(Quantum Mechanics) Fayyazuddin, Riazuddin-Quantum Mechanics-World Scientific (2012)

I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
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
¡¡¡
Publishers' page
iv
¡ July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
Publishers' page
V
Publishers' page
vi
To the memory of our parents, brother mbdul .Sfatim and sister

3ohra
vii
viii Quantum Mechanics

Preface
Preface for the 2nd Edition

Since the publication of the first edition of our book on Quantum Me-
chanics 22 years ago, physics has progressed in many directions. To catch
the flavor of these developments where quantum mechanics has faund appli-
cation, we have added three new chapters. One is on the Two state prob-
lem, where a unified treatment of Rabi oscillations, Particle mixing and
Neutrino oscillations is given, providing a beautiful illustration of quantum
mechanical phenomena of superposition and interferometry. The second is
on Quantum computing which is again an application of two state physics.
The third one is a beautiful application of the Dirac equation in (1+2)
dimensions to Graphene, where there is now evidence far the existence of
massless quasi particles obeying the Dirac equations. This is of topical in-
tere ·t as currently a lot of work, both experimental and theoretical, is going
on in this field. ~
The other major changas we have rnade are as follows:7 C!.Y
In. fhapter 10, a new s ction on th Path integral formul!ttion of quanturn
mechanics was added. In chapter 20 (chapter 18 in the first edition) on
the Dirac equation, we have adopted a more commonly used metric and
notation far gamma matrices. We have also expanded the chapter so that
it now contains a section on the Lorentz group and its representations. In
chapter 9 (operators), a section on supersymmetric o ·cillatobs added, as
an example of a supersymmetric algebra. In chapter 12 ('.rime indep ndcnt
perturbation theory), a section on the "variatíonal principle" and its ap-
plications was added. Sorne portions of the chapters, namely 5 (Harmonic
oscillator), 8 ( Collision theory), 9 ( Operators), 17 ( old 15, l~i;era.ction of
radiation with matter) have been rewritten far more simplicity ~a.ter clar-
ity. Another important feature of the second edition is the addition of a
ix
X Quantum Mechanícs
substantial number of new problern .

In this and the fust ~ w have followed what Freeman J Dyson1
has called a third legacy of Dirac {not so well known to people) namely
"Dirac's disengagement from verbal dispute about the meaning of quan-
tum mechanics. When expressed in mathematical equations, the laws of
quantum mechanics are clear and unambiguous." This is what is required
of a theory.
Finally we wish to express our deep appreciation to Dr. Farhan Saif
for writing the chapter on Quantum computing for us, and to Dr. Kashif
Sabeeh for critically reading chapter 21, on Graphene. We wish to express
our deep thanks to Ishtiaq Ahmed (our gradúate student) for managing and I
doing the typing, drawing figure , ~·eparing the index. Thanks are also a I\O
dueto M. Tahir Iftikhar Muhammad Usman, Faisal Munir, Ghulam Farid,
Bilal Tariq and Saadi Ishaq for assistance with typing the manuscript.
Fayyazuddin
Riazuddin
Preface for the lst Edition

Quantum Mechanics was discovered in 1920's to explain the stability
S of atorr{.)t is undoubtedly one of the greatest achievements in theoretical
physics in the century. It is the basic theory for atomic and subatomic
phenomena.
The concepts of quantum theory are radically different from the classical
theory, which describes the everyday phenomena successfully. The quantum
mechanical concepts are described in mathematical language. This is the
approach we have followed in this book. We, however, do not assume any
advanced knowledge of mathematics. The knowledge of differential and
integral calculus and familiarity with matrices are sufficient to understand
this book. The mathematics needed beyond this is developed in the text.
We have tried to keep the presentation well motivated and to provide
sufficient details in order to facilitate the understanding of the subject.
Our emphasis is on the basic theory rather than on specific applications in
atomic, molecular, solid state and nuclear physics.
The book could be divided into 3 s rnester courses. hapter l-7 Chap-
ter 8 (Secs. 1-7) and Chapter 12 (S cs. 1-2) could form/ee semester under- (A
graduate course. Chapter 8 (Secs. -L3), Chapter 9 Chapter lO (Secs. 1-3),
1F. J. Dyson Silent ua.nL,1m cnius.
New York Review of Book : . 1 S rungest Man:
The Hidden Life of Paul Dirac, Mystic of the Atom, by G. Farmelo.
Preface xi
Chapter 11 (Secs. 1-5), Chapter 12 (Sec. 4), Chapters 13 and 14 should be

suitable for the second semester undergraduate course. The rest of the
sections and chapters could formeie semester graduate course.
This book is based on~ur e of lectures which we have given at the
Punjab University, Lahore The Quaid-I-Azam University, Islamabad, The
King Fahd University of Petroleum and Minerals, Dhahran(R), The King
Saud University, Riyadh(F) and the Ummal-Qura University, Makkah Al-
Mukarramah(F) at various times. In fact we have been encouraged by our
students to write these lectures in a book form. We would like to express
our thanks to them and acknowledge the respective universities for their
support.
In particular we are grateful to our former students Dr. M. M. Ilyas
and especially Dr. Sajjad Mahmood for help in preparing this book for
publication. We are grateful to our colleagues Dr. Fahim Hussain and
Dr. Pervaiz Hoodbhouy for reading the first draft of the manuscript and
for sorne useful suggestions.
We also wish to express our thanks to Mr. Shbahat Ullah Khan for
typing the first draft of the manuscript.
We were first introduced to this subject by Prof. Abdus Salam. We
would like to take this opportunity to express our deep sense of gratitude
to him for the encouragement throughout our careers.
Fayyazuddin
Riazuddin
March 4, 1999
xii Quantum Mechanics

Contents
Prejace ix
l. BREAKDOWN OF CLASSICAL CONCEPTS 1

1.1 Introduction . . . . . . . . . . . . 1
1.2 Breakdown of Classical Concepts 7
1.3 Problems . 14
2. QUANTUM MECHANICAL CONCEPTS 17

2.1 Uncertainty Principie . 17
2.2 Illustration of Heisenberg Uncertainty Principie . 19
2.3 Schrodinger Equation . 20
2.4 Physical Interpretation of Wave Function 23
2.5 Stationary States . 27
2.6 Eigenvalues and Eigenfunctions 29
2.7 Problems . 30
3. BASIC POSTULATES OF QUANTUM MECHANICS 33

3.1 Basic Postulates of Quantum Mechanics . . . . . 33
3.2 Formal Properties of Quantum Mechanical Operators 34
3.3 Continuous Spectrum and Dirac Delta Functions . . . 38
3.4 Uncertainty Principie and Non-Commutativity of Observ-
ili~... ~
3.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . 48
4. SOLUTION OF PROBLEMS IN QUANTUM MECHANICS 51

4.1 One Dimensional Potential Step . 51
xiii
xiv Quantum M echanics
4.2 The Potential Barrier . . . . . . . . . . . . . . . . . . 58

4.3 Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.4 An Example of Bound State (Discrete Energy Levels) 63
4.5 Bound State (Square Well Potential) 65
4.6 Problems . . . . . . . . . . . . . 68
5. SIMPLE HARMONIC OSCILLATOR 71

5.1 Introduction . 71
5.2 Quantum Theory of simple Harmonic Oscillator. 72
5.3 Properties of Hermite Polynomial . 78
5.4 The Average Values 81
5.5 Problems . 83
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
7. MOTION IN CENTRALLY SYMMETRIC FIELD 103

7.2 Coulomb Potential . 106
7.3 Wave Functions of Hydrogen-like Atom 110
7.4 Problems . 114
8. COLLISION THEORY 117

8.1 Two Body Problem: (Centre of Mass Motion) . 117
8.2 Collision Theory . . . . . . . . . . . 120
8.3 Kinematices of a Scattering Process . . . . . 121
8.4 Scattering Cross Section . . . . . . . . . . . . 125
8.5 Scattering by a spherical symmetry potential 127
8.6 Phase Shifts . . . . . . . . . . . . 131
8. 7 Optical Theorem . . . . . . . . . . . . . . . . 136
8.8 Resonances, Dispersion Formula . . . . . . . 141
8.9 Bound States and Resonances: Square Well Potential. 144
8.10 8-function Potential , . . 147
Contents XV
8.11 The Born Approximation . . 150

8.12 Qua.si-Cla.ssicalApproximation 157
8.13 Problems 168
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
10. HEISENBERG EQUATION OF MOTION, INVARI-

ANCE PRINCIPLE AND PATH INTEGRAL 201
10.2 Heisenberg Equation of Motion 203
10.3 Free Particle Propagator . 207
10.4 Unitary Transformation . 209
10.5 Invariance principles, Coservation laws 211
10.6 Discrete Transformation . . . . . . . . 218
10.7 The Path Integration Formulation Of Quantum Mechanics1 221
10.7.1 Introduction . . . . . . . . . . . . . . . . . . . . . 221
10.7.2 Path Integral Method . . . . . . . . . . . . . . . . 222
10.7.3 Free Particle Propagator from Path Integral For-
mulism . . . . . . . . . . . . . . . . . . . . . . . . 225
10.7.4 Motion with a Source: Generating Functional for
Time-ordered Products of Heisenberg Operators fi.i ( t) 226
10.8 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
11. ANGULAR MOMENTUM AND SPIN 237

11.1 The Zeeman Effect 237
11.2 Angular Momentum 240
11.3 Matrix Representation of Angular Momentum . 244
11.4 Spin 246
11.5 Splitting of the Ground State of Hydrogen Atom 250
11.6 Addition of Spin 251
xvi Quantum M echanics
11. 7 Addition of Angular Momenta 254

11.8 Rotations: Rotation Matrices 261
11.9 Vector and Tensor Operators 268
11.10 Problems . 276
12. TIME INDEPENDENT PERTURBATION THEORY 283

12.2 Stationary Perturbation Theory . 283
12.3 Degeneracy . 287
12.4 The Stark Effect . 290
12.5 Non-Perturbative Method (Variational Principle) 295
12.6 Problems . 299
13. TIME DEPENDENT PERTURBATION THEORY 303

13.1 Transition Probability 303
13.2 Problems . . . . . . . . . . . . . . . . . . . . . 309
14. STATISTICS AND THE EXCLUSION PRINCIPLE 311

14.1 Introduction . . . . . . . . . . . . . . . . . . . . 311
14.2 Permutation Operator and Exclusion Principle 312
14.3 Non-interacting Particles and Exclusion Principle . 314
14.4 Two Electrons System (Helium Atom) . 316
14.5 Scattering of Identical Particles 319
14.6 Problems . . . . . . . . . . . . 321
15. TWO STATE SYSTEMS 325

15.1 Introduction ..... 325
15.2 Magnetic Spin Resonance 329
15.3 Particle Mixing . 332
15.3.1 Introduction . 332
15.3.2 General Formalism for Particle Mixing 334
15.3.3 K - ~e º
K omplex . . . . . . . . . . . 337
15.3.4 Bq - Bq complex ( q = d, s ) .
O -O
339
15.3.5 Neutrino Oscillations 340
15.4 Problems . 344
16. Quantum Computation 345

16.1 Introduction .... 345
Contents xvii
16.2 Qubit . 346

16.3 Multiple qubit systems . 347
16.4 Matrix representation of qubits 348
16.5 Single Qubit Logic Gates . . . 349
16.6 Quantum Logic Gates at work 351
16.7 Two Qubit Logic Gates 351
16.8 Quantum Circuits . 353
16.9 EPR Entanglement . 354
17. PERTURBATION INDUCED BY ELECTRO MAG-

NETIC FIELD 359
17.1 Interaction of Electromagnetic Field with Electrons . 359
17.2 Gauge Principle : Aharanov-Bohm Effect 362
1 7 .3Landau Levels . . . . . . . . . . . . . . . . . . 367
17.4 Quantization of Radiation Field . . . . . . . . . 369
17.5 Perturbation Induced by Electromagnetic Field 373
17.6 Stimulated Emission and Absorption of Radiation 374
17.7 Spontaneous Emission of Radiation: 378
17.8 Decay Width . . . . . . . . . . . . . . . . . . . . . 383
17.8.1 E1 transition. . . . . . . . . . . . . . . . . 383
17.8.2 Magnetic dipole transition dueto intrinsic spin: 386
17.9 Dipole Sum Rules, Two Point Function 387
17.10 Application of Dipole Sum Rules 392
17.11 Dispersion Relation. 394
17.12 Problems . . . . . . . . . . . . . 396
18. FORMAL THEORY OF SCATTERING 403

18.1 Introduction . . . . . . . . . . . . . . 403
18.2 Interaction Picture . . . . . . . . . . . . . . . . . . . . . . 405
18.3 Formal Solution of Schródinger Equation in the Interac-
tion Picture . . . . . . . . . . . . . 407
18.4 Scattering States - The S-Matrix . 412
18.5 The Scattering Cross-section ... 418
18.6 Properties of the Scattering States 420
18.7 The Coordinate Representation 422
18.8 Unitarity of the S-Matrix 428
18.9 Optical Theorem . 431
xviii Quantum Mechanics
19. S-MATRJX AND INVARJANCE PRINCIPLES 435

19.1 Introduction . . . . . . 435
19.2 Rotational Invariance 437
19.3 Parity ..... 442
19.4 Time Reversa! 445
19.5 Problems ... 450
20. RELATIVISTIC QUANTUM MECHANICS:DIRAC EQUATION451

20.1 A Brief Review of Theory of Relativity ..... 451
20.2 Relativistic Quantum Mechanics: Introduction 456
20.3 Dirac Equation . . . . . . . . . . . . . . . 458
20.4 Covariant form of Dirac Equation . . . . . . . 461
20.5 Relativistic lnvariance of Dirac Equation . . . 464
20.6 Transformation Properties of Dirac Bilinears 468
20.7 Discrete Transformations . . . . . . . . . . . 470
20.8 Generators of the Lorentz Group and their Representations 475
20.8.1 Matrix representation of the generators of the
Lorentz group . . . . . . . . . . . . . . . . . . . . 475
20.8.2 Representations of Lorentz Group . . . . . . . . . 477
20.9 Gauge Invariance: Dirac Equation in the presence of Elec-
tromagnetic Field . . . . . 481
20.10 Constants of Motion . . . . . . . . . . . . . . . . . . . . . 483
20.11 'Velocity' in Dirac Theory . . . . . . . . . . . . . . . . . . 486
20.12 Two Component Pauli Form of Dirac Equation: Existence
of Magnetic Moment . . . . . . . . . . . . . . . . . . 488
20.13 Dirac Equation for the Hydrogen-like Atom . . . . . . . . 491
20.14 Role theory and existence of Positron (anti-matter) . . . . 496
20.15 Free Particle Solutions, Plane Wave Solutions of Dirac
Equation . . . . . . . . . . . . . . . 497
20.16 Charge Conjugation . . . . . . . . . 504
20.17 Large and Small Bilinear Covariants 506
20.18 Weyl Equation . . . . . . . . . . . . 508
21. DIRAC EQUATION IN (1+2) DIMENSIONS: APPLI-

CATION TO GRAPHENE 517
21.1 Dirac Equation in (1+2) Dimensions . . . . . . . . 517
Contents xix
21.1.1 Dirac Equation in Hamiltonian form and Plane

Wave Solutions . . . . . . . . . . 519
21.2 DIRAC EQUATION AND GRAPHENE . 521
21.2.1 Introduction . . . . . . . . . . . . 521
21.2.2 Electronic properties of graphene 522
21.2.3 Cyclotron mass . . . . . . . . . . 527
21.2.4 Dirac Hamiltonian and Dirac Equation for Quasi-
particles in Graphene . . . . . . . . . . . . . . . . 528
21.2.5 Tunneling and Klein Paradox . . . . . . . . . . . 530
21.2.6 Dirac Fermions in an Externa! Gauge Field (Elec-
tromagnetic): Landau Levels . . . 533
21.3 Problems . . . . . . . . . . . . . . . . . . 537
21.4 Selected References and Further Reading 538
Index 541
Chapter 1
BREAKDOWN OF CLASSICAL
CONCEPTS
1.1 Introduction
Quantum mechanics is the theory which describes phenomena on the atomic

and molecular scale. An event in this domain is not visible to the human
eye. The concepts of classical physics which have been developed to de-
scribe phenomena on the macroscopic scale may not be applicable for pro-
cesses on the microscopic scale of dimensions 10-6 to 10-13 cm. There are,
however, sorne macroscopic quantum systems, e.g. superfluids, supercon-
ductors, transistors and main sequence stars.
We first outline the~cal theory and then describe
how, for the microphysical world, the necessity for departure from clas-
sical physics is clearly shown by experimental results.
In classical physics, matter is treated in terms of particles of definite
mass, and radiation is described as wave motion. The two great disciplines
of classical physics are Newtonian Mechanics and Maxwell's theory. The
former describes the mo ion of mat rial particles according to Newton's
Laws. Classical mechanícs successfully describ ·S electrically neutral }-,
macroscopic systems. Energy E and momentum p of a particle are two
important dynamical variables.
The electric and magnetic phenomena are described in terms of electric
and magnetic fields E and B which satisfy Maxwell's equations
divE = 41rp (1.la)

18B
curlE = --- (1.lb)
e 8t
47r. 18E
curlB = -J + -- (1.lc)
e eot
divB = O. (1.ld)
1
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)
In free space, E and B satisfy the wave equation
( e1 fJt2 -
[)2 )
v'2 E, B = O. (1.3)
A solution of this equation shows that E and B propagate through space

as waves with speed c. For appropriate frequencies, these waves should
be identified with visible light. The whole spectrum of radiation from ex-
tremely long wave length region of radio waves, through visible range, to
extremely small wave length region of X-rays and -y-rays is described in
terms of electromagnetic waves as given by Maxwell's theory.
As we have seen, electromagnetic radiation is regarded as consisting of
waves which propagate through space with velocity c. A typical wave in
x-direction is expressed as:
'l/J(x, t) = Aei(2;-x-21rvt)
>. : wave length

T = 1/ v : periodic time ( v frequency) (1.4)
21r
k = -:X : wave number
w = 21rv : angular frequency .
k can be regarded as vector. A wave in 3 dimensions can be written as
'l/J(r, t) = Aei(k·r-wt). (1.5)
We define a phase () =k · r - wt. A surface of constant phase is called a
wave front. The velocity with which this surface moves is called the phase
velocity. To calculate the phase velocity we note that () is constant on this
surface:
d() =o
dt
or
dr
k. dt -w = o.
This gives the phase velocity [k = 1 k In]
w
Il·U= lkl = VA. (1.6)
BREAKDOWN OF CLASSICAL CONCEPTS 3
We cannot send a signal in the form of a monochromatic wave. However,

C, what we do, in pract~, is to send a signal in the form of a wave packet or
group of waves. The only velocity which can be experimentally measured is
the group velocity which we define below. The wave packet can be generated
by superposition of a number of simple harmonic waves with wave numbers
centered round the mean wave number. Consider first the superposition of
two waves
'!/Ji = Aei(kx-wt)
'1/J2 = Aei[(k+.6.k)x-(w+.6.w)t]
'1p = '1/J1 +'1/J2

= Aexp [i( k + ~k)x - (w +
;)t]
6
x [ exp [ - i ( ~k x - ~wt) J + exp [i ( ~k x - ~w t) J J

~ Aei(kx-wt)2cos ( ~k x - ~w t) (1.7)
where we have put
6k
k+-~k
2
6w
w+
2~w.
The first term in the last line of Eq. (l. 7) is carrier wave of frequency 2: and
wave length 2;,. The second term gives the modulation of the amplitude of
the carrier wave.
The phase velocity as befare is given by
w
u= k = t/);
On the other hand, the maximum of the amplitude moves with velocity
6w 6v
Vg = 6k = 6(1/ .\)"
(1.8)
This is the group velocity.

The wave length of modulation is given by
27T' 41r
>-m = l6k = 6k . (1.9)
2
The plot of Eq. (1.7) is shown in Fig. 1.1. The width of the wave packet is
evidently given by (half the wave length of the modulation)
1 21r
6x = 2>.m 6k
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) )
'1/J(x, O)= l A(k)eikxdk

/
(1.llb)
)
j L
A(k) = _!__
2rr _00
00 '1/J(x)e-ikxdx (1.llc)
The wave packet is locaLised within a dístance t::..x and has a spread t::..k in
wave number as shown in Pigs, .2 and 1.3. t can be shown by Fourier
ü
analysis that t::..x and t::..k ar uch ha
t::..xt::..k ~ 1.
Fig. 1.2 A wave packet pictured at t = O
A(k)
Fig. 1.3 Picture of 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)
To reduce the dispersion, we take

6.k « ko ,
¡ July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
and expand Eq. (1.12) around k0:
w(k) = w(k0) + (!~) ko (k - ko) + ... (1.14)
Now let us rewrite Eq. (1.llb) in the followingway

1/)(x, O) = f~ooA(k)eikox+i(k-ko)xdk
= eik0x f~oo A(k)ei(k-ko)xdk (1.15)
or
1/J(x) = eikox X(x) , (1.16)
where
X(x) = ¡_: A(k)ei(k-ko)xdk (1.17)
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)
in the exponential. This is justified only if
t(~:~) ko (.6.k)2 « 1
or
or
(1.20)
The wave picture of radiation satisfactorily explains the phenomena of

interference and diffraction of light.
1.2 Breakdown of Classical Concepts
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.
(a) Dual Behaoiour of Radiation
(i) Phenomena of interference and diffraction, which are characteristic of

waves.
(ii) Phenomena of emission and absorption of radiation which cannot be
explained by the wave nature of radiation. Examples of phenomena indi-
cating particle aspects of radiation are as follows.
(1) Blackbody spectrum
This concerns the thermodynamics of the exchange of energy between

radiation and matter. It was shown by Planck that the correct thermody-
namical formula, namely
E(w) = 81rliw3 (e(ñw)/kT -1)-1 , (1.21)

(81rc)3
where E(w) is the energy per unit volume ofwave, with angular frequency w
and k is the Boltzmann constant, is obtained only if radiation of frequency
w exchanges energy with matter in discrete units of Iu», Here as well as in
the formula (1.21), n = 2~, h being a universal constant, called Planck's
constant, has dimensions of [E] [t] and a numerical value
n= 1.055 X 10-27erg s = 6.582 X 10-22 MeVs. (1.22)
Incidently the Planck's formula (1.21) reduces to the classical Rayleigh-

Jean's law for low frequencies liw « kT while the total emitted energy
E= fo 00 E(w)rku ,
8 Quantum Mechanics
remains finite and gives the Stefan's law

E= 4a- T4 (1.23)
e
• 271"5 k4
with O" = 15h3c2.
Thus, according to Planck's hypothesis, the only available energies for
the wave of frequency w will be O, ñu), 2ñw, ... , i.e. radiation of frequency
w behaves like a stream of particles (photons) of energy E = ñw.
Now according to the special theory of relativity
E2 /c2 = p2 + m2c2 . (1.24)
Since photons travel with the velocity of light e, their rest mass should be
zero, so that
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.
(2) Photoelectric effect
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.
(3) Compton effect
The particle aspect of radiation is most clearly shown in the Compton

effect, the scattering of X-rays by electrons bound in atoms. The binding
energy which is in eV range can be neglected as X-rays are in the MeV
range. The change of wave length .6.,\ observed by Compton depends on
the angle of scattering and not on its wavelength. This cannot be explained
classically since in terms of radiation waves one expects ,6.,\ proportional
to ,\. This can be easily explained as an elastic scattering of a photon of
energy E = ñw and momentum p = ñk with an electron at rest as shown
in Fig. l. . It is easy to see from energy-momentum conservation that
Fig. 1.4 Collision of a photon with electron at rest in Compton effect.
.6.~ = ( -\2 _ -\1) = (!!:_ _ !!:_) = ~ sin2 !l.

271' 27!' 27!' P2 PI me 2
= 2~
271'
sin2 !l.2 (1.27)
where
-Ae = -ñ ~ 3.86 x 10- 11 cm (1.28)

271' me
is called Compton wave length of the electron.
{b)Dual Nature of Matter
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
For group velocity we have

dw dE
Vg = dk = dp (1.33)
Now
or
2E dE = 2 e2p.
dp
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
1/J(x, t) = j A(k)ei[kx-(lik 2 /2m)t]dk ,
since for a free non-relativistic particle

E p2 ñk2
w(k) =--¡¡= 2mñ = 2m · (1.35)
For a wave packet, we have shown that

l:ik!:ix ¿ 1 .
Since p = ñk, we have
/:ip!:ix ¿ ñ • (1.36)
This is an example of the Heisenberg uncertainty principle, which states
that uncertainty !:ix in the position of a particle and uncertainty !:ip in its
momentum is given by the above relation. The association of a wave packet
with a particle implies that it is not possible to determine simultaneously
the position and momentum of a particle precisely.
As we have discussed previously A(k) is appreciably different from zero
in the range ( !:ik « k)
1
ko - -2 !:ik <
-
k <
-
ko + -21 !:ik .
Now w(k) can be written in the form
w(k) =
2~
[k5
Thus the wave packet 1/J(x, t):

+ 2ko(k - ko) + (k - ko)2]; !:i = (k - ko) L o
1/J(x, t) = j A(k)ei(ko+k-ko)xe ;;,;'[k5+2ko(k-k0)+(k-k0)2]
~ e'"(k ox- ñk5)

2m
ñ
X(x - vt - -1:ikt)
m
where
X(x-vt- ~ Llk) ~ 1: A(k)e'(kx-w(k)<)[x-•<--,!.ak<)dkL(l.37)
Thus, we see that the wave packet propagates without appreciable

O
spread with the velocity v, the particle velocity. This is true only if ac-
cording to Eq. (1.20)
ñ
t(-)ók
m
« óx'
or
tóv « óx. (1.38)
Now tóv gives uncertainty in position of the particle at time t over and
beyond the initial uncertainty óx. Thus the wave packet spreads as it
propagates, this spread is negligible provided that
tóv « óx.
To sum up, what are conventionally labeled as particles have wave prop-
erties, which can be exhibited under suitable conditions and conversely light
is as much an elementary particle as an electron, a proton or a neutron.
Thus, for phenomena on the atomic scale, sorne basic change was needed
in our ideas. In the new theory, essentially non-classical features of particle
aspects of radiation and wave aspects of particles should appear naturally.
(c)Hydrogen Spectrum; The Bohr atom

The spectrum of hydrogen is very simple. It has discrete frequencies
and the corresponding wave lengths satisfy the general law
~ - R(2_ -
.X - nr 2-)
n~ ' (1.39)
where n2 = 3, 4, 5, ... , n1 = 2 gives the Balmer series and R is the Rydberg

constant given by
mee4 mee ~1
= 4,,r-ñ,3c = 41rñ ( a 2 ) = 109, 677, 581 cm ( = 13.595 volts) ,
R
where a = (e2 /ñc) = 117 is a dimensionless constant and is known as the

fine structure constant.
According to Rutherford, an atom consists of a heavy nucleus with
a charge Ze, about which Z electrons revolve. According to Maxwell's
theory, an accelerated charge emits electromagnetic radiation. An orbiting
electron constitutes a rapidly accelerating charge and should emit radiation
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, ....
(ii) The radiation is emitted or absorbed when electron makes a transition

(quantum jump) from an orbit of energy En to one of energy E~, say, and
the resulting angular frequency Wnn' is given by
nwnn' = IEn - En, 1 . (1.40)

Let us apply these rules to the hydrogen atom. For a circular orbit of atom
a, we get by equating Coulomb force to centripetal force:
e2 /a2 = maw2 .
Now
L = ma2w = nñ .
This gives discrete values for a and w:
(1.41)
where
- ñ,2 -11
ao = --2 = 5.29 X 10 m (1.42)
me
is called the first Bohr atom.
Now
En= -1 2m(anwn)2 - -e2 = ( 1 e2) 1

- -2- (1.43)
an ao 2,
n
14 Quantum Mechanics
and from Eq. (1.40)
e2 ( 1 1 ) (1.44)
Wnn' = 2fíao n2 - n'2 .
This explains the discrete spectrum of the hydrogen atom; in particular we

get the Balmer series with n = 2 and n' = 3, 4, 5, ....
The Bohr atom is stable, since no further radiation of energy is possible
once it has reached the lowest level (ground state) E1 = ( - !e2 /a~).
The success of the Bohr theory for hydrogen-like atoms is impressive but
clearly unsatisfactory in the sense that the motion of the electron is de-
scribed in classical terms by the superposition of additional arbitary rules.
To summarise, the non-classical features which have emerged are:
(i) Particle aspect of radiation.
(ii) Wave aspect of particles.
(iii) Sorne dynamical variables such as energy and angular momentum hav-
ing discr te vaJues.~ese features should appear as natural consequences of
the new theory - quantum mechanics.
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)
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.
Chapter 2
QUANTUM MECHANICAL
CONCEPTS
2.1 Uncertainty Principle
Classical mechanics breaks down when applied to small systems, provided

they are small enough, and scienceis concerned with observablephenomena.
For classical systems, an assumption is tacitly made that the operation of
observation does not appreciably disturb the system and the disturbance
which may be caused by measurement can be corrected for exactly, at least
in principle. On the other hand, we can observe phenomena on atomic scale
only by using microscopic objects (atoms, nuclei or photons) of the same
dimensions so that the process of measurement disturbs the system to be
observed and for small enough objects this disturbance is not negligible.
If for example, the momentum of an electron is known initially and we
try to measure its position, then at least one photon must be scattered by
the electron to come through the microscope onto the screen. This would
disturb the electron by an amount which cannot be predicted so that we
cannot observe an electron without disturbing it. As Dirac has put it "there
is a limit to the fineness of our powers of observation and the smallness of
the accompanying disturbance - a limit which is inherent in the nature
of things and can never be surpassed by improved technique or increased
skill on the part of the observer. If the object under observation is such
that the unavoidable limiting disturbance is negligible, then the object is
big and we may apply classical mechanics to it. If on the other hand, the
limiting disturbance is not negligible, then the object is small and we have
to apply a new theory, i.e. quantum mechanics to it". (P. M. A. Dirac, the
Principles of Quantum Mechanics, Oxford University Press, 4th Edition,
p. 3.)
The limit to the fineness of our powers of observation and the smallness
17
of the accompanying disturbance is expressed by the Heisenberg Uncer-

tainty Principie.
Suppose we wish to look at an electron. To do so we must touch it with
something else, e.g. shine light on it. Let the wave length of light be >..
Each photon of this light has momentum Q = h/ >.. In arder to "see" the
electron, one of these photons must bounce off the electron, thus disturbing
the electron and changing the electron's original momentum. The exact
amount of change 6.p cannot be predicted, but it cannot be greater than
the photon momentum h/ >.. Hence
6.p ::; h/ >.
The longer the wave length, the smaller the uncertainty in momentum.
Because light is a wave phenornenon, we cannot expect to determine the
electron's position with perfect accuracy; however the uncertainty 6.x can-
not be less than >., i.e.
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.
QUANTUM MECHANICAL CONCEPTS 19
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.
2.2 Illustration of Heisenberg Uncertainty Principie ~
Consider an experiment in which~mentum P of an electron in the x-

direction is accurately known at the beginning and the position is measured.
We look at the electron through a microscope by means of light of wave
length >.. Each photon of this light has momentum P = h/>. = hv/c. In
order to 'see' it, one of the photons must be scattered from the electron,
thus disturbing the electron and changing its original momentum.
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
Fig. 2.1 Microscopic experiment
From conservation of momentum (see Fig. 2.1)

hv tu/ .
P+- = -sm8+px (2.3)
e e
where P is the original momentum of electron, Px its momentum in~ ~
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
direction, () the angle of scattering, v and v' ¡bei.¡:¡,~frequencies of the photon

before and after scattering because of Compton effect. Now () :e; e in order
to observe the electron (i.e. light goes through the microscope). Now we
can determine the position of the electron with accuracy (given by the
resolving power of the microscope)
>..'
6.x ~ - (2.4)
sin e
Now the quantum may have scattered anywhere within the angle ±E, there-
fore from Eq. (2.3) we see that Px the momentum of the electron in x-
direction has uncertainty
h .
Ó.Px = 2 )..' sm s. (2.5)
From Eqs. (2.4) and (2.5), we have

Ó.pxÓ.X ~ 2h >h .
2.3 Schrüdínger Equation
Since wave aspects of particles are not a result of classical mechanics, it

is not possible to derive the basic equations of quantum mechanics from
there. We are here dealing with a new field of physics. We shall however,
use the de Broglie rule as a guide to guess the basic equation of quantum
mechanics. The de Broglie rule is that with each particle of momentum p
is associated a wave of wave length >. given by
>. = 21r = !_!_ = _!!__ = 21r. (2.6)
k p mv k
The Einstein relation for energy E is
E = hv = nJ.JJ ( wave aspect). (2.7)
For a free non-relativistic particle
p2
E = - (particle aspect). (2.8)
2m
Thus,
E P2 !ik2
w=-=--=- (2.9)
!i 2mñ 2m
so that
dw !ik p
Vg =- =- =- =V
dk m m
QUANTUM MECHAN/CAL CONCEPTS 21
Consider for simplicity a particle moving in the x-direction. We now in-

troduce a quantity '1/J(x, t), called the wave function which specifies all that
can be known about the system it represents. We put two requirements:
(i) to find the physical interpretation of '1/J, i.e. to have a set of rules which
will enable us, knowing '1/J, to predict what results a certain measurement
will giv~ (ii) to find an equation which '1/J satisfies so that this equation
' 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,
'1/J(x, t) = ¡_: A(k)ei[kx-w(k)t]dk (2.10)
where w is given by Eq. (2.9). Hence the wave packet associated with the
particle1 is
'1/J(x, t) = ¡_: A(k)ei[kx-(ñk2/2m)t]dk. (2.11)
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
so that the differential equation

. 8'1/J n2 82'1/J
in-=----,
ot 2m 8x2
(2.14)
follows. This is the simplest form of the Schrodinger equation. This is
the basic equation of quantum mechanics and describes phenomena on the
1 Note the important difference. In quantum mechanics, the wave function 1/J(x, t) is in
general complex, whereas in classical electromagnetism, 'ljJ(x) represents a real wave.

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)
where the Hamiltonian H is:

fi ñ2 a2
H= T+V = + V ( x) = - 8x2 + V (x) (2.20)
2m 2m
In general we denote the momentum and position operators by p and
x. The important representation of these operators in x-space is provided
by Eq. (2.16) which is known as the Schródinger representation.
The correspondence principie will ensure that in the limit of the ob-
served system becoming large and the disturbance becoming negligible, it
must go over into classical mechanics.
2.4 Physical Interpretation of Wave Function
The wave funciton '1/J(x, t) has a statistical interpretation. It is related to

the probability, P(x, t) of finding the particle at the position x and time
t. This is because particle is associated with a wave packet which has a
spread. The probability is positive definite : P(x, t) 2:: O. Therefore P(x, t)
should be proportional to '1/J*(x, t)'ljJ(x, t). We take it
P(x,t) ='1/J*(x,t)'ljJ(x,t) = l'I/J(x,t)l2. (2.21)
Now
+00 P(x, t)dx 1+00 l'I/J(x, t)l2dx
1 _00 = -oo = 1 (2.22)
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.
From Schri:idinger'sequation, we have

. ª'1/J n2 a2
iñ at = ( - 2m ax2 + V(x) )'1/J. (2.24)
Taking its complex conjugate
. ª'1/J* n2 a2 *
-ili7ft = ( - 2m ax2 + V(x) )'1/J . (2.25)
From these two equations, we get
ª'1/J*
7j¡'I/J + 'ljJ *ª'8t1/J -- li ( 82,¡_,•
2mi '1/Ja;, - '1/J *ª2"')
ax2
- ....!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)
if 1(x, t) --+ O as x --+ oo. Hence the probability is conserved if we use

definition (2.21) of the probability.
Next, we define a quantity called probability current density denoted
by S(x, t):
S(x, t) = ~
2mi
ª 1*).
(1* ax1 - 1 ax ª (2.28)
Then from Eq, (2.30), we have

aP(x, t) ~S( ) = O (2.29)
at + ax x, t
which is analogous to the equation of continuity viz.,
ap V . (2.30)
at+ ·J=O
of electricity. It is because of this reason that S(x, t) given by Eq. (2.31) is
called Probability current density.
Remarks:
l. We have shown that the probability density
P(x, t) = l1(x, t)l2 (2.31)

and that
d
-d
¡DO P(x, t)dx = O (2.32)
t -oo
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
1= exp[i(kx - wt)] (2.33)

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)
px'lj;(x, t) = -iñ :x ( x'lj;(x, t))

. . o'lj;(x, t)
= -iñ'lj;(x, t) - iña: ax . (2.36)
From these equations we see that

xp=J px. (2.37)
Note whenever we say A= B, we mean that for all functions f(x)
Af(x) = Bf(x).
But for brevity we write A = B. In order to see how badly two operators
do not commute, we define a quantity called the "commutator":
[x,Pl = (xp - px). (2.38)
From Eq. (2.36), we get
[x,pJ'lj;(x,t) = iñ'lj;(x, t)
or
[x,Pl = in. (2.39)
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
QUANTUM MECHAN/CAL CONCEPTS 27
2.5 Stationary States
By stationary state of a system we mean one where energy is constant in

time. Now we have the Schrodinger equation
iñº'I/J1:• t) = H'lj;(x, t). (2.52)

We may look for a particular solution of the form
'1/J(x, t) = u(x)f(t). (2.53)
Substituting into Eq. (2.43), we have
iñu(x) %/(t) = Hu(x)f(t) (2.54)

where H Harniltonian is of the form
(
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) -
H( x, -iñ:x)u(x) - Eu(x). (2.59)

From the Correspondence principle, it followsthat the number E appearing
in Eq. (2.58) may be interpreted as the energy.
Equation (2.58) is such that Hamiltonian operator H acting on u does
not alter u apart from multiplying it by a number E. Such an equation
is an example of what is called an eigenvalue equation. Here E is the
eigenvalue of the Hamiltonian operator or simply energy eigenvalue, while
u(x) is called the eigenfunction belonging to the eigenvalue E and is often
written as uE(x). Thus we write Eq. (2.58) as
H(x,-iñ:x)uE(x) = EuE(x). (2.60)

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)
This is the Schródinger equation for stationary states for which

'1/JE(x, t) = uE(x)e-iEt/ñ, (2.62)
since solution of Eq. (2.58) is

f(t) = e-iEt/ñ. (2.63)
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)
i.e. the probability of finding the particle in the interval dx is constant in

time for a stationary state.
We can develop certain continuity properties of the wave function u(x),
which will help to salve the Schr6dinger equation (2.61). Let the potential
energy V(x) have only finite discontinuities. Since energy is finite and
we want the wave function u(x) to be finite, otherwise there would be
l- Q\ 4lL
a,.""íl.(tiJ,..._
trouble with the probability interpretation of the wave fupction. Therefore,
it follows from the Schr6dinger equation (2.61) that ~ must be finite.
-,¿) Hence i~ and u are continuous. As a result of the two wave functions
shown in Fig. 2.2 (1) is permissible while (2) is not.
(1)
Fig. 2.2 Illustration of permissible and not permissible wave functions.

QUANTUM MECHANIGAL CONCEPTS 29
2.6 Eigenvalu ,s and Eigenfunctions
An operator A is a mathematical entity which, in general acts on any

function of x and can be written as x, X( fx).
A simple example is the
operation of differ ntiation, A being any function of /J:.,
~
A= A
~( ox{)) ~ ~
e.g. A = p = -in
·te {)
ox . (2.65)
Another example is
~ ~( {)) 8
A = A x, ox = Dx x
which gives
(º~~x) x )1/J(x) = 1/J(x) + x :x

= (1 + x :x)1/J(x).
However, if to each operator x, X( fx) there belongs a set of numbers
an and a set of functions such that
A(x, :x)un(x) = anun(x) (2.66)
then an is called an eigenvalue of A and un(x) is called an eigenfunction

of A belonging to the eigenvalue an, The eigenfunctions of an operator are
thus those special functions which remain unaltered under the operation
of the operator, apart from a multiplication by the eigenvlaue. Equation
(2.66) is called the eigenvalue equation for the operator A.
Examples:
~ d2
A= dx2·
Eigenvalue equation becomes the differential equation
d2
dx2 u(x) - au(x) = O. (2.67)
We want to find a and u(x) so as to satísly Eq. (2.67) subject to the

conditions
u(O) = O
u(7r) = O (Boundary conditions).
Solution of Eq. (2.67) is

u(x) = C1 sin(Fax)
u(O) = O ==* C2 = O
+ C2 cos(Fax)
L.
L } '
u(n) =O==* O= C1 sin( Fan).
This is only true if Fa = n, n beinglinteger:
a= -n2•
herefore the only eigenvalues of the operator are
ªn = -n2
nd eigenvalue equation becomes
d2
dx2 Un(x) = -n2un(x)
where eigenfunctions are
fn(x) = C1 sin(nx) L o
Note that the eigenvalues depend very much on the boundary conditions
imposed on the solution to the eigenvalue equation (2.66). It is not nec-
essary that eigenvalues should always be discrete. They may be whole
continuum.
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 •
Find the corresponding eigenvalues.

2.5 Show that for a wave function represented by a plane wave

'lj;(r) = eip·r/fi. (2.68)
the probability current density S is given by
p
S(r,t) = - =v,
m
(
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.
Chapter 3
BASIC POSTULATES OF
QUANTUM MECHANICS
3.1 Basic Postulates of Quantum Mechanics
We have introduced two concepts: (i) the state of a particle or a quan-

tum mechanical system is described by a wave function or state function
1/J(r, t) which satisfies Schródinger equation; (ii) the dynamical variables
like momentum p and energy E are operators.
Now important properties of a physical system are quantities like p, E
which can be measured or observed; such quantities are called observables.
There must be a means of predicting the values of observables from the
state function and the procedure for doing this is given by followingset of
postula tes:
(1) To every observable there corresponds an operator A.
(2) The possible result of a measurement of an observable is one of the
eigenvalues an of A given by the equation
(3.1)
where A is an operator andan is eigenvaluecorresponding to eigen-
function Un.
(3) A measurement of A on a system in an eigenstate certainly leads
to the result an, the eigenvalue.
(4) The average value of a large number of measurements of an ob-
servable on a system described by an arbitrary state 1/J is given
by
(.A) = a,¡, =a= j 1/J* .A'I/Jdr (3.2)
provided that J 1/J*'I/Jdr = 1, and there exist suitable boundary conditions.

Thus for example, the average value of momentum in x-direction for a state
33
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
'lf;(x, t) is given by
p = j 'lf;*(x) ( - iñ :x)'l/J(x)dx. (3.3)
Applying to the special case of A= x, we get the average value of position

x =J 'lf;*(x)x'lf;(x)dx
=J x'f/;*'f/;dx = J xl'f/J(x)l2dx. (3.4)
This is consistent with our interpretation of 'f/;(x), that l'l/J(x)l2 determines

the probability density P(x) of the particle in space, since l'l/J(x)l2 appears
as the weighing factor appropriate to x in the calculation of the average
position.
It is convenient to introduce a compact notation for the matrix element
(<PIA'l/1) or (<PIAl'f/J) to mean J <P* A'lf;dr. Furthermore, the integration may
not always be over space and it may be necessary to imply integration over
other continuous or discontinuous variables.
3.2 Formal Properties of Quantum Mechanical Operators
The quantum mechanical operators (observables) possess certain properties

which are important and we discuss sorne of them briefly.
(a) They are linear, i.e. if (Cn are numbers)
(3.5)
n
then we have
(3.6)
n
(b) They obey the laws of association and distribution. Thus if A, B

and e
are three operators, we have
A(BC) = (AB)c (3.7)
A(B+C) = AB +le (3.8)
(e) An observable corresponds to a hermitian operator. We define the
hermitian conjugate or adjoint .At of an operator A by the equation
(3.9)
BASIC POSTULATES OF QUANTUM MECHANICS 35
i.e.
J 'lj;* A.cpdr = J (A.t 'lj;)*cpdr
An operator is said to hermitian if Xt = A, i.e.

('1/JIA</J) = (</JIA'I/J)* = (#1</J) (3.10)
Theorems:
(i) The eigenvalues of a hermitian operator are real.
(ii) The eigenfunctions of a hermitian operator corresponding to dif-
ferent eigenvalues are orthogonal:
We have eigenvalue equation
(3.11)
and then
~ *
(umlAun) = a~(umlun)* = a~(unlum)· (3.12)
But from Eq (3.9):
~ * ~t ~
(umlAun) = (umlA Un)= (unlAum)
= am(unlum)· (3.13)
Hence from Eqs (3.12) and (3.13):
(am - a~)(unlum) = O.
Therefore if
i) m=n, an=a~,since (unlun)-/=-0 (3.14)
ii) m-/=-n, (unlum) = J U~Umdr=O. (3.15)
If the eigenfunctions are normalized,
(unlun) = J U~Undr= l. (3.16)
Hence Eqs (3.15) and (3.16) can be written in a compact form :
(unlum) = J U~Umdr= Ómn (3.17)
where
(3.18)
m=n.
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
= I:: I:: cr.c;«; f u:nundr

m n
m n
(3.22)
m
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.
Since 'l/; is arbitrary, [A, .B] = O. Thus if a set of simultaneous eigenfunctions

of two observables A and B exist, then A and B commute.
(í) If [.A, .B] = O and [.A, C] = O but [.B, C] =/= O, then it is not possible
to fud functions which are simultaneously eigenfunctions of A, B and C.
It is only possible to find eigenfunctions for A and B or for A and C.
Corresponding to this, we can only make simultaneous measurements of
the observables corresponding to the pair of operators A and B or to the
pair X ande, it is not possible to measure ~together. b
/
3.3 Continuous Spectrum and Dirac Delta Functions
(a) Continuous Spectrum

W Q.. So far~ave considered the case when the eigenvalues of an operator A
are <lis r te. In case eigenvalues of an operator take on continuous values,
the sum in the completeness relation (3.19) takes the form of an integral
'l/;(x) = J C(a)ua(x)da, (3.28)
where the label a corresponds to continuous set of eigenvalues and replaces

the discrete label nin Eq. (3.19). For simplicity we first consider 'l/; to be a
..., function of a single variable x onlyl"\ the generalisation to three dimensions 1
V
is straight)orward. I
Now using (3.28)
J u;,(x)'l/;(x)dx = (ua' l'l/J)

= J daC(a)(ua,lua) , (3.29)
where a' lies in the domain of integration of a. Now the orthogonality

condition (3.18) becomes
J u;,(x)ua(x)dx = (ua lua)

1
= O when a=/= a' . (3.30)

For a = a', it does not have to vanish. In fact it must be infinitely large
at a = a' because if it is finite at a = a', then the integral on right-hand
side of Eq. (3.29) vanishes but 'l/;(x) <loes not vanish in general. But this
infinity must be such that
J u;,(x)'l/;(x)dx = C(a') J da(ua,lua) , (3.31)

so that in analogy with Eq. (3.20), we have
C(a') = j u;,(x)'lf;(x)dx (3.32)
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)
Fig. 3.1 The Dirac delta function.
Such a function is called Dirac ó-function and is written as

ó(a- a')= O for a =f. a'
= oo for a= a' (3.34a)
such that
J daó(a - a')= 1 . (3.34b)
Thus we write
(ua lua) =
1 J u~,(x)ua(x)dx
= ó(a - a') , (3.35)
corresponding to orthonormality relation (3.17) for discrete set. Substitut-

ing Eq. (3.38) into Eq. (3.29), we have
j u~,(x)'f/J(x)dx j daC(a)o(a - a') .

=
Comparing it with Eq. (3.35), we have
j daC(a)o(a - a')= C(a') . (3.36)
This is the fundamental property of the 8-function which we require.

The generalisation to three dimensions is obvious
x _____,, r, a _____,, a, a' _____,, a'
Thus Eqs. (3.28), (3.32), (3.34), (3.35) and (3.36) respectively become
'f/;(r) = J C(a)ua(r)da (3.37)
C(a') = J u:,(r)'f/;(r)dr (3.38)
{~:-a') =0 for a =1- a' ~

for a= a'.
(3.39a) l
j dao(a-a') = 1 ~ (3.39b) ~
(ua lua) = o(a - a')~

1 (3.40) j
j C(a)o(a - a')da = C(a') . (3.41)
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
= J C(a)ua(r)da (for continuous set) . (3.43b)
Then
C; = Ju~ (r')'f/;(r')dr' (3.44a)
C(a) = J u;(r')'f/;(r')dr' . (3.44b)
Substituting in Eq. (3.43a), we have for discrete set
1/J(r) = ¿ ( J u~(r')1/J(r')dr')un(r)
n
= J (¿ n
un(r)u~(r') )1/J(r')dr' .
Thus it follows that

¿ un(r)u~(r') = ó(r - r') . (3.45a)
n
The corresponding relation for continuous set is
Ir:~ J u0(r)u0(r')da = ó(r - r') . (3.45b)
~~hese ar known a ~losure Relations. These are equivalent to completeness

relations, since one dan writ ( e.g. discrete set)
1/J(r) = J 1/J(r')ó(r - 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
(e) A Simple Representation of ó-function Consider the function
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 ·
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)
Its generalisation to three dimension is
-1
(21r)3
-f eia·(r-r')da = ó(r - r')
.
(3.47b)
(d} Properties of 8-function
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
J ó(a - x)dxó(x - b) = ó(a - b) .
These equations have meaning only in the sense of integration; for example,
the first one means
J f(x)8(x - b)dx = J(b) .

ó(bx) = lil ó(x) means
J ó(bx)dx = m.
(e) Fourier Transform
lfompleteness relation, in terms of eigenfunctions keiax becomes
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
f(r) = (27í~3/2 J C(a)eia·rda (3.50a)
1
C(a) = (27r)J/2. ¡· f(r)e-ia·rdr.. (3.50b)
(f) Momentum Eigenfunctions (An Example of Continuous

fü Spectrum of Eigenvalues
~~ ~mentum operator in Schrodinger representation is
~
p= -i
'li - f) .
@ ,
~genvalue equation is
&
pup(x) = pup(x)
-ili! up(x) = pup(x). (3.51)
A solution of this equation is
up(x) = Bexp (~px) . (3.52)

Here eigenvalue p is a continuous variable and takes on any value. Above

we have taken the momentum in x direction. For three dimensions
p = -iñV
and
pup(r) = pup(r)
up(r) = Bexp (~P · r) (3.53)
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)
Thus normalised momentum eigenfunctions are
u (r) = 1 e(i/n)p·r (3.55)

p (21rñ)3/2
Since these eigenfunctions form a complete set we can write
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
'1/;(r) and C(p) are the Fourier transform of each other.

Now average value of the momentum operator is given by
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
= (2:n)3 fff dp'dpC*(p')C(p)

xpe(i/ñ.)(p-p')·r dr
= ff dp'dpC*(p')C(p)p8(p - p")
= J IC(p)l2pdp · (3.57)
Thus IC(p)l2 is the probability of momentum operator p having eigen-

value p when the system is in state '1/J(r). In other words, in a measurement
of the momentum of a particle, the probability of finding the result p is
IC(p)l2. Thus C(p) may be regarded as the wave function, in momen-
tum space justas 1/J(r) is the wave function in r space. C(p) is sometimes
written as <;b(p).
3.4 Uncertainty Principie and Non-Commutativity of Ob-

servables
In quantum mechanics, two observables A and B do not necessarily com-

mute and obey a commutative law, i.e. in general
[A,B]#O. (3.58)
This statement is essentially equivalent to the uncertainty principie which
expresses the limi!:tions on our knowledge impo ed by mutual disturbances
of observations. ~e{ Eq, (3.58) i_rrplies t!:!_at ~it is not possible to
find simultaneous eigenfunctions of A and B i.e. we canno have an exact
knowledge of the result of measurernent of A and B simultaneously.
We now show explicitly that Eq. (3.58) leads to the uncertainty prin-
ciple. A and B being observables are hermitian. Let a and b denote the
average values of large number of measurements of A and B respectively.
Define
A-a= 8A= a
Íi- b = 8Íi == fJ. (3.59)
July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
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
(6.a)2 = ((A- a)2) = (e?)

