Beruflich Dokumente
Kultur Dokumente
ions 1
i n /
NONRELATIVISTIC
QUANTUM
MECHANICS
Anton Z. Capri
*
1 i
•
World Scientific
cpfo 6kwis
Soíupons
NONRELATIVISTIC
QUANTUM
MECHANICS
Anton Z. Capri
Department of Physics
University of Alberta, Canada
V
\
<y 2 M
Published by
For photocopying of material in this volume, please pay a copying fee through the Copyright
Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to
photocopy is not required from the publisher.
A.Z.Capri
Edmonton, Alberta
July, 2002.
Contents
2 R e v i e w of Classical Mechanics 8
2.1 Lagrangian and Hamiltonian for SHO 8
2.2 Lagrangian and Hamiltonian: Simple Pendulum 9
2.3 Bohr-Sommerfeld Quantization: SHO 9
2.4 Bohr-Sommerfeld: Particle in a Box 10
2.5 Larmor Frequency 11
2.6 Applicability of Bohr-Sommerfeld Quantization 12
2.7 Schrodinger and Hamilton-Jacobi 12
2.8 WKB Approximation 13
2.9 Dumbbell Molecule: Bohr-Sommerfeld 14
3 Elementary Systems 15
3.1 Commutator Identities 15
3.2 Complex Potential 16
3.3 Group and Phase Velocity 17
3.4 Linear Operators 18
3.5 Probability Density 19
3.6 Angular Momentum Operators 20
3.7 Beam of Particles 21
3.8 Time Evolution of Wave Function 22
3.9 Operator Hamiltonian 23
3.10 Zero of Energy 24
3.11 Some Commutators 25
3.12 Eigenfunction for a Simple Hamiltonian 26
4 One-Dimensional Problems 29
4.1 Potential Step 29
4.2 Deep Square Well 30
4.3 Hydrogenic Wavefunction 32
4.4 Bound State Wavefunction, Current, Momentum 33
4.5 Time Evolution for Particle in a Box 34
4.6 Particle in Box: Energy and Eigenfunctions 35
4.7 Particle in a Box 36
4.8 Particles Incident on a Potential 36
4.9 Two Beams Incident on a Potential 38
4.10 Ramsauer-Townsend Effect 39
4.11 Wronskian and Non-degeneracy in 1 Dimension 40
4.12 Symmetry of Reflection 41
4.13 Parity and Electric Dipole Moment 43
4.14 Bound State Degeneracy and Current 44
4.15 Car Reflected from a Cliff 45
5 More O n e - D i m e n s i o n a l P r o b l e m s 48
5.1 Motion of a Wavepacket 48
5.2 Lowest Energy States 50
5.3 Particle at Rest 51
5.4 Scattering from Two Delta Functions 51
5.5 Reflection and Transmission Amplitudes: Phase Shifts 53
5.6 Oscillator Against a Solid Wall 55
5.7 Periodic Potential 56
5.8 Reflection and Transmission Through a Barrier 57
5.9 Hermite Polynomials: Integral Representation 58
5.10 Matrix Element Between Degenerate States 60
5.11 Hellmann-Feynman Theorem 61
6 M a t h e m a t i c a l Foundations 63
6.1 Cauchy Sequence in a Finite Vector Space 63
6.2 Nonuniqueness of Schrodinger Representation 64
6.3 Degeneracy and Commutator 64
6.4 von Neumann's Example 65
6.5 Projection Operator 67
6.6 Spectral Resolution 68
6.7 Resolvent Operator 68
6.8 Deficiency Indices 69
6.9 Adjoint Operator 71
6.10 Projection Operator 72
6.11 Commutator of Lz and <p 74
6.12 Uncertainty Relation: Lz and cos tp, siny? 74
6.13 Domain of Kinetic Energy: Polar Coordinates 75
6.14 Self-Adjoint Extensions of p 4 76
7 Physical I n t e r p r e t a t i o n 79
7.1 Tetrahedral Die 79
7.2 Probabilities, Expectation Values, Evolution 79
7.3 (Lx) and (Ly) in an Eigenstate of Lz 80
7.4 Free Particle Propagator 81
7.5 Minimum Uncertainty Wavefunction 82
7.6 Spreading of a Wave Packet 84
7.7 Time-dependent Expectation Values 85
7.8 Ehrenfest Theorem 87
7.9 Compatibility Theorem 89
7.10 Constant of the Motion 90
7.11 Spreading of a Gaussian Wavepacket 90
7.12 Incorrect Time Operator 92
7.13 Probability to Find a Particle 92
7.14 Sphere Bouncing on Sphere 93
7.15 Cloud Chamber Tracks 95
7.16 Spin 1 Measurement in Two Directions 95
7.17 Particle in a Box: Probabilities and Evolution 97
7.18 Free Wave Equation: Translation Invariance 99
7.19 Free Wave Equation: Accelerated Frame 101
7.20 The Wigner Function 103
7.21 Properties of the Wigner Function 105
7.22 Uncertainty Relation and Wigner Function 106
8 D i s t r i b u t i o n s a n d Fourier T r a n s f o r m s 108
8.1 Properties of the Delta Function 108
8.2 Representation of Delta Function 110
8.3 Normalization of Scattering Solution 112
8.4 Tempered Distribution 114
8.5 Fourier Transform of V 114
8.6 Tempered Distribution of Fast Decrease 115
8.7 A Useful Identity 116
8.8 A Representation of ¿(a:) 118
8.9 Fourier Transform of ¿ ' " ' ( x ) 118
8.10 Value of x m S < - n \x) 119
8.11 Distribution Occurring in Fermi's Golden Rule 119
9 Algebraic M e t h o d s 122
9.1 An Operator Identity 122
9.2 Expectation Values: Simple Harmonic Oscillator 123
9.3 Angular Momentum Matrices 124
9.4 Displaced Oscillator 127
9.5 Dipole Matrix Elements 128
9.6 Scalar Operator 129
9.7 Probability to Obtain I, m 129
9.8 Probability to Obtain I, m Along Different Axis 130
9.9 Commutators of x and p with L 132
9.10 Some Eigenfunctions of Angular Momentum 132
9.11 Expectation Value of Lx 134
9.12 Rotation Invariance of the Hamiltonian 136
9.13 Uncertainty Relation for SHO 137
9.14 Baker-Campbell-Hausdorff Formula 138
9.15 A Useful Commutator 139
9.16 Uncertainty in Lz 140
9.17 Expectation Values of Angular Momentum 140
9.18 Validity of Ehrenfest's Theorem 142
9.19 Wigner Problem: Annihilation and Creation 143
9.20 Wigner Problem: SHO 145
9.21 Identity for Pauli Matrices 147
9.22 Operator Identity - Spin Rotation 148
9.23 An Operator Identity 149
9.24 Commutator with Inverse Operator 150
9.25 Schwinger Method for Angular Momenta 150
9.26 Minimum Uncertainty in Jx 152
9.27 Unsold's Theorem and its Application 153
9.28 Rotation Matrix for j= 1 154
9.29 Algebra and Constants of the Motion 155
9.30 Coherent State and Normal Ordering 158
9.31 Normal Ordering of xn 158
12 N o n - d e g e n e r a t e P e r t u r b a t i o n T h e o r y 232
12.1 Expansion of l / | r \ — r^\ 232
12.2 Second Order Correction to State 234
12.3 1/2 Ax2 Perturbation of SHO 235
12.4 1/4 Ax4 Perturbation of SHO 237
12.5 1/4 \x4 - Brillouin-Wigner Perturbation 238
12.6 Two-level System 239
12.7 Approximate SHO 242
12.8 Two-dimensional SHO 243
12.9 Kuhn-Thomas-Reiche Sum Rule 246
12.10 Electron in Box Perturbed by Electric Field 248
12.11 Positronium 249
12.12 Rigid Rotator in Electric Field 250
12.13 Electric Dipole Moment Sum Rule 251
12.14 Another Sum Rule 253
12.15 Gaussian Perturbation of SHO Bosons 254
12.16 Gaussian Perturbation of SHO Fermions 255
12.17 Polarizability: Particle in a Box 256
12.18 Atomic Isotope Effect 258
12.19 Relativistic Correction to H atom 260
12.20 van der Waals' Interaction 260
13 D e g e n e r a t e P e r t u r b a t i o n T h e o r y 263
13.1 Stark Effect for n = 2 Level in H 263
13.2 Perturbation of Particle in a Box 264
13.3 Perturbation of Isotropic Two-dimensional SHO 266
13.4 Two-dimensional SHO with Off-diagonal Term 267
13.5 Non-diagonal Two-dimensional SHO 270
13.6 Particle in a Box Perturbed by Electric Field 272
13.7 Unusual Particle on Interval 273
13.8 Rigid Rotator in Magnetic Field 274
13.9 Axp y Perturbation of SHO 276
13.10 Paschen-Back Effect 277
13.11 H Atom: Weak Field Stark Effect 281
14 Further A p p r o x i m a t i o n M e t h o d s 284
14.1 Variational Ground State of SHO 284
14.2 Variational Ground State of H2 Molecule 285
14.3 Square Barrier: W K B Approximation 287
14.4 Variational Ground State in Gaussian Potential 289
14.5 Variational Ground State: Quartic Potential 292
14.6 WKB: Ball Bouncing on a Floor 294
14.7 Ground State of H * 295
14.8 Variational Solution: Particle in a Box 298
14.9 Hydrogen Atom: Variational Technique 299
14.10 Bound State in One Dimension 301
14.11 Field Emission: W K B Approximation 302
14.12 Deuteron: Variational Principle 303
14.13 Bouncing Ball: Variational Calculation 305
