=O.
Hence
ñ,2 e e2
H = --'12 - -(2A· (-iñ"V)) +ec/J+--A2. (17.17)
2m 2mc 2mc2
We note from Eq. (17.13) that the Schrodinger equation for a free par-
ticle
the entire charge is there. The probabílíty is conserved and we have the
equation
(17.20)
where
j = - ('v1/;*)1/J) + _e_A1/;*1/J
_E!,_ (1f;*'v1/J (17.21)
2mc 2mc
is the probability current. If we have a situation in whích 1/J is the wave
function for each of a large number of particles with the same charge which
are all in the same state, 1/;*1/J can be interpreted as the density of particles,
then we can regard 1/;*1/J as electríc charge density and j as electric current. tk
We note from Eqs. (17.18) and (17.19) that if in !.!:::e particle <J
Schródinger equation (17.18), we make the substitutions
(l 7.22a)
(l 7.22b)
we get the Schrodinger equation (17.19) for a particle of charge e moving in
an electromagnetic field. Also we note that Eq. (17.19) is invariant under
gauge transformation (17.4) (with 'v2 A= O) provided that simultaneously
we make the transformation
'ljJ-----. 1/J'(r, t) = e+(ie/lic)A(r,t)'I/J(r, t). (17.23)
The last transformation is necessary in order to maintain the gauge
invariance. Lastly, we note that the electromagnetíc field energy
__.!:__(E2 + B2)
8?T
is also invariant under gauge transformation (17.4) and (17.23).
1
r 17.2 Gauge Principie:} Aharano"l'l8ohm Effect t:
V
In the previous section, we started from classical electromagnetíc theory
and wrote down the classical Hamiltonian for a charged particle in an
electromagnetíc field. We then went over to quantum mechanics by the
correspondence principie. In order that physícal laws as expressed by the
Schródinger equation be invariant under gauge transformations, the wave
function 'ljJ must undergo the space-time dependent phase transformation
(17.23).
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
(Here D and Do are called co-variant derivatives) in Eq. (17.18). Then the
resulting Schródinger equation
2
n2 ( i e )
- 2m v' - ,¡:~A iJ! = in ( at8 + ficp
ie )
iJ! (17.26)
This is the gauge principle. It gives the form of the interaction of a non-
relativistic particle with an electromagnetic field correctly. This has now
been generalised that all particle interactions at the fundamental level are
of gauge character.
Taking the vector potential A to be independent of time and putting
V= ecp, note that if wº(r, t) is the solution of the free particle Schródinger
equation, then
iJ!(r, t) = wº(r, t)eh(r) (17.27)
where
Dl];t~v' - ie A))
ñc
- é·y(r) v'iJ.!º J_
D2iJ.! = ér(r)y72iJ.!o. (17.29)
Thus (17.27) is a solution of Eq. (17.26) where A(r) -=f. O if wº(r, t) satisfies
..
y
~ (1)
-:,.-,-:------.-----~---J- p'
·~o . -- ---
*
p _,,.... d
\
- ~ '' j,
\.D.
.,, - J
_y
(2"
~ )
where
11 = 11O +~
e1 ( ')
ne
P'
(l)P
A r · dlI ~ (l 7.32b)
?
+~1 ( ') / P'
12 = 12O e A r · dl . (l 7.32c)
ne (l)P
g
Here 1f and 1 are the phases of the wave functions \[!~ and in the ab- wg
sence of A. The interference pattern is determined by the phase difference
8(B # O) = 11 - 12
= 1f - 1g + e i A(r') · dl,t
J
= 8(B = O) +a, (l 7.33a)
where C is the closed path P P' P and
In Eq. (17.33b) we have used Stokes theorem and put B = V x A and <I>
is the magnetic flux through the surface S bounded by the closed path C.
Note the important fact that the phase difference is gauge invariant. This
is because
f V A · dl' = f dA =O
while the individual phases 11 and ')'2 are not. Note also the remarkable
fact that the amount of interference can be controlled by varying magnetic
flux even though in the idealized experimental arrangement, electrons never
enter the region B # O.
Now referring to Fig. 17.1
Phase differe~ne = Path difference
2rr A
8 a 1 . dy
- = -dsmO ~ -- (17.34)
2 A A AL
where Lis the distance of the screen from the slits. Thus from Eq. (l 7.33a)
we see that the diffraction maximum of the interference pattern for B # O
is shifted from that for B = O by the amount t:J..y given by
=
e<I>
ne
(Ld A )
2rr
(17.35)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where we have used Eq. (17.34). This shift in the diffraction maximum,
being gauge invariant, should be measurable. In fact the existence and
magnitude of Aharanovj'[Bhom effect has been confirmed to within 5% of
the theoretical prediction (17.35) by two qualitatively different experimen-
tal arrangementsjone involving an electron biprism interferometer while the
second used a Josephson-junction interferometer. The following comments
are in order.
These are of great theoretical interest in the study of quantum Hall effect
(QHE). Considera particle of mass m and charge e in a uniform magnetic
field along the zt,=faxis for which the Hamiltonian is given in Eq. (17.13)
with </J = O.
First we note that the vector potential A is given by [see problem 17.1]
1
A= x r. (17.38)
2B
This satisfies the relation B = 'v x A and the gauge condition 'v · A = O.
For B = (0,0,B), A= !(-By,Bx,O).
The selection of A is not unique. A new vector potential
A' =A+'vA
with
also gives the same B. This gauge freedom can be exploited to eliminate
y-dependence. Select A= f
xy so that
.a .a a) 1
'vA= ( 1-+J-+k-
ax ay ez -Bxy
2
= [y1 + XJ']B
.
2
and then,
A' = A + 'v A = Bxj.
Removing the prime, we have
A= (O, Bx, O). (17.39)
Then, the Hamiltonian (17.13) is given by
,2 1
H = E._ - wxp
2m Y
+ -mw
2
2x2 (17.40)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where w = eB.
me
Now
[H,1\] = O, [H,Py] = O.
Thus it is possible to construct wave functions which are simultaneous
eigenfunctions of Py, Pz and H so that we can write
'1/J(x, y, z) = eipzzf ñeiPuYlñu(x) (17.41)
where
¡32 p2 p2 1 )
H'l/J = _____'E_+_]!..._ +-z -wxp +-mw2x2 '1/J
( 2m 2m 2m Y 2
=E'l/J (17.42)
or
p~ -wxp
( 2m Y
+ ~mw2x2) u(x) =
2
(E - 2mp~ - 2mp;) u(x)
or
p~ + ~mw2(x -
( 2m 2
Py )
mw
2) u(x) = (E - p;) u(x).
2m
(17.43)
Hence
* A (17.49)
n = 27rl1
where
(17.50)
is the magnetic length and A is the area of the system, which can accom-
modate one electron per energy level so that there is a total of n* = 2
electrons that fill the lowest Landau level. This has relevance to the Integral
:h
Quantum Hall Effect.
We confine ourselves to a pure radiation field so that we put (jJ = O and the
Lorentz condition then becomes \7 · A = O. The vector field A(r, t) then
satisfies the equation
where EA(k) is called the polarisation vector. In what follows we will write
simply EA instead of EA(k).
Suppose that the field A(r, t) is enclosed in a cube of volume V= L3,
then the functions Jveik·r, with ki = ni(21r/L) (i = 1,2,3) forma complete
set (ni= an integer). Thus we can expand A(r, t) as
A(r, t) = Jv L k
L
A=l,2
(EACA(k)ik·re-iwt + EA. c¡(k)e-ik·riwt),
(17.58)
where we have also summed over polarisations.
We note that
)
A A"
= ÓAA' Irf (17.59)
f
E • E
k· EA = k· EA• = Ü. (17.62)
We now calculate the electric and magnetic fields in terms of CA (k) and
c¡(k).
(17.64)
Hence the Hamiltonian of the radiation field H is given by
H = 2._
81r
f (E2 + B2)d3x
= 2._ ¿¿(w: (CA(k)C¡(k) + C¡(k)CA(k))
81r k A C
+k2(C.x(k)C¡(k) + c¡(k)C.x(k))
+oscillating terms), (17.65)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
H = :
7r
L Lx w:e (C.x(k)C;:(k)
k
+ Cl(k)C.x(k)). (17.67)
.
If we write
C.x(k) = --
v'47rc (P.x(k) - iwQ.x(k))
. (17.68)
2w
Ct(k) = ~
J41rc (P.x(k) +iwQ.x(k))
. (17.69)
then
H = i LL k .X
(P}(k) +w2Q¡(k)). (17.70)
H = i ¿¿lk.J
k .X
(a.x(k)al(k) +al(k)a.x(k)). (17.73)
Note that a.x (k) and al (k) are now operators which satisfy the commutation
relation
a.x(k)IO) = O L (17.75) )
al(k)ln.x(k)) = v~n.x~(k~)-+-11n.x(k) + 1) L (17.76) )
a.x(k)ln.x(k)) = v'n.x(k)ln.x(k) - l)v (17.77) )
al(k)a.x(k)ln.x(k)) = n.x(k)ln.x(k)), (17.78)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
ln.\(k)) = ~ 1
n.\(k)!
( a1(k) r" IO). (17.79)
(17.80)
where
(17.81)
is called the number operator, because it has eigenvalue n.\(k), where n.\(k)
is an integer. The n number of photons of wave number k and polarisation
>. is represented by a state ln.\(k)):
(17.84)
represents a state with two photonsJone with wave number k and polarisa-
' tion >. and the other with wave nu~er k' and polarisation X. Also from
Eqs. (17.76) and (17.77), we have
2m&
e A e2 2
=Ho--A·p+--A. (17.88)
me 2me2
We call H0 the unperturbed Hamiltonian. The other two terms in
Eq. (17.86), we regard as a perturbation. Since 2~c2 A 2 is of higher or-
der as compared with the term .;:cA ·
f> (because e2 is small) we neglect
this term. Thus the perturbation potential is given by
1 H inL =- ec./4irli ~
~ ~~ --1 (
ª><
(k')
e
ik'·r
e
-iw'L
mcv'V k' >,.' ..fiwi
X E>..'. f> + al,(k')e-ik'·riw'té'*. f>). (17.90)
We take the matrix elements of (17.89), between initial state la, n>..(k)) and
final state lb, n>,(k)-1). The initial state contains n photons ofwave number
k and polarisation >. and the final state contains ( n - 1) photons with the
\ same wave number and polarisation. We thus have, using Eq. (17.90)
Tba(t)E:),n>..(k) - llHintlb))
= _ _!:._(b, n>..(k) - llA · f>la, n>,(k))
me
_ _!:._ eJ41rn ¿ ¿ -1-, (b, n>..(k) -11
me JV k' >..' -.fiwi
X (aN(k')eik'·re-iw'tE>..'. f> + al,.(k') -----~
xe-ik'·re+iw'tE>..'* · f>)la, n>..(k)). (17.91)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
2nc2ñn>.(k) x i(k·r-wt)
Vw € e (17.94a)
(17.95a)
we have from Eqs. (17.95)(with >.=-e), (17.94a) and (17.94b), for absorp-
tion or emission of a photon of frequency w and polarisation >.:
Similarly
w:;1isijiQII
41r2e2
= -2-l(bleik·r€.>- · f>la)l28(Eb - Ea - nw).
m wc
~ ~ p(R)dEg{,(w)hd,.,, (17.98a)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
376 Quantum M echanics
where
_ V ( 2 dp) V 2 (17.98b)
p(E) = p(w) = (21rfi)3 p dE df2 = (21rc)3fiw df2
27r:~nc3
(nA(k) + l)l(ble-ik·r EA•· :f>la)l2df2 (17.99a)
with
(17.99b)
Similarly the total transition probability from an atomic state a to a state
b with the absorption of a photon is given by
(17.lOOa)
with
(17.lOOb)
which is the Bohr frequency condition for the absorption of radiation of
frequency ñi», Using the above formulation, we can derive Planck's ra-
diation law from the point of view of quantum field theory. Suppose we
have atoros and a radiation field which can freely exchange energy by the
reversible process
(17.101)
in such a way that thermal equilibrium is established. In the above tran-
sition we have taken the atoro to have two states of energy, Ea and Eb
with Ea > Eb. If the population of the upper and lower atomic levels are
denoted by N(a) and N(b) respectively, we have the equilibrium condition
N (a) W,;',:_nission = N ( b) w::sorption. (17.102)
Now from Eqs. (17.99) and (17.100), we have
W,;',:,nission (nA(k) + l)l(ble-ikr€A* . f>la)l2
(17.103)
w::sorption - nA(k)l(aleik-ré. f>lb)*l2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
But
(ble-ik-rE.>-*. f>la) = (alf>t. E>-eik-rlb)*
= (aleik-r E>- • f>lb) *. (17.104)
From statistical mechanics, we know
N(b)
(17.105)
N(a)
But using Eq. (17.104), we have from Eqs. (17.102) and (17.103)
N(b) = w::ission = n>,(k) + 1
(17.106)
N(a) w::sorption n>,(k)
Using Eqs. (17.105) and (17.106), we obtain
1
n>, (k) = eñw/kT _ (17.107)
1.
The above result is for photon states for which ñw = Ea - Eb. Suppose
the radiation field is enclosed in a cavity. The walls of the cavity are made
up of various kinds of atoms and are capable of absorbing and re-emitting
photons of any energy. The energy of the radiation field, per unit volume,
in the angular frequency interval w and w + dw is given by
1 V 3 1
U(w)dw = V (2rr)3 d k2 eñw/kT - 1 ñw
= _l_w2dw41r 2ñw
(2rrc)3 eñw/kT _ 1
The factor of 2 in the above expression comes from the two polarisation
states and in deriving the above formula we have used Eq. (17.98a). The
energy distribution per unit frequency per unit volume (w = 2rrv)
dw
U(v) = U(w) dv
8rrhv3 1
= c3 ehv/kT _ 1. (17.109)
decay in a very short time, thus giving rise to an intense, coherent, and
monochromatic pulse of radiation. In a laser, we have just this. But at
equilibrium, we have
N(a) = N(b)e-ñw/kT (17.110)
and it is not possible to have N(a) > N(b). Sorne special techniques are to
be used for population inversion in which the number N(a) in the excited
state is very much greater than the number N(b) in the lower state so that
the number in the lower state is practically zero. Then light which has the
frequency ñw = Ea - Eb will not be strongly absorbed, because there are
not many atoms in state b to absorb it. On the other hand, when that light
is present, it will induce the emission from this upper state. So, if we have
many atoms in the upper state, the moment the atoms begin to emit, more
would be caused to emit and we will have a strong stimulated emission.