(6.b)2 = ((B - "6)2) = (/32) (3.60)
Now
(6.a)2 = (a2) = J '1/J*(x)am/;(x)dx
= j(at'l/J)*a'l/)dx
= j (a'l/J)*a'l/)dx
= j la'l/Jlzdx.
Then
(a2) (/32) = ( j la'l/Jl dx) ( j l/3'1/Jl dx)

2 2
~ 1 j (a'l/J)* /3'1/Jdxlz . (3.61)
This follows from the Schwartz inequality
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
= j '1/J* a/3; f3a ibda: + j '1/J* a/3; f3a 'l/)dxl2

J
= JP - iQr (3.62)
BASJC POSTULATES OF QUANTUM MECHANICS 47
where
p = J 'l/¡* a{3 ; {3a 'l/¡dx
Q = j 'l/¡*ia{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
= 1 j '1/J*iAB; BA'l/Jdxl2 (3.63)
~11rln since from Eq. @-59) af!. - {3a = AB - BA.

particular if A= P, B = x,
[p, x] = -iñ
(~a)2 = (~p)2
(~a)2 = (~x)2
(~p)2(~x)2 ~ ~ 1 j 'l/¡*'l/¡dxJ2 = ~
The root mean square deviation is often called the uncertainty (standard
deviation), i.e.
~p = Jc~p)2 , ~X= Jc~x)2 . (3.64)

Hence
(3.65)
where ~p and ~x denote the uncertainties in the measured values of p
and x. If
[A,B]=O
there is no mutual disturbance, and the result of simultaneous measure-
ments of observables A and B can be known exactly.
Example
1
r
Momentum and energy operators for a free particle are given by
~ ~ ·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
which is true for an arbitrary function v,(x). Thus
[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
'
Hence show that

(~x)2 = ((x - (x) )2)
02
=
2
(~p)2 = ((p- (p))2)

1 li2
=
2 02
so that
1
~x~p=
2n.
(ii) Using the relations
(p) = J pJ</>(p)J2dp
(p2) = J p2J</>(p)J2dp
show that
3.3 Show that for a particle of mass m, moving in a potential V(r),
[H, r] = -ili.!!__
m
(p, H] = [p, V(r)] = -ili"vV
Using the above results and the fact that H is hermitian, show that
m!(r) = (p) l '

d
dt (p) = -(VV) .
(Note that Newton's law is valid for expectation values.)
3.4 Using the result
[H, x]-in]!_ ,
=
m
for a particle of mass m moving in x-direction in a potential V(x),
show that the average value of its momentum in a stationary state
with discrete energy is zero.
3.5 A particle is in a state
1/J(x) = )a sin (5:x), lxl ::=:; a

=O elsewhere,
show that the probability for the particle to be found with momen-
tum p is given by
l001ra 2
sin (pa/ñ)
2n (p2a2 ¡¡¡,2 _ 2fr,r2)
3.6 A particle of mass mis confined by an infinite square well potential
V(x) = O, lxl ::=:; a; V(x) = oo, lxl > a. If the particle is in the state
1/J(x) = x, lxl ::=:;a,
1/J(x) = O, lxl >a,
find the probability that a measurement of energy wíll give the
result
¡¡,2 2
En= 2m c:a2)n2.
3.7 For a free particle, find a wave funcion which is a simultaneous
eigenfunction of both momentum and energy. This is not so for a
particle moving in a potential since then IP, HI f=. O [cf. Problem
3.3].
3.8 a) If a particle is in a state
2 ) t ,,2
1/J(x) = ( 7rt52 e-~
find the probability of finding it in the momentum eigenstate
-
up ( x ) - y{12rli,e _ipx
x .
b) If it is in a state
uv(x) = f!¡. sin nx,
show that the probability of finding it with momentum pis given

by
Chapter 4
SOLUTION OF PROBLEMS IN
QUANTUM MECHANICS
We are now in a position to undertake the solution of physical problems in

quantum mechanics. A typical problem in quantum mechanics is to find
the allowed energy levels for a given system, which involves the following
three steps: (i) What is the potential (physics)?, (ii) Write the Schrodinger
equation and solve it (mathematics), (iii) Results and their interpretation
(physics).
First we shall consider a potential which is constant, i.e. V(x) = V
( constant). It may be that V is a different constant in different regions of
space.
4.1 One Dimensional Potential Step
The potential step is represented by the function
= V for x > O
V ( x)
V(x) = O for x < O.
L ( 4.1)
Let E denotes the total energy of the particle. Then
E= T(x) + V(x) = constant. (4.2)
First we discuss the motion classically. As shown in 'g. 4.1 here

are two cases: (i) E - V(x) > O . This means that a particle with kinetic
energy T = E = p2 /2m approaches the potential barrier from the left.
As the particle moves across the barrier, it enters the region of potential
V(x) = V. The kinetic energy T1 of the particle in this region is given by
T1=E-V=-. P?
2m
51
E0 - - - - - - - - - - - - - - - - - - - - - - (i)
E0- - - - - - - - - - - - - - - - - - - - - - (ii)
Fig. 4.1 The energy diagram of a potential step
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.
Here again we consider two cases: (i) (E - V) >O. We define constants ko

and k1 by the equations
k2 _ 2mE _ 2m(E- V)
O - ñ,2 '
k 21 - ñ,2 . (4.4)
SOLUTION OF PROBLEMS IN QUANTUM MECHANICS 53
From Eq. (4.3) we have

a2
(0x2 + k5)uo(x) = O x <O (4.5a)
a2
(0x2 + k?)u1(x) = O x > O. (4.5b)
The general solutions of Eq. (4.5) are

uo(x) = Aeikox + Be-ikox (4.6a)
ui(x) = Deik1x + oe==, (4.6b)
The first and second terms on the right-hand side of Eq. ( 4.6) represent
plane waves moving in positive and negative directions of x-axis respec-
tively.
We are interested in the situation where the particle approaches from
the left and may be either transmitted or reflected. Thus,
\~ uo(x)~-
~
incident )
+~
(reflected)
L (4.7a) i
q U1 (x) D eik1x
'-v--"
. (4.7b)
( transmi tted)
The constant C must be zero as there is no wave in negative direction of

x-axis in the region x > O.
The continuity conditions at x = O require
uo (O) = u1 (O) ; u~(O) = u~ (O)
which gives us
A+B=D (4.8a)
and
(4.8b)
i.e. we have two equations for three constants. One constant can be deter-
mined from the normalisation condition
¡_: lu(x)l2dx = l. (4.9)
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)"
A can be determined from the normalisation condition but it is not very

interesting. In order to find reflectivity and transmitivity of the potential
step, we have to find the incident, transmitted and reflected probability
currents.
Incident Current: (Subscript i denotes incident)
Si(x) = ~
2mi
(u'!' OUí
i 8x
- Ui out)
8x
(4.11)
= ~ (e-ikoxikoeikox _ ikox (-iko)e-ikox) IAl2

2mi
= ~klAl2 (4.12)
m
= Incident flux (pvo, p = IAl2 = density of states).
Transmitted Current:
St(x) =
n 2 k1
-IDI
m
n: 4k5 2
(4.13)
m (k + k1)2 k1IAI .
Reftected Current:
n
Sr(x) = --IBI ko
2
m
= -~ (ko - k1)2 kolAl2. (4.14)
m (ko + k1)2
Transmitivity
T= IDl2k¡
IAl2ko
Llkok1
(4.15)
= (ko + k1)2 ·
Refiectivity
IBl2
R= IAl2
(k - ki)2
(4.16)
(k+k1)2"
Therefore
R+T= (ko-k1)2 + 4kok1 =l
(ko + k1)2 (ko + k1)2 •
SOLUTJON OF PROBLEMS IN QUANTUM MECHANICS 55
Now ko is always greater than zero since E > V. k1 however, can be

zero and in that case T = O. As E becomes smaller and smaller, the
transmittivity goes down and when E = V, then T = O, and R = 1 and
everything is reflected.
In general R is not zero. lt is in this respect that the quantum me-
chanical result differs from the classical theory. If E » V, then ko ::::::: k1,
B = O, D = A, and R : : : : O and everything is transmitted - which is, of
course, the classical limit. In general, a part will be reflected and a part
will be transmitted. In classical mechanics there is no reflection at the
boundary if E > V, but in quantum mechanics there is reflection. The
extreme quantum limit is for E« IVI (i.e. when kr = (2m/h2)(E + IVI)
and k1 » ko). In this case as is clear from Eq. (4.10) B::::::: -A and D::::::: O
and there is total reflection. Thus most interesting case is to take V large
in magnitude, but negative (Fig. 4.2) so that we have a sudden large po-
tential drop, through which classical particles pass with greatly increased
momentum. However, in quantum mechanics, as seen above their is total
reflection - the exact opposite of the classical prediction. This essentially
quantum mechanical effect may be observed in nuclear physics when a low
energy incident neutron, say, is reflected by the sudden onset of)tmf highly ""'
attractive potential, as it approaches the surface of a nucleus.
- - - - - - - - - - - - - - - - - -Eo
X
Fig. 4.2 The energy diagram of a potential step with negative V

(ii) (E - V) < O, for x > O. For x < O, we have, as befare

uo(x) = A eikox + B e-ikox x < O. (4.17)
For x > O, we put
K2 = 2ñ:1 (V - E) .
Then we have the Schródinger equation as
(::2 - K2)u1(x) =O x>O

and the solution is
(4.18)
As x--+ oo, the second part of (4.18) is infinite, Since the wave function is
to be finite, D must be zero. Therefore
U1(x) = e e:">. (4.19)
Continuity conditions at x = O give L
A+B=C (4.20a)
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
uo(x) = A[eikox+ e-2io(E)e-ikox¡

= 2Ae-iº cos(kox + ó) (4.26)
u1(x) = Ae-i(.;--o)e-Kx = -Aieiºe-Kx. (4.27)
In deriving Eq (4.27) we have used Eqs (4.21) and (4.25). Thus we
conclude that the reflected wave is shifted in phase, at x = O from the
incident wave at x = O by a factor e-2ia(E).
4.2 The Potential Barrier
We now discuss the quantum mechanical analysis of a particle going across

the potential barrier as shown in
V(x)
.-a-- ....
--------- --- ---------
(O) (1) (2)
Fig. 4.4 The rectangular potential barrier.
The potential is given by

V(x) = O x : '.'.'. O Region (O)
= V O< x < a Region ( 1)
= O x ::=:: a Region (2). (4.28)
The treatment for the case when E > V is similar to that carried out for
the potential step. We, therefore, discuss only the case when E < V.
Define
k2 = 2;,;E , K2 = 2; (V - E) .
The Schrodinger equation in the three regions is
[::2 + k2]uo(x) = O, x::; O
[:;2 - K2]u1(x) = O, O< x < a
[::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
= u1(0), u1(a) = u2(a) l

These give
u0(0)
u~(O) = u~ (O), u~ (a) = u;(a). '
A+B=C+D (4.30a)
ik(A- B) = -K(C - D) (4.30b)
e e-Ka+ D eKª = A' e-ka (4.31a)

-KC e-Ka+ KD eKª = ikoA' eika. (4.31b)
From Eqs. (4.31), we get
and
D = ~1 (l + tr:: (4.32a)
(4.32b)
From Eqs.
(4.33a) J
(4.33b)
= u; then from Eqs. (4.32) and (4.33)

L
Let us put K/k
A= !A' ikª!(2µcoshKa - i(l - µ2) sinhKa) (4.34a) )
2
1
B = -A'
2
eika
µ
-i .
µ
(1 + µ2) sinhKa L (4.34b) C)
Let us put
F = (2µcoshKa - i(l - µ2) sinh Ka)
Then from Eqs. (4.32) and (4.34), _we have L l (4.35) o
A A 2µe-ika / F (4.36) )
A(-i)(l+µ2)sinhI<a/F L (4.37)
e A(-i)(l + iµ)eKª I F
(i)(l - iµ)e-Ka / F.
L (4.38a)
(4.38b)
1
Now reflectivity R and tra' smittivity T are given by
\B\2 (1 + µ2)2 sinh2 Ka )
R = IAl2 = [4µ2 cosh'' Ka+ (1 - µ2)2 sinh2 Ka]~
(4.39) )
IFl2 4µ2
( 4.40)
T = IAl2 = [4µ2 cosh'' Ka+ (1 - µ2)2 sinh2 Ka]·
First we note that when E < V, then the region (1), i.e. O < x < a is
inaccessible classically and thus T = O and R = 1 classically. But here we
see that in general T -/=- O and there is a leakage of the particle through the
barrier. The fact that something is transmitted at ali is a purely quantum
mechanical effect ( this is called the tunnelling through the barrier).
If the barrier is broad, so that Ka » 1, then
1 4
sinh2 Ka - e2Kª + e-2Ka - 2
_, 4e-2Ka
and
T = 4µ,2 / inh2 Ifo

(1+µ2)2+ .¡t2/sinb2 Ka
¡:::j (4µ/(1 + µ2))2e-2Ka. ( 4.41)
Using µ2 = K2 /k2 = (V - E)/ E, we obtain the transmitivity or the tun-
nelling probability when the barrier is broad as
T(E) = 16i ( 1 - i) exp ( - v~~m- 2_(V E_) a). (4.42)

SOLUTJON OF PROBLEMS IN QUANTUM MECHANICS 61
Thus we see that in quantum mechanics, it is possible for the particle to

penetrate the barrier although its energy is less than the potential of the
barrier. From Eq. ( 4.42), we see that the transmission probability is much
less than one, and decreases exponentially with the increase of the width
of barrier. This has important applications e.g. o-decay where o-particle
having less energy than the potential barrier of the nucleus comes out of the
nucleus. The shape of such a potential and the shape of the wave function
is shown in ig. 4.5.
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
Suppose we have a potential such that

V(x) = V(-x) ( 4.44)
that is, the potential is symmetric about the ongm and then H(x) =
H(-x). Let uE(x) be the solution of eigenvalue equation
ñ2 82 )
H uE(x) = ( -
2m
ax2 + V(x) uE(x)
= EuE(x). ( 4.45)
Now changing x -+ -x, we have
f¡,2 {)2 )
( - 2m 8x2 + V(-x) UE(-x) = E UE(-x) (4.46)
so that using Eq. (4.44),
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
UE(x) = r]UE(-x). (4.48a)
Changing x-+ -x
(4.48b)
Substituting Eq. (4.48b) in Eq. (4.48a), we have
UE(x) = 7]2UE(x). ( 4.48c)
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)
each of which has a definite parity. Thus it is always possible to construct

solutions which are either symmetric or anti-symmetric about the origin
provided that V(-x) = V(x).
4.4 An Example of Bound State (Discrete Energy Levels)
In the previous examples we have considered potentials in which the particle

could move off to infinity at least in one direction both classically and
quantum mechanically. We now consider examples of a bound system in
which this is not the case. Consider a potential of the form (see Fig. 4.6)
V(x) = O, !xi~ a
V(x) ----+ oo, lxl ~ a. (4.50)
Classically the particle is confined to the region !xi ~ a, and whatever
its energy, it bounces elastically off the potential walls. Since V becomes
infinite for !xi ~ a, but other terms in the Schródinger equation
r,,2 a2 )
( - 2m 8x2 +V u = Eu
for this region remain finite, we must impose the boundary condition
u(x)=O lxl~a. (4.51)
+ 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
Thus the boundary condition at a surface at which there is an infinite

potential step is that the wave function is zero.
For !xi < a, the Schródinger equation is
ñ,2 d2u 2mE
---2m dx2 = Eu, k2 = --ñ,2 (4.52a)
or
(4.52b)
The general solution of this equation is
u(x) = A sin kx + B cos kx ( 4.53)
Boundary condition at x = ±a gives L

u(a) =O=' (-a)
(4.54a)
i: (4.54b)
Therefore either
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)
Thus a classically bound system when treated quantum mechanically has

yielded discrete energy levels. This is a general feature as we shall see.
1 July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
- -- - l
f
1
- - - a X
T Vo
1
1
1
1
1
l
1
1
1
(1) (2) (3)
Fig. 4.7 The energy diagram of a square well potential.
4.5 Bound State (Square Well Potential)
The square well potential is an important concept in physics. The square

well potential may be a reasonable approximation to the actual potential
in sorne physical problems. Therefore a solution of this problem will give
qualitative results for the actual problem.
Now
V(x) = -V - a «; x < a
O elsewhere. ( 4.57)
Here we consider the case E = T > O for lxl > a but E < O in [z] < a
so that the particle is confined in the potential well in the region [z] ::::; a
i.e. the particle is in a bound state (E = T - V < O or T = E+ V < V
for !xi < a). [The other case when E > O in both the regions lxl > a and
!xi < a is similar to the cases considered in Secs. 4.1 and 4.2 (see Problem
4.4).]
Let
€=-E = IEI.
Define the wave numbers
2mE 2mé
K 2 = -~ = f¿2 (for regions (1) and (3)) (4.58)
k2 = 2m(E + V) 2m(V - é) (for region (2)). ( 4.59)

fj,2 fj,2
The Schródinger equation for eigenfunctions and eigenvalues is

n,2 {)2 )
( -
2m
ax2 + V(x) u(x) = Eu(x). (4.60)
This can be written for regions (1), (3) and (2) as
L
6
(::2 - K2)u1,a(x) = O (4.61a)
C~2 + k2)u2(x) = O L (4.61b)

The following are solutions:
ua(x) = Ae-Kx L
(no term of the form eKx should appear since the wave function must not
(4.62a) )
become infinite for large distances, i.e. x --+ oo)

u2(x) = Beikx +ce-ikx L (4.62b)
u1(x) = DeKx_ (4.62c)
(no terms of the form e-Kx should appear since wave function must be
finite for x--+ -oo). We can write u1(x) as
u1(x) = De-Klxl_ (4.62d)
The potential ( 4.57) satisfies
V(x) = V(-x) ,
i.e. it is symmetric or even and we may confine ourselves to eigenfunctions
of definite parity. Above solutions do not have definite parity, i.e. they are
not of the form
u(x) = ±u(-x) .
Consider first u2(x). The solutions oí definite parity, i.e. of even and odd
parity are respectively
1
2(u2(x) + u2(-x))
U2e(x) =
= ( B + C) cos kx = B' cos kx l (4.63a) )
1
U2o(x) = - u2(-x))
2(u2(x)
= i(B - C) sin kx = C' sin kx. (4.63b)
The requirement of parity for u1 and u3 is
ua(-x) = ±u1(x) (4.64)
1 July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I
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)
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
4.1 A particle of mass m, moves in a potential

V(x)=oo x:$0
V(x) = O O< x < a
V(x) = V x ~ a
For E< V, find the condition for allowed energy values.
4.2 A particle of mass m approaches a potential barrier
V(x) = O x <O
=V OSxSa
= Ü X> a
from x = -oo. For E < V, determine reflection and transmission
coeffi.cients for Ka« 1, where
K2 = 2m(V-E)
f¿2 .
4.3 Consider a particle of mass m in a potential

V= O x<O
= -V OS x S a
Ü X> a
If V -+ oo and a -+ O, such that V a = µ, find the energy for the
bound state (E< O).
SOLUTION OF PROBLEMS IN QUANTUM MECHANJCS 69
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.
Chapter 5
SIMPLE HARMONIC
OSCILLATOR
(( ~The career of a young theoretical physicist consists of treating the har-

monic oscillator in ever-increasing levels of abstraction." - Sidney Coleman.
5.1 Introduction
In classical mechanics, a harmonic oscillator consists of a particle moving

under the action of a restoring force
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
E~· ~mw'a' L (5.5)

The simple harmonic potential is important in reality for the following
O
reason:
71
Suppose a potential V(x) has a minimum somewhere, say at x = x0•

We can then write the potential V(x) as a power series around x0
8V(x) 82V(x) (x - x0)2
V(x) = V(xo) + ~lx=xo(x - xo) 2
+ (5.6) élx2 lx=xo
Now V(x0) is a constant anda constant <loes not matter in a potential and
can be put equal to zero. But 8~~x) lx=xo = O because V(x) has a minimum
at x = x0, therefore
V(x) = K + C(x - x0)2 + C'(x - x0)3 ... (5.7)
If we are interested in the neighbourhood of x0, higher terms will be small
and hence
V(x) ~ C(x - x0)2 (5.8)
Thus in this case it is a good approximation to approximate V(x) by
C(x - x0)2 which is a harmonic oscillator potential.
5.2 Quantum Theory of simple Harmonic Oscillator
The hamiltonian for simple harmonic oscillator is

·2 1
2m 2
H = F: + -mw2 2 x
(5.9)
where in the classical Hamiltonian, the dynamical variables p and x have
been replaced by operators (correspondence principie). In Schrodinger pie-
ture:
X=X,
A
p=
•ts
-z,iax·
a (5.10)
It is convenient to change the variable x to y:
fmw
y= y----¡¡:-x. (5.11)
Using this variable:

ñw ¿¡2 2
H=-(--+y ). (5.12)
2 ély2
Define, the operators:
at = J 2~ñw(-ip+mwx) = ~(-:Y +y) L (5.13) )

a = /1(,ipA + mwxA) = J2
y~ 1 ( a )
oy + y . (5.14)
SIMPLE HARMONIC OSCILLATOR 73
Now
(5.15)
Since w(y) is arbitrary;

(5.16)
In terms of the operators a and at:
nw 1
H = -(aat +ata) = nw(aat - - )
2 2
= nw(ata+ ~). (5.17)
From Eqs. (5.16) and (5.17):

[H,a] = nw[ata,a] = nw{at[a,a] + [at,a]a}
= =luoa. (5.18a)
Similarly
(5.18b)
Equations (5.18a) and (5.18b) are useful to determine eigenvalues and eigen-
states of H:
(5.19)
Suppressing y, write the state
(5.20)
Using the notation introduced in (ch. 3):
(xlx) = (aunlx) = (unlatx)

= (unlataun)
1 1
= nw (unl(H -
2nw)un)
1 1
= nw (En - (5.21)
2nw)(unlun)·
The eigenstates of H, viz un(Y), forma complete orthonormal set:
(5.22)
Hence the state x(y) can be expressed in terms of them:

(5.23)
n
and
n m
(5.24)
n
Hence from Eq. (5.21):

1
(En - 2: O
2nw)
1
En 2: (5.25)
2nw.
Thus En has a mínimum value Eo = !nw and let the corresponding eigen-
function be 'Uo(y). E0 = !nw i called the zero point energy. Classícally it
is zero, It is important to note that the vacuum (ground state) energy rAi j--
~in._M zero has far reaching consequences and is known a· the quantum
fluctuation of the vacuum.
Now
Huo(Y) = Eouo(y). (5.26)
On using Eq. (5.18a)
Hauo(Y) = (-ñwa+aH)uo(Y)
= (Ea - ñw)aua(y). (5.27)
If aua(Y) is not zero, then aua(Y) is an eigenstate of H with eigenvalue
Eo - ñw which contradicts that Ea is the mínimum value. Hence
aua(Y) = O. (5.28)
Eqs. (5.27) and (5.28) show that a acts as an operator which annihilates
energy in the system by tu» and is called the annihilation operator.
However, using Eq. (5.18b)
Hatua(Y) = (nwat +atH)ua(Y) = (Ea+ñw)ua(Y)
3
= ( 2nw)atua(y). (5.29) /
Hence if atua(Y) 1- O, it is an eigenstate of H with eigenvalue Ea+ Iu» = {:;,

!nw. Thus at acts as an operator which creates energy in the system by
ñw and is called the creation operator. Continuing this process n-times, we

get
(5.30)
i.e. (atrua(Y) is an eigenstate of H with eigenvalue (n + !)ñw. Hence

Un(Y) = Nn(at)nuo(Y) (5.31)
where constant N¿ is to be determined by norrnalization such that
(unlun) = l. Hence
Hun(Y) = Enun(Y)
1
En= (n+ n = O, 1,2 ... (5.32)
2)ñw
i.e. simple harmonic oscillator has discrete eigenvalues in units of tu», First
we note (on using Eq. (5.18b)):
Haun(Y) = (-ñwa + aH)un(Y)
= (En - ñw)aun(y). (5.33)
Hence if aun (y) -1- O, it is an eigenstate of H, with eigenvalue
1
(En - ñw) = En-1 = [(n - 1) + 2]ñw. (5.34)
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)
Now (cf Eq. (5.31))

Un(Y) = Nn(atrua(Y)
= Nn(at)n-latuo(Y)
= Nn(atr-1u1(Y)
= Nn(at)n-2atu1(Y)
= NnVl.2(at)n-2u2(y)
= Nnv'l.2.3(at)n-3u3(y).
Continuing this process, we get
Un(Y) = Nnv'l.2.3 ... nun(Y)
= n; Vnlun(y). (5.41)
Note in deriving Eq. (5.41), we have repeatedly used Eq. (5.40). From the
normalization:
we get
1
Nn = .JnIº (5.42)
Hence
(5.43)
In order to obtail,iexplicit form of un(Y), we use following procedure. We

note that the eigenvalue equation
Hun(x) = Enun(x) (5.44)
gives, on using the Eqs. (5.12) and (5.19),
d2Un 2)
dy2 + ( En - Y Un =O (5.45)
where En = ~ En, so that

En= (2n + 1). (5.46)
For large values of y,
(5.47)
l_
o
2
which has solution e=F. Therefore we write
1 2
un(Y) = e-,y Hn(Y) (5.48)
Substituting back in Eq. (5.45), we obtain

H~ - 2yH~(y) + 2nHn(Y) = O. (5.49)
This is known as Hermite equation and its solutions are known as Hermite
polynomials.
For n = O,
~ 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)
Continuing this process, we get

1 2
Un(Y) = AHn(y)e-,Y (5.54)
where Hn(Y) is a polynomial of degree n. From the above equations, we

note that
Hn(-y) = -Hn(Y)
Hn(-y) = Hn(Y) (5.55)
To conclude: for a harmonic oscillator, the energy is quantized; the energy

eigenvalues are given by
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
e-Y2 Hn(y)Hm(y)dy = ( V1f2nn!)8nm (5.58)

-oo
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.
5.3 Properties of Hermite Polynomial
i) Hermite polynomial has a generating function

S(y, s) = ey,-(s-y)2 = e-s2+2sy
= ~ Hn(Y) sn. (5.59)
L n!
n=I
Taylor expansion of S(y, s) gives:
8 8n sn
S(y, s) = S(y,O) + -s8S(y, s)ls=os + ... + -
s"
8S'(y s)ls=o-1 ...
n.
Thus
(5.60)
From Eq. (5.60):
Ho(y) = 1
H1(Y) = 2y
H2(y) = 2(-1 + 2y2)
ii) Recursion relation

Hn+l - 2yHn + 2nHn-1 = O.
Differentiating Eq. (5.59) w.r.t s:
+ 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)
Now left-hand side
Joc
S(y, s)S(y, t)e_Y2 dy = e2st J

oc
e-(s+t-y)2 dy
-oc
= e2st,/ii
= '-~ (2str (5.63)
v1r L n.1 ·
n
(5.64)
-(X)
u,
Fig. 5.1 Wave functions of stationary states of the harmonic oscillator far n = 1 and
n= 2.
iu/
Fig. 5.2 Position probability density in the states n = O, 1, 2.

average of
lu n12
Fig. 5.3 Position proba.bility density in a. state with large n.
5.4 The Average Values
First, we note from Eqs. (5.13) and (5.14).
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
The average value of energy in the state w(x, t) is given by
(E)= E= j w*(x, t)Hw(x, t)dx
= ¿¿ czc,« i(En-Ekl• j u~(x)Huk(x)dx

n k
~ ~ i(En-Ek)t
= e: e: C~Cke Ek8nk
n k
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):
(x) =x= j w*(x, t)xW(x, t)dx
= L L C~Ckei(n-k)wtXnk
n k
where
Xnk = j u~(x)xuk(x)dx = (unlxuk)

= {-;f;(unl(a + at)uk)·
On using Eq. (5.65). From Eqs. (5.39) and (5.40).
auk
atuk =
= vkuk-1
vk + luk+l·
L )
Hence, we have
and
(x) =X=¿¿ C~Ckei(n-k)wt¡{;f¿( v'n + lók,n+l + VrLók,n-1)]

n k
= {;[;(¿ C~Cn+1e-iwtv'n +1+L C~Cn-1eiwtvn)

n n
where we have written C¿ = ICnlé5n.

We conclude that Xnk can be expressed as a matrix, but x is a sum of
functions, each of which represents a classical motion for a simple harmonic
oscillator.
5.5 Problems
5.1 Show that
(x)n = J u~(x)xun(x)dx = (unlxun) = O.
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:
Hint: First show that

(unl(a + at)2un) = (2n + 1)
(unl(a - at)2un) = (2n + 1).
1
b.xb.p = ñ(n +
2)
b.x = [((x - (x))2)]!
L
b.p = [((p- (p))2)]!
5.5 Show that

Pnk = -iVmwñ[vn + lók,n+l - vnók,n-1]
)
n
Further show that
(x)t = (x)o+ ~~ i~ ~p~

(P)t =
é- 7
(p)acoswl - mw(x}osiriwt.
in complete correspondence with classical equations.
5.6 Verify that
\J!(x, t) = yfi3 [( V7f)

{3 1
2 e-2-
-iwt
+ ( 2(3
V7f)3 1 -i3wt
2 e-2-Je_2_
-/j2x2
is a solution of time dependent Schródinger equation for a particle

of mass m in a simple harmonic oscillator potential. Calculate
expectation values of x and p and show that
dx i5
=
dt m
Hint, note that lst and 2nd term in \J!(x, O) are normalized eigen-
functions of the time independent Schrodinger equation H\J! = E\J!
with eigenvalues n;
and 3~w respectively.
5. 7 A particle of mass m moves in a potential

1
V(x) = x >O
2mw2x2
=oo x<O
find the energy eigenvalues.
SIMPLE HARMONIG OSCILLATOR 85
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
__. <Po(P) = (-1-)i e-2!:w

nmiu»
</J1(P) = (i 1r
l
(mñw)3
) ~ e-.!f e-2!:w.
The phase factor can be omitted.

Chapter 6
ANGULAR MOMENTUM
6.1 Introduction
Classically we define angular momentum by the equation

L = r x p. (6.1)
In quantum mechanics, by the correspondence principle L is an operator
L= f X p. (6.2)
In the Schródinger representation f = r and
p= -inV, (6.3)
where we know that Pi and Xj satisfy the commutation relation
[pi, xí] = -iMij· (6.4)
From Eqs. (6.2) and (6.3), it follows that the components of angular mo-
mentum can be written as
t; = -in(y.!!_ - z.!!_),/ (6.5a)
Bz ay~ ,)
L = -in(z.!!_ - x.!!_) ¡ (6.5b) )
y ax 8z Y'--'"
t., = -in(x:Y -y:x). (6.5c)
6.2 Properties of Angular Momentum
(1) We now proceed to show that

[Lx, Ly] = inLz.
87
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)
= -(iñ)2 (x!_ - y!_)'l/J

oy ax
= iliLz'l/J.
Hence (since 1/J is arbitrary)
[Lx, Ly] = uu.; (6.6a)
Similarly
[Ly, Lz] = iliL:4{
[Lz, Lx] = uu.;
(6.6b)
(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·
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
(LxLy - LyLx)'l/J = inLz'l/J

= inAz'lp-
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.
(X, y, z)
Fig. 6.1 The spherical polar coordinates of a point (x,y,z).
(3) It is convenient to express angular momentum operators in terms

of spherical polar coordina tes ( these are natural coordinates for angular
momentum),
x = r sin é' cos </> L )

y= r sin sín é é t.__
z =reos().
Then we have
t; = -iñ:</> (6.9a)
t; = :e
iñ (sin</> + cot cos </> :</>) e (6.9b)
Ly = -in( cos </> :</> - cot ()sin</>:</>) (6.9c)
L2 = -n 2(1--- 8(. sme- 8) + ---1 82) (6.9d)

sin () 8() 8() sin 2
() 8</>2
= -te«,
where
18. 8 1 82) (6.9e)
n = ( sin() 8/sm o 8()) + sin2 () 8</>2
is the angular part of the Laplacian V2 in spherical polar coordinates.
Proof.
r = rf1
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
</)1 = - sin Bi + cos Oj.

In polar coordinates
o Ala A 1 o
V = ri ar + 81-:;: {)(} + <P1 r sin(} acp
2 1 e 2 o 1
y' = 2!:l(r !:) ) + 20((), <P), (6.10)
r=or or r
therefore [r = rr1' r1 X 01 = <P1, rA1 X <P1 = -01]
L = -ilir x V
A {) A 1 {))
= -ili ( <P {)(} - (J sin(} acp
= -i!i( (- sin epi+ cos </Jj) :(} - ( cos (} cos epi
+ cos e sin </Jj - sin Bk) si~(} :<P)
= -iñ(i(-sin</J :o - cotBcos </J :<P) + j(cos</J:(} - cot (}
x sin </J : ( {)(} a + Z. o OqJ
a)
COt
· l (6.lla)
)
L _ -= L X -
ºL y -- ne
Z
i, -i</> a + . o acp
( - {)(} a)•
Z COt (6.llb)
Then
L+L- = L; + L~ - i[Lx, Ly] = L; + L~ + L; - L; + ñl.;
L2 = L+L- - L; - ñ.I.¿
- i,2
- t t: e
icj, ( o + Z. cot e f)cpo )
f)(} e
-i<j, ( {)
- {)(} + Z. cot o OqJ
{) ) -
i,2 82
tt.OqJ2
· i,2 o
+ Zn OqJ
{)2 o 1 82
= -ñ2[{)()2 +coto{)(}+ sin2(} f)cp2]
giving
L
2
= =ñ
2 ( 1 a .
sin() a/sm () 8())
a + sin12 () 8cp2
a2 )
·
So Eq. (6.9d) is proved.
6.3 Eigenfunctions and Eigenvalues of Angular Momentum
We now determine the simultaneous eigenfunctions of L2 and L¿ and their

eigenvalues.
Since L2 and L¿ commute, it is possible to find simultaneous eigenfunc-
tions '1/J>.m satisfying the eigenvalue equations
Lz'I/J>.m = mli'I/J>.m (6.12a)
L2'1/J>.m = An2'1/J>.m, (6.12b)
where mñ and >.!i2 are eigenvalues
/ indexEigenvalue!of L2 of t; and L2 respectively corresponding to eigen-
functions '1/J>.m· We have extracted out factors ñ and !i2 for convenience.
Using Eq. (6.9a), we have from Eq. (6.12a)
-in :c/J '1/J>.m = lim'I/J>sm· (6.13)
The solution of this equation is of the form

'1/J>.m = feim</>, (6.14)
where f is independent of cjJ but is in general~nction of r and (). Since the C\._
same physical position is denoted by cjJ + 2mr, ti being an integer, we must
impose the boundary condition that '1/J>.m is periodic in cjJ with a period 21r
that is
or
e2immr = l.
This is possible only if

m = 0,±1,±2, ...
Thus eigenvalues of L; are
mñ, m = O, ±1, ±2, ...
ANGULAR MOMENTUM 93
The factor depending on </; which characterises the eigenfunction of L¿ is

denoted by
(6.15)
This function is normalised so that

{2,r
Ío if>;,.(<J;)if>m1(</;)d</; = Ómml· (6.16)
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 -ml dn 2 = 1. (6.21a)
Now
dO = sin 8d8d</;.
This leads to (using Eq. (6.15))
11r IP>-m ( 8) 12 sin 8d8 = l. (6.21b)

To find the eigenvalues >. of L2, we proceed as follows. We have already

introduced in Eq. (6.11)
L± = Lx ±iLy
L+L- = L2 - L; + ñLz, L_L+ = L2 - L; - ñl.¿ (6.22)
where L± satisfy the commutation relation,
[Lz, L±] = [Lz, Lx] ± i[Lz, Ly]
= inLy ± i(-inLx)
=±nL± (6.23a)
[L+, L_] = [Lx, Lx] - i[Lx, Ly] + i[Ly, Lx] + [Ly, Ly]
= -i(inLz) + i(-inLz)
= 2nLz (6.23b)
[L2,L±] = [L2,Lx] ±i[L2,Ly] = O. (6.23c)
Theorem: If (Lx±iLy)Y>-.m = L±Y>-.m are not zero then they are simul-
taneous eigenfunctions of L2 and L; belonging to eigenvalues >.ñ2, (m ± l)ñ
respectively.
Proof.
L2(L±Y>-.m) = L±(L2Y>-.m) 1
= >.n2(L±Y>-.m)¡,.__, )
Lz(L±Y>-.m) = {[Lz,L±] +L±Lz}Y>-.m
= ±nL± Y>-.m + mñL± Y>-.m
= (m ± l)n(L±Y>-.m)· (6.24)
L+ is called the raising operator and L_ is called the lowering operator.
We can write
(6.25)
cf
where m are constants and can be fixed from the normalisation condition
which gives
or
IC~1.l2 = .! Y;'m(L:¡:)(L±)Y>-.mdD
= .! Y.x"m(L2 - L; =f nLz)Y>-.mdD
= (>. - m2 -F m)n2. (6.26a)

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)
Hence we have the inequality
(6.27b)
Thus m is bounded on both sides. Therefore m has a mínimum and a

maximum value which we respectively denote by m2 and m1. Then we
must have
(6.28a) 7
(6.28b)
since otherwise Eq. (6.24) shows L_Y.>-rn2 is eigenfunction of L¿ with eigen-
value (m2 - l)ñ which contradicts that m2ñ is the mínimum eigenvalue.
Similar argument holds for L+Y>-rni. Thus from Eq. (6.25)
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)
It is customary to designate m1n the maximum eigenvalue of L¿ by lñ;

where l must be an integer including zero as the eigenvalues of L¿ are
integral multiples of Also from the fact that m1 = l = -m2,
ñ.
-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)
where N1m is the normalizing constant determined from Eq. (6.2la).

Yim(O, </J) are called normalised spherical harmonics. In Eq. (6.36) Pr
are given by [see problem 6.7]
PIm( cos O) = 2ll!

(-1)1 . m O(
sm
d
d( cos O)
)l+m( . ())21
sm
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.
6.4 Properties of Associated Legendre Polynomials
The following properties of Associated Legendre Polynomials are given

without proof
(1) P{n(w)are orthogonal, that is
¡+1 m m 2 (l + m)!
L )
}_1 P1 (w)Pz, (w)dw = 8w (2l + l) (l - m)! (6.39)
(2) p1-m(w) = (-l{'/raj!m)! P1m(w) L (6.40)

)
(3) Ir(-w) = (-1)1+m Pt(w) ¡- (6.41)
)
(4) Yi-m(B, </J) = (-l)m [(2l + 1)
4~(7-m!)! J 112 P1-m(w)e-imcf,
= (-l)mY¡;',JP, </J) L (6.42)

2l + 1 1/2
(5) Yio(P, </J) = (---¡;;:-) Pz ( cos O). (6.43)
Equation (6.41) follows from Eq. (6.37) and Eq. (6.40) follows from the
Leibnitz Theorem:
pm(w) = (-1)= (l + m)! (1 - w2)-m/2 dl-m (w2 - 1).
1 21z! (l - m)! dwl-m
For l = O, m = O, Y00 = constant. In other words the wave function of the
state of a particle with zero angular momentum depends only on r, that is
it has complete spherical symmetry.
The first four spherical harmonics corresponding to l = O and l = 1 are
given by
m=O
O Yoo = -1-1
(41r)~
3 1 •et, 3 1 3 1 • •et,
1 Y11 = -(81r) 2 sin Be' , Y10 = ( 41r) 2 cos O, Y1_1 = ( s1r) 2 sm ee::
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
6.5 Parity of a State
Previously we introduced the concept of parity in Sec. 4.3. We saw that in

one dimensional problems if the potential is symmetric
V(x) = V(-x)
about the origin, then 'lj;(-x) is also a solution of the Schródínger equation
ñ,2 82
(- 2m 8x2 + V(x))'lj;(x) = E'lj;(x)
and the parity of 'lj;(x) is defined by r¡
'lj;(x) = r¡'lj;(-x)
with r¡ = ±l. The transformation x - -x corresponds in three dimensions
to
r - r' = -r.
ANGULAR MOMENTUM 99
Such a transformation is called the reflection of coordinates. Now in

spherical polar coordinates
x = r sin B cos <P l )

{
y= rsinBsin</J L (6.44) )
z = reos B.
If the axes are reflected to t' = ( x', y', z') and if r', B' and </J' are defined
in the usual way, but with respect to x', y', z', then it is easy to see that
r -----t x' = -r corresponds to{?'~r L L )

B' rr - B (6.45)
,)
</>' 7r + </J.
Consider now a particle in an angular momentum state Yim ( B, <P) =
constant x P1m(cos B)eimc/> which specifies the dependence of the particle's
state on angles B and </J. When </J-----t rr+</1, eimc/> -----t eim1r émc/> = (-l)meimc/>.
en O , - .,_E¡'.,..Í!E_s()) -1~"" -cosB) = (-1)1+~
c=..-x Pt ( cos B). Thus under reflection of coor ura-tes~, --
Yim ( B, </1) -----t Yi.m(B',</1') = constant X (-1)1+m(-l)mp1m(cosB)eimc/>
= (-1)1Yim(B, </>). (6.46)
Since, as is clear from Eq. (6.45) the dependence of the wave function on
r is unchanged by reflection, the parity of any state specified by angular
momentum l, mis equal to (-1)1 and is determined by the l value only.
( -1 )1 is often called the orbital parity of a state with an angular momentum
specified by l and m values.
6.6 Problems
6.1 Consider an operator V+ which satisfies the commutation relations
[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).
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
by using the result

[xi,.Pj] = iMij
6.4 Show that for a state 'l/J1m, such that
Lz'l/Jlm = mn'l/J¡m L 7
L2'l/J1m = l(l + l)fi2'l/J1m,
the average values of L';,, L~ are given by
(L;) = (L~) = l(l + l)n;- m2ñ2

6.5 Show that for an eigenstate 'l/Jzm of Lz, the average value of Lx is
zero.
6.6 Show that
(L2, (L2, x]] = 2!i2(xL2 + L2x).
Hint: Use the commutation relation
and the definition
to show that
[Li, Xj] = +ihEijtXt ~ )
[L2,x¡J = (-ih) LCijdLixj +xjLi]
ij
Using these relations, derive the final result.

ANGULAR MOMENTUM 101
6.7 Using the relation derived in the text

L±Yim(IJ, cp) = J(l =f m)(l ± m + l)nYi,m±1 (IJ, </>)
and the expressions for L± in spherical polar coordinates, show
that
Pt±1(1J) = [(l=fm)(l±m+l)J-112(±81)
a -mcotlJ)Pt(IJ)
= [(l =f m)(l ± m + l)J-112(±) sin±m I) :O

(sin:¡:m IJP1m(IJ)).
X (6.47)
Putting m = -l and taking the lower sign, we must have
d~ (sin-1 IJP¡-1(1J)) = O.
This is a consequence of the fact that there is no state with m < -l.
Thus
P1-1(1J) = asin1 IJ.
The constant a can be determined from the normalisation condition
Fa2 fo1r (sinlJ)21+1dlJ = 1
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-m+l B{(l - m + 2)(l +m - 1n-112
x(-1) sinm-l g d (sin-m+2(B)Pm-2

1
(1J)))
d( cos 8)
= (-1)2 ((l - m + l)(l - m + 2)(l + m)(l + m - 1))-112
x sin
. ma d(coslJ)2
d2 (. -m+2apm-2(8))
sm l •
Chapter 7
MOTION IN CENTRALLY
SYMMETRIC FIELD
7.1 Introduction
Consider a spinless particle of rnass µ in a potential V ( r). The Hamiltonian

is given by
-2
H = !!_
2µ
+ V(r), p = -ilf\1. (7.1)
If the potential is spherically syrnrnetric, V(r) = V(r)
H = --'v
r,,2
2µ
2 + V(r). (7.2)
The Schrodinger equation for such a systern is
iño'lj) = H'lj)
8t
= (-~:'v +V(r))'l/J.
2 (7.3)
Such a system represents a hydrogen like atoro in which an electron of mass

µ is in a potential of the nucleus (which is regarded as fixed, because it is
very heavy as compared with the electron), so that V(r) = - z;2 where Ze
is the charge of nucleus. We are interested in the stationary states of the
systern, that is we want to find the energy eigenvalues of the system. For a
stationary state
'l/J(r, t) = u(r)e-i(E/fi)t_ (7.4)
Substituting in Eq. (7.3), we have
(- ~: V2 + V(r)) u(r) = Eu(r). (7.5)
103
From a mathematical point of view, it is important to find a coordinate

system in which the equation is simplified. Here V is a function of r = [r]
alone, so it is convenient to use spherical polar coordinates r, (), </J. Then
2
\7 u=
(1r2 a/r
a .»ar)+ 1
r2D,((),<f>) u
)
-- !~( ) + 0((),2 </>) u, (7.6a)
r r ru
a 2 r
where
(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
- ñ2 _!_~(r28R) + l(l + l) ñ2R + V(r)R = ER(r). (7.10)

2µ r2 ar ar 2µr2
MOTION IN CENTRALLY SYMMETRIC FIELD 105
We note that in this equation Lz = mñ <loes not enter Hthis is in accor-

dance with the (2l + 1) fold degeneracy of the levels with which we are
already familiar. We transform this equation into a more convenient form
by making the substitution
R(r) = x(r). (7.11)

r
Thus Eq. (7.10) becomes
ñ2 a2 ñ2 z(z + 1) )
( - 2µ 8r2 + 2µ r2 + V(r) x(r) = Ex(r). (7.12)
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)
and that for x(r) by
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.
(3) Equation (7.12) is formally identical with Schródinger's equation

for one dimensional motion in a field of potential
ñ2 l(l + 1)
Vz(r) = V(r) +-µ
2 2 •
r
(7.15c)
The term ~: l(l~l) is called the centrifuga! potential ar barrier; this

gives a repulsive potential. For small r, it is likely to be large. It
prevents a particle with nanzero angular momentum from getting
too clase to the origin, hence the name centrifugal barrier. We
assume r2V(r) --+ O as r --+ O which is satisfied by the Coulomb
potential.
7.2 Coulomb Potential
We can go as far as Eq. (7.12) without specifying V(r). From now

onwards we consider the specific problem of hydrogen like atoros for which
Ze2
V(r) = -- (Coulomb potential)
r
where Ze is the charge of nucleus. This is an important problem in atomic
physics. From Eq. (7.12) for x(r), we have for V(r) = -2:2
á2x 2µ ( ze2 ñ,2 )
dr2 + ñ,2 E+ -r- - 2µr2 l(l + 1) X = O. (7.16)
We want to solve this equation under the boundary condition x( r) --+ O, as

r --+ O. Let us put
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)
First we examine the asymptotic behaviour. As x --+ oo

K and l(l + 1) -----+ O
X x2
and then
(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
MOTJON IN CENTRALLY SYMMETRIC FIELD 107
x = e+"x - oo as x - oo and therefore, is rejected. Thus the only possible

solution is x = e-ex. We now examine the behaviour of Eq, (7.18) for small
values of x. Multiply Eq. (7.18) by x2 throughout. We have
2d2x
x dx2 - l(l + l)x = O. (7.20)
We seek a solution of the form
x(x) = constant x x", (7.21a)

Substituting this in Eq. (7.20) gives
s(s -1) - l(l + 1) = O

s = l + 1 or s = -l. (7.21b)
Now l 2: O, therefore s = -l is not a permissible solution as it contradicts
the boundary condition x(x) - O as x - O. Thus the boundary condition
requires that
Therefore x(x) can be written in the form

s = (l + 1) L
x(x) = e-"xxt+1g(x), (7.22)
where the expression (7.22) for x(x) takes account of the behaviour of the
function for large and small x and g(x) is an unknown function. Substitut-
ing this expression for x(x) in Eq. (7.18), we obtain
d2g dg
x dx2 + 2(l + 1 - ex) dx + (K - 2cl - 2c)g = O. (7.23)
We now look for a solution of this equation in the form
g(x) = ¿ CsX8 =Ca+ C1x + C2x2 + · · · (7.24)

s
Then substituting Eq. (7.24) into Eq. (7.23), we get
¿ (Css(s - l)xs-l + 2(l + 1- cx)C8SX8-l + (K - Zel - 2c)C5x5) = O,

s
(7.25)
which gives on equating the coefficient of xs-l to zero, the following rela-
tions between successive coefficients C8
C5 (s(s - 1) + 2(l + l)s) = Cs-1 (2cs + 2cl - K). (7.26)

Equation (7.24) represents an infinite series. For large values of s, the
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
whose ratio of the coefficients of successive powers of x is

-\(2c)8 2c
s. =-
_1_(2c)s-1 s
s-1!
which is the same as the behaviour of 08/08_1 for large s. Series (7.27)
converges to e2éx. Thus the series for g(x) converges and it asymptotically
behaves as e2éx that is
g(x) rv e2éX
for large x. Thus
,,/2µ xl+le-éx
R(x) = - g(x),...., x1eéx
fi X
for large x. Thus as x --+ oo, R(x) --+ oo, as e is real. It will thus not
represent a physically permissible solution. In order to avoid this, the
series (7.24) has to be terminated at sorne finite value of s. Then R(x) will
be of the form
(7.28)
where P(x) is a polynomial in x. This solution will satisfy all the require-
ments of the physical situation. The condition for termination of the series
can be found as follows: Let the highest power of x in the series for g(x) be
xn' (ni being an integer ~ O). Then the coefficients Cn,+1, Cn1+2 etc. must
be zero so that from Eq. (7.26), putting s =ni+ 1, we get Cn1+1 = O, if
2c( ni + 1) + Zel - K =O
or
K K
é=----- (7.29)
2 ( ni + l + 1) 2n '
where n = ni+ l + 1 is a positive integer and for a given l is ~ (l + 1) as
w > O. With the help of Eq. (7.17), Eq. (7.29) becomes
K2
E=-c2=--
4n2
=- (µe4z2). (7.30)
2h2n2
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
l and m are called azimuthal and magnetic quantum numbers (l = O, 1, 2, · ·

·, (n - 1)) and (m = -l, · · ·, l) i.e can take (2l + 1) eigenvalues. n' is called
the radial quantum number. That energy levels do not depend on m is
a consequence of rotational symmetry of the potential. That they do not
depend on lis peculiar to 1/r potential like Coulomb potential. This means 11 " _
that energy levels are degenerate. Since corresponding to each val~:1of l, m ~~~
can take (2l+ 1) values, therefore degree of degeneracy is given by ¿ = n2. ,tl'vV)Í ~
In other words for one energy value E, there are n2 different eigen~~~ctions /
usually a degeneracy is associated with a symmetry and it is known that
in this example there is an external symmetry [R, H] = O, with
1 e2
R = -(p x L -L x p)- -r.
2m r
The operator R given above follows from the Lenz's vector in the classical
Kepler problem by h correspondence principie.
Now n = 1, 2 . . and far a gíven l, n ~ (l+l). Thus for E< o/that is for
the bound states formula (7.30) gives discrete set of energy lev~. These
are in agreement with experiment. We now introduce the spectroscopic
notation. The states with different values of l are denoted by
'
l = O, 1, 2, 3, 4, 5, · · · s, p, d, f, g, h, · · · .
The principal quantum number is written on the left of ~ove letters. Thus
for n = 1, l = O we have ls level; for n = 2, l = op;we
have 2 and 2p
levels which are degenerate. Since for l = O, m = O J]! and 2 level are
not further degenerate, but 2p level is three fold degenerate corresponding
to m = -1, O, l. Thus for n = 2, we have four fold degeneracy.
We end this section, with following remarks L
l. The energy levels given in Eq. (7.30), can be written as
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.
2. Define,.+the radius aot1led the Bohr radius
___!__ = µe2 = (µe2)a. En=_ Z2a (ñ)

ªº n2 en ) 2n2 ªº
cñ = 1.97 x 10-13 Me V - m
eñ -10
ao = ( = 0.528 X 10 m
a mee 2)
3. Define the Rydberg constant R:
R = (mee2)a2 = (mee2)a2• En= _z21r(ne)R

2ne 41rnc , n2
R ~ 1.0978 x 105cm-1, heR = 21reR ~ 13.61eV
7.3 Wave Functions of Hydrogen-like Atom
The wave functions of stationary states and the corresponding eigenvalues

are given by
(7.31)
Rn1 being defined (except for a normalisation constant) by xe1e-exg(x) with
the series being terminated at s = n' = n - l - 1 and recurrence relations

for Cs being given in (7.26). Thus we can write
R nl = B nl X l e -EX gnl ( X) , (7.32)
where gni(x) is the polynomial

n-l-1
gn¡(x) = Ca ¿ (Cs/Co)X 8 (7.33)
s=O
and constant Bn1Co is determined from the normalisation condition
J R';1r2dr = l. (7.34)
Now from (7.26)
C8s(2l + 1 + s) = Cs-1(2(s + l)c - K), (7.35)

July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
with e = K /2n, so that

Kª (s + l - n)(s + l - n - 1) · · · (2 + l - n)(l + l - n)
e [O,0 - --'----'---'---~-~------
ª - nª s!(s + 2l + l)(s + 2l + 1-1) · · · (2l + 2)
Kª(-l)s ((n-l- l)(n-l-2) · · · (n-l-(s-l))(n-l-s)) (2l+l)!
=
nª s!(s + 2l + 1)!
Kª s (n - l - 1)!(2l + 1)!
= -;;(-l) s!(2l+l+s)!(n-l-s-1)!º
Thus
_ C nL-l-l(- )ª (n - l - 1)!(2l + l)!(Kx/n)ª
9nl (X ) - O 1 ) . (7.36)
s=O s!(2l + 1 + s)!(n - l - s -1 !
We now introduce what are called associated Laguerre polynomials L~1 t/ (y)
defined as
n-l-1 [( l)l] s
L21+1( ) - ~ (-l)s+l n+ .Y (7 37)
n+l Y - L (n-l-1-s)!(2l+l+s)!s!º ·
s=O
From Eqs. (7.33), (7.34), (7.37) and (7.38), it is clear that we can write
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)
is the normalisation constant and can be determined from normalisation

condition (7.35). It can be shown that the associated Laguerre polynomial
L~ (y) satisfies
yL?(Y) + (p + 1 - y)L'J'(y) + (q - p)~ =O, (7.40)
this is the same equation as the differential equation for g(x) with y= ~x,
q = n + l, p = 2l + 1 ande = ::;;_ as discussed previously. This is as it should
be. The Laguerre polynomial Lq(y) satisfies
yLqlf + (1 - y)L~ + ql.¿ = O (7.41)
and the associated Laguerre polynomial is
dP
L~(y) = dyPLq(y), (7.42)
where L¿ (y) can be written as

dq
L (y) = eY-(yqe-Y). (7.43)
q dyq
A further property of L~1 t/ (y) is that
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.
2s level:
z3/2
U200 = (2 - Zr/ao)e-Zr/2a0
4v2rra~12
corresponds to
Símilarly usíng expressíons for

Y1-1, Y10, Y11
we can wríte the expressions for
corresponding to 2p levels. All these wave functions correspond to the same

energy E2 = tE1. Thus for n = 2, we have in all 4 wave functions, one
for 2s and three for 2p level, all corresponding to the same energy, i.e.
degeneracy is 22 = 4 fold as it should be.
It is useful to introduce a quantíty, called radial probabílity density.
Now the probability of finding the electron in the volume element dr =
r2 dr sin ()d()dq> is
u~lm (r )unlm (r )dr = R~1 (r )Y¡;,.((),</> )Rnt (r )Yim ( (), </> )r2dr sin ()d()dq>.
Then
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
unless n' = n, l' = l, m' = m.

Physically this means that if the particle is in eigenstate (n, l, m), the prob-
ability of finding it in another eigenstate (n', l', m') when n', l', m' are all
different from n, l, m is zero.
Radial probability density Pn1(r) is defined by
which is the probability of finding the electron in a spherical shell of radius

r and thickness dr. Let us now calculate the expectation value
(rnz) = 1 00 rPnz(r)dr = 1 00 1Rnz(r)l2r3dr
{l + !{1- l(l + 1) }}.