14.14 Beta Decay of Tritium 306
14.15 Anharmonic Oscillator 307
14.16 Nonlinear SHO: Variational Calculation 308
14.17 WKB Solution and Parity 311
15 T i m e - D e p e n d e n t P e r t u r b a t i o n T h e o r y 314
15.1 Transition Probability: Bound State to Free 314
15.2 Photo-disintegration of Deuteron 316
15.3 Excitation of SHO 319
15.4 Excitation of SHO by Stiffer Spring 321
15.5 Periodic Perturbation 322
15.6 Excitation of H-atom 323
15.7 Expanding Box 325
15.8 Sudden Displacement of SHO 326
15.9 Sudden Perturbation of Two-level Atom 327
15.10 Berry's Phase 329
15.11 Neutron in Rotating Magnetic Field 333
- -1 • JO Jí ' 11NI I
16 P a r t i c l e i n a U n i f o r m M a g n e t i c F i e l d 342
16.1 Estimate of Magnetic Energies 342
16.2 Radii of Landau Levels 343
16.3 Equation of Continuity 344
16.4 Gauge Invariance 345
16.5 Gauge Transformations and Observables 346
16.6 Spin 1/2 Particle in Magnetic Field 348
16.7 Spin 1/2 in Magnetic Field: Heisenberg Equations 349
16.8 Separation of Spin and Space for Spin 1/2 352
16.9 Spin 1/2 in Time-dependent Magnetic Field 352
16.10 Spin 1/2 in Rotating Magnetic Field 354
17 A n g u l a r M o m e n t u m , E t c . 358
17.1 Operator to Lower Total J 358
17.2 Energy Shift Due to a Magnetic Field 359
17.3 Coupling of Spin 1 to Spin 1/2 361
17.4 Example of Wigner-Eckart 362
17.5 Rotations for Spin 1/2 and Spin 1 364
17.6 Spin 1/2 Coupled to Spin 3 / 2 366
17.7 Coupling of Spin 1 or 0 and Spin 1/2 367
17.8 Identity for Constant Magnetic Field 368
17.9 The State \n,j, m,l) 369
17.10 Landé ^-factor 369
17.11 Spin and Space Coordinates 372
17.12 Clebsch-Gordon for 3/2 373
17.13 Rigid Rotator in a Step Potential 373
17.14 Spin Dependent Operators for Two Particles 375
18 S c a t t e r i n g - T i m e D e p e n d e n t 377
18.1 Cross-section from Experiment 377
18.2 Green's Functions for Free Particle States 378
18.3 Dispersion Relations 379
18.4 Kalien-Yang-Feldman Equations 380
18.5 Born Approximation 381
18.6 Scattering in CM and Laboratory Frame 382
18.7 Propagator for a Free Particle 385
18.8 Propagator for Simple Harmonic Oscillator 386
19 Scattering - T i m e I n d e p e n d e n t 389
19.1 Equations for Spherical Bessel Functions 389
19.2 Rodrigues Formula: Spherical Bessel Functions 392
19.3 Wronskian for Spherical Bessel Functions 393
19.4 Superposition of Yukawa Potentials 393
19.5 B o r n Approximation for Gaussian Potential 394
19.6 Born Approximation for Square Well 395
19.7 Phase Shifts for Delta-function Potential 397
19.8 Phase Shifts for Yukawa Potential 399
19.9 Low Energy s-Wave Amplitude 400
19.10 Spherical Potential Shell 402
19.11 Expressions for j0{x) and n0(x) 403
19.12 Effective Range, Scattering Length 404
19.13 Effective Range, Scattering Length: Yukawa Potential 405
19.14 Shape-independent Parameters 406
19.15 Phase Shifts for Hard Sphere 407
19.16 Resonance for Square Well 411
19.17 Double Slit 412
19.18 Born Approximation: Spherically Symmetric Potential 413
19.19 Scattering from a Separable Potential 414
19.20 Generalized Optical Potential 415
19.21 Free Particle Eigenfunctions 416
19.22 Scattering from an Inverse Square Potential 418
19.23 Neutron-Proton Scattering: Spin Flip 419
19.24 Reflectionless Potential in One Dimension 420
19.25 n-p Scattering: Singlet and Triplet States 422
19.26 Phase Shift, Scattering Length, Etc 423
19.27 Scattering off a Diatomic Molecule 425
19.28 WKB s-Wave Phase Shift: Attractive Potential 426
19.29 WKB s-Wave Phase Shift: Hulthén Potential 427
19.30 WKB Approximation for Phase Shifts 428
19.31 Zero-Range Potential 431
19.32 Calogero Equation 435
21 Q u a n t u m S t a t i s t i c a l M e c h a n i c s 467
21.1 Average Energy of Assembly of SHO's 467
21.2 Properties of the Density Matrix 469
21.3 Expectation Values for Spin 470
21.4 Expectation Value for Number of Particles 471
21.5 Spin 1/2 Polarization 472
21.6 Density Matrix for Spin s = l 473
21.7 Polarization Vector for Spin j 475
21.8 Composite Density Matrix 476
21.9 Arbitrariness of Composite Density Matrix 477
21.10 Two Energy Levels Bose Gas 479
21.11 Density Matrix: Particles Coupled by Spring 481
21.12 Particles: Dissimilar, Bose, and Fermi 483
21.13 A Three-Level Laser 484
21.14 Integrals from Q u a n t u m Statistical Mechanics 486
Index 489
Chapter 1
T h e B r e a k d o w n of Classical
Mechanics
Solution
The classical energy of the earth in its orbit about the sun is
where
m is the mass of the earth = 6 x 10 27 g ,
M is the mass of the sun = 2 x 10 33 g ,
G is the gravitational constant = 6.67 x 1 0 - 8 dyn cm 2 /g 2 and
R is the sun-earth distance = 1.5 x 10 13 cm .
Proceeding as for the hydrogen atom we find that we need only replace e 2 by
GmM. Thus, we find
GmM (1.1.2)
GMm2 = 18 x 10 (1.1.3)
Therefore
n « 4 x 10 73 . (1.1.4)
1
Also we see t h a t since
1 /2tr \2 An An ,
AE — - I —-GmM ) m —3 = E\ (1.1.5)
2 \ h J n n
we get AE = 6.4 x 1 0 - 3 4 erg .
Solution
The energy U of a molecule at temperature T is given by the equipartition
principle
U = \ k B T = y . (1.2.6)
6 63 X 10 3
A = - 6 2~ ^ o 3 _ Q 0 x 1 Q 8 = 3.20 x 1 0 - m . (1.2.9)
This is considerably longer than visible light which has a wavelength of about
5 x 10~ 7 m .
1.3 P h o t o n s in a B e a m
For a monochromatic beam of electromagnetic radiation of wavelength (A «
5000 Á) , intensity 7 = 1 w a t t / m " , calculate the number of photons passing an
area of A = 1 cm 2 normal to the beam in one second.
Solution
The energy of a photon is given by
-^photon —- —y • (1.3.10)
Solution
The de Broglie wavelength is given by
A = h/p .
2nr= n\ .
it P2 e2
E =
to " 7 - <L412>
We substitute for p and r in terms of A and introduce the fine structure constant
2ne2
« = -¡¡r (1.4.13)
to get
h2 hca
E = (1.4.14)
2mA2 n\
We now equate the Coulomb force with the mass times the centripetal acceler-
ation
mv2 2xe2
— = ^x- ( L415 )
This yields
h2 27re2
S P = 7 T ' C-4-16'
We can now solve for A.
1 mac 1
A=—~n- (1.4.17)
Solution
The shortest wavelength is given by
(1.5.19)
A m , n = 5.8 x 1 0 - 8 cm = 2 x lattice spacing
»o = l , 5 x 105 cm/s .
(1.5.20)
a)
«o
A '
(1.5.21)
Therefore,
vmax = = 2.6 x 10 12 Hz .
Amin
(1.5.22)
b)
u = \kBT .
(1.5.24)
Assume u « Emax. Therefore,
T ~ ^Emax on
(1.5.25)
3k B - 8 3 K •
to
l2n(kBT)4 iXmax x3 dx
^powder — h3y
0 Jxmtn ex - 1
rr 12Tv(k B T) [*""" x3dx
4
" — V S T J. — r f'-"7»
Here, we have introduced
h =
h Do
•Emm — ^ ^min ~j ™ T (1.6.28)
k B A m a x
where Á m a x = 1 0 - 5 cm « size of the powder particles. If we now estimate x m i n
at room temperature by using that vo ps 500 m / s , we get x m ¡„ « 8 x 1 0 - 4 < < 1.
Therefore, we can approximate the exponential in the last integral by ex ss 1 + a:
and get
rxmtn x3dx rrm,„ j
Jo e ' - l ™J0 X dX
~ ~Z X
min- (1.6.29)
A u
12n-(*Br)4 x /I hv2 o \ 3 4tt kBT
= itur- (£rrr --) = ^ - (1-6.30)
h3V¡¡ X
1^3 kBXmax ) ~ K
Therefore, the specific heat per unit volume at constant volume
\_dU_
V dT
is decreased by a constant amount, namely
Anm ~ Bnm .
có
Solution
In equilibrium, at a temperature T, if the number of atoms in the state n and
m is given by Nn and Nrn respectively, we have t h a t
Also at equilibrium the number of transitions from n —»• m equals the number
of transitions from m —> n. Thus, we have
or
<•<">=^rJBBm\B
c
• <17.38)
J^nm/
But, at equilibrium the radiation density is given by Planck's Law
Therefore, comparing these two equations we see that we have
and
871" , 3 n
Aim — ' (1.7.41)
Bibliography
[1.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 1.12, (2002) .