Dipole Approximation
From Eq. (17.99a), we see that even in the absence of radiation field
(n.>-(k) = O), the probability for radiative transition from an atomic state
a to a state b is not zero. This is called spontaneous emission of radia-
tion. From Eq, (17.99), the total transition probability per unit time for
spontaneous emission from an atomic state a to a state b is given by
2
Wba = e w [ (b[e-ik·r i>.*) · p[a)l2dD. (17.111)
21rm2ñc3
We now expand the exponential e-ik·r
e-ik-r = 1 - ik · r · · · (17.112)
[kr[el (17.113)
then it is a good approximation to use only the first term. This is called
the dipole approximation and it holds if Eq. (17.113) is satisfied. Now r is
of the order of the atomic radius (Z is the atomic number)
ñ,2 ñ
rrva=--=-- (17.114)
mZe2 mc Z a
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and ñw is given by
ñw = Ea - Eb = - mc2z2a2 (__!__ - __!__)
2 n~ n~
(17.115a)
so that
(17.115b)
Hence
1
kr rv -Za
2
rv !z- 1-
2 137'
(17.115c)
where a = ~: = 1~7 is the fine structure constant. Thus the dipole approx-
imation is valid, unless Z is very high.
The transition probability Wba in the electric dipole (El) approximation
is given by
Wba =
ca» [(b[E >.* · pfa)f 2 dD..
A
(17.116)
21rm 2 e2
Now
in~!= =
[r,HJ
[r,Ho + .XV]
= [r, Ho] + terms of arder .X or e. (17.117)
We can neglect these terms since we are calculating Wba to arder e. Thus
we have
(bf f>f a) = : (bf [r, Ho] [a) + O( e)
im
~ --¡;(Ea - Eb)(bfrfa)
= -imw(bfrfa). (17.118)
Hence
~j
o
(17.119) '
where
(17.120)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Since €(1), €(2), ek ( = l~I) form an orthonormal set, satisfying the relations:
2
¿é') .é>.) = 2,
A=l
'°'
L
2
€i
(A) • (>.) _
€j -
A ..
UiJ
_ kikj
k2 , (17.121)
A=l
C we can express b4 r as
J Thus
r = r(sin () cos <pE(l) +sin() sin <p€(2) + cos ()ek].
O
we take € along z-axis and k in x - y plane and
Since Yim has parity ( -1) 1 and l = 1 in Eqs. ( 17.122), we can say that dipole
photon carries an angular momentum of one unit, and negative parity. Thus
parity conservation give:
(17.123)
(The initial and final states have opposite parities).
On the other hand from the rules of addition of angular momentum in
quantum mechanics r>\
IJa - Jbl :=:; 1 s i, + s, L (17.124) U
The above relation implies that the change
(17.125)
must be restricted to l:!,.J = O, ±1; J¿ = Jb = O is strictly forbidden since the
emitted or absorbed photon (spin 1) cannot have zero angular momentum.
The above are the parity and angular momentum selection rules for an
electric dipole transition or El transition.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
The above selection rules also follow directly from the electron dipole
matrix elements given in Eq. (17.119). We note that initial and final states
la) and lb) for an atomic electron for a hydroge~e atom can be specified "
by
la) = lnilimi) ~ (17.126a) )
lb)= jn¡l¡m¡) L (17.126b)
)
(ria)= Unilimi(r) = Rn;i.(r)YtimJ(),rp) L (17.126c) __.,
(rlb) = Un¡l¡m¡ (r) = Rn¡l¡ (r )Yt1m1 ( (),rp). (17.126d)
Hence, using Eqs. (15.122),
(blr · Etla) = (n¡l¡m¡lr · Etlnilimi)
But
since 1oC1), 1oC2) and ~ form an orthogonal coordinate system. Also we have
=i~L±.
Thus
k · ( rp + pr) · E>...
(blk · (r:f> + :f>r) · E>..la) = ~ (blk · {r[r, Ha]+ [r, Ho]r} · E>..la) + O(e)
We see that first term of Eq. (17.130) gives rise to electric quadrupole (E2)
transition. The selection rules for E2 transition are t::.l ::::; 2, no change in
parity of atomic states (this excludes t::.l = 1). Thus the selection rule for
E2 is t::.l = 2. The matrix elements for E2 transition are suppressed by a
factor Za: as compared with El transition as is the case for Ml transition.
The decay rate for the radiative transition from a state a to state b is given
by
(17.133)
I' is called the decay width. The mean life time for this decay is given by
(17.134)
17.8.1 E1 transition
For El transition, we have from Eq. (17.119)
(17.135)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
The sum over two polarisation states gives, using Eq. (17.121e)
= (blrila)(bhla)*(óij - k~:j)
12
= 1 (birla) 12 - lk. (~;la)
r a-->b
aE3 r {21r
lo lo (1 - cos2 Bk) sin fhdfhd</Jk
=
21rñ2:21 (birla) 12
= 4a 3 12 . (17.137 )
3c2ñ2E7l(blrla)
-
¡\
As a concrete example, we calculate the mean life-time for the transition
2p--+ ls for a hydrogenb_ke atom. From Eq. (17.137), we have
r2p--,1s 4a 31
= 3c2ñ2E73 ~
L.,
m=-1,0,1
1
¡ * ( r)ru21m(r)d 3 r ¡2,
u100 (17.138 )
where we have averaged over the three possible values of m for the 2p state.
Now
X
r /2,r
lo le Y0ii
(sin(} cos </J)
sin sin <P Y1m sin (}d(}d</J.
é
o o cos (}
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
Thus
r,,_,. - 9~~,E~(¡=
Rw(r)R,,(r)r'dr)2~) L (17.139a) )-1,
where
Hence
(17.141)
Now
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
(17.142)
Thus [using Eq. (17.141)]
rzp-.is = (~)
8
mt a(Za)4 ~ )
T2p-->ls = ( 23) 8 ñ
mc2
1
a(Za)4 (17.143)
r
For the hydrogen atom, Z = 1, therefore
T2p-->ls = ( J ~2 ~5 (17.144)
ha =
e w
2 4
27rC3
¡ "°"'
~
.\=l
2
1 (birla) · E.x• I 2 dD., (17.146)
i.e. the classical formula for the intensity of light emitted by an oscillating
dipole.
L
side of Eq. (17.131) is replaced by
This is the decay rate for the radiative transition from the staJ a to b due
to intrinsic magnetic dipole moment µ Now, using Eq. (17.lfi}
8•
e. ~
¿lo-· (k x E*>.)12 = aiai ¿(k x E*>.)i(k x E>.)j
>. >.
t- L Aº7
The transition magnetic moment [mp is the proton mass used in nucleon
(17.150)0
magneton]
o
µ(¿-Aº)=
o
r(¿-Aºry)
® 17.9
T=
n
-
r
.5 X 10-20s
i = ¿ ln)(nl (l 7.157a)
n
Hin)= Enln), (l 7.157b)
we have
(O/xi(t)xj(t')/0) = ¿(Olxi(t)/n)(n/xi(t')/0). (17.158)
n
= f 00 dEe-iE(t-t')/ñ ¿(O/xi/n)(n/xi/0)
Ío n
(17.162)
This is known as the spectral representation of the two point function de-
fined in Eq. (17.157) and Pij (E) is called the spectral function.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
4. The spectral function Pii(E) has the following propertíes: ( C's are "
complex numbers) ....,
So far we have not used any dynamics. In order to relate Pii with Pii,
we have to put sorne dynamical information. Now we have
!!:_x·(t) = !!:_(eiHtfñx·e-iHtfñ)
dt i dt t
(17.167)
Hence
!!:_X. (t) = ]_eiHtf ñp'-e-iHt/ñ
dt ' m '
= ]__)3i(t). (17.168)
m
Using Eq. (17.168), we obtain
= ;22
Jo
t" dEe-iE(t-t')fñE2Pii(E). (17.169)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
i: fo l )
o
(Olfii(t)i;j(t')IO) = - 00 dEe-iE(t-t')/ñEp;j(E)
(17.171b)
Under time reversal t - -t, r - r, :f> - -:f> and we see that commuta-
tion relation (17.154) is not invariant under time reversal unless either we
change the order of operators in a product, or take the complex conjugate
of complex number [see chapter 10]. Therefore, if we assume time reversal
invariance, we get
(17.172)
Therefore, from Eqs. (17.171a,b), we have
(17.173)
Hence we have
Pji(E) = Pij(E). (17.174)
7. Now from Eqs. (17.171) and (17.174)
(OI [1\ ( t), Xj ( t')] IO)
= _ im {00 dE(e-iE(t-t')/ñ + eiE(t-t')lñ)Ep;j(E). (17.175)
n lo
At equal time t = t'
[ih(t),xj(t)] = eiHtfn[p;,xj]e-iHt/n
= -iMij· (17.176)
Hence from Eqs. (17.175) and (17.176), we have the sum rule
00 ñ2
1o
dEEp;j(E) = -2óij·
m
(17.177)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
Note that in deriving Eq. (17.177), we have not used any detailed form of
potential in the Hamiltonia~xcept that it does not involve the momentum
p.
The ground state lü) is arbitrary. Consider the simple case when the
ground state IO) is spherically symmetric, i.e. it is an s-state. Then we
have rotational invariance so that
(17.178)
If the ground state has non-zero spin, then we sum over spin and again we
have rotational invariance. In this case the sum rule (17.177) reduces to
t" n,2
lo dEEp(E) = 2m. (17.179a)
3m2 t"
(p2)o = ,i,2 lo dEE2p(E). (17.179c)
To sum up, sum rules (17.179) have been obtained on very general grounds,
namely, completeness, translational invariance, existence of the ground
state, and the basic commutation relation of quantum mechanics. In ad-
dition, we have used time reversal invariance and the form (17.166) of the
Hamiltonian to derive Eq. (17.168). Equation (17.168) follows essentially
from the fact that the potential in the Hamiltonian does not depend upon
the momentum.
We can easily show that Eqs. (17.179) lead to the uncertainty principle.
For this purpose, we use the followingmathematical result.
Define an inner product
en ( 00 ) = ie
- mñ y¡-;¡;¡¡:-
M21r1w"'( En - Ea "· ·) ( neI ik·r p · ex I O.)
-1WJ
A
(17.184)
Now
2
= m2w2pij(w)
= m2w28ijp(w). (17.186b)
Hence
e241rli 1 2 2
W = m21i2 w p(w)
2wV21rñ3(6ijÓij)m
e2 41rli
= li2 21rñwp(w). (17.187)
2V
The total cross-section O"(w) for the process (17.183) is given by
w
O"(w) = Flux' (17.188a)
1
Flux= pv = Ve. (17.188b)
Hence
e2
O"(w) = 1r(41r-)wp(w)
e
= 41r2ñawp(w) (17.189a)
or
__ 1 _0"(w)
P ( w ) - 41r2oo w · (17.189b)
Rewriting the sum rules (17.179) in terms of w, we have
f'XJ d!..tJwp(w) = ___!._L (17.190) )
lo 2m
lot:d!..tJp(w) = 1
(r2)0 31i L (17.191) )
{00 2 1 2
lo dux» p(w) = 3m2 li(p )o. (17.192)
Now using Eq. (17.189b), we have
¡ L
~
1
4 1r a
00
d!..tJO"(w) = -2Tim (17.193) )
L
O
_1_
41r2a
r: dwO"(w)
lo w
= !(r2)º
3
(17.194) )
_1_
2
4tt a O
¡ w..uWO" ( W ) = t1s2~
00 ,J..
t: 3
(p2)02 .
m
(17.195)
Since O"(w) is the total cross-section and can be determined experimen-
ally, the above sum tules can be testad. Since we have used the dipole
approxirnation.Íthird um rule (17.195) cannot be tru tedas in this sum
rule b cause of w factor in the numerator, large valúes of w be ome lrnpor-
ant, but for large values of w,~pole approxirnation breaks down. In this
resp et, th sum rule (l 7.19tl) is the best one, as w is in the denominator,
(r2)0 is related to the radius of the atomic system. This sum rule relates
the radius of atomic system with the total cross-section.
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
We have seen that (Oli:i(t)i:j (t'))o has the following spectral representation
(cf. Eqs. (17.162) and (17.178)), E=mv
(Olxi(t)i:j(t')IO) = Óijñ 1= dwe-iw(t-t')p(w). (17.196)
If v < O or Im(t - t') < O, ewv acts as a damping factor. We can then
take derivative of the above expression with respect to v, since the spectral
representation of
(Olfii (t)pj ( t') IO)
contains the spectral function
{Jij(w) = m2w2pij(w). (17.198)
Further, the quantity of physical interest is the time ordered product of two
point functions and their Fourier transformations. We define time ordered
product as follows:
T(xi(t)xj(t')) =
=
xi(t)xj(t')
xí(t')xi(t)
t > t'
t < t'.
L (17.199)
1
Let us consider the Fourier transform of
(OIT(pi(t)pj (t')) IO),
= r
lt,
dte-iw(t-t'l(Olfii(t)pj(t')IO) + r dte-iw(t-t')
Loo
t'
x(Olfij(t')pi(t)IO). (17.200)
Using Eq. (17.169), we have
Gij(w) = m2Mij(1= dte-iw(t-t') ¡= dw'e-iw'(t-t'lw'2p(w')
Ío
+ ¡ t'
-oo
t'
dte-iw(t-t') ¡=
o
dw'eiw'(t-t')w,2 p(w')). (17.201)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o 1
+ 00 dw' w'2 p(w')
-oo
dte-i(w-w')(t-t')). (17.202)
Ít'
~
t" dte-i(w'+w-ie)(t-t') = ,
W +w
-i
.
-iE
. L }
r
Í-oo
dte-i(w-w'+ie)(t-t')dt =
w - w'
i
+ ÍE'
(17.203)
Gij(w) = ¡_:
where we take the limit E --+ O at the end. Thus
dte-iw(t-t')(OIT(fti(t)pj(t'))IO)
= -Wi•,: ·m 2 (1,
J
100
O
dw ' W ,2 p ( W ')(
w' +w -
1 . -
iE
1
w - w'
.
+ iE
).
(17.204)
Hence
Gij(w) = i8ijm ñ
We note that
w 2 -w'
12
+i1:
. = P 1
w 2 -w'2
.
-i7ró(w 2 -w ) <: (17.207a)
)
8(w2 -w'2)
2~
= (ó(w -w') + 8(w + w')) (l 7.207b)
where P denotes the principal value integral. Thus we obtain
. m,2 w2o(w)
Im (iGij(w)) = ói;221r
=
;:2t w
8ii--o(w)
2
l_ W
(17.208) )
Re(iGij(w))
. m2
= Óij-2
2
P
7r a
47r
1=
O
w'20"(w')
w'2 -w 2dw
·
,
(17.209)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
By optical theorem
w
Im f(w,0) = Im f(w) = -o-(w),
47rC
(17.210)
we see that
(m~c iGij(w)) = Óijf(w) (17.211)
represents the forward scattering amplitude for the photon scattering.