= n2ao
Z 2 n2
This characterises the radius of the shell. This depends primarily on n while
dependence on l is weak beca use of the factor ~ in l dependent term. This
e
may be compared with the radii of circular orbits of Bohr ( r n) Bohr = n ;·0 .
All electrons in eigenstates with common n values have approximately the
same (rni), independent of values of l or m. Such electrons are said to
be in the same shell. Each shell also has the property that the associated
eigenvalues have exactly the same value En. Note that for r < a0/Z,
Unlm rv r1, r ---t Ü, SO that the probability density U~lm Unlm rv r21, r ---t Ü.
Thus the value of probability density in a volume element near r = O

is relatively large only for l = O and decreases rapidly with increasing l.
Other useful expectation values are
2) = a5n4[5 1-3l(l+l)]
(r nl 2z2 + n2
1 Z
(-) ¡- -
r n - aon2
1 z2
(r2)n1 = a5n3(l + !)
7.4 Problems
7.1 In spectroscopic notation, a state is specified as 2s+1zj, where s

is the spin and j is the total angular momentum. In particular
for hydrogen atom, s = !
i.e the spin of the electron. Thus j =
l - !, l+ t List all the possible states for n = 1, n = 2, n = 3.
Find (..l. - ..l.) (..l. - ..l.) (..l. - ..l.)
.>.3 >-2 ' .>.3 >-1 ' >-2 >-1 •
Show that for the transitions from the level
n = 3 to n = 2,
n = 3 to n = 1,
n = 2 to n = 1,
O,
the photons of wavelengths 6559A 1025Aº and 1027Aº are
emitted respectively.
7.2 A particle of mass m moves in a square well potential

)
V(r): ~Va ~~·:a~L I\\~"'
If bound s-state exists, show that eli:rgy eigenvalues are given by
-, = kcot ka
where
, = J2mW/n2, E= -W
k = J2m(Vo - W)/n2.
If there is only one such state and it is very weakly bound, show
that
1r21i2
Voa2 rv 8m .
7.3 Assume that interaction between the neutron and proton that make
up a deutron can be represented by a square well potential with
a rv 2 x 10-13 cms. If l = O energy level of the system (binding
energy) is 2 MeV, calculate Va in MeV.
7.4 The radial wave function x(r) satisfies the equation
d2x
dr2 +
(2m(E _V)_ l(l
r,,2
+
r2
1))
Take V(r) = (1/2)mw2r2 and by expanding xexp(l/2p2) in powers
of p = r/b with b2 = n/mw, show that energy levels are given by
X
= 0 l o
En= (1/2)ñw(4n + 2l + 3)
where n is an integer 2': O.
7.5 A particle moves in a spherically symmetrical potential V= a/r2-
(3/r, where a and (3 are constants. Find bound state energy levels.
7.6 Find the expectation values (r) and (~) for the ground state ofL P\.
hydrogeni!k atom.
7.7 Find the energy eigenvalues and eigenfunctions for a three-
dimensional harmonic oscillator. The potential for such an oscilla-
tor is given by
V(x, y, z) = ; (wfx2 + w~y2 + w~z2) L Ú
Discuss the degeneracy of energy levels. Hint: Write the wave
function as
u(x,y,z) = u1(x)u2(y)u3(z)
and show that each of the functions u1(x), u2(y) and u3(z) satisfy
ce dimensional harmonic oscillator equation.
7.8 For
p2 A Ze2
H = - + V(r) = T- -
2m r
Show that
Chapter 8
COLLISION THEORY
8.1 Two Body Problem: (Centre of Mass Motion)
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)
Hence we can write

n2 n2 n2 n2
---V2-1 --V~= --V2
2m1 2m2 2µ
- -v'h
2M
n2 n2
H =
2M
--v'h--V2 + V(r)
2µ
(8.7)
where M = m1 + m2 is the total mass of the system and µ is the reduced

mass
1 1 1
-=-+-.
µ m1 m2
Equation (8.3) can now be written as
. o'l/J
iñ- =
(-1t2 2
-VR
n2 2
- -V + V(r) ) '!/; (8.8)
at 2M 2µ
with
'!/; = 'l/;(r, R, t).
COLLISION THEORY 119
For a stationary state belonging to energy eigenvalue Er of H, the time

dependence will be of the form e-iErt/ñ so that
'if;(r, R, t) = '1/Jo(r, R)e-iErt/ñ.
Substituting in Eq. (8.8), we have
ñ,2 ñ,2
(-2M 'vh-
2µ
'v2 + V(r))'I/Ja(r,R) = Er'if;0(r,R). (8.9)
We note that no term in the parentheses depends on both r and R. Hence

we can write a solution of the form
'if;0(r, R) = u(r)U(R).
Substituting in Eq.(8.9) and dividing by u(r)U(R), we get
ñ,2 1 ñ,2 1
2M U(R) 'vkU(R) + (- 2µ u(r) vzu(r) + V(r)) = Er. (8.10)
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.
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).
Fig. 8.1 lmpact parameter b.
8.2 Collision Theory
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
in collision problems in which case a particle arrives from infinity (large

í distanc ty scattered by a field of forces and goes off to infinity (large
distance). For these states the wave function rnay .remain flnite at infinity.
For bound states, for example hydrogen-like atoms, the boundary condi-
tions at great distances were used to determine the discrete energy levels of
the particle. In a collision problem, the energy is specified in advance, and
the behaviour of the wave function at great distances is found in terms of
it. This asymptotic behaviour can then be related to the probability that,
as a result of the collision, the particle will deviate ( or as, we say, will be
scattered) through a given angle. The purpose of the theory is to calculate
this probability, i.e. to say the probability that the bombarding particle will
interact in a certain way with a target particle. Thus scattering processes
are very important for studying mutual interactions of particles. For ex-
ample to investigate nuclear potentials, one studies the mutual interaction
of nucleons (proton or neutron) by scattering one o~e other.
f
We first consider the kinematices of a cattering procese.
8.3 Kinematices of a Scattering Process
Consider a collision between two particles with coordinates and masses

given by
The wave function 'lj; satisfies the Schrodinger equation
. {)'lj; ( fí,2 2 fí,2 2 ( ))

in-¡:;-= ---V1 - -2-V2 + V r1 - r2 'lj;.
ot 2m1 m2
By introducing the variables
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)
@
-:: 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
ql = CÍ1 + CÍ2· (8.17)

Velocity of centre of mass in this frame is given by
. 1
V=R= Mm1r1
(8.18)
In the centre of mass frame, let initial momenta of particles 1 and 2 be Pi

and p2 and let final momenta be p1 and p2. In this frame, centre of mass
is at rest
MR=O, (8.19)
this gives
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
/
/
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.
The collision process in two frames is shown in

We have
(8.20c)
where u is the initial velocity of the particle 1 in the centre of mass frame.
Equation (8.20c) can be written as
~ = ql + .R.. (8.20d)
m1 M m1
or
(8.21)
Similarly ( after collision r·1 = ~, R = ,& ( q~ + q~) = ,& q1, u' = ~)

~ ' = ql + R.._
'
m1 M m1
or
Also
L (8.22)0
or
(8.23) D
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
</JL = </Je. (8.24b)

Now kinetic energy Te in centre of mass frame is given by
1 m2
Te =p2/2µ = 2µM;q2
(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
8.4 Scattering Cross Section
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 number of particles detected by the detecto~gives the probability of (2.

scattering at angle e. Thus the probability which is related to the scattering/
cross section is proportional to the number of particles detected at an angle
e e
between and + dO divided by the incoming flux.
Now incoming flux:
pv = 1Aeikzl2v = IAl2v. (8.29)
Outgoing probability density:
p = IAJ(O, </J)!eikr¡2. (8.30)

r
Therefore the number of particles scattered in volume element between r
and r + dr in the solid angle d0. is given by
PdT = Pr2drd0.
(8.31)
Flux received in volume element dr:
PdT]__ = IAl2lf(O </J)i2d0.dr
dt ' dt
= IAl2IJ(O, </J)l2d0. v. (8.32)
Therefore, the number of particles scattered into d0. per unit time per unit
incident flux:
da = lf (O, </J )12d0.. (8.33)
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)
Scattering is an important tool to probe the structure of matter. Ever
<°&~
~
,_____ • z-axis
Fig. 8.3 The incident plane wave and the scattered wave moving radially outwards.
since Lord Rutherford discovered the atomic nucleus by scattering of a

particles on atoms; the structure of matter has been probed using scattering
of electrons, protons and photons on various targets. Major discoveries
Wil~made in atomic conden · d matter nuclear and particle physics. The
scattering cross s ction gives ntinforrnation about th interaction which
cau ed the deviation of in ident particle to the dete tor.
The relation between the differential cross section in the centre of mass
frame and that in the laboratory frame follows from the fact that the effec-
tive cross section in the two frames must be same:
(!~)e dD(Bc,<Pc) = (!~) L dD(BL,<PL), (8.36)
From Eq. (8.24b)

<Pe= <PL
(1 ++ cos Be) sin BcdBc

mi
sin()Ld()L = m2. • (8.37)
(~ + cosBc)2[1 + 2~ cosBc + (~ )2]í
Hence from Eqs. (8.36) and (8.37)
( ddu)
H
n
L
= (~ + cos Bc)2[1 + 2~ cosBc + (~ )2]!
( m
1 + m~ cosBc
)
(du)
dn e.
H
(8.38)
8.5 Scattering by a spherical symmetry potential
Mo t of the potentials in physics are sphericJ svmmetric vW V(r) = V(r).

For a spherical potential f(O,</>) = f(B), ckl
= 21rsinBdo.'""Our aim is
6lo
to calculate f (O) in terms of phase shífts, This is particularly convenient
at low energies, where only ~w partial waves are important. It is also
useful, when in a scattering process, a resonance with a definite angular
momentum is excited. I•. -;-t .
We start the section by semiclassical argument:i..fl1gular momentum of 1 (....1_
a particle traveling with momentum pin the íncídent bearn ata distance ~
do ·est to the scattering centre is pbt i_:,,called the impact parameter (Fi . ~
r
8.3. Thus
pb = tñ.
If R is the radius of interaction, the scattering takes place if
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.
l hove argument shows that it is convienent to analyse the scattering

p itude f(()) in terms of its angular momentum components.
The incident wave
(8.39)
includes all the components of angular momentum about the origin.

eikrcose = ¿ii(kr)Pi cos(O)
e
1
= ~L e JuTIªdi(kr)Yeo(e,
2f + 1
</J) (8.40)
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)
For s-wave; project out f = O component in Eq. (8.40)
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,-
J eikr cose dO 21r f eikr cose sin ede

f dO 41r
= _l_[eikr - e-ikr]. (8.46)
2ikr
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)
The flux of incoming wave is proportional to le-ikrl2 = l.

The flux of outgoing wave is proportional to !Seikr¡2 = ISl2.
For elastic scattering, the flux of incoming wave must be equal to flux of
outgoing wave:
1s12 = 1, s = e2il\ (8.48)
where ó is real. ó is called ~ase shift.
Reexpressing Eq. (8.47) in the form
e2ió eikr _ eikr
u(r) rv 2ikr
= ---;-[(e2ió _ l)eikr + (eikr _ e-ikr)]
2ikr
eikr _ e-ikr e2ió _ 1 eikr
(8.49)
= '"-...----"
2ikr + 2ik r
incident wave
and comparing it with Eq. (8.28):

e2ió - 1 é5 ( eió - e-ió) eió sin ó
f = 2ik = 2ik = k (8.5SO)
Thus for s-wave scattering
~f'-
di f(ll)l2 = si~:ó (8.5
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
{Bf = klfl 2 = 4: O-total (8.56)

called the optical theorem.
The generalization to any partial wave is straightforward. For large r Ú-:-
i.e. r --> oo, using Eq. (8.40~d the asymptotic form of ie(kr) given just ~
after Eq. (8.40), the incident wave behaves as
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)
_ 2t' _+ 1 [( e2i5e _ l )eikr + ( eikr _ (- l)ee-ikr)]Pi( cos B)

( 2ikr
ik.rcoslJ + L 2t': 1 (e2i5e - l)~eikr Pi(cos B). (8.58)

i 2ik r
Hence comparing above equation with Eq. (8.28), we get
f(B) = ¿(2l + l)(e2i5e - l)Pe(cosB)

2\i e
= ¿(2l + l)fe(k)Pe(cosB). (8.59)
e
where fe(k) is the scattering amplitude for the gth partial wave:
fe(k) = 2~k(e2i5e(k) -1)

1 °{J
= -e' e sin Se (8.60)
®Now
k
@!e(k) = i sin2 8e(k). (8.61)
f(O) = ¿(2l + l)fe(k)Pe(l) = ¿(2l + l)fe(k).

e e
Thus
(8.62)
The total cross section:
(T = J lf(B)l2dn = J IJ(e)l221rsinOde
= ¿ ¿(2í! + 1)(2€ + l)fdl, jP1.(cosB)P1,,(cosB)dD

i i'
= ~ ~(21! + 1)(2€ + l)fdl, j 1! V 1!;:

2 4: 1 2 1 YeoYe odD
1
= L L 41r)cu + 1)(2t + 1)Jdli81.t = 41r ¿(21! + 1)1h12

1. e 1.
= 41r ¿(21! + 1) sin2 h (8.63)

R.
Hence from Eqs. (8.62) and (8.63), we obtain the optical theorem:
lr:;:;;y(O) = ..5._(T. (8.64)

L.7 47f
Each term in Eq. (8.59) corresponds to a definite value of the angular
momentum í!. Eq. (8.63) shows that each angular momentum contributes
independently to the total cross section. Since sin2 ó¡_ ::; 1, each í! con-
tributes to the sum in Eq. (8.63) at most a partial cross sectiottJ J
(ue)m= - 1: (21 + 1)
an important result. Eq. (8.59) for the differential scattering eros sec-
L (8.65) )
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)
8.6 Phase Shifts
By determining the phase shifts at various energies empirically from the

experimental data using Eqs, (8.59) and (8.64), one can get information
about V or the interaction between the bombarding and target particles.
The method of partial wave for expressing the scattering amplitude is quite
general and <loes not depend on the detailed form of V(r) provided that it
falls faster than ~ at lar ge r.
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)
Now from Eqs. (8.68) and (8.70)

G d2G1 _ Gº d2Ge] = -Gº 2µ,V(r) G (8.72)
[ e dr2 e dr? e ñ,2 e·
Integrating Eq. (8.72), from O to r:
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
Now Gt(r) and G~(r) vanish at r = O. Thus we have,
(8.74)
Now from Eq. (8.58), we obtain

Gt(r) _ -4- [e2ióteikr _ (-l/e-ikr]
r--+oo 2,k
=
O
eióe[eió,,eik1' _ eibre-ió,,e-ikr]
2¡k
= ~75)
2;k ei(ó,.+~l[ei(kr+ót-~) _ ei(kr+óe-~)J
Hence, for r-----+ oo,

1 br
G2(r) - - sin(kr
k
+ 82 - -)
2
(8.76)
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
Rt (r)V(r)j2(k-r)~2-d;, (8. 78)
Eq. (8.7 ) is no very useful, since Re(r) is not knoin.However if the J

potential V ( r) is wcak11 and regarded as perturbation then to first order j l)
in perturbation, Ri(r·) can be replaced by the fr partic e solution Jt(kr).
For this case
sinót ~ -
2~k
V(r)j¡(kr)r2dr. J (8.79)
This is called Born approximation.

Now J2(kr), for kr << 1, is given in Eq. (8.71). Thus it follow",,lthat
if ka :::; 1, where a is the radius so that V(r) is very small for a 2: r, then
t
in the region r:::; a<< t;, that is the regían where V(r) is appreciable, the
phase shift ót for the eth wave, e > ka, will be very small. It follows from
Eq. (8.59):
emax
f(B) = ¿ (2/! + l)h(k)Pt(cos B) (8.80)

i=O
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,
For very slow velocities si~;r ---. 1, so that

L o
óo ~ sinóo ~ - ~k J V(r)r2dr
-~ is called the s-wave scattering length:

. óo
l lffi
as - k--+O -- (8.83)
k
where
as = 2µ
r,,2 J V(r)r 2
dr
a8 < O, if V is attractive and > O if Vis repulsive.

The general solution of Eq. (8.69):
R~ = (A¡,j¡,(kr) + Ben¡,(kr))
The following limits for the spherical Bessel functions are useful:
L (8.840
. (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
As an application of the above limits, we consider the behavior of the

phase shift óR. near k = O. First we note that V(r) ~ O, for r > a. We
consider the radial equation (8.68) in these regions:
i)t.::;: r .::;: a: k2' is negligible:
d2Gt _ [2µ,V(r) + f(f + 1)] GR.= 0
dr2 ñ2 r2
ii)b s r << s
k-l.(ka kr << 1):
'C, d2Ge _ f(f + 1) GJI =
0
dr? .,.2
which has solution
GR.(r) = C1rJl+1 + C2r-J1
where 01 and 02 are independent of k.
iii)t k-1 up to oo
d2GR.
dr 2 +
2_f(f+l)]G
r2
[k
R.
=O
which has general solution given in Eq. (8.84)
GR.(r) = rRR.(r) = r[AóR.(kr) + BR.nR.(kr)]
krrl (2f1)!
--------, r [Al (2f + 1)!! - BR. kJl+l r .
-R.-l]
for kr ----+ O. Then connecting the solutions in regions (ii) and (iii) for
kr << 1, we obtain
AR.kl _ BR.(2f + 1)!! .-Jl-l _ C
(2f + 1)!! = C1, kH1 1 - 2·
Thus using the asymptotic limits (8.85)
GJI rv k~ sin(kr -
1
2br) - k~ cos(kr -
1
2br).
Putting AJI = q cos óR., BR. = -q sin ÓJI, so that the last equation takes the
correct asymptotic form:
q sm(kr
. br
GR. rv k -
2 + óJ1)
where
BJI
tan8J1 = AJI =
º2 k2Jl+l
C1 (2f+ l)!!(2f1)!!
L (8.86)0
m
Note that 02 and 01 are independent of k. Thus as k ----+ O, k2H1 cot ÓJI ----+
constant which depends on f. In particular k cotó ----+ -.1..., where a; is
Ú _ . called the s-wave scatterin~ength.( ot that in contrast"to (8.83), this
~ ~ result does not depend on Born approximation). Since Óf rv k2H1, for
small k, all phase shifts wit e-/=- O are small in comparison with 80 i.e. at
low k,pave dominates.
8.7 Optical Theorem
We now discuss the optical theorem:
f~l 1t:) Im/ (o~ t,cu + 1) sin' 5,
~ 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 =
Thus, since for absorption {3 is positive,

lr¡e(k)l2, r¡e(k) = e-2f3e(k)
L (8.92)
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
Thus the total elastic cross section is

ae1 = 41r ¿(2f + l)lft(k)l2
e
= k24 '°'
471" 1 L..-(2t + 1)11- s,,¡ 2
e
= 471" ¿(2f
l
+ 1) 1 + 'f/1 -4~~i cos 2ót
There is also a cross section for inelastic processes. Now

L (8.96)
o
1 2
(Flux)in l(Flux)in - (Flux)outl = 1 - IStl
and thus
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)
Maximum absorption occurs when Se = O, then
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.
p > a, í! > L, no elastic scattering and no absorption:

tte = 1, e-2/3e = 1, SR. = l.
Hence we have from Eq. (8.91), Eq. (8.96) and Eq. (8.97) [using L =
ka>> 1]
a-el = ; t,(21!
e=O
+ 1) ~ ;L2 = 1ra2 L )
a-abs = ; t,(21! + 1) ~ 1ra2 [_/ )

l=O
a-tot = a-el + a-abs = 21ra2 •

On classical grounds we expect that the cross section cannot exceed the
area presented by the disc. The above result can be understood as follows:
The absorptive disc takes flux proportional to 1ra2 out of the incident
beam and this leads to a shadow behind the disc. The incident beam is
diffracted from the edge of the disc and that gives rise to elastic scatter-
ing. The elastic scattering that accompanies absorption is called shadow or
diffraction scattering. It is strongly peaked forward. This can be seen as
follows:
Since elastic scattering is coherent,

dael 1 2
dD = (8.102)
4k2[~)2e+l)(l-St(k))Pt(cosB)[.
For a black disc
s, = O e < L, p <a
St = 1 e > L, p > a, (8.103)
hence
Fig. 8.4 Scattering by a black disc of radius a.
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
Now we use the formula [valid for large ti and small BJ
Pn(cosB) = Jo(nsinB)
= Jo(nB), (8.106)
1
2
do"! 1 ka
dD = k2 xdxJa(xB) (8.107)
1
where J0 is the Bessel function of zeroth order. Using the formulae

d
zJo(z) = dzzJ1(z) (8.108a)
1 d
Jo(xB) = xB d(xB) (xBJ1(xB)), (8.108b)
we get,
dael = ___!__
dO k2
1 r:
lo
xBd(xB)Jo(xB)
02
12
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.
(= a4k2) when ()=O; it is zero when

ka()= 1
or
l/ka. () =
duel 4 2
Thus has the approximate value a k in the forward direction and
dO.
decreases rapidly in the angular region
O<():::; l/ka.
Thus
ael = a4k221r lor': sin ()d()
= 1ra2.
These features are observed in nuclear scattering and in particle physics
scattering at high energies.
8.8 Resonances, Dispersion Formula
A resonance is characterized by its mass (energy E0) and width r.

A decaying particle produced at time t is described by the wave function,
'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)
The Fourier transform of '1/J(t):
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:
P(E) = l'I/J(E)l2 = l'I/J(O)l2ñ2 . (8.114)

(E-Eo)2 + ~ 2
The unknown constant l'I/J(O)l2 is fixed by the normalization:
1-: P(E)dE = l. (8.115)

Using the formula

1ró(E - Ea) = lim ( Ef. )2 (8.116)
€-+a E- 0 +E 2
00
./
-oo
P(E)dE = l7/J(O)l2n2 f
-oo
00 -r2 [(
E-Eo)2
f + c2]
4
dE
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
ft(k) = -k1e•óe sinóe = k cot : e - i.k (8.120)

At resonance Ea, fp_(k) should be maximum i.e. at E= Ea, cot óe = O.
Let us expand cot 81', about E - E0:
cot óe(E) = cot ól'(Ea) + cot' óe(E) 1 (E - Ea). (8.121)

(E=Eo)
Put
cot' óe(E)I
(E=Eo)
= -i
so that near resonance
(8.122)
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 --; Eo, cot 8e --; O, (8.124)
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
J(()) = (2€ + l)fe(k)Pp(cose). (8.125)

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]
and i.> (2f + l)(~)2

(8.128)
V,Yf(O) = k[(E - Ea)2 + ~] ·
Using the optical theorem, near resonance, G'tot:
41r 41r (2f + 1)(~)2 (8.129)
G'tot = -k Imf(O) = -k2 r2 ·
(E- Ea)2 + 4
In general if resonance of spin J occurs in a scattering process
41r (2J + 1)(~)2
k2 (E_ Ea)2 + ~ ·
-t-J
\ G'tot =
This is called Breit.{Wigner formula. If both the projectile and the target
has spin J1 and Jz respectively, then
(8.130)
1 41r (2J + 1)(~)2

(J' - --~~~---:-= (8.131)
tot - (2J1 + 1)(2h + 1) k2 (E - Ea)2 +~ 2 •
8.9 Bound States and Resonances: Square Well Potential
We illustrate the concept of bound states and resonances, for scattering by

a square well potential.
E> o, V= -IVI, k2 = 2mE K2

= 2m(E + IVI) (8.132)
ñ,2 ' ñ,2 .
The Schródinger equation for a stationary state is
[ d~2 + k2] ua ( x) = O x < -a L (8.133) )
[d! +I< ]u.1(x)=O

2 2 -a<x<a~ (8.13tl) )
d2 +k-"] ·u2(x)
[ dx2 = O x > O. ( .135)
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)
B Boundary condition gives: (see

F
A = [2 cos 2ka - i(
= e-2ikªS(E)
zi· 4)
2e-2ika
f + -f<) sin 2ka]
(8.139)
where S(E) is called the S-matrix:
2
S(E) = (8.140)
[2 cos 2ka - i( f + -f<) sin 2ka] .
Now the transmissivity is given by
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)
Clearly k must be imaginary, otherwise no solution exists. Since k = v2-;;E

i.e. for poles, E must be negative. Now
cos2ka 1
. = -(cotKa - tan ka)
sm 2k a 2
1 cot2 Ka - l
----- (8.143)
2 cot ka
Hence from Eq. (8.142)
K k
cot2 Ka - i( k + K) cot Ka - 1 = O. (8.144)
The two solutions of the quadratic Eq. (8.144):
\~ ) i) kcoLKa = il( l- )
~ U(J'- ti) K cot,Jfo = ik. (8.145)
These solutions can be put in the form
i) KtanKa = -ikL
)
ii) KcotKa=ik. (8.146)
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)
Now near resonance

2
tan8(E) = f(E - Er)
d 2
dE[tan8(E)] = r· (8.156)
Thus
d8(E) zr
dE = [r2 + 4(E - Er)2]
d8(E) 1 = ~- (8.157)
dE E=Er r
8.10 8-function Potential
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)
Hence, we have the boundary conditions: u(x) is continuous, but ~~ is not

1 '{\
as give·n~q. (8.162).
In tlie regions x < O and x > O, V(x) = O:
u(x) = Aeikx + Be-ikx, X< O (8.163)
u(x) = ce=, x > O. (8.164)
The boundary conditions given above give
u(O) = A+B = C (8.165)
. . 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):
u(x) = A[eikx + B e-ikx]

A
= A[eikx + mVo e-ikx] x<O (8.169)
ikñ2 -mVo '
ikñ2
u(x)=A.kñ2 Vr' x>O. (8.170)
i -m o
The transmitivity and reflectivity are given by
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 ·
(¡,.,f\ S(E) regarded as~alytic function of k has a pole at
ikñ? - mVo = O
or
E=_ mVa2 (8.174)

2n2 •
Pole is at negativo valué of E. Mat

state. This can be een as foll.ow :
negative ~ E conespondb a bound
For
E<O, E=-IEI
k = iK, K = J ":i~EI.
2 (8.175)
Then the solutions in Eqs. (8.163) and (8.164) are replaced by

)
u(x) = Ae-Kx, x>O~
u(x) = Be-Klxl, X< Ü. (8.176)
Boundary conditions, namely u continuous and the one given in Eq. (8.162)
give
u(O) =A= B (8.177)
2m
-KA-KA= -V
ñ,2
0A (8.178)
so that
A#O, K= _ mVo (8.179)

f¡,2 •
Hence
E=_ mVo (8.180)

2n2 ·
i.e. ó-function potential has only one bound state.

8.11 The Born Approximation
Suppose we have a particle in momentum state IPi) at sorne early time

,2
ti(- - oo) i.e. it is in an eigenstate of Ha = with momentum Pi, andfm
we turn on a potential V(r) for t > ti and then we observe the particle at
a later t ¡ - oo when it is again in an eigenstate of Ha with momentum
P¡· This is what happens in a scattering process. Thus for ti < t < t¡,
the system is in a time dependent state 'ljJ(t) which satisfies the Schródinger
equation
rn8~~t) =(Ha+ V)'l/J(t) (8.181)

with the boundary conditions
'if;(t) = t1.p¡(r)e"Jtt?:p¡t
= up1 (r)e"
-•e
PJ
t
l< ti
t > t¡ (.12)~
~
where Ep = :m.
2
It is convenient to write the olution of Eq. (8.181) as
'l/Ji(r, t) = ~ appi(t)up(r)e -i/:pt L (8.183)8
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
j u;,(r)up(r)d r = Ó¡;p, 3 (8.184)
we obtain
iñapp;(t) = L ª~PJt)eT(Ev-E,,)t(p'IVI-+ p) (8.185a)
p
where
(PIVIP) = J d3rup(r)Vup(r). (8.185b)
Now integration over t, with the boundary condition

a~p(t) = Ó¡;p, t < ti
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,
and to the first order [known as Born approximation]
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)
where momentum eigenfunctions are given by

1 'I.Pi·r
u (r) = e-"-
Pi (2r,/i) ~
1 ip¡.r
uP¡(r)= 3e_"_· (8.192)
(21rñ)2
Hence from Eqs.(8.191) and (8.192):
TB = ---
1
(2rrñ)3
J ipf.r
e- " V(r)e_n_d3r
ipi.r
= - (2:ñ,)3 Je iirv(r)d3r (8.193)
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
0 dimension of length vil,L. Hence f(B, </J) in the Born approximation:

µ 1
!B(B, </J) = - ñ,2
J iq.r
e---,;:-V(r)d r
3
(8.195)
21r
= _ .r: J
21rñ2
eik.rv(r)d3r.
Now
q2 = PT + PJ - 2Pi·P¡ = 2p2(1 - cos (})
= 4p2 sin2 ~ (8.196)

2
where (} is the angle between Pi and P¡, i.e. the scattering angle.
For elastic scattering
2 2 2 2
Pi= P¡ = P =p · (8.197)
For a spherica(,,m:rr1etr.ic potential:
V(r) = V(r), ÍB((}, </J) = ÍB(B). (8.198)
Now r in polar co-ordinates: é, J
r = r(sinJsinÓ,cosJsinÓ,cosÓ).
Take q along Ó = O axis, 1{Jl = rq cos Ó
d3r = r2 dr sin ÓdÓdJ = 21rr2 dr sin ÓdÓ.
Thus, we have
fa(O) ~ -:, f V(r)r' [! ,<kroo,é,inÓdÓ] dr
__ i:
- ñ,2
J [ékr _ e-ikr]
ikr
2
r V(r)dr
00
= -;:,,, j sin(kr)rV(r)dr. (8.199)

o
For the limiting case of small velocities, si~;r ,. . , 1
fB(B) = -
2µ
ñ,2
t"
Ío V(r)r2dr. (8.200)
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
e,= 21r larr sinBdBlf(B)l 2
21r la,r ()d()jcI>(p 0)12. (8.202)
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(x)l2
ñ,2 ñ,2
(J"(X-=--
p2 2µE
or
lim (c,E) = constant, (8.203)
E->oo
154 Quontum Mechamics p

which gives the hight energy limit of the cross-section. As a first applica-
tion of the Born approximation, we apply it to scattering by the Yukawa
potential:
V(r) = g2e(-r/ro)
1
__ (8.204)
r
For this potential, from Eq. (8.199):
(8.205)
For Rutherford scattering i.e. scattering of a charged particle with charge

Z1e on a target of charge Z2e,
V(r)
Z1Z2e2 = (8.206)
r
For the Coulomb potential given in Eq. (8.206) take ro -..; oo, i.e. fo- -..; O
and g2 -..; Z1Z2e2 in Eq. (8.205) and hence from Eq. (8.206) we obtain for
the Rutherford scattering
()) = _ 2µZ1Z2e2 n2 = _2 Z1Z2e2
fB ( f¡,2 q2 µ q2 (8.207)
Note the remarkable fact, that for the Rutherford scattering, the depen-
dance on Planck's constant is canceled. Hence Rutherford scattering cross
section is given by
da 4(µc2)2 z12 z22e4 z21 z2e4
2
~~--=---=-- = µ2 (8.208)
dO. (cq)4 4p4 sin" !
This formula was first derived by Lord Rutherford in classical mechanics.
This is probably the only formula which is not modified in quantum me-
chanics. It is important to remark here that the formula (8.208) also agrees
with the result obtained from the exact treatment of the Coulomb scatter-
ing in quantum mechanics. However in the exact treatment, the scattering
amplitude is different from (8.207):
I
J(())= ( n ())e2i8bexp[in'ln(l-cos())] (8.209a)

pl-cos
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics
2i8 r(l - in')

e º=---- (8.209b)
I'(I +in')·
~~ gives the scattering cross section of a spinless particle of charge Z1 e
with a point particle of charge Z2e.
For o-particle scattering of heavy nucleus of mass M and charge Ze:
1 1 1 1
-=-+-~-
µ ma M ma
da Z2e2
dfl = ma p4 sin'' !. (8.210)
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
We now consider the scattering ofh,lectron from a charge distribution of

a spinless nucleus or proton, having structure. First ignoring the spin of
electron, from Eq. (8.193)
T = TB = - (2:n,)3 J e''::rv(r)d3r L (8.214)

'
V2V(r) = 41rZe2p(r). (8.215)
From Green's theorem:
J (</J'\7 2'1/J - '1(JV2</J) e- = j [cpV'l/J - '1/JVcp]&,S· (8.216) \,J

If '1/J and V'!/J vanish on large surface:
! [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):
je'':;' 'v2V(r)d3r = j V(r)V e '':;' 2 d3r
= (i:)2 J V(r)e '':;' d3r
= -
q2
n,2
J ,q.r
e----¡¡;-V(r)d r.
3
(8.218)
Hence, on using Eq. (8.215):
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
F(q2) = J e'1' p(r)d3r (8.222)
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)
For qR << 1 (R nuclear radius):

iq.r ( .q.r (iq.r)2 )
e ñ = l+zn+~+ "". ,
(8.225)
Now
.,..
j q.r dD. = O.
o
Thus, we have
F(q2) = 41r J p(r)r2dr - ~ J (~:)2 p(r)r2dD.dr
= 41r J p(r)r2dr - ~ ~: (4;) J r4p(r)dr. (8.226)
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"
8.12 Quasi-Classical Approximation
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 ~ ~
( - -:: V' +v(r) )u(r, t) ~ @u(r, t) (8.228) . l.f!._ej

.. .
Let us write the solution of Eq. (20.183) as

u(r) = w(r)eiS(r)/ñ, (8.229)
where w(r) and S(r) are real functions. Substituting Eq. (20.184) in
Eq. (20.183), we have
ñ2 1 .
- (- ñ2 (VS)2w + ~(2Vw · VS + wV2S) + V2w) + V(r)w = Ew.
2m
(8.230)
Separating out the real and imaginary parts, we get
2Vw·VS+wv2 sf10L (8.231)
V2w - ; (VS)2 + w(k(r)/{, O, (8.232)
where
k(r) = ./(2m/ñ2)(E - V(r)). (8.233)
Let us suppose that both w(r) and S(r) remain finite as ñ=:« O. We obtain
the classical result if we put = O, viz.
ñ
(VS)2 = (E- V(r)),

2~
so that if p = VS, we get
E= P2 + V(r).
2m
Thus, the semi-classical approximation is valid if
ñ2v'2w < < w(V 8)2. (8.234)
For one dimensional motion in the x-direction Eqs. (20.186), (20.187) and
(20.190) give,
d2S/dx2
--'---
dS/dx
= -2-dw
dx L (8.235)
)
ddx2
2w w (dS)
- ñ,2 dx 2 + w[k(x)] 2
= O, (8.236)
and
ñ2d2w/dx2 << (dS)2 (8.237)

w dx
The solution of Eq. (20.190) is
dS
ln dx = -2lnw + lnA
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)
Cbn of Eq. (8.228) for the one dimensional problem is
(nk1x))
112
{ C1 exp(i 1: k(x')dx')
+ C2 exp(-i 1: k(x')dx')}
= k {1: sin k(x') +a}, E> V(x) l (8.242) G

The region where E< V(x) is called the classically unattainable region. In
that region k(x) is an imaginary function and we have
k(x) = ilk(x)I
and requiring that the wave function be finite everywhere, we get the solu-
tion in this region
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.
160 Qt1antt1m M echanics
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)
whereas in the region x < a, it is given by

C" ¡a
u(x) = Jlkf exp(- Íx lk(x)ldx). (8.246)
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)
where F = -i~lx=a is a positive number. Thus near x = a, we can write

[absorbing ,¡ 2 in F]
(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)
one can easily see that cp(r¡) satisfies Bessel's equation
d2cp + _!_dcp + (1 - ___!__ )cp = o. (8.251)

dr¡2 r¡ dr¡ 9r¡2
The general solution of Eq. (20.205) is given by
(8.252)
This is the solution of the Schródínger equation near the turning point
x = a to the right of a. To obtain the solution near x = a to the left of a,
we simply change tt :» ilr¡I so that the solution is now given by
(8.253)
For large r¡, the behaviour of the Bessel function is given by
h1¡3(r¡)--+ (1/21rr¡)-1!2 cos(r¡ -F i- - ¡) (8.254)
i=r=1!3 h1;3 ( ilr,I)

--+ e 1/3 ( _!_ i1rlr¡l)-1;2 ( e-1111 eC=r=1r /6-1r/4)i + el11I e-C=r=1r /6-1r/4li) _!_
2 2
= (21rlr¡I)-! (elr¡I + e-l11le-(1/2±1/3)1ri). (8.255)
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
u= y112<jJ(r¡) -----* 3~(2mFy)-1!4 cos(r¡ - ~)

V'ff 4
A1 .
= 3 ~ sm(r¡ + n/4) (8.256)
vnk
provided A1 = B1, whereas the solution (20.207) takes the forro of Eq.
(8.246), i.e,
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) = ~sin(1x k(x)dx+ ¡), (8.258)
whereas to the left of a it is given by [c.f. Eq. (8.246)]
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
with C = (-l)nC'. Hence we have the quantum condition

b 7f
or
l a
k(x)dx = (n + l)rr - -
2
b 1
or
l ª p(x)dx = (n +
2)rrñ
f p(x)dx = 21rñ(n + i), (8.261)
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) ,
V(x)
Fig. 8.8 The plot resembling the effective potential V¡(r).
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
=(E: Vo) 1/2 exp (-2 lb (l(l ~ 1) - ~rr: E)1/2dr).

(8.267)
Now each time the particle hits the wall of the well, the probability of
!
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)
where T is the lifetime of the bound system. We note that
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)
(
= (2ma2)112 (Z(Z + l)n2)y'!(!+i).

(8.275)
E 2mEa2
We now apply the WKB approximation to s-wave bound states. For
the s-wave, the radial Schródinger equation is
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
x(r) = [N!v'kTr1] sin [Jr 1 k(r')dr' +1r/4], (8.276)
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)) ]
The boundary condition x(O) = O gives the quantisation condition.
for 1 k(r)dr + 1r/4 _J:}n1r (8.277a)
or for
1 [2µ(E - V(r)]1/2 drQ(n- i:)1rñ, (8.277b)
where n = 1, 2, 3, ... is the principle quantum number.

We now apply the above results for a linear potential viz. V(r) = (Jr.
This potential is relevant for the resonances J/'lj; (3100 MeV) and 'lj;' (3684
MeV). The conventional interpretation of the resonances is that they are
regarded as s-wave bound states of charmed quark and antiquark (ce called
charmonium). Furthermore, since quarks have spin !,
e ande spins may be
combined to form a total spin S, which is either O or 1 (Sec. 11.6). Since
resonances 'ljJ and 'lj;' have quantum numbers of a photon viz. JP = 1-, they
\ can be identified with 13 81 and 23 81 states of the charmonium. The mass
(¡v ~ of\d~anned quark is Lúgh (1500 MeV\1\2000 foV), therefore it i a good
\~ ~pSr'oximation to treatTu9und ce systern nou-relativistícally. It is assumed
\
~ that the potential V(r) = f3r kee ~em bound together.
For the linear potential the turning point r1 is given by
r1 = (1/ (J)E. (8.278a)
Let us put
r = (1/ (J)Ex. (8.278b)
Then we have
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)
Hence the mass spectrum for s-states of the charmonium is given by (µ =

!me)·
(8.280)
Forme= 1850 MeV and f3 = 1/(b2nc) with b = (1950 MeV)-1, one gets
from Eq. (20.230)
to be compared with the experimental value of 580 MeV.

We now determine the mean square radius of charmonium using the
WKB approximation. For this purpose we first determine the normalisation
constant N. Now
INl2 lorl k(r)

1 {
sin
[1r1 k(r')dr' + 47r] }2 dr
r = l. (8.281)
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)
The mean square radius is given by
(1Pslr211Ps) = lar¡ r2lx(r)[2dr

- N2 {r1 r2dr
- 2 lo k(r)
N2 E3 r,¡:;- ¡1 x2
= ~v 2µE lo (1 - x)l/2dx
= N2 E3 r,¡:;- r(1/2)r(3)
2(33 y 2µE r(7 /2)
8 E2
- 15 (32. (8.284)
Hence [with 2µ = me] and using Eq. (20.229),

) 8 E;,_
( 'l,Uns 1 r 21 'l,Uns = (32
15
(8.285)
8.13 Problems
8.1 The wave function
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
v:º ~,>~·L r-lt"' ¡

Treat the case for which the particle moves sufficiently slowly so
that C 2: 2 phase shifts are negligible. Show that
a~óo
- - tan(ka)
·a- tan ka
tanó1
· L + ka tan ka ,
k2 = 2mE / ñ2. If ka is small and the phase shifts 80 and 81 are

small, show that
80 ~ -ka,
Show further that
da
drl "'a2
( 1 - -3-
(ka)2 + 2(ka)2 cos () + O(ka)3 ) .
8.4 Consider the scattering of a particle at low energies (i.e. small k,

so that one can confine to s-wave only) by an attractive potential
V(r), the range of the potential being r = a. Suppose there exists
among a discrete spectrum of negative energy levels a bound state
(with angular momentum /i, = O) at energy E= -EKE >
O) where E
is very small. Show that the s-wave phase shift is l(v'en by
tan 80 = -VE/E
and
21rñ2
u(E)----
- µ(E+ E)
Sketch u(E).
8.5 Use Born approximation to find the angular distribution (that is
behaviour of u(()) with respect to ()) of the electrons in the elas-
tic scattering of electrons by an atom represented by a shielded
Coulomb potential
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
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
Hence show that

a=
2 ;n J 2
'l/J(r)V(r)d3r.
lim 80/k = -a.

k-+O
8.8 The scattering amplitude in terms of phase shift is given by
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.
For small phase shifts, e2i8e - 1 ~ 2ióg. Using the formula

sinqr ~ . 2
-- = ¿__,(2f+l)Pg(cos0)(Jt(kr)),
qr e
show that for small óg:
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 ~
l)
8.9 Consider the scattering of a particle of mass m by the potential /.ll('kv,..

V::º {>:~L . \l
Estimate the differential scattering cross sec ion for h2 k2 > > 2m V0
and discuss the validity of the approximation used.
In the limit ka<< 1, show that the cross section takes on the form
du
dD. = A+Bcos()
with B << A.
8.10 Consideran electron confined inside a sphere of radius R. What is
·I the pressure exerted on the surface of the sphere, if the electron is
in (i) the lowest s state (ii) the lowest p-state?
~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
where V is the potential between two particles and µ is their re-

duced mass.
Solution:
-? For s-wave, we have 'ljJ8(r) = R(r)Yoo = (l/v'41r)R(r) =
(1/v'47r)x~), where x(r) satisfies the radial Schréidingerequation.
d2x 2µ
dr2 +[,i2CE-V(r))]x=O.
Writing x" = ~:}, we have
.!!__( "/ ) = 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.
Integrating the left-hand side by parts, we have
j x[d/dr(x" /x)]xdr = xx" - (x')2.
Now
x'(r) = R(r) +rR'(r)L )

x"(r) = 2R'(r) +rR"(r).
Using the boundary conditions far a bound state wave function
R(r);
R(r) -r-,Q R(O) finite

R(r)-o,
r-,o
we have
1= x[d/dr(x" /x)]xdr = IR(O)l2.
Hence we have
IR(O)l2 = (2µ/n,2) 1= R(r)(dV/dr)R(r)r2dr
= (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,
where me is the mass of charmed quark. Show that the mean

square radius of the charmonium for the logarithmic potential is
given by
Hint: The quantisation condition is
larl [mc(E - Cln(r/ro))] 1 dr 1 2 = (n -1/4)1rn,
where r1 = r0eE/C. In order to perform the integration, put

(E/C - y2) = In(r/ro).
Chapter 9
OPERATORS
9.1 State Vectors
In developing quantum mechanics in the previous chapters we have intro-

duced two concepts: (i) State of a system, which has no classical analogueL
(ii) Dynamical variables, which are represented by hermitian operators,
called observables. Although they have a classical analogue ¡imtare treated
here quite differently. For example, in the Schródinger representation x is
an algebraic variable, the dynamical variable A is an operator A = A ( x, ;x),
( for example, the momentum operator p = -iñ ;x) and the state of a system
is represented by the wave or state function 'l/J(x).
We now want to evolve a more general formulism. In this formulism, it
is convenient to represent states in quantum mechanics by vectors, called
state vectors in a certain vector space, usually with an infinite number of
dimensions. This is because vectors have simple transformation laws. The
state function 'l/J(x) is then a particular representation of the state vector.
Following Dirac, we shall denote the state vector by 1), called the ket
vector and we shall label a particular one by i'lj.,). We also introduce abra
vector corresponding to the ket vector l'l/J) and denote it by ('!/JI.
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
í' (<phW = (,PI</>) L (9.1)

o
t? 175
Operators
:
In quantum mechanics, dynamical variables are linear operators. A

linear operator A is defined by the relation
(9.2)
i.e. a linear operator operating on a ket produces a new ket. Its linearity
is specified by the followingrelations
(A+ñ)17/J& = Al1/Jéi+Bl1faé) (9.3a)
A.(11/1) + I</>)) = A.11/1) + Al</>) (9.3b)
Á.(011/1)) = CA.11/1) (9.3c)
where C is a number. The product ÁÍJ of two linear operators A and ÍJ is
defined by
(ÁB) 11/1) = Á.(Bl1/J) ). (9.4)

In general
(Á.ÍJ)l1/I) i- (ÍJÁ.)11/1) (9.5)
i.e,
ÁÍJ i- ÍJÁ. (9.6)

In a special case when ÁÍJ = ÍJÁ, we say A and ÍJ commute.
Consider an operator A, such that
IF) = Al</>).
Perform the scalar product
(1/IIF) = (1/IIAI</>). (9.7)
In particular if C is a number
(1/JICI</>) = 0(1/11</>). (9.8)
Quantities like
are called matrix elements.

The Hermitian Conjugate .At of A is defined by
(1/JIF)* = (1/IIAI</>)* (9.9a)
= (</>I.Atl1/J). (9.9b)
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:
If there is an operator A such that operating upon sorne state tan), it

gives the same state, i.e. to say
(9.14)
where an is a number, then tan) is called an eigenvector of A belonging to
the eigenvalue ªn·
Theorem:
(i) The eigenvalues of a hermitian operator are real.

(ii) The eigenvectors of a hermitian operator are orthogonal.
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:
We assume that eigenvectors of an observable form a complete set so

that a state vector lvi) can be expressed as a linear expansion of the eigen-
states lan) of an observable A. Thus
lvi) = ¿ Cnlan), for the discrete case (9.15a)

n
= J C(a)la)da for continuous eigenvalues. (9.15b)

OPERATORS 179
Then for the discrete case
(9.16)
Therefore
(9.17)
Substituting in Eq.(9.15a)
(9.18)
n
Thus formally one can put

(9.19)
n
where i denotes the unit operator. Equation (9.19) is a formal statement

of the completeness condition.
For the continuous case we have from Eq. (9.15b)
(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)
l'l/J) = j la)(al'l/J)da. (9.21)
Thus in this case, the formal statement of completeness is
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)
where
S¿ = ¿ if a= an, discrete
= J ···da if a is a continuous variable, (9.22b)
EqlC9.17) or (9.20) provides a representation of the state 17/J) in the space

de~d by the basic vectors
(9.23)
and completely determine the state.
Similarly in this basis, an operator & is represented by the matrix elements
(anl&lam) = O:mn for the discrete case. (9.24)

Thus
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
Thus the matrix representation of product of two operators is the product

of matrix representations of the two operators.
In terms of state vectors the average value of repeated measurements of
an observable Á. when the system is in a normalised state 17/J) is given by
a,¡,= (7/JIAl?/J). (9.26a)
Using the completeness relation, we can write
(9.26b)
n m
n m
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.
9.2 Schrüdinger Representation
Let us take the basis vectors to be eigenvectors of the position operator x:

:i:J. ·) = ·lx) L (9.28a) )
{x1Jx) = 6(x' - x) L (9.28b) )
(x1l:i:Jx) = xó(x1 - x). (9.28c)
In this basis, the state lvi) is determined by quantities
(xlvi)
I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-Quantum Mechanics
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)
= JJ ('l/Jlx) (xlAlx') (x'l'l/J)dx' dx. (9.32a)
This reduces to our previous definition of a,¡, in the Schrodinger represen-

tation, namely
a,¡,= j 'lj;*(x)A ( x, !) 'lj;(x)dx, (9.32b)
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
f(x) :x 8(x - x') = -J'(x)8(x - x'). (9.41a)
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)
f(x')~8(x - x') = - of(x') 8(x - x'). (9.41c)

ox' ox'
Take f(x) = x - x'. Then from Eq.(9.41a)
(x - x')~8(x - x') = -8(x - x'). (9.41d)

OX
Therefore we can write Eq.(9.38) as

(x - x') (xlplx') = (x - x') (-in :x 8(x - x')) .
Hence
(xlplx'Q-inffx8(x - x')
''l:3-inffx(xlx'). (9.42)
The meaning ofthis equation is as follows [use ffx8(x-x') = - 8~,8(x-x')
and q. (9.41c)]:
(xl73Jx'}'l/J(x')J:1-i/i.· ( f;,; 8(x - x')) '1/)(:t')
tJ-ilió(x-x')~·l/;(x'). (9.43a)
Therefore when (xlplx') is multiplied with a function 'lf;(x') on the right, its
effect can be represented by [see Eq. (9.33)]
(xlplx') = (x - x') (-in 0~,)
. (9.43b)
Consider now more general matrix elements
(xlpl'l/J) = j dx' (xlplx') (x' l'l/J)
= j dx'fx-x') (-in0~,) (x'l'l/J)
= -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.
OPERATORS 185
9.3 Relation between the Momentum and Schródínger

Representations
Let IP) denote eigenvectors of the momentum operator

PIP) = PIP) (9.5la)"
(p'IP) =JP' - p) (9.51b)
(p'lfiip) = P¿J' - p), (9.51c)
so that in the momentum represcntation pisan algebraic variable p. Now
(xlfilp) = p(xlp). (9.52)
Using Eq. (9.44)
(xlfilP) = -iñ :x (xlp), (9.53)

therefore Eq. (9.53) gives
-iñ :x (xlp) = p(xlp). (9.54a)

The normalised solution of Eq. (9.54a) is
(xlp) = _l_e(ip/li.)x. (9.54b)
~
This is an eigenfunction of the momentum operator in the Schrodínger rep-
resentation. Now we can expand the state 11/J) in terms of the eigenvectors
of the momentum operator:
11/J) = J dplp)(pl1/J). (9.55)

Then taking the Schródinger representation
(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
11/J) = j dxlx)(xl'I/J). (9.57)

Therefore taking the momentum representation
(pl'I/J) = j dx(plx)(xl'I/J)
= -1
~
-f dxe-i(p/ñ)x(xl1/J). (9.58)
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
Yim(e, </J) = (e, </Jllm), (9.59)
where the simultaneous eigenvector llm) of L2 and L; satisfy
L21Zm)r:ln2Z(Z + l)IZm)
Lzllm) '-G)hmllm), (9.60)
l = O, 1, 2, · · · , m = O, ±1, ±2, · · · , ±l. In particular
(</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
Unlm(r,O,</J) = (r,O,</Jlnlm), (9.62)
where n specifies the energy eigenvalues

me4z2) (9.63)
En= - ( 2ñ2n2 '
ti = 1, 2, · · · ; n ~ (l + 1) for a given l. lnlm) are simultaneous eigenvectors

of H, L2 and t.;
Also we note that in the Schrodinger representation, the orthonormality
relations and completeness condition
(arnlan) = Ómn
L lan)(anl = .i (9.64)
n
beco me
or
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
9.4 Matrix Mechanics
An operator is diagonal in the space defined by its eigenvectors. Consider

for example the Hamiltonian H. Introduce two sets of basis vectors. The
eigenvectors IEn) form one set of basis vectors. In this basis H is diagonal:
HIEn) = E11IEn) L (9.68) )
(EmlEn) = 8,,..,, L (9.69) J
(EmlHIEn) = E,iÓmn.•
Using the completeness relation
(9.70)
n
H = LL IEn)(EnlHIEm)(Eml
n m
= LL IEn)Emómn(Eml
n m
(9.71)
Consider another set of basis vectors

n
_ L
(ailaj) - Óij (9.72) }
¿ iai)(ail = l. (9. 73)
In this basis
H =¿ ¿ laj)(ailHlai)(ail
i j
= ¿ ¿ laj)Hij(ail (9.74)
j
and since basis vectors IEn) and iai) form a complete set, one can write
IEn) = ¿ lai) (aj IEn)

i
(9.75)
where
u: = (ailEn)
Un/= (ailEn)* = (Enlai)· (9.76)
Now
n
(9.77)
Thus on using Eqs. (9.75) and (9.76)
(EmlEn) = L Uni(Emlai)
i
(9. 78)
Hence, we have
L Uni(ut)im = Uniutim = s.:

i
tnr' = 1. (9.79)
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
OPERATORS 189
Therefore
(EnlHIEm) = ¿¿(EmlaJ)(aJIHlai)(ailEm)
i j
= L L Un/ HjiUmi = L L o.; t HjiUmi

'ª
i j i j
= (Ut FIU}m.n· (9. 1)

Hence ifkamiltonian (or any other observable) is given in a basi: in whlch
it is not diagonal it can be diagonalised by a unitary transformation U
which is an overlap between the two basis vectors as given in Eqs.(9.76).
We note that
(ailHIEn) = En(ailEn)
or
¿(ailHlaj)(ajlEn) = En(ailEn)· (9.82)
j
Therefore
(9.83)
j
Hence the eigenvalues of H are given by
detlH - EII = O. (9.84)
This is also true for any other observable F(H -l F,E -l F).
Finally since both sets of basis vectmµ~n} and iai} form complete set~ ~ (:)
1 "'i
any arbitrary ket 11/J) can be written m terrns of them: !
IVJ) = L
!En) (EnlVJ) = ,6nlEn) L
(9.85)
n n
and
IVJ) = L lai)(ailVJ) = L ailai)·
.¡
(9.86)
Now, using Eqs. (9.80) and (9.85)

lYi = (ailVJ) = L ¿(EnlUin,6mlEm) = L Uin,6n
n m n
or
a= U,6 (9.87)
where a and ,6 are column matrices
ª1
a2
a= ,6=
9.5 Simple Harmonic Oscillator
The Hamiltonian operator is given by

A2 1
H
2m
= }!____ +
2
2. -mw2 x
(9.88)
We want to salve the eigenvalue problem, i.e. to find eigenvalues and eigen-
vectors of H without introducing any explicit representations for the oper-
ator.
Let us introduce
t
a= 1
~p+imwx (A . AL
v2mñ.w
l (A
a= ~p-imwx. A) o
(9.89)
v2mñw
Note that the operators a and at are not hermitian. Using the basic com-
mutation relation
[x,fi] = in, (9.90)
we see that
11.
[a,at] = .• {imw[p,x]-imw[x,p]}
2m11.w
1
= --2mñ.w = l. (9.91)
2mñ.w
It is easy to see that the Hamiltonian operator H is related to a and a t as
ñ.waat = H + ~ñ.wL )
1
ñ.wat a -- H - -ñ.w
2 . (9.92)
From Eq. (9.91), we can prove by induction

a(att - (at)na = n(att-1. (9.93)
Now
[a,H] = aH - Ha
= ñ.wa. (9.94a)
Therefore
aH = (H + ñ.w)a. (9.94b)
Similarly we have from Eq. (9.94a), by taking its hermitian conjugate,
[at,HJ=-ñwat. (9.95a)
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
= ¿ 1 (mlaln) 12 2:: O. (9.97)

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)
Let JO) denote the eigenstate of H belonging to the least eigenvalue of

H which we have denoted by A= Thus !nw.
HIO) = .\IO). (9.101)
Now
HaJO) = .\alO). (9.102)

But using Eq. (9.94a), we have from Eq. (9.102)
HaJO) = (aH - ñwa)IO) = (.\ - ñw)aJO), (9.103)

showing that unless alO) = O, aJO) is an eigenstate of H with eigenvalue
A - ñw < >., which contradicts that A is the least eigenvalue. Thus we must
have
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)
showing that (at)nJo) is an eigenvector of H belonging to eigenvalue
(9.109)
OPERATORS 193
Consider [use Eqs.(9.93) and (9.104)]

(Olan(atrlo) = (Olan-1a(atr10)
= (Olan-l(n(at)n-1 + (at)na)IO)
= n(Olan-1(atr-1IO). (9.110)
Repeating this process we find
(Olan(atr10) = n(n -1) · · · l(OIO)
= n!n
Therefore the eigenstate (at)nlO) is not normalised, but }.;-y(at)nlO) is nor-
malised. We denote this normalised eigenstate of H belonging to eigenvalue
(9.111)
f
!)
En = (n + ñw by
(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)
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).
9.6 Supersymmetric Oscillator
The simplest example of a supersymmetric system is provided by simple

harmonic oscillator for which the Hamiltonian is [ñ = 1, m = 1, w = 1]
1
HB = -(p2 + q2)
2
1 . . 1
= -./2(q + ip)(q - ip) - 2[q,p]
1
=ata+2. (9.117)
where at and a are creation and annihilation operators which satisfy com-
mutation relation [a, atJ = l. The eigenstates of HB are IO)B, ll)B, l2)B, ....
alü)B = Ü~ /
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)
Ylflf (A) 'rlf + lflf (A!) (rn) \y (g) '?Y (!)