[1.2] F.K. Richtmyer, E.H. Kennard, and J.N. Cooper, Introduction to Modern
Physics, 6th edition, sec 13.12, McGraw-Hill, New York, (1969).
Chapter 2
R e v i e w of Classical
Mechanics
Solution
For a simple harmonic oscillator we have
1
2
T -mv
(2.1.1)
Therefore,
dL
p = -x- = mv . (2.1.3)
ov
(2.1.4)
8
2.2 Lagrangian and Hamiltonian: Simple
Pendulum
Repeat problem 2.1 for the simple pendulum. Interpret the momentum p con-
jugate to the angle variable 9 .
Solution
Here, the kinetic energy is given by
T=^ml292 . (2.2.6)
H = d-±9-L
09
= ml2 9 — -ml2 92 + mgl(l — cos 9)
But the Hamiltonian is a function of the generalized coordinate and the conju-
gate momentum. So we get
Pe
H = + mgl( 1 — cos9) . (2.2.11)
¿ml¿
Furthermore, the expression for the conjugate momentum (2.2.9) shows that it
is just the angular momentum.
p= ±V2^E]Jl-'^ • (2.3.14)
f p d , = l-k-^-d,
2E
— 2\/2 mE J (2.3.15)
We now call w2 = k/m and, after evaluating the integral, set it equal to nh to
get
2nnh = 2 — \ . (2.3.16)
UJ 2
So, the quantized energy levels are
En = nhu> . (2.3.17)
Solution
As in the problem above we use Bohr-Sommerfeld quantization
(jipdq = nh
= V2 mE 11: dx — J dxj
= \ZlmE x 2L . (2.4.20)
h2
£ = !
s ^ " - I2-4'21)
¡1 = ML .
V = -¡I B .
Assume that B is constant and derive the equation of motion for L. Show that
the gyroscope precesses with the angular Larmor frequency
u)L — MB .
Solution
For a gyroscope with angular momentum L we have t h a t if the angular velocity
is Q and the moment of inertia is 7 then
L = Id . (2.5.22)
2
T=^IJ = ^lé2. (2.5.23)
Also,
Solution
The system of quantization proposed by Bohr and Sommerfeld is applicable to
systems with repetitive (periodic) motion. It is not applicable to systems with
unbounded motion.
xjr = A e ' s / h
h2 a 2 * 3*
2m dx2 + VV = ik- dt
1
fdS\2 , 8S
+ V { q )
2m f e j -~dT-
Solution
If we put
V = AeiS'h
2.8 W K B Approximation
In the problem above set S = W — Et and let W = Wo + hW\ + h¿W^ + • • • for
the case of a one-dimensional Schrodinger equation. Find the equations for Wo
and Wi and solve the equation for Wq . This is the so-called Wentzel-Kramers-
Brillouin or W K B approximation.
Solution
Clearly if in the problem above we set S = W — Et (before letting h —>• 0) we
get
(2.8.31)
(2.8.32)
So,
(2.8.33)
and
(2.8.34)
Solution
a) If we take the plane in which the molecule is rotating as the plane z = 0 then
the Lagrangian is just the kinetic energy
L = iI<p 2 . (2.9.35)
But,
where E is the total energy which is conserved and therefore a constant. Thus,
/* 27T
V2IE / dip • 27t\/2IE = nh . (2.9.39)
Jo
So,
2
1 , / h \ n2h2
E = — (2.9.40)
21 \2n 21
Bibliography
[2.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 2.4, (2002) .
Chapter 3
Elementary Systems
[x~,p] = 2 ihx
and
Hence prove that for any function g(x) analytic at the origin
[g{x),p] = •
Solution
a) To verify the identity we simply write it out
[AB,C] = ABC-CAB
= ABC - ACB + ACB - CAB
= A[B,C] + [A,C]B. (3.1.1)
Hence,
[g{x),P\ = l ^ c n [ z n , P ] (3.1.5)
[ffW.p] = • (3.1.9)
l + v.j = \w„.
Solution
We begin with the Schrodinger equation and its complex conjugate.
d$! h2
dV*
—ih + (U - iW)y . (3.2.10)
dt 2m
Next, we multiply the first equation by and the second by $ and subtract
the two equations to get
=
> (3-212)
v= I9±
2tt
v =
\J 1 - (w0/w)J
Solution
a) The phase velocity is quite generally given by
^ /9 ^ /—rr
= =
=I vs ^ • <3314»
So we see that w = \fgk. Therefore,
( 3 3 1 6 ,
- - '
So, c2k2 = ui2 — uIq. Differentiating this equation with respect to k we get
= . (3.3.17)
£ = cjt = £ (3.3.18)
* dk uj vp
Solution
In all cases, to see if the operator is linear, we have to check whether
=
A(X\ipi + \2ip2) ^1 Aifii + \2A1fi2 (3.4.19)
a) Linear since
b) Linear since
K3ifi{x) = J K(x,y)dy J K{y,z)dz J K(z,w)ifi{w) dw . (3.4.21)
c) Linear since
d) Linear because the identity operator is linear and we have BB~l = 1. Thus,
A"
(exp A)i> = ^2 • (3.4.25)
n=0
But by an argument the same as for part b) above and induction we see that
An is linear. Furthermore, a sum of linear operators is again linear. Therefore,
exp A is linear. In fact, by the same sort of argument, if A is linear then any
function of A say f(A) is also linear.
f) Not linear because
where
_ „-i 2 fc 2 /2
A(k) =
Solution
We first compute the wavefunction $!(t,x) explicitly.
OO
/
dk exp [~k2(L2/2 + iht/2m) + ikx] . (3.5.27)
•OO
The contour of integration may now be shifted to the real axis since the integrand
has no singularities (poles or cuts) in the region between the real axis and a line
parallel to the real axis. Also, for real q oo, the integrand vanishes. Thus,
tf(i,x) exp
2 (L2 + iht/m) / OO
dq exp [—q2(L2/2 + iht/2m)]
2tr -OO
• exp (3.5.31)
(L2 + iht/m) 2 (L2 + iht/m)
2n LJ2X^ 2
P= exp (3.5.32)
\JLA + (ht/m)'- (L + (ht/m)2
4
To see the spreading of the wave packet we find the values of x such that p has
dropped to roughly 1/e of its maximum value. These points are obtained by
setting the argument of the exponential equal to —1. Then we find
L22x™ 2
1 . (3.5.33)
(L4 + (ht/m)2)
h2t2
w = L\ 1+ (3.5.34)
m2L4
This shows how the wave packet gets wider with time.
L = r xp = -fx V
i
then
as well as
[Lz, Lx] = [xpy - ypx, ypz - zpy]
= [xpy,ypz] + [ypx,zpy]
= —ihxpz + ihzpx
= ihLy . (3.6.38)
3.7 B e a m of Particles
A beam of free particles is moving along the x-axis with velocity v such that
there is one particle in a volume V.
a) W h a t is the corresponding normalized, time-dependent wave function for
such a particle?
b) W h a t is the number of particles crossing a unit area, normal to the x-axis,
per unit time?
Solution
a) For a free particle moving with velocity v along the x-axis we have
f(i,x) = Ae^kx~^ (3.7.39)
where
mv mv2
(3.7.40)
22 CHAPTER 3. ELEMENTARY SYSTEMS
A = (3.7.42)
W
b) The current density is given by
, x) dy*(t, x)
~ V(t,x)
2 im dx dx
h 1 —Hkx—ujt) ^ i(kx-ujt) i(kx-cot) ^ -i(kx-uJt)
c a ^ a
2im V OX OX
hk mv v
= (3.7.43)
mV ~ mV V
The number of particles passing through an area S normal to the x-axis is
therefore
jS = v
V
-«(*)«] M
/ OO i[kx w
f{k)e -
-OOthe energy for a free particle is given by the equation
and remember that
(hk)2
hw =
2m
Solution
For a free particle the wavefunction at any time t is of the form
OO
(3.8.44)
/ dk f(k)e^kx-^k^
3.9. OPERATOR HAMILTONIAN 23
where
oo
/ J k
k T ^ 2 - (3.8.46)
This means that
/(*) = • (3-8.47)
Therefore,
oo gi[kx — hk2t/2m]
/ J k
k2 + a2 • (3.8.48)
Fortunately we have been told not to attempt the integral. Therefore, we are
done.
H = £-e~ax
2m
where a is a constant. Find an acceptable Hamiltonian operator for this system.
Notice that the answer is by no means unique.
Solution
The classical Hamiltonian that we are given is
Pop=-~ (3-9-50)
i ax
and
b) rearrange the order of the operators P o p and xop so that the resultant Hamil-
tonian is hermitian. The easiest, and by no means unique, way of doing this is
to symmetrize the operators. Therefore we propose
3 9 5 1
• i - - »
24 CHAPTER ó. ELEMENTARY SYSTEMS
Solution
Suppose we have the Hamiltonian H and we add a constant Vo to it to get the
Hamiltonian H'. Then we have that
ih— = HV (3.10.53)
as well as
d^'
ih—= H'V = {H+ V0)V' . (3.10.54)
If we try adding a constant phase —a/h to $ we have
V = e-ia/n^ . (3.10.55)
Then,
-ia/h i t . t w (3.10.56)
th——=:e ' + ih——
dt [ dt.