Hence we have from Eqs. (17.207) and (17.210)
f(w) = ~
7r
t" dw'w'
lo
Im f(w').
w'2 - w2 - u:
L (l 7.212a) )
Re f(w) = ~p f
00 dw'w'1~ f(w') I (l 7.212b)
rr lo w' - w2 l---- )
Re f(w) = _l_p {oo w'20-(w') dw'.
(l 7.212c)
27r2C w'2 - w2 lo
The above equations are known as dispersion relations. In actual physical
cases, the forward scattering amplitude f(w) <loes not vanish as w - oo.
Therefore the above integrals do not converge. In such cases, we make a
subtraction at w = O, so that we get
r= !(~') . L (l 7.213a)
2w2 1:
f(w) - f(O) =
7r 1o
dw' (
w' w' - w - u,)
)
2w2
Re (f(w) - f(O)) = P [" dw' Im {(w') ¡ (17.213b)
7r lo w'(w' - w2) L )
Re (f(w) - f(O)) = __:!_p f 00 o-(w')dw'.
(l 7.213c)
27r2c
w'2 - w2 lo
From the above equations, we see that f(w) is an analytic function of w
in the cut plane starting from O to oo. Since cr(w) can be determined
experimentally, by the dispersion relation (17.214c) we can determine the
forward scattering amplitude.
17.12 Problems
Show that
A 1 e
Q = -(fix - -Ax) (17.214)
mw e
and
A 1 A e
P= -(P --A) (17.215)
mw Y e Y
satisfy
[Q,fi] = iñ. (17.216)
Show that Hamiltonian
H = __2_(p - ~A)2 (17.217)
2m e
can be written as
p2 1
H= -+-mw2Q2 (17.218)
2m 2
and hence the energy eigenvalues, called Landau levels are given
by
1
E= (n+ (17.219)
2)ñw.
Show that the other canonical pair
A e
(Py - -Ay) QA' =
e
,, 1 e
p = -(Px + -Ax)
A
(17.220)
mw e
are cyclic i.e. they commute with Q, P and do not occur in the
Hamiltonian.
17.3 For a particle of charge e and mass m, in an electromagnetic field
(c/>,A):
H = --v'
2m
ñ2iñ
+ -eA·
me
The expectation value of H is
2
v' +--A
2me2
e2 2
+ec/>
.. eo
Show that
'~
'Ji= .!!__ \11/J . V'ljJ + ieñ ('ljJ"V'f/, -1/JV'if;*)
2m 2mc
2
+ _e_A2'lj)*'lf; - e'lj)*'l/Jcp
2mc2
= _!!_V'lf;* · V'l/J + e(pcp -j · A)
2m
where
P = 'l/J*'l/J,
j = ~('!f;*V'f/J-'1/JV'I/J*) = _e_'l/J*'f/JA2.
2mci 2mc2
The Lagrangian density .C = - 'H. Verify that
ºP
ot+v·J=. t"'7 • o
17.4 Show that for a hydrogen-like atom, parity operator P gives
(n¡l¡m¡JrJnilimi) = (-1f+l¡-l(n¡l¡m¡JrJnilimi/·
Hence parity conservation implies that (-lt f= (-1)1! i.e. atomic
state must change parity for E1 transition. This is the parity se-
lection rule. Note also that (l¡ - li) must be odd.
17.5 By using the relation
[L2, [L2, rl] = 2ñ.(rL2 + L2r),
show that for a hydrogen-like atom
l¡ - li = ±1
or l¡ = -li· However l¡ = =l, only if l¡ =O= li but then
(n ¡OOJrJniOO).
vanishes.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(rlb) = JveiPe·r/ñ.
Using these results, we can express the total transition probability
per unit time for the absorption of photons of frequency w as
e2 21rñ V
Wba = m2ñ2 Vw (21rñ)3
J 2mpedEedf!21rñ
xó(Ee + Ea - ñw)l(bleik·rE>. · f>la)l2.
Now we can write
(ale-ik·rE.>-. f>lb) =E>.. Jv J d3ruioo(r)e-ik·rpeeiPe·r/n
da 32,,/2Z5a8a2(Ee/mc2)112 sin2 (}
4,
dn {(a.Z)2 + ~(1-ve/ccosB)}
where we have neglected the binding energy and have put ñw ~
p~/2m, a= e2 [he and Ve is the electron velocity.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
17.7 Show that (OIEi(r, t)Ej(r', t)IO), the expectation value of electric
field at 2 points in vacuurn (i.e. no photons), is given by
lim
o-+O
roo e-ok sinklRJdk,
lo
show that
and D1 (R) --+ oo when R --+ O, i.e. r = r'. Since what we rneasure
by a test body is the field strength averaged over sorne region in
space, it may be more realistic to consider the average field operator
about sorne point, e.g. defined as
E= ó.lV t: Ed3r,
Chapter 18
FORMAL THEORY OF
SCATTERING
18.1 Introduction
Ha= p2/2µ
403
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
404 Quantum Mechanics
(µ is the reduced mass for scattering of a system of two particles) and the
scattering interaction causes transitions from an initial state with momen-
tum Pi to final states characterised by momentum P¡·
We write the Hamiltonian as
H=Ho+V, (18.1)
where H0 is the kinetic energy operator of two colliding particles. For this
simple case in the centre of mass system, H0 = p2 /2µ. For the case of two
colliding systems Ho also contains their separate interna! energies. V is the
interaction between the colliding particles responsible for their scattering.
H0 is so simple that solutions to the eigenvalue equation
are known.
The Schrodinger equation is
The eigenstates In) of H0 forro a complete set so that we can write the
solution 17/J( t)) in terms of them, i.e.
(bln) = Óbn,
we get
(18.5)
where
( )<: ).....
It is convenient to introducMn' and°1ut' state. Thus Ja)¡71 = la, t - -oo),
la)out = la, t=» oo). Comparing Eq. (21.8) with Eq. (21.4), we can write
the initial conditions of the problem:
Ca(-oo) = 1,Cn(-oo) = O, n-=/= a. (18.9)
We are interested in calculating Cb(t) for large t, so that
d
dt ICb(t)l2 for large t
gives the transition probability per unit time from initial state la) with
propagation vector ki(Pi = ñki) to final state \b) characterised by propa-
gation vectors k ¡ (p ¡ = ñk ¡). As we shall see the above quantity is related
to the cross section.
To fírst arder in p rturbation theory (i.e. put Cn = O, n # a, Ca = 1 M._ ....._J,\J lVAc. kt f-
on the right-hand side of Eq. (~we get a solution of Eq. (21.5) ~ ,-~ 1 .J
The integral in Eq. (18.lOb) is not defined for t' = -oo. To give a meaning
to the integral, we use the following prescription. Introduce a factor erit' in
the integrand and write
Cb(t) = - ~
. Vi,ª
ft
-oo eÍWbat'+r¡t' dt' + Óba, (18.11)
Iim n-s O
where the limit r¡ ----+ O is taken at the end, i.e. lim r¡ ----+ O is taken after the
lim t0 ----+ -oo. The above prescription will hold for a time t which satisfies
1
\ti«-. r¡
=Ho (18.15a)
and
H¡(t) = eiHot/fiHe-iHot/fi
Assume that [1/;(t))r is generated from [1/;(t0))r by a linear operator U(t, t0)
[1/;(t))r = U(t, to)[1/;(to))r, (18.18)
where
Therefore
U(t, t')U(t', to)= U(t, to). (18.21a)
This is known as the group property of the operator U. In particular,
Substituting back, the expression U(t, t0) from Eq. (21.22), we have an
iteration procedure:
= 1- ~
Áo
r dti V¡(ti) + (-*)2 Áor dt'V¡(ti) Áor: V¡(t2)dt2 + · · ·.
(18.23)
This can be used as a basis for a perturbation expansion, each successive
term giving one higher order of expansion. For t--+ oo, t0 --+ -oo
= L Cn(t)ln),
n
so that
(bl1/J(t))r = L Cn(t)óbn
n
(18.25)
1see J. J. Sakurai, Advanced quantum mechanics, Addison-Wesley (1973) p.186
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
Cb(t) =(bl'I/J(t))1
= lim (blU(t, to)l'I/J(to))!. (18.26)
t0-,-oo
But
Therefore,
= lim e-iEatofñeiEato/ñla)
to-,-oo
= la). (18.27)
,r Thus
----
From Eq. (21.23), we see that to first order in perturbation
Therefore
= Óab - i (bl
n
r dt' eiHot' /nv e-iHot' /ñja)
= Óab - ~ t. -oo
(18.30a)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
But
11/l(t)) 1 = U(t, to)l1/l(to)) 1
= U(t, to)eiHoto/ll.11/l(to)). (18.30b)
Comparing Eqs, (18.30a) and (18.30b), we have
U(t, to)iHoto/ñ = eiHot/ñe-iH(t-t0)/ñ
or
(18.31a)
Hence
J1/l(t))¡ = eiHotfñe-iH(t-to)/ñe-iHotofñJ1/J( to)) I. (18.31b)
From Eq. (18.31a), we have
u'«. to)= eiHotofñeiH(t-to)fñe-iHot/ñ
= iHotofñe-iH(to-t)fñe-iH0t/ñ
= U(to, to)
= l. (18.32b)
Hence
ut(t,to) = u-1(t,to), (18.32c)
Thus U(t, t0) is an unitary operator.
(\ Note that H0 and H do not in general commute and therefore we cannot
add up the exponents in Eq. (16.31). Since ultimately we are interested in
~: lim U(t, t0) and lim lim U(t, t0),
to--= t-+= to-,-0<¡
we first define the above limiting procedure. We shall use the following
prescription. Consider first a function J(t). Then we define
and
lim J(t)
t-+=
= lim r¡
11-0
r=
}0
e-r¡t' f(t')dt'. (18.33b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= f (O) - f (t')l~oo
= f(-oo).
If J(t) oscillates at t = +oo, then at large !ti, the corresponding expo-
nential factors e±r¡t will average out such oscillations to zero. In such cases,
we get finite r¡ dependent expressions. We can then use these expressions
for further calculation and take the limit r¡ - O in the final result.
Example: (i)
1/
=---
1/ + iw
Similarly
-r¡
lim eiwt = -------'-
t---++= -r¡ + iw
Thus we have a finite but r¡-dependent expression.
(ii)f.:0 eiwtdt. This function oscillates at t = +oo. Now
} _=
t" eiwtdt = ;.- ( lim eiwt - lim eiwt)
iw t-,= t-,-=
= r¡_,O
lim _.!.._ (-
iw -r¡
1/
+ iw - --11-)
r¡ + iw
211
= lim
r¡---+O r¡2 + w2
= 2·m5(w).
We now extend the above prescription to the operator (18.31a) and
write /
U(t, -oo) = lim 1/ {0 er¡t' eiHot/ñe-iHt/ñeiHt' /ñe-iHot' /ñdt' .(l::fs.34a)
r¡-+O 1-=
In particular,
U(O,-oo) = limr¡ {º er¡t'eiHt'/ñe-iHot'/ñdt'. (18.34b)
r¡---+O j _=
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
--'
{'{\
18.4
We have
Scattering States ~ The S-Matrix
Thus
.
= l 1m . 1J 1 a)
11-+a 1J + (i/ñ)(H - Ea)
.
= l 1m iñ17 1 )
a
r¡-+OEa - H + iñ17
. ie
hm E H (18.35)
=
e-+O a - + u:. la),
where e = ñ17. The meaning of Eq. (21.35) is as follows
(Ea - H + ic)U(O, -oo)la) = icla),
i.e. in the limit e ----, O,
(Ea - H)U(O, -oo)la) = O
or
HU(O, -oo)la) = EaU(O, -oo)la). (18.36)
Thus U(O, -oo)la) is an eigenstate of the total Hamiltonian H belonging to
the eigenvalue Ea which is also the eigenvalue of Ha. The above procedure
can succeed if in the limit e ----, O, Ea is an eigenstate of both Ha and H.
In other words, it will be assumed that the continuum part of the energy
spectrum of Ha and H extend from O to oo and to each state la) in the
continuum part of the Ha-spectrum which has energy Ea, there belongs a . /
corresponding state U(O, -oo) la) in the continuum part ofÍiwectrum of the .,e:--
J~éJ
}
•
H having the same energy Ea. This assumption of no 'leve! shifts' in the
continuum is not satisfied in sorne problems and then we have to drop this
restriction. However, we shall not consider such cases.
We denote the state U(O, -oo)la) by
la)in = la(+)) = U(O, -oo)la)
ic
= Ea-H+icla), (18.37)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or
(Ea - H + Íé)la+) = iéla)
= (Ea - Ha+ ié)la) - Vla) + Vla)
= (Ea - H + Íé)la) + Vla). (18.38a)
Alternatively, noting that H =Ha+ V, we have
(Ea - Ha+ ié)la+) = iéla) + V¡a+)
= (Ea - Ha+ ié)la) + Via+). (18.38b)
We can write Eqs. (18.38a ~.38b) as
1
la(+))= la)+ . Vla)L
Ea -H +ié
(18.39a) }
laC+l) = la)+ l . VlaC+l). (18.39b)
Ea -Ha +ié
The notation laC+l) is intended to emphasise that the solution of the above
equation will tend to the 'unperturbed' state la) as t ~ -oo.
In exactly the same manner, we can show that
U(O, oo)lb) = lim r¡
r¡---,a lo
rX) el- +(i/ñ.)(H-Eb)]t'
17 dt'lb)
(18.42a) }
(18.42b)
The notation I b(-)) is intended to emphasise that the solution of the above
equation will tend to an 'unperturbed' state lb) as t ~ +oo.
The Eqs. (16.39) and (16.42) are the fundamental equations of the for- J_
mal theory of scattering and are known as Lippmann Bchwinger equations. ¡J
Our purpose is to calculate the transition amplitude
(blU(t, -oo)la) for large t, viz.
lim (blU(t, -oo)la)
t->oo
= Sba· (18.43)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o
Therefore
s: = (blU(oo, -oo)la) L.
= (bC-llaC+l) = out(bla)in (18.44)
The operator S = U(oo, -oo), with matrix elements given above is called
the scattering matrix or S-matrix. Note that this operator is unitary.
Now we want to calculate (bC-) laC+)). For this purpose, we first calculate
(blaC+)). Now, on using Eq. (18.39b)
or
l . (bl = (bl l . (18.45b)
Ea - Eb + te Ea - Ha + te
Hence
1
(blaC+)) = (bla) + . (blVlaC+l). (18.45c)
Ea -Eb +zc
Now from Eq. (18.42a), we also have
therefore
1
(bC-)laC+)) = (blaC+)) + (blV . laC+)).
Eb-H +ie:
The argument is similar to that used in the derivation of Eq. (18.45b) gives
1 la(+))= 1 la(+)).