'rlf - ?Y
srorarado '.iJU!MO{{OJ
cll['.) JO l[:)!l[M_ 'SlO'.)'!llcldO U'RrHUilcll[ Al'Rl'.)!ql'R OM'.) cll'R lf pu'R lf 1'6
swa¡qo.1d L"6
' ( :)!J'.)clUIUIASladns)
s-? - 1
g-d +1
1_(g_c> - 1)(g_c> + 1) =
. . + F)¡;-'J('; + F)¡;-'JZ + F)-'JZ + 1 = HFJ_'JHJ,J,
[(TZ1'6)·b3: U! pa!ldUI! A:)'8lc1Ucl1Jap '.)UTIO:):)'tl 0'.)U! '.3U!)f8'.l] pu'R

(:)'Rl!cM!Uila._iI) F)-'J +1= .!IHFJ-c>ªH.1,J,
'1 'o = du san¡'RAUa'.il!a cll['.) U! «n lOJ AJl'RI!UI!S

o-? -
-=----- =
1
1
d Aq UclA!'.il S! lO'.p'RJ UU'RUIZ'.)lºH J'R:)!'.)8!'.)'8'.)S aq'.) , ... 'z '1 'o

= flu ¡.m¡'eAIIa~!"'l e (ul B!S'Rq aq'.) U! xuraur l'RUO'.il'R!P '8 S! vn
a:)U!S AI{'RU!._iI
·a'.)nUIUIO:)-!'.)U'8 l[:)!l[M smw1aua~ sarmbar dnoas aUI!'.)
-a:)'8dS aq'.) JO '.)Uaura'.arnyua AU'R '.)'Rl['.) MOl[S SUO!'.)'RlclP!SUO:) aAoq'R aqJ, ·aAoq'R
pauonuaur A:)maua~ap U!'RJdxa 'paApap AT!S'Ra aq U'R:) l[:)!l[M 'suon'tl¡a1 asaqJ,
\ '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
U!9 X U!6 - ~009 ~!J!lUaps PIJOM 6l'vl ztoz '6 ,í¡nr I

(e) have pure imaginary expectation values,

( d) are purely numerical operatorq-í
9.2 If an operator A has the property that
cv ~hat are its eigenvalues?

9.3 The Hamiltonian
H =E_+ V(x)
Á.4
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
OPERATORS 197
(i) N = Á_t A is hermitian.
(ii) N2 = N. Hence show that N has eigenvalues O and l.
(iii) [N, Á_tJ = Á_t.
(iv) Let IO) and 11) be eigenvectors of N belonging to eigenvaluesO and

l. Show that
A.tia) = 1
.A_t11) = o.
9.7 Let ÍJ and 6 be two anticommutating operators
Let lvi) be a simultaneous eigenstate of both ÍJ and C. What can

be said about the corresponding eigenvalues? If 02 = 1 holds,
what <loes it imply for the eigenvalue of ÍJ?
(lx2IO) = {h/2v'2mw, for n = 2 (9.124)

O, for n-=/= 2
Further show that
(Olx210) = _n__
2mw
9.9 In the three dimensional vector space with an orthonormal set of
7 basis vector {11), 12), 13)} ~
(a) Find the matrix representation of the following operators
11)(11 ' 12)(21, 11)(21-12)(11, 11)(31- 12)(31 + 12)(11

211)(11 - ~13)(21 + il2)(21.
(b) Which of the above operators are hermitian?
(e) Let A= il1)(21 - il2)(ll + 13)(31. Show that
find the eingevalues and e~genstatesof A~
(d) Show that eigenstates of A are orthogonal.
A is hermitian and o
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics 1
9.10 For the Hamiltonian
[pi,H] = -in88 V(r).

Xi
Using the above result, show that
[pj, [pi, H]] = (-iñ)2 V(r).

88
Xj 88
Xi
From the relation:
(ml[pj, [pi, H]]lm) = (-iñ)2(ml (a!j a!i V(r)) lm),
tv the sum rule
¿(Em - En)(mlftiln)(nl1\lm) = - ~2 (mi V2 Vlm),
n
then show that
¿(En - Em)l(P.lmnl2
n
= ~2 f l'l/im(x')l2 'v2 Vd3x
= ~
L
41re2 Zl'l/im(O)l2
where for the Coulomb potential
v2V = 41re2 Z83(x)
9.11 Show that 3 x 3 matrices
satisfy the relation

(Gi)jk = -iriEijk
o
(GiGj - GjGi)mk = -iriEijn(Gn)mk
i.e.
[Gi, Gj] = ÍriEijkGn,
showing that matrices G's satisfy the commutation relation of an-
gular momentum and as such gives the adjoint representation of
group 03 and represent spin l. Write these matrices explicitly:
G1.=fi(~ o b ~-i), G2=ñ(~ ~~)

-i () oo -i
e, = li (~ ~i b)
o o o
OPERATORS 199
Show that the eigenvalues of G3 are ñ, O, =ñ.
J In the basi~ in which G3 is diagonal the corresponding eigenvectors

lx) are:
jl) - m. m. m.JO) - 1- 1) -
Show that the eigenvector!:M IY) in the basis in which G3 is not
diagonal are:
]GF+l)- ~ (~!) JG,-0)-G)
JG,--1)- ~ G).
Now the matrix which connect the two bases is
Show that
-1 i O )
U=~
(
O o_J2 .
-1 -i O

Chapter 10
HEISENBERG EQUATION OF
MOTION, INVARIANCE
PRINCIPLE AND PATH
INTEGRAL
10.1 Introduction
In quantum mechanics, dynamical variables are operators which do not in

general commute e.g.,
[x,ft] = in.
In the Schródinger representation x ---+ x (i.e, it is treated asan algebraic
variable)
(10.1)
Now dimensionally
[x][p] = ( Ac~on).
Also we see that dimensionally

[tl[E] = [n].
We would expect time t to be always treated as an algebraic variable, thus
in analogy with Eq. (10.1) it is plausible to postulate
H ---, in_!!__tt (10.2)
at• 1,
where t is treated as an algebraic variable and the sign chosen is conven-
tional but turns out to be convenient. The above equation is an operator
equation which operates on time dependent state vectors lllT(t)). Thus
in :t lllT(t)) = HlllT(t)). (10.3)

'
We now show that in the Schréidinger representation it leads to the ordinary
Schréidingerequation for the state function
w(x, t) = (xlllT(t)). (10.4)
201
From Eq. (10.3), we have
in:t (xl\J!(t)} = (xlHI\J!(t))
= j dx(xlHlx)(xlw(t))
where
(xlHlx) = t5(x - x)H(x, =iñ :x) (10.5)
so that we have in the Schródinger representation
in :t (xl\J!(t)) = H(x, =iñ. :)(xl\J!(t)). (10.6)

A particular solution to this equation is
(xl\J!(t))En = e-iEnt/liun(x) (10.7)
where un(x) is the energy eigenfunction:
H(x, -in :x)un(x) = Enun(x). (10.8)

We may write
(10.9)
so that from Eq, (10.7)
(xl\J!(t))En = e-iEnt/li(xlEn)· (10.10)
A general solution of Eq. (10.6) is the linear sum of the above solution
(xl\J!(t)) ¿ e-iEnt/li(xlEn)a(En) (10.lla)
(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
HEISENBERG EQUATitO~, INVARIANGE PRJNGIPLE AND PATH INTEGRAL203
10.2 Heisenberg Equation of Motion
Schrodinger's equation of motion is
in !t [w(t)) = Hlw(t)). (10.15)
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
= dx'ljJ*(x, t)A(x, ax)'I/J(x, t). (10.16)/
This is of course a function of time.

Now in classical theory we have no concept of state, and we deal with
dynamícal variables which change with time. We now want to go to a dif-
ferent picture of motion, in which states correspond to state vectors which
are independent of time and dynamical variables correspond to time de-
pendent linear operators. We shall see that this new picture has a classical
analogue and is more in the spirit of classical description.
The trick to go from the Schródinger picture to the new picture is simple.
We note that Eq. (10.15) has the formal solution
[W(t)) = e-iHt/ñ[w(O))
= e-iHt/ñl'I/J). (10.17)
Note that the Hamiltonian H is a hermitian operator. Thus
ªw(t) = (w(t)IAJw(t))
= ('1/JleiHt/ñ .Ae-iHt/ñl'I/J). (10.18)
Let us now define the time dependent linear operator
A(t) = eiHt/ñ.Ae-iHt/ñ. (10.19)
Then
ªw(t) = ('I/II.A(t)17P) = a(t),¡,. (10.20)

The quantity on the right-hand side represents the average value of the
time dependent operator, for a state 17P) specified at t = O. This picture of
motion is quite different from the Schródinger picture. Here the operator
A(t) represents the making of an observation at time t on a state which is
specified at t = O. This new picture is called the Heisenberg picture. This
is much closer to the classical description, and one may expect the time
dependent operators A(t) to be rather closely related to the corresponding
time-dependent classical dynamical variables.
Let us differentiate Eq. (10.19) with respect to t. For this purpose we
rewrite Eq. (10.19) as
e-iHt/ñ .A(t) = Ae-iHt/ñ
l ó
Thus
- iH e-iHt/ñ .A(t) + e-iHt/ñd.A(t)

n dt
= .A~i He-iHt/ñ.
Multíply both sídes on the left by inéHt/ñ
eiHt/ñ H e-iHt/ñ .A(t) + in d.A(t)

dt
= eiHt/ñ .AHe-iHt/ñ
= eiHt/ñ Ae-iHtfñeíHt/ñ H e-iHt/ñ.
l
But
eiHtfñHe-iHt/ñ = H(t) )
eiHt/ñ Ae-iHt/ñ = .A(t). (10.21)
Therefore
ind~~t) = .A(t)H(t) - H(t).A(t)
= [.A(t), H(t)]. (10.22)

In particular for .A(t) = H(t). Then
ind~?) = [H(t),H(t)] = O (10.23)

HEISENBERG EQUATION OF MOTION, INVARIANCE PRINCIPLE AND PATH INTEGRAL205
thus
H(t) = H(O) = H. (10.24)
i.e H is constant of motion.
Thus we can write
iñd~?) = [A(t), H]. (10.25)

This is known as the Heisenberg Equation of Motion.
Any observable which commutes with the Hamiltonian is a constant of
motion, since then
iñd~?) = [.F,H] = O.
This implies that
!(7/JI.F(t)17/J) = (7/JI: 17/J) = o

i.e. the average value of an observable which commutes with the Hamil-
tonian <loes not change with time. In particular if the system is in an
eigenstate of F at t = O, the state will be an eigenstate at any subsequent
time.
To sum up
S-picture H-picture
States represented Moving or time dependent fixed or independent
by vectors of time
iñcftl\[!(t))s iñcftl\[!)H = Ü
= Hl\[!(t))s 17/J)H = l\[!(Q))s

·tcdA(t)
Dynamical iñ~{ = o in--¡¡¡-
variables are A= A(O) = [A(t),H]

represented by
linear operators
The Heisenberg equation of motion has a classical analogu€ff in partic- )

ular it is analogous to classical equations of motion in the Hamiltonian
form.
One way to make/ transition from classical mechanics to quantum me-
chanics by correspon8ence principle is discussed in ch. 2. An alternate (Ár_s
6i)
:::.-
way to make this transition is as follows. Consider a dynamical variable

F(pi, Qi, t), which is a function of generalized coordinates Qi and conjugate
momenta pi:
dF aF er . er .
dt = 8t + ~(aqi Qi + ap/i)· (10.26)
i
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)
We note that for

(10.29) o
F=H, dH =O
dt
i.e. the energy is conserved. (F, H) in Eq. (10.29) is the classical Pois-
son Bracke~B). In general we define the Poisson bracket between two
dynamícal variables A and B:
8A8B 8A8B)
(A,B)PD '"""' ( --- - -- (10.30)
~ aqi Bp;. ªPi aqi
L.!::-==',,.........:::;
(A,B)p -(B,A)PB·
We note that the Eq. (20.32) is analogous to Heisenberg's equation of
motion (10.25) if the classical PB is replaced by the quantum PB defined
by (Dirac)
A 1 A
(A(t), H) = ln [A(t), H] (10.31)
and the dynamical variable .A(t) and the Hamiltonian are regarded as op-
erators.
Thu · the rule to go from classical mechanics to quantum mechanics is

to tr at dynamical variables A and B as hermitian operators and replace
~t;3-ssicaJ poisson bracket by the quantum commutator:
1 A A
(A, B)PB --t ¿/¡, [A, B]. (10.32)

Let us apply this rule to canonical coordinates Qi,Pi· Then from
Eq.(20.33):
(qi(t),qj(t))PB = 0 = (Pi(t),pj(t))PB
(qi(t),pj(t))PB = ~
L.,
( ~~
oqi ºPí - ~~
oqi opí ) = ~
L., (óirÓjr) = Óij· (10.33)
Thus the rule (20.33) gives

r UQr op;
[qi(t), 1/j(t)]
op; uqr
=o= [pí(t),Pj(t)]
r
l )
[qi(t),Pí(t)] = iMij· (10.34)

These are known as canonical quantization conditions. Note ~\impor ant
fact that these hold for equal timtprob. 10.lJ
10.3 Free Particle Propagator
As an application of the Heisenberg picture, we calculate correlation func-

tion
(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,
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
('if;, -tlq, t) = ('if;lef.H(-t)ef.Htjq) = ('if;jq) = (qj'if;)*

= 'lj;*(q)
while
(q, tj'if;) = (qle--t.m¡'if;)

= (qj'lj;(t)) = 'lj;(q, t). (10.38)
Thus
K(t2, t1) = (q2(t2)lq1(t1))H

= (q2(t2)lef¡Hti jq¡).
~2
Now for a free particle, H = }!___ and using i>IP) = PIP), we get
2m
Then
Further using
we get
=
1
21rñ J dpe-K
'p2
2"'
t '
eKPQ, (10.39)
where we have used (q2lp) = -1-e-kPq2 (plq1)

' and t
= -1-e--kPQi =
~' ~
t2 - ti, q = Q2 - Q1. Thus
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 .
10.4 Unitary Transformation
The transformation (10.17) or (10.19) vizL

1 \[I (t)) =
-1/Jt.
e ---¡¡-----11,b}
(J
= ut¡~) (10.41)
A(t) = U.Aut (10.42)
where
iHt
U =e" (10.43a)
ou' = 1 = iro (10.43b)
is an example of unitary transformation. Here H is the generator of the
unitary transformation:
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics !
Under the above unitary transforrnation we go frorn one physical picture

(Schrodínger) to another physically equivalent pictureg!eisenberg).
In general a unitary transformatíon can be written as
(10.44)
where E is real and F is hermitian so that
ut = e-ie.frt
= e-iEF' = o+.
Consider a state represented by the state vector l'lj!). We can form a
new state by making a unitary transformation.
l'lj!u) = Ul'lj!) L J
('lj!ul = ('lj!!Ut. (10.45)
If l'lj!) is normalised so that
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.
HEISENBERG EQUATION OF MOTION, INVARIANCE PRINCIPLE AND PATH INTEGRAL2ll
The unitary operators enable us to make transformatio1!1'om one de- .S

scription of a system to another physically equivalent description. Under
unitary transformation we define the transformation law on an operator as
Au = UAUt. (10.49)
Equations (10.45) and (10.49) define unitary transformations.
Theorem 1:~der a unitary transformation eigenvalues of an operator
remain unchanged.
Consider the eigenvalue equation
(10.50)
Then
U Autu¡an) = anUlan)
Aula~)= anla~)
proving th theorem,
Theorem 2:t91e av rage valu of a large nurnber of measw·emen{.gr A on S.
a system in tate IV') remain unchanged under a unitary transforrnatíon:
a,¡, = (1/JIAl'l/J)
= (1/JIUtuAutUl'l/J)
= (1/JulAul1/Ju)
=a,¡,v.. (10.51)
Consequences:
A,1 unitary transformation does not change the physics; it is just a
transformation from one description to another physically equivalent one.
10.5 Invariance principies, Coservation laws
In quantum mechanics, a transformation is associated with a unitary oper-

ator:
U = eieF pt = F. (10.52)
'
F is called the generator of transformation.
Consider the matrix element (JIHli) ora transition from an initial state (~
li) to a final state lf) by an operator S vi~e matrix elements (JISli). \:..)
Now
(JISli) = u1utusutu¡i)
= (r1usut¡iµ) (10.53)
which is an identity. However if

uso' = s (10.54)
then
(JISli) = (rlSliµ). (10.55)
O Hence invariance of S, viz~q. (10.54) implies that the result of an exper-
iment remains unchanged in a situation where the states are transformed.
For example, for a closed system, due to homogeneity of space, there is no
preferred origin and we have the invariance under unitary transformation
corresponding to translation in time and space implying that the result of
an experiment will be same now, in the past and future; here and elsewhere.
Thus if we examine the light emitted by a quaser, we find that the atoms
were behaving (i.e obeying the same laws) billio1t9f light years away as they
are now. Similarly for a closed system there is no preferred direction due
to isotropy of space and hence we have r~tational invariance. As a result
the generators of these transformations vi2t..91ergy, momentum and angular
momentum (see next section) are conserved for a closed system. To see
this, from Eq. (10.54):
o
usoso = su
or
US = SU; [U,S] =O= [U,H] (10.56)
i.e S-matrix and H commutes with U. Under infinitesimal transformation
U= 1 + iEF, (10.57)
and Eq. (10.54) gives
i.e,
[F,S] =ü. (10.58)

Hence the invariance under transformation U means, the generator of trans-
formation F commutes with S-matrix or Hamiltonian.
FIFi) = FilFi)
FIF¡) = F¡¡F¡)
(F¡l[S, F]IFi) = (F¡ISF - FSIFi)
(Fi - F¡)(F¡ISIFi)· (10.59)
Hence, t<Ji1m Eqs. (10.58) and (10.59)

Fi = F¡, if (F¡ISIFi) i- o. (10.60)
i.e eigenvalues of generator P are conserved.
Examples
1- Under translation in time
17/J(t - T)) = e-iH(t-T)l7/J) = eilff-l7J;(t))

-iH,
17/J(t + T)) = e-----,;:-17/J(t)) (10.61)
and
iH, " -iHT
A(t + T) = e-r.. A(t)e--¡¡-. (10.62)
Thus Etary operator corresponding to the translation in time t ------, t + T is
iH..-
U = e1'.
Invariance under translation in time implies the generator of translation in
time viz!Jr
is conserved.
2- Translation in space: [for simplicity, consider one dimensional case;
then generalize to 3-dimensions]:
x-x+a.
Associated with translation in space is a unitary operator Ur:
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.e
(xlUTIVJ) = (xlVJT) = (xll + a :x + ···IVJ). (10.66)
But -in tx = p, the momentum operator. Thus

1.ap
UT = e--,¡-
i.e the momentum operator p J generator of translation in space [For an ~

alternate approach see Prob. 1cf.l5r
For 3-dimensional case
~
X-----+ X+ a; UT =e--,. (10.67)
Hence if
[f>, H] or [f>, S] = O (10.68)
then momentum is conserved i.e translational invariance implies conserva-
tion of momentum.
No [u ing prob. 10.3]
UTxut = e~ xe -~ap (10.69)

ia h2
x + h[p,x] +
21¡p, [p, x]] + .....
A A A A A A
= 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)
HEISENBERG EQUATION OF MOTION, INVARIANCE PRINCIPLE AND PATH INTEGRAL2l5
is also translational invariance, [f>, H] = O.

The third example is the periodic potential V(x) = V(x + a) a being )
lattice constant. Let us write unitary transformation Ur as T(a) for lattice
T(a)lx)f=\lx - a}}__ (10.75) )

rt(a)lx6{x + a). (10.76)
Obviously the Hamiltonian with periodic potential is invariant und r attice
translation.
(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
For this state, using Eq. (10.78) and (10.79)

Hlka) = ¿ einka Hin)= Eolka) (10.82)
n
while
()(]
T(a)lka) = ¿ einkªln - 1)
n=-oo
n=-oo
(10.83)
Hence !ka) is simultaneou~enstate of H and T(a) with eigenvalues Ea

and e•ka respectively. Now from Eq. (10.83):
(xlT(a)ika) = eikª(xlka). (10.84)
But on using Eq. (10.76)
(xlT(a)ika) = (x + aika). (10.85)
Hence
(x + alka) = eikª(xlka) (10.86)
or
'1/Jk(x +a)= eika'l/Jk(x). (10.87)
Eq. (10.87) implies that '1/Jk(x) can be written in terms of Bloch function,
Uk(x):
(10.88)
To see this, we note that Bloch function has the property
(10.89)
Thus from Eq. (10.88)
'1/Jk(x +a)= eik(x+a)uk(x + a)
= eikxeikauk(x) = eika'l/Jk(x).
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
x/ = R.jxjRikXk
= (R)ij(RT)kiXjXk
= (RT)kiRijXjXk = x?. (10.92)

Hence
(RT)kiR;.j = Ókj L )
RTR= 1
(RT R)kj = Ókj
= RRT.
L- (10.93)
)
For example, for rotation about z-axis, by an angle ()

X= + y sin ()
X COS ()
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)
'l/JR(x) = (xlUR['l/J) = (x - w x xl'l/J)

= 'l/J(x - w x x) (10.99)
By Taylor expansion
'l/JR(x) = 'lj;(x) - w.(x x V)'lj;(x), (10.100)
so that
UR(w) = l -w.(x x V)
z
= 1- ,¡:w.L.
Exponentiation gives for finite rotation,

UR(w) = e--;;w.L. (10.101)
where L = -iñx x V is the orbital angular momentum operator. It is easy
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)
which is invariant under rotation, so that [L, H] = O
10.6 Discrete Transformation
In section (10.5), we have discussed the continuous transformations. Space

reflection and time reversa! viz x ---+ -x, t ---+ -t are discrete transfor-
mations:
i- Space reflection
X---+ -X (10.106)
HEISENBERG EQUATION OF MOTION, INVARIANCE PRINCIPLE AND PATH INTEGRAL2l9
Associated with this transformation is a unitary operator P, called the

parity operator. Thus
Flx) = 1-x)
PP!x) = P¡ - x) = [x) (10.107)
Hence
p2 = 1 (10.108)
i.e P has two eigenvalues + 1 and -1. P is both hermitian and unitary:
(10.109)
Now
F1jJ(x) = 1/J(-x) = (-xl1/J) = (xlFl1/J). (10.110)
Any ket 11/J) can be expressed as
11/J) = ~[(1+ P)11/J) + (1- P)11/J)l
= l'l/J+) + 11/J-) (10.111)
where
11/J+) = (1 + P) l'l/J) (10.112)

2
11/J - ) = (1 -2 P) 11/J) (10.113)
and have the property (using Eq. (10.108)):

Fl1/J±) = ±11/J±)·
Hence 11/J+) and 11/J-) are eigenstates of P, with eigenvalues +1 and -1
respectively. Thus 11/J+}. and [,¡j,-/ forma complete set, so that any arbitrary :JJ.
state can be expressed .¡; 1em\ofthem.~m Eq. (10.110), it follows that '/
tC.~\ A F1/J+(x}H+(-x}H+(x)j__ (10.114) )
~ P1/J_(x--Gh-(-x~1/J-(x). (10.115)
If Hamiltonian H is invariant ;~;-y'~~ reflection, then ~
PH&IIP
i.e
[F,HJ=O (10.116)
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
t -----+ -t. (10.117)
®
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. ~
Under time reversal, for a ket vector l-rp(t)): / ¿;¡.l>ff ~

¡,,¡/t)(t }--1. r 17P(~'l/J(-t)I~ (10.119)
(xl7j;Ctl(t)) 7j)(-t)lx~l1Ji(-t))*. (10.120)
Thus
7j;Ctl(x, t) = 'lj;*(x, -t).
or
7j;Ctl(x, -t) = 'lj;*(x, t) (10.121)
HEISENBERG EQUATION OF MOTJON, /NVARIANCE PRINCIPLE AND PATH INTEGRAL221
The time dependent Schrodinger equation is

iñ :t '1/J(x, t) = H'lf;(x, t). (10.122)
Taking the complex conjugate (H is hermitian)
-iñ:t'I/J*(x,t) = H'lf;*(x,t).
Changing t ---, -t
iñ :t '1/J*(x, -t) = H'lf;*(x, -t).
Then using Eq. (10.121)
iñ :t '1/Jt(x, t) = H'lf;t(x, t)
i.e. the Schródinger equation is invariant under time reversal.
10.7 The Path Integration Formulation Of Quantum

Mechanics1
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¡
S([q], ti, t¡) = J

t¡
dtL(q, q). (10.124)
All fundamental laws of classical physics can then be understood in terms of

the action S. Furthermore there is one to one correspondence between the
symmetries of the action and existence of conserved quantities. Another
important point in this approach is that time is treated ~ equal footing
with the other coordinates.
1 For Further reading and more details, see
(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 $.
10. 7.2 Path Integral Method

Just as in classical mechanics one can go from Hamiltonian formulation
to the Lagrangian, or vice versa, we start from the Schródinger formal-
ism, based on Hamiltonian, and make a transition to one based on the
Lagrangian. An important quantity to calculate in quantum mechanics is
the transition amplitude.
(10.125)
where states involved are in the Schródinger picture at time i t and tF, and
i'lj!(t)) satisfies the Schródinger equation.
ifi :t i'lj!(t))s = Hl'lj!(t))
which has a a formal solution
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.
Then the transition amplitude from a point q1 to a point QF in time T =

tF - t s is given by
x= (qF¡e--i\:HT¡q¡)
= (qF, tFIQ1, t¡) (10.126)

where qt = q(O) and QF = q(T). Divide T into N segments, each lasting
T
ót = -. Then
q
()
x= rr J
N-1
i=l
dqi(QF¡e-*Hót¡QN-1)(QN-1le-iH8t¡QN-2) .....
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)-
To proceed, a convention is selected that after the expansion, all p operators

will be moved to the left of all q operators, the effect of the commutator
[132, V(q)] is 0(8t2) and is irrelevant for 8t ----+ O. Then using (qH1lp)
_1_ej¡Pqi+1 (plq·) = _l_e--f¡pq;
,/2rli ' J ,/2rli ,
I·
1
21rñ
= - 1 J
dp eKP(qi+1-q;)e-¡¡ót[2=le-,;8tV(q;)_
i i p2 i
Then using the Gaussian integral
(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)
where q(T) = q¡ and q(t) = QF· Then

lC = (qF leif (t-T) lqr)
= J Dq(t)efi- {[!mri2-V(q))
{'f L(q,q)
=
J Dq(t)e T (10.131)
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.
HEISENBERG EQUATION OF MOTION, INVARJANCE PRINCIPLE AND PATH INTEGRAL225
10. 7.3 Free Particle Propagator from Path Integral For-

mulism
To illustrate how one applies path integral formalism, we calculate the free
pa.rtic propa .ator which we have already calculated in ec. (10,3) .,
C::: · e now calculate it by t C pa 10 ·egral rnetho where = ,¡mq and from
Eqs. (10.129) and (10.132)
-im . )/f
!(
i ~ 1n ( )2
K(qp,tp,q1,t1) = dq, .... dqN_1eKL, m QH1···Qj .
21rMt
Consider
J dqje2;h~,((qj+1-qj)2+(qj-Qj-1)2l = J dqjefim[q;-!(qj-1-qH1)l2.e~(qH1-qj-1)2
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
K(qp,tp,q¡,tr)= ( 21ri8tñ (j+l)}1

m )~N-l(21rij8tñ)! im
e'J.Not(qp-qr)_
2
Now N8t = tp - t i = T, q = qp - qt while
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)
226 Quantum Mechanics Í,~

which agrees with the one found in_§.ec. (10.3) ~
This example also brings out th? role of classical action. The classical
equation of motion for a free particle is
ij =o
which with boundary conditions q(O) = qi and q(T) = qp has the solution
[T = t r -t1]
q(t) -- qp T- q¡ t + q¡ .
Thus the classical action is
{T 1
Se= lo 2mq~dt
= {T !m (qp
- q¡)2 dt
lo 2 T2
m (q¡ - q1)2
2 T
so that
en
i.5
e = en
i. m(qF-q¡)2
(10.139)
2T
L O.~é§)
It is apparent that the argument of the exponential appearing in the free
particle propagator Eq. (10.138), is given precisely by the time the value *
o
of the action along the classical trajectory qc(t). It can be shown that this
is general feature of all quadratic actions [see problem 10.8] L
10. 7.4 Motion with a Source: Generating Functional for
Time-ordered Products of Heisenberg Opemtors
'li(t)
We now discuss the path integral describing a particle moving in the po-
tential V(q) and undergoing external time dependent force J(t) for which
1
L(q, q) = - V(q) + J(t) (10.140)
2mq2
and
H
Following the procedure in~ec. (10.7.2),
= r:
2m
,2
+ V(q) - J(t)q. (10.141) §)
~~
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)
We put V(q) = O and then

ZFI(J) = C1r:c5t) N/2 j dq, ....dQN-1e* I:f=</[*,(qJ+1-q;)2+q;J(t;)8tJ.
(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
N-1 N-1 N-1 1

= q1 L J(tj) + L (qj+1 - qj) L J(tk) l-_J10.144)
j=O j=O k=j+l
Thus the full exponent in Eq. (10.143) is
* t { %' Q¡ .l(t;)8t + ;;,tQ;H - Q;)' + (<¡;+; - Q;) k~k J(t,)Jt}
]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)
We now take the continuum limit
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)
Then
- 2~ X; (.E, J(t,)ót)'
ót
= --
1 ~
~ [J(tj+i)] 2 ót = JT dtI 2 (t). (10.148)
2m
j=O O
Then from Eqs. (10.143) and (10.145-10.148)
.
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
J dqjeí 2ót {(qH1-qi+f;fI(tH1))2+(qi-qi-l +f;f I(tj))2}.
We can write the coefficient of ~ ;;t as
Thus using the Gaussian integration (10.128), the dqi integration gives
Therefore after q1 ..... , QN-l integrations
Taking the continuum limit ¿f=-;

1 I (tj) ót = J; I (t) dt, as in free particle
case (Nót = T), Eq. (10.150) becomes
(10.151)
Thus finally from Eqs. (10.7.4) and (10.151)
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)]
I(t) = [T J(r)dr (10.153)
which can be written as
I(t) = foT () (r - t) J(r)dr (10.154)
where
(} (r - t) = 1 if T > tl )
= Ü if T < t.
Thus
I(O) = fot () (r) J (r) dt: (10.155)
= 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).
For this purpose, we define the functional derivative
__!_L = lim f [g(x) + E(x - y)] - f [g(x)] _

(10.156)
óg(x) e-+O E
If
f [g(x)] = 1b dx [g(x)f,
then
ó:rx) = 1b dxn [g(x)f-1 ó (x - y) = n[g(x)in- 1.
Then from Eq. (10.154)
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)]
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
= [q¡ + QF; q¡ (ti - t)] Z(J = O)
= (qF, tFltz(t1)lq1, t¡)J=O·

where
Z(J =O)= (~) ! e-k m(q1;.;q¡)2 (10.161b)
21riñT
óJf: rg¡t
first for t1 > t2 and then for t1 <h
u
Next we calculate (-lñ)2 1 2[,
For ti > t2, using Eq. (10.160) t1 ----+ t2]
T
j \
-
2~
j dtB(t2 - t)B(t1 - t)]
o
where we have used (10.157), (10.159) and (10.160). Thus
where we have used that [note T = tF - t1]
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
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
[q(t1), (j(t2)] !~ sinw(t2 - t1).Y/O,) if t2 :#: t1.

tf
=
[q(t2),p(t1)] = ificosw(t2 - ti'{¡/m. t2 / t¡

Thus canonical quantum conditions hold only for equal time t2 =
t1 = t
10.2 Show that for the angular momentum L = x x p, L; = Ei!nXlPn,

the classical PB is
(Li,Lj)PB = fi!nfjrs L (ózkPnÓskXr - ÓnkX!ÓrkPs)
k
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(>..) =
to get the required quantity f (>..)

10.4 Let x and Px be the coordinate and linear momentum in one di-
mension. Evaluate the classical PB
(x, F(px))PB
Let x and Px be corresponding quantum mechanical operators.
Then using the rule to go from classical mechanics to quantum
mechanics, evaluate the commutator,y
[
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 [
memb~rAxl±) = ±i±)]. What is the corresponding eigenvaluE\'iThus

exp ( ia:x) gives the unitary operator corresponding to the trans-
lation.
10.5 Consider the Hamiltonian describing a free particle in an external

field E ( constant in time)
·2
p 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}
where (.A(t)) = (1Pol.A(t)l1Po) in the Heisenberg picture. Then prove

the following:
(i)
(x(O))
(x(O)p(O)
= O, (p(O))
+ p(O)x(O))
= ñk,
= o.
(p(0)2) = 2m(E) l )
(ii)
(~x), = { (~x)5 +~((E) - r,,:~ 2)} !

L )
(~P)t = { 2m ((E) - ñ:~ 2)} !
10.6 For the Hamiltonian

= (~P)o L
!
H = 2Pª2V(ª)q,
o
show that the transition amplitude is given by
where
is the Lagrangian.
HEISENBERG EQUATION OF MOTION, INVARIANGE PRINGIPLE AND PATH INTEGRAL235
10. 7 If T < t1, t2 < O, S = Jt Ldt, and

j Dq q(t1)q(t2)e"8 = (qrlT(q(t1)<i(t2))lqo)
where QT = q(T), q0 = q(O), show that for the Harmonic oscillator

(qrlT( q(t1)<i(t2)) lqo)
={ i~ [o(t1 - t2) sinw(T - t1)

mwsmwt
sin wt2 + O(t2 - t1) sin w(T - t2) sinwt1]
+ in. ; wT [qo sinw(T - t1) + QT sinwt1]

[ Qo sinw(T - t2) + qr sin wt2] } (qp lq1)
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
with the boundary conditions

q(t) = q¡, q(O) = Qi
show that
S = /Jq¡ t - f!__t3 + /Jqi t + m( q¡ - Qi)2.
el 2 24m 2 2t
Show that the propagator K(q¡, t; Qi, O) is
K(q¡,t;qi,O) = m
( 21rhit
)! exp (iSc1)
T .

I July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics l
Chapter 11
ANGULAR MOMENTUM AND

SPIN
11.1 The Zeeman Effect
Considera hydrogen atom in a constant weak magnetic field B. Then the

Hamiltonian is given by
HB = H + interaction energy between the atom and
the magnetic field,
where H is the Hamiltonian of an unperturbed atom (i.e. in the absence of
a magnetic field)
p2
H=-+V(r). (11.1)
2me
Let us first calculate the interaction energy between the atom and the
magnetic field classically. We may regard the proton as fixed and the elec-
tron moving round it to be in a circular orbit. Let the radius of the orbit
be r, and the momentum of the electron be p. The velocity of the electron
in the orbit is given by
p
Ve= -. (11.2)
me
The electron orbit can then be regarded as a current loop. Now a current
loop which endoses a surface S acts like a shell of strength j/c where j is
the current flowing in the loop. Now any element of the shell of area dS
acts like a dipole of moment j/c dS. Therefore a circular current loop can
be regarded as a dipole of moment
(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
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)
ANGULAR MOMENTUM AND SPIN 239
where the eigenvalues of H, L 2 , Lz are given respectively by En = m4

21i2;';_2 ,
1(1 + l)ñ2 and mñ. For a given l,m has (21 + 1) values
-1, -1 + 1, , o, 1, 2, ... , 1 - 1, 1
n = 1, 2, , and n 2: 1 + l.
Thus
But
HBUnlm(r, e, <P) = E;;unim(r, (), <P) L (11.12)
HBUnim(r, B, <P) = (H - _e_BLz)UnLm(r, B, </))d

2mec V\
= ( En - mñ )un1m(r, B, </)). (11.13)

2~c
Thus
EnB _ en
= En m = En - --Bm 11.14 ( )
' 2mec
where En are the unperturbed hydrogen atom energy levels. Equation
(11.14) shows that for given values of n, l, the (21 + 1) different states
(corresponding to (2l + 1) values of m) which previously corresponded to
the same level, En, are now split into (2l + 1) separate levels with spacing
according to Eq. (11.14)
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.
11.2 Angular Momentum
As we have discussed previously, orbital angular momentum is defined as
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)
where >..lt2 and mñ denote the eigenvalues of J2 and Jz,

We now introduce new operators
J± = Jx ±iJy, (11.20)
Then it is easy to see using the commutation relations (11.18b) that
J+J_ = J; + J; - i(JxJy - JyJx)
= J2 - f; + fiJz
= J2 + ~ñ,2 - (Jz - ~ñ,)2. (11.2la)
Similarly
(11.2lb)
Thus
[J+, L] = 2fiJz. (11.21c)
Also
[Jz, J=f]
(J2,h]
=
=
=r:-ñJ=f
o.
L (11.2ld)
(11.2le)
)
Then using Eqs. (ll.21a) and (11.2lb)

JzLI.Xm) = (-ñJ_ + J_Jz)I.Xm)
= (-ñJ_ +mñJ_)I.Xm)
= (m - l)ñJ_ I.Xm) L (ll.22a) _)
J2J_¡.Xm) = J_J2!.Xm)
= .xn? J_¡.xm). (ll.22b)
Thus if J_¡.Xm)-/=-O then J_I.Xm) is an eigenstate of Jz belonging to eigen-
value (m - l)h but to the same eigenvalue .Xñ2 of J2. Again using Eqs.
(11.21) and (ll.19b)
JzJ+I.Xm) = (hJ+ + J+Jz)I.Xm)
= (m + l)ñJ+I.Xm) (11.23a)
J2J+I.Xm) = J+J2I.Xm)
= .xñ2J+I.Xm) (ll.23b)
Thus if J+!.Xm) -/=- O then J+I.Xm) is an eigenvector of J2 with eigenvalue
(m + l)h but same eigenvalue .Xñ,2 of J2. Continuing in this way we see
that the eigenvalues of J2 corresponding to eigenvalue .Xñ2 of J2 are
· · · (m - 2)ñ, (m - l)ñ, mñ, (m + l)ñ, (m + 2)ñ · · · , (11.24)
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
IF) = Ll>.m), then

(FI = (>.mlJ+·
Thus
(FIF) = (>.mlJ+Ll>.m) 2:: O. (11.25)

The equality sign holds when J_ ¡>.m) = O. Using Eq. (11.21a), we have
1 1
(>.mlJ2 + 4ñ2 - (Jz -
2ñ)21>.m)
2:: O
i.e. on using Eq. (11.19) we get
1 1
[(>. + - (m - 2:: O.
4)ñ2 2)2ñ2](>.ml>.m)
But
therefore
1 1 2
(>.+¡) 2:: (m-2). (11.26)
Thus ( +>. i) 2:: O and we have the inequality

-J>. + 1/4 + 1/2::; m::; J>. + 1/4 + 1/2. (11.27)
Thus m is bounded on both sides. Thus the series (11.24) must terminate
on both sides. Therefore m has a minimum and maximum value which we
respectively denote by m2 and m1. Then we must have
(11.28)
since otherwise Eq. (11.22) shows that J_ l>.m2) is an eigenvector of Jz with
eigenvalue (m2-l)ñ which contradicts that m2ñ is the minimum eigenvalue.
Thus Eq. (11.28) holds and because of it we must have an equality sign in
Eq. (11.26) for m = m2, i.e.
(.X+ 1/4) - (m2 - 1/2)2 = O
or
m2=1/2±J-X+l/4. (11.29)
Similarly we can show that

(11.30)
because otherwise, from Eq. (11.23), J+l>.m1) would be an eigenvector of
Jz belonging to eigenvalue (m1 + l)ñ which contradicts that m1ñ is the
maximum eigenvalue. Thus Eq. (11.30) holds and we have
(>.m1IJ_J+l>.m1) = O
or using Eq. (11.21b)
2 1 2
(>.m1IJ + -
4ñ CJz + 21 ñ) 2 i>.m1) = O
i.e. on using Eq. (11.19)
( >.ñ.2 + ~ñ.2 - ñ2(m1 + ~)2) (>.m1l>.m1)= O

or
1 1
(>. + - ) - (m1 + - ) 2 = O.
4 2
Therefore
1
m1 = -2 ± J>. + 1/4. (11.31)
To avoid min>max we must have a negative sign in Eq. (11.29) and a
positive sign in Eq. (11.31). Thus
m2 = (m)min = ~ - ·J>. + 1/4 (11.32a)

1
m1 = (m)max = -2 + J>. + 1/4 = -m2 (11.32b)
and the series (11.24) for the eigenvalues of Jz becomes
m2ñ, (m2 + l)ñ, · · · , (m1 - l)ñ, m1ñ. (11.33)
Now because the successive terms of the series differ by one unit of ñ,
(m1 - m2) must be an integer or zero, i.e.
m1 - m2 = 2J>. + 1/4 -1 = 2m1 (11.34a)
or
2J>.+1/4-1=2j (11.34b)
where 2j is an integer or zero. Thus j is restricted to integral or half integral
values, including zero. Since from Eq, (11.32b) m1 = -m2, it follows from
Eq. (11.34a) and (11.34b) that
m2 = -j, m1 =i (11.35)
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)ñ?;j = o, 1/2, 1, 3/2, · · · (11.37)

mñ;m = -j,-j + 1, · · · ,j-1,j. (11.38)
The eigenvalue Eq. (11.19) now becomes
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.
11.3 Matrix Representation of Angular Momentum
We take eigenstates jjm) to be normalized. Then the matrix elements of

J2 and J¿ in the basis defined by the vectors jjm) can be written as
(jmjJ2jj'm') = j'(j' + l)ñ2(jmjj'm')
= j(j + l)n2Ójj'Ómm' L (11.40a) )
(jm!Jzl/m') = mMjj'Ómm'· (11.40b)
Let us now find the matrix representation of J±. Since
therefore
O= (jmjJ2hl/m') - (jmlhJ2jj'm').
Using Eq. (11.39a) we have
O= j(j + l)ñ2(jmlhl/m') - j'(j' + l)ñ2(jmihl/m').

Hence
(jmlJ±lj'm') = O unless j = j'

j July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics
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
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) )
= nJ(j + m)(j - m + l)óm',m-1· (11.50)

From Eq. (11.43), in view of the fact that J± lim) is an eigenstate of J¿
with eigenvalue (m ± l)fi, it is clear that
J± lim) = C(±)j li, m ± 1).
Hence from Eq. (11.47)
J+lim> = nJ(j + m + 1)(j - m)li, m + 1> L (11.51a) ~--- ... ). o'·
J_ljm) = ñ/(j + m)(j - m + l)lj,m - l)~L (ll.5lb) \::.)
J+ and J_ are called raising and lowering operators. Rewriting Eqs. (11.40)
we have
(jmlJ2ljm') = ñ2j(j + l)ómm' L (11.52) }
(jmlJzljm') = nmómm'· (11.53)
11.4 Spin
(i) j = O; all above matrices are null.

(ii) spin (1/2)fi, j = 1/2 : The vector space is two dimensional with basis
vect ors 1 2,1 21) , 112, 2-1)·.
}
(11.54a)
J
Therefore J_ can be represented by the matrix
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 .
Also from Eq. (11.52), J; is represented by
~ 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
Ux = (~ ~), CTy = (~ ~i), O'z = G ~l) · (11.56)
These are known as Pauli matrices. We note that

)
ª2X = ª2y = ª2 = 1 L
Z
UxUy + UyUx = Ü ~ (11.57a) l

CTyCTz + CTzUy
d
= Ü ~
O'zUx + UxO'z = Ü [,_

248 Quantum Mechanícs
Writing
we can rewrite Eqs. (11.55) and (11.57) as

[O"i, O"ji = 2iE:ijkO"k L
[O"i,O"jl+ = (O"iO"j +O"jO"i) = 2óij· (11.57b)
Since for j = 1/2, the eigenvalues of Jz are ±(1/2)ñ, it follows from Eq.
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
~O"zl ± 1/2) = ±1/21 ± 1/2). (11.59)

In the basis defined by the eigenvectors 1 + 1/2), 1 - 1/2), it is very easy
to check by direct substitution of Eq. (11.56) into Eq. (11.59) that the
eigenveejo s of Jz = !nrrz are
¡(G~~t "' 1 + 1/2) =

i G) . 1 - 1/2) = (~) . (11.60)
By taking the commutation relations

J X J = inJ
for the angular momentum J, we have uncovered the possibility of
j = 1/2, m = ±1/2.
As far as the orbital angular momentum is concerned, the eigenvalues l and
m are restricted to integers. Thus j = 1/2, m = ±1/2 cannot be associated
with orbital motion, it must be the intrinsic angular momentum or spin
of the particle itself. In particular, the electron has spin (1/2)ñ. In fact
many elementary particles (e.g. electrón ~oton,bt,ron, neutrino) have
spin (1/2)ñ. Particles with spin O, or ln or 3/2)ñ or even higher values also
exist in nature. We shall, however, confine ourselves here to j = 1/2 or spin
(1/2)ñ. In this case, in addition to a factor 'lj;(r) specifying its probability
distribution in space, the state function of a spin 1/2 particle has a factor
lx), specifying its spin state. In the basis defined by the eigenvectors 1 + 1/2)
and 1-1/2) of (1/2)ooz belonging to eigenvalues ±(1/2)ñ, we can represent
the general spin state lx) as
lx) = 1 + 1/2)(+1/2lx) + 1- 1/2)(-1/2lx)
=al+ 1/2) + bl - 1/2), (11.61)
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
1 + 1/2) = G). 1-112) = (~). (11.65)
we can write the state function ,¡/J(r)lx) in the form
i.e. it has two components.

(iii) j = l; the vector space is now 3 dimensional with basis vectors 11, 1),
JI, O), [L, -1). The operators are now 3 x 3 matrices.
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: = 1rnlí
v2
(º º)
1 O1 1
O1O L )
)
1 O O)
Jz = /i O O O .
( O O -1
11.5 Splitting of the Ground State of Hydrogen Atom
The total angular momentum of an electron in an atom is given by

1
J = L + 2ñn = L + S, (11.66)
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).
Thus it follows that
H(Tu10o(r)f ± 1/2) = (E1 ::¡:::

n
b2 )urnol ± 1/2) (11.69)
i.e. u1oo(r)I ± 1/2) are eigenstates of H(T with eigenvalues
Ef = E1 ::¡::: b~. (ll.70)
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
11.6 Addition of Spin
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)
J We first show that11 these states are eigenstates of S z:

Sz\X++(l, 2)) = (S1z
=
+ S2z)\X++(l, 2))
(S1z + S2z)\x+(l))\x+(2))
1 1
=(2n+2ñ)lx +(1,2))
= ñ\x++(l, 2)) L.. (11.75a)
Sz\X+-(1, 2)) = O L (11.75b) )
Szlx-+(1, 2)) = O L (11.75c) /

Sz\X--(1, 2)) = -ñ\x--(1, 2)). (11.75d)
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) )
lx:¡:1(1, 2)) = lx--(1, 2)) = 1 H) \_ (11.78c) )

o
Jx_ (1, 2)) =
1
v'2 (lx+-(1, 2)) - lx-+(1, 2))) =
I i 1) v'2
- 11 i)
(11.78d)
Hence we have the result that 82 has eigenvalues s(s + l)h2, where s takes
the values
s = 1,0.
Sometimes we will write the spin states simply as L
xf(l,2) = x++(l,2) = x+(l)x+(2) (11.79a) )
xic1, 2) = ~Cx+(l)x-(2) + x-(lh+C2)) l, (11.79b) )
x:;1(1,2) = x--Cl, 2) = x-(1)x-(2) L- (11.79c) )
~(1, 2) = ~Cx+(l)x-(2) - x-(l)x+C2)). (11.79d)
We conclude that the addition of two spin 1/2 angular momenta give the
angular momenta with j = O and l.
11.7 Addition of Angular Momenta
The normalized eigenstates lj, m) of J2 and J form basis vectors 2
lj,j), lj, j - 1) · · · lj, -j)

in a space of (2j + 1) dimensions. In the basis defined above, the angular
momentum operators J2 and J2 are represented by (2j + 1) x (2j + 1)
diagonal matrices for all j =O,!, 1, !, ...
j(j + 1)
j(j + 1)
J2 = ñ?
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
e-ooiOj e-ioºoc:j-1) : ::_: ~ )

9j(()) = e-j¡Oh =
(
eiOj
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)
Thus the eigenvalues of Jz are

(11.83)
Therefore the basis liihm1 m2) of the space R1 x R2 is a system of or-
thonormal eigenvectors of Jz with eigenvalues m = m1 + m2. Since
m1 =ji,··· , -j-L (11.84a) )
m2 = i2, · · · , =i». (11.84b)
therefore m have (2j1 + 1)(2i2 + 1) eigenvalues which run from
(11.85)
What are the possible values of j? Now (m)max = i, + j2. The máximum
possible value of j must also be (j1 + h) since otherwise, there must be a
corresponding m = j > [: + j2 = (m)max· Thus
j ~ i, + i2-
Now the possible j values differ by an integer and the total number of
eigenvalues of Jz are (2ji + 1)(2j2 + 1) for given i, and j2. But for each i,
m has (2j + 1) values, therefore
jmax
¿(2j + 1) = (2j1 + 1)(2h + 1).

Jmin
This is an arithmetic progression with common difference 2 and number of

terms Umax - imin + 1) so that its sum is
Umax - imin + 1)((2jmin + 1) + Umax - imin))
= (2j1 + 1)(2j2 + 1).
This gives, on using that imax = i, + Í2, i!in = (j1 - i2)2, i.e.
imin = li1 - J2I-
Thus, the possible j values must obey the inequalities
(11.86)
As an example consider
J =L+S,
so that i, = l and i2 = !. Thus, we have
l l - !¡ < J < l + !2'

2 - -
0
i.e. j only has two possible values

1
Z=r:2, l#O
!
and for l = O, j = only. The following table summarizes the situation for
sorne low values of l:
l Possible value (2j1 + 1)(2j2 + 1)
of j m = 2(2l + 1)
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
Basis vectors of the product space have been labeled as

l}1m1)l}2m2)= l}d2m1m2),
corresponding to the commuting set J¡, J'.f., J1z, hz. A second way of
labeling a vector of product space is to write it as
l}mji}2)
which corresponds to the commuting set J2, Jz, J¡, J'.f. (note that J2 <loes
not commute with J1z and hz). We have
}
Jzl}mji}2) = mñljm}1J2)
J2ljmj1}2) = j(j + l)ñ2ljmji}2).
Now, the states ljmji}2) or l}i}2m1m2) form complete sets so that we
can write
l}mji}2) = ¿ l}i}2m1m2)(ji}2m1m2l}mji}2) L (11.87a) )
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
(jmj1}2 lj' m' ji}2)

= ¿ (jmj1}2i}i}2m1m2)(jd2m1m2l/m'j1}2)
(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.e.
J and zero otherwise.

The elebsch\,\Gordan coefficients can be calculated, but the calculation
~
is quite cumbersome and it is easier to find the e.e. coe:fficientsfrom books
of tables.
Here we give two tables for J2 = 1/2 and }2 = l. For j2 1/2,
m2 = ±1/2, the e.e. coefficients are given by the following Te.o 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
For J2 = 1, m2 = +1, O, -1, the e.e. coefficients (J1lm1m2IJmJ1l) are
Table II
j m2 = 1 m2 =0 m2 = -1
.!. l
J =Ji+ 1 [(.ti+ml(j1+m+J}] 'I [(j,+l-mH,;1+m+llJ "2 [(j1-m)(j1-m+l)] ~

(2j,+1)(2j,+2) (2j,+l}{j¡+l} (2j1 +I)(2j, +2)
1 1
J = Ji _ [ (ji +m}(it-m+l) J 2 m [ (j, -m)(ji +m+l)] 2
2j¡(j1+l) [j, (j1 +1)]'/2 2j1 (j, +1)
1 1
J = J1 -1 [(i1-'l'n)(j,-m+ll] 'i _ [<j1-ml(j,+ml]~ [(j1+m+I)(j,+m)J 2
2j¡ (2j¡ +1) j¡(2i1+l) 2j¡(2j1+l)
As an application, consider the case

J =L+S,
where L is the orbital angular momentum and S represents the spin !
operator. As already seen J can only take two values viz. l + and l - ! !.
Thus
J2IJ = l + 1 m) = (l + 1
2,
3 1
2)(l + 2)ñ-21J = l + 2, m) (11.90a))
L
1 1 1 1
J 2 IJ=l-2, m) =(Z-2)(l+2)ñ 2 jj=Z-2, ml-(11.90b))
JzlJm) = mhjJm), (11.90c)
where j = l + !, l - !, m =me±!· Using the C.G. coefficients (Table I)

we can write
1
ll + -, m)
2
= l+m+!¡l.:m_.:.:)+
2l + 1 2 22
l- m +
2l + 1
! l -m+-
1
1
2
1
-
2
-1)
2
L }
(11.91)
ll-2 m)=- l-m+!¡z.:m_.:.:)+ l + m + ! ll.:m + .: - .: ).