So, if we choose
<* = Vo
we get that
ilr' = e- , v °/ A *J . (3.10.57)
we find that
•hdV _ [ e-iv0t,nih<W
ih— = Ko + e i>i m
= e~iV°t/n[H + V0]V
_ e -iVot/h{ E + V [ ^ _ (3.10.61)
So, we have found that
tf'(i,x) = e~iV't/h9(t,x) (3.10.62)
and
E' = E+V0 . (3.10.63)
So, we only get the same constant added to the energy and a constant phase
shift to the wavefunction.
3.11 Some C o m m u t a t o r s
Given the Hamiltonian operator
" = £ + vW
where the operators x and p satisfy the commutation relation
[x,p] = ih
and let the expressions [x, H] and [x, H] operate on an arbitrary function /(ab-
solution
a) Using the hint we have
[»,«]/<.) = ¿
1 2
2 h ^ = - h ^ — . (3.11.64)
2m dx m i dx
So,
\p,H]f{x) = -—1V(x)f(x)}-V(x)*$L
(3.11.66)
Thus,
dV
(3.11.67)
lP H] =
' ~ i H
^ •
a) Show t h a t both
ipi = Ai e-*2/ 2
and
x
rp2 = A2xe ¡2
Solution
a) To verify that the given functions are indeed eigenfunctions we simply act on
them with the given Hamiltonian operator. Thus,
H Aie-x2/2
[1 - x2 + x2\A\ e l2/2
2 2
1 .A i e ~* ' . (3.12.68)
r2/2
H A2xe = +x2 A2xe x2/2
= [2 - x2 + x2\A2x e~x*,2
= 2 . A2X e-*2/ 2
. (3.12.69)
3 1? EIGENFUNCTION run A SIMPLE HAMILTONIAN 27
CO
/
\ip¡\2 dx i =1,2. (3.12.70)
• OO
Jo JO e~ dz
7T (3.12.72)
a
Therefore,
'(a) = ^ • (3.12.73)
e
.
-¿r = ~L 2^-
Jd—
/ (Ca ) f00 2 „ 1 . Af
xdx=
~ (3-12-75)
So, again setting a = 1 we get
A
2=(^) • (3.12.76)
OO
/
= 0 e~x2x3dx (3.12.78)
-OO
since in this case also the integrand is an odd function.
Bibliography
[3.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 1.12, (2002) .
f
Chapter 4
One-Dimensional Problems
/ VE + VO-VE\2
V VE + VO + \ÍE J
In view of this result, what will happen to a car moving at 10 km/hr if it meets
the edge of a 200 m cliff? Is this answer reasonable? If not, why not?
Solution
The Hamiltonian for this problem may be taken to be
where
x < 0
x > 0
(4.1.1)
nn
Since the incoming beam is from the left we have only a transmitted beam
on the right. Thus, the appropriate solutions are
where
2m{E Vo)
k¡ = ^ f , k\ = ~ . (4.1.3)
k1
R = ~ *2 . (4.1.5)
K
ki+k2 '
Cancelling a factor of 2 m / h 2 we get
2
Regarding the car, the question is a red herring. The car going over a cliff
has nothing to do with the problem we have just solved. For the car going
over the cliff the potential is not a step function but varies smoothly from a
potential that is 0 for x < 0 to a potential which is mgy for x > 0. This is a
step function in the configuration of matter, not in the potential energy of the
car. The car problem involves a potential that acts perpendicular to the car's
original direction of motion.
This explains why a probability of |ii| 2 = 0.84% is unreasonable, apart from
the fact that we know that cars don't as a rule bounce away from unobstructed
cliff edges.
The problem of a car going over a cliff is considerably more complicated and
is handled at the end of this chapter (problem 4.15).
V = < 0
^ ~^a "
Find the equation determining the energy eigenvalues of this system. Solve it
approximately assuming a is very small. What happens in the limit a —> 0?
Solution
Here,
x / 0 for \x\ >a
_ for
^ I -f£ l®l < a '
We define U = vo/2a. The problem now is exactly the same as the problem dis-
cussed in [4.2]. The solutions divide into positive and negative parity solutions
and read
f AeKx x < —a
1
B cos kx |x| < a.
1
iP+(x) =:
{ A e~Kx x >a
AeKx
ip-(x) =
=
|r
^ B sin
sinkxkx
x < —a
|x| < a (4.2.8)
A
o
I{ -Ae-Kx x >a
where
,4.2.9)
k tan ka = K (4.2.10)
Negative parity
kcotka = —K . (4.2.11)
The odd parity solution requires that k/ka = —K. This leads to an energy that
diverges as I/a2. Thus, we get only one bound state in the limit as a —> 0. The
*+&={!?-*• itr
corresponding wavefunction is clearly
=^-"w.
Furthermore in this limit we find that U —>• oo. The result is a J-function
potential.
4.3 Hydrogenic Wavefunction
The wavefunction of an electron in the ground state of an hydrogen-like atom
is
Zr a
ip(r) = Ae ! , a = —-z
Solution
a) To normalize we simply integrate and impose
1 = J\ip\2dv. ( 4 . 3 . 1 5 )
a'
Jo 3
"=v5-
b) Let the probability of finding the electron between r and r + dr be P(r) dr.
Let the probability of finding it in the volume between-!) and v + dv be pdv.
Due to the spherical symmetry we again have dv = 4irr2 dr and so the two
probabilities describe the same thing.
P(r) dr = 4np(r) r2 dr . ( 4 . 3 . 1 8 )
Hence,
P = |V>|24ttR 2
= —-J- e~2Zr'a 4ttr2 . ( 4 . 3 . 1 9 )
4.4. t f U U i y u s I'A'IHJ RAVCRUIV^KLILV, O U Í T I T W I , MUMCJVI UM 33
We now maximize P.
dP n
^ = 0. (4.3.20)
This yields
2 r e - 2 Z r / a [1 - Zr/a\ = 0 . (4.3.21)
So,
2
- Z e
(V) = \ e -Zr!a r2dr
Jo r
Ze2
(4.3.24)
a
Then, the average kinetic energy (T) is given by
Solution
a) To obtain the wavefunction for a bound state we have to solve a differential
equation with all the coefficients real and real boundary conditions. As a conse-
quence, we may always choose the normalization constant so that the solution
is real. This is never the case for scattering problems where the boundary con-
ditions describe travelling waves and are of necessity complex.
34
*-XJ ± Jl
7 = — = 3{V-*VV>} • (4.4.26)
J
2im K
-r > l m
Since V is real, this current vanishes. This means, using the equation of conti-
nuity, that the probability density remains constant in time. Thus, a particle
in a certain region of space, determined by this wavefunction, remains in that
region of space. We have a bound state.
c) Writing out the expression for the current we have
7=o = ^ - ( r ^ - ^ r ) = ^ - [ r ^ + ( p r ) i ] . (4.4.27)
2im lm
Therefore,
= x
(p) \ J ^ Wi»l> + (pV'*)^] = 0 . (4.4.28)
Solution
We are given the wavefunction at time t = 0 as
£ 5 1
= £ • <4-5-30'
4.6. PARTICLE IN vu*- '1VJ OO
or
1 . . , ,25hn2t
V(t,x) = -^sin(57rx/a)exp(-z-^^-) . (4.5.32)
Solution
The Hamiltonian for this problem is
P~
H = ^ - V0 0 < x < a . (4.6.33)
The fact that the potential is oo outside the interval 0 < x < a means that the
wavefunction must vanish at both x = 0 and x = a. If we call
2 2m(E + Vo)
¿
* = — ^ (4-6.34)
the Schródinger equation becomes
d2ip „
— + k V = 0 0 < a; < a . (4.6.35)
The physically acceptable solution is therefore of the form
ip(x) = ^ s i n ( ^ x ) 0< x < a (4.6.36)
and we require that
Ha) = 0 (4.6.37)
so that
ka = nn n = 1,2,3,... . (4.6.38)
Thus, the (unnormalized) wavefunctions are
« * ) = ( *.»"<»«/«) for 0<«<a
v
^ U tor z < 0, z > a
The corresponding energies are
„ h2Tr7n2
EN = ~ 2 Vq n = 1 , 2 , 3 , . . . . (4.6.40)
The normalization is given by
An = y j l . (4.6.41)
4.7 Particle in a B o x
If a particle, in a box with sides at x = ±a, is in a state described by the
wavefunction
f(x) = Acos(nx/2a) + B sm(7Tx/a)
a) Choose A and B so that the particle is in the lowest possible energy state.
b) Choose A and B so that the state has parity = + 1 .
Solution
a) For a particle in a box with sides at x = ± a the (unnormalized) eigenfunctions
are:
parity = + 1
h2k2
ip = coskx |x| < a , ka = (n + 1/2)7t , En = —— (4.7.42)
2m
or for parity = — 1
h2k2
ip = sin kx |a:| < a , ka = nir , En = —— . (4.7.43)
2m
Thus, for the wavefunction
f(x) = A cos(7rx/2a) + B sin(nx/a) (4.7.44)
the lowest energy is obtained with 5 = 0.
b) The solution with positive parity requires that f(—x) = f(x). This also
requires B = 0.