Eb - H + ie Eb - Ea + ie
Therefore
(bC-)laC+)) = (blaC+)) + l . (bJVJaC+l)
Eb - Ea+ ic
- . - 2ic ( +))
- Óba + e:-->O
hm (E b - E a )2 + e2
1
bV a
1 (
Sba = (blSla)
= (bC-)laC+))
= Óba + 21rió(Eb - Ea)Tba· (18.48)
T is called the transition matrix or T-matrix.
This gives the amplitude for
the transition from state la) to lb).
Now from Eq. (18.39b), introducing a complete set of states le), we have
1
laC+)) =la)+" lc)(cl . VlaC+))
L.,
e
Ea - Ho + te
1
= la) - L Ea -
e
E . lc)Tca
e - ic
= (la) - L Ea -
e
! . O+ Zé
le) (clTla))
= (1 - E•,.- ~
o+u:. r) la). (18.49)
follows:
na = -(bfVla+)
= - ¿ (bfVlc) (cla+)
e
Finally we want to relate lnal2 with the transition probability per unit
time from the state la) to the state lb). For this purpose we note that
the probability of transition from the state la) to the state lb) is given by
ICb(t)l2, where
Cb(t) = (blU(t, -oo)la). (18.53)
Now from Eq. (18.31a),
e-iHotfñU(t, to) = e-iHtfñeíHtofñe-iH0t0/fi,
therefore
But
Therefore
Cb(t) = ei(Eb-Ea)tfñ(bla(+))
= ei(Eb-Ea)t/ñ (ó _ ba Ea -
1
Eb +.ir:; 7]ba )
on using (18.45b) and (18.49). Hence
ei(Eb-Ea)t/ñ
Cb(t) = =Ti¿ (Ea_ Eb + ic) + Óba· (18.56)
where the limit r¡ ---+ O is taken in the end. The perturbative expression
(21.11) can be obtained from the exact expression (16.57) by replacing na (
by -Vi,ª.
For the state b-/=- a, the probability of transition from state la) to lb) is
given by [on using Eq. (21.56)]
--..:..:.
2 1 . eiwt 2 2
ICb(t)I = .,2 I lim . 1 lnal . (18.58)
n r¡->O -W + ir¡
We are interested in this probability of transition for large t. For large t
eiwt
lim --.- = 21ri8(w).
r¡->0 W - ir¡
Thus for large t, we have
1Cb(t)l2 = (21r)28(w)8(w)lnal2
8(0) =2
1 ¡t,/2 dt
71" ~t/2
1 .
= - lirn t.
27r t .....
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
w= J :t ICb(t)l2p¡(Eb)dEb
=
2;
J 8(Eb - Ea)lnal2P¡(Eb)dEb
27í 2
= ,¡-ITbal p¡(Eb)· (18.61)
If we use the first order perturbation theory, then
27í 2
W = ,¡-IViial p¡(Eb)· (18.62)
This is the famous Fermi Golden Rule.
In this case, we take the unperturbed states la) and lb) to be normalised
momentum eigenstates with initial and final momentum to be Pi = hki,
P¡ = hk¡ so that
la)= [p.), lb)= IP¡)·
We note that the number of states between k¡ and k¡ + dk¡ in volume
V is given by
V 3 V 3
(21r)3d k¡ = (21rh)3d P¡·
I July 9, 2012 14:29 World Scientífic Book - 9ín x 6in RF-QuantumMechanics I
Now
and
fv 1up(r)l2d3r = (2:)3 = l.
= IP¡l2dlP¡ldO
-1 12dlP¡ldE dO (18.63a)
- P¡ dE¡ f .
Thus
(18.63b)
and
271"
w =--¡:;
=
271"
J
--¡:;ITJil
ITJil2 d3p¡~E¡ - Ei)
2
p¡(E¡ ). (18.64)
J
Now the incident flux = PinV where Pin is the incident particle density and
v is the incident velocity. Therefore,
Flux = PinV
1
(18.65)
= (21rn)3V
is the probability that a particle is incident on a unit area perpendicular to
the beam per unit time. Let da denote the differential cross-section. Then
da(Flux)= Transition probability per unit time W.
Hence
w
da= Flux
1 271" 2
= Flux --¡:;IT¡ilp¡(E¡ ), (18.66)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
E- _ p¡ E _ PJ dE¡ IP¡I
i - 2µ' f - 2µ' dlP¡I µ
Flux 1 )3 p
= ( 21rri --¡;,·
Hence
(18.68)
We have derived
(18.69) )
(18.70)
= (bla) + Ea - ~
b + Zé. (blVla+) + (blV E b - ! .
a - Zé
laC+l)
= (bja) = Óab, (18.78a)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(/+llB) = oL (18.80a)
)
(cC-)IB) =0. (18.80b)
G±(E) - l (18.Sla)
O a-E a - u±·
no Zé
(18.Slb)
= (ria)+ JJ (rlGt(Ea)lr")
We can write
(rlGt(Ea)lr") = lim G±c(r, r")
E:---+Ü
= G±(r,r") (18.83a)
and
(r"IVlr') = 8(r' - r")V(r'), (18.83b)
so that Eq. (18.82) becomes
= - J 7/J;(r')V(r')7/Ji+l(r')d3r'. (18.85)
L
7/Ji+)(r') by 7/Ja(r'), so that
H=Ho+V(r)
A2 PA2
l ,2
)
Ha = _!!_E_ + _fJ_ = P._
2ma 2m¡3 2µ
in the c.m. system and µ is the reduced mass. Hence la) and lb) are
normalised momentum eigenstates with initial and final momenta
so that
(
r 1 a) -
_ ( 1 ) _
r Pi - (2n1¡,)3/2e
1 ipi·r/ñ l (18.87a) )
( r l b) -- ( r 1 P¡ ) -- (27rn)3/2
1 ip¡-r/ñ (18.87b)
e ,
( p IP·)
J •
= _l_
(27rn)3
J ei(pi-P¡)·r/ñ
= J d3qeiq·(r-r )/fi 1
Ep -Eq ±ic
1 .. (18.88)
µi=µ¡=µ, E, = E¡ = E = Ep,
q2
Eq=-
2µ
and Eq. (18.88) gives
<S--
·
lim G± (r, r')
"__,º
= (rlGt (E) Ir') = ___'!:_!!__
(27rn)3
j eiq·(r-r')/ñ
(18.90b)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
This shows that the solutions 7/J~+) ( r) of the above integral equation asymp-
\ 1 totically represent outgoing spherical waves in addition to incident plane
(1'\'v-' waves. Hence laC+l) is usually called the outgoing eigenstate of H, if la)
. _ \ O~ is a momCR}um eigenvector IP) while laC-l) is called the incoming eigen-
~ \ state of H.~ansition Amplitude in Born approximation: We have derived
~ Eq. (18.86b), which on using Eqs. (18.87) gives
T/;;, =-Vi,ª= -(P¡IVIPi),
i.e.
T~Pi = -VPJPi = - I (27rln)ªe-iprr'/nv(r')
xeip,·r' fnd3r'
(18.91a)
where
(18.91b)
with
~ = Pi - P ¡ = momentum transfer.
Now the scattering cross-section is given by
da = W/(Flux)¡n, (18.92a)
where the golden rule gives
21r 2
w= --,¡:IVp¡p,Ip¡(E¡ ), (18.92b)
= µ¡IP¡ldO. (18.93a)
Now
1 IPil
(Flux)in = PinV = (21rñ)3 µi . (18.93b)
ip~·r'/ñ
L
- <Pn r¡3 (21rfi)3/2 e (18.96c)
)
and
(18.97)
(r~r~IV!r~r~) = o(r~ - r~)t4·~ - r~)V(r'),
with r' = r~ - r~. Then /
T{/1, = --1
(21rn)3
-f J e-ip~-r~fñ<j)* (r~)V(r')</>o(r~)
n
x eiP"' ·r~ fñd3r~d3r~. (18.98a)
We change the variable of integration r~ = r' + r~ and treat r' and r~ as
independent variables. Thus we can write:
T'iB = __
ba
l_{j eiq·r' /ñV(r')d3r'.
(21rfi)3
J eiq·rn,1..* (r')
'1-'n f3
X ,i.. (r' )d3rl}
'1-'0 f3 f3 ,
(18.98b)
where q = p0 - p~ is the momentum transfered to particle o:. Now define
V(q) = J eiq·rfñV(r)d3r L (18.99a) )
Fn(q) = J eiq·rfñ<j)~(r)</;o(r)d3r. (18.99b)
Equation (18.98b) becomes
B 1 -
Tba = - (21rfi)3 V(q)Fn(q). (18.100)
We have the following picture that the particle a is scattered, with ini-
tial momentum p0 and final momentum p~. Therefore if we compare
Eq. (18.100) with Eq. (18.91a), we get from an analogue of our formula
(18.94b), i.e. with µi = µ¡ = mo, Pi= P0, P¡ = P~,
= ¿(blSlc)(alSlc)*
e
= ¿(bC-l¡cC+l)(cC+l¡aC-l) + ¿(bC-llB)(BlaC-l),
e B
(18.108)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
since the bound states IB) are orthogonal to lbC-l) and laC-l).
But
(18.109)
e B
(18.llüa)
or
ss' = 1. (18.llüb)
We now write the same relation in terms of the transition matrix T. We
denote the initial and final states by li) and lf) instead of la) and lb).
Now
T¡i = (JITli)L
S¡i = ó¡i + 21ri8(E¡ - Ei)TJi·
Thus
gives
(18.111)
e
Now
(clStli) = (ilSlc)*
= Óci - 21ri8(Ec - Ei)(ilTlc)*
= Óci - 21ri8(Ec - Ei)Ti:, (18.112)
where
Ti:= (ilTlc)* = (clTtli) = T}i. (18.113)
From Eqs. (18.111) and (18.112), we have
¿ (ó¡c + 21ri8(E¡ - Ec)T¡c) (óci - 21rió(Ec - Ei)Ti:) = Ófi
e
or
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore
io(E¡ - Ei) (Tt¡ -T¡i) = (21r)o(E¡- Ei) ¿Ti~T¡co(E¡ - Ec)·
e
(18.114)
If E;-=/=- E¡, this is a trivial identity. However, if E¡ = E¡, we can cancel
o(E¡ - Ei) on both sides and get the relation
i(Tif - T¡i) = 21r ¿ Ti~T¡co(E¡ - Ec)·
e
(18.115a)
If we had started with
Let us now take the index i to refer to momentum Pi and f to the mo-
mentum p t : There may be other quantum numbers necessary to spec-
ify the states li) and lf) but we')mrld\ not write them explicitly. For dó
Pi = P¡, we have forward scattering (O = O, or momentum transfer
D.2 = (pi - P¡ )2 = p; + p}- 2IPillP¡I cos O= 2p2(1- cosO) = O). Thus we
have
(18.119)
where O'total denotes the total cross-section of the process starting from the
initial state i, since we have not only summed over angles but all possible
states consistent with the energy conservation. Thus Eq. (18.118).tgives
Im Tii
p
= 2r,,2 (2rr )3 µ O-total J
or
k
Imfk(O) = (18.122)
47fO'total·
This is the generalised form of the optical theorem. It should be emphasised
that on the right-hand side of Eq, (18.122), we have O'total rather than ael
because of the summation over e which involves all possible channels, both
elastic and otherwise.
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
18.10 Problems
L'im'
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and note
j1(qr) rv (qr)1 as q -4 O.
In calculating the partial wave cross-section, remember that we do
not distinguish final states which only differ in magnetic quantum
number m so that the observed cross-section involves the sum
L IYim(éi)l2 = 2l4: 1 _
m
Va « pfl, when
µa
pa ~ 1
ñ
and
fí,2 pa
Vo « -2, when ~«l.
µa n
Show further that the first condition is satisfied if
Voa2
µ ----,¡:¡--- « 1
which is also the second condition. Thus B.A. is justified at all
energies if Vo « ~.
Hint: In the first approximation
1/J+(r) = 1 eip·r/lí
p (27ífl)3/2
In the second approximation
'l/l+(r)
P
= 1
(27ífl)312
1 2
{eip·r/lí - _ __.!!_
47í fí2
¡ Ir - r'I
e
iilr-r'I
V(r')eip·r' /ñd3r'}.
ii
where
n
()( ) - _1 Joo eiwt dw
t - 21ri -oo w-ie ·
Chapter 19
19.1 Introduction
Equntion (19.3) also follows from the result [H, U] ~ q. (see Sec, 10.5)~ }- ( }
noting that the transition rnatrix T can be formally written as
1
T = -V - V V (19.6)
Eo-H+is
where
V=H-HoL (19. 7) C)
We consider two cases:
(a) U continuous: U can be built out of infinitesimal transformations. We
consider an infinitesimal unitary transformation
u= l -iEF, (19.8)
where F is hermitian. F can often be identified with an observable of
the system. For example, corresponding to an infinitesimal rotation of
coordinates, the unitary operator U is given by
i
U= 1- -€ · J (19.9)
n
where J is the angular momentum operator. From Eqs. (19.3) and (19.8),
it follows that
[S,F] =O. (19.10)
In case F is one of the observables of the system, we have
i.e. li) and lf) are eigenstates of F, with eigenvalues Ff and F¡.
The transition from the state Ji) to state Jf) is given by the S-matrix
(JJSJi) =f. O. Invariance under transformations (19.4) and (19.5) imply
(JI [S, fr] Ji) = o. (19.13)
Hence
(FI - F¡)(JISli) = o (19.14)
or
Ff = F¡, (19.15)
o
the system is invariant under U
[S, U] = O l (19.18)
Then Uf= U'¡ in any transition from li) to lf), provided that (JISli) -=fa O.
From symmetry or invariance principles, (i) we can derive selection rules
viz. which processes are allowed and which are forbidden. (ii) We can relate
different processes
In a scattering process, we specify the initial and final states by the initial
and final momenta Pi and p ¡ in the centre of mass frame. If rotational
invariance holds (as for example, is the case for central forces V = V(r)),
then
(P¡ISlpJ = (P¡IUkSURIPi)
= (p¡UklSIURPi). (19.20)
Thus the scattering matrix can depend only on those quantities which are
invariant under rotations, i.e. on the scalars
2 2
Pi
Ei = -2µ = -P¡ = E¡ = E (19.22)
2µ
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(19.32)
or
(19.33)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
Now
(19.42)
and we can write
\o(pi - p¡)
P
r:
lo 1-1
{+i 8(cosB)8(</>)P1,(cosB)d(cosB)d</>
Therefore
Hence
(19.46) o
8(pi - P¡) =
8(pi - p¡) ~ 2l +1
Z:: -4-Pi(cosB). (19.47)
2
p l=O tt
8(pi - p¡)-
p2 '°'
¿(2l + l)IF1(P)l2P1(cosB) = 8 (p · - p ¡) ~ 2l+ 1
' 2 ¿-4-Pi(cosB)
4rr l p l=O 7f'
(19.48)
or
1
IF1(P)I 2 = p4 L (19.49) o
Hence
F¡(p) = \e2ió1(Pl, (19.50)
p
where ó¡ (p) are real functions of momentum or energy.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or
M=O. (19.85)
u----;-u ~
[qi,Pj] = iMij
------; -iMij· (19.89)
(ii) If more than one operator is involved, then order of factors is inverted
and ket vectors go into bra vectors. Thus
rrÁrr-1 = Át L (19.90) )
rrfm-1 = f3t l (19.91) )
IIÁÍJII-1 = i3t AL~ (19.92) )
IIl1/J) = (1/Jtl I> (19.93) )
Al1/J) = iJCl<fJ)Ct nt (19.94)
==> (1/JtlAt = (</Jt¡ctf3t. (19.95)
The transformation (i) or (ii) is called antiunitary.