2' 2l + 1 2 2 2 2l + 1 2 2 2
(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)
11.8 Rotations: Rotation Matrices
As we have shown in Sec. 10.5, rotations in space are most conveniently

described in terms of angular momentum operators J. A rotation of co-
ordinates through angle w about the direction w in the positive sense is
described by the operator
Since J2 commutes with

u; = e-'i!-w·J
J«, Jy and J it follows that

2,
L (11.101) ó
[Rw,J2] = Ü (11.102)
i.e. a rotation does not change the total angular momentum of the system.
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,
In this case the desired rotation is given by

Ro = e-j¡On·J L (11.105)
Let us take Po along the z-axis (Po unit vector along z-axis).
Now () and cp are polar and azimuthal angles of p in the system in which
Po is along the z-axi5l-j,
L-- I )
J p = (sin cos cp, sin é' sin cp, cos ())
é
Po x p = - i sin () sin cp + j sin () cos cp V

(11.106a)
)
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
n · J = -Jx sin ó-} Jy cos <fa. (11.107)
Let us look at
(11.108)
and
Generalizing these results, we see that
(11.109)
Hence we have the important result
Re= R(</>, B, -</>) = e-iln·J

= e-if:J.e-1fl:Jye1J. (11.110)
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)
We have seen that rotation operator R¿ = e--kw·J commutes with J2

and Rwljm) is also an eigenstate of J2 with eigenvalue j(j + l)n2. Thus,
we can write
m'
j
= L ljm')d~)m(w) (11.112)
m'=-j
where
(11.113)
d(í)(w) is a (2j + 1) by (2j + 1) matrix whose matrix elements are d~!m(w);

d(í)(w) is called a rotation matrix.
•I
The rotations form a group, i.e. they satisfy the followingfour properties
i) Identity rotation operator exists
Rw=O = l.
ii) Inverse of Rw exists
)
u; = e-i¡¡!,·J L
ti.; = ei¡¡!,·J L J
RwR-w = l.
iii) Two successive rotations made in a row say w1 followed by w2 are

equivalent to a single rotation
Rw2 Rw1 = e --·ñ e --w1·n.

iw2 J i J
Rw =
-=/=- Rw1 Rw2, in general,

since w2 · J and w1 · J do not commute in general.
iv) The product operation is associative
(Rw1 Rw2)Rw3 = Rw1 (Rw2Rw3)·

We now show that the set of matrices dí(w) gives a representation of
the rotation group. Now
(jm'IRwljm) = (jm'IRw2Rw1 ljm) = ¿(jm'IRw2ljm")(jm"IRw1 ljm)
m/'
(11.114)
m"
264 Quantum Mechunics
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
[d(íl(w)t] m'm = (jm'le(i/li)w·J¡jm)

= (jmje-(i/li)w·J jjm') *
= iilmm' (w)* = iilm'm (-w)
d(j)t (w) = d(il(-w)

or
(11.118)
For the infinitesimal rotation w = e, di (w) has a simple r.epresentation.
Writing E± = Ex ± iEy, we can write E· J in terms of raising and lowering
operators J±, i.e.
1
E· J = + cJ+) + Ejz,
2(E+L (11.119)
so that d(J) ( t:) has a simple representation
d~?m(E) = (jm'l(l - ii · J)jjm) + 0(E 2)
= 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
d~?m(t:) = Ómm' - i~+ J(j +m)(.j-m+ l)óm',m-1
- i~ j(j - m)(j + m + 1)8m',m+l

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)
d~!m(8) = (jm'le-i{Jy ljm). (11.123)

Thus to construct d(j), we need only the matrix elements for a rotation
about the y-axis by an angle 8; cp dependence is trivial. We will write
d~!m(cp,8,-cp) = d~!m(cp,8).
The matrix elements d~!m(8) turn out to be real and have the simple
property
ij) (8) = (-1)2j-m-m' d(j) (8) (11.124)
-m',-m m'm ·
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)
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)
An applícation of this result is given in problem (11.18).
Rotation Matrices far the Combination of Two Angular Momenta

Consider J = J1 +J2, then we show that using the C.G. coefficients, we
can express the rotation matrix corresponding to J in terms of products of
those for J 1 and J2.
Now
(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 )
Hence
(J. 1 J. 2 m'1 m'2 le-(i/ñ)w·J11· 1 J. 2 m 1 m 2 ) = d(ji) m;m2 (w).

m~ m1 (w)i12l (11.134)
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'
Note that for w = 8n and j an integer [c.f. Eq. (11.130)]
(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
J ~*.,,, ( (}, <p )Yi1m1 ( 8, <p )Yi2m2 ( (}, <P )dn
[ ( 2·J1+ 1)(2"J2+ 1)]112 (.. ¡· .. )(·o··¡· ·oo)

= 4-rr(2j + l) J1J2m1 m2 JmJ1]2 J J1]2 Jl]2 .
(11.142)
11.9 Vector and Tensor Operator

s
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
Consider an operator A. Then under rotation, A will transform to a new

operator
To see this we note that

(1/!IAl1/!)
= (1/!IR;:;-1RwAR;:;-1Rwl1/J)
= (1/Jw!RwAR,::;-111/Jw)
therefore (11.143)
Now we study the transformation properties of operators under rotation.

l. Scalar operator: A scalar operator is invariant under rotation, i.e.
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
i.e. Wij = EijkWk, we have

V/= V;+ EijkVjWk =[V+ V x w]i.
We see that using Eq. (11.144) we get the correct transformation law
for the vector operator.
Let us go to the "spherical" or "canonical" base in which instead of Jx
and Jy, we consider J+ = Jx + iJy, J_ = Jx - iJy. Then we have
Pz, h] = ±nhL /
[J+, L] = mr;
These commutation relations can be written compactly as
[Jz, Jq] = ñqJq, q = O, ±1

[h, Jq] = ñy'l(l + 1) - q(q ± l)Jq±l, q= o, ±1
Jo=Jz t-. )
J1 = - ./2.(J,., + iJy) L )
1 .
J_1 = ,/2.(Jx - iJy)·
Thus in a "canonical" or a "spherical" base we will write the component of

a vector as
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
[Jz, Vq] = ñqVq L- (11.148) )

[ h, Vq] = li,J~i(_l_+_l_) q_(q-±-l)
Vq±l ·
Now we generalize; an irreducible tensor operator TJk) has q = -k, · · · , k,

so it has 2k + 1 components. A vector is a tensor for which k = 1 and
thus it has three components Vq = TJ1). We will always be working in
the representation in which Jz is diagonal. An irreducible tensor operator
of order k will satisfy the following commutation relations with angular
momentum
[Jz, Tt)J = ñqT?) L (11.149)

[h, TJkl] = ñy'k(k + 1) - q(q ± l)Td~i ·
Noting that
(jm'[J'f[jm) = /iJj(j + 1) - m(m -F 1)8m',m'fl,

we can write
(kq'[h[kq) =
(kq'[Jz[kq) = ñq(kq'[kq)
!iy'k(k + 1) -
=
q(q ± l)óq ,q±l
ñq8qq'·
1 L )
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
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.
11.lOa Wigner •ckart Theorem !Tl\

First we note that rJk\:~jim1) is an eigenstate of Jz with eigenvalues (m1
q)n. Note that o: is an eigenvalue of another observable which commutes
+V
with J2 and Jz. To see this, we consider
JzTJkl lo:j1m1) = (nqTJkl + TJkl Jz)lo:j1m1)
= n(q+m1)TJkllo:j1m1). (11.154)
Let us consider how the state rJkl¡o:j1m1) transforms under rotation
[using Eq.(11.112)]
RwTJkllo:j1m1) = RwTJklR-;:/Rwlo:j1m1)
T(k) ik)
= "\;""' (w) "\;""' lo:J· 1 m' )iiil (w)
L q' q'q L 1 m~m1
q' m'1
= "\;""' "\;""' T(k) lo:J· 1 m' )ik) (w)iii) (w) (11.155)

L L q' 1 q' q m~m1 ·
q' m~
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
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
J+ lajm) ~ ( J+ ~ lfijm)) (fijmlajm)
= nJ(j + m + l)(j-m) ¿ lf3jm + l)(f3jmlajm).

{3
(11.162b)
Comparing Eqs. (11.162a) and (11.162b), we have

([Jjmjajm) = ([Jjm + ljajm + 1),
i.e. it must be independent of m. Hence we have the result that
(a'j'm'laj'm') is independent of m'.
This is a remarkable result, it shows that the matrix elements of a tensor
operator Eq. (11.161) between two angular momentum states factorize
into a C.G. coefficient containing all the angular momentum details but
independent of the dynamics and a coefficient independent of m, m' and q
but containing all the dynamics.
Thus it is possible to write Eq. (11.161) as
( ci j' 1 IT(k) l laj)
(a'j'm'ITJk)lajm) = J2}'TI (j'm'kjlkjqm), (11.163)
where it is customary to write

, -r 'I
-r ') (a'fllT(k)jjaj)
( a J m aJ m = ..,/2j' + .
1
The quantity (a'j'jjT(kljjaj) i known as the reduced matrix element J_
and the result (11.163) is the Wigner~ ckart Theorem. Because of the
reality of C.G. coefficient
N
(j'm'kjjkjqm)* = (kjqmjj'm'kj) = (j'm'kjlkjqm)
( a ,., 'IT(k)I . )
J m q cqm. =
(a'j'IIT(k)llaj)(kº
..,/2j' +
1·' 'kº)
[qm. J m J . (11.164)
1
J_
11.lOb Application of Wigne~Eckart Theorem N
As a simple application of Wigneij\Eckart theorem, we derive the angular J__
momentum selection rules for the el ctric dipole (El) transition. fJ
First we note that ( for j integer)
(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
= (2k + l)(2j + l) (k · mi., 'k ')( ·1ok ·¡k ·oo) (11.165)

41r(2j' + 1) Jq J m J J J J '
on using Eq. (11.142). By the Wigneil1EckartTheorem,
(J"'m'IYikq 1 J·m) = (J·'IIY(k)II J") (kjqmlj'm'kj).

J2j' + 1 (11.166)
Hence we have the result
(j'IIY(kllli) = (2k + 1];2j + l) (j'OkjlkjOO). (11.167)
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.
Now using Eq, (11.165)
(l' m' JY1q ( ()q:¡) l lm)
= lf ~:-:+-+-~
, (llqm¡z'm'll)(l'OlllllOO). (11.169)
The C.G. coeffi.cient
(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.
The parity selection rule forbids !:::,.l = O transitions.

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 )
(l - lm - 1/Yi-1 /lm) = a (l+ m)(l + m - 1)

2(2l + 1)(2l - 1) .
Thus we know completely the angular momentum part of the matrix
elements for the El transition.
11.10 Problems
11.1 Consider an operator ú, which obeys the commutation relations
[u, Jz] = ~líil L )

[u,J+J = o.
Show that
u/j,j) = constant /j -1/2,j -1/2).
11.2 Show that
(i) CTxCTyCTz = Í.
(ii) (O'· a)(O' · b) = (a· b)I +iO' ·(ax b), where a and b are ordinary
vectors and I is a 2 x 2 unit matrix.
(iii) (O'· n) has eigenvalues ±1, where n is a unit vector.
11.3 The eigenvectors of CTz corresponding to eigenvalues ±1 are denoted
by / ± 1/2). Show that
(i) F2(/ + 1/2) ± / - 1/2)) are normalized eigenvectors of CTx with
eigenvalues ±l.
(ii) 1(1 + 1/2) ±il -1/2)) are normalized eigenvectors of ay with

eigenvalues ±l.
Write the above eigenvectors as column vectors.
11.4 Sx, Sy and S, are components of spin 1/2 operator S and abey the
commutation relations
[Sx, Sy] = inSz, etc.
Show that for S± = Sx ± iSy
S+I + 1/2) = O
S-l -1/2) = O
s_¡ + 1/2) = ni -1/2)
S+l -1/2) =ni+ 1/2),
where 1±1/2) are eigenstates of Sz belonging to eigenvalues ±1/2 Ii.
11.5 If a particle is in an eigenstate of ax, find the probability of finding
it in the eigenstate of a z belonging to eigenvalue + l.
11.6 Consideran electron in a uniform magnetic field B in the positive z-
direction. The result of a measurement has shown that the electron
spin is along the positive x-direction at t = O. An arbitrary spin
state at time t can be written as
lx(t)) = a(t)I + 1/2) + b(t)I - 1/2),
where lx(t)) satisfies the Schrodinger equation
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) m' = m+ 1/2L

(ii) (j'(j' + 1) - j(j + 1))2 = !(j'(j' + 1) + j(j + 1)) + 136.
11.8 Consider a system of two distinguishable particles, each with spin
1/2 and magnetic moments µ1 = µ10-1, µ2 = µ20-2 in an externa!
magnetic field B in the z-direction. The spiq./spin interaction of the
particles is bo , · o-2, where bis a constant. Find the exact energy
eigenvalues of this system.
11.9 In the normal Zeeman effect, an energy level characterized by quan-
tum number n is split into (2l + 1) different levels:
eñ
Enm = En - --Bm,
' 2meC
where m = -l, · · · , l. The equal spacing between the levels is given
by
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
¡¡ Find the dipoU}lc!ipole magnetic interaction energy of an electron

and a positron at a fixed distance a, in eigenstates of total spin.
11.11 Show explicitly that the state [j = l - 1/2, m) as given by Eq.
(11.92) is an eigenstate of the operator (L + S)2 with eigenvalue
j(j + l)ñ,2.
11.12 Consider an electron in an external magnetic field B. The energy

of the spinning electron is given by
eg
Hspin = -µ · B = --S · B,
2mc
show that
dS =-~SxB.
dt 2mc
Further show that the expectation value of the spin operator S at
times t, when initially the particle is in the spin state with Sx = !tiL )
is given by
ñ ñ. .
(S) = coeune¿ + smwtey,
2 2
where ex and ey are the unit vectors along the x- and y-axis re-
spectively and w = ;:;! .
11.13 The Hamiltonian for a particle of mass mis given by
A2
H + V(r) + UñL -a ;
= }!___
2m
where U is sorne constant. Show that L is not a constant of motion,
!
but J = L + ñu is a constant of motion.
11.14 Show that for the spin !n case (J = S = !ñu), the rotation oper-
ator can be written as
UR(B) = e-(i/2)1:1n·uti\ ;-
=
() ..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
Compare it with the transformation law for a vector under the

above rotation. For the same rotation, show that
u'= e-i(9/2)n·uui((l/2)n·u
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 >
Hence, show that for the states

IP i) = e-(i/2)9n·u¡ + !)
2
= cos ~ 1
2
+ !) + icf> sin ~
2 2
1 - !
2
) L )
IP 1) = e-(i/2)9n·u¡ _ !)
2
= cos -() 1 - -1 ) - e -icf, sm
. -() 1 + -1 )
2 2 2 2 '
we have
(T . PIP i) = IP i) V )
(T · PIP 1) = -lp 1).
<T · p is called the Helicity Operator.
11.16 Show that the rotation matrix for the operator R¿ = e-(i/ñ)w·J for
w = ()n and J = !M (spin operator), !
dw (</J, (), -</J) = ( .cos_ ! -e-icf> sin!) .
e•cf> sm !!.2 cos !!.2
11.17 (a) Starting from Eq. (11.138), show that
'°'
L iii)m~m1 m;m2 (w)(1· 1 1· 2 m 1 m 2 IJ'm1·1 1· 2 )
(w)ih)
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.
coefficients and replace at the end J", m" by i. m.]

(b) Use the above relation and f,:ble of C.G. coefficieni_;,o show
that
/ 5
,Vr;+;;i-,!1
2J m -2,m-2 1(w)dL(w)+J1-
22
2.mi-,!1
J m -2,m
+i(w)dÍ 1(w),
2 ,¡-,¡
= .~d;,.,m(w).
11.18 Show that

Pj(cosO') = ~:
2J 1
¿
m
~:i(,6,a)Yjm(O, </>),
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).
Take the r-representation and make use of d~b(a, ,6)

¡;f;;~:i(,6, a) (c.f. Eq. (11.130)).]
11.19 Consider the 4 state system consisting of two spin-! particles. The
vector space of the system is spanned by the 4 orthonormal states:
I rn = 1 ihl i)2 1 u: = 1 i)1l l)2 I H) = l lhl i)2 I H) = l lhl l)2
where the arrows refer to the direction of the spin along the z-axis
and the subscript 1 and 2 refer to the particle. Suppose that the
Hamiltonian of this system is given by
H=,(81,z+ 2.:)+*81 ·82 _y
a. Write the above Hamiltonian in terms®,±, 81,z, 82,±, 82,z· 'l
b. Using the form of the Harniltonian found in part (a) find the
matrix of H in the basis given above.
c. Write the Hamiltonian in terms of the 8total where Stotal
S1 + S2.
d. Find the energies and stationary states of the Hamiltonian.
r: e. If the syst m is in the state CT
!) at time t = O what is the 9
J probability of finding the : ystem in the ingl ·tat· at time t¡,
11.20 Considera system of three particles. Particle 1 has spin particle !,

!
2 has spin and particle 3 has spin l. This system has 12 states:
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:
11.23 Show that:

a)
ef.o:JyAxe "'o:Jy = Axcosa+Azsina L o
)
b)
[Jx, (J,, TJ'll] + [.J,, [J,, TJ'l]] + [J,. [J,. TJ'll] - k( k + 1 )ñ'TJ'l Le)
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.
12.2 Stationary Perturbation Theory
Suppose we want to find eigenvalues and eigenfunctions of the Hamiltonian

H for stationary states :
H'I/Js = Es'I/Js· (12.1)
Assume that H can be decomposed as
H=Ho+>.V, (12.2)
such that we can solve exactly the eigenvalue equation
Houn(x) = EnUn(x). (12.3)
>.V is called the perturbation. >. is sorne parameter whi h is small. We
assume that V <loes not depend explicitly on time~e ame is always as- ~
sumed for H0). Since the eigenfunctions un(x) forma complete set, we can
283
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
Making use of Eq. (12.3), we get

¿ Cns(Es - én)un(x) = ¿ AVCnsUn(x).
n n
We multiply both sides by u~(x) and integrate over x:
¿
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)
TIME INDEPENDENT PERTURBATION THEORY 285
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)
Now óm8(€8 - €m) = O, thus to order >., we have

>.Ei1loms + >.cgl(és - ém) = >.Vms·
When m = s, óms = 1, we get
(12.8)
When mi= s, Óms = O, we get
cC1J _
ms -
Vms i=
€s - €m, m s. (12.9)
Note that ci!l remains arbitrary. u1~rst order

Es=€s+>.Yss L (12.10)
"Ps = ¿)óms + >.cgl)um

m
(12.11)
m
We choose cW such that

J 1/J;(x)'I/J8(x)d3x = 1, to first order iol
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)
For an observable F we have matríx elements of the form

'lj;5*.F'lj;rd
Up to first order, we have

Ísr =
J 3
X. (12.13a)
(12.13b)
We wríte
J u*s Fu r: d3x = ¡(o)

sr · (12.13c)
Also note
Thus up to first order

V~8 = (mlV[s)* = (slV[m) = V..m L o
V, f(O)
Ísr = Ji~) +). L f (O)
V..m mr· +). L mr sm . (12.13d)
m;,!s és - ém m;,!r Er - ém
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
Equating terms of order >.2

cC2l(c
ms s + E m )+EC2l5
s ms +cCms1lECs1l = '°"'c<
L...,, ns1lv;mn· (12.15)
n
Thus for m = s [since d!l = O and ó88 = 1]

E(2)
s
= '°"' 0<1Jv.
L...,, ns sn
n
Using Eq. (12.9), we get
(12.16)
Now
Thus to second order in A

E s = E:s + /\'V.ss + A,2"""
L l"Vinl2 · (12.17)
o
ncf-s E:s - E:n
Similarly we can calculate c~l

from Eq. (12.15) and hence 'l/Js to second
order. From Eq. (12.15), for m-=/=- s, we have
)
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·
They forma complete set, so that we can write

2
1Ps(x) = ¿¿ Cns,jUn1(x). (12.19)

j=l n
Substituting Eq. (12.19) in the eigenvalue equation
(Ho + .X V)'l/Js = Es'l/Js (12.20)
we get
(12.21)
j n j n
Multiply both sides by u;,,,i and make use of orthonormality relations
J u;,,,Jx)un1(x)d3x = ÓmnÓij, (12.22)
then we get
Cms,i(Es - cm)= A~¿ Cns,j J u;,,,iVun1d3x. (12.23)
, _,- i
J n
We now make the expansions
/ Cms,i Umsdi + .xcms,i
(1) + A 2 cms,
(2) i + ... (12.24a)
Es =es+ .XEi1> + .X2 E12) + ... (12.24b)

and correspondingly we write
'l/Js = 'l/J~ + -X'l/JF) + .x2'l/JF) · · · . (12.24c)
Note that on the right-hand side in the first term ofEq. (12.24a), we cannot
simply have óms as in the non-degenerate case since the index i is not then
balanced. Therefore we have to introduce di which we determine later.
Substitute Eq. (12.24) in Eq. (12.23) and keep terms up to order .X:
é---- (ómsdi + .xc~L)(cs + .XEF) - cm)
= AL¿
J n
Ónsdj J u;,,,i Vun1d3x. '----/
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
~ AdiE11l8ms + .xc~L(cs - cm)

=A~
J
dj J u;,,,i Vu5jd3x. 7 (12.25)
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)
12.4 The Stark Effect
Consider an electron bound in an atom in electric field
E=-V<p=-~!~r (12.33)
where <p is the electrostatic potential thus

<p(r) =-E· r.
Origin is taken at the position of the nucleus. Perturbation potential is

then
e V = -e<p = eE · r.
Energy of the system to second order in e is given by
Es = és + e v.ss + e2""L..., IVsnl2

ncfs és - én
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)
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
Do = -er88 = -e J u;ru8d3r (12.40)
Di = -e2 L rnsrsn + rsnrnsE (12.41)

ncfs Es - En
=ii·E,
where
(12.42)
Do is called the permanent electric dipole moment of the system as it is

determined by the unperturbed state of the system and is entirely indepen-
dent of the applied field. D1 is called the induced dipole moment for the
states. The tensor ii is called the polarisability for the state s.
Theorem: If the unperturbed Hamiltonian H0 is invariant with respect

to the reflection of coordinates, i.e. H0(r) = Ho(-r), then Tss = O if the
non-degenerate eigenstate u8(r) of H0 has definite parity.
('
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
and the parity of Unlm(r) is (-1)1.

Due to Coulomb interaction, all energy levels are degenerate except the
ground state since only the ground state has definite parity + l. Thus to
first order, the Stark effect vanishes for the ground state of H atom and we
have to go to the second order. However the radially excited states are de-
generate, hence the first order Stark effect does not vanish for such states.
First order Stark Eff ect

Consider the first excited state which correspond to n = 2:
l=O m=O 2s U200

l=l m = -1,0,1 2p U21-1, U210, U211
Take the electric field along the z-axis so that

E = \E\ez, V = e\E\z (12.45a)
Yij = J u;,i Vu8,jd3r i, j = 1, ... , 4. (12.45b)
Instead of i, j we can use l, m; which specify the above matrix elements.

Thus we have to calculate the matrix elements
(2, l, m \V\2, r, m') = e\Ef J u;1mzuwm,d3r.

Now
(12.46a)
Therefore
[Lz, V]= O. (12.46b)
Henct = 2 is understood]
(l,m f[Lz, V]fl',m') = O (12.46c)

or
imñ> m'n)(lm\Vll'm') = o. (12.46d)
Thus
(lm\Vflm') =O unless m = m'. (12.46e)
Parity ofstates \lm) and ll'm') are (-1)1 and (-1)1'. Thus
P¡tm) = (-1)1\lm)L (12.47a) )
Ffl'm') = (-1{ fl'm'). (12.47b)
Now
FV(rl'l/J) = efE\F{z(r\'l/J)}
= -e\E\zF(r\'l/J) = -VF(r\'l/J). (12.48)
Therefore
Pv+VF=O. (12.49)
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
(lOIVIOO) = (OOIVl10)* = elEI J u;10(r)zu20o(r)d3r. (12.53)

But [here a is the Bohr radius] L
Z = r CDS() (12.54a) 7
_ 1 -r/2a. ()1
U210 - 4(21r)l/2a5/2 re CDS lA...- (12.54b) J
1 ( r) -r/2a
(12.54c)
u200 = 4(21r)l/2a3/2 2 - ~ e .
Therefore
(lOIVIOO) = elEI
16(2~)a4
21r 11r fo 00 r2 cos2 () ( 2 - ~) e-r/a sin()d()r2dr
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.
12.5 Non-Perturbative Method (Variational Principle)
rn[Eerturbative approach we have a series expansion in terms of a small

parameter. Inµ3on--perturbative approach we have no series expansion. A
well known example of this is the variational principle. This is based on
the observation that1
(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
(1/!IHl1/l) 2: ~(1/JIEn)(Enl1/!)Eo = (1/!11/l)Eol (12.58) O

Hence (~~fJf> 2: Ea, the equality sign holds when 11/l) is the ground state
11/lo). In other words the absolute mínimum of (~~fJf>
is the ground state
energy and corresponds to 11/l) = 11/la)·
In applying the variational principle, one has to pickup a state 11/l(a)),
which depends on a parameter a. Then choose the value of a which rnin-
imizes (~~c~:J'tift,the 11/l) for that value of a is the closest of all the
11/J(a)) to the ground state and as such the least value of has the (~~fJf)
most accurate approximation to the ground state energy and wave function.
1See G. Baym, Lectures on Quantum Mechanics, Benjamen, 1969, p. 242
We illustrate this approach for He atom. The Hamiltonian of two electron

atom is
f>i f>~ Ze2 Ze2
H=-+-----+---
e2
(12.59)
2m 2m "i r2 lr1 - r2I
where the last term is the electron Coulomb energy; r1 and r2 are the posi-
tion vectors of two electrons measured from the nucleus. It is not possible
to solve this problem exactly for energy eigenvalues and eigenfunctions.
The simplest approximation is to consider the electro~electron interac-
tion as perturbation
(12.60)
The ground state is
(12.61)
where xoo is spin singlet. This is required by the Pauli principie as the
spatial part of 7/Jo(l, 2) is symmetric under r1 - r2 and therefore the spin
wave function must be antisymmetric in spin coordinates to make 7/Jo(l, 2)
antisymmetric. Here
7/Jis(r) = 7/Jioo(r)
= Yoo(B, rp)R10(r)
= ~ (Z/ao)3f2e-Zr/ao, (12.62)
7r
for r = r1 or r = r2. Now the energy of each electron in the Coulomb
potential (-Ze2/r) is [see Hydrogen atom sec. 7.2, and 7.3]:
z2e2 2
( - -) = -Z (-13.605eV)
2a0
= -Z2 Ry (12.63)
so that the ground state energy to the zeroth arder
z2e2 _z2e2
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
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)
Thus there is a discrepancy of a about 0.31Ry(5.45%) between the per-

turbative result and the experimental value of the ground state energy of
He. One reason is that the wave function (12.62) ignores the fact that one
electron tends to screen the charge on the nucleus as seen by the other
electron. That is to say, when one electron looks at the nucleus it sees not
only positive charge of the nucleus but also sorne negative charge density
around the nucleus from the other electron. This suggests that one can
improve the accuracy by choosing a trial wave function of the form (12.62)
with an effective Z' value in 1P1s :
1Pls = ~(Z' /ao)3l2e-Z'r/ao (12.73)

vi
and treat Z' as a variational parameter. Then
Ea(Z') = (H)
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')
To find the minimum of E0(Z'),

= (Z'2 - 2ZZ' + ~Z') e2
8 ªº L (12. 77)
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
,.__,,.,.....
value of E0(Z') is the ground state energy (Z = 2)
Eaªr = -(Z - ~ )2 e2
16 a2
= -5.695Ry
:::::! -77.46eV (12.79)
which is great improvement on the perturbative value as the discrepancy is

now reduced to 0.115Ry("' 2%).
12.6 Problems
12.1 Considera harmonic oscillator of mass m on which acts a uniform

time-independent external force F. Call its ground state
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 •
12.3 If the perturbation potential is given by

eV = eF · r = eFrcos()
where F is a constant, show that for the hydrogen atom
(lOOIVlnlm) = O for m -=f. O
and (lOOIVlnlO) is non-zero only for l = 1 and in this case it is
given by
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
It is clear that the eigenstates of H0 are lnjm),

Holnjm) = En1lnjm)
Treating V as a perturbation, and using Eqs. (11.95,96), show that
the energy shift
!),.E;;¡ = (njmlVlnjm)
= _.::!!_(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.
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 .
Chapter 13
TIME DEPENDENT
PERTURBATION THEORY
13.1 Transition Probability
Let us discuss how the perturbation method works for the time dependent
Schródinger equation
in8'1jJ =H'ljJ (13.la)

at
where we assume that
H =Ha+ >.V(r; p, t) (13.lb)
i.e. the perturbation V may depend on time while the unperturbed Hamil-
tonian Ha <loes not explicitly depend on t.
The Schródinger equation involving Ha is
inacp = Hac/J (13.2)

at
which can be solved exactly. >.V is assumed to be small compared to Ha.
In general, solution of Eq. (13.2) is
cp(r, t) = L Cnun(r)e-icnt/ñ (13.3)
where un(r) are eigenfunctions of Ha belonging to eigenvalues én·
We now assume that solution 'ljJ of Eq. (13.1) can be written as
'ljJ(r, t) = ¿ Cn(t)un(r)e-i1ont/ñ (13.4)
n
where the Cn are in general dependent on time. Everything in Eq. (13.4)

is known except Cn(t). We substitute Eq. (13.4) in Eq. (13.la) with Has
l
given in Eq. (13.lb) and make use of
Haun(r) =
303
énUn(r)
.:
Then we obtain
n n
or
n n
We multiply both sides by u~(r)é"rnt/n¡.a.n ntegrate over r and make use

of
)
()
Then we get
(13.6a)
n
where
Vmn(t) = J u:r,(r)V(t)un(r)d3r
= (mlV(t)ln). (13.6b)
These are coupled equations.
Suppose that for t < to, V= O, so that our system is unperturbed and
is in an eigenstate of H0• Then
'lfJ(t < to) = </>(t < to)
Comparing Eq. (13. 7a) with (13.3), we get

= Us(r)e-ic.t/n
L (13.7a)
o
C; = 1, C¿ = O, n-/=- s for t < to, (13.7b)
Suppose that the perturbation is effective from t0 to t. Under the action of
the perturbation, the system may pass from its initial stationary state u8 to
another state. According to the general rule, the probability of a transition
from the original eigenstate s of H0 to another state nis given by
P(s, n) = ICn(t)l2. (13.8)
We now rewrite Eq. (13.6a) as
iñÓm(t) = .XCs(t)e*m.-és)t/nvms(t) +AL Cn(t)
xe'í'--,,),~:~mn(t) L (13.9)
o
TIME DEPENDENT PERTURBATION THEORY 305
In view of Eq. (13.7b), in the first approximation we may replace on the

right-hand side of Eq. (13.9), Cn(t) by O for n -=I= s and equal to 1 for n = s.
Thus we get in the first approximation
(13.10)
and for m = s
i!iC8(t) = AC8(t)Vss(t)
The solutions of Eqs. (13.10) and (13.11) are
L (13.11) o
Cm(t) =-~AJ
t
ei(e:m-t:s)t' fñVms(t')dt', m -=I= S L (13.12) )
o
to
Cs(t) = e-i>./ñf,~ Vss(t')dt' L (13.13)

Note that ICsl = 1, only its phase has changed. Now
'1/J(r, t) = ¿ Cn(t)un(r)e-it:nt/ñ. (13.14)
n
But from Eq. (13.13)

(13.15a)
where
- = -- 1
Vss
t - to
1.t v';.s (t ') dt ·
to
1 (13.15b)
Thus separating out the n = s term in Eq. (13.14) and using

Eqs. (13.15) and (13.12) (for m = n), we have
~ '1/J(r, t) ~ Us(r)e-i(es+>.Vss) t ~ to
+.X¿(-~) 1.t ei(en-es)t' fñVms(t')dt'
X .:;;e-i,,</: k_ 7:-? (13.16)

o
Originally 'ljJ is given by
<p = Us(r)e-iest/ñ l-
Time dependence of the s-th state after the perturbation to the lowest order
is
(13.l 7a)
o
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
i.e. we consider transitions caused by a perturbation independent of t.

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
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
= lim t, far Wns = O.

t-+oo
(13.23)
o
Further
L
Thus
.
hm (
t-+oo
Hence far large t, we may replace

. 212Wns t
Slil
Wns
/2)2
t
1
= m5( -Wnst)
2
L (13.24) o
sin2 ~Wn8t (1 )
(~ )2 = 8 2wnst 1ít
= 21rM(cn - €8). (13.25)

Thus Eq. (13.21) becomes far large t
21r>.2
P(s, n) = -ñ-1Vnsl28(€n - és)t. (13.26)
Hence the transition probability per unit time (in the limit of t ----> oo) is
given by
(13.27)
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
Fig. 13.1 Plot of 4 sin2w :'.;'0•t verses wns

ns
We note that width is proportional to f.

Height of the middle peak: t2•
P(s,n) is large only for those final states that satisfy
2rr
t
21rñ
(13.29)
i.e. !}.E!}.t ,....., ñ.

13.2 Problems
13.1 A harmonic oscillator is perturbed by a weak, time dependent but

spatially uniform force F(t), where F(t) = O for t < O. Initially,
(i.e. t < O) the system is in its ground state IO). Up to first order
in F, show that the probability of finding the system in state l 1)
at time t is given by
P(O, 1) = 2~nw llt eiwtF(t)d{
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
STATISTICS AND THE

EXCLUSION PRINCIPLE
14.1 Introduction
Identical particles cannot be distinguished by means of any inherent prop-

erty; however in classical mechanics, they can be labelled in sorne way so
that they do not lose their individuality despite the identity of their physical
properties. This is because in classical mechanics the existence of sharply
definable trajectories far individual particles makes it possible, in princi-
ple, to distinguish them by their paths and each particle can be fallowed
during the course of an experiment. In quantum mechanics, it is assumed
that far identical quantum particles such as electrons, no such distinction is
possible. Consider a quantum mechanical system of two identical particles
described by the state function
'l/J(l, 2) (14.1)
where each of the numbers represents all the coordinates (position and
spin) of one of the particles. The Hamiltonian of the system H(l, 2) must
be symmetric far exchange of particles
H(l, 2) = H(2, 1). (14.2)
This is because an energy measurement cannot distinguish between the
particles. Also jiµ(l, 2)12 determines the probability far finding one particle
here and the other there, but can make no distinction as to which particle
is which. Thus the probability distribution is unchanged by exchange of
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.e. the wave function is either symmetric or anti-symmetric under the

interchange of the particles' coordinates. It is found that the choice of pos-
itive or negative sign is a property of the type of particle. For an electron
(in general for half integral spin particles, called fermions), the sign must
always be negative. As will be seen below this leads to the Pauli Exclusion
Principle. The state function for two electrons (in general for two identical
\ fermions) is always anti-symmetric for exchange of their coordinates. Elec-
tJ trons are then said to satisfy Fermil¡Dirac statistics. On the other hand,
a state function for two photons ( or in general for integral spin particles,
called bosons) must always have positive sign in Eq. (14.2) or be symmet-
~ if 1 and 2 are interchanged. Photons are said to satisfy Bosq{Einstein J..
statistics. Thus there is an intimate connection between spin and statistics. fJ
14.2 Permutation Operator and Exclusion Principie
We now introduce an operator F12, which interchanges all the coordinates

of particles 1 and 2.
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
Í'i2H(l, 2)1/;(l, 2) = H(2, 1)1/;(2,1)

= H(l, 2)Í'i21/J(l, 2). (14.8)
Hence
(14.9)
In other words, Í'i2 is a constant of motion. It means the symmetry which

a state initially has, does not change with time.
The situation with more than 2 particles is more complicated. Suppose
there are 3 particles described by an arbitrary wave function 1/;(1, 2, 3).
STATISTICS AND THE EXCLUSION PRINCIPLE 313
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
Pi2P13'!/J(l, 2, 3) = Pi2'!/J(3,2, 1) = '!/J(2, 3, 1)[_ )

Pi3Pi2'!/J(l, 2, 3) = Pi3'!/J(2,1, 3) = '!f;(3, 1, 2), (14.11)
so that
Pi2Pi3 -# Pi3Pi2 , i.e. [P12, P13] -# O L_ (14.12)

0
Thus we cannot find a complete set of eigenfunctions belonging to all the
P's simultaneously. But this <loes not mean that there are no simultaneous
eigenfunctions at ali. For example, even though angular momentum opera-
tors i.; L; and t; do not commute, the function Y00(0, '1/J)Rno(r) f(r), =
which is spherically symmetric is an eigenfunction of all three of them with
eigenvalue zero. In the present case, there exist exactly 2 kinds of '!f;'s that
are eigenfunctions of all the P's, one totally symmetric and one totally
anti-symmetric:
Pij'I/Js,a(l, 2, 3) = ±'1/Js,a(l, 2, 3) (14.13)

where
'1/Js,a(l, 2, 3) = { '1/J(l,
2, 3) ± '!/J(2, 1, 3) ± '1/J(l, 3, 2)
±'!f;(3, 2, 1) + '!/J(3, 1, 2) + '!/J(2, 3, l)}. (14.14)
Out of 3! = 6 independent states, there are 4 further states with more
complicated permutation properties than the above two states.
We need not concern ourselves with such states because of the following
experimental fact. Experimentally for each system of n identical particles,
the states are either completely anti-symmetric or symmetric under the
interchange of coordinates of any two particles:
Pij'I/J(l, .... , i, ... j, ... n) = '1/J(l, ... .i, ... i, ...n)
= -'1/J(l, ... , i, i. n), n identical fermions l )
= '1/J(l, ... , i, i, n), n identical bosons. (14.15)
This property depends only on the species of particles and not on the num-
ber n.
The anti-symmetry of the wave function for identical fermions implies,

for example
'l/J(2, 1, 3, ... ) = -'l/J(l, 2, 3, ... ) (14.16)
so that if 1 = 2, i.e. if 2 identical fermions have the same values of their co-
ordinates then 'ljJ vanishes. This shows that identical fermions obey Pauli's
exclusion principle, which states: Two identical fermions cannot occupy the
same point if they have the same spin orientation; nor can they have the
same value of momentum if they have the same spin orientation; in general
two (or more) identical fermions cannot occupy the same state.
14.3 Non-interacting Particles and Exclusion Principie
We first consider a system of 2 non-interacting identical particles so that

we write the Hamiltonian as
Ha= Ha(l) + Ha(2). (14.17)

We now write a solution in the forro
'l/Ja,{3 = 'l/Ja(l)'l/J{3(2)L (14.18)

)
E= Ea +Ef3, (14.19)
where 'l/Ja(l) and 'l/Jp(2) are solutions of
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
Ha= Ha(l) + Ha(2) + · · · Ha(n). (14.22)

Then a possible eigenfunction of Ha is
(14.23)
I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics !
with eigenvalue
E= Ea+ E¡3 + · · · + Ev. (14.24)

The other possible eigenfunctions can be obtained by permutation of the
n numbers 1, ... , n. There are n! possible wave functions corresponding to
the same eigenvalue E. We thus have n! fold degeneracy.
In general Eq. (14.23) is not an admissible solution for n identical par-
ticles since it lacks symmetry under the interchange of any two particles. J._,
For particles which obey Fermip\l)irac statistics (e.g. electrons) the an- ¡J
tisymmetric state functions must be constructed from the products of the
above type. This is most easily expressed in the form of the determinant:
'lf'o.(l) 'lf'f3(1) · · · 'lf'v(l)

_ (~) 1/2 'lf'o.(2) 'lf'f3(2) · · · 'lf'v(2)
'lf'A(l,2, ... )- (14.25)
n.I .
..
.
.. ...
'lf'o.(n) 'lf'f3(n) · · · 'lf'v(n)
This has the required property of anti-symmetry, since exchanging any
pair of particle coordinates is equivalent to interchanging two rows of the
determinant which changes its sign. Equation (14.25) has the important
property that it vanishes if two or more of the 'lf''S are the same since then
two or more columns of the determinant are the same e.g.
'lf'A(l, 2, ... , n) = Ü if 'lf'o. = 'lf'f3 l (14.26) .:

Thus no two ( or more) electrons can be in the same tate. In other words,
there cannot be more than one electron in any of the states a, /3,· ··,v.
This is known as the Pauli exclusion principle which was postulated as an
explanation of the periodic system of the chemical elements. \
j For a ,,vst rn of n photon µ,(or any partícle which satisfies Bos~instein
statistics)ene must construct symmetric stat functíons from products of
7J
J type Eq. (14.23). In this case any number of the particles may be in the
same state.
Illustration:
Consider a system of two particles 1 and 2 with two possible states 'lf'o. and
'lf'{3.
'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)
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.
14.4 Two Electrons System (Helium Atom)
The Hamiltonian of the system can be written as

Pi f>~ 2e2 2e2
H--+-----+-
e2
(14.28)
- 2m 2m r1 r2 r12 '
where r12 is the distance between the two electrons. r1 and r2 are the
distances of electrons 1 and 2 from the nucleus which is taken to be very
heavy. We write
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)
where Uc, (ri) is a solution of the eigenvalue equation

Pi - -2e2)
( -2m r1
uc,(ri) = Ec,ua(r1). (14.31)
We can also write the solution

(14.32)
Here electron 1 is in state f3 and electron 2 is in state o:, but the energy
E= Ea+ E13. Thus we have two fold exchange degeneracy corresponding
to two eigenfunctions Ua.f3 and Uf3,a with eigenvalue E. This is for the
unperturbed Hamiltonian. When the perturbation is turned on, the wave
functions become different; there is a shift in the energy and degeneracy is
removed. The matrix elements Yij are given by
XVi1 = e2 Ju~ _!_u1d3r1d3r2

r12
L (14.33a) )
>.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)
The energy eigenvalues corresponding to eigenfunctions 'l/J± are as given

in Eq. (14.36). Note that 'l/J+ is symmetric with respect to interchange of
electrons 1 and 2 while the other 'l/J- is anti-symmetric. If the electrons
are in the same state, only 'l/J+ is possible. Also if the electrons are at
the same place, the anti-symmetric wave function vanishes. The electrons
represented by the symmetric wave function 'l/J+ are said to be in the para
state and electrons represented by anti-symmetric wave function 'l/J- are said
to be in the ortho state. So far we have neglected the spin of electrons. We
now take into consideration the spin of electrons. Electrons are fermions,
thus the total wave function must be anti-symmetric under interchange of
electrons 1 and 2. For para state, the spin wave function must be anti-
symmetric [spin singlet]:
1
x-(a1, a2) = v'2 (x+(l)x-(2) - x+(2)x_(l))
= -x-(a2, a-1), (14.39)

Thus for para state the wave function is given
(14.40)
For ortho state, the spin wave function must be symmetric [spin triplet].
In this case there are three linearly independent states viz.
(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
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.
14.5 Scattering of ldentical Particles
Consider the scattering of two particles a and b in the centre of mass system.
b
b
n-e
~ 2
Fig. 14.1 Scattering of particles at an angle (} and the angle 1r - (}.
Let f(B) be the scattering amplitude for scattering of particle a at an

angle B. Let f(1r - B) be the scattering amplitude for scattering of particle
a atan angle 1r - B. Now f(1r - ()) = the amplitude for particle b scattering
through the angle (). Thus the probability of sorne particle detected at D1
is given by
O"(e) = IJ(e)l2 + lf(1r - e)l2. (14.43)
The two amplitudes f(()) and f(1r-(}) are distinguishable in principle when
particles a and b are not identical.
Consider the scattering of two identical particles. First consider the

case of scattering of two identical Base particles of spin zero, far example,
a particles. In this case there are two ways to get an o-particle into the
counter D1: by scattering the bombarding o-particle by an angle () and or by
scattering it by an angle 1r - (). Since the particles are indistinguishable, we
cannot tell whether the bombarding particle ar the target particle entered
the counter. Thus the two amplitudes f(()) and f(1r - ()) can interfere. The
wave function must be symmetric. Thus
(]"(()) = IJ(()) + f(1r - ())12,i (14.44)
or
(]"(()) = lf(())l2 + lf(1r - ())12 + 2Re J(())f*(1r - ()). (14.45)
'
This differs from the classical result by the interference term
2Re f(())f*(1r - ()). Far example, far()= 1r/2, the classical result gives
O"( 1r /2) = 21! ( 1r /2) 12 (14.46)
compared to the result obtained using Base statistics
0"(1r/2) = 4lf(1r/2)12 (14.47)
Now consider the case of scattering of two identical Fermi particles of spin
1/2, far example an electron or proton. Here the total wave function must
be anti-symmetric. If the two particles are in a spin singlet state (spin wave
function anti-symmetric), then the spatial wave function is symmetric and
O"s(()) = IJ(()) + f(1r - ())12 L (14.48) O
When the two particles are in a spin triplet state then the spatial wave
function must be anti-symmetric
O"t(()) = lf(()) - J(n - ())12 k_ (14.49)0
For unpolarised particles, noting that there is one singlet state and three
triplet states
1 3
O"(())=
40"s(()) + 40"t(())
1 3
=
41f(()) + f(n - ())12 + 41J(()) - f(n - ())12
1 1 1
=
21J(()) - f(n - ())12 + 21f(())l2 + 21J(rr - ())12 (14.50)
= lf(())l2 + lf(rr - ())12 - Re f(())f*(n - ()). (14.51)
Again we see that this result far scattering of two identical Fermi particles
is different from the classical result.
14.6 Problems
14.1 A deuteron consists of a neutron anda proton. Considera system

of 2 deuterons as a composite of 2 neutrons and 2 protons with the
wave function
where ti and p denote all the coordinates of a neutron and a proton

respectively. How does the above wave function behave under the
interchange of 2 deuterons, i.e. under
14.2 For a system of 2 identical particles, is
where
'l/J10o(r) = ~¡33/2e-í3r,
'l/J2oo(r) = ~¡33/2(2 - ,6r)e-f3r,

4v21r
an acceptable wave function? If not, find an acceptable wave func-
tion. If the particles have spin !,
write down acceptable wave
functions which include the spin wave functions also.
14.3 For a helium atom (Z = 2), show that the ground state energy to
first order in perturbation is
Ea = Ebº) + 6.E
where
Z2e2 Ze2
Ebº) = ---, 6.E = (5/8)-
ao ªº
Hint: Ground state wave function is
where
Uioo(r) = ~( !._ )3f2e-Zr/ao

vn ªº
Apply Eq. (14.33a) with the above wave function.
14.4 Consider an electron confined in a cubica! box of length L. Show

that the energy eigenvaluesand normalised eigenfunctions are given
by
)
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,
where the factor 2 is due to two spin orientation. Consider now

a non-interacting electron gas in the box. A quantum state is
specified by three quantum numbers (n1, n2, n3). Since an electron
has spin 1/2, only two electrons can be in the same state specified
by (n1, n2, n3). Now the Pauli principle forces the electrons to
occupy all the states between O and EF, where EF is the Fermi ( L~
energy and for non-relativistic case is given by EF = ~ :howthat O~
EF = 7í2 n2 (~) 2/3
2m m1r
where p = {Ji m, is the density of electrons in the box. Find the
pressure as a function of p viz. the equation of state for the de-
generate gas. Obtain also the equation of state for a relativistic
degenerate electron gas ( E F ~ cp F) .
14.5 Consider two electrons confined in a box whose sides are of length
L. There exists an attractive potential of strength V0 between pairs
of electrons whenever they are very close to each other. Using per-
turbation theory, calculate the ground state energy and the wave
function. Hint: Take the potential { )
V(r1 -r2): ~Vo ,J;,::JL.

1
(Á ~VV )
Then approximate this potential aJ r >
J V(r1 - r2)d3r2 = -Vo 4; a3
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.

Chapter 15
TWO STATE SYSTEMS
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)}:
(ilj) = Óij, i,j = 1, 2. (15.1)
Let ÍI be the Hamiltonian of the system which in the above basis is a

O'\ 2 x 2 hermiti¡~n matrix
H = (Hu H12), H21 = H[2; Hu,H22 real. (15.2)

H21 H22
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
gives two eigenvalues
(15.3)
which are possible energies of the system. We denote the corresponding

eigenstates by I+) and 1-), which are the stationary states of the system if
H <loes not depend explícitly on time.
325
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)
(' This gives [n = lníl

V "' (15.8)
which fixes only u±/V± i.e the relative phase between U± and V±, Using
the normalizing condition
(15.9)
then apart from an overall phase, the normalized eigenstates are
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]
= ~J(Hn - H22)2 + 4H12H21 L )

co () = HIJ
2n
-H22L /
H12 +H21
cos <P = 2IH12\ . (15.11)
TWO STATE SYSTEMS 327
Then the normalized eigenstates of H are given by ( i~os. 6) ! and
7
e sm2
(-e::: !n !)with the eigenvalues E± = no ± n,,¡respectively.

It is easy to see that above states are eigenstates of helicity operator
fcit7 with eigenvalues ±l.
Since 11) is G) and 12) = (~),
() ·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
iñ. :/1/J(t)) = Hl'l/J(t)). (15.14)

In the basis provided by {11), 12) }, Jl)(ll + 12)(21 = 1, we can express l'!/l(t))
as
l'!/l(t)) = ll)(ll'!/J(t)) + l2)(2l'l/J(t))
= 'l/J1(t)ll) + 'l/J2(t)l2). (15.15)
Thus
in!_ ('l/J1(t))
=H ('l/J1(t)) (15.16)
8t 'l/J2(t) 'l/J2(t)
where His given in Eq. (15.2 ). Using Eq. (15.3 ), we can also express l'!/l)
in the orthonormal basis { 1 +), 1-)}:
l'!/l(t)) = '!/l+(t)I+) + '!/l-(t)I-) (15.17)
where
(15.18)
In other words
where U is the Unitary matrix

(~~~!D u (~:~:D
=
(15.19)
cos « e-ict, sin fl_)

u= ( -e•ct,. sin2 cos
2
! , uo' = trt: = 1.
! (15.20)
Thus the two representations are related by a unitary transformation. Then

from Eqs. (15.16) and (15.19)
in.!!_ ('l/J+(t)) = ontr:' ('l/J+(t))

at 'l/J_(t) 'l/J_(t)
-mu .!!_u-1 ('l/J+(t)) (15.21)

at 'l/J_(t)
where
UHu-1 = (E+O O) E_
(15.22)
i.e. in the basis {I+), 1-)} the Hamiltonian is diagonal.

If U is independent of t, then
'l/J±(t) = 'l/J±(O)e-iE±t/li. (15.23)
'l/;1 (t)) - -1 ('l/J+(O)e-iE+t/ñ.) (15.24)
( 'l/J2(t) - U 'l/J-(O)e-iE_t/li .
l'l/J(t)) = [e-iE+t/li'l/J+(O)I+) + e-iE-t/li'lj;_(O)I-)]
= e-iEt/li¡e-i6Et/ñ.'l/J+(O)I+) + e+i6Et/ñ.'l/J-(O)I-)] (15.25)
where E= E+!E_, 6..E = E+;E-. Suppose at t = O 'l/;1(0) = 1, 'l/;2(0) = O,
then from Eq. (15.18)
'l/J+(O) = cos 9 'l/J-(0) = -e'~ sin 9 (15.26)

2, 2.
Then using the Eq. (15.12), in the basis {11), 12)},
l'l/J(t)) = e-iEt/ñ.{(cos2 ~e-i6Et/li + sin2 ~ei6Et/ñ.)ll)

2 2
+eict, sin~ cos ~(e-i6Et/li - ei6Et/li)l2)}
2 2
. 6..Et 6..Et
e-•Et/ñ.{(cos -- - i cos Bsin --)11)
ñ ñ
-ieict, sinBsin 6..:t 12) }. (15.27)
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)
The survival probability l

. 2 u{) sin
P1_1 ( t ) = 1 - sin • 2 !:iEt
T L (15.29)
C)
The above equations illustrate the quantum mechanical phenomenon of
interferometry which provides a sensitive method to probe extremely small
effects. Eqs. (15.28) and (15.29) are plotted in F'g. 15.l.
P,_, = 1-},'_,
Fig. 15.1 Thel'P'robability far spin flip.
This shows that a state produced as pure 11) state at t = O continuously

oscillatelJ;etween 11) and state 12) with a period of oscillation T = 1:;.';;¡ñ..
Thus by measuring T, one can measure !:iE/n.
15.2 Magnetic Spin Resonance
The spin system is a two state system vi~e states ()

1 1 1
11) = 12), Clzl2) = 12)
1 1 1
12) =1- uz 1 - =1- (15.30)
2), 2) 2)
forma complete set. In electron spin resonance experiment!Jhe spin moves

under the joint action of a large uniform magnetic field B inf;:direction
which is constant in time and a small oscillatory magnetic field b(t) in the
) Xfb, plane:
yJ
B=BezL
b(t) = b(excoswt) +eysinwt). (15.31)
)
Spin Hamiltonian
H = -µ · [B + b(t)]
eñ 1
µ = 982mc 20' (15.32)
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
iñ ! lx(t)) = Hlx(t)). (15.35)
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
-
explicitly depends on t, to deal with it we write it as
H = r,,vt(t) ( ~ -~,~) V(t)

~
= vt(t)HV(t) (15.36)
where
e;2t o )
V(t) = -iwt
( O e----Y-
(15.37)
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)
H' = iñ. dV(t) yt (t) + ÍI

dt
·2
= Z n2Wuz + H- = Ws -W
fi(-2-uz+
Wb
2ux)·
(15.42)
and is independent of t. H' is not diagonal and the unitary matrix which
diagonalizes it is given in Eq. (15.11), where now [cf Eq. (15.20)]
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)
lx(t)) = vt(t)ix'(t)) (15.46)

o
= e-iwt (cos ~t - i cos B sin ~t) 11) - ieiwt sin B sin ~tl2).
(15.47)
Therefore the probability that spin has flipped by time t i.e. the particle
has spin down is given by
P_! = l(-~lx(t))l2 = sin2 Bsin2 ~t
= w~
(w - w8)2 + w~
. 2 [((w-ws)2+w~)!
sm
2
t l (15.48)
At resonance w = W8
P_t = sin2 wbt (P_t)max = 1 (15.4.9)

Wb is known as the Rabi frcquency ~o first solved this problem. Above is a
model for pin resonance experiment ·. f the fi Id B is known, a measurement
*
¡r
of resonanc frequency w = w8r,measme · the param r 9s· The interese
in 9s is to detect any deviation from the value 9s = 2 predicted by Dirac
equation. The experimental value for the parameter:
1
a= ± 2)
2 (gs - 2) is aexp = (1159652.4 X 10-9•
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
called lepton number L = 1; bl;ilepton has L = -1. Thus corresponding

to leptons (e+, ve), (µ,+, vµ.), (r+ V-r ), antileptons are (e-, De),(µ,-, Dµ.) and
( T-, IJ-r). In any process lepton number and charge is conserved.
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:
( 7r- ,7r o ,7r +) , I = 1, h = -1,0, 1, Y=O
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
1r± ---> µ,± + vµ.(Dµ),

1rº --+ 2,,,
Kº ___, 7r+1r-.
«: is antiparticle of n+; (Kº,K-) are antiparticles of (Kº,K+). Mesons

and antimesons have same mass, spin and parity, but opposite !3 and hy-
~ percharge. In weak decays, isospin and hypercharge are not conserved.
i\. We conclude this section with the remark that baryon number and elec-
tric charge are absolutely conserved; we can consider particle mixing only
ror K o-:-;()
e
and K and heavy neutral mesons Bdo=° - Bd and B8o- - B08• It is
also possible to consider mixing between neutrinos which are electrically
neutral.
In a production process involving strong hadronic (or electromagnetic)
interaction hypercharge Y is conserved and Kº and K0 appear as two
distinctly different particles. In the presence of weak interaction Y is no
longer conserved and transitions between Kº and K
can occur. Both Kº
and K° decay into n+n-:
Kº -t n+n- t- K° , ~Y=~lt- )
-o
Kº -t 7r+7["- -t K, l~YI =2.
Hence weak interaction can mix Kº and K°:
(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.
15.3.2 General Formalism for Particle Mixing

We now discuss particle mixing for µ<º, R°)
complex or .in general (xº, x°).
For an unstable particle, the cornplex mass M = m - !r, where m is the
mass of the particle and r is decay width of the decaying particle. In
lxº), lxº) basis (ñ =e= 1), we can express a state l'l/J(t)) as
L
l'l/J(t)) = a(t)lxº) +a(t)lxº) (15.50)
where t is measured in the rest system of the particle xº. Then the time
evolution of the state
'lfJ(t) = (;~:D
is given by
i :t '1/J(t) = M'l/l(t). (15.51)

The mass matrix M given in Eq. (15.51) is not diagonal in this basis
and is given by
12).
:22
(15.52)
Hermiticity of the matrices maa' and r aa' gives (a, a'= 1, 2)
(m)aa' = (mt)aa' = (m*)aa'i

m21 = mi2,
r cea'
r21
=
=
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)]
m'f - ;r 'f = [(m11 - ~r11) =f (m12 - ~r12)(mi2 - ~q2) ]
= (m11 - ~r11) =f VM12M21- (15.55)

We can write
M=no+u·n (15.56)
where
i
no = m11 - 2T11
n = ( M12 + M21, i(M12 - M21), O) (15.57)
so that the magnitude of n is

n = (M12M2i)112. (15.58)
and from Eq. (15.10), eigenvectors corresponding to eigenvalues (15.55) are
(15.59)
Thus the mass eigenstates are [c.f Eq.(15.13)]
IX:¡:) = ~ [lxº) =f ~ lx°)] (15.60)
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):
lxº) = ~ [a+(O)(lxº) + ~lxº)) + a_(o)(lxº) - ~lxº))].