( 0 x < 0
Vo 0 < x < a
0 x > a
Find the number of particles that are reflected in one second. How many par-
ticles per second are transmitted?
Solution
For a beam of particles with energy E > Vo incident from the left we have only
a transmitted beam on the right and the wavefunction is given by
iA(eikx + Re~ikx) x<0
ip=\ A {CeiKx + D e~iKx) 0 <x<a (4.8.45)
I ATeikx x > a
The constant A is determined by the condition that the incident beam A e'kx
has a current density of N particles per second. This means that
Hk . - o
N = —\A\ . v( 4 . 8 . 4 6 )
m '
The wavevector k is defined by the energy of the incident beam
h2k2
£ = — (4.8.47)
and K is defined by
h2I\2 „
= +
~2m~ ' <"'48>
Matching the wavefunction and derivatives at x = 0 and i = « w e get
1 + fl = C +D
ik{l - R) = iK(C — D)
iKa iKa
Ce + De~ = Teika
iK (CeiKa - De~iKa) = ikTeika. (4.8.49)
N\R\* = N (k2-I<2)2sin2K a
(4.8.53)
(k + K ) sin 2 Ka + 4k2K2 cos 2 Ka
2 2 2
3$
J
V_v_i J / 1 I -1. JUI1/ ~r • vxj * ^ i jj x" X l/V-***—* -
and the number of particles per second t h a t are transmitted are given by
dí»^ Íí'2
2
TV i n = N r——— . (4.8.54)
2 2 2
(k + K ) sin Ka + 4k2K2 cos 2 Ka
Notice t h a t
N\R\2 + N\T\2 = N (4.8.55)
At the same time another beam of free particles with intensity NR particles
per second and energy ER > Vo is incident from the right. Calculate the total
particle current travelling to the right. [4.1]
Hint: The waves from the left and right are completely independent and their
scattering from the potential can be handled independently.
Solution
To solve this problem we simply use the results of problem 4.8. Then the total
particle current to the right is given by the transmitted part of the current
incident from the left plus the reflected part of the current incident from the
right. Corresponding to these two currents we have a transmission amplitude
TL for the current from the left given by
T 2ikLKLe^k--2K<•)"
L
(k2L + A'2) sin K i a + ' H k i K i cos K ^ a
where
h2k2
EL = ^ (4.9.57)
and also
h2 KL2
E l = ^ -
+ V0 . (4.9.58)
2m
Similarly, we have a reflection amplitude RR for the current incident from the
right given by
Rr = — (k2R + K2R)smKRa g
K K
( R + R ) s i n KR(I + 'LIKJI KR C O S KRQ
wianw xw j i u u i i~ x » 3y
where
h?kf>
f * = - j — (4.9.60)
and also
VKl Tr
ER = — — — h Vó • (4.9.61)
Í 0 if x < 0
V(x) = ^ V0 if 0 < x < a .
[ 0 if x > a
Show t h a t if the wavelength of the particle in the region 0 < x < a is such t h a t
n\ = a then no reflection occurs. Give a physical explanation for this result.
The actual Ramsauer-Townsend Effect was observed in the scattering of
electrons off atoms of the noble gases since due to their closed shell configurations
these atoms have very sharp outer boundaries.
Solution
This is exactly the case of resonance transmission discussed in [4.3]. The
Schródinger equation reads
d ip k ib x < 0
—K¿xp 0 < x < a (4.10.63)
dx2
—k~ib x > a
where
k2 ^ K> = 2
""E - F
°> . (4.10.64)
h¿ ñ¿
Assuming that we have an incoming wave from the left so that for x > a we
have only a transmitted wave, the solution may be written
1+ R = A + B
ik(l — R) = iK(A — B)
Ae iKa
+ Be -iKa Te ika
%K (A JKa Be -¡Ka
ikT eika (4.10.66)
Solving these equations for the transmission amplitude T and computing the
transmission probability T = |T| 2 we get
r= i +
4E{E - V0)
i2 Ka (4.10.67)
Ka — nn (4.10.68)
nX = 2 a (4.10.69)
is a constant.
Hint: Write out the Schrodinger equation for the two solutions and multiply
each of the equations by the other solution.
b) Use the result of part a) to show t h a t in one dimension the energy eigenvalues
of a Hamiltonian
H = £ . + V(x)
p iz + (k2 + U(x)) u = 0
dx
2
^ l + ( k + U(x))v = 0, (4.11.70)
where, as always,
*> - , u{z) =
n2 ' w
h2 '
If we multiply the first of (4.11.70) by v and the second of these by u and
subtract, we get
<4n-7i>
This may be rewritten as
dW _ d du dv
V 0 . (4.11.72)
dx dx di ~ Udi
Therefore, W is a constant and we have shown the desired result.
b) Since, the Wronskian is a constant, we can evaluate it at any point such as
x = ±oo where we have that
u(±oo) = v(±oo) = 0 .
Thus, we have that W = 0. From this it follows that we can write
1 du 1 dv
(4.11.73)
u dx v dx
Integrating this equation we obtain that
l n u = lnv + ln^4 (4.11.74)
where A is an integration constant. Thus,
u = Av (4.11.75)
so that u and v are linearly dependent, contrary to our assumption that there
are two linearly independent eigenfunctions corresponding to the same energy
E.
4.12 S y m m e t r y of Reflection
Consider a repulsive potential V(x) > 0 such that it vanishes outside a finite
region a < x < b . Show t h a t the reflection coefficient is the same for a particle
incident from either the left or the right. Assume that the solutions in the
interval a < x < b have been normalized such t h a t their Wronskian (see problem
4.11) is 1.
Solution
Let u(x) and v(ar) be two linearly independent solutions of the Schrodinger
equation in the interval a < x < b . For a wave incident from the left the
desired scattering solution of the Schrodinger equation is
where
C — u(a)v'(b) — v(a)u'(b)
D = v(a)u(b) — u(a)v(b)
E = u'(a)v'(b) — u'(b)v'(a)
F = u(b)v'(a) — u'(a)v(b) (4.12.81)
are real numbers. Therefore,
,„l2 (E + k2D)2 + k2(F — C)2
' ' ~ (-Ek2D)2 + k2(F + C)2 • (4.12.82)
E ^ - E (4.12.85)
so t h a t
C=F
F = C
D = —D
E = - E . (4.12.86)
Solution
If the wavefunction for the state of the system is
V » ( « i , x2, x N ) (4.13.87)
PV = ± 1 V • (4.13.88)
We also have t h a t
PxP = - x P
^2 2
= - x .
q Xi
'
(4.13.89)
(4.13.90)
If we write out the expectation value on the right hand side of (4.13.89) and
apply the operator P2 — 1 we have
N
( N \
pp
pip,p^2^> ^j
( N \
± v , 5 > ( p i i p )(±)v>
\ i=i /
= - Y^qiXi^j
= -d . (4.13.91)
Solution
a) The current is given by
/= — — . (4.14.92)
Next we use the fact t h a t the time-reversal operator commutes with H to act on
(4.14.93) from the left with T and write (remember that the energy eigenvalue
E is real)
Using the properties of the time reversal operator we can rewrite this equation
as
In this case the current (4.14.92) clearly vanishes. Thus, we are left with the
second possibility.
2) T h e state is degenerate so t h a t
ip*¿eiaip . (4.14.97)
v>* = e i a I p . (4.14.98)
a2
Now, since the phase of ip is arbitrary we can multiply ip by e~' I . In t h a t
case we get
ia/2 ia ia 2 ia 2
e = e e ~ ' ip = e ' ip (4.14.99)
= ip . (4.14.100)
Solution
To solve this problem we need the potential energy V^(x) of the car t h a t is
travelling in the x-direction. We assume t h a t the edge of the cliff is located at
x = 0 , y = 0. T h e n for x < 0 the potential energy of the car is a constant
which we set equal to zero. For x > 0 the potential energy is m g y . But, the
classical trajectory of the car as it goes over the cliff is given by
2
X = vt , y = - ^ g t (4.15.101)
or
2
3 / = - ¿ z - (4.15.102)
V { x ) = ~ T ^ ¿- x 2 x>0 . (4.15.103)
¿v
So, the H a m i l t o n i a n is
H = { % 2 * - ° (4.15.104)
2
1 *>°
If we now set
k 2 = ^ ¿f , <y2 = ^ ¿ 4¿ (4.15.105)
ñ hv
t h e n t h e Schrodinger equation for t h e effective H a m i l t o n i a n becomes
d2ib ,
¿
+ k ip = 0 x < 0
ax
ip = e i k x f ( x ) x > 0 . (4.15.108)
T h e n , f ( x ) satisfies
/" + 2ik f + a2x2f = 0 . (4.15.109)
Therefore,
2a 2 + 2ika\ — 0
603 + 4ika-i — 0
2
6ika3 + a a o = 0 . (4.15.112)
T h u s , we find
• 2
ia
«i = - ^ « 0
a2 = ~ a 0 • (4.15.113)
So, the wavefunction near x = 0 is given by
1 -(- R — a0
ik(l-R) = ika0
r a2 1
I-TT4 (4.15.115)
ñ
a2
«o«l , = ¿ 4 - (4.15.116)
= | j R | 2 =
^ 1 ^ ' (4.15.117)
o m2q2 l9 2m(mv2/2)
to get
K 1 ( gh
=ió(^J • < 415U8 >
An order of magnitude calculation now shows t h a t
This shows why classical mechanics works so well when applied to macroscopic
objects.