Using alternative (i)
(19.96)
hence the commutation relation remains invariant. Similarly using alterna-
tive (ii)
II[qi,Pj]II-1 = IIqipjII-1 - IIpjq;II-1
= (-pj)Qí -q;(-pj)
= [qi,Pj]·
(19.97)
We also note that
IIJrr-1 = -J (19.98)
and the commutation relation
(19.99)
is preserved either in (i) or (ii).
Let us now discuss how the transition matrix T transforms under II. If
H0 and V are invariant under time reversal, then
IIHII-1 = HL, )
rrvrr-1 = v.
Now
1
T=-V-V V (19.100)
Ea -H +ic
and we have
1
IITii-1 = -V - V V
Ea -H -ic
=Tt, (19.101)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= WITIP). (19.103)
In deriving Eq. (19.105), we have used definition (i). The same result also
follows from definition (ii)
UITli) = u1rr-1rrrrr-1rr1i)
= (itlTIP). (19.104)
Hence we see that the two definitions are completely equivalent.
We now consider the scattering states under II. We specify a state la)
by its momentum Pa, z component of spin ma anda, which denotes other
quantum numbers which may be necessary to specify the state
o
where
(19.105)
m
Iljo; Pa, ma)out = la, -Pa, -ma) + E
a -
H
-
. Vio:,
Zé
-Pa, ma)
(19.108)
1
I V
IIla, Pa, ma)out = (a, -pa,-mal + (o, -Pa, -malVE H .
a +ic
(19.109)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 20
RELATIVISTIC QUANTUM
MECHANICS: IRAC EQUATION
From the solution of this equation come details about the hy-
drogen atom, the spin of the electron, and the existence of anti-
matter. Poets bring us fresh insights with the right sequence of
words, Dirac brought us fresh insights with the right sequence
of~.
(i) Laws of physics take the same form in all inertial frames (Principle
of relativistic invariace).
(ii) In any given inertial frame, the speed of light e is~e whether
light is emitted by a source at rest or in uniform motion (light
principle).
((
From these postulates, it followsthat the ength" (c2t2 - x2 -y2 - z2)
is invariant.
In relativity, space and time are treated on equal footing and as such
we deal with four vectors in four [1+3] dimensional space:
( contravariant vector)
f
1John
Pt·
S Rigden, Hydrogen: Trie Essential Element (Cambridge University Press) Cam-
bridge, Massachusetts, 2005,
451
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where µ = O, 1, 2, 3; µ = O, i = 1, 2, 3, and
Xµ= (xa, Xi)= (et, -x), ( covariant vector)
so that
x2 = xµxµ = xºx0 + xixi = c2t2 - (x2 + y2 + z2).
Introduce metric tensors gµ", 9µv to raise or lower the indices:
(20.1)
where
~ ~1 ~ ~ ) (20.2)
=9µv
( O O -1 O
O O O -1
- .
detg = (detAT)det(gA)
= (detA)2detg
1 = (A~)2 - (At)2
i.e. (A~)2 :::=: 1, which implies A~ :::=: 1 or A~ < -1. If A~ :::=: 1, the time
direction is unaltered and the LT is called orthochronus, written as ts.
We shall confine ourselves to Lt,
which excludes space reflection and time
reversal.
Lorentz group is a six paramete~ group; three parameters correspond
to rotation in three dimensional space(x, y, z) and the remaining three pa-
rameters correspond to Lorentz velocity transformation (Lorentz boost).
In order to see this, first consider the transformation
x'º = xºL )
x'i = Ajxi. (20.9)
Consider for simplicity rotation around z-axis by an angle w:
A=
O1 cosw
O O O
sinw º) (20.10)
( O - sin w cos w O
O O O 1
The matrix
giving
;,
)
z
c. (20.11)
J
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
O O
~~el
10
= (A(v)
O
º)
l
( 0.l
2 2)
O O 01
giving
t'=, (t - *x)
x' = 1(x - vt)
y' =y
z' = z (20.13)
where
,= J1-~· 1
-===
- This is the Lorentz transformation, when the primed frame is moving with
uniform velocity v along x~xis with respect to unprimed system.
For infinitesimal Lorentz transformation:
l 7
(20.14)
Then Eq.(20.6) gives
Thus
(20.15)
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(20.18)
(20.19)
V
o
E= ffi"(C2, p = ffi"(V = E-c2 (20.20)
or
c2p
v = E" (20.21)
Thus in analogy with space-time transformation (20.13),
E' = "f[E - v.p]
p = p + [( "( - 1 ) V- v.p - "(V- E]
,
v2 c2
where the primed quantities are in the frame moving with velocity v relative
to one in which E and p are measured. It is convenient to write the vector
p in terms of its longitudinal and transverse components
P = (Pn,P1_). (20.22)
Under pure Lorentz transfo~mation (Lo~entz \)oost)
Pn - 'Y [ Pn - c2 E] /,, )
p'i_ = P1_ L )
Finally
E' = 'Y [E - v.pn] L-- (20.23)
o
L )
L- (20.24)
)
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
After the discovery of quantum mechanics in 1926ili7 the big problem was
that the Schréidinger equation f
in 8<P = _.!!.__ v2<P (20.25)
8t 2m
was not consistent with the special theory of relativity as time and space
coordinates are not treated on equal footing.
Now for a single free particle in non-relativistic classical mechanics
p2
E= - (20.26)
2m
To go over to quantum mechanics replace the dynamical variables E and p
by differential operators
(20.28)
(20.31)
bove equation is known as the Klein Gordon equation for a free particle
o( rnass m. In this equation, space and time coordinates appear on the
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
l
giving
Thus for E > O, p is positive but for E < O it is negative and thus cannot
be interpreted as probab!1!ty. }Ilewe~~e may arbitrarily omit all solu-f ~
tions of the form rp(x)e-,et//1 for E < O ;;hu this is not allowed since the )
solutions witb E > O do not forrn a complete set.
(20.39)
The most general equation one can write satlsfying conditions (1) and
(2t· put e= 1 and /i = 1] is
o
operate by
°
[ 8mn fJt2 1 [(a í) mR. ( a i) R.n
2 - 2 + Z· +-+ J·J
O O
. (( a í) mR. f3ln 8xi
axi axí - zm a + /3mR. ( a i) R.n Bxia )
+m2/3mR.f3R.n] 'l/Jn = O (20.41)
where the term involving aiaí has been symmetrized which is permissible
since / i and {j~J commute. The above equation reduces to the Klein~
-4 Gordo1fke equation,
/\
(
t)2
Bt 2 -
2
v' - m
2) v,« = O, (20.42)
provided that
~ [(ai)mR.(a1)R.n + (a1)mR.(ai)R.n] )
(ai)mt(f3)R.n + (f3)mt(ai)er )
f3mtf3.P.n Óm.n
It is clear that a's and (3's are not algebraic numbers. They are N x N
matrices with the properties
L- (20.43)
o
)
(20.44)
Thus the matrices a1, a2, a3, /3 anticommute in pai0d their square is ~
unity.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Thus the Eq.(20.40) takes the matrix form [inserting back ñ ande]
(20.46)
H = a.p+(3m, (20.47)
(20.49)
so that
But
(20.50) o
Tr((3a'(3) = Tr((3(3ai) = Tr(ai). (20.51)
a'=
. ( o (J"i)
ui O , with I = G ~) (20.52)
(20.53)
where each of 'lj;A and 'lj;s has two components: when written in full
f3= (
10
O 1 O
00-1 0
O O O -1
Oº)O
' 01 =Qx= (fü~) L )
1000
ª
2
= ay =
o0 o0 oÍ
O
O -i O O
-i) ' a3 = O:z = (~ l O ~ ~1)
O O O . (20.54)
(
i o o o O -1 O O
,82 = 1,
or
(20.56)
where
,B = 't", ,Bo: = ,; )
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
)
(20.57)
Hence, the Dirac equation in covariant form is given in Eq. 20.56), which
-Ir
can also be written as
We note that
(·/)t = (,Bo/)t = ci,B = -,Bo/ =
Pµ = -iño,,L
-'i
(20.58)
o
i.e. ,yi is antihermitian,
('i)2 = ,Bai,Bai = -,B2(ai)2 = -1, i = 1, 2, 3.
Noting that gºº = 1, gii = -8iJ, the anticommutation relations (20.44) can
be written in the compact form
{,/1.,1"}= (,µ,v +,"1/1.) = 2g/1.". (20.59)
Thus there are four independent -v-matrices ,º, ,1, 12, 13• The fifth one is
defined as
L
,5 = Í'Yo,n2,3 L
,! = ,5 i-
(,5)2 = ,g = l. (20.60)
It is convenient to define
[,µ, ,"] = ,,,," - 'Y",,, = -2ia-/1."
or
(20.61)
From the properties of ,-matrices, it is easy to see that there are sixteen
independent -y-matriceslisted in able l. I'he followingidentities are useful
1µ1v 1>. = s" 1>. _ gµ>.1v + g">.'Yµ + ÍEµv>.p15TP L )
['Yµ, a-">.] = 2i(gµv "(>. - gµ>.1") L (20.62)
)
{'Yµ,O"v>.} = -2Eµv>.pf5TPL
)
i µv>.p
5 _ _
'Y --4,E "/µrv"f>.rp-¡5,
where we see the convention
i ·1c
E0123 =1= -E0123, E J = -Eijlc·
L )
Define
'Yº = (º 1) ''Y'. = ( º
1O
o" =
-(J'i
(1,0"i)
(J'O i) ''Y
= (1,u)
5
L
= (-1 º)L
O 1 (20.64)
)
o
O'µ.= (1, (J'i) = (1, -u)L )
so that
(20.66)
where each of </JA and </J B has two components. In the chiral representation
where
(20.67)
where each of </J L and </J R has two components. Thus (1 :::¡:: 'Y5)'lj; respectively
pick <PL and <PR· These are known as chiral projections of 'lj;.
where
Lµ11 = i(xµav - x11 8µ)
are the generators of the Lorentz group as previously noted. The most
general representation of the generators of the Lorentz group are given by
(20.73)
where the hermitian 8µ11 satisfy the same commutation relations as in Eq.
(20.17) and commute with Lµ11. Thus
[Mµ11, MP<YJ = -i[gµp M11u - g11u M''" + µ - v] (20.74)
which form the Lie algebra of the Lorentz group. For a spinor 'lj;,
1 i
3µ11 = 2uµv = :;¡h'µ, 1'11]
where on exponentiation
U(A(E)) = e-!€µ.,Mµ". (20.76)
For the rotation subgroup of Lorentz group
E µ11 ----, f·.
2J l e µ11v~µ11 ~
~ E·2J.~ij
V l
f
µ11 Lµ11 ~
~ f··Lij
2J l
E12 = E,
~12
v
= º'"V1'"V2,
• , ,
L12 = i· ( x 1~8x2 _
x 2~)
8xl = -~L3
li '
M12 = L12
8 3 = __iñ ,, 1 í' 2
(20.78)
ó
2 Tí
and
(20.79)
l July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(20.80)
so that
(20.81)
so that
(20.82)
Also
.. 2
EijV1 = -¡¡:w.L. (20.83)
'l/;'(x)= [1+~w.(L+~u)]'l/;(x)
[1 + ~w.J] 'l/;(x)L
The angular momentum J is a vector sum of the orbital angular momentum
L and S = ~a, where S is spin angular momentum, with eigenvalues S =
(20.84) o
!,83 = ±!.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
UR = ef.w.J.
Here UR is a unitary operator, which rotates the co-ordinate1.system by an
angle w. f
t
The usual operator
We have seen that under Lorentz transformation the Dirac spinor satisfies
the transformation law
'lf;(x) --t '1/J'(x') = S'lf;(x) (20.85)
where for infinitesimal transformation
S -- 1 - f'µvO"
i /J,V -
- 1 + 41 e:µ,v'Y µ, ')' V
(20.86)
so that
(20.87)
Then
(20.88)
Thus
if;(x)'lf;(x) --t if;'(x')'I/J'(x') = if;(x)S-1S'lf;(x) = if;(x)'lf;(x) (20.89)
i.e. 1/J(x)'lf;(x) is a scalar.
We want to find the transformation properties of the bilinears if;r'lf;,
where r is any of the independent products of 'Yµ, given in Table l. Thus
(20.90)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
(i)
- i
- - 4! Eµv>.p Aµo "f °'Av{3"f f3A>..u "f uAPó"fs
- i
- - 4! Eµv>.p Aµ.AvA>..AP °' f3 u ó
o f3 u ó"f "f "f "f
L CJ
But
Therefore
L o
s-1"158 =~A) (-i) Eaf3uó'Yª'l'Yu'Yó
=~Ah5l_
Thus for proper Lorentz transformation L+
(20.91)
G
s-1"f5S = "fs. (20.92)
(iii)
@A)A~1µi l (20.93)
G
(iv)
r = (Tµ.v
Now
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
ih5'1/J:
i/J1/J,'I/J:
i/J1/J,,,/''if;:
nsor
(Pseudoscalar)
where
since
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Further
if;(x)''t15'1/J(x)----> i[;P(xPhµ15'1/JP(xP)
=if;(x)1º1µ151º'l/J(x)
= -(-ltif;1µ15'1j;(x): (Axial Vector)
using
1°1µ151° = -1º1µ1°15.
Both if;1µ'1j; and if;1µ15'1j; transform as vectors under proper Lorentz
transformation L+, but transform differently under space reflec-
tion. Now
if;(x)crµv'l/J(x)----> i[;P(xp)crµv'lj;P(xp)
=if;(x)1°crµv1º'l/J(x)
= if;(x)crµv'lj;(x) (Tensor).
This follows from the fact that
i
1°CTµv = -1ºbµ1v _ 1v1µho
2
i
= -bº1µ1º1º1v - (µ +--+ v)]
2
= (-1)µ(-ltcrµv = ol'",
(2) Time Reversal:
X ----t XT = (-xº,x),
i.e.
A''=
V
( -1ººº)
O 100
O O 1 0
o ool
L (20.104)
..