This gives
1
ª±(O)= v2· (15.65)
Hence from Eq. (15.62) after time t,
11/J(t)) = ~[exp(-im_t - ~r _t)lx-)
+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~¿
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)
where the CP phase T/CP is arbitrary (lr¡cpl2 = 1) but we select it to be

!
-l. Then we see from Eq. (15.60) that if = 1 the mass eigenstate IX:¡:)
will be the eigenstates of CP with eigenvales ±1 respectively, which we
designate as lxi,2). From Eq. (15.61) it is clear that !
is unity if m12 and
*
r12 are real. Writing = 1 + E, then E is a measure of CP violation and
arises due to mismatch between mass and flavor eigenstates:
lx-) = lx~) + Elxg)

lx+) = Jxg) + Elx~) (15.71)
-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
probability of finding Y = -1 component at t in the the kaon produced at

t = O in apure Y= 1 state is given by Eq. (15.68) (l~I « 1, see below)
P(Kº--+ R°, t) = l(Kl1/;(t))l2

t
=
1
-{ exp(--) + exp(--)t 1 t
- 2 exp[--(- + -t )] cos(b.m)t}
4 rs TL 2 rs TL
where b.m = mL - ms and since TL is much larger than rs
P (K o -:-:O)
K , t = 1( 1 _ rs cos(b.m ))t .
4 + e ""S 2e_!_..L
__L
--+ - 2
(15.72)
If kaons were stable (rL, re=- oo), then

o -O 1
P(K --+ K , t) = - cos(b.m)t) (15.73)
2(1
which shows that the state produced as pure Y = 1 state at t = O continu-
ously oscillates between Y = 1 and Y = -1 states with frequency w = f\m
and period of oscillation,
(15.74)
Kaons, however, decay and their oscillations are damped. By measuring

the period of oscillation, b.m can be determined:
b.m = ttvt. - ms ~ (3.5)10-12 M eV. (15.75)
Such a small number is measured as a consequence of quantum phenomena

of interferometry
Finally from Eq. (15.71)
IKs) = IKP) + EIKg)

IKL) = ¡Kg) + EIKP) (15.76)
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
which experimentally is of order 2 x 10-3, showing that !El « l.

15.3. 4 B qo - Bºq comp lex I

Iq = d, s )
f For Bº - Bº;1complex (suppressing the subscript q) it is known that f12

and m12 have the same phase and lf12 « lmd. Thus from Eq. (15.61)
q/p = e-2if3
where m12 = lm12le2if3. The mass eigenstates B'f', which we now denote as
BL,H are:
o
IBf) = ~[IBº) - e-2if31IJ°)JL
IBt) = ~[IBº) + e-2if31IJ°)J L (15.77)
In the limit (3--. O
IB2)--, IBf) = ~[IBº) - [B0)], CP = +1 l )

[B~)--. [Bg) = ~[[Bº) + [F)], CP = -1. (15.78)
The mismatch between mass and CP eigenstates is a possible source of

CP-violation in Bº decays. For Bº - Bº, complex:
~r = r1 - r2,
G ~m = m2 - m1 (15.79)
If we start with [Bº), then from Eq (15.66)
l'I/J(t)) = e-frnte-!rt[cos( ~m )t[Bº) - ie-2if3 sin(~m t)[F)J (15.80)

2 2
where m = m1tm2.
The probabilities of finding [B0) and [Bº) at time t are given by
P(Bº--. B0, t) = l(B01'1/J(t))l2

~m
= e·-rt sin 2 (-t)
2
= ~e-rt(l - cos(~m)t) L. (15.81)
)
P(Bº--. Bº,t) = l(Bºl'I/J(t))l2
~m
= e -rt cos2(-t)
2
1
= + cos(~m)t). (15.82)
2e-rt(l
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.
The mixing parameter
(15.83)
Experimentally, far B~ ,B~

6.mBo~3.34x10-
d
10MeV, TBo=l.52x10-
d
12s L )
6.mB~ ~ 117 x 10-10MeV, TB~ = 1.47 x 10-12s L )
(6.mBº)
Xd =( T)
6.mBo
B~
= li d TBD ~
d
0.77
(6.mBo)
Xs = /i • TB~ ~ 26.2.
-/\ Nonbro values of Xd and

of quantum interference.
Xs clearly show mixing between Bº, Bº; a proof
15.3.5 Neutrino Oscillations

The neutrinos occur in three flavors, Ve, vµ, V-r which are produced in
decays of charged leptons. If neutrinos are massless, then these are also
the mass eigenstates. If any of them have a mass, then mass eigenstates
which we denote by 1/i (i = 1, 2, 3) are in general different from the flavor
eigenstates Vo: (a= e,µ,T). We note that two set of states are connected
with each other by a unitary transformation
(15.84)
where
(15.85)
The mismatch between flavor and mass eigenstates gives rise to neutrino
oscillations.
For example, for the conversion of Ve to V» (x = µ, T), we have from Eq.

(15.27), where states 1±) are mass eigenstates lvi), i = 1, 2 and states 11)
and 12) are flavor eigenstates lve) and lvx) and we change to~ !- () CM~f IV\
to the usual convention used in neutrino physics [4> = O]:
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
P(ve - Ve, t) ~2 Wsin2( l~:;" L) {i,~, V\

~n' 20 sin'({) L (15.91) )
P(ve - Ve, t) = 1- sin2 Wsin2 t}). (15.92)

t
1
+
sin2 2()
t
-L- X= et
Fig. 15.2 Neutrino oscillations.
The probability P oscillates with frequency f ~ amplitude sin2 2(). The

amplitude of the oscillation sin2 2() is given by the mixing angle. These
oscillations are shown in ig. 15.2.
It is clear from above analysis, that oscillations are between neutrinos
of same momentum but different masses. To look for oscillations, one needs
factors, which enhance tiny ffects: a coh rent sourc (there are many, the ('
sun, co ·mic rays, reactor etc) low energy neutrinos large base line (;ize J
í of sun, and that of the arth)0 large mixing angle and large flux.
For 3 flavors, it is customary to parameterize mixing matrix elements
Uai (not all independent) by three mixing angl~ fh:1, 023 and one com-
plex phase 8
1ue212 2 1uµ312 2 1uea12 2 \
IUe1l2 = tan e12, IUTJ!2 = tan 023, IUe.112 = tan 813 h_
The mass eigenstate lví) (i = 1, 2, 3) has a well defined mass mi and it is
customary to order the mass eigenvalues such that
A 2 2 2
D.ffi12 = m2 - m1 >Q
and
.tf- where tlm2 >o}< O) correspond to normal (inverted) mass spectrum hier-
archy. We also fuí'nmarize the various sources which provide neutrinos
(1) Reactor: n--, p +e-+ Ve.

(2) Solar: in pp cycle 4p _,4 He+ 2e + 2ve + 26.7MeV.
The depletion of Ve flux from the sun would indicate the conversion
of Ve to another species of neutrinos, e.g. Vµ- This is observed
experimentally.
(3) Atmosphere: Interaction of Cosmic rays with atmosphere
p+A-, 1r± +A'
_, 1r±vµ(vµ)
- e±ve(ve)vµ(vµ)·
Thus one would expect ~~~ej ".: : '
2, if there is no osillation i.e. no
flavor change. But experimentally it is substantially reduced {9qm
'""2, providing an evidence for oscillation.
(4) Acceleratorl
Typically, a single experiment is mainly sensitive to only one of the
above mass gaps and to one mixing parameter, although subleading effects
driven by remaining parameters may become relevant in precision oscilla-
tion searches.
So far, solar and long-baseline reactor neutrino experiments have mea-
sured the mass-mixing parameters (flmr2, 812) in the Ve --; Ve channel,
while atmospheric and long-baseline accelerator (LBL) experiments have
measured ( tl m 2, 023) in the vµ --, vµ. channel. Conversely short-baseline
().,{l reactor experiment~mainly sensitive to ( tlm 2, 013).
To summarize detailed combined analysis of all neutrino data give best
fit values:
tlmi2 = 7.65 X 10-5eV2
flm2 = 2.40 X 10-3eV2 ~
L )
)
812 '::::é 341- )
023 -::::é 45º[_.
813 ".: : ' 9°.
Currently, there is no constraint on the sign of tlm2. To conclude there is
'
compelling evidence that neutrinos change flavor, have non zerolrpasses and /\.
that neutrino mass eigenstates are different from weak eigenstates. Note
again that the quantum mechanical phenomena has enabled us to measure

differencesof arder 10-3 to 10-4e V2.
To conclude this chapter. We have provided a unified treatment of
quantum interference and~shown that this have been observed in variety
of cases with no connection with each other, except that all are two state
systenl In the next chapter we discuss another application of two state
systeiñp:n quantum computation.
15.4 Problems
15.1 Show that ( i~o~!

e sm 2
e)
and (-ei'Ps~n!) are eigenstates ofthe he-
cos 2
licity operator l~Iwith eigenvalues ±1 respectively. Here
l~I = (sin B cos cp, sin B sin cp, cos O).
15.2 Show that the probability of finding the particle in eigenstates of
<rx with eigenvalues ±1 is
11 .n. .n n ..
Pxil =
2{21 ± 2(sm2
2tsmBcosBcoswt + sm 2tcos 2tsmBsmwt)}. /
What is the corresponding result for Py¡ ivl

Hint: Eigenstates of <rx and <ry are respectively
1 1 1 1 1 1
-[I \/'2 + -)
v'2 + -2 ) ± 1 - -2 ) l ' -[1 2
±il - -)].
2
Chapter 16
Quantum Computation
16.1 Introduction
Theoretical and conceptual understanding of the fundamental principles of

quantum mechanics, as explained in earlier chapters, lead to develop a tool
box to engineer computers and their algorithms based on quantum laws.
The discussion begins with the basic elements of computers, such as, bits
and logic gates, ultimately leads to new concepts like super dense coding,
/
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
16.2 Qubit
In conventional computation the information storage and data manipula-

tion is done by means of the basic building block of hardware, that is, a
bit. Hence, a data written in binary form is mapped on a series of bits
as a sequence of zero and one, that conventionally corresponds to low and
high voltages, respectively. A quantum mechanical bit, that is a qubit or
qbit, requires a two state quantum system which may be spin of electrons,
interna! states of an atom or two modes of an electromagnetic field. Hence,
the system can be found in either of the two states. In Dirac notation we
label the two state as lü) and 11). As an example, we consider them indi-
cating ground and excited state of an atom, that is, lü) = lg) and 11) = le),
respectively. However in contrast to the conventional classical bit the qubit
appears in a superposition state. This feature marks a major difference in
classical and quantum computers and introduces a lot of philosophical and
conceptual debate. Hence, in general we express the qubit as
(16.1)
whereJf1 and c2 are complex numbers and define probability amplitudes,
such that
lc112 + lc212 = l.
Physically the superposition state may arise as the atom is performing a
transition 'lfi1m its ground state, that is, state vector IO) to 11).
Since the probability amplitudes, c1 and c2, are complex numbers we
may write then as
C1 = lcilei-rt
C2 = lc2lei'2•
Hencee_uation(l6.l) becomes . .
l'l/J) = lc1 le''1 IO) + lc2le''2 ll)
where 11 and ,2 are phases corresponding to state vector IO) and 11), re-
spectively. We may further simplify the above expression
l'l/J) = ei'Y1 (lc1IIO) + lc2leí'll)).
Hence, we identify 1 = 11 - ,2 as the relative phase and ,1 expressing a
global phase on the qubit. Without loss of generality we may express the
qubit as
(16.2)
Quantum Computation 347
Another useful representation of a qubit comes as we write the real valued

lc1I and lc2I in the parametric relation, that is,
lc11 = cos ( !) )_ )
lc2 I = sin ( !) .
This helps us to express the qubit in the most general form as:
11/J) = cos ( ! ) IO) + sin ( !) eÍ"fll). (16.3)
Hence, the definition of the qubit is based on 'Y and (), which makes the
qubit a point on a sphere oí uni radiu r in spherical coc rdinates (r 'Y, B).
It is known as the Bloch sphere and shown in Fig.~ . Frorn the above
equation, it i evident that Ior O= O and f r 6 = 1r the qubit, respectively,
is in )ID) and 11) state, whereas far any other value B it appears as another
poinh5n the Bloch sphere, hown in F ig. 16.1. It is irnportant to note that
IO) z
11}
Fig. 16.1
the qubit expressed as a general vector on the basis of two dimensional

Hilbert space is defined on a three dimensional sphere. From where one
extra dimension is appearing? The riddle is answered as we note that the
qubít is effectively deflned on the two dimensional surface of the Bloch q_,,
sphere as its radius is always fíxed to{th unity. J
16.3 Multiple qubit systems
In order to express a physical system which consists of a large setup of

many qubits, we express the state of the system as a tensor product of each
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.
16.4 Matrix representation of qubits
A very useful expression of the qubit comes from matrix representation.

We define the basis vectors IO) and 11) in matrix form, such that:
JO)= (~),11)= (~)

which leads us to write the qubit as
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
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)
We conclude that a single qubit or a multiple qubit system is described as

a column matrix.
16.5 Single Qubit Logic Gates
In conventional computation, manipulation of data is made by means of

classic logic gates. For this reason we define logic gates as operators and
transforrns which change the state of a systeT. of qubits according to re-
quirement. The mo t sirnplest among th m is tüT gate which transforms
a particular binary code into its compl!m nt.
In the quantum domain, we may define a Quantum NOT gate as an
e..
operator, X, such that
XIO) = [1) X[l) = [O).
We define it figuratively asWhe corresponding definition as the outer prod-
11
1>
X
Ü>
Fig. 16.2
uct in Dirac notation becomes

X= IO)(ll + ll)(OI. (16.8)
As we take the basis vector as column vector, as discussed above, we obtain

the matrix representation of the NOT gate, which is
(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
We define the phase gate in the Dirac notations as:

y= IO)(OI - 11)(11 l-
As we consider the matrix representation for the basis vectors, the phase
gate is
o
Y=(~ ~1). (16.10)
{¡\ Engineerínglsuperpo ition on a qubit is performed by means ofb1perposi-

t!on gat Z, whi h i. referred sorne times in Iiterature ~amard gate
H. The logic gate operates on either of the ba is vectors IO) or 11), and
( generat~up rpo sitlon b tween the two with equal probabílity, such that
[ ee íg 16.4;
BIO) =
A 1
1210}
+ 1
1211)
L )
, 1 1
Hll) = -10) - -11).
/2 /2
Quantum Gomputation 351
(IO>+ll>)/2
/\
H (IO>+ll>)/2
Fig. 16.4
In the outer product we define the Hadamard gate as

H = IO) + 11) IO) + IO) -11) 11)
v'2 v'2 '
whereas the corresponding matrix notation for the Hadamard gate is
~
H-- 1 (1 1 ) (16.11)
- y'2 1 -1 .
16.6 Quantum Logic Gates at work
In order to develop a physical insíght in the operation of quantum single

bit logic gates we consider the simple example of atoml\field ínteraction
in cavity QED (quantum electrodyna.mics). The atom\p&.fo~
between ground and excited state in the presence of field. The atom, there-
fore, acts like a qubit so that l1P) = c1JO) + c2ll), where IO) corresponds
to ground state and jl) defines excited state. Under the condition of res-
onance between the atomic transition frequency and field frequency, the
square of the probability amplitudes lc1(t)j2 and lc2(t)j2 change in time
periodically and alt · rna ively, as shown in figw· 16.5. Hence, cho sing a
time of interaction provides the desired opch.tion H
of a NOT gate and
R gate. on-dísper ive interaction of the atom with the field leads to tbe
phase gate, which <loes not change the state but the phase in a controlled
fashion.
16.7 Two Qubit Logic Gates
Control not logic gate (CNOT), Ucont, is a transformation which acts on

two qubits, one is the controlled qubit and the other is the target qubit.
In its operation the CNOT gate acts on a set of two qubits, such that+¡ it
t
does not change the controller qubit but modifies the target qubit using
modulus two operation. We define the CNOT gate mathematically as:
'llcont[jq1) ® Jq2)] = lq1) ® lq1 EB q2),
ICg 2
1
ICe 12
t
Fig. 16.5
here notation EB represent modulus two operation, rnoreover, lq1)

is the
controller qubit and lq2) is the target qubit [see Fº . 16.6]. Thus, for
Fig. 16.6
different input we have following results: L

'Ucont[IO) ® II)] = IO) ® IO EB 1) = IO) EB 11) )
'Ucont [IO) ® IO)] = IO) ® IO EB O) = IO) EB IO) V )
'Ucont[ll) ® IO)] = 11) ® 11 EB O)= 11) EB 11) (..,, /
'Ucont[ll) ® 11)] = 11) ® 11 EB 1) = 11) EB IO). (16.12)
Further analysis shows that the CNOT gate flips the target qubit if the
controller qubit is l. The CNOT gate together with a single qubit logic
gate acts as a universal logic gate. In quantum computation we perform
various different tasks by using the quantum logic gates and developing
corresponding circuits.
16.8 Quantum Circuits
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
cuitry is a schematic diagram of series of three control not gates in a certain

combination. Let's see that what is the output generated at each stage. We
consider the initial state l1/J1) which is in fact a combined state of la), con-
troller bit, and lb), target bit. As we operate control not operation on the
initial state l'lf'i), to generate state l1/J1), as follows
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
354 Qv.antv.m M echanics
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. )
l·t/;011t) = ~contllb} ® la EB b)] >

IV'out) = lb) © la EB b EB b) [_ )
IV'out) = lb) © la)·

Hence this circuit performs the swapping operation as it swaps the state of
two qubits.
Here we discuss possible quantum cloning circuitry using quantum logic.
The generalized schematic diagram for cloning operation is as follows.
Simply we ~e!{ control not gate and specify the target bit as zero. Hence,
'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.
16.9 EPR Entanglement
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
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
l '1'1 > J 'l'ou t >
r: 1\
\.. l./
Fig. 16.9
!1/!1) = [ !O) ~11)]
and output state is
11/Jout) = Ucont [11/!1) ® IO)] L )

+ lll)J
[[10) .,/'j, IO) ] t--._
1
11/Jout) ®
)
A
= 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.
It is important to note that for two different values of controller qubit

la) = IO) and la) = 11), we have two output states la+) and la-), viz.
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.
{'\
16.2 Show that Bell states, given as .. I

_ !O, O) f ------ , .
1 ª± ) - v'2, ~
) 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.
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.

July 9, 2012 14:29 World Scientific Book - 9i n x 6i n RF-QuantumMechanics I
Chapter 17
PERTURBATION INDUCED BY
ELECTROMAGNETIC FIELD
../
17 .1 Interaction of Electromagnetic Field with Electro ns
Classically the electromagnetic field is described by Maxwell's equations,

which in vacuum are given by
v' · B(r, t) = O (17.la) )

v' · E(r, t) = 47rp(r,t) (17.lb)
)
y' x E(r t) = _! 8B(r, t) (17.lc) .,)
, e at
n
V X
B(
r, t
) _
- ! 8E(r,
O
t) + 471" J'( r, t ) , (17.ld)
e t e
where p(r, t) and j(r, t) are the charge and current densities, that is, sources
of electromagnetic fields E(r, t) and B(r, t). Charge conservation is ex-
pressed by the continuity equation
ap~, t) +y'. j(r, t) = o. (17.2)
It is convenient to introduce a scalar potential cp(r,t) anda vector potential

A(r, t), such that
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
</> __, </>' = </> - ~ :t A(r, t) L (17.4a) )

A__, A' = A+ v' A(r, t), (17.4b)
359
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)
18( 'v · A ) = -41rp(r, t).

'v 2 <P + ~ (17.6)
8t
For static Coulomb field for which p(r, t) = p(r) and ~f = O, we can select
the gauge such that
'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)
Here we confine ourselves to static scalar potential.

PERTURBATION INDUCED BY ELECTRO MAGNETIC FIELD 361
By the correspondence principle, the Hamiltonian in quantum mechan-

ics is given by
H = ].___ (f> - ~A(r,

2m e
t)) 2 + ecp(:r), (17.13)
where in the Schri:idingerrepresentation

:r-----+ r, f>-----+ -iñV.
Hence
A2 2
H = .!:_ecp(r) - _e_(f> ·A+ A· f>) + _e_A2
2m 2mc 2mc2
ñ,2 2 e . . e2 2
=--V --(2A·(-zñ"V)-iñ"V·A)ecp+--A,
2m 2mc 2mc2
(17.14)
where we have used the fact that
f> · A = A · f> - in"V · A. (17.15)

This follows from the commutator
[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
_!!_V2'!/J = inº'I/J (17.18)

2m 8t
goes over to
2
-- ñ,2 ( i e ) 'ljJ
V - --A = in ( -8 ie )
+ -cp 'ljJ (17.19)
2m ñe 8t ñ
for a particle of charge e in an electromagnetic field.
We also note that p = 'ljJ*'ljJ = l'!/Jl2 gives the probability per unit volume
of finding the particle at r and when we find the particle at sorne place
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).
Suppose we <lid not know how to write the interaction of a charged

particle described by wave function '¡/; with the electromagnetic field (A, cp).
Can we then formulate a physical principle which would lead to the correct
law for this case? The answer is obviously yes from the previous discussion.
Therefore let us demand now that physical laws as expressed by the
Schródinger equation be invariant under the space-time dependent phase
transformation (gauge transformation of the second kind) as given in
Eq. (17.23) for an electrically charged system.
We see that the free-particle Schródinger equation (17.18) is not in-
variant under the gauge transformation (17.23). In order to restore gauge
invariance, we have to postulate the fields (A, cp) and replace
. -+
v'-+ v' - K~A = D (17.24)
íecp
at8 {)ta
--+-+-=o nr >:
D (17.25)
(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)
is left invariant under gauge transformation (17.23) provided that we as-

sume A and cp undergo transformations simultaneously given in Eq. (17.4).
This is because
º"A iJ! = e'c1,.
Do iJ! -+ Doe'c1,. º"A Do iJ! L
DiJ! -+ eíf,¡ADiJ!
D2iJ!-+ eíf,¡AD. DiJ! = eíf,¡AD2iJ!.
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
1(r) = nee ¡r A(r') · di', (17.28)

364 Quantum M echanícs
is a solution of Eq. (17.26) This can be seen by noting that
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
- (ñ2) v2wº + vwº - in awº. (17.30)

2m at
Now iJ.!(r, t) can be regarded as a wave function of a particle that goes
from one place to another along a certain route where there is a field A
present. iJ.!º(r, t) is the wave function for the same particle going along
the same route but with A = 0.The solution (17.27) has sorne striking
physical consequences as shown in the two slit electron interferometry ex-
periment proposed by Aaharonov and Bohm [ ig J. In this experiment
..
y
~ (1)
-:,.-,-:------.-----~---J- p'
·~o . -- ---
*
p _,,.... d
\
- ~ '' j,
\.D.
.,, - J
_y
(2"
~ )
Fig. 17.1 Two paths enclosing a non-zero B field.
the magnetic field B (pointing in a horizontal direction out of the paper)

is produced by a long solenoid of small cross-section and is confined to the
interior of the solenoid so that the two electron beams (1) and (2) cango
above and below the B -=f. O region but stay within the B = O region and
finally meet in the interference region P'. In the interference region, the
wave function of the electron is
(17.31)
so that
(17.32a)
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
a= e i A(r') · dl' =e is B · da = ;e . (l 7.33b)
In Eq. (17.33b) we have used Stokes theorem and put B = V x A and 
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
t:J..y = dLA [8(B # O) - 8(B = O)]

2rr
=
e
ne
(Ld A )
2rr
(17.35)
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.
Measurement of AharanovJ13homeffect not only verifies the gauge

principle in electromagnetism but also quantum mechanics itself
since classically the dynamical behavior of electrons is controlled by
Lorentz force which is zero when the electrons go through magnetic
field free region; yet in quantum mechanics observable effects are
seen and depend on the magnetic field in a region inaccessible to
the electrons.
(ii) The vector potential A rather than the fields plays a crucial role
as the basic dynamical variable in quantum mechanics.
(iii) By varying B (and henc <l>) we chango the relation phas between
the contributions from the 2 paths and moveÍjiterference pattern
up and clown. When ,~/1? = 2mr or <l> = n<P<t[to = 21r/ie/e =
4.135 x 10-1 gaus cm2) the pattern will r turn to it initial form
as if there were no field. In other words, an integral multiple of the
flux quanta </Jo = 2; ñc will not make any observable difference to
the quantum mechanics of the particle.
The quantization of the magnetic flux [cf. remark (iii) above]

<l> = 21rñc n
e '
n = O, ±1, ±2rt'\
was first predicted by F. London. This was confirmed experimentally in
(17.36)
ó
superconductors. In a superconductor, due to interactions of the electrons
with the vibrations of the atoms in the lattice, there is a small net effective
attraction between the electrons. As a result of this attraction, the electrons
form bound pairs; such a correlated pair of electrons behave like a quasi
Bose particle of charge 2e. When a superconducting ring is placed in a
magnetic field B, inside the material of the ring there is no magnetic field
( the absence of the magnetic field in a superconductor is called the Meissner
effect). The flux enclosed by the superconducting ring is quantised and
instead of Eq. (17.36), we have
<l> = 21rfic n n = O, ±1, ±2, · · · . (17.37)
2e '
Thus we see that the magnetic flux trapped by a superconducting ring

should be quantised.
17.3 Landau Levels
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
'v · A' = 'v · A + 'v2 A = O

or
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)
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)
This is an eigenvalue equation for a simple harmonic oscillator. The

energy eigenvalues are given by
p;
= (n + -)ñw
E - -
1
(17.44)
2m 2 '
For the motion in x -y plane, we can put Pz = O. The energy levels labeled
by n are called Landau levels. Now a charged particle of charge e in a
magnetic field moves in a Larmer circle of radius r
mv2 evB
(17.45)
r e
Classically radius is not fixed. But if we quantize the angular momentum,
then
mvr = n*h = n*(21rn) (17.46)
so that
eB 2 *t-
-r = 21rn u: (17.47)
e
Thus a quantum electron takes up an area A of order
A= 7fT2 = 21rn*n. (17.48)
eBc
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.
17.4 Quantization of Radiation Field
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
\72A - __!__ a2 A = O. (17.51)

c2 8t2
The plane wave solution of Eq. (17.51) can be written as
c.>..(k, t)éik·r, (17.52)
where A specifies the polarisation. ,\ = 1, 2 for linear polarisation and
A = 1 ± i2 for circular polarisation.
Substituting Eq. (17.37) in Eq. (17.36), we get
c.>..(k,t) = -c2k2C.>..(k,t). (17.53)
With w = clk], we get the solution
c.>..(k, t) = ct(k)e-iwt + C~(k)e+iwt_ (17.54)
Hence
A(r, t) = éik·r ( ct(k)e-iwt + C~(k)e+iwt) . (17.55)
Since A(r, t) is a real field, we can write the plane wave solution of
Eq. (17.51) as
(17.56)
Let us write
(17.57)
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
-Vl eik·r e -ik' ·rd3 r r

= ukk' J._ (17.60) )
,
~f eik·r eik' ·r d3r = 8k,-k' l (17.61)

)
k· EA = k· EA• = Ü. (17.62)
We now calculate the electric and magnetic fields in terms of CA (k) and
c¡(k).
xe-ik·r ( +iw)e+iwt). (17.63)

B = curlA
= .i: L ¿(k X EA)CA(k)eik·re-iwt - (k X EA*)c¡(k)
Jv k A
(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)
PERTURBATJON INDUCED BY ELECTRO MAGNETIC FIELD 371
where we have used Eqs. (17.57)-(17.60) and the fact that

A .X'*
(k X E ) • (k
I
X E
,X'*
) = k · k I EA • E
,X'*
- k· E •
I
k ·
A
E • (17.66)
If we average over time so that oscillating terms drop out, we have
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)
Comparing Eq. (17.70) with the Hamiltonian of an oscillator 2~ (p2+w2q2),

we see that Hamiltonian for radiation field behaves like a collection of os-
cillators [cf. Sec. 9.5].
The comparison with Sec. 9.5 suggests that, in order to quantise the radi-
ation field, we replace
C.x(k) - .!fc-/Viwa.x(k)l.- (17.71) )

C;:(k) - 1;c-/Viwal(k), (17.72)
so that
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), al, (k) J = 8kk'8.xA'. (17.74)
In analogy with Eqs, (9.120a,b) of Sec. 9.5, we have
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)
where
ln.\(k)) = ~ 1
n.\(k)!
( a1(k) r" IO). (17.79)
Now using the commutation relation (17.74), we can write
(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)):
HJn.\(k)) = Iu» ( n.\(k) + ~) Jn.\(k)). (17.82)
It is thus clear that we have the following interpretation. The state
al (k) IO) (17.83)

represents one photon with polarisation >. and wave number k. The state
( 17.79) represents n photons each with polarisation >. and wave number k.
1
The operators a (k) and a.\ (k) act as creation and annihilation operators
respectively. We note further that
(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
(n.\(k) + lla1(k)ln.\(k)) = Jn.\(k) + 1L (17.85a) )

(n.\(k) - lJa.\(k)ln.\(k)) = Jn.\(k). (17.85b)
Finally Eq. (17.58) in terms of creation and annihilation operators is given

by
A(r t) = fww L ¿-1-(€.\a.\(k)e-ik·re-iwt

/V k x J2w
+€.\* al(k)e-ikreiwt). (17.86)
17.5 Perturbation Induced by Electromagnetic Field
Let the Hamiltonian of the atomic electron ( charge e < O) in a static

potential V ( r) be
A2
p
Ho =
2m
+ V(r). (17.87)
Then the Hamiltonian of the atomic electron in an electromagnetic field

A(r,t) (pure radiation field) is given by (see Eq. (17.17))
A2 2
p e e 2
H = -+
2m
V(r)- -A·p+--A
me
A
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
>.V(t) = Hint = _ _!:._A· p. (17.89)

me
Substituting Eq. (17.86) in Eq. (17.89), we have
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)
Using Eq. (17.77), we have

(n>.(k) - lla1,(k) = Ó>.>.'Ókk'Vn>.(k) - l(n>.(k) - 21
a>.'(k')ln>.(k)) = Ó>.>.1Ókk' /n>.(k)ln>.(k) -1),
so that Eq. (17.91) gives
e 4nlin>.(k)(bl e ik-r€ x · pAl)
na(t) = -- ----- a e -iwt . (17.92)
m V 2w
These are the matrix elements for the absorption of a single photon of
frequency w by a charged particle from an initial state that already has n
photons of frequency w and polarisation >..
Similarly, using (n>.(k) + lla1,(k') = Ó>.>.'Ókk'Vn>.(k) + l(n>.(k)I, etc,
Tba(t) = ~(b, n>.(k) + llA · f>la, n>.(k))

m
e
= -·
4n/i(n>.(k)+l)(bl e -ik·r € >.* · pAl)
a eiwt , (17.93)
m V2w
gives the matrix elements for emission of a single photon of frequency w by
a charged particle into a final state that has n + 1 photons, that is, from
an initial state with n photons of frequency w.
Thus we see that the vector potential A(r, t) acts as an effective potential
2nc2ñn>.(k) x i(k·r-wt)
Vw € e (17.94a)
for absorption of a photon of frequency w and for emission of a photon it

acts as
2nc2ñ(n>.(k) + 1) >.* -i(k·r-wt)
Vw € e · (17.94b)
17.6 Stimulated Emission and Absorption of Radiation
The transition probability from the state a to state b is given by

ICb(oo)l2,
where [see Eq. (13.12)]
(17.95a)
Using the formula
1-: ei(Eb-Ea~liw)t/lídt = 21f'M(Eb _ E =f ñw), (17.95b)

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 >.:
c:bsorption(oo) = ::ñ n>-~~1rn21rM(Eb - Ea - nw)
x (bleik·r €>- · f>la) L (l 7.96a) _)
cbemission(oo) = ::n (n.>-(k~~ 1)21rñ21rM(Eb - Ea+ nw)

x(ble-ik-r€.>-• · :f>la). (17.96b)
The transition probability per unit time from the state la) to the state lb)
with the absorption of a photon is given by
wb;s = ~ 1 qibs ( 00 )12
= _!!__ n.>-(k) 21rñ[21rM(Eb - Ea - nw)]2
m2ñ2 wV t
x 1 (bleik·r€>- · f>la)l2. (l 7.97a)
Now
1
[8(Eb - Ea =f nw)]2 = ,i,8(0)8(Eb - Ea =f nw)
. t
= hm tc.8(Eb - Ea =f nw).
t-too 2 7rn
Thus
Wbbs = [n.>-(k)cj47r2e2l(bleik-r€A. f>la)l2ó(Eb - Ea - nw). (17.97b)
ª V m2wc
Now n.,.~k)c is the incident proton flux so that absorption cross section is
given by
ag! (w)
8 = Wbªd'ª/(Flux)in
Similarly
w:;1isijiQII
41r2e2
= -2-l(bleik·r€.>- · f>la)l28(Eb - Ea - nw).
m wc
= [n>..(~ + 1] 47í:e2 (ble-•k·r€A

(17.97c)
e
.. f>la}l26(Eb - Ea+ Iu» 17.97d)
(jJ
mw
S In actual physical situatioLthe photon momentum fik does not
uniqu ly specify the photon state. The photon i in general · rnitted or
absorbed in sorne momentum interval nk and ñ(k + dk). The number of
~. \ates in ;2::;lki:gi::h>P}r ( :h)' (v' :; )

this 2 dEdll
~ ~ p(R)dEg{,(w)hd,.,, (17.98a)
where
_ V ( 2 dp) V 2 (17.98b)
p(E) = p(w) = (21rfi)3 p dE df2 = (21rc)3fiw df2
is called the number of states per unit energy interval.

Hence using Eqs. (17.97) and (17.98), the total transition probability
per unit time from an atomic state a to a state b with the emission of a
photon is given by
w emission
ba
= 47r2e2
m2
J (nA(k)
wV
+ 1) V w2dwdn
(21rc)3
X 8(Eb - Ea + nw) 1 (ble-ikr €A •. f>la) 12
2
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
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
= S;li (:f (eñw/k;, _ 1) dw. (17.108)
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)
This is Planck's law.

We note from Eqs. (17.99) and (17.100), that in the presence of a large
number of photons of a given wave length (n.>s(k) large) transition rates
(stimulated emission of light) corresponding to that wave length will be
enormously enhanced. Thus if many atoms can be raised to a given ex-
cited state in the environment of the right kind of photons, then they will
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.
17.7 Spontaneous Emission of Radiation:
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)
If [k · r] ""' kr is small, i.e. if
[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
PERTURBATJON INDUCED BY ELECTRO MAGNETIC FIELD 379
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)
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].
r · €*(+) = -r · i-l = _..!_ sinOe-irf> = - {4;rrY1 -1

v'2 V3 ·
r · €*(-) = -r · €(+) = ..!_ sin()eirf> = - {4;rrY11 (17.122a)
J2 V3 ·
for circularly polarized light, €(±) = =f«P>~f2>; for plane polarized light,
O
we take € along z-axis and k in x - y plane and
r. ,,· ~ reos O~ f¡rYw l (!7.122b)
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.
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)
= J drr2 Rn111 (r)R,,,ii.(r) J dnY¡;m¡ ((), rp)Y1m((), rp)YtimJ(), rp).

(17.127)
Now from Eq. (11.165), the integral involving spherical harmonics is pro-
portional to the Clebsch(¡Gordon coeffi.cient(llimmill¡m¡lli) which van-
ishes unless
l~li ± 1
m¡Y-m+mi (17.128)
i.e. the selection rules for El transition are
Á)¡-li = ±1
D.m~m=±l,O, (17.129)
and parity must change.
Sometimes, the selection rules do not allow El transition between cer-
tain states. This occurs when (birla) = O, between states a and b. In this
case, we have to take the term -ik · r in Eq. (17.112) seriously, so that,
transition matrix elements for the radiative transition are given by
(bl(k · r)(t:.x• · :f>)la).
We can write
(k · r)(t:.x• · :f>) = (bl~ ((k · r)(t:.x• · p) + (t:.x· · r)(k · :f>))
+ ~ ( (k · r) (E.X· · :f>) - (E.X· · r) (k · :f>)) 1 a) .

(17.130)
The second term can be written as (since k · é• = O) follows
~(bl(k · r)(1o>.* · p) - (1o>.• · r)(k · í'>)la) = (bl(k x é) · (r x í'>)la)

1
= X€>.)· Lla),
2(bl(k
(17.131)
where Lis the orbital angular momentum operator. Now µL = 2,':,cL, gives
the orbital magnetic moment, hence the radiative transition due to this
term is called a magnetic dipole (Ml) transition. From Eq. (17.115), noting
that kr is of order Za we see that the matrix elements for Ml transitions
are suppressed by a factor of Za as compared with El transitions. The
selection rules for a1
Ml transitíon are t::,,.J = ±1, O, no change in parity
of states, For circJarly polarised light this can be een as follows for a
hydrogenfüke atom.
For Ml transition (see Eq. (17.131)), for circularly polarised light, we
have the matrix elements of the form
But
since 1oC1), 1oC2) and ~ form an orthogonal coordinate system. Also we have
(k X€±)· L = ~lkl(Lx ± iLy)
=i~L±.
Thus
(n¡l¡m¡l(k x €(±)) · Llnilimi)

= i ~(n¡l¡m¡IL ± lnilimi)"' (n¡l¡m¡lnilimi ± 1).
These matrix elements are non-zero only for
6.l ~lt - tl\o

t::.m(j =¡ - mk',)±1
giving us the required selection rules. Further since (k x E-*) is invariant

under parity, then parity conservation gives (-1)1i = (-l)1t or (-l)A1 = 1
i.e. no parity change.
The first term in Eq. (17.130) can be written as
k · ( rp + pr) · E>...
Thus we can write
(blk · (r:f> + :f>r) · E>..la) = ~ (blk · {r[r, Ha]+ [r, Ho]r} · E>..la) + O(e)
= ~ (blk · [rr, Ha]· E>..la)

= -imwk · (birria) · E>.._ (17.132)
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.
17.8 Decay Width
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)
The sum over two polarisation states gives, using Eq. (17.121e)
¿ l(blrla) · €A*l2 = ¿(blrila)(blrjla)*€( €]

A A
= (blrila)(bhla)*(óij - k~:j)
12
= 1 (birla) 12 - lk. (~;la)
If we take k along the z-axis, this gives

L !(birla)· €A· ¡2 = l(blrla)l (1 2 - cos2 fh). (17.136)
A
Integrating over d0., we have from Eq. (17.135),
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)
- This rate depends on E~ which i typical of El emission. To get an order

of magnitude for r, for bydJ:ogenlJike atom (ao is Bohr radius to 6::r-
-(blria)I "'ao,E7 ~ e:
2 0--+ ! (~!) ,giving
T = -ñ. ~ -
6 4-
~ 3.5 x 10 -10 sec, ªº
a e r
which gives time of a typical spontaneous El atomic transition.
-
¡\
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
1 (r) = (100lrl21m) = (1 00 R10(r)rR21(r)r2dr)
X
r /2,r
lo le Y0ii
(sin(} cos </J)
sin sin sin() = ~v'ªf-Yi,±1,
Thus
and we have from Eq. (17.138)
r,,_,. - 9~~,E~(¡=
Rw(r)R,,(r)r'dr)2~) L (17.139a) )-1,
where
R10(r) = 2 (!) 3/2 e-zr/ao L (17.139b) )
R (r) = _1_ (~) 5/2 re-zr/2ao (17.139c)

21 v24 ªº ,
and ao is the Bohr radius
ñ,2 ñ
ao=-=--. (17.139d)
me2 mea
Now
Hence
(17.141)
Now
(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)
':::::' 1.6 x 10-9 sec. (17.145)

Finally, we note that if we multiply Eq. (17.119) for the decay rate, by
the photon frequency w, we get a formula for the intensity of radiation
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.
17.8.2 Magnetic dipole transition due to intrinsic spin:

Here all one has to do is to replace the orbital magnetic moment µL = 2!cL
!
by th intrinsic magnetic moment of spin particle of charge e, namely [see
_;erl 11.5 and 20.12]
~ ¡_::_. e
µs = 9s(2m)s
where
s=
n
-(T
2
and g8 = 2 for an elementary particle =c= Then the right hand
L
side of Eq. (17.131) is replaced by

g; ñ~(blu · k x 1:*>-la) (17.147)
o
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
>. >.
= ""'(k X€*>.)· (k X€>.)

~
x
= 2k2 =
w2
2-
c2 lO
Hence E'Y = nw
We now apply the formula (17.1491)\ to the radiative transition from a Q

(17.149) o
baryon Z::º to Aº: /
t- L Aº7
The transition magnetic moment [mp is the proton mass used in nucleon
(17.150)0
magneton]
o
µ(¿-Aº)=
Hence from Eq. (17.150):

gs (~),
2 2mpc
g2s 1.61 (17.151) = L o
Here E'Y = 74MeV so that Eq. (~ gives
L (17.152)
o
r(¿-Aºry)
® 17.9
T=
n
-
r
.5 X 10-20s
Dipole Sum Rules, Two Point Function

(Exp: 7.4 ± 0.07 x 10-20s)
Consider a particle in quantum mechanics described by dynamical variables

(17.153)
x and p. We want to discuss this dynamical system without assuming any

form for the Hamiltonian and for the ground state as far as possible. This
method is particularly suitable for the systems in which the detailed form
of the Hamiltonian or dynamics is not known.
Let us start with the basic commutation relation of quantum mechanicsr( )
[pi, xJ] = -iMiJ· (17.154)
388 Qua.ntum M echonics
In the Heisenberg picture . . L

Xi(t) = eiHtflixie-iHt//i (l 7.155a) )
J\(t) = eiHt/lipie-iHt/li. (l 7.155b)
Define the two point function
(Olxi(t)xj(t')IO) = J d3ru~(r)xi(t)xj(t')uo(r), (17.156)
where uo(r) is the ground state wave function.

We now want to obtain a "spectral representation" for this function (see
below). For this purpose we use the following steps:
l. Using the completeness
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
2. Using the translational invariance viz. Eq. (17.155), we have

(nlxi(t)IO) = (nleiHt/lixie-iHt/lílO)
= e-i(Eo-En)t/lí(nlxilü). (17.159)
3. Using Eq. (17.159), we have
(Olxi (t )xj (t') IO) = L ei(Eo-En)(t-t')l!i(Olxi In) (nlxj IO)
n
= f 00 dEe-iE(t-t')/ñ ¿(O/xi/n)(n/xi/0)
Ío n
xó(E - En+ Ea). (17.160)

Let us define
(17.161)
n
Thus we get the following representation of the two point function
(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.
4. The spectral function Pii(E) has the following propertíes: ( C's are "
complex numbers) ....,
c;Pijcj = L(o¡c;xiln)(nlxjCjlO)ó(E - En+ Ea)
(jo, t l(nlx,C,IO)l'O(E -En+ Eo)

(17.163)
for arbitrary constants C¿ · Pii is a positive matrix
P11,P22,p33 > O.
5. Similarly, we have
(011\(t)_pi(t')IO) = fo00 dEe-iE(t-t')/ñPii(E). (17.164)
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
= ~eiHtfñ[H, xi]e-iHt/ñ. (17.165)

Let us now assume that the Hamiltonian H has the form
'2
H =R._+ V(r), (17.166)
2m
so that
(17.167)
Hence
!!:_X. (t) = ]_eiHtf ñp'-e-iHt/ñ
dt ' m '
= ]__)3i(t). (17.168)
m
Using Eq. (17.168), we obtain
(Olfii(t).Pi(t')IO) = m2(01!xi(t) d~'xi(t')IO)
= ;22
Jo
t" dEe-iE(t-t')fñE2Pii(E). (17.169)
Comparing Eq. (17.169) with Eq. (17.164), we get

2
Pij(E) = ;2 E2Pij(E). (17.170)
6. We also note from Eq. (17.168) and Eq. (17.169)
i: fo l )
o
(Olfii(t)i;j(t')IO) = - 00 dEe-iE(t-t')/ñEp;j(E)
(Olxi(t)jjj(t')IO) = i: fo 00 dEe-iE(t-t')lñEPij(E) LJ:2.1na)

Changing t - t', i - i. we have
(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
i: laoo dEe-iE(t-t')/ñ Epj;(E) = i: 1a= dEe-iE(t-t')/ñ Ep;j (E).
(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)
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)
This is known as the spectral function sum rule.

Two additional sum rules are obtained from Eqs. (17.162) and (17.164),
by putting t = t', i = j and using Eqs. (17.170) and (17.178)
(r2)0 = 31= dEp(E) L (17.179b) /
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
(!, g) = J dEp(E)f(E)g(E) (17.180a)
in a certain function space. The Schwartz inequality gives
l(f,g)I < llfll llgll, (17.180b)

where
11!11 = vV,7). (17.180c)

Choose f(E) = 1, g(E) = E. Then we have
/00 dEElp(E)l2::::; (1 00 dEp(E)) (1 00 dEE2p(E)). (17.181)
Now using Eqs. (17.179), we have

ñ4 1 ñ2
4m2 < 3 (r2) 3m2 (p2) (l 7.182a)
or
(r2) (p2) :2: ~ ñ2, (17.182b)
which is the uncertainty relation.
17.10 Application of Dipole Sum Rules
Consider the reaction of the type

'Y+ IO) -+ In), (17.183)
where IO) is the ground state and In) denotes all possible final states con-
sistent with the conservation laws. Then to the lowest order in e, the
transition probability is given by
ICn(oo)i2,
where (cf. Eq. (17.96a) with n>.(k) = 1)
en ( 00 ) = ie
- mñ y¡-;¡;¡¡:-
M21r1w"'( En - Ea "· ·) ( neI ik·r p · ex I O.)
-1WJ
A
(17.184)
Using the dipole approximation, the total transition rate is given by

(average over initial polarisation)
W = ICn(oo)l2 = _!C_ 41rñ 27rn (~ ~ €.>.€.>.)

t m2ñ2 2wV 2 ~ • J
>.=1
X L ó(En - Ea - mu) (Olf>i In) (nlf>j lü). (17.185)
n
Now
2
~ L €t€; = ~Óij, (l 7.186a)

>.=l
Pi1(w) = ¿(Ol1\ln)(nlfi1I0)8(En - Ea - mu)
n
= 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.
17.11 Dispersion Relation
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)
Let us consider its analytical continuation for complex (t - t'). Put

t -t' = u+ iv. (17.197)
Then right-hand side of Eq. (17.197) becomes
Óijñ 1= dwe-iwu+wv p(w).
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),
Gi;(w) = .L'º dte-iw(t-t'l(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)
Interchanging the order of integration ( this is justified if the integrals exist),

we get
= m2Mij (fo00 ¡,00

r
Gij(w) dw'w'2 p(w') dte-i(w'+w)(t-t')
o 1
+ 00 dw' w'2 p(w')
-oo
dte-i(w-w')(t-t')). (17.202)
In order to give meaning to time integration, we use the following prescrip-

tion:
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
Now using Eq. (17.190b), we have

. 2 100 dw '
O
2w'3p(w')
W2 -W1
2 .
+ Z€
L (17.205) o
(17.206)
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)
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
17.1 Let B be a uniform magnetic field. Verify that A= !B x r satisfies

the relation B = V x A and the gauge condition V · A = O. Take
=
B (O, O, B) and show that A= !(-By, Bx, O).
17.2 Consider the motion of a particle of charge e and mass m in x - y
plane with a uniform magnetic field along z-axis, B = V x A,
A= !B x r.
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
2(1/JIHl7/J) = J [- .!!!_(1/J*v'21/J + 1/Jv'21/J*)

2m
+ (~~) (1/J*(A · v')?jJ-1/J(A · v')?jJ*)

e2
+-
me
2A2 + 2ecp]d3x.
Using the relation

V· (V'1/;*'!f;) = '!f;V2'!f;* + '!f;*V2'!f; + 2V'!f;* · V'!f;,
and the fact that by Gauss's theorem, the left hand can be written
as surface integral for V('!f;*'!f;), which vanishes for a large surface,
one gets
('!f;JHJ'!f;) = j d x[:: V'!f;*V'!f; + ;:::)'I/J*(A

3 · V)'!f; - '!f;(A · V)'!f;*)
2
+
_e_'I/J*1j;A2 +e1/J 1/J<f>].
2mc2
Hence the Hamiltonian density L 1
'~
'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.
17.6 Photoelectric Effect: Consider the ejection of an electron from the

ground state of hydrogen like atom by absorption of a single photon
viz. the reaction
Photon + Atom = e + Atom'.
Calculate the cross-section ~~ for this process by neglecting the
recoil of atom.
Hint: The number of states when the electron has momentum
between Pe and Pe + dpe = c2,Yn_)3 d3pe = (2;11,)3 2mpedEedf!,
E¿ = Pe/2m. The ground state of hydrogen like atom is repre-
sented by the state function
(ria) = (rllOO) = U10o(r)
= ~(Z/ao)3f2e-zr/ao
7í
and for free electron the wave function is given by
(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
= _l_E>.. p _l_(z/a )3/2

\!Y e ,/ir O
J d3re-i(k-pt )·r e-zr/ao.
.
The flux of incoming photon is -o-e and ~~ = Fl~x ~· Note Ea i

the binding energy of the electron in the ground state. In evaluating
( E>- · Pe) 2, sum over polarisations; for this purpose it is convenient
to select as unit vectors E1, E2 and
The cross-section can be expressed as
1t
and take k along Z-axis.
1
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.
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
41rnc 2 ( ó·n·n , - --a a ) D1R

•3 aXi aXj' '
where
D (R) = _1_ / d3k ik·(r-r') R = r - r'

1 (21r)3 2w e '
1 roo
= 41r2JRlc lo dksinkJRI.
By defining the last integral as
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,
where ó. V is a small volurne containing the position in question.

Then show that
(DIE· EJO) "' nc/(ó.P)4, (I)
where ó.P is the linear dirnension of volume ó. V. The above ex-

pression characterizes the fluctuations in the electric field when
no photons are present. Compare it with the square of the field
strength for a classical electrornagnetic wave of wavelength 21r A
where the time average of E2 can be equated with the energy den-
sity of electromagnetic wave so that
(E2)average = nn(c/A) , (II)

where n stands for the number of photons per unit volume. Thus,
for the validity of classical description, purely quantum effects such
as (I) (with D..f.,...., >.) must be completely negligible in comparison

with (II). For this, we must have
1
ñ » ).3'
i.e. the description of physical phenomenon based on classical elec-

trodynamics is reliable when number of photons per volume >.3 is
» l. Show that this condition is satisfied by number of photons
per volume >.3 ata distance of 10 km from an antenna with a power
of 135, 000 Watts giving waves of >. ~ 50 cm.
17.7 For quantised radiation field, show that
[Ei(r, t), Bj(r', t)] = ic(41r)ñEijk 8(r - r').

88Xk
What does this say about the simultaneous measureability of B
and E?

Chapter 18
FORMAL THEORY OF
SCATTERING
18.1 Introduction
We have discussed scattering theory befare, but we restricted ourselves to

potential scattering involving elastic collisions of two simple structureless
particles. We used coordinate representation explicitly in our discussion,
whereas in sorne problems, other representations may be more advanta-
geous.
We now discuss the general but formal theory of scattering because it
has the followingadvantages:
(i) It is not restricted to any particular representation.
(ii) It can be easily generalised to complex systems such as nuclear
reactions where there are many channels, elastic as well as inelastic.
(iii) Symmetry properties of the system can be easily incorporated in
this formalism and the selection rules for the scattering process can
be easily derived.
It is useful and natural to look at a scattering process as a transition
from one unperturbed state (initial state) to another (final state). The
scattering process corresponds to the situation where the energy spectrum
is continuous and the perturbation is such that
V=O for t «O and t »O
and independent of time in between.
If the scattering region is of finite extent, the initial and final states can
be taken, to a good degree of approximation, as plane wave eigenstates of
definite momentum of the unperturbed Hamiltonian
Ha= p2/2µ
403
(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
Holn) = Enln) (18.2)
are known.
The Schrodinger equation is
iñ ! 17/J(t)) = Hl'l/;(t)). (18.3)
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.
l'l/J(t)) = L Cn(t)e-iEnt/ñln). (18.4)
' Substituting E~ in Eq~) and using the orthonormality of eigen-

vectors
(bln) = Óbn,
we get
(18.5)
where
Vi,n = (b!Vln) L (18.6) )

(/J,¡Jbn = Eb - En. (18. 7)
Now for t « O, V= O so that our system is in an eigenstate of H0
l'l/J(t «O))= la, t)

= la)e-iEat/ñ. (18.8)
FORMAL THEORY OF SCATTERING 405
( )<: ).....
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
Cb(t) =s]: dt'eiwbat'v¡,a(t') + s: (18.lOa)
with the initial condition of Eq. ~

If V is independent of time as we shall assume, then
Cb(t) =-*Vi,ª ¡_too eiwbat' dt' + s-: (18.lOb)
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¡
18.2 Interaction Picture
To get a general solution of Eq. (21.3) or Eq. (21.5), it i e nvenient to go

to the interaction picture. In the br6dinge1:-'picture, the stat v ctor i
time dependent:
iñ :t \'l/J(t))s = H\'l/J(t))s (18.12)
H=Ho+V,
where Ha is independent of time, but V may in general depend on time,

although most of the time we shall take it to be independent of time.
We define a state vector
11/J(t))¡ = eiHot/1il1/J(t))s L (18.13)
iñ,!!:_¡1/,,(t))¡ = -HoeiHot/1il1/J(t))s + eiHot/fiifi.!!:_11/J(t))s
dt dt
= -H0J1f,,(t))¡ + eiHot/1iHe-iHot/1iJ1/,,(t))¡. (18.14)
Now define
HJ (t) = eiHot/fi Hae-iH0t/1i
=Ho (18.15a)
and
H¡(t) = eiHot/fiHe-iHot/fi
=Ha+ Vj(t), (18.15b)

where
Vj(t) = iHot/five-iHot/fi. (18.15c)
in !11/J(t))1 = Vj(t)11/J(t))¡. (18.16)
We see that Vj(t) is the effective Hamiltonian for the state vector J1/;(t))¡.
This is the reason it is called an interaction picture. Since the transforma-
tion is effected by a unitary operator eiHotfñ, J,,¡,(t))¡ describes the same
quantum state 11/J(t)) s but in a different picture. The unitary transforma-
tions Eq. @_.13) and Eq. (J_§.15) define the ínteraction picture.
An arbitrary operator A in the Schroding r picture is transformed into
Á.¡(t) = eiHot/fi A.e-iHot/fi J_
(l8.l ?a) )
ifidÁ.;/t) = [Á.1(t), Ha]. (18.17b)

To summarise, we compare the three pictures
Schrodínger picture Interaction picture Heisenberg picture
inftl1P(t))s = Hl1/J(t))s inftl1P(t))1 inftl1P(t))H = o
in~1 =o = V¡(t)l1/J(t))1 11/J(t))H = 11/J(O))s
A.= Á(O) Á.¡(t) = eiHot/fiÁ_e-iHot/fi Á.H(t) = eiHt/fiÁ_e-iHt/ñ
dA A dA A
ifidtA1(t) = [A1(t), Ha] ifidtAH(t) = [AH(t), H]
H = Ha + V H¡(t) = Ha + V¡(t) HH = H.
The interaction picture is intermediate between the Schrodinger and
Heisenberg pictures. Here both the state vectors and operators change
with time but the time development is governed by different portions of the
Hamiltonian, the state vectors by the interaction V and the operators by
Ha.
18.3 Formal Solution of Schródínger Equation in the In-

teraction Picture
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
U(to, to) = 1, (18.19)

by definition. Substituting Eq. (21.18) in Eq. (21.16), we have
iñau~; to) 11/J(to))r = Vr(t)U(t, to)l1P(to))r,

-
which implies that
.. 8U(t,to) () ( ) (18.20)
ih at = V¡ t u t, to .
The important point is that the operator U(t, t0) depends only on the struc-
ture of the physical system and not on the particular choice of initial state
11/J(to))r.
We see that
11/J(t))r = U(t, to)l1P(to))r

= U(t, t')l7/J(t'))r
= U(t, t')U(t', to)l7/J(to))r.
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,
1 = U(to, to) = U(to, t)U(t, to),

Hence
U(to, t) = tr+«, t0). (18.21b)
Sometimes it is convenient to replace the differential equation by an
integral equation. Thus integrating Eq. ~1 t0 to t and making
use of the condition (16.19), we have
----- . t
U(t, t0) f
= 1- ~
t; V¡(t')U(t', t )dt'.0 (18.22)
Substituting back, the expression U(t, t0) from Eq. (21.22), we have an
iteration procedure:
U(t, to)= 1- * i;r dt1 Vi(ti)[l - ~

i;t: V¡(t2)U(t2, to)dt2]
= 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
U(oo, -oo) = 1 - ~ ¡_: dt1 V¡(t1) + ( ~i)2 ¡_: ¡_t: dt1

We can write it in more compact form. By interchanging t1 --+ ta, we
dt2Vi(ti)V¡(t2) +··· . -.
can write the
second term as (,/)2 J~00 dt2 t~
dt1 V¡(t2)Vi(t1). It can be shown this
t
term is equally well written as ( hbdr)2 J~00 dti 100 dt2 V¡(t2)Vi(ti). Using
this identity we can write the second term as....__ _
Chi)2"ti f~00 dt1 f~00 dt2T(V¡(ti)V¡(t2)), where the time ordered product is
defined as
T(V¡(t1)Vi(t2)) = Vr(t1)Vi(t2)
= V¡(t2)Vr(t1)
Thus
U(oo, -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
If the experimental arrangement is such that the system is known to be

in an eigenstate la) of Ho at t = to - -oo, then the transition amplitude
to sorne eigenstate lb) of H0 is given by
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
Cb(t) = lim (blU(t, to)la) = (blU(t, -oo)la). (18.28)

to--+-oo
----
From Eq. (21.23), we see that to first order in perturbation
U(t, -oo) = 1 - ~ -oo

. ¡t dt'V¡(t')dt'.
Therefore
= Óab - i (bl
n
r dt' eiHot' /nv e-iHot' /ñja)
= Óab - ~ t. -oo
dt' eÍWba t' v¡,a ( t') l

,
which is a re-derivation of Eq. (21.11) if V is independent of time.