Bibliography
[4.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 4.3 and 4.4, (2002) .
More One-Dimensional
Problems
oo
/
f(k)e^kx~u^ dk
-oo
is normalized such t h a t
r oo
/ mt,- , x)|2 dx = 1 .
J — oo
Assume /(A:) is a smooth function vanishing rapidly at infinity. Show that the
velocity of the centre of mass x defined by
oo
/
x)|2 dx
-oo
is given by
oo
/
«.»'<*-«)* dx = 2irS(k -q) .
-oo
then
dx duj
dt ~ cf/:
Solution
We first write out j*?!2 in detail.
i[kx—u>(k)t]
dk
/ oo /»00
r ( f f ) c — d q f ( k ) e'
-oo J — OO
OO pOO (5.1.1)
/
Now, / diwg / * ( g ) / ( a ) e '(fe-9)^ e -'^(fe)-w(g)]i
-oo J — OO
/ oo /«oo /«oo
•oo J — OO J — OO
/ OO roo poo
/ dkdqr(q)f(k)e-i^-u^t xe^-^dx. (5.1.2)
But, -oo J — oo J — OO
Ri(k-q)x _ ei{k-q)x
(5.1.3)
dq
Therefore,
/ oo /«oo i /»oo
/ dkdqf*(q)f(k)e~i^k'>-w^ti-j- / jV-ti'dx
-oo J—oo dq J—QQ
OO roo I
/
/
If we differentiate underd kthe
d q integral
f * ( q ) f {we - i ^ - uget
k ) enow ( 9 ) ] « f í - 2 j r í ( * - «) . (5.1.4)
-oo «/ — oo dq
dx
2n
dt / OO poo
/ dfcfg/*(?)/(*)[«(«)
•oo J — OO
(5.1.5)
Solution
The potential is
V = VQ exp (5.2.11)
The lowest lying states will be centred near x = 0. In the vicinity of this point
we have
1
l + -az2 (5.2.12)
This approximate potential is just the potential for a simple harmonic oscillator
with its zero shifted and with a spring constant k = Vo a. Therefore,
m = x/Voa/ra . (5.2.13)
Then,
E i « V0 + - hw (5.2.15)
corresponding to the first excited state of a simple hamonic oscillator with parity
= -1 .
5.3 Particle at Rest
Find the wavefunction for a particle at rest at the origin of a coordinate system
fixed in space.
Solution
We have to solve the eigenvalue problem
(5.3.16)
(5.3.17)
Solution
The potential is an even function of x so parity considerations m a y be used for
the phase shifts. But first we need the correct boundary conditions. These are:
1) xp(x)is continuous at both x = ± a .
2) The S-functions indicate t h a t there is a discontinuity in dip/dx at x = ± a .
As usual, we find this discontinuity by integrating the Schródinger equation
over a tiny interval t h a t includes the discontinuity. T h e Schródinger equation
in terms of the reduced variables
2 _ 2mE 2mAo
(5.4.18)
h2 ' K
~ h2
reads
(5.4.20)
x < —a
—a < x < a (5.4.21)
x > a
Here we have already imposed the physical boundary conditions t h a t for x < -a
we have an incoming and a reflected wave and for x > a we have a purely
transmitted wave. For — a < x < a we have a combination of waves travelling
to the left and to the right.
Applying the boundary conditions at x = —a we get
and
K + 2ik „ K ^ „
A = + R
^ r 2 T k ' <5-4-24>
and
r, K ->ikn „—I\+2ik
= + (5.4.25)
and
A B =
= ' ~ékTe2lka • (5-4-28)
Equating the two expressions for A and B we get two equations for R and T.
and
[ * ' + ( * • + 2,•*)»].-»»•
2iK2 sm(2ka) — 4k2 e~2,ka
and
{ cos(kx —
ylcos(fcx)
x < —a
{ isin(kx
Bsin(kx)
— S-)
—a<x<a
x < —a
(5.4.34)
isin(kx + ¿-) x > a
Writing out the boundary conditions at either x = a or else x = — a we get
Therefore, we get
tan (ka) = tan(&a + 6+) + — . (5.4.37)
K
So,
Proceeding in the same way for the negative parity solution we find
Therefore,
cot(fca) = cot (ka + J _ ) — — . (5.4.41)
k
So,
T= i [e2iS+ +e2iS~]
= arctan tan KA ka
k
and
1
S- = cot — cot KA — ka
k
to
r
(k2 — K< 22\-
)e — 2ika
T = 2 2
(k + A' ) + ikK(cot Ka — tan Ka)
and
ikK (cot Ka + tan I\a)e' -2 ika
R =
( k + A' 2 ) + ikK (cot Ka — tan Ka)
2
Solution
Our starting point is
Also,
á + = a r c t a n f — tan(A'a)] — ka
k
and
JS
S_ = arccot[—cot(A'a)] — ka .
t a n ( J + + ka) = — tan Ka .
K
But,
ei(i++ka) _ e-i(S++ka)
t a n ( í + + ka) = ^ei(i++ka) + g_i(i++/ca)j
e2i(S++ka) _ i
j[e2i(S++ka) _|_
Therefore,
i[e2'(s-+ka^ + 1]
c o t ( o _ + ka) — e2i{S_+ka) _ i]
K
— — cot K a . (5.5.50)
k
Therefore,
T = - (e2iS+ + e2i6~)
and
R = e
)
1 k + iK tan Ka k — iK cot Ka — 2 ika
k — iK tan Ka k + iK cot Ka
ikK [cot Ka + tan Ka\ 2ika
(5.5.53)
k2 + K2 + ikK [cot Ka — tan A'a]
i
oo if x < 0
V(x)
kx2 if x > 0
~{
Compare the zero point energy in this potential with that for the simple har-
monic oscillator with the same force constant k. Explain any differences you
find.
Solution
For x > 0 we again have just the simple harmonic oscillator solutions
This means t h a t the lowest energy is now the same as that of the first excited
state of the simple harmonic oscillator. Since the particle is more confined,
the ground state energy is greater. Furthermore all the even parity states are
missing.
5.7 Periodic P o t e n t i a l
Find the equation t h a t determines the allowed energy bands for a particle mov-
ing in the periodic potential (see figure 5.1)
V(x)
V0
a 2a + b 3a + 26 4a + 36 x
(a -(- 6) 2(a -|- 6) 3(a -(- 6)
0 0 < x < a
V =
Vo a < x < a+ 6
Solution
We first solve the problem in the interval —b<x<a and then use Bloch's
Theorem [5.1] to extend the solution to a < x < a + 6.
For 0 < x < a we have with
k\ = 2mE
V
that
The wavefunction and its derivative must be continuous. Imposing these con-
ditions at x = 0 and x = a we get
C j + C2 = C3 + C4 .
mci-c2) = mc3-c4) •
ikia ikia
Cle +c2e~ = eiK{a+b)[c3.e-ik*b + c4eik*b] .
kx [Cleikia -c2e-ikia] = k2 e
iK a+
( V [c 3 e~ik>b - c 4 eik*b] . (5.7.59)
To obtain a nontrivial solution for the parameters c¿ we require that the deter-
minant of their coefficients vanish
1 1 -1 -1
k1 —k 1 —¿2 k2
Jk ia _eiK(a+b)e~ik2b _eiK(a+b) eik2b = 0
e-ikia
kxeikia -kxe~ikia —jfc2 eiK(a+b)e-ik3b ¿ 2 GiK(a+b) e ik 2 b
(5.7.60)
Expanding this out and after some tedious simplifications we get that
¿fe? + k 2
cos I\(a + b) = cos k\a cos k2b—-—-— sin k\a sin k2b . (5.7.61)
ZK1K2
If E > Vo this equation determines the energy bands since it requires that
k\ + k2
— 1 < cos k\a cos k2b — • sin ¿ j a sin ^26 < 1 for E > Vo • (5.7.62)
2k\k2
For E < Vó we have to replace k2 by Í K 2 . In this case the energy bands are
determined by
k\
— 1 < cos k\a cosh k2b • sin k\a sinh K2b < 1 for E< Vq .(5.7.63)
1k\K2
ip = Teikx . (5.8.65)
J l ^ ~ ^ d x = 0=j{L)-j(-L). (5.8.68)
H'n(x) = 2n Hn_1(x)
b)
2x Hn(x) = Hn+i(x) + 2n Hn-i(x)
and finally
c)
H'¿(x) - 2xH'n[x) + 2n Hn(x) = 0 .
f h i s last equation shows that the functions Hn(x) defined by the integral
(5 9 71) satisfy the differential equation for the Hermite polynomials.