Under time reversal, suppose
(20.105)
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
and all the complex numbers go to their complex conjugates [see ¡hap. 10].
Invariance under time reversa! implies
(in''f"'a; - m)'l/Jr(xr) = 0. (20.106)
Taking the complex conjugate of free particle Dirac equatioe.56),
(20.107)
However
giving
S r'Y 's::
r = -_,,i ,
= ,,..
f. .
1,
= 2
= i 3
Now
)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
or
Thus
D
(20.110)
On the other hand from Eq (20.105)
. ['l/JT(xT)]* = TJrB'l/J(x) .
Hence from Eq (20.110) invariance of probability density 'l/Jt'l/J giveL s
ITJTl2 = 1
Thus, using Eqs. (20.105), (20.109) and (20.110), for bilinears:
L o
if;(x)'l/J(x) --4 'l/Jf(xT)'l/Jr(xT) = if;(x)B-1 B'l/J(x)
= if;(x)'l/J(x). (Scalar)
@ x'1•~-
Ae 8t + Ee (20.111)
!E
2 a(Aª,6)µ 1/
C,µ
= Eµ
1/
= gµ>..E>..v
1 >..
= (E>..v - Ev>..)
2gµ
-- 1 g µ>.. [ E,,,euJ.C, J.,6 - Ec,,BUv
x uv J.C, J..6]
"»
2
= 2 !E a [gµ>..5a5.B
C,µ x 1/ - gµ>..5a5.B]
1/ x
so that
(20.113)
Let us write
(20.114)
are given by
(20.115a)
)
= (HE)
ooi o
:St =S!l) L (20.115b) )
= (n H)
o -i o o
= s2 = s; (20.115c)
Thus S = (Sx, Sy, Bz) gives the matrix representation of rotation subgroup
of Lorentz group.
Now
(Mº1)t = i(gµ08~ - gµ18~)
=-(~~~~) =
i 0000
0000
-"'
1
1l,,- (20.116a) )
(Mº2)i = i(gµ08~ - gµ28~)
(20.116b)
)
(20.116c)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
r;, = (r;,x, K,y, K,z) give the matrix representation of Lorentz boost.
The matrices Si, Sj, K,i and K,i satisfy the followingcommutation relations
L
iEijksk
We note that 4 x 4 matrices S's are ~ermitian but matrices r;,'s are anti-
hermitian. Thus finite representation of Lorentz group is not unitary.
so that
(20.118)
From Eq. (20.74), we can easily write down the commutation relations for
J's
(20.119)
Thus the generators Ji satisfy the commutation relations of the rotation
group 0(3). Hence out of 6 generators of Lorentz group, three generators
f
Mii belong to the subgroup 0(3). The other three generators Mºi give
the Lorentz boosts. Define
(20.120)
Then from Eq.(20.74), we get the commutation relations for K's:
[Ki, Jl] = iélm Km l--
[Ki, Ki] = [Mºi, MºJJ = -igºº Mij = -iEijk Jk. (20.121)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Note that the minus sign in Eq. (20.121) is the manifestation of the non-
compactness of the Lorentz group. Note also that K's are anti iermitian. /\.
The negativo sign originates Erom the Minkowski matrix, g°º = 1, _gij =
-Óij. As a consequence the irreducible representations are radically difficult
in nature from those of the rotation group. Further
(20.122)
Writing
' i ik: k
EQj = T/j = -r¡1, Eij = E J W (20.123)
so that U(A) given in Eq. (20.76) can be written
U(A) = éw.J+ir¡.K
= UA(R)UL(v) (20.124)
where
w = On
V
1J = r¡-. (20.125)
V
explicitly showing that Lorentz group is a direct product of b,tatfon group
in three dimensions and Lorentz boosts. It should be noted that UL (v) is
not unitary unlike unitary operator eiw.J : UL(v)Ul(v) -=!= 1 but is in fact
hermitian Ul(v) = ei.,.K = UL(v) L
The mathematical designation for proper Homogeneous Lorentz group
Lt is 80(3, 1) which refers to the fact that signature of the metric gµv is
1, -1, -1, -1.
For the classification of irreducible representation of L¡,
it is useful to
introduce hermitian combinations
Mi= !(Ji+ iKi) (20.126)
2
u,v Representation
0,0 Lorentz scalar
(1/2,0) 2-component Lorentz spinor of first kind
(0,1/2) 2-component Lorentz spinor of second kind
(1/2,1/2) Lorentz four vector
This algebra is identical to the Líe algebra of group SU(2)M ®SU(2)N with
two Casimir operators
)
N2 =Ni Ni l- )
[M', M;] - [N' N;] - O L (20.128) C)
Since in Eq. (20.126) M' and N' are not linear combinations of basic
elements, the commutators in Eq. (20.127) do not define the (real) Líe
algebra of L¡.
However, due to analogy with angular momentum, it is
convenient to use the eigenvalues of M2, M3, N2, N3 which are respectively
u(u + 1),v(v + 1),u,v = O, 1/2,3/2, ... to label the elements of irreducible
representations. This provides the basis lkl), where
lkl) = {lu,k)lv,l)}
Ji - 1 i
-2ª' Ki -
--20'i il )
i
Mi=~ ,Ni= O. (20.130)
)
(20.133)
~
(!:) ------+
boost
(
e~~-2a
V
e-11~-~
o ) (!:) (20.134)
ú
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
o
since
8µ,'l/J(x)--+ :~eiei\/cn.'l/J(x)8µ,A + eieA/cñ8µ,'l/J(x)J.4
To compensate the first term on the right hand side, we must introduce a
vector gauge field, in this case electromagnetic potential Aµ ( x), and replace
8µ,'l/1 by
or
(20.146)
i.e. the current jll = eijj11l?jJ is conserved. Further
i'" = e?jJt('Yµ)t1º1/J = e?/Jt1º1ºbµ)t1º1/J = eijj'Yµ?/J = i" (20.147)
i.e. jll is hermitian and jº is given by
jO = eijj1° 1/J = e?/J t 1/J = ep (20.148)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
~ = co:.f> + (3mc2
r<c:»
i>----. -iliV. (20.150)
The Heisenberg equation of motion for Pi is given by
iñ dti = Wi, HJ
= Wi, COjPJ + (3mc]
=0
since [pi, Pí J = O and ªi and f3 commute with Pi. Thus for a free particle,
momentum is a constant of motion and we can find simultaneous eigenval-
1
ues of both energy and momentum. For a spin particle in a centrally
2
symmetric field, the Hamiltonian is
H = ccc.p + (3mc2 + V(r). (20.151)
In a central field, the angular momentum should be conserved, i.e.
dJ = o.
dt
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
L;<}(~::~
The Heisenberg equation for angular momentum L is given by
)
dL·
iñ d/ = [Li,H]
= [Li, COfPi + (Jmc2 + V(r)]
= cai[Li,Pi], (20.152)
since L¿ commute with (Jmc2,a-i and V(r). But
[Li,Pj] = Eitm[XtPm,Pj]
= Eiim{[xi,Pj]Pm + xi[pm,Pjl}
= ÍnEilmÓljPm
= inEijmPm·
Therefore
-~dLi "L A
indt = i,icEijmÜ'.jPm
or
dL
dt
l (Ji)
A
= ca xp, (20.153)
iti d;
da,
= [rri, H]
= h, caiPi + (Jmc2 + V(r)]
= CJJj[rri,aj] +mc2[rri,f3].
We can easily calculate these commutators by using the Pauli representation
of a and (J. Since (3 is a diagonal matrix
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore
Hence
,~dui 2'
dt = 2CE·iJ"kP J·ak
A
in-
or
ñ dc A
Thus
(20.156)
We see that the velocity operator Xi = có(t) is not a constant of motion in
spite of the fact that the particle is free.
Noting that H and Pi are constants of motion, we can find the solution of
the differential Eq. (20.156) as follows. First we note that
iooi(t) = 2ói(t)H.
Therefore
ói(t) = cii(O)e-2iHt/1t. (20.157)
Substituting Eq. (20.157) in Eq. (20.156), we have
inai(O)e-2iHt/li = 2ai(t)H - 2cfJ;/\ )
ai(t) = cpiH-1 + in2 ái(O)e-2iHt/ñ H-1 .
In particular, for t= O
i; ói(O)H-1 = (ai(O) - cpiH-1). (20.158)
Thus the solution of Eq. (20.157) is
ai(t) = cpiH-1 + (ai(O) -cfiiH-1)e-2iHt/1t.
Therefore
Xi = cai(t)
= c2piH-1 + c(ai(O) - cpiH-1)e-2iHt/1t. (20.159)
ThAus, ~he ith comp~nent of velocit)í,j consists of t,~o pw·~s, a constant_ part ~'\ L ~
The first and second terms in Eq. (20.160) are understandable since their
expectation values give the trajectory of the wave packet according to the
classical law:
x;1assical(t) = X~lassical(O) + t(c2pd E)classical.
The presence of the third term in Eq. (20.160), which is a consequence
of the second term in Eq. (20.159) appears to imply that the free electron
executes very rapid oscillation in addition to the uniform rectilinear motion.
= O l (20.165) o
ie
[Dµ,Dv] = cñFµ,v, (20.166)
given in Eq. (20.140), the above equation becomes
eñ
ñ2 Dµ, Dµ + -aµ" Fµv + m2c2)1/J = O. (20.167)
2c
This differers from the Kleir~ordon equation (20.162) in the term
e µ,v
2a Fµv· Now
ü
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
(20.168)
where
(20.169)
(20.170)
and
a
i = 'Yo 'Y = (-(J"iO O"io)
i
(20.171)
1
where '1/JL,R = ~ 'Y51/J, the equation (20.168) splits into two equations
which are completely decoupled.
ieñ en
(n2 Dµ DµI -F -O".E - -O".B + m2c2 I)</JL R = O (20.173)
e e '
\ where I is a 2 x 2 unit matrix, which we will not write explicitly in what
11 5
follows. 'I he above equation
tert.E and u.B.
diffei¡-
frorn l<leinvl]ordon equation in the
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
In order to interpret these terms and see their physical significance we first
consider only the radiation field:
(20.175)
Thus
imc2 O ) .!,
-{) (e -imc2t/li.¡,)..,,, _- e -imc2t/li
n + -m . , ,.
( ---
m
Using it again
2
~ [e-imc2 tn (-imc2 /n + ~) 'l/J] = e-imc2 rn (-imc2 + ~) 'lp
m & n m
-imc2 li ( [)2 . c2 {) mc2)
=e / [)t2 - 2imh ot - ,i;l 'l/J.
2
in81 = -- 1t2 ( V--Aie ) en
--a.B ) 'lj;. (20.177)
at ( 2m ne 2mc
Now
But
1
using V.A= O. Further taking B to be uniform, and noting A= xr
2B
satisfies B = V x A, we get A.V= where L = -in(r x V),1is the/
2inB.L,
orbital angular momentum. Thus finally
iné)'lj; =
8t
[-n 2 V2
2m
- _e_(L
2mc
+ 2S).B + __!j:___A2] 'lj;
2mc2
(20.178)
where S = !!:<,
denotes spin ~ operator. Now _e_L is the magnetic mo-
2 2 2mc
ment dueto the orbital motion of a particle of charge e. Note the important
factor of 2 in front of S, known as magnetic spin anomaly which naturally
comes from Dirac equation. The second term gives the intrinsic magnetic
1
moment of a particle of charge e with spin i.e. the magnetic moment of
e
2
electron is µe = 1 or g = 2, in terms of Bohr's magnetron -- in agree-
2mc
ment with experiment. This naturally comes out from Dirac equation while
earlier it was calculated experimentally. This was regarded as~eat success 0-.
of Dirac equation.
Dµ. D
µ.
= Dº2 - 02 = _!__
c2
(!!_8t + i:.A
n 0)2 - V2. (20.179)
v2--+----
82 2 a L2
(20.183)
- 8r2 r ar nr2 '
and L2 is the square of orbital angular momentum L. The total angular
momentum
1
J = L + S = L + =a:
2
commutes with L2 as both L and u commutes with L2. It also commute
• h O".r •
wit tr.t: = -, smce
A
r
[Ji, er.r] = [Ji, akxk]
= crk[Li,xk] + ~xk[cri,crk]
= crkü.ikj xi + ixk iki cri
=0.
Thus J commutes with the operator acting on <!>L,R· J2 and Jz have eigen-
values j(j + 1) and mñ respectively, where
. 1 2
m = -j, ... ,+j; (20.184)
J = 2'3'""')
and can be used to label the states. Now L2 has eigenvalues f(f + l)n2 and
the integer f takes the values:
e= J. + 21 = f+,
. 1
f=J-2=L,
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
tr.t: is hermitian and has its square (rr.f)2 = 1 so that it has eigenvalues ±1.
In the two dimensional sub-space provided by e±, (rr.f) has non-diagonal
matrix elements (it is like dipole operator for which the selection rule /j.f, i=- O
holds), so that 1,
(20.185)
* }(~n Eq~0.182~
ñ,2 ,,
ñ
2 ( L2
¡¡,2r2 -
± ~ie
-)
dr a.r
!!&
1L1-
,
)
(20.186)
(20.187)
( ± ie ~ (j-!)~i+4J _ e2 A2
lic dr r li-~c'i O
l l
0
- [ r2 ñc2 r4 ñ2c2 dr '
= r\ [
[ (j + ~) 2
- Za2
l (j 1/2 [ [
+ ~) 2 _ za2 1/2
±1 , (20.189)
o - - e2 A5 -
where we have used (A - z ~)
r
, i.2 2 -
e
z2a2
r
2
ñ~:
tb
2
and 2 ( dtº) = z::4; where a = ~:. Thus the eigenvalues of the
operator
ñ2 ( -L2 -
Z2a2 ± i-rr.f
--
Za )
(20.190)
r2 r2 r2
are
:X(:X + l)ñ,2
(20.191)
r2
1see
C.Itzykson and J.Zuber, Quantum Field Theory, McGraw Hill,~O~e, 2.3 -#7
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where
with
(20.193)
[
n2 (~ + ~~) + 2 z ne2 E+ (E2 - m2c2) - >.(>. + l)n2] R(r) = O.
8r2 r ar c2 r r2
(20.194)
Writing
if
E
e2-----, --e2
mc2 l )
E
E2 -m2c4
2rrw2 L )
f. -1 5. = (.1 + ~) - Ój
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
an¡i
6
'IÍ =
= ti - (j + ~) + ºj - 1H
Thus n must be shifted by Ój.
n-+n=n-ój.
Thus
(20.198)
2mc2
or
2 1
Enj = ±me 112 , (20.199)
[
l + z202
(n-8;)2
]
catastrophy first occurs for j = ~ i.e. for Za > 1 or Z > 137. Since a is
small therefore for Za << 1, we can expand equation (20.199) in powers
of Za. This gives
(En. - mc2)
J
= -mc2 z2a2
2n2
[1 + z2a2
n
(-.