We shall now assume that Vis1naependent of time. We note that the
formal solution of Eq. (21.3) or (21.12) is
-
l'I/J(t)) = e-iH(t-to)/ñl'I/J(to)) (18.29)
or
(18.30a)
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, t)) (18.32a)

or
ut(t,to)U(t,to) = U(to,t)U(t,to)
= 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
tE~= f (t) = ~~ r¡ j_ºoc e11t' f (t')dt' (18.33a)
and
lim J(t)
t-+=
= lim r¡
11-0
r=
}0
e-r¡t' f(t')dt'. (18.33b)
FORMAL THEORY OF SCATTERING 411 (\\\
If f(t) is a smooth function having well defined limits J(+oo) as t - ~

+oo, then the above are identities as can be easily seen by integrating by
parts:
lim J(t) = lim (r¡er¡t' f(t')I~= - {º er¡t' J'(t')dt')

t---+-0() r¡---+O 1) } _O()
= 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)
lim eiwt = lim r¡ {º eiwt' er¡t' dt'

t---+-0() r¡---+O } _O()
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 _=
--'
{'{\
18.4
We have
Scattering States ~ The S-Matrix
Thus
U(O, -()() )la) = lim 1J

r¡-+O
jª
-oo
el71+(i/ñ)(H-Ea)]t' dt'la)
.
= 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)
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)
= lim -Íé . lb). (18.40)

e:->0 Eb - H - Zé
We denote U(O, oo)lb) by lbC-l), so that
lb)out =lb(-))= U(O,oo)lb)
Íé lb). (18.41)
Eb -H =ie
Equivalently
(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)
414 Q-uant-um M echanics
Now using the group property of U,

Sba = (blU(oo, -oo)la)
(blU(oo, O)U(O, -oo)la).
Now
U(O, -oo)la) = laC+l)
lbC-l) = U(O, oo)lb),
so that
(bC-ll = (blUt(o,oo)
= (blU(oo, O).
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)
(blaC+l) = (bla) + (bl l . VlaC+l). (18.45a)

Ea =H¿ +zc
But
(bl(Ea - Ho + ic) = (bl(Ea - Eb + ic).
Operating on both sides by E
a-
¡¡o +.ic , we have
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
(bC-)1 = (bl + (blV l . ,

Eb -H +ic
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
= (bla)+ (Ea-!b+ie: + Eb-!a+ie:) (blVlaC+))
- . - 2ic ( +))
- Óba + e:-->O
hm (E b - E a )2 + e2
1
bV a
1 (
= Óba - 21rió(Eb - Ea)(blVla(+)). (18.46)

We define an operator T with matrix elements
Ti¿ = (bJTla) = -(blVla(+)), (18.47)
so that
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)
Furthermore, in order to obtain perturbative approximation for we na,

need an equation for Tba in terms of Via· This can be easily obtained as
follows:
na = -(bfVla+)
= - ¿ (bfVlc) (cla+)
e
where we have used Eq. (18.45a). Hence

~ VbcTca
na = -Via + L.,e E _ E + u::. . (18.50)
a e
These give a system of simultaneous equations from which perturbative ap-

proximations to the transition amplitude Tba can be obtained. Equivalently
in operator form Eq. (18.50) can be written as
1
T = -V+ VE. T. H (18.51)
o +ic a -
The iterative solution to this equation is given by
1
T=-V-VE H . V ... (18.52)
a- O+ Zé
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
e-iHotfñU(t, -oo )la) = e-iHt/ñ lim r¡ {º dt' er¡t'e(ífñ)(H-Ea.)t' la)

r¡-+O 1-oo
= e-iHt/ñ ic la)
Ea -H +ic
= e-iHt/ñ.laC+l)
= e-iEat/ñla(+l). (18.54)
Hence
(18.55)
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)
Note that equivalently it can be written as
Cb(t) = -~(-Tba) ¡too dt'ei(Eb-Ea)t'/ñ+r¡t' +óba, (18.57)
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
= (27r)28(0)8( Eb ~Ea) ;2 lnal2
= (21r) 2 ó(0)8(Eb - Ea),¡:II'ba

1 12 ·
Now
l ¡t/2 iwt' d"t'
8(w) = - lim
27r t-+ -t/2
8(0) =2
1 ¡t,/2 dt
71" ~t/2
1 .
= - lirn t.
27r t .....
Therefore, for large t,

2 (27í)2 1 2
ICb(t)I = (21r) 8(Eb - Ea)nlTbal t (18.59)
or
(18.60)
The same result is obtained from Eq. ~) if the limit r¡ -t O is taken
at the very end. It is assumed that Tba has no singularity as a function
of energy at Eb = Ea. The above result implies a constant transition
rate. This is what we expect to be the effect of the the scattering causing
transition from state la) to state lb) and the above transition rate is closely
related to the scattering cross-section as we shall see.
Equation (18.60) explicitly exhibits the fact that energy is conserved in
the transitions. If the transition is to a continuum of states about state b,
then p¡(Eb)dEb is the number of final states with energies between Eb and
Eb + dEb, where p¡(Eb) is the density of final states, i.e. the number of
states per unit energy. In this case, the transition probability per unit time
is given by
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.
18.5 The Scattering Cross-section
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
FORMAL THEORY OF SGATTERING 419
Now
and
fv 1up(r)l2d3r = (2:)3 = l.
Thus we take V = (21rn)3. Hence the number of final states between p f

and P¡ + dp¡
p¡(E¡)dE¡ = d3p¡
= 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)
where p¡(E¡) is given in Eq. {Zl.65b).

For the case of non-relativistic elastic scattering in the centre of mass
system
E- _ p¡ E _ PJ dE¡ IP¡I
i - 2µ' f - 2µ' dlP¡I µ
!Pil = IP¡I = P and V= 'f!..,

µ
Therefore
p ¡ (E¡) = p2 !!_ dO
p
=µpdn (18.67)
and
Flux 1 )3 p
= ( 21rri --¡;,·
Hence
(18.68)
18.6 Properties of the Scattering States
We have derived
(18.69) )
(18.70)
We may consider Eq. (18.69) to be a formal solution of Eq. (18.70) in the

sense that la(±)) are given in terms of the known states la). Ifwe substitute
the formal solution in Eq. (18.47), we get the transition matrix elements
na= -(blVla+)
1
= -(blVla) - (blV E H . Vla). (18.71)
a - +ic
In this way, 1Tbal2 or the scattering cross-section for a scattering process can
in principie be obtained. However, in practice, not much is gained since the
effect of the operator Ea-k+ic is not known unless eigenvalues of H have
already been determined but this is the problem we wish to solve. Hence,
in practice, one has to resort to approximate methods to solve Eq. (18.70).
For this purpose, it is convenient to introduce the Green's operators

1
Gt(Ea) = E
a -
H ±.
O Zé
(18.72a)
a±(Ea)= Ea-~±ic' (18.72b)

so that
(18.73a)
with solutions
(18.73b)
Using the identity
1 1 1 1
- - - = -(B-A)-
A B B A
with
A=Ea-H±ic L )
B = Ea - Ho ± ie,
we have
(18.74)
We can solve a±(Ea) by the perturbative expansion
G±(Ea) = Gt(Ea) + Gt(Ea)VGt(Ea) + Gt(Ea)VGt(Ea)VGt(Ea) + · · · ,
(18.75)
so that
This can be regarded as an expansion of

1
la±)= 1- Gt(Ea)Vla), (18.77)
where l-Gl(Ea)V is expanded in power series.

The formal solution Eq. (18.69) can be used to demonstrate the orthog-
onality of the scattering states. Now
1
(b(+l1a<+l) = (bla+) + (blV . laC+l)
Eb - H -Zé
= (bla) + Ea - ~
b + Zé. (blVla+) + (blV E b - ! .
a - Zé
laC+l)
= (bja) = Óab, (18.78a)
where we have used Eq. (18.45b). Similarly
(bHlaC-l) = Óba· (18.78b)

Corresponding to the orthonormal set la), we have obtained two sets laC+l)
and laC-l) of orthonormal eigenvectors of H. Are these sets complete?
The answer in general is no, since H may also have discrete eigenvalues
corresponding to bound states produced by the interaction V. These dis-
crete state~hich have no counterpart in the spectrum of H, are never
found among the solutions of Eq. (18.70), as these solutions refer to the
continuum part of H spectrum. These bound states are orthogonal to the
scattering states as they have energies lower than energies in the scattering
process and must be added to the scattering states laC+l)[or laC-l)] for the
completeness relation. Thus we have the completeness relation
¿ lcC+l)(c(+)I+ ¿ IB)(BI = 1, (18.79)
e B
where IB) refer to the bound states and
(/+llB) = oL (18.80a)
)
(cC-)IB) =0. (18.80b)
18.7 The Coordinate Representation
Let us consider the Green's operators
G±(E) - l (18.Sla)
O a-E a - u±·
no Zé
We can represent them as follows:
(18.Slb)
We now take the r-representation of our fundamental equation (18.73a)
(ria(±)) = (ria)+ (rlGt(Ea)VlaC±))
= (ria)+ JJ (rlGt(Ea)lr")
x (r"!Vlr') (r'la(±l)d3r" d3r'. (18.82)

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
7/Ji±l(r) = 7/Ja(r) + J G±(r, r')V(r')7/Ji±l(r')d3r'. (18.84)
On the other hand, the r-representation of the transition matrix is given

by
Ti¿ = -(blVlaC+))
= - JJ (blr") (r"IVlr') (r'la+)d3r" d3r'
= - J 7/J;(r')V(r')7/Ji+l(r')d3r'. (18.85)
In the Born approximation, we can replace in the integrals above
L
7/Ji+)(r') by 7/Ja(r'), so that
,¡,¡+l(r) ~ ,P.(r) + j G+(r,r')V(r')V, (r')d''r'0 (18.86a) )
Tt/, = - J 7/J;(r')V(r')7/Ja(r')d3r'. (18.86b)
We consider two applications of the above considerations.

Application I: Consider two spinless and distinguishable particles o: and
(3 which interact via a potential
V(r) with r = ra - r¡3.
No other potential is present in the problem. Thus
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
Pi= ñki, P¡ = ñk¡,

July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics' 1
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
= ó(p¡ - p.). (18.87c)

Now
(rlGt(Ea)lr') = ¿ (rlq)(qlr')_
Ep-Eq ±ic
q
= 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
1 d3 2µ ( 1 ) 1 ±iE. lr-r'I :'") ( )

x p2 - q2 ± ie q = f¡,2 - 471" Ir - r'I e " . ~ 18.89
(See Eqs. (8.171) for d3q integration, here p = nk.) Then Eq. (18.84)
beco mes
2 ±.':Elr-r'I
eip·r/ñ - -(1 _!!:_)
(±)
'l/Jp
-
(r) - (27rn)3/2
1
471" f¡,2
x V ( r')'l/J~±)(r')d3r'.
J _e,---;,,-,---
Ir - r'I
(18.90a)
Now
Ir - r
'
1 =
2
(r - 2r · r
'
+ r ,2 ) 1 I 2 '::::' r[l - -r r. r' + 0(-)]
1
r2
so that
(18.90b)
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)
while p¡(E¡) is given in Eq. (21.65b), i.e.

2dlP1I
p¡(E¡) = (IP¡I dE¡ )dO
= µ¡IP¡ldO. (18.93a)
Now
1 IPil
(Flux)in = PinV = (21rñ)3 µi . (18.93b)
Hence Eqs. (18.91), (18.92) and (18.93) give

da B 1 1 IP¡I - 2
(dO) = 47r2 ñ,4 IPil µiµ¡IV(~)I . (18.94a)

µi = µ J = µ, IP ¡ 1 = IPi 1 = P,
we have
dO")B
( dO. = _1_µ2IV(~)l2
el 47r2 ñ,4
= (~)21V(~)12. (18.94b)
21rñ2
Application II: 2 spinless and distinguishable particles a and (3 interacting
via V(ra - r13) = V(r), with r = r°' ·- r13 but there is also a fixed attractive
potential U(r13) acting on (3 only: Thus
H =Ho + V(r),
A2 A2
Ho = + + U(rt3). (18.95a)
2p°'
m¿ 2Pf3
mp
U(rt3) is strong enough so as to possess a number of bound states
A2
( Pt3 + U(rt3))(rt3ln) = En(rt3ln). (18.95b)
2mt3
In other words, the particle (3 is bound in the potential U(r13). Let Yn stand
r: for the nth bound state. We consider the following processes:
J + Yo ----; a + Y 0 ~ tic catt ring
Iª + + Yo - Q Y,, n-:/; o, rnelastic scattering,
for example
a = e", Y0 nucleus (A, Z),
(3 = proton, Yn excited nucleus (A, Z)*.
In this case, states la) and lb):
la) = IO) IPa), lb) = In) IP~)
are eigenstates of H0 with energies
Ea= Ea +Eo 1_ )
Eb =E~+ En.
Then from Eq. (18.86b)
Tt/i = -(blVla)
= - JJJ j(blr~rJ)(r~rJIVlr~r~)
x (r'o r'/3 la)d3r"o d3r"/3 d3r'o d3r'¡3, ~/ (18.96a)

where
(r0r13la) = (rolP0)(r13IO)
_ <P ( ) 1 ip"'·r/ñ (18.96b)
- o r¡3 (21rn)3/2 e
(rorplb) - (rolP:)(rpln)
_ ( ) 1
,
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~,
dc,;f 1 1 IP~I 2 - )121 ( )12

dO~ = 47r2 /¡,4 IPol·malV(q r; q
= IP~I ( mo )2IV(q)l2IFn(q)l2, (18.101)
IPo 1 27rfi2
For the elastic case, we have n = O, p~ = p0 and formula (18.101) becomes
!~t = (2:~2 )2IV(q)l21Fo(q)l2, (18.102)

where Fo ( q) is the elastic form factor
Fo(q) = f iq·rl</Jo(r)l2d3r. (18.103)
We can rewrite (18.102) as
daB
d[t = a (q)IFo(q)l
<>
8 2 (18.104a)
where
aB(q) = (2:~2)2IV(q)l2 (18.104b)
is the usual Born approximation to the differential cross-section for scat-
tering through momentum transfer q for a fixed target. A better approx-
imation is to replace a8 by ael, the exact elastic differential cross-section
off a fixed ,B-particle target:
!~~
<>
= CTez(q)IFo(q)l2, (18.105)
the so called impulse approximation.
18.8 Unitarity of the S-Matrix
We have defined the 8-matrix

S = lim U(t, to), (18.106)
to--oo
t-+ oo
where
U(t, to)= eiHotfñe-iH(t-to)fñe-iHoto/ñ. (18.107)
Then the unitarity of U(t, t0) and hence of 8-matrix follows from the
hermiticity of the Hamiltonian.
The unitarity property can also be seen as follows
(sst)ba = (blSStla)
= ¿(blSlc)(clBtla)
e
= ¿(blSlc)(alSlc)*
e
= ¿(bC-l¡cC+l)(cC+l¡aC-l) + ¿(bC-llB)(BlaC-l),
e B
(18.108)
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)
¿ (ó¡c + 21ri8(E¡ - Ec)T¡c) (óci - 21rió(Ec - Ei)Ti:) = Ófi
e
or
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
we would have obtained

i(T/¡ - T¡i) = 21r ¿ T;¡Tcio(E¡ - Ec), (18.115b)
e
for E¡= Ei.
Equations (18.115) are frequently used unitarity relations. In the above
relations, Í::c means summation over e extended to all possible states.
We now show that conservation of probability is a consequence of the
unitarity relation, which in turn is a consequence of the ~rmiticity of the e
Hamiltonian. We recall that the probability of transition from a state li)
to any state lb) is given by
1Cb(t)l2 = l(blU(t, -oo)li)l2.
We sum over all possible states b
¿ ICb(t)l2 = ¿(blU(t, -oo)li)*(blU(t, -oo)li)
b b
= ¿(ilUf(t, -oo)lb)(blU(t, -oo)li)

b
= (ilUt(t, -oo)U(t, -oo)li)
because the states lb) form a complete set. Thus
¿ ICb(t)l2 = (ilUt(t, -oo)U(t, -oo)li)
b
= (ilU(-oo, t)U(t, -oo)li)
= (illli)
= 1, (18.116)
which is the statement for the conservation of probability.
18.9 Optical Theorem
Let us now specialise to the case f = i. Then

Ti: - Tii = -2ilm Tii· (18.117)
Therefore, from Eq. (18.115), we have
ImTii = 7f ¿ 1Ticl28(Ei - Ec)· (18.118)

e
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 fk(O) is the elastic (since i = f) scattering amplitude for an energy

2 f;,2k2
E; = E = fµ. = ,¡¡¡- at the momentum transfer D. = O.
Now
where ¿ denotes summation over all other quantum numbers necessary to

specify the state le). From Eqs. (18.120) and (18.64), the quantity
2rr ~ 2 p
-¡;- L ITcil 8(Ec - Ei) = (2rrn)3µ O'total, (18.121)
e
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.
18.10 Problems
16.1 A system is prepared in the state la') at t = O. Show that the

probability of finding it in the state lb') at time t is given by
1 (b' le-iHot/nu(t, O) la')l2,
where U is the time displacement operator in the interaction picture
la(t))¡ = U(t, to)la(to))¡.
16.2 Show that the unitarity of S-matrix follows from the conservation
of probability.
16.3 If the S-matrix is represented as
S = l -í/2 K
l+í/2K'
where K is a hermitian operator, show that it satisfies ss' = l.
Derive the relation (analogue of unitarity relation)
T¡i = -K¡i - it: ¿K¡aTaió(Ea - Ei),
a
where
(JIKlí) = 21r8(E¡ - Ei)KJi·
16.4 Show that for the scattering process
a+ Yo --ta+ Yn,
if the potential U(r), which possesses the bound states Yn, is cen-
tral, the inelastic form factor (n =f. O) given in Eq. (18.99b) is
F11n(q) = 4rri1y¡;,.,(q) fo00 u~(r)j¡(qr)u (r)r dr,

0 2
where l(l+l)ñ2 and mñ are eigenvalues of L2 and Lz. Show further

that
(i) Firri(q) '""'q1 as q --t o,1(ii) the partial wave cross-section is
da
~: = (2l + l)~O'B(q)I
p' ~ ¡00
u~(r)jz(qr)uo(r)r2drl2,
d~G0 Po O
where un(r) are the solutions of the radial Schródinger equation.
Hint: Write
<Pn(r) = (rln) = Yim(r)un(r)
eiq-r = 4rr L i j¡,(qr)Y¡:m,(<l)Yi,m (r)
11 1
L'im'
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
16.5 Show that for

V=Vo r~a
=0 r>a
the Born approximation (B.A.) is valid if
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'}.
For B.A. to be valid, the modulus square of the 2nd term in { }

should be « l. Then calculate { } for the potential given for
r = O. Put R = r' - r and take p along the z-axis,
16.6 Starting from the equation
c(+l(E) = 1
E-H +ic'
show that
ii
c+(t) = 2~ J e-iEt/nc+(E)dE = -iO(t)K(t),
where
is called the propagator.

iii
K(E) = J eiEt/liK(t)dt = 271" ¿

n
1En)(Enl8(E- En),
where IEn) are eigenstates of H, HIEn) = EnlEn)·

iv
K(r, r', t) = ¿ e-iE,,t/liun(r)u~(r')
n
K(r, r', E)= 21r ¿ Un(r)u~(r')o(E - En)-

n
v For a free particle

Ko(r r' t) = (_!!!'..__)3/2e- 2~, (r-r')2
' ' 21riñt
mv'2mE
Ko(O, O, E) = 1rñ3 .
Hint:
n
()( ) - _1 Joo eiwt dw
t - 21ri -oo w-ie ·
(1) For a free particle

~ ( ) _ 1 ip·r//i
(18.123)
Un ( r ) - Up r - (2rrñ)3/2 e .
Chapter 19
S-MATRIX AND INVARIANCE

PRINCIPLES
19.1 Introduction
We have seen in Sec. 10.5 that an invariance principle is connected with a

conservation law. By invariance we mean the following.
If the result of any experiment on sorne system is unchanged by a physi-
cal transformation of the apparatus, then the system is said to be invariant
with respect to that transformation. This is not an empty statement but
gives information about that system.
In quantum mechanics, a transformation is described by aJ unitary
transformation U. The result of an experiment involving a transition from
an initial state ji) to a final lf) is described by the matrix elements of an
operator S called the scattering matrix (S-matrix) between states li) and
lf). Invariance means
(JISli) = (rlSliu)
= u1utsu¡i) (19.1)
or
s= tr su (19.2)
or
[S,U] =0. (19.3)
Here
jiu) = Uli) (19.4)
Ir)= Ulf) (19.5)
are the transformed states. Equation (19.3) shows that invariance under
ªID unitary transformation U means that a S-matrix commutes with it.
435
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
Fli) = Ff li) (19.11)

Flf) = F¡lf), (19.12)
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)
that is, the eigenvalues of F are conserved in the transition. F is then

said to be a constant of motion. Operators like F, whose eigenvalues are
S-MATRIX AND INVARIANCE PRINCIPLES 437
conserved, are called 'good operators'.~b) U discrete (for example space

reflection) U is of course unitary. If
U2 = 1 (19.16)
then the eigenvalues of U are given by
U'=±l L
U is also hermitian. Therefore U itself can be regarded as one of the
(19.17) o
observables of the system, for example, the parity operator, provided that
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
UIBli) = uut¡s¡ui) L (19.19)

(iii)We can derive restrictions on the form of the transition amplitudes.
19.2 Rotational lnvariance
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
P7,PJ and Pi. P¡ L (19.21) o

If we specialise to two body elastic scattering,
2 2
Pi
Ei = -2µ = -P¡ = E¡ = E (19.22)
2µ
Pi · p ¡ = p2 cos (), (19.23)

then we have two independent variables (E, B) or (p, B), i.e. energy E and
scattering angle ()
(P¡ISlpi) = (p¡, (), </>IBIPi, o, O)
= ~Pi -p¡)(B,</>IS(p)IO,O) (19.24)
= ~Pi - P ¡ )A(p, B) (19.25)
because the 8-matrix has non-vanishing matrix elements only on the 'energy
shell', i.e. between states of same energy E¡ = E¡. We note
(B,</>llm) = Yzm(B,</>) = (-ir (C2~!U~~~)')

112
P¡m(cosB)eimc/> l )
(O, Ojlm) = Yzm(O, O). (19.26)
Because of rotational invariance, there is no dependence on m and we can
put m = O and thus we have
(B, </>ll, O) = (B, <t>ll) = ( 2~t1)

112
P¡ ( cos B) L (19.27) )
(O, Ojl) = (2l4: 1 r/2 (19.28)

(P¡ISlpi) = 8(p¡ - Pi)(B, </>IS(p)IO, O)
= 8(p¡ - Pi)¿¿ (B, </>ll) (llS(p)ll') (l'IO, O)
l l'
2l + 1
= 8(p¡ - Pi)¿ --F¡(p)P¡(cosB), (19.29)
l 41f
where we have put
(llS(p)ll') = 8wFz(p), (19.30)
because due to rotational invariance, angular momentum is conserved.
Now from the unitarity of the S-matrix
ss' = 1 (19.31)
(19.32)
or
(19.33)
Now
(19.42)
and we can write
8(pi -p¡) = ;28(pi -p¡)8(cosB)8(</>) l (19.43) ú

Therefore
1
28(pi - p¡ )8(cos B)8(</>) =
p
¿ l
a¡(p)Pi(cos B). (19.44)
Multiplying both sides by P¡,(cosB) ~integrating over angles and using /

the orthogonality of Legendre polynornials, we have
\o(pi - p¡)
P
r:
lo 1-1
{+i 8(cosB)8(</>)P1,(cosB)d(cosB)d</>
= ¿a (p) lot" 1-1{+1 P (cosB)P¡,(cosB)d(cosB)d</>

1 1
1
2
= ~a¡(p)2l + 18w2rr. (19.45)
Therefore
Hence
(19.46) o
8(pi - P¡) =
8(pi - p¡) ~ 2l +1
Z:: -4-Pi(cosB). (19.47)
2
p l=O tt
Therefore from Eqs. (19.40) and (19.47), we have
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.
Because of rotational invariance, M can be a function of scalars Pi · p2,

Pa · p3, Pa · Pi only; it cannot be a function of Pi · (p2 x p3) because of
Eq. (19.83). Hence rotational and space inversion invariance imply
or
M=O. (19.85)
Thus the above decay is forbidden if we assume invariance under space

inversion. In other words, a spinless particle with even parity cannot de-
cay into three spinless particles each having odd parity if paritY,tinvariance Q....
holds. This is an example of how an invariance principle leads to selection /
rules.
19.4 Time Reversal
Under time reversa!
t=» -t, r-- -;rh._ (19.86)0
Thus under time reversal
p----;-p, L=rxp----;rx(-p)=-L (19.87)

}
u----;-u ~
Let II denote the operator which transforms quantum mechanical states

(19.88)
G
and operators under the above transformation t=» -t.
First we show that II cannot be an unitary operator. Under II, the
commutation relation
[qi,Pj] = iMij
------; -iMij· (19.89)
Hence transformation generated by II cannot be unitary. But we want the

above commutation relation to be invariant under II. A way out of this
difficulty is as follows.
(i) All c-numbers are simultaneously transformed into their complex
conjugates, or
446 Quantum M echomics
(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)
if we use definition (i).

Using definition (ii),
IITir1 = T. (19.102)
Now invariance under II implies
(!ITli) = u1rr-1rrrrr-1IIli)
= (¡t 1rt lit)*
= 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
la) = la, Pa, ma)

la(+))= la, Pa, ma)out
o
where
(19.105)
Using definition (i)

(19.106)
m
Iljo; Pa, ma)out = la, -Pa, -ma) + E
a -
H
-
. Vio:,
Zé
-Pa, ma)
If we use definition (ii)

= la, -Pa, -ma)in ·ij
Iljo, Pa, ma) = (a, -pa, -mal·W ,

(19.107)
(19.108)
1
I V
IIla, Pa, ma)out = (a, -pa,-mal + (o, -Pa, -malVE H .
a +ic
(19.109)
Chapter 20
RELATIVISTIC QUANTUM
MECHANICS: IRAC EQUATION
(( ~Dirac's equation can be written in one line as
i!i %/1/J) =(cap+ ,Bmc)l1P)
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~.
20.1 A Brief Review of{:_heory of Relativity
The special theory of relativity is based on two postulates:
(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
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
g°º = gºº = 1, lj = -óij = -Óij = 9ij

From (i) and (ii), it follows that the length of the vector x is invariant:
L ó
x'µx~ = xµxµ L
The vectors x'!' and xµ are related to each other by a linear transformation
( called Lorentz Transformation)
(20.3)
o
(20.4)
which leaves xµxµ invariant:

Xµ9µvx" = x'ª9a.fJX1fJ
=A~Xµ9a.(J
=A~x". (20.5)
The above equation gives
9µv -- Aªµ9a.(J AfJv · (20.6)

In matrix form Lorentz transformation [LT] is
x' =Ax, x = ATx'

and the matrix
g = (AT)µa.9a.(J(A)fJ'Y
=ATgA (20. 7)
RELATIVISTIC QUANTUM MECHANICS:DIRAC EQUATION 453
where A is a 4 x 4 matrix. Further
- .
detg = (detAT)det(gA)
= (detA)2detg
(detA)2 = 1, detA = ±1. (20.8)

~
If detA=l, LT is ca d proper, denoted by L+. This excludes spatial
-tJ\ reflection for which detA=lfl. Further from Eq.(20.6)
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
For Lorentz boost, for simplicity consider Lorentz velocity transformation

¡\
along x~axLs (v -::·(º)_ ~:r ~}
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
where the hermitian operators

Lµv = i(xµav - XV aµ) (20.16)
are six generators of the Lorentz group correspondingbparameters Eµv·
They satisfy the commutation relations
[LµV, LP = i(gµp F
17] gv Lµp + µ - v).
17 - 17(20.17)
In relativistic mechanics, energy and momentum are treated on equal
footing just like the space and time. Accordingly we introduce energy-
momentum 4-vector
(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)
)
20.2 Relativistic Quantum Mechanics: Introduction
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
E-----> in:t' p--> -ihV, (20.27)

which ac on the state fuuction 1>. We get the Schrodinger equation given
in Eq.~0.25)
In order to getr,;lativistic equation, it is natural to replace the non-
relativistic energy equation by the corte ponding relativistic energy qua-
tion
(20.28)
Replacing E and p by operators given in Eq.(20.27), we obtain the rela-

tivistic equation in quantum mechanics
- r,,2 82if! = _r,,2v2cp + m2c2éf!. (20.29)

c2 8t2
Now
,:,µ,:, - 1 82 T72
V Vµ - c2 é)t2 - V , (20.30)
Hence, we can write the differential equation (20.29), in the form
( 8µ8µ + m;t) if! = O
(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
same footing as required by the theory of relativity. But this equation as a

single particle equation faces difficulty.
The equation of continuity@L
aµjµ = O or e;:+ Vl)i = O (20.32)

must be satisfied as it is equivalent to conservation of probability. A natural
generalization for }µ from the Schr8dinger theory [Eq.(2.46)] is
Jµ = _ ___!):.._ [*aµ - aµ*J (20.33)

2im
l
giving
i; = = __ n_ [* a - a*]

[!_
e 2imc at at (20.34)
7
j = ___!):.._ [*V - V*] . (20.35)
2im
It is easy to see from Eqs.g9.29) and (20.33) that
aµJµ = o. (20.36)
The requirement that probability current density p must be positive definite
is not satisfied as the expression for p is not absolute square of the wave
function, but contains the time derivative of the wave function. Thus for a
mode with energy E which should oscillate as e-iEt/ñ, we have
(x, t) ,. . ., </J(x)e-iEt/ñ
so that
E
p = mc2 l</J(x)l2 (20.37)
where
E= ±Jc2p2 + m2c4 ±Ep,
l~
=
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.
The lGeinJ1Gordon equation cannot be considered as a quantum mechan-

ical equation for a single particle. One encounters two problems: Energy
can be negative for which probability density is negative. It is important
to realise that two problems are interlinked.
We may note that, Klein-Gordon (KJfG) equation

1 02 ) m2c2]
[( c2 8t2 - v72 +~ (t, x) = O (20.38)
for m = O, is similar to the second order differential equation
(20.39)
which the electromagnetic fields E, B satisfy. Pauli and Weisskopf ínter-

preted I<i,G equation as a field equation, where <l> is a field, the quantum of
which describes the spin zero particle such as rr-mesons.
Dirac made a breakthrough in 1928. He did it from pure logic by in-
troducing spinors in addition to more familiar scalar, vector and tensor
quantities.
By doing so the two problems are decoupled: pis always positive definite
but negative energy problem remains. Dirac solved it by proposing Hole
theory, thereby predicting antimatter.
20.3 Dirac Equation

\
~ We notice that in the Klein~ordon equation the problern wi h p aríses
\ ~ due to time derivativa in Eq.~0.34) which in turn arises because of I<lein~
Gordon equation being second order in time derivative. Dirac started with
a linear equation both with respect to time and space derivatives. Here
an analogy with the Maxwell's equations, which are fully relativistic, will
be helpful. One can write basic equations of electromagnetism in terms
of electromagnetic fields E and B which have six components and satisfy
Maxwell's equations, which are linear in time and space coordinates, or
in terms of vector potential Aµ. which have four components but satisfy a
-(~\ second order Klei~ordon equation
8µ.8µ.Av = O.
As noticed previously each component of E and B also satisfies such a
second order equation (20.39).
The above analogy suggests that
(1) We should expect severa! components of wave function VJn·
(2) Each component should be related to other by first order linear
equations both in space and time derivatives. ~
(3) Each component should satisfy the Klein Gordon-equation. ('J
The most general equation one can write satlsfying conditions (1) and
(2t· put e= 1 and /i = 1] is
(iÓtngt +i(ai)eno~i -f:Jbim) 1/J,i = O (20.40)
where €, n = 1, 2, ...N and i = 1, 2, 3. In order to satisfy condition (3)
o
operate by
(-""'' ! + i(a')m, a!, _ mil,,,,) L

. .
One obtams, smce ata an
d a
ox commute,
°
[ 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.
Thus the Eq.(20.40) takes the matrix form [inserting back ñ ande]
(in c~t + iña.V - (3m) 'ljJ = o (20.45)
where 'ljJ is a column matrix with N components. Or
iñ~~ = H'lj; = (ca.(-iñV) + (3mc2)'lj;,
(20.46)
where p = -iñ'v is the momentum operator. Since the Hamiltonian oper-

ator
H = a.p+(3m, (20.47)
should be hermitia~each of the four matrices ci and (3 must be hermitian:

7
at = a, (3t = {J~ (20.48) O
From the anticommutation relations ai (3 -(3ai, det( ai (3)
(-l)Ndet((3ai) i.e. det(ai)det(3 = (-l)N(det(3)det(ai). Thus (-l)N = 1
so that N must be even. Further
(20.49)
so that
But
(20.50) o
Tr((3a'(3) = Tr((3(3ai) = Tr(ai). (20.51)
Thus Tr(ai) = O; similarly Tr(f3) = O. For N = 2 matrices 1 and O"i form

a complete set but since the unit matrix commutes with O"i it cannot be
identified with (3. Thus the lowest possible rank is N = 4. One can choose
a particular representation
a'=
. ( o (J"i)
ui O , with I = G ~) (20.52)
where O"i are 2 x 2 Pauli matrices, which satisfy

It is easy to show that the representation given in Eq.(20.51), known as

the Pauli representation, satisfy the relations (20.44). Then 'lj; is a column
matrix:
(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
20.4 Covariant form of Dirac Equation
We rewrite Dirac equation given in Eq.(20.45) by multiplying on the left

by ,B:
ifi,B~~ - [c,Bo:.(-iliV + ,Bmc2)]'1j; = O. (20.55)

Noting that
,82 = 1,
we can write Eq.(20.55) as
[in (,ºo~º + ,iai) - me] .,¡; o =
or
(20.56)
where
,B = 't", ,Bo: = ,; )
)
(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·
The properties of the -y-matrices discussed above are independent of any

representation of ,-matrices. Two representations of -y-matricesare useful:
Type No. of independent components

1 1
'Yµ 4
6
4
1
(i) Pauli representation:
L )
This representation is particularly suitable to go to the non-

(20.63)
o
relativistic limit.
(ii) Weyl or Chiral Representation:
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 )
Then we can write

O"-µ.= (1, -u), ifµ. = (1, u) L o
'Yµ= (Jµ (J';)L
This representation is specially suitable in taking the limit m----t O i.e. for
chiral fermions.
(20.65)
o
In the Pauli representation, one can write the 4-component wave func-
tion 'ljJ as 'ljJ = 7/JA + 7/JB, where
7/JA = 1 ~ 'Yº 7/J = ( cpOA) ' 7/JB = 1 - "(º 7/J = ( 0 )

2 cpB
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;.
20.5 Relativistic Invariance of Dirac Equation
Under Lorentz transformation

x'!' = Atx"
a:= A~a"
and suppose that
'lj;(x) -4 '1/J'(x') = S'lj;(x) (20.68)
Then
('Yµa: - m)'l/J'(x) = ('Yµ A~8 11 - m)S'ljJ(x) =O
or multiplying left by s »,
[(s-1'YµS)A~8 11 - m]'l/J(x) = O
which reduces to
['Y"811 - m]'ljJ(x) = O
as required by the invariance, provided that

s-i'YµSA~ = 'Y"
or
S -1 'Y µSA"µg11(3 Af3a = 'Y " g11(3 Af3a·
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]
as is clear from Eq~0.72). Thus the relation (20.72) can be written as
'lf;'(x) = [1 - ~Eµ11Mµ11] 'lf;(x) = U(A(E))'lf;(x) (20.75)
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
In particular, for rotation about x3-axis,
E12 = E,
~12
v
= º'"V1'"V2,
• , ,
L12 = i· ( x 1~8x2 _
x 2~)
8xl = -~L3
li '
L3 is the third component of the orbital angular momentum along x3~axis.

Now /\
( i·l·/)2 = l.
Thus the operator Í'y1í'2 has eigenvalues ±l. Let us write
L
2
Hence the eigenvalues of 83 are±~. Thus
M12 = L12
8 3 = __iñ ,, 1 í' 2
+ ~a-12 = -~(L3 + S3)

(20.77)
(20.78)
ó
2 Tí
and
(20.79)
Thus we see that change in the fundamental form of 'l/; induced by an

/'- infinitesimal rotation about :d1ax:is is determined by the operator
(20.80)
where J3 = L3 + 83 should correspond to the third component of the

angular momentum operator, since it generates an infinitesimal rotation
around the third axis. As already seen L3 corresponds to the orbital angular
momentum while 83 = -~'y1,2 which has eigenvalues ±~ represents an
additional angular momentum associated with the particle and is called
1
the spin angular momentum. Thus the Dirac equation represents a spin
2
particle.
In general
ij _ i ( i i i i) _ i (ªiªj - aJai O )
o - 2 'Y'Y -, 'Y - -2 O aiai -aiai
} where we have usect,
so that
(20.81)
But w~ can write
so that
(20.82)
Also
.. 2
EijV1 = -¡¡:w.L. (20.83)
Hence from Eq.(20.72), we have
'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 = ±!.
Hence Dirac spinor '1/J represents a spin particle. !

Remarks: J is the generator of rotation group
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
refers to the rotation of vector by an angle w.
20.6 Transformation Properties of Dirac Bilinears
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)
Using Eq.(20.69), we see that

Jµ(x) = {;(x),,t'lj;(x)---+
J~(x') = At{;(x)''('lj;(x) = AtJV(x)i,1

i.e. it transforms like a vector.
(ii)
r = ,.,r5
We note that
3-l,.,,153 = _!:_E
4! /J,VAp
s-l,.,,µ.33-l,.,,v33-l,.,,>..33-l,.,,PS
I 1 1 ¡
- 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)
s-1,,~ ...¡5s = s-17 ., ss-1~y5s1
@A)A~1µi l (20.93)
G
(iv)
r = (Tµ.v
Now
s=-»: = i¡s-11,,ss-11vS- s-11vss-11,,S]

2
= ~[A~1>.A~1P - A~1v A~y'J

= A~A~c,AP. (20.94)
Hence using Eq. (20.90) under proper Lorentz transformation L+ the bi-
linears transform as
ifnp: Scalar
ih5'1/J:
i/J1/J,'I/J:
i/J1/J,,,/''if;:
nsor
The proper Lorentz transformation L+ <loes not distinguish between ;j;'lj;

and i/;15'1/J and between if;1µ'lj; and i/;1,,15'1/J. The appearance of det(A)
suggests that we should consider space reflection to distinguish between
these bilinears.
20.7 Discrete Transformations
(1) Space Reflection:
X----+-X, t----+t (20.95)

Under space reflection
Aµ= (~ ~1 ~ ~) det(A) = -1. (20.96)

V o o -1 o
O O O -1
Under space reflection
x----+ xp = (xº,-x) (20.97)
and suppose that
(20.98)
l July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics l
where Sp is a 4 x 4 matrix. Invariance under space reflection means

that for free particle Dirac Eq. (20.56) we should have
(in1µa: - m)'lj;p(xp) = o. (20.99)
Let us write Eq.(20.56) in terms of xP :
(in('Yº8{ - 'Yi8¡ - m)) S¡;1'!f;p(xp) = 0.
Multiplying by Sp on the left
(in(Sp1ºS¡;1a{ - Sp,yiS¡,1af') - m) 'lj;P(xP) = O
i.e. it reduces to Eq.(20.99) as required by the invariance if
sP'Yº s¡; 1 = 1ºL )
S P'Y i3-1p = -'"( i · (20.100)
These can be satisfied if we choose
Sp = rJP'Yº
where r¡p is a multiplicative constant. We note that
{;(x)--, i[;P(xp)'Yºr¡p. (20.101)
Thus Jr¡pl2 = 1 if we want the probability density 'lj;t(x)'lj;(x) to
be invariant i.e. to be the same in both coordinate systems. Thus
under space reflection.
'lj;P(xp) = r¡p'"(º'!f;(x), (20.102)
Since 15 anticommutes with ,yº, it follow then
'!f;f (xP) = 'I/JR(x)l- )
'I/Jk(xp) = 'lj;L(x). (20.103)
Thus the bilinears, under space reflections transform as
J if;(x)'lj;(x) __, if (x)'!f; (x

1 1 1) ~(xhº1º'!f;(x)
- '1/;(x)'lj;(x):
(Pseudoscalar)
where
since
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:
t ----t -t, X ----t X
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)
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),
[-iñ ('Yº*(8a) + ('yi)*ai) - mc]'I/J*(x) = o.

Writing it in terms of xr, it becomes
[-iñ ('Yº*(-8'{;) + ('i)*aT) - mc]Si1'1/Jr(xr) = O.
Multiplying on the left by Sr,
[iñ (Sr'Yº* s;::1(8'{;) - Sr('YÍ)* s;::1) - mc]'l/Jr(xr) = o,
which reduces to Eq. (20.106) as required by invariance if
Sr'Yº* s;::1 = 'Yº L )
(20.107)
However
giving
S r'Y 's::
r = -_,,i ,
= ,,..
f. .
1,
= 2
= i 3
The matrix which satisfies these conditions is

Sr= r¡rB
where
(20.108)
and in: is a phase factor
Now
)
BT = (-y1,,3f = b3f b1f = ,,3,,1 = nt = ts:', (20.109)

From EqE.105):
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)
if;(xh5'l/J(x) --4 'IÍJr/'5*(xT)'l/JHxT)

= if;(x)n-1,,5 B'l/)(x)
= if;(x),'51/J(x). (Scalar)
if;(xh,,'l/J(x) --4 'IÍJf(xThµ,*'lj;r(xT)

= (-1),,if;(x)B-1¡'µ B'lj;(x)
= (-l)µ'lf;[z;),'µ'lj;(x). (Vector)
if;(xhµ,1'51/J(x) --4 'lfHxT)bµ,)*(,'5)*1/Jr(xT)

= if;(x)B-1(,'µ,)*(,'5)* B'lj;(x)
= if;(x)B-1(,',,)* BB-1(,'5)* B'lj;(x)
= (-l)l'if;(xhµ,1'5'!/J(x). (Vector)
20.8 Generators of the Lorentz Group and their Represen-

tations
20.8.1 Matrix representation of the generators of the

Lorentz group
We can construct the matrix representation of Lorentz group as follows.
For an infinitesimal Lorentz transformation viz
@ x'1•~-
let us introduce an operator (matrix) U(A)

Aµxv
1/ '
with
Ae 8t + Ee (20.111)
U(A) = 1 + 21 Ea,eAª.6. (20.112)
Then A gives the matrix representation (adjoint) ofLorentz transformation.

From Eqs (20.111) and (20.112), we have
!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
so that we can write
(20.114)
Then six matrices

are given by
(20.115a)
)
= (HE)
ooi o
:St =S!l) L (20.115b) )
(M31 )i = i(gµ38~ - gµ18~)
= (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)
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
ÍEijk/'i,kL (20.117) '

)
-ÍEijkK,k·
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.
20.8.2 Representations of Lorentz Group

As already seen in Secs. 20.1 and 20.5, the most general representations
of the Lorentz group are given by Mr", which satisfy the commutation
relations given in Eq. (20.74). Just as L¡
transformations has two sub-
groups, the six generators of the Lorentz group split into three generators
Mi} which belong to the rotation group 0(3). This can be seen as follows.
Define
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)
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
Ni= !(Ji - iKi)

2
which satisfy the commutation relation.
)
[Mi, Mi] i(/ik Mk L
)
r. . .
~ [N',M']=O. (20.127)
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)}
k = -u, ... ,u, l = -v, ... ,v.

Suppressing u, v on the basis vectors of the product space:
J3lkl) = (M3 + N3)lkl)
= M3lu,k)lv,l) + lu,k)N3lv,l)
= (k+l)lk,l) L )
J±lk, l ;= k ± 1, l)[u(u + 1) - k(k ± 1)]112
'+lk,l±l)[v(v+l)-l(l±1)]1/2 L - :;:;, )
lk, l)i(l - k) L- )
K±lk, l) = lk, l ± l)i[v(v + 1) - l(l ± 1)]112
lk± 1, l)[u(u+ 1) - k(k ± 1)]112.
-
(20.129)
We see that in the ( u, v) representation matrix representation of K3 is di-
agonal and imaginary. This implies that finite dimensional representations
of the Lorentz group are non-unitary. We can label the representation of

Lorentz group as followL
Under space reflection Ji ---+ Ji, but Ki = Mºi change sign so that
Mi - Ni and thus (O,!) - (o,!). Hence if a left handed spinor belongs
!,
to representation 7/J L ( O) , the right handed spinor belongs to representa-
!,
tion 7/J R ( O) .
Now for the fundamental spinor representation (!,O),
Ji - 1 i
-2ª' Ki -
--20'i il )
i
Mi=~ ,Ni= O. (20.130)
Thus under rotation
and under boost
q>L---+ e~~.u (!:) · (20.131)
For the spinor representation (O, 1/2)

. 1 . . i .
J' ---+ -a' K' ---+ -a' (20.132)
2 ' 2
which give
)
)
(20.133)
The Dirac spinor '1/JD transforms in (~10) EB (O,~) representation of the

Lorentz group, which is a reducible representation
( eiOn. ~
q>L)
q>R rotation O eiº~-~) (!:)~ )
~
(!:) ------+
boost
(
e~~-2a
V
e-11~-~
o ) (!:) (20.134)
ú
20.9 Gauge lnvariance: Dirac Equation in the presence of

Electromagnetic Field
The Dirac equation

(20.135)
is clearly not invariant under local gauge transformation [A is a function of
x]:
(20.136)
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
8µ,'l/1--+ (8µ +:~Aµ) 'ljJ = DµVl·

Then the resulting Dirac equation
[ili (8µ, +:~Aµ)- me] 'l/J(x) =O

or
(in,µ Dµ - me)'l/J(x) = O (20.137)
is invariant under the combined transformation (20.136) and
Aµ(x)--+ Aµ(x) + 8µ,A. (20.138)
It is easy to see that, under gauge transformations (20.136) and (20.138)
Dµ'l/1--+ eiei\/cñ.Dµv;(x) (20.139)
ensuring the gauge invariance of Eq. (20.137). Now
[Dµ,Dvlf(x) = [8µ + :~Aµ,8v + :~Av] f(x)

ie ie e2
= [8µ, 8v]f(x) + en [8µ, Av]f(x) + en [Aµ, 8v]f(x) + e2n2 [Aµ, Av]f(x)
so that
ie íe
[Dµ, Dv]f(x) = en [8µ, Av]f(x) + en [Aµ, 8.,]J(x)
ie ie
= enf(x)[8µ,Av - ovAµ] = enf(x)Fµv·
Hence we have (since f(x) is arbitrary).

ie
[Dµ, D.,]= efiFµv· (20.140)
In terms of momentum operator TÍµ
¡5µ = iñAf ~Mµ + ~ G}) Aµ

~'!.Aµ= Pµ - -Aµ/r'I
e e
The Dirac Equationl (20.135) is
[1µ (fµ - ~Aµ) - me] 1/J = O. (20.142)
Let us take the hermitian conjugate of Eq. (20.135):

( -ifi( 8µ1/J t) ('Yll) t - mcu: t (x)) = O.
Multiplying on the right by 1° and using 1µt = 'Yº"fµ'YO,
(-ifi(8µi¡j)'Yµ - me-0) = O
which is the Dirac equation for the adjoint spinor i¡j(x). The above equation
can be written in the form
- +-
1/J[-ifi8µ1µ - me] = O (20.143)
and in the presence of electromagnetic field:
ijj [ (-ifi8µ - ~Aµ) 1µ - me] = O. (20.144)
We now derive the equation of continuity. We have already seen that

(20.145)
is a vector. Now ifwe multiply Eq. (20.135) on the left by ijj and Eq.(20.143)
on the right by 1/J and subtract, we obtain
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)
where p is the probability density. The Eq~.146) is equivalent to the -#-

equation of continuity. This can be seen as follows:
18 ·o
a a
µJ·µ -- oJ·o+{), ,J ·i -- _e _J_ +v.. at
~
= ! 8p + .=o (20.149)
e at
where
or
and
i" = cif;,º'lf; = c'lf; t 1/J = cp ;:::: O
i.e. the probability density is always positive and is delinked from the
problem of negative energies.
20.10 Constants of Motion
With the help of the Heisenberg equation of motion, we can determine

whether an observable is a constant of motion or not. Consider the Dirac
!
r )
Hamiltonian for a free spin particle:
~ = 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
1 ow the orbital angular momentum
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)
i.e. the orbital angular momentum is not a constant of motion in sharp

contrast to the corresponding situation in Sch1~dinger theory. Now let us 'Í ~
consider J..'
u=(~~)·
Then
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
On the other hand
Therefore
Hence
,~dui 2'
dt = 2CE·iJ"kP J·ak
A
in-
or
ñ dc A
-- =-ca xp. (20.154)

2 dt
Thus the operator u for an electron in a central field is not a constant of
motion. However, it follows féíln
Eqs. (20.153) and (20.154) that
d 1
dt (L + = O,
2ñu)
i.e,
d
-J=O, (20.155)
dt
where
1
J= L+
2ñu.
Thus although L and ~ ña taken separately are not constants of motion,
the sum
is a constant of a motion. We can interpret this result by saying that the

1
electron has a spin angular momentum S = ñu, which must be added to
2
the orbital angular momentum L, so as to get total angular momentum J
1
which is a constant of motion for a spin
2 particle in a central field of force.
( 20.11 fyelocity' in Dirac Theory
For a free particle, the Hamiltonian is given by

H = ca.p + {Jmc2.
Using the Heisenberg equation of motion, we have
ifixi = [xi, H]
= [xi, caj'Pj + {Jmc2]
= cajiMij
= iñco¿
or
which shows that a. is the velocity of particle in terms of c.

This result is rather surprising, as it means an altogether different relation
between velocity and momentum from what one has in classical mechanics
(v = c2p/E). Note, however, that the eigenvalues of ai are ±1 [as a;= 1].
Thus Xi has eigenvalues ±e corresponding to the eigenvalues ±1 of ªi· We
conclude that a measurement of a component of the velocity of a free elec-
tron is certain to lead to the result ±c. We may also note that because
aj and ak do not commute for j =f. k, a measurement of the x component
of the velocity is incompatible with a measurements of the y component
of the velocity; this may appear strange as we know that p1 and p2 com-
mute. In spite of these peculiarities, there is actually no contradiction. The
theoretical velocity in the above conclusion is the velocity at one instant
of time while the observed velocities are always average velocities through
appreciable time intervals. We shall find upon further examination of the
equation of motion that the velocity oscillates rapidly about a mean value
which agrees with the observed value.
We have
ifia.i = [ai, H]
= 2aiH - (aiH + Hai)·
But
aiH + Ho¿ = c[aiaj + ajai]'Pj + mc2(aif3 + {Jai)
= 2cóij'Pj
= 2cf3i·
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 ~
c2piFI 1 andan o cillatory part of tí·equency 2H/h~ 1 known a· z1tter-/<tN).\

bewegung term. This frequency is at least 2mc2 / Ii and hence high. The first
part of the velocity is rea.sonable since, for an eigenstate of momentum and
energy, it gives c2Pi/ E, in agreement with the classical relativistic formula.
Taken literally, the second part shows that the velocity of the electron has
an additional term that fluctuates rapidly about its average value in the
absence of any potential. Only the constant part would be observed in a
practica} measurement of velocity, such a measurement giving the average
velocity through a time-interval much larger than n/2mc2• The oscillatory
part insures that instantaneous value of Xi shall have the eigenvalues ±c.
As for the coordinate operator, the relation can be easily integrated to give
xi(t) = xi(O) + c2piH-1 + i~n(ai(O) - cpiH-1)H-1e-2iHt/1t. (20.160)

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.
20.12 Two Component Pauli Form of Dirac Equation: Ex-

istence of Magnetic Moment
The Dirac equation in the presence of electromagnetic field is given in Eq.