Solution
a) If we differentiate the integral representation under the integral sign we im-
mediately get
on roo
H'n{x) = = / n(x + i s ) " - 1 e~s ds = 2n Hn-i(x) . (5.9.72)
v71" j—oo
2 n +i r 0 0 2
2xHn(x) = —7=- / x ( x - f ¿ s ) " e _ s ds
v*" —00
on + l /-oo
= tfn+i(x) 7=r- / ¿s(x + ¿s) n e _ s ds . (5.9.73)
71
v " ./-oo
But,
oo /«oo
/
(x -f- zs) n de~s = / —2s(x + is)" e _ s ds . (5.9.74)
-oo J — OO
After an integration by parts this yields
oo ,00
/
is(x + is)" e~ s ds = — — / (x + i s ) n _ 1 e _ s ds . (5.9.75)
•oo ¿ J —OO
So, combining this with (5.9.73) we have t h a t
as desired.
c) Now, using (5.9.72) together with (5.9.76) we have that
Solution
Using the hint we have t h a t
[ x +
7/ ^ ^ l{xh)]dx =
-U [ x
^ +
i{xri)]hdx- (5 10 80)
- -
Using this we can rewrite the m a t r i x element as
(xp + px)<j>2)
1h d<f>2 d<j>\ „ d d
<PiX— x-j—<p2 + <p1—{x(p2) - -- yx<pi)(p2 dx
27 ax ax ax dx
h 2 d
Í l * ^ ^1 dx . (5.10.81)
~ i J
[ dx x
dx < 01
Furthermore, using the fact t h a t the two eigenfunctions are degenerate we have
h2 d2<f>\
+ V<¡>\ = Eft
2 to dx2
h2 d2(j>2
+ ^02 = E(p 2 . (5.10.82)
2m dx2
So, multiplying the first of these equations by ^>2 and the second by <j>\ and
subtracting we get
d2<
- *1 ( A d
^ l*
&1
t>2
= 0 (5.10.83)
2 m V 2
c t e 2
rfx2
(5.10.84)
c?x dx dx
From this it immediately follows t h a t
( 5 10 8 5 )
•
/ 2X
~ ^ dx =
° ' (5.10.87)
5.11 Hellmann-Feynman T h e o r e m
Consider a Hamiltonian H(A) t h a t depends on a parameter A and let A be some
operator such that [A, H] is well defined. Prove that for any stationary state
^(A) of this Hamiltonian we have that
dX " ' \ 5a / v 5a
This is known as the generalized Hellmann-Feynman theorem. For A = 1 it is
the Hellmann-Feynman theorem.
Solution
The equation for the stationary states reads
= + + A
lfx^ ' (5.11.89)
But,
e
»-4i> =
= (ip, HA--) . (5.11.90)
Bibliography
[5.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 5.8, (2002) .
Chapter 6
Mathematical Foundations
Solution
We are given a sequence of vectors {vj j = 1 , 2 , 3 , . . . } where each of these
vectors is an element of an r-dimensional vector space and may therefore be
written for a fixed basis in terms of its components as
Vj = {vji,vj2,...,vjr) . (6.1.1)
Furthermore, we are told t h a t these vectors form a Cauchy sequence. This means
that given an e > 0 we can always find a number N such t h a t for n,m > N we
have
(l^nl I "i- | ^ n 2 ^m21 "I" ' ' ' ~t~ | ^ n r Vmr \ ) ^ ^ * (6.1.2)
is finite since we would now have to sum over an infinite number of components.
This is w h a t makes completeness in an infinite dimensional space non-trivial.
64 ( J H A f l ' E t í O. MAL nrAVl/l t iua L tXJUSXUAí IV1\S
Solution
T h e proof is by displaying the operator S. In the Schrodinger representation
we have t h a t
P = - 4 - , x = x. (6.2.5)
i ax
Thus, we try with a real d i f f e r e n t i a t e function h(x) ,
5 = , 5_1 = S* = eih(-x) . (6.2.6)
Then, for a function g in the domain of p we have
' s = Is
= eih^\p+h'(x)}g(x) (6.2.7)
where
v w - | w .
Thus,
SpS-1 =P+h'(x) . (6.2.8)
6.3 D e g e n e r a c y and C o m m u t a t o r
Show t h a t if all eigenvalues of the self-adjoint operator A are non-degenerate
and t h a t if the self-adjoint operator B commutes with A, i.e.
[A, B] = 0
then B is a function of A.
74. VWr* I*i^uJnni*IU O ü A / l i r u
'
Solution
We have
The kets |n) form a complete set. Also, using the commutator we have
Thus,
Now, corresponding to any ket |n) there is a unique eigenvalue an. Con-
versely, corresponding to any eigenvalue an there is a unique ket |n) . This
means that there is a one-one correspondence between n and an. So, we can
write
c„ = c ( a n ) . (6.3.14)
and since, as stated, the |n) form a complete set we have that
B = c(A) . (6.3.16)
An = px2n+l + x2n+1p , n = 1, 2, 3, . . .
where
h d
P =
~idi '
Find the eigenvalues and eigenfunctions of An. W h a t are the deficiency indices
°f An? The Hilbert space in this case is C2(—oo, oo). For n = 1 this example is
due to von Neumann.
uu wiini J. JJJJ.i v. lVAsxx njuiris*.± ik^Í\LJ U¿M- lUISS
Solution
We write the eigenvalue as Ah. Then, the eigenvalue equation for An reads
2n-f 1 d/ _=
? a ( vx 2 « + 1 / ) + ?. Ah f . (6.4.17)
dx i dx
Separating we get
1 = !ax-(2n + l)dj, _ . (6.4.19)
/ 2 2x
This requires t h a t 9(A) < 0 so t h a t the singularity at x = 0 does not cause the
function to blow up too strongly at the origin. Therefore, all eigenvalues are of
the form Ah with
This means that the deficiency indices are (0,1) and there are no self-adjoint
extensions of this operator. This example shows that, given two observables A
and B, with corresponding self-adjoint operators A and B, it is not generally
true that the operator AB + BA is self-adjoint and corresponds to an obervable.
6.5. PROJECTION OPERATOR 67
X
~ H / N<F>N(X)<T>N{Y)LP(Y) DY = ^ ANPN1¡J .
n n
This is called the spectral resolution of the operator A.
Hint: Assume completeness of the eigenfunctions.
Solution
We are given
Pn1> = <t>n(<t>nA). (6.5.25)
^ = Y^an(j)n (6.5.27)
n
so that
Then,
V = (6.5.29)
n
and
= y^tA<í)n(<t)n,ip)
n
^ ^ (fan (<ftn ? *0)
n
X n P n
= J 2 ^ - (6.5.30)
68 CHAPTER 6. MATHEMATICAL FOUNDATIONS
6.6 Spectral R e s o l u t i o n
Find the spectral resolution (see problem 6.5) of the operator
A _
— (, 0- O
3 a
i + ia2
~~ ai~ — ia io
2 -a3
Solution
The operator is
as a
¿ = ( . i + m2 ) (6.6.31)
a\ — la 2 — a3
The eigenvalues are given by
- ( a 3 - x ) ( a 3 + x) - (a 2 + a 2 ) = 0 (6.6.32)
or
x2 = a 2 + a 2 + a | = a 2 . (6.6.33)
So that
x = ±a . (6.6.34)
P- = = ,a as
v («1+^2) \ (6.6.36)
2a V ~ ( a i - ía
2) a + a3 J
So, it is easy to see t h a t
aP+ + (—a)P_ = A . (6.6.37)
Solution
If A is a square n x n matrix then
may be written as
matrix of cofactors o f M — z)
1 ( 6 9 )
= d^m-z) ' "
det ( 4 - *) = 0 . (6.7.40)
H = ~*
2m dx2
defined on the interval (a,b).
Hint: It may be useful to express the boundary conditions on a function / £ DH
in terms of 2-component quantities
_ ( /(«) ^
and
_ ( m
F(a) = { W J = { fib)
Solution
The Hamiltonian is
h2 d?
H (6 8 41
= • - - »
To find the deficiency indices when the Hilbert space is L2(a,b,dx) we have to
find solutions of
f± — A exp(k±x) (6.8.43)
70 CHAPTER 6. MATHEMATICAL FOUNDATIONS
where
t 22 1.2
H k± = ±i . (6.8.44)
2m
Therefore, we find
_ v w e ,- 3 i r / 4 ^ _a/2^^/4 (6.8.45)
h h
So, we need
The same boundary conditions must apply to both f ( x ) and 5(2:). We therefore
define
F
«=((i-'$'(«)) ' GW = U-$-<«)) ,6 8 49)
--
(the factors of 6 — a are simply to have both components with the same dimen-
sion) and look for boundary conditions of the form
G(b) Ft(a)
FHb) ( J V ) = ( l "o1 ) g ( a )
' (6 8 51)
' '
A A =
' ( 1 ~ o ) ( i ~o ) • <6 8'52)
A = eia ( a° + 0,3 ai
~ ta2 ^ (6.8.53)
\ aj + ia2 a0 — a3 J
6.9. ADJOINT OPERATOR 71
0 -1 \ , f 0 - a 2 + a\ + al + al
1 0 ) a = 0
0 -1
(6.8.54)
1 0
This fixes the matrix A and since it has exactly 4 parameters as required we
are finished. Thus, the domain of H consists of all twice differentiable functions
f(x) on the interval (a, 6) such that
ih i { % ( k \ ) = e
'" ( a ° l a 3 a i
~ t a 2 ) (ih \ v6-8-56)
(b-a)f(b) J \ai + ia2 a0 - a3 J \ (b - a)f (a) )y '
where a , a i , a2, a3 are real and ao = y/l + a\ + a^ + a§.
Find the adjoint operator at by explicitly giving its action on this basis set.
Also find the commutator [a, a*] .
Solution
We have that the set {«„} is complete and orthonormal. Also,
Therefore, we get
This last term on the right may be rewritten as y/m+ 1 i m + i , n Using the defi-
nition of the adjoint operator we therefore have t h a t
Therefore,
aat un = Vn + 1 a u n + i = (n + 1) u„ (6.9.61)
and
a t a u „ = y/na) w„_i = nu„ . (6.9.62)
Therefore,
Thus,
[a, at] = 1 . (6.9.64)
A|n) = a„\n)
then
A
II ~üm (6.10.65)
dn 0-m
m^n
is a projection operator onto the eigenspace of a n , that is it projects onto the
space spanned by the states belonging to the eigenvalue an.