J
_l_ - ~)]
+! 4n
. (20.200)
where a0 is the Bohr radius. Note the important fact that states with the
same value of n but different j (e.g. 2p1¡2 and 2p3¡2 which were degenerate
in Schrodínger theory) now split.:
"\.
\.
mc2
\onllnuu m of Positiva Energy States
- ,_º
I
J
Continuum of Negativa Energy States
Sz=-112 Sz=1/2
in the vacuum. The constant mc2 produces energy gap between E > O
and E < O states. The vacuum charge vanishes. A possibility exists that
one of the electrons is lifted by a radiation quantum (photon) to a positive
energy state, where it becomes observable as an ordinary electron, and the
\) gap or hole in the infinite sea ~}the positron (posítíve chru·ge{ppposit to
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
!Y\ l
(20.209)
Thus
or
Hence
where
(20.212)
Let us write
u(p) = (uA(P)) {
UB(P) ~
(20.213)
CJ
Now in the Pauli representation of "(-matrices
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
(20.215b)
ú '
E= ±mc2
and from Eqs. (20.215), we get
(E+ mc2)uB(O)
UB(O)
=
=
O
=
L
L
o
(20.216)
(20.217)
b
and for E = -mc2:
UA(O) = 0
o
(20.218)
E = -me2 : (20.219)
Thus
u('l(o) - U,)) - m 'Spin down L o
and ( O ) eimc2t/li. (20.220)
xCll
xo(2)) e-'mc2t/li.
· and ( o ) e•mc
. 2 t ;~" (20.221)
( xc2J
are positive and negative energy solutions for spin down for a particle at
rest.
For a particle with momentum p, we note that
ur(p) = N'(!p + mc)ur(o)
satisfies Eq. (20.207), since
(/p - m) (/p + m) = p2 - m2 e2 = O
E>O, pº = _E = -1!..
E = Jp2 + m2e2 U
(r)
(p) = N
( x(r)
cu.p (r)
)
, r = 1,2
e e Ev+mc2X
(20.224)
E<O, p
o
=--=-
IEI Ep
r = 3,4
e e
(20.225)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where N = IVPº + mclN'. It is easy to see that for p-/=- O, the above spinors
are not eigenstates of 4 x 4 matrix I:;3 = ( ~3 :3) as is the case for the
spinors with p = O. The reason is that rr.n does not commute with one
Dirac Hamiltonian, where nis an arbitrary direction. However, for n =
1;1,
the operator
(20.226)
=
INl2
INl2 E
(1 +
2Ep
v+mc 2
c2(u.p)2
(Ev+mc2)2
Órs
) x(r)t x(s)
r, s =
Ev+mc2
1, 2
l (20.227a)
Similarly
2EP
uCrlt(p)uCrl(p)=INl2
e; +mc2 , b- "°':"'>
'l
r,s=3,4. (20.227b)
Now
(20.228) (_)
Therefore
r=12
r=3,4
N=. /Ev+mc2
Y 2mc2
L (20.230) G
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Hence the normalized spinors for positive and negative energy states are
= Eµ+mc2 ( x<r) )
u<r)(p) ca.p x<r) ' r = 1, 2, E > O (20.231)
2mc2 Ep+mc2
E+ mc2 ( x<r) )
u<"l(p) = 2 me 2 ~ (r)' r = 1,2. (20.236)
E+mc X
E+ mc2 ( ~X(r))
v<"\p) = 2mc2
E+mc
x(r) '
r = 1,2 (20.237)
Hence for positive energy solution i.e. for particle, the plane wave solution
is
1/J(x) = u(p)e-ifñp"x,,. (20.238)
whereas for the negative energy solution
1/J(x) = v(p)eifñp"x,,.. (20.239)
One way to interpret the negative energy solution [c.f. Sec. 20.14] is to re-
gard a particle with negative energy as an antiparticle with positive energy
but travelling backward in time. uCr) and vC"l(r = 1, 2) are called particle
and anti-particle spinors. Substituting Eqs. (20.238) and (20.239) in Dirac
equation (20.203), we obtain
(/p - mc)u(r) = O L (20.240) )
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
A+ +A-= 1
m2c2 m2c2 _ p2 m2c2 - m2c2
A+A- = -- = ---- 2mc = O. (20.246)
2mc 2mc
Now
A+u(p) = u(p), A+v(p) = OL )
A_v(p) = v(p), A_u(p) = O. (20.247)
Thus A+ and A_ project out positive and negative energy solutions respec-
tively. Further we note that
E;= 1 uCrl(p)uCrl(p) = A+(P)l
(20.248) )
To show this
..-,2
L,r=l Ua(r) (P )-(r)
U{3 ..-,2
= L,r=l
(A + ) aaUa(r) (P )-(r)
U{3 P
( )
= (A+)aaE;=1u~l(p)uy\p)
= (A+)aaÓaf3 = (A+)af3
where we have used Eq.~.243) and the relations (20.247).
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
The operation which takes a particle state function 'l/;(x) toan antiparticle
state function 'l/;c is called charge conjugation. Let us consider the Dirac
equation in the presence of electromagnetic field given in Eq. (20.137)
(20.249) ()
We fnow take the complex conjug~te of this equation:
r: ( 8µ. -
-i'y2 [ ( -ifi, .. ~:Aµ.) - me] 'l/;* = O
or [(,y2)2 = -1]
'Yº* ='Yº'
- me] 'l/;* = O L (20.251)
o
so that
'Y2'YJJ.*'Y2 = 'Yµ. (20.252)
and thisb,he rea.son for using -i'Y2 as above. Hence, from Eq~.251), we
get
(20.253)
(20.263)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Therefore
I 3 *
-'Y 'Y u (p) = ( ·2º) u
-UT
O _.
ur
2
*
(p) =N ( ·2) -UT
-iu2u· .p
e E+mc" X
-ia.2x )
=N • . (20.264)
( e E~.;,Z:,2 ( -iu2x)
Now
Thus
-ia2x(r) = rx(-r), (20.266)
1
For not too relativistic spin particles, sorne of the covariant bilinears
2
if'l/J, -0"!µ'1/J, '0"ts'l/J, ifi'Yµ'Ys'l/J and if;a µv'l/J
are 'large' while others are 'small'. To see this recall from Sec. 20.4 that in
the Pauli representation we can write the four component wave function 'l/J
as
(20.269)
where each <PA and <PE has two components so that from the Dirac equation
(20.56), using Eq (20.63)
i.e.
E~mc2 L (20.270) )
=iña · V f>
<PE ~ 2mc
<PA ~ CT • 2mc <PA, (20.271)
i.e. <PE is of order v/c compared to <PA· Using the Pauli representation of
-v-matrices, we have
i[J'if; = '1/} 'Yº'i/J
ifi'Yº'i/J = 'i/Jt'i/J = <P~<PA + <P1<PE ~ <PA t <PA = 'i/Jl'i/JA + O(v2 /c2) Ll20.273) )
i[J')'k')'5'i/J='i/Jt')'O')'k')'5'i/J
(20.276)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
In contrast to 1, 'Yo, i'Y5'Yk and O"ij which connect 1/Jl to 1PA, the nonl§agonal /\.
matrices 'Yk, i')'5'Yo, O"ko and ')'5 connect 1/Jl with 1PB or 1/!1
with 1PA· Hence
the corresponding bilinears are small and are of order v / c. For example
1P'Y51P = 1P t 'Yº ')'51/J
~ = (!~) (20.279)
. ., {j
i,i 8x0
(<lcp¡,
>n) + ui"" ·( -u.V
o u.V)
O
(<PL)
<PR - me (<PL)
<PR =
o· (20.280)
It is the mass which links <PL (or equivalently 1/JL) with <PR (or VJR)~ te
the important fact that when the mass of a fermion is zero, Eqs-t20.2 1)
are completely decoupled:
(20.282a)
(20.282b)
It is easy to see that these equations are not disconnected. In fact using the
usual representation of Pauli matrices O'i [0"2* = -0"2, 0"2<J"i<J"2 = -O'Í, i =
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
If we multiply the above equation on the left by ")'5 and use the fact that
")'5anticommute with 'Yµ, we see that 'Y5VI is also a solution of the Dirac
equation. In other words, the Dirac equation for a massless spin 1/2 particle
is invariant under the transformation
(20.284)
The above transformation is called the chiral transformation.
~ We now discuss the physical significance of the Weyl Eqs. (20.282). Con-
sider first Eq. (20.282a). The plane wave solutions of this can be written
as
cpL(x) = w(p)e(ifñ)(p·r-Et), (20.285)
where w(p) is a 2-component spinor. Substituting Eq.e.285) in
Eq.~0.282a) we find that w(p) satisfies the equation
[o-· p + E/c]w(p) = O. (20.286)
Non-trivial solutions of this equation exist only when E2 = c2p2. There
are, therefore, two solutions, one corresponding to positive energy E = clp],
and the other to negative energy E= -clpl. We denote the positive energy
spinor by u(p) and negative energy (E= -mlpl) spinor by v(p). Thus we
Problems
Show that
d1t
dt = e(E + a x B)
where B = V x A is the magnetic field (A does not depend ex-
plicitly on time) and E is the electric field.
20.2 From the fundamental properties of 'Y matrices, derive the following
relations:
5_ i µ,n >,,pt--·
( i) 'Y - -
4,éµ,v>,,p'Y
'Y u,y I )
1
(ii) 'Y5 í'>,, = 3 ! éµ,vp>,, 'Yµ, 'Yv 'Ylf1
)
(iii) [1'1-L,av>,,J = 2i(gµ,v'Y>,, - gµ,>,,í'v)L
)
(iv) {'YI-', O'v>,,} = -2Íéµ,v>,,p'Y5'Yp L
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
512 Quantum Mechanics
_ 1 vp >.
nµ, - -2-éµ,vp>.ª p
me
has the value ni= a¡ i.e. n = a":
20.5 The Dirac Equation is Lorentz invariant provided that 'ljJ satisfies
the transformation law 'l/J'(x') = S'ljJ(x), where for the infinitesimal
Lorentz transformation
S -- 1 1
+ ¡éµ,v')' /J, V
')' .
Choose the primed system in such a way that it coincides with the
rest frame of the electron so that in the primed frame
The above function for the same physical situation in the unprimed
frame is given by
'ljJ(x) = s-1'ljJ (x').
1
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Show that
E+mc2
2mc2
(~)
I
20.6 In Dirac theory of the electron, one identifies the electron intrinsic
spin,se
S=!{xxo:
4i
L o
·
20.7 The large and small components of the solution of the Dirac equa-
tion
o
while
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
20.9 A positive energy Dirac spinor (using the Pauli rep. of 'Y matrices)
is given by
(r) ( x(r) )
U (p) = _c_O', PX(r) ,
E+mc2
As m - O, show that the 'Y5 operator and the helicity u · l~I oper-
ators have the same effect on uCrl(p).
20.10 In the approximation E::::;; mc2, leA0I << mc2, show the equation
for 'ljJA is given by
1 e 2 en (NR)
-(p--A) --u·B+eAo'l/JA=E 'l/JA,
2m e 2mc
where E(NR) = E-mc2 and A= V x A.
Take A= O= Ao. Consider the probability current 4 vector
i" = eihµ'l/J.
Show that in the lowest order (E::::;; mc2)
jo= c'lj;1_'l/JA
l = c'lj;+ai'lj;
= _.!!!__ {'l/J+B'l/JA - B'lj;1'l/JA}
2m A 8x' 8x'
plus a term icijk a~i ('l/J-Ji.crk'l/JA) which can be ignored since when /
is integrated over d3x, it gives zero.
c/J'l = -icr2c/JÍl l )
L
On the other hand under parity
cpR = icr2c/Jt
L o
Thus under C P
cfJL <---- c/JR
G
c/JL - -icr2c/Jr, l )
c/JR - icr2c/JÍl L
o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
where
1/J' = (!:).
Obtain the equivalent form of ,,,,, and 1~ = i,b,i 1~13 and check
that
(,'µ,1,v +1,v11µ,) = 2gµ,v.
Verify that the Weyl set { ,,,,, } and the Pauli set { ,,,, } are related
by
where
1
s = v'2 (i i )
i -i .
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Chapter 21
The Dirac equation in (1+2) dimensions has sorne interesting features. For
one thing it has found applications in graphene, which we shall consider in
sorne detail. Furthermore, being an equation in odd space-time dimensions,
it has sorne novel features as we shall see. We start by writing the Dirac
equation in (1 + 2) dimensions: in covariant form [xµ = (z", x1, x2)]
(i,µ8µ - mc)'l/J = O (21.1)
where 80 = i !Jt.
We may replace e by Fermi velocity v¡ for a possible
application to graphene. For the moment, we put e = 1 as we shall use
natural units (e= 1, ti= 1). Here
b.
We note that the product of 'Y matrices 'Yº, ')'1, ')'2: r = 'Yº'Y1'Y2 commutes
with each of 'Yº, ')'1, ,..y2 and is thus proportional to the unit matrix. On the
other hand r2 = -1 so that
r = 1º11'Y2 = ±ih
where J2 is two dimensional identity matrix. Thus there exists two in-
equivalent representations for the I matrices in (1+2) dimensions, which
are given by the choice of sign. An explicit matrix realization of these two
L
representations is given as follows:
'Yº = (T3. 11 = ur", 'Y2 = iu2 (21.3)
}
517
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
1 O O O)
A= O -1 O O
( O O -1 O
O O O -1
We will make the second choice and thus define xP = ( xº, x1, -x2). Now
(21.4)
o
using the first representation in Eq. (21.3), the Dirac equation given in Eq.
(21.1) takes the form,
(21.8)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
so as to get back Eq. (21.5). One might therefore argue that the theory
cannot be invariant under parity. However, comparing Eq. (21.7) with Eq.
(21.6), one sees that a natural assumption would be
'l/Jt(xP) = -r¡p'l/J-(x)
i 'ljJ'l!_(xP) = -r¡p'l/J+(x)
(21.9)
L
where under the parity operation
'l/J~,B(xP) = r¡pu2'!/JB,A(x)
The peculiarity of the parity transformation (21.12) is that it changes A
(21.12) o
states into B states.
where :f>
i,º ª=º
'lj; = bef> + m)'lj;
is the momentum operator -iV. Thus
(2113) t0!:})
i!!_'lj; = H'lj; (21.14a)
8t
where the Hamiltonian L
H = ,y°(';j. +m)A
have only one rotation, therefore the orbital angular momentum operator
has only one component
We see that
(21.15)
o
[H, L3] = -a2 [p1, x1 ]p2 _ al ¡p2, x2]
= iuof> =1= º L
On the other hand
Thus for J3 -
-
L3 + la
2
3
'
(21.16)
i.e J3 is a constant of motion.