(20.137):
(20.161)
The Kleinfordon equation in the presence of electromagnetic field is
(n.2 Dµ,Dµ + m2c2)</> = O. (20.162)
The equation (20.161) can be transformed into second arder equation sim-
ilar to Klein~ordon equation (20.162) by multiplying on the left with
(iñ," D.,+ me), (20.163)
so that
(-ñ21µ,,v Dµ D., - m2c2)1/J =O (20.164) \
Using the anti commutation relations (20.159) and o+" given in!Q_-~.161) ~p -.:tt
One can reset the above equation as
Using the relation

[ñ2DµDµ - ~ñ2aµ"[Dµ,D.,] +m2]1/J
::;,
= 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
But "fº'i = ci and

Foi =(8oAi - OiAo)
=Ei
and O"ij = EijkEk so that
i-i1 i "k
(J" Fij = E 1 E k (8i)Aj - OjÁi)
= -~.(V X A)
=-~.B
Thus in terms of electromagnetic field E and B the Eq. (20.167) is written

as
(20.168)
where
(20.169)
Now in the Pauli representation o: is non-diagonal while every other term

is diagonal. It is then preferable to use chiral representation for "(-matrices
given in Eq. (20.64), where
(20.170)
and
a
i = 'Yo 'Y = (-(J"iO O"io)
i
(20.171)
are both diagonal. Then in terms of two components wave functions

<PL, <PR:
(20.172)
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
In order to interpret these terms and see their physical significance we first
consider only the radiation field:
Aº= O, A=/= O, B = V x A and
We take A independent of time t, so that E= O and the equation (20.173)

E~ - a: L (20.174) o
reduces to
(20.175)
where D2 = (V - Á:A)2 and '1/J is </JL or </JR·

Now a mode with energy E = mc2 +E should oscillate in time as e-iEt/li.
In the non-relativistic limit, kinetic energy E « mc2, so that for a slowly
moving particle:
'1/J(r, t) = 'l/J(r, t)e-imc2t/li. (20.176)
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.
Dropping :t22 compared to 2imc2 :t, equation (20.175) becomes
2
in81 = -- 1t2 ( V--Aie ) en
--a.B ) 'lj;. (20.177)
at ( 2m ne 2mc
Now
But
V.A'lj; = (V.A)'l/J + A.V'l/J

=A.V'lj;
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.
20.13 Dirac Equation for the Hydrogen-like Atom
Now consider the Hydrogen atom for which

A=O, B =0
so that
Dµ. D
µ.
= Dº2 - 02 = _!__
c2
(!!_8t + i:.A
n 0)2 - V2. (20.179)
Here Aº is the Coulomb potential Aº= _z:_ so that

r
eAº = -Z-er2 = -Z-,ar

and
E= -VAº= - dAº f , (20.180)

dr
W Then equation (20.173), on using Eqs~0.179) and (20.180), takes the form
[ ;: ( :t + i; Aº r- n2V2 ± i:n d:º u.r + m2c2] <PL,R = 0. (20.181)

For stationary states

.i;
'l"L,R = e -iEt/ñ</>L,R ( r ) '
where
<PL,R(r) = Yim(O, </>)RL,R(r).
Thus equation (20.181) [see ._
sec. 7.1] becomes
eEAº
[ 2 ------,i
c2
e2 A~
c2
,.2t"'7v2 ±i---u.r-
..... en dAº
e dr
A (E 2 - m2
c2
c4)] R LR ( r ) _- 0
'
(20.182)
where
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]
= [L', akxk] + ~[a', 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,
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)
Thus in the above subspace the operator
* }(~n Eq~0.182~
ñ,2 ,,
ñ
2 ( L2
¡¡,2r2 -
represented by the 2 x 2 matrix
(Hf )(i+~) _ e.2

/L&
A2
0
e2
fic2 Ao ±
2 ie dAo
ne
± ~ie
-)
dr a.r
!!&
1L1-
,
)
(20.186)
(20.187)
( ± ie ~ (j-!)~i+4J _ e2 A2
lic dr r li-~c'i O
which has eigenvalues >..ñ2 given by
The solution of the quadratic equation gives
)... _ ( J. + 21) 2 _ ~ 2 A2 l [(.

± J + 21) 2 2
__ e_ ( dAo ) 2] 112
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
where
which may be written as

[ (; + ~ )' - Z'a' r- 1, (20.192)
with
(20.193)
Thus our equation (20.182) becomes
[
n2 (~ + ~~) + 2 z ne2 E+ (E2 - m2c2) - >.(>. + l)n2] R(r) = O.
8r2 r ar c2 r r2
(20.194)
Writing
R(r) = x(r), (20.195)

r
we have
n2 d2x [E2 - m2c2 2ze2 >.(>. + l)n2]
2m dr2 + 2m + 2mc2r E - 2mr2 x(r) = O. (20.196)
Thi is formally identical with the equation for Hydrogen-like atom in

Sch:@.unger theory i.e. [see Eq.(7.16)]
!!!__ d2x ( Ze2 _ f(f + l)n2) _
2m d r 2 + E + r 2 mr 2 X - O, (20.197)
with energy eigenvalues,

z2e4
E= -m2n2n2i ti = n +e+ 1 is a positive integer,
if
E
e2-----, --e2
mc2 l )
E
E2 -m2c4
2rrw2 L )
f. -1 5. = (.1 + ~) - Ój
an¡i
n - (f + 1) --+ n - (>; + l)~
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
]
where n = 12, , ... J· and Js

... , n - 2.1 = 2,13 2,
We note from Eq (20.193) that s.
becomes complex for Za> j + ~- This
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)
For Hydrogen atom Z = 1,

ma2c2 me4 e2
Eº
n
= ----2n2 = ---
2ñ2n2
= ---
2a0n2
(20.201)
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.:
[E(2p3¡2) - E(2P1¡2)] = mc2 ~; = 4.53 x 10-2eV = 10.l,g..Hz, (20.202) #="

in agreement with experiment; on the other hand states with same n and
j (e.g. 2s1¡2 and 2p1¡2) are still degenerate (Lamb shift). This requires
interaction of electron with quantized radiation field.
20.14 Hole theory and existence of Positro~ti-matter) _S

We have already seen in ec. 20.9 that there is no trouble with the prob-
ability density as it is al~ays positive definite. But the interpretation of
negative energy problem which the Dirac theory admits remains. The diffi-
culty with ths negative energy states is sol ved by Dirac's Role theory. ·ir t
we elaborate what is th difficulty? Consider[&drogen atom. An lectron
bound in an atom can emit radiatíon spontaneously and thus en- love
ergy. It can make a radiative tran ition and fall into the contlnuum of
energy of negative energy states by spontaneously emitting a photon of
energy ~ 2mc2. But the ground state of atom in stable. Clearly such a
catastrophic transition must be prevented. Dirac solved it by giving a new
picture of vacuum regarding it an infinite sea of electrons occupying nega-
tive energy states (see figure 20.1). The Pauli principle then forbids such a
transition. We see that E > O is empty in the vacuum but E < O is filled
"\.
\.
mc2
\onllnuu m of Positiva Energy States
- ,_º
I
J
Continuum of Negativa Energy States
Sz=-112 Sz=1/2
Fig. 20.1 Dirac's picture of the vacuum
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
electron due to charge conservation). This is what happens when a radia-

tion quantum of high energy (~ 2mc2) disappears with this process, giving
rise toke+ pair. To describe the above phenomena, Feynman regarded
~o itron and electrón as essentially one pa:ticle,differing only in the sense
in whlch they move with re ·pect to tim see · 20.2). A positron was
an electron moving backward in time. Int ·raction wi.th radiation quantum
turns the path of e+ around and then it appears as an electron.
!Y\ l
Fig. 20.2 Feynman's picture of electron and positron
20.15 Free Particle Solutions, Plane Wave Solutions of

Dirac Equation
Dirac equation for a free particle is given in Eq~.56)
(in,µ flf) - mc)'I/J(x) = o. (20.203)

uxµ
Each component of Dirac equation satisfies the Klei~Gordon equation
( e~ !22 - v72) '1/J(x) = m2c2'1/J(x) (20.204)
The plane wav olution of E<t~-204) is

'l{J(x) "'e-,.p'':i:,. = e-/¡(Et.-p.x)
L
(20.205) ~
Thus th plan wave solution of Eq.g0.203) can be written as
'1/J(x) = u(p)e-iv"x,, = u(p)e-ifñ(Et-p.x). (20.206)
Hence ·twyn Eq.g0.203):
(/p - mc)u(p) =O (20.207)

where
(20.208)
Now u(p) has four components:
(20.209)
Thus Eq.(20.207) is matrix equation: L

(/p- me)0¡3u¡3(p) = O
L
(20.210)
D
The hermitian conjugate of Eq.(20.207) is
utp[(,,µ)tPµ - me]= O
By multiplying the above equation by 'Yº on the right, writing ut(p)'yº as
u(p) and using 'Yº'Yµt 'Yº = 'Yµ, we get the equation for the adjoint spinor
u(p):
o
u(p)['YµPµ - me]= u(p)[/p- me]= ol (20.211)
()
First, we note that for a particle on the mass shell:
p2 = PµPµ = p~ - p2 = m2e2.
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
/p = 'YµPµ = 'Y0Po - 'Y-P = ( Po -u.p)

ap =Po
l (20.214)
o
Thus from Eqs. (20.207),(20.213) and (20.214), we obtain

(E - mc2)uA(P) = C<T.puB(P)(
(E+ mc2)uB(P) = cu.puA(P)L

In the rest frame of the particle (p = O), we have
(20.215a)
(20.215b)
ú '
E= ±mc2
and from Eqs. (20.215), we get
Thus for E = mc2:

(E - mc2)uA(O)
(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)
where x(r) are two component Pauli spinors: r = 1, 2,
xC1l = G) : Spin up a3xC1l = xC1l, xC2l = (~) : Spin down
The four solutions in the rest frame are:

E=mc2:
uC>l(o) ~ (x~'') ~ m ,Spln up~
u<'l(o) ~ (x~'') ~ m Spin down l o

E = -me2 : (20.219)
uC3l(o)- U,i)-m .Bpin up ( )
Thus
u('l(o) - U,)) - m 'Spin down L o
and ( O ) eimc2t/li. (20.220)
xCll
are positive and negative energy solutions for spin up and
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
Written in full in Pauli representation of ,-matrices, we have

L (20.222) o
u(p) = N' (pº + me -u.p ) u(O) (20.223)
ap -pº +me
where N' is a normalization constant. Thus the four independent solutions
are:
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)
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)
callad he helicity operator, commu es wíth the Dirac Harníltonian. ow

from Eq.~0.224)
uCr)t (p)u(s) (p) = INl2 ( x(r)t, E:~;;c2 x(r)t) ( cu\(s)x(s))
=
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
We normalise the spinors, so that

uCr)(p)u(s)(p) =0~
=19 ~~~,;4 (20.229)
I
L cJ
This normalisation gives
INl2 = (Ev + mc2)
2mc2
Selecting the phase, so the N is positive
N=. /Ev+mc2
Y 2mc2
L (20.230) G
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
u<rl(p) =. s; +mc2 ( -ca.p 2 x<r))

Ep+mc r = 3, 4, E < O (20.232)
2mc2 x<r)
Finally, we have
1/J(x) rv u<1,2)ei/!i.(-Ept+P,x), 4' >O (20.233)
rv u(3,4)ei/!i.(Ept+p.x), {E< Q (20.234)
Change p ------t -p in the negative energy solutions and write
v<ll(p) = -u(4)(-p)
v<2)(p) = u<3)(-p)
then we can write the above solutions as
1/J(x) = 'U(r)(p)e-i/hp''x,-, ~-=1,2 E> ol
'1/;(x) = vM(p)i/llp":1:,., \1· = 1,2,E < O. (20.235)
Without any ambiguity, we take E to be positive i.e. E= Ep, so that
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) )
(/p + mc)v(r) = O. (20.241)

From Eq.(20.229), it follows that u(r) and vCr) are normalised as follows
U(r) (p )u(s) (p) = s.. L }
v(r)(p)v(s)(p) = -s: L )
(20.242)
The spinors u(p) and v(p) also satisfy the completeness relation (in spinor
space):
E;=1[utl(p)uy\p) - vtl(p)vt)(p)] = 8af3· (20.243)
In matríx notation:
(20.244)
Define the projection operators:
A±(P) =A±= me± /p (20.245)

2mc
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).
20.16 Charge Conjugation
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:
[-ifi'YJJ.* ( 8µ. - ~:Aµ.) - me] 'l/;* = O (20.250) ()

By multiplying Eq.(20.250) by -i,2 on the left, we get
r: ( 8µ. -
-i'y2 [ ( -ifi, .. ~:Aµ.) - me] 'l/;* = O
or [(,y2)2 = -1]
-i,2 [in')'JJ.*'Y2'Y2 ( 8µ. - ~:Aµ.)
First, we note that
'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)
Thus if the Dirac equation is to be invariant under the charge conjugation,

then
[in'Yµ. ( 8µ. + ~:A~) - me] 'l/;c = O. (20.254)
Now the Eq. (20.254) is identical to Eq. (20.253) provided that

(20.255)
Note that A~= -Aµ., as it should be, since
8"8,,Aµ. = i; (20.256)
and electromagnetic current Jµ must change sign under charge conjugation.

The matrix -h2'Yº is called the charge conjugate matrix and is written as
(superscript T stands for transpose]:
C = -Í'y2'YO, c 2 = -1 L. )
et = -c = e" = c-1 L )
e-re:' = -('Yµ)r. (20.257)

Then we can write '1/Jc given in Eq. (20.254) as
'1/Jc = ciV. (20.258)
Now under charge conjugation, '1jJ-----+ '1/Jc, so that for plane wave solution
(20.259)
implying
(20.260)
In ~ection 20.7, we have discussed the discrete symmetries ~ space reflec-
tiorf and time reversal, Under space reflection
'ljJ(t, x)-----+ '1/JP(t, -x) = 'Yº'1/J(t,x)
(20.261)
where
,ºu(p) = N G ~1) (E~Jc2x)

=N (_ c;;-.p2 ) = u(-p).
E+mc X
Similarly
'Yºv(p) = -v(-p) (20.262)
i.e. particle and antiparticle have opposite parities.
Under time reversa!
(20.263)
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
. 2 (1)(º -1) (1) = (º) =

= 1 O (2)L
ur X O 1 X
'
_.
ta X
2 (2) = (º -1) (1) = _ (1) = _
1 O O O X
(1)
. (20.265)
Thus
-ia2x(r) = rx(-r), (20.266)
where r = ±1 corresponds to ~ (spin i) and -~ (spin l) respectively.

Hence
(20.267)
so that under time reversal
(20.268)
20.1 7 Large and Small Bilinear Covariants
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)
in[!~ (<!>A) + ( O u.V) (<PA)] _ me (<!>A) = 0

e at <PE -u.V o <PE <PE
i.e.
inn.V</JE = _!(E - mc2)</JA, iñcr.V</JA =--(E+

1 2
me )<PE·
o
e e
Thus for
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
= (<P~<P1) (~ -~) (pt~ )

~ Ó~ÓA - OtÓB ~ Ó~ÓA ~ j/,~j/,A + O(v' /é') l {20.272) )
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
= (<P~<P1) (~ ~1) (~k -~k) (!;)

= [</J~rrk</JA + </J1uk</J1]
= -i{J')'5')'k'i/J L- (20.274)
)
i[Jrrij'if; = Eijk(<P~<P1) G-~) (~k :k) (pt~)
= Eijk[<JJiuk<PA - <JJ1uk</JE]- (20.275)
All the above bilinears are seen to be 'large' if we neglect terms of order
(v/c)2• Thus
(20.276)
(ijk cyclic). (20.277)
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
= (<P~<Pt) G ~1) (~ ~) (::)

= -(<P~<PB - <P1<PA) ~ O(v/c). (20.278)
20.18 Weyl Equation
Recall from Sec. 20.4, that in the chiral representation of -y-matrices, we

can write the four component wave function 1/J as
~ = (!~) (20.279)
so that using q~20.6'1); the Dirac Eq~0.56) becomes
. ., {j
i,i 8x0
(<lcp¡,
>n) + ui"" ·( -u.V
o u.V)
O
(<PL)
<PR - me (<PL)
<PR =
o· (20.280)
Thus <PL and <PR satisfy the coupled equations
[iu. V - 8~J ó f\- ~e L

i ÓR (20.28la)
[-iu.V-ia:,J ·~- L, ~cÓL (20.281b)
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 =
1, 2, 3] one verifies that, if c/JL(x) is a solution of Eq. (20.282a), a2c/JÍ., is

a solution of Eq. (20.282b). The above equations are known as the Weyl
equations for massless spin 1/2 particles. Now [(l=t='Y5)/2]VIrespectively pick
up c/JL and c/JR· These are known as the chiral projections of VI· We further
note, that for a massless spin 1/2 particle, the Dirac equation becomes
o
'Yµ -¡:;-VI = o. (20.283)
UXµ,
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
Hence we conclude that a 2-component Weyl wave function c/JL(x) satisfy-

ing Eq. (20.282a) or equivalently positive chiral projection [!(1- 'Y5)Vl(x)]
of a 4-component wave function Vl(x) satisfying the Dirac equation for a
massless particle describes a left handed (negative helicity) spin 1/2 par-
ticle and a right handed (positive helicity) spin 1/2 antiparticle. In an
entirely similar manner one verifies that the wave function cp R or equiva-
lently [(1 - 'Y5)/2]V1(x) describes a right handed particle anda left handed
antiparticle. We know that the neutrino mass is consistent with zero. If

we take its mass to be exactly zero, the above results are relevant to the
neutrino (antineutrino). A theory ofneutrincJ.Eased on <PL =f. O, <PR = O (or ~
vice versa) is called the two component theory of the neutrino. Actually in
nature, the helicity of neutrino is determined to be negative and thus we
only have the first possibility.
Problems
20.1 In the presence of electromagnetic field, Dirac Hamiltonian is given

by
H =ca· (p- :.A)+ eAo + {3mc2•

e
Write
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:
,yl-L('y · a),µ, = -2,y · a

,-,¡l-'('y · a)('Y · b),µ, = 4a · b,
where summation over the µ-index is implied and where
the summation convention is being used.
20.3 Using the anti-commutation relations for -y-matríoes, show that
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
20.4 Show that in the rest frame of kn ~ particle, the vector
_ 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
')' .
Considera 'Pure Lorentz Transformation' in which the primed sys-

tem is moving with velocity v in the direction x3-axis and that x11,
x'2 are parallel to x1, x2
x'1 = x1
x'2 = x2
x'º = x0 cosh w - x3 sinh w
x'3 = xº sinhw + x3 coshw,
where
coshw = (1 - {32)-1/2
. h {3 {3 = 1}__
sm w = (1 - (32)1/2' e
Take w to be infinitesimal and find S for the above transformation.
From this show that for finite w, S can be written as
S = coshw/2 - i')'3,·y° sinhw/2.
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
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
·
Using only the anti-commutation rules for Dirac matrices a1, a2

3¡¡,2
and a3, show that the electron has spin 1/2 i.e. show 82 =-
4
20.7 The large and small components of the solution of the Dirac equa-
tion
(Í'YµPµ + mc)'f/J = V'l/J

!
are defined as '1/J A = (1 ± ,º)'1/J.
B
(We choose a rep. in which ,o
= {3 is diagonal.)
Obtain the equation for 1PA and show that when the kinetic energy,
T, of the particle (cp0 = mc2 + T) is such that T << mc2, 'l/JA re-
duces to the non-relativistic Schrodinger wave function multiplied
by e-imc2t/fi.
20.8 Show that, using the Dirac equation,
where s = 1,2 is the spin index. Show also that
o
while
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.
20.11 Show that under charge conjugation
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
20.12 Consider the coupled 2-component equations

. ,,,8 me _
UíL µ,</JL- /icc/JR-Ü
where
Write the above equation in the 4-component form
( i,',,,8,,, - ":ic) 1/J' = O

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 .

Chapter 21
DIRAC EQUATION IN (1+2)

DIMENSIONS: APPLICATION TO
GRAPHENE
21.1 Dirac Equation in (1+2) Dimensions
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
and as usual ['yº is hermitian, ')'1,

bµ,'Y"}
')'2 are ant~rmitian]:
= 2gµ" L
(21.2)
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
They are inequivalent because they cannot be related by a similarity trans-

formation. This in fact is a feature of an arbitrary number of odd time
o
space dimensions.
'vVe now discuss the role of the parity operation vis a vis above
repr sentations of the 'Y matrices. In (1+3) dimensio~ S
tw[i@r¡ i)
xP _. (xº, -x1, -x2 -x3); recall for the Lorentz group defíned by th
matrix A, for L¡, ~ = +l and consists of three rotations and three
boosts. For the parity transformation, which is a discrete transformation,
1 O O O)
A= O -1 O O
( O O -1 O
O O O -1
n (1+2) dimensions, there exist only one rotation an~voboosts. The

as a discrete symmetry (not rotation) which has et~= -1, can be
parity ,.....,,..
defined as G
A=
(O
Oº)
O1 -1 O
O 1
or (~~~)L
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,
(i$ - m)'I/J+(x) = (i(o-380 + io-181 + io-282) - m) '1/J+(x) = OU21.5) o

If, however, we use the second representation in Eq. (21.3), we have
(i$- m)'I/J-(x) = (i(o-380 + io-181 - io-282) - m) '1/J_(x) = O LE.1.6) o

Now if Eq. (21.5) is to be invariant under parity, we should have
(i(o-38f; + io-laf + io-2~) - m) '1/J~(xP)

Now since 8~ = (80, 81, -82), there is no matrix M that simultaneously
= O L (21.7) o
anticommutes with o-2 and commutes with o-1 and o-3, therefore we cannot
find a relation of the type
(21.8)
DJRAC EQUATJON IN (1+2) DIMENSJONS: APPLICATJON TO GRAPHENE5l9
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)
where r¡P is a phase factor, lrJpl2 = l. Thus to preserve parity, we have to

take the two representations together. To put Eqs. (21.5) and (21.6) in
symmetric form, we transform to new 1/J' s:
i.-
Then Eqs. (21.5) and (21.6) take the form.
1/JA 1/J+=
'lpB = i'y2'l/J_ L (21.10) o'
(i{p - m)'l/JA = Ü
(21.11)
(i{p + m)'l/JB = Ü
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.
21.1.1 Dírac Equatíon in Hamiltonían form and Plane

Wave Solutions
We can write Eq. (21.1) in the formn [n = 1, e= 1]
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
= (-u2p1 + u1p2 + mu3) (21.14b)

Note that when m = O, the H is not identical with ug:f>, or uc-3'v, which is the
case for 2-component Weyl spinor. Since in aplane, as already noticed we
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,
UA(P) = u1(p) = (p2o~ipl)

p +m
pº =E> O (21.20)
' Pº = -IEI < o (21.21)

DIRAC EQUATION IN {1+2) DIMENSIONS: APPLICATION TO GRAPHENE52l
Thus for antiparticle spinor
(21.22)
We note that for p2 = O = p1,
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
for particle (electron), antiparticle (hole) spinors. Since 'l/JB is related to

'l/JA by the parity operation (21.12), we have
= 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
Fig. 21.1
realization of a single layer of graphene has led to extensive exploration of

electronic properties in this system. Experimental and theoretical studies
have shown that the motion of quasiparticles in these 2D system are very
different from those of conventional 2D electron gas systems realized in the
(\semiconductor heterostructures.
~\v- .: "t'1n fact graphene is a unique system in many ways:

(A): It is a semimetal. Since it is a 2D system exposed to the environment, it
can be easily modified chemically and/or structurally in arder to change its
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.
21.2.2 Electronic properties of graphene

The starting point in studying the electronic structure of graphene is the
tight binding approach, which we briefly discuss. Graphite, a 3D carbon-
based material, has 4 valence electrons: 3 of these in 2s, 2px, 2py states form
tight bonds with neighboring atoms in the plane and as such will not play
a part in the conduction. The fourth one is in 2pz state. We therefore treat
graphite as having one conduction electron in the 2pz state. In graphene,
the in-plane sigma(a) bands are formed from the 2s, 2px and 2py orbitals,
hybridized in a sp2 configuration, while the 2pz orbitals, perpendicular to
the layer, form the out of plane pi (1r) bands. The a bands give rigidity
DIRAC EQUATION IN {1+2) DIMENSIONS: APPLICATION TO GRAPHENE523
.--------.
,, ''
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)
a = l.42Aº is the carbo~arbon distance. The reciprocal lattice vectors

are given by
27í . r,, 27f . r,,

b¡ = (1, v3), b2 = (1, - v3). (21.26)
3ª 3ª
The first Brillouin zone is a hexagon [see g 1.3]. Of particular importance

for the physics of graphene are 2 points K and K' at the corners of the
b f .•,·, .... ~
.
/
..-- . ..l
-· --
...... ·
.. -··...
...
----~
. I
t
•• ••
-.
.~· K
....
o lf
'·\ .,,.
··--.................. _...
\ .- .-·· n
.
-.;. '•,
·'"\,,,. -··· ·...
·'
Fig. 21.3
graphene's Brillouin zone. Now
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) ()
DIRAC EQUATION IN {1+2} DIMENSIONS: APPLICATION TO GRAPHENE525
The tight binding Hamiltonian of the electrons in graphene is

1{ = E ¿(at(Rj)a(Rj) + bt(Rf)b(Rf))
j
-t ¿(at(Rf)b(Rf) + h.c) (21.29)

(i,j)
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
b(Rf) = ¿ bk ei~. (21.30)

k
Thus we can write the above Hamiltonian as
H = ~ ( al bt) H (::) ·
where 1{ is
H=
t.
-t ~
-t ~
eil(;rje=['
e;'t)', )
Here u, is a set of triad of vectors connecting an atom A with its B nearest

L (21.31) ~" 1t /o
neighbors, in our case 61, 62, 63 [see fig 21.2] and vi the triad of their
respective opposites. e is the 2pz energy level, taken as the origin of energy.
Graphene's electronic states are described as a linear combination of atomic
orbitals from the two atoms forming the primitive cell O(A) and e(B) [see
fig 21.2]
(21.32)
The eigenfunctions and eigenvalues of the Hamiltonian are obtained from
the equation
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.
~'-".>
where é and t are found by fitting experimental or the first-principles data.

The most common practice is to adjust the tight-binding dispersion to a _J.
correct description of the pi-bands at the K point. This yields E: = O, t~ 2.8 if
eV). Thus, using Eqs. (21.28) and (21.34)
E±(k) = ±t{3 + cos kyv r;ua + 4cos kx23a cos ky-2-}

v'3a 1¡2
= ±ty'3 + f(k). (21.35)

It is instructive to calculate f(k) clase to one ofthe Dirac points: k = K+p,
p is the momentum measured relative to the Dirac points
!( k) "' -3+ i•'(P! + p;)

Thus the energy band around the Dirac point is
L o
(21.36)
where v¡ = 3;ª represents the Fermi velocity. It is seen that the disper-
sion is conical [fig 21.4]. The most striking difference between the result
'
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
DIRAC EQUATION IN (1+2} DIMENSIONS: APPLICATION TO GRAPHENE527
depend on the energy or mass. The energy dispersion (21.38) resembles

the energy of ultra-relativistic particles, (although v¡ is 300 times smaller
than the speed oflight); these particles are quantum mechanically described
by the massless Dirac Equation as we shall see. The most interesting as-
pect of graphene is that low energy excitations are massless, chiral Dirac
Fermions. This difference in the nature of the quasiparticles in graphene
from conventional 2D electron gas system has given rise to a lot of new
and unusual phenomena such as anomalous Hall effect in graphene. The
transport experiments have shown resulta in agreement with the presence
of Dirac fermions. The 2;(p Dirac@<e speetrum was confirmed by cyclotron
resonance measurements which we describe below.
21.2.3 Cyclotron mass

The cyclotron mass is defined, within the classical approximation as
* 1 8A(E) (21.37)
m = 21r --¡¡jj;
where A(E) is area in k-space enclosed by the orbit. Using the Energy
dispersion E2 = vjp2
E2
A(E) = 1r-2
V¡
so that
m*=~ L
One may contrast it with Schrodinger dispersion E = frn which would
2
imply a constant cyclotron mass. The electronic density, n, related to the

(21.38)
o
Fermi momentum is
k2
n = ____E_
7r
which implies
m" = yin -Jñ,. (21.39)
v¡
Fitting the above equation provides an estimation to th Fermi velocity
[v¡ = 3at] and the hopping parameter t respectively [see g 2l.5:
v¡ ~ 106ms-1 and t ~ 3 eV.
The experimental observation of the ,jn dependence of the cyclotron mass,
as shown in Fig 21.5, provides evidence for the existence of massless Dirac
quasi-particles in graphene.
I July 9, 2012 14:29 World Scicntiflc Book - 9in x 6in RF-QuanturnMechanics [
528 Qua.ntum Mec!&aníc.s
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,
21.2.4 Dirac Hamiltonian omd Dirac Equation for Quasi-

particles in Graphene
In Eq. (21.31), we may take u, or v1 as two independent lattice vectors
e.g a¡ and a2 as given in Eq. (21.25). Thcn Eq. (21.31) gíves
H = L ( at bt) 7-l-D (::)

k
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
DIRAC EQUATION IN {1+2) DIMENSIONS: APPLICATION TO GRAPHENE529
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•
1-lv'l/J± = E'I/J±, E= ±v¡ IPI (21.44)
or in its time dependent form
or, on using Eq. (21.43),
/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
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
p - pP = (pº ,p1, -p2):

p1 +-----+ p1,p2 +-----+ -p2 and riK +-----+ riK'·
Thus the parity operation connects Dirac Hamiltonian around the Dirac
point K and K'.
21.2.5 Tunneling and Klein Paradox

We consider the scattering of massless Dirac forrnions in two spatial dimen-
sions by a square barrier shown in ig 21.6). . be Dirac equation is
( vl¡ 'Yº(i ! - Va)+ i"}'1éh + i"}'282) '1T(x) = O (21.48)
where 'Yº = a-3, "}'1 = ia-1, "}'2 = ia-2. Now
A Energy
y o
---
/JI
D I
-,//
(~-> 1.:~~-- ¿_1.
Fig. 21.6
w(x) rv e-ip.x u(p)

= e-iEti
= e-iEt
PO-
'1/J(x)
Then from Eq. (21.48), by replacing pº ------; E - Va, the solution of Eq.
e
u(p)
(18.49) can be written
where [see Eq. (21.20)]
(21.49)
where the energy dispersion is

E= ±lplv¡.
We can write
Py=~
where
~=Py.
Px CJ'
Then the wave functions in the three regions can be written as
'1/J¡(r) = _!..._ (. 1. ) ei(p.,x+v,,y)

y'2 ise•c/>
+--2:._ ( .1 ) ei(-pxx+p,,y) (21.51a)
y'2 ise•C1r-c/>)
where s is sign(E). For regions II and 111:
'1/Ju(r) = ___!!'.____ (, 1. ) ei(q.,x+p,,y)

y'2 is'e'9
+~ ( .1 ) ei(-q.,x+p,,y) (21.51b)
y'2 is' e•C1r-9)
, ;).,-- ,Pm(r) ~ ~ (;,!i>) e;(,,x+,,,) (21.5lc)
)6 where s' = t:.i.91 (E-Vo) ta ifJ =~ qx = -jq -vi= J(Vo - E)

2 2~ -pi.
The coefficients r, a, b and t are determined from the continuity conditions
'1/)1(x = 0,y) = '1/Ju(x = 0,y)L )
'1/Ju(x = D,y) = '1/Jru(x = D,y). (21.52)
If we define A± = a ± b, then the above conditions give

l+r=A+L )
i + e-ic/> -A+e-ic/> = {[iA+sinO + A_cosB] ~ )

s
A+cos(Dqx) + iA_sin(Dqx) = teiDp., l_. ,)
1
~[iA+sin(O + Dqx) + A_cos(B + Dqx)] = tei<l>eiDp.,. (21.53)
s
The last three equations can be put in matrix form
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)
_ cos2 IJ cos2 </J

(21.57)
- [coslJcos<jJcos(Dqx)]2+ sin2(Dqx)[l - f sin sin é']? ·
é
( ~ Notice the following:

i}
'f~
r (l)'JFor Dq¿ = mr,n an integer, T(<f¡) = T(-cp) and T(¡f,) = 1 independent
of the value of ,P and~1'tier becomes compl ely transparent.
DIRAC EQUATION IN (1+2) DIMENSJONS, APPLICATION TO GRAPHENE~ {)
0.1
~0.6
0.4 v.,=zoomevl
V =21l'ámeV
0
O.?
-75 -IX> -4S -30 -rs u se -1$ ~o

,..
O.! I
,,
~o.e I
/
I
~ 0.4
;
~ I
0.2 _.., ;
qgo - -'15 -30 -15 o 15 90

.angre+
Figure IEi.7 (Color online) Angular b eha'vior of T(cp) for two

diflerent yaJues of \'ó: Vi, = 200 llleV dashed line, Vo = 285
llleV scrlirl l ine , The rernaining pararnet.ers are D = 110 nrn
(top), D = 50 nm (h<>tt.om) E - 80 meV, kF = 2~/>.., >.. - 50
Cllll.
(2)1y For normal incidence O, () -t O, T(O) = 1 for any value of Dqx.

</> -t
Tl;·rs result is the manifestation of Klein paradox and <loes not occur for
non-relativistic particles. In this latter case for normal incidence, T is
always smaller than one.
(3)'(\In the limit IVol >> IEI, () -t O and
{dis'j<jJ
T(</J) ---4 I -~Z(Í)qx'f.§2<!>· (21~
The angular dependence ofT(</J)is plotted in Fig (21.7), which shows that
there are several directions for which transmission is one.
The above result has several applications, e.g. in filtering electron

beams01 transmission through graphene quantum dots.
21.2.6 Dirac Fermions in an External

tromagnetic): Landau Leuels
Near the Dirac point K'; we have}Qirac equation for massless pa:rticle}__
Gauge Field (Elec-
d
<f f-.Jl,
l _5
Í"fµ'Oµ,'l/JA = Ü.
Near the other Dirac Point K, ~irac equation is

Í"f.µ.8µ.'1/JB = Ü.
As already discussed 1/J A and 1/J B are related by parity transformation. We

now introduce a gauge field Aµ. in a gauge invariant way by replacing 8µ.
by covariant derivative Dµ.:
ie
8µ --+ Dµ. = 8,,, + -Aµ
e
so that
i,µ. Dµ'I/JA = 0. (21.59)
Multiplying by ( -i··t Dv) on the left, we have
Using Dirac algebra:

(-i,vDv)(i''tDµ.'1/JA) =Ü L
1
CTµv = 2i [,µ. ,v - ,v ')'µ]
'Yµ'Yv + ,v 'Yµ = 2gµv
we can write the above equation as
{gµv DµDv + ~crµv[Dµ, Dv]}'I/JA = O. (21.60)
One can easily evaluate
[Dµ,Dv] = i:Fµvl (21.61) }

Fµv = 8µAv - 8vAµ-
Thus we obtain
[Dµ.Dµ - ;e CTµv Fµ.v]"PA = O, µ, V= O, 1, 2
or
ie
\ (8µ + CAµ)(8µ + ie
CAµ)"PA -
e v
2c (1µ Fµv"PA = o. (21.62)
N This differs from the Kleir~ordon equation

DµDµ"PA = O (21.63)
in the term 2ecCTµv
Fµv· It is easy to verify that we can write
o!" = f.>.µv ')'>.. (21.64)
DIRAC EQUATION IN (1+2} DIMENSIONS: APPLICATION TO GRAPHENE535
Thus we can write
(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
2~<P, <P is the magnetic flux.

= (21.65e)
e
We now return to Eq. (21.54) and explicitly evaluate
_..!!_(J'µ.v F = _..!!_/,µ.v F 'Y'
2c JJ,V 2c JJ,V A
=-~{a3(81A2 - 82A1) + (ia1)(-8oA2 - 82Aº)

e
-(ia2)(-8oA1 - 81Aº)}
e
=--{u. +iuxE} (21.66)
e
where
B=VxA E= -8oA- VAº.
Thus Eq. (21.54) becomes
[Dº2 - D2 - ~u - i~u X E]1/!A = Ü (21.67a)
e e
where
ie
DO = 80+-Ao
e
1 8 ie
= --+-Ao,
v¡ at e
(21.67b)
Eq. (21.5a) becomes

-ÍE ie · e .e
[(-v¡ + -Ao)
e
2
- D 2 - -u, B - z-u x E]1/JA
e e
= O (21.68)
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
By the change of variable Y = k + ~ B1,t\ we can write the above equation )

as
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)
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)
le = Jli, lB being the magnetic length: Thus the spectrum of Landau

levels is given by [inserting back ñ]
. 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
21.1 Show that

(i)
a':" = E>.µv 'Y>.
µ,V= O, 1, 2. E012 = 1, ar" = ibµ'Yv - 'Y"'Yµ)
(ii)
Tr['YIL,",>.] = -2iEµv>.
16.2 For the static magnetic field:
Aº=O ,W¿¡)=0 , E=O
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
()
Show that
Define Canonical variables

[D1,D2] =
.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
E±(n) = ± Wc lnl n = O, ±1, ±2 · · ·
21.4 Selected References and Further Reading
i. For 21.1: Lecture notes of a short course given on Horigin of Mass" {(

by Adnan Bashir at National Center for Physics, Quaid-i-azam
university, Islamabad, in Dec. 2005.
ii For the rest of the sections, the main reference is A. H. Castro Neto,
F. Guinea, N. M. R. Peres, K. S. Novoselov and A. K. Geim, The i (
electronic properties of graphene, Rev. Mod. Phys. 81, 109162
(2009) (arXiv:0709.1163 [cond-mat.other] ~p. 2007). GZ)
All figures, except 21.2 has been taken tromthi reference.
iii see also: Francisco Guinea, M. Pilar Lopez-Sancho and Maria A.
H. Vozmediano, Interactions and disorder in 2D graphite sheets,
arXiv:cond-mat/0511558 [cond-mat.str-el], Nov. 2005.

f
.::1J. )(~ iv M.~cklenburg, J. oo and B{2.: Regan, Electronlé'hoton In- /V
teractions in Grap ene, Phys. Rev. B 81, 245401 (2010)
(arXiv:1003.4419 [cond-mat.mes-hall], March. 2010).
v Carbon Nanotube and Graphene Device Physics by H.-S. Philip
Wong and Deji Akinwande, Cambridge University Press, 2011.
SELECTED GENERAL REFERENCES

l. G. Bayern, L cture 01 Quantum M han.iSJ!, W. A. Benjamin, New
York, 1969.
2. D. Bohm, Ouantum Theory, Dover Publ., New York, 1989.
3. P. A. M. Dirac, i;-he Principle of Quantum ~hanics (~ edi-
tion), Oxford University Press (Clarendon), Oflford, 1958.
D. 4. S. Gasio~owicz, Quantum Physics (Second Edition), JnGn Wiley
fl.__ /§Pns, Inc., 1996
5. K. Gottfried, Quantum Mechanic¡s, Vol 11 Fundamentals, W. A.
Benjamin, New York, 1966.
6. L. D. Landau and E. M. Lifshitz, .Quantum 11echanics ( onrel-
.a.:tivistic Theorx) (2nd edition), Addison-Wesley, Reading, Mass.,
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.

Index
o-function potential, 147, 149 of momentum, 33, 44

'Y matrices in (1 +2) dimensions, 517 of operator, 33, 36, 47, 48
of position, 34, 82
Kleiu;pordon equation, 458 Azimuthal quantum number, 109
Absorption cross section, 375 Balmer series, 12

Adjoint representation, 198 Baryon, 333
Aharonvr,3ohm effect, 362 Baryon number, 333
Angular Irequency, 71, 340 Basis vectors, 187
Angular momentum Bell's inequalities, 356
cartesin coordinates, 87 Bessel equation, 132, 134, 161
spherical polar coordinates, 90 Bessel function, 161
Anharmonic oscillator, 301
Bilinear covariants, 506
Annihilation operator, 74, 194, 525
Binary form, 346
Anti matter, 496
Blackbody spectrum, 7
Antilepton, 332
Planck's formula, 7
Antimesons, 334
Planck's hypothesis, 8
Antiparticle spinor, 521
Bloch function, 216, 523
Antisymmetric state, 315
Antiunitary transformation, 446 Bloch sphere, 34 7
Approximate methods Bohr atom, 12-14
non-perturbative, 283 Bohr frequency condition, 376
perturbative, 283 Bohr magneton, 251
Associated Laguerre polynomials, 111 Bohr orbit, 15
Associated Legendre's polynomials, Bohr postulates, 13
96, 97 Bohr radius, 13, 110, 294, 300, 385,
Asymptotic behavior, 121, 128 495
Atom, 12 Boltzmann
Average value, 81 constant, 7
of t.; and Ly, 100 factor, 195
of L; and Lt, 100 statistics, 316
of Energy, 82 Born amplitude, 170
541
Born approximation, 133-135, Complete orthonormal set, 36, 73

150-152, 154, 423 Complete set, 44, 404, 415
transition Amplitude, 425 Completeness, 178, 179, 184, 207
Born formula, 1 70 Completeness relation, 38, 41, 43, 422
BoselEinstein statistics, 312, 315 in spinor space, 503
Bosons, 254 Complex conjugate, 175
Bound state, 63, 65, 109, 120, 144, Compton effect, 9, 20
146, 422 Compton wave length, 9
Bound states , 24 Conjugate momenta, 206
Bound system Conservation
life time, 165 flux or probability, 137
Boundary condition, 24 Continuity conditions, 59, 67, 531
Boundary conditions, 29, 63 Continuity equation, 2, 24, 26, 359,
Breiijl)Wigner formula, 144 457, 483
Brillouin zone, 523 Continuous spectrum, 38
of energy eigenvalues, 120
Canonical Continuous transformation, 218
coordinate, 207 Coordinate representation, 422
equations of motion, 206 Correspondence principie, 22, 27
quantization, 207, 221 Coulomb gauge, 360
Canonical variables, 538 Coulomb interaction, 292
Carbon nanotubes, 521 Coulomb potential, 106, 109, 169,
Casimir operators, 479 198, 296, 300, 491
Centrifuga! barrier, 106 Coulomb scattering, 154
Centrifuga! potential, 106 Covariant vector, 452
Charge conjugate matrix, 505 CP eigenstates, 339
Charge conjugation, 337, 504 CP violation, 339
plane wave solution, 505 Creation operator, 75, 194, 525
Charmonium, 166, 172 Cross section, 125
Chiral Dirac fermions, 527 Cyclotron mass, 527
chiral projections
positive and negative energy, 509 Damped harmonic oscillator, 340
positive and negative helicity, 509 ~e Broglie rule, 20
Chiral projections of 'lj;, 464, 509 ~e Broglie wave length, 10, 15
Chiral representation of ')'-matrices, Decay width, 383
463, 489, 508 Degeneracy, 287
Chiral transformation, 509 four fold, 109
Classical turning points, 159 three fold, 109
Clebsch\+Jordan coefficients, 257 Differential cross section, 126
Closure Relations, 41 Differential cross-section, 419, 428
Commutation relation, 445, 446, 455 Differential equation
of L±, 94 radial part, 104
Commutation relations Dipole approximation, 378
of angular momentum, 88 Dipole sum rules, 387
Commutative law, 37 Dirac bilinears, 468
Commutator, 25, 37 Dirac delta function, 38
DIRAC EQUATION IN (1+2) DIMENSIONS: APPLICATION TO GRAPHENE543
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
Gauge transformation, 360 expectation value, 114

Gaussian integral, 209, 223 hamiltonian, 300
Gaussian wave packet, 14 normalisation condition, 110
Generalized coordinates, 206 raial functions, 112
Generating function, 79 wave functions, 110
Generators of the Lorentz group, 455 Hypercharge, 333
Global phase, 346
Golden rule, 308 Incident current, 54
Good operator, 437 Induced charge, 535
Graphene, 521 Induced dipole moment, 291
electronic states, 525 Inelastic scattering, 136, 426
Graphite, 522 Infinitesimal Lorentz transformation,
Green's function, 209 465
Green's operator, 421, 422 Infinitesimal rotation, 217
Green's theorem, 155 Infinitesimal transformation, 212, 468
Ground state energy, 295 Infinitesimal unitary transformation,
Group of waves, 6 436
Group velocity, 3, 10 Initial conditions, 405
Integral equation, 407
Hadamard gate, 350 Intensity of light, 8, 386
Hadrons, 332 Interaction picture, 405, 406
Hamiltonian effective Hamiltonian, 406
for massless fermion in (1+2) operator Á, 406
dimensions, 529 solution of Schrodinger equation,
central field of force, 104 407
Hamiltonian operator, 22, 26, 27 state vector, 406
for simple harmonic oscillator, 72 Interferometry phenomena, 325, 329,
Harmonic oscillator, 71 338
Heisenberg equation, 203, 205, 232 Intrinsic parity, 333
angular momentum L, 484 Intrinsic spin, 218, 333
for Pi, 483 of an electron, 513
Heisenberg picture, 204, 207, 406 Invariance principies, 211
Heisenberg uncertainty principie, 11, Ionization energy, 297
19, 26 Isospin, 333
Helicity operator, 327, 501 Isospin doublet, 333
Helium atom, 316 Isotropic scattering, 128
Hermite polynomial, 78
Hermitian conjugate, 176
Hermitian operator, 34, 35, 47, 48,
Klein paradox, 530
Klei~Gordon equation, 456, 457, 488,
J_
177, 455 497 N
conjugate or adjoint, 34 in the presence of electromagnetic
Hydrogen spectrum, 12 field, 488
general law of, 12 in the term o-.E and rr.B, 489
Hydrogen-like atom Pauli and Weisskopf interpretation,
energy eigenfunctions, 112 458 {
energy levels, 112 KleinYpordon like equation, 459
f-J
Kronecker delta, 36 Matrix representation

of qubits, 348
Laguerre polynomials, 111 generators of the Lorentz group,
Lamb shift, 495 475
Landau levels, 368, 533, 537 of Hadamard gate, 351
Large bilinears, 507 of Lorentz boost, 477
Larmer circle, 368 of Lorentz transformation, 452
Lattice translation, 215 of phase gate, 350
Left handed particle, 510 of the NOT gate, 350
Legendre polynomial, 97, 128 Maxwell's equations, 1, 359
Legendre's equation, 96 Mesons, 334
Leibnitz theorem, 97 Minkowski matrix, 478
Lenz's vector, 109 Momentum operator, 44
Lepton number, 333 Momentum operator, 25, 26, 43, 48,
Leptons, 332 175, 182, 185, 519
Linear operator, 176, 407 Momentum operator , 22
group property of U, 407 Momentum representation, 185
Lippman~chwinger equation, 413 Mott scattering, 155
Local gauge transformation, 481 Multiple qubit systems, 347
Logic gate
Superposition gate or Hadamard Neutrino, 332
gate, 350 Neutrino oscillation, 340
Lorentz boost, 453, 455 No cloning theorem, 355
generators, 4 77 Non-Commutativity Law, 45
Lorentz gauge, 360 Non-degenerate eigenstate, 291
Lorentz group, 453 Non-perturbative method, 295
Generators, 466 Normalised eigenstates, 178
Líe algebra, 466 Normalised spherical harmonics, 96
Lorentz boost, 478 Normalization condition, 53
rotation group, 477, 478 Normalization integral, 25
rotation subgroup, 476 Normalized eigenfunctions, 35, 44
subgroup, 466 NOT gate, 349
unitary, 477
Lorentz transformation, 452, 464 One-dimensional harmonic oscillator,
commutation relation, 477 536
Lorentz velocity transformation, 454 energy levels, 538
Operators, 176
Magnetic interaction energy, 300 Optical theorem, 130, 131, 136, 144,
Magnetic length, 537 431
Magnetic moment of electron, 491 Orbital angular momentum, 491
Magnetic quantum number, 109 Orbital magnetic moment, 382
Magnetic spin anomaly, 491 Ortho state, 318
Magnetic spin resonance, 329 Orthochronus Lorentz
Mass matrix, 335, 336 trnasformation, 453
Matrix element, 34, 176 Orthogonality
Matrix mechanics, 187 of Legendre polynomials, 440
Orthogonality condition, 38 Phase velocity, 2, 3

Orthonormal basis, 325 Photoelectric effect, 8
Orthonormal eigenfunctíons, 36 Photon momentum, 18
Orthonormal set, 422 Planck's law, 377
Orthonormality, 186 Plane wave solutions, 520
Orthonormality of eigenvectors, 404 Poisson Bracket, 206
Position operator, 25
Para state, 318 Positive and negative energy, 509
Parametric relation, 347 Postulates of special theory of
Parity, 61, 98, 442 relativity, 451
definite parity, 67 Potential
even, 146 attractive, 169
even and odd eigenfunctions, 66, 67 electrostatic, 290
odd, 146 linear, 166
orbital, 99 logarithmic, 173
Parity of S-matrix, 442 repulsive, 106
Parity operator, 219, 292 spherical symrnetric, 152
Parity transformation, 519 spherically symmetric, 103, 115
Partial cross section, 131 square well, 115, 144
Partial wave, 170 symmetric, 98
Partial wave cross-section, 432 Potential barrier, 58, 68
Partial wave scattering amplitude, Potential step, 51
142 Primitive cell, 525
Particle mixing, 332 Principie of detailed balance, 449
Particle on mass shell, 455 Principie quantum number, 109
Path integral, 221, 222 Probability, 56
Path integral formalisam, 225 Probability amplitudes, 346
Pauli matrices, 326, 460, 508, 529 Probability current, 362
Pauli representation of 'Y matrices, Probability current density, 26, 31
514 Probability density, 24, 26, 34, 474
Pauli representation of -y-matrices, for a stationary state, 28
463, 507 radial, 113
Pauli spinors, 499 Probability of an operator, 36, 45
Pauli's exclusion principie, 312, 314 Probability of transition, 416
Permutation operator, 312 Projection operator, 503
Perturbation potential, 373 Propagator, 434
Perturbation theory, 405 Proper Homogeneous Lorentz group
potential, 290 Lt is 80(3, 1), 478
Stationary, 283 Proper Lorentz transformation, 453
time dependent, 303 Dirac bilinears, 470
time independent, 283 Properties of
Perturbative expansion, 421 Angular momentum, 87
Perturbed hamiltonian, 290 Associated Legendre's polynomials,
Phase factor, 519 97
Phase gate, 350 Dirac bilinears, 468
Phase shift, 129, 131, 135, 168 Dirac delta function, 39, 42
Small bilinears, 508 Transition matrix elements, 444

Space reflection, 218, 219, 470, 505 Transition matrix or T-matrix, 415,
Dirac bilinears, 471 436
Spectral function, 388 Transition probability, 303, 304, 306,
Spectral function sum rule, 391 307, 374, 405, 418
Spectral representation in the electric dipole, 379
of the two point function, 388 Transition probability amplitude, 151
Spherical Bessel function, 128, 132, Transition rate, 418
134 Transmission
Spherical harmonics, 97 coefficient, 164
Spherical polar coordinates, 90 probability, 164
Spherical symmetry, 97 Transmission coefficients, 68
Spherical waves, 425 Transmitivity, 54, 60, 69
Spin singlet, 318 Transmitted current, 54
Spin singlet state, 320 Transverse momentum, 105
Spin triplet, 318 Tunneling, 163
Spin triplet state, 320 Tunneling probability, 532
Spontaneous emission of radiation, Tunnelling, 60
378 probability, 60
Square barrier, 530 Two body problem, 117
Square well potential, 65 Two component theory of neutrino,
Standard deviation, 47 510
Stark effect, 290 Two components wave functions, 489
State vector, 175 Two electron system, 316
Static magnetic field, 536 Two qubit logic gates, 351
Stationary state, 103, 110, 119, 120, control not logic gate (CNOT), 351
144 Two state system, 325
Stationary states, 27, 507, 536
Stefan's law, 8 Uncertainty principie, 17, 18
Superposition gate, 350 Unitary matrix, 328
Superposition of waves, 3 Unitary operator, 181, 406, 410, 436,
Supersymmetric oscillator, 194 468
Unitary transformation, 209-211,
Taylor expansion, 78, 146 328, 435
Tight binding Hamiltonian of the Unperturbed Hamiltonian, 403
electrons, 525 Unperturbed hamiltonian, 289, 291,
Time ordered product, 394 303
Time reversa!, 218, 220, 445, 472, Unperturbed state, 291
505, 506
Time-ordered product, 226, 232 Variational principie, 295
Total cross section, 126, 131, 137 Velocity in Dirac theory, 486
Total cross-section, 431 Virial theorem, 84
Total energy of a particle, 22
Transition amplitude, 222, 409, 413 Wave, 2
Transition matrix, 423 in 3 dimensions, 2
Transition matrix element, 420 in x-direction, 2
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
Yukawa potential, 154
Zeeman effect, 300

Zero potential, 229
Zitterbewegung term, 487

(Quantum Mechanics) Fayyazuddin, Riazuddin-Quantum Mechanics-World Scientific (2012)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Quantum Mechanics) Fayyazuddin, Riazuddin-Quantum Mechanics-World Scientific (2012)

Uploaded by

Copyright:

Available Formats

I July 9, 2012 14:29 World Scientific Book - 9in x 6in RF-QuantumMechanics I

To the memory of our parents, brother mbdul .Sfatim and sister

viii Quantum Mechanics

Preface for the 2nd Edition

substantial number of new problern .

Preface for the lst Edition

Chapter 11 (Secs. 1-5), Chapter 12 (Sec. 4), Chapters 13 and 14 should be

xii Quantum Mechanics

l. BREAKDOWN OF CLASSICAL CONCEPTS 1

2. QUANTUM MECHANICAL CONCEPTS 17

3. BASIC POSTULATES OF QUANTUM MECHANICS 33

4. SOLUTION OF PROBLEMS IN QUANTUM MECHANICS 51

xiv Quantum M echanics

4.2 The Potential Barrier . . . . . . . . . . . . . . . . . . 58

5. SIMPLE HARMONIC OSCILLATOR 71

7. MOTION IN CENTRALLY SYMMETRIC FIELD 103

8. COLLISION THEORY 117

8.11 The Born Approximation . . 150

10. HEISENBERG EQUATION OF MOTION, INVARI-

11. ANGULAR MOMENTUM AND SPIN 237

xvi Quantum M echanics

11. 7 Addition of Angular Momenta 254

12. TIME INDEPENDENT PERTURBATION THEORY 283

13. TIME DEPENDENT PERTURBATION THEORY 303

14. STATISTICS AND THE EXCLUSION PRINCIPLE 311

15. TWO STATE SYSTEMS 325

16. Quantum Computation 345

16.2 Qubit . 346

17. PERTURBATION INDUCED BY ELECTRO MAG-

18. FORMAL THEORY OF SCATTERING 403

xviii Quantum Mechanics

18.10 Problems . 432

19. S-MATRJX AND INVARJANCE PRINCIPLES 435

20. RELATIVISTIC QUANTUM MECHANICS:DIRAC EQUATION451

21. DIRAC EQUATION IN (1+2) DIMENSIONS: APPLI-

21.1.1 Dirac Equation in Hamiltonian form and Plane

Quantum mechanics is the theory which describes phenomena on the atomic

divE = 41rp (1.la)

In free space, E and B satisfy the wave equation

A solution of this equation shows that E and B propagate through space

>. : wave length

BREAKDOWN OF CLASSICAL CONCEPTS 3

We cannot send a signal in the form of a monochromatic wave. However,

'1p = '1/J1 +'1/J2

x [ exp [ - i ( ~k x - ~wt) J + exp [i ( ~k x - ~w t) J J

This is the group velocity.

'1/J(x, O)= l A(k)eikxdk

BREAKDOWN OF CLASSICAL CONCEPTS 5

Fig. 1.2 A wave packet pictured at t = O

Fig. 1.3 Picture of A(k)

To reduce the dispersion, we take

and expand Eq. (1.12) around k0:

w(k) = w(k0) + (!~) ko (k - ko) + ... (1.14)

Now let us rewrite Eq. (1.llb) in the followingway

in the exponential. This is justified only if

The wave picture of radiation satisfactorily explains the phenomena of

BREAKDOWN OF CLASSICAL CONCEPTS 7

1.2 Breakdown of Classical Concepts

(a) Dual Behaoiour of Radiation

(i) Phenomena of interference and diffraction, which are characteristic of

(1) Blackbody spectrum

This concerns the thermodynamics of the exchange of energy between

E(w) = 81rliw3 (e(ñw)/kT -1)-1 , (1.21)

Incidently the Planck's formula (1.21) reduces to the classical Rayleigh-