For the plane wave solutions
'lj;(x) rv u(p)e-ip.x (21.17)
which when substituted in Eq. (21.11),
(p - m)uA(P) = O. (21.18)
Writing
6(p)) (21.19)
UA(P) = (6(p)
the above equation gives
pº-m p2+ip1)(6(p))=o
( -p2 + ipl -(pº + m) 6(p)
which gives
p2 _ ipl
6(p) = p o+ m 6(p) 6(p) = p2 + ipl 6(p)
pº-m
Therefore we have two solutions,
(21.22)
a3u1 (O) = 1
2u1 (O) L )
(21.23)
3
a v1(0) = - 21v1(0).
Thus we have particle with spin clockwise (spin up) and antiparticle with
spin anticlockwise (spin down). Finally ~ ':)G l/
'l/JA = UA(p)e-iP(F ~
= VA(p)ei1U
= UB(p)e-ip¡p:
'l/JB
= VB(p)ei~x L
(p;f~
UB(P) =
VB(P) = (p2~·ip') .
L
1)
(21.24) '
E+"l1i
Now particle spinor has spin down and antiparticle has spin up.
I
21.2 DIRAC EQUATION AND GRAPHENE
21.2.1 Introduction
Natural Carbon can exist in several form~ost familiar are graphite and) ~
diamond. But there are other types.
Graphene is a single sheet of graphite and consists of a honeycomb
attice of carbón atoros. Graphite is a stack of graphene layers. Carbon
anotubes [see f¡g 16~ are rolled up cylinders of graphene. Fullerenes
C60) are molecules consisting of wrapped graphene by the introduction of
pentagons on the hexagonal lattice. Recent progress in the experimental
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Fig. 21.1
I J- functionality and henceH its potential applications., (B): Its low energy
excitations are massless Di.rae Fermions as we shall see: It mimics physics
of quantum electrodynamics (QED) for massless fermions except that in
graphene the Dirac Fermions move with the speed 1/300 the velocity of
light which implies many of the unusual properties of QED can show up in
~graphene but at much smaller speeds.
.--------.
,, ''
I I
, I '
''
I
'' \
\
'I'
, I
, ,,
------'
'\ \
\
'- --- --
Fig. 21.2
to the structure, while 7r -bands give rise to the valence and conducting
bands. The electronic properties of graphene can be described by a tight
binding model with only one orbital per atom. Within this approximation
a ~s set is provided by the Bloch functions made up of 2pz orbitals
from the 2 inequivalent carbon atoms A and, B which form the unit cell
of the graphene hexagonal lattice [see g 21.2 . Graphene is made out
of carbon atoms arranged in a hexagonal structure shown above. For a
hexagonal layer the unit cell contains two atoms A and B, belonging to the
2 triangular lattices. The lattice vectors are
(21.25)
b f .•,·, .... ~
.
/
..-- . ..l
-· --
...... ·
.. -··...
...
----~
. I
t
•• ••
-.
.~· K
....
o lf
'·\ .,,.
··--.................. _...
\ .- .-·· n
.
-.;. '•,
·'"\,,,. -··· ·...
·'
Fig. 21.3
KM
OM= b1x = -
21r
3a
J_ _
L
21r
OM = tan30Lv'3
1
l )
~ KM=
3a y'3 \--- .·é)
K and K' are called Dirac points. Their positions in momentum space are
27í 1 ¡ 27í 1
K = -(1, ¡;,), K = -(1, - ¡;,)· (21.27)
3a v3 3a va
The Dirac eones sit at the K, K' points. For an atom A of a sublattice, the
3 nearest neighbors B which belong to the other sublattice have position
vectors [see fig 21.2]
O,~ ;(1, v'3), O,~ ;(1, - v'3), O, ~ --a(l, - O) l (21.28) ()
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
where the (i,j) stands for summation over pairs of nearest neighbor atoms,
i, j on the lattice. at(Rf )(bt(Rf)) creates an electron on site Rf (Rf)
.ublattice (B) with a corresp nding annihilation operator a(R¡4)(b(Rf)).
t~~ 2.7 eV) is the nearest neighbour hopping energy (hopping between
Jr'fferent sublattices). Fourier transformi11~nihilation and creation oper-
atorl. w hav
j:..J a(~)=
A '""
~ak
,.,
'!<ORAL
e1
H = ~ ( al bt) H (::) ·
where 1{ is
H=
t.
-t ~
-t ~
eil(;rje=['
e;'t)', )
n Q~ft) aD =E (2133)
where 1{ is given in Eq. (21.31). The eigen.functions are determined by the
coefficients CA and CE in Eq. (21.32). The eigenvalues of this equation are
obtained by diagonalization of 1{ which gives the energy bands
E_ e= ±t{lei~1 + eiJco>2 + eiktl3¡2}1/2 {21.3 )n;¡) 0.
~'-".>
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
'
Fig. 21.4
2
(21.36) and the usual case 1::(p) = Jrn , m electron mass, is that vt changes
r- substantially with energy wherel"las Fermi velocity in Eq. (21.36) does not
\I
V
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
O.DI
{ 004
oca
Ftg, 21.5 Cyclotron mass of charge oarriers as a funcl.íon oí their conoentra.tion n. Pos-
itive and negat.ive n corresponda to electrons and holes, yrnbols o are the experimental
data,
where
?-lv = ( ~ 7-{AB)
7-{AB Ü
wíth
'HAB = -t[eik.ar + ék.a2]. (21.40)
Let us expand around the Dirac point K' given in Eq, (21.27):
k=K'+p,
keeping terms linear in ]PI· Then
'HAB = 3at (.'iPx + ipy
. )
. (21.41)
2
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
This gives
rfo = (º .
-ipx +Py O
ip¿ + Py)
= v¡[-p1a2 + p2a1] (21.42)
( ¡ where v¡ = 3gt is the Fermi velocity, a1 and a2 are Pauli matrices, pi =
~q,1 (p1,p2) and in relativistic notation: pJJ. = (pº,p1,p2). This does not give
°<)P which is clear from Eq. (21.42). Eq. (21.42) can be put in the form
1-lv = v¡[ia3alpl + ia3a2p2]
/ 'xd) = ,º(l)>Vf) (21.43)
~ where ,0 = cr3, -y1 = ia1, -y2 = ia2 are Dirac matrices in (1 +2) dimensions.
This is the Dirac Hamiltonian for massless fermion in (1+2) dimensions near
the Dirac point K The corresponding Dirac equation is
1•
/N
8t
=,º(-y· (-iVv¡ )'l/J (21.45)
where pis the momentum operator -iV. Writing 80 = v11 t. we can write
Eq. (21.45) in covariant form
i(,080 + ,181 + ,.../82)'!/J = o
or
(21.46)
which is the Dirac equation in (1+2) dimensions for a massless fermion. It
is important to remark that if we expand around the Dirac point K, we
obtain
Ho = v¡[-p1cr2 - p2cr1]
= v¡[ia3a1p1a1 - ia3a2p2]. (21.47)
Now as discussed in section (21.1), in 3 space-time dimensions there exists
two inequivalent representations for -y-matrices given in Eq. (21.3). We
have used the first of these representations for the expansion around the
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Dirac point K'. We take the second representation for the expansion around
the Dirac point K, which is obtained from K by the parity operation: 1
Thus the parity operation connects Dirac Hamiltonian around the Dirac
point K and K'.
A Energy
y o
---
/JI
D I
-,//
(~-> 1.:~~-- ¿_1.
Fig. 21.6
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
= e-iEt
PO-
'1/J(x)
Then from Eq. (21.48), by replacing pº ------; E - Va, the solution of Eq.
e
u(p)
(21.49)
M A+)
A_
( t
= (@) O
o
where
(21.54)
B)¿1
Then
(21.55)
G
Writing M-1 = N, to find t what we need is N31, the cofactor of M13
which is
(21.56)
0.1
~0.6
0.4 v.,=zoomevl
V =21l'ámeV
0
O.?
0.2 _.., ;
The angular dependence ofT(</J)is plotted in Fig (21.7), which shows that
there are several directions for which transmission is one.
(21.65a)
where the induced current is
JA = _..!!_/,¡w Fµ.v (21.65b)
2c
which gives the induced charge
Q= j d2xJº(x) (21.65c)
where
e .· e ..
J O (x) = --fYFij = --E'1 (aiAj - OjAi)
e e
= 2~ (V x A)3 = 2~B (21.65d)
e e
so that
Q = 2~ J d2xB
where
ie ie
D2 = (81 + -A1) 2
+ (82 + -A2) 2
.
e e
We now consider a static magnetic field:
Aº=OQ=O, E=O.
We can select the gauge
A1 = -A1 =-A:"= By (21.69)
A2 = O (21.70)
and we can write the wave function 1/J A as
1PA(x,y) = 'PA(y)eikx (21.71)
where the 2-component spinor cp A (y) is
)
'f)A -
_ ('P+)(,
'P-
ua - -1+1l. 3 =
3 -
~
(21.72)
l
Then Eq. (21.60) gives
--E2 -i2(k e
+ -By)2 - -
82 eB
'P+ - -cp+ = O. (21.73)
[ VJ C 8y2 C
e 2 8y2
[ (~B) + Y 2] 'P+(Y) = ( VJ
E2
+ ~B
e ) 'P+(Y)
or
1 2 1 e
+ -(-B) 2
Y 2] 'P+(Y) 1 e 2 ( -E2 e )
+ -B 'P+(Y) _ ,
[ --8y
2 2 e
= -(-)
2 eB v¡ e
= E 'P+·
(21.74)
which is the equation of 1-dimensional harmonic oscillator, so that energy
levels are given by
EI =
n
(
n+-21) -eB
C
(21.75)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
or
1 e 1 e
-E
v2
2
n
+ -B
e
= 2(n+-)-B.
2 e
f
Thus
En= sgnlnl wclnl!, n = O, ±1, ±2 · · · (21.76)
where
(21. 77)
. v2
En = sgnlnl 2lnlñ_l_eB , n =O± 1, ±2 · · · (21. 78)
e
These are Landau levels near the Dirac point K'. The Landau levels at the
opposite Dirac point K have exactly the same spectrum and hence each
Landau level is doubly degenerate. This particular Landau spectrum has
been verified by several experimental groups. The result (21.69) is relevant
to half integer (anomalous) quantum Hall effect in graphene.
21.3 Problems
L
show that Eq. (21.60) of the text takes the form for u3 = 1
2
[-Di - D2 -
2 e
~BJef >+ = v¡ </J+
f.2
()
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Show that
.D2
Q, P:
,
---¡;B
ie l o
-i eB = Q,
e
so that
[Q,PJ =i
and
where
E = ! ( ~ + eB)'
2 v¡ e
w = eB
e l
The above is the equation of/:!_ne-dimensional harmonic oscillator,
and give energy levels
7.
196!andl,
Quantum !ech~cs, .JIÍ<l¡, WHey/,wns,~2
8. P.T. Matthews, ~duct10n to QuantlllQ., Mecha;mcs ( h1rd Ed1-
. . ISy~ ~ 1 L,.c_ ")
.,.!!_on)_lli!cGraw-l-lill Book Company (UK) Limited, 1974.
9. A. Messiah, ~ntum Mechanics (in 2 volume~, John Wiley &
sons, New York, 1968.
10. E. Merzbacher, Quantum Mechanics (2nd edition) John Wiley &
sons, New York, 1970.
11. J. J. Sakurai, Advanced Quantum 1eChanl~, Addision-Wesley
Publishing Company, 1973.,¡ Modern Quantum Mechanics (S. F.
Tuan, Editor), Addison-Wesley, Reading, Mass, 1994.
12. L. I. Schiff, Quantum Mechanics (3rd edition), McGraw-Hill, New
York, 1968.
13. R. Shankar, Principle of Quantum Mechanics, Plenum Press, ew
York, 1980.
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Index
541
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Dirac equation, 330, 461, 482, 506, Eigenvalue, 27, 29, 30, 33, 36, 37, 49,
529, 534 62
in (1+2) dimensions, 529 of L2, 96, 492
covariant form, 461 of t.; 92, 96
for a massless spin 1/2 particle, 509 of unitary operator, 437
for hydrogen-like atom, 491 Eigenvalue equation, 27, 29, 30, 33,
for massless particle, 533 35, 43, 52, 61, 73, 76, 404
Graphene, 521 Eigenvector, 1 77
in (1 +2) Dimensions, 517 Einstein relation for energy, 20
in the presence of electromagnetic Elastic scattering, 124, 129, 152, 424,
field, 481, 504 426
plane wave solution, 497 form factor, 428
positive and negative energy Electric charge density, 291
solution, 502 Electric dipole moment, 291
positive and negative energy states, Electromagnetic radiation, 2
502 Electron coulomb energy, 296
two component Pauli form, 488 Electron scattering, 155
under the charge conjugation, 504 Electronic density, 527
Electronic properties of graphene,
Dirac fermions in an externa! gauge
522, 523
field, 533
Energy dispersion, 527, 531
Dirac Hamiltonian, 483
Energy eigenfunction, 78, 202
in the presence of electromagnetic
Energy of a free particle, 22
field, 511
Energy operator, 26, 48
Dirac matrices in (1+2) dimensions,
EPR entanglement, 354
529
EPR states; 356
Dirac notation, 346
Dirac point, 530
entangled states, 356 J_
Dirac points, 524
Eulerp.[.,agrange equation, 224 IV
External ymmetry, 109
energy band, 526
Dirac quasi-particles, 527 Fermi energy, 322
Dirac's Role theory, 496 Fermi golden rule, 418
Discrete Fermi velocity, 526, 527, 529 f
spectrum of energy eigenvalues, 120
Discrete energy levels, 63
FermtDirac statistics, 312, 315
Fermion, 312
ts
Discrete transformation, 218 Fermionic oscillator, 194
Discrete transformations, 470 Fine structure constant, 109, 379
Dispersion formula, 141 Finite rotation, 218
Dispersion relation, 394 Flavor eigenstates, 340
Displacement operator, 216 Flux, 419, 425
Dual Behaviour of Radiation, 7 Fourier transform, 43, 44, 141
Dual nature of matter, 10 Free particle Dirac equation, 473
Free particle propagator, 207, 225
Effective cross section, 126 Fullerenes, 521
Eigenfunction, 27, 29, 184
momentum, 152 Gauge, 536
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Wave equation, 2
of electric and magnetic field, 2
Wave front, 2
Wave function, 21, 23, 24
Wave length, 3, 6, 10
Wave number, 168
Wave packet, 4-6, 11, 12, 21
Weak interaction, 220
Weyl equation, 508
Weyl representation of ,-matrices,
463
Weyl spinor, 519
Weyl wave function, 509
WKB approximation, 157, 159, 163,
165
WKB solution, 162
X-rays, 9