Hint: Prove that
Solution
To use the hint we consider a eigenstate |i) , I ^ n. In that case
m
Pn |/) = J J ' |/) = 0 (6.10.67)
an ctm
m^n
since in the product we arrive at a term with m = I. Now consider the eigenstate
I ft). Then,
Pi |/> = SnlPn\l)
= snl\l)
= Pn\l) • (6.10.70)
Since the eigenstates of A form a complete set we may expand any state in
terms of these
Then,
pn\4>) = X>P„|Z>
i
=
i
= bn\n) . (6.10.72)
Pn=Pn • (6.10.74)
n n I
n I
6 Z
= E ' I > = I<¿> • (6.10.75)
I
Thus, we have shown that
P = 1
E « (6.10.76)
n
6.11 C o m m u t a t o r of Lz and (p
Formally one can derive t h e relation
[Lz, <p] = ih
and deduce f r o m it t h a t
Explain this a p p a r e n t p a r a d o x .
Hint: E x a m i n e the d o m a i n of Lz on which it is self-adjoint. A similar a r g u m e n t
also holds for [p, ar] = ih and a particle confined to a finite interval on the line.
See also [6.1] and [6.2].
Solution
Lz is defined as (h/i) (d/dtp) for f u n c t i o n s f(ip) differentiable and periodic in ¡p.
T h i s defines the d o m a i n of Lz and m e a n s
is incorrect .
Solution
We know from general considerations [6.3] that
h,
A L z A sin ip > —(cos ip) (6.12.81)
This does not lead to any contradictions. For example, if we are in an eigenstate
of Lz s 0 that the wavefunction is
m
^ * (6-12.83)
This time, however, no contradiction occurs. The results simply mean that the
uncertainty relations read 0 > 0, a statement t h a t is clearly true.
T = -(/i2/2m)V2
then, near r = 0 the function ip(r, 9, ip) must be bounded by A ra , a > —1/2
where A is independent of r.
Hint: Examine the condition that
Ti¡>(r,.,.) G £ 2 (0,OO)
near r = 0 by writing
where the a,¡ are independent of r, i.e. they depend only on 9 and ip.
Solution
Using the hint we have t h a t near r = 0
where the bj are new functions of 6 and <p and are independent of r just like the
a j . Since the integration over r has the measure r2 dr it now follows t h a t near
r = 0 we have
a > ~ \ • (6.13.88)
H =T + V
we have a potential V(r) such that V(r) is less singular than r~2 near r = 0
then the condition (6.13.88) obtained also sufRcies for ip to be in the domain of
this Hamiltonian, at least as far as the behaviour near r = 0 is concerned.
Solution
For these computations it is convenient to choose units such t h a t h = 1. Then,
2 d2
P
dx!
í>4 = • (6.14.89)
To find the deficiency indices we need to find the solutions belonging to £2(0, 00)
of
^ = (6.14.90)
We try
f± = ea±r (6.14.91)
and find that
Thus,
i ± n 8
Q± — e / ^ ei±*/a+iv/2 ^ e«±jr/8+»ir ei±7r/8+»3jr/2 (6.14.93)
^ 6.14. SELF-ADJOINT EXTENSIONS OF p4 77
and
—in/8+in —in/8+i3ir/2
Ct — — 6 J c (6.14.95)
So, the deficiency indices are (2, 2) and we have a 2 x 2 = 4 parameter family
of self-adjoint extensions.
To find the self-adjoint extensions we consider the matrix element of p4 and
integrate by parts several times. Thus,
oo d4ip ,, d3ip loo
d
d<p* dd3rprp ,
[°° d<f>*
—-4 ax =
T dx * dx3 J0 3
oo
0 •°° d2r d2i/>
00
xpw ipW + dx
0 0 dx2 dx2
[°° d3<j>* dip
v>(3) v<2> 00 +
0 0
, V
dx
JQ dx3 dx
0 0
v>(3) 00 v-<2> 00 + .*<»>• v r
0 0
I J
d
+ 'v 1 (6.14.96)
<l>*(0)ip(3\0)—<f>^*(0)ip(2\0)+<p^*(O)ip^(O)—^3^*(0)ip(0) = 0 .(6.14.97)
To satisfy this condition we insist that all ip G £2(0,00) should be of the form
If we now cause all three of the square brackets to vanish separately we have
that
^x)(0) = a v ( 0)
V)(2)(0) = -6*V( 0)
c = ab* . (6.14.102)
Writing,
we have that
c=\ab\ei^~l) . (6.14.104)
It is also clear t h a t the four real parameters |a|, |6|, /?, and 7 label the self-adjoint
extensions.
Bibliography
[6.1] P. Carruthers and M.M. Nieto - Phase and Angle Variables in Quantum
Mechanics, Rev. Mod. Phys. 40, 411, (1968) .
[6.2] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., chapter 6, (2002) .
Physical Interpretation
Solution
Since each side of the die is equally probable and the sides are numbered from
1 to 4 we get
( n ) — 4 ( 1 + 2 + 3 + 4) — - (7.1.1)
Similarly,
An = (n2)-((n»2
\ [ l 2 + 22 + 32 + 4 2 ] -
(7.1.3)
, 1 . i\J2 y/E
tp = -U 0 (x) + — Ui(x) - — U 2 (x)
80 CHAPTER 7. PHYSICAL INTERPRETATION
Solution
Since uo, Mi, «2 are normalized and the moduli of their coefficients add up to
unity, the whole wavefunction is normalized.
a) The most likely value of the energy corresponds to the eigenfunction with
the largest coefficient ie. «2- This value is 5 / 2 huj with a probability of
I — a/6/31 2 = 2 / 3 .
1 lt 2 3. 6 5, Q7
(E) = x x
= ' <7'2-4)
9 * 2 "*"9 2 9 2
The probability of observing this value is zero since it is not one of the eigen-
values.
c) The resultant energy value obtained is 3/2 hui. The wavefunction is
d) The wavefunction is
Solution
The commutation relations for the angular momentum operators are
[Ly, Lz] = ihLx . (7.3.6)
[Lz , Lx] = ihLy . (7.3.7)
FREE PARTICLE PROPAGATOR 81
7-4-
= 0 . (7.3.9)
In exactly the same manner, taking the expectation value of (7.3.7) we find that
(m\Ly\m) = 0 . (7.3.10)
$ (0, x) — S(x — a) .
Find the wavefunction for t > 0. This solution is called the free particle propa-
gator.
Hint: To evaluate an integral of the form
f
J —C
exp{j'Ax 2 + i/3x} dx
Solution
The eigenfunctions for a free particle are
k x
uk(x) = — e ' (7.4.11)
v27t
with energy
h?k 2
• p-4-12)
Therefore,
1 i°° — ika
dx = (7.4.14)
A { k ) = 8 [ x
^ L ~a) e
~,kx ~
Then, we find that
/ °°
1 /*°° 2
= _ / e>'fc(i-a) e-i7ifc í/2m ^
2?r 7 _ o o
This integral is not obviously convergent. To remedy this we replace the co-
efficient —iht/2m of A:2 in the exponential by —iht/2m — f ,e > 0 , with the
intention of letting e go to zero after the computation is done. This process is
called, "regularizaron". Thus,
i r°°
w(t,x) = — eik(x-a) e-i(htl2m-U)k* dk
2?R J — OQ exp
2ix y e + iht/2m — a)2
,m(x 2ht
1—i ,m(x — a) 2
exp 1 (7.4.16)
~2^ 2ht
It is in the last step t h a t we used the fact t h a t t —> 0 + to obtain the correct
square root .
II/II2II<HI2>I(/,<7)I 2
and put
f = xip , g = pip .
Solution
We s t a r t with f ( x ) = xip(x) , g(x) = pip(x). For t h e Schwarz inequality to
become an equality requires t h a t
g = Xf (7.5.17)
where A is a c o n s t a n t . T h i s m e a n s t h a t
Furthermore,
$t(xip,pip) = 0 . (7.5.20)
This m e a n s
or
Now,
Adding we get
Therefore,
h dip
— — = iaxip(x) . (7.5.26)
idx
So finally,
2
ip(x) = Aexp ( )l (7.5.27)
2h
$ ( 0 , x) — [27tL 2 ] 1/i4
exp — (x/2L)2
Solution
We first express ^ ( 0 , x) as a superposition of free particle states. Since the most
general solution for a free particle is
(7.6.28)
/ OO
, fe:E fifc2f 2m
A(k) e '( - / ) dk
•OO
we write
(7.6.29)
/ OO
A(k) eikx dk .
•OO
Then, the coefficient of A(k) is determined from the explicit expression for
tf(0,x).
i r°°
A(k) = — <I<(0 ,x)e~ikxdx. (7.6.30)
2n
J-OO
Inserting this in the integral and completing the square in the exponent, we can
evaluate A(k).
(87ri2)1/,4exp
4(L2+iht/2m)
*(*,*) =
2n
ix
e x p -(L2 + iht/2m)
C { - 2(L2 + M/2r
(87TZ,2)1/4 exp [ - 4 ( L ! + x ¡ t / 2 r n )
(7.6.32)
27r L2 + iht/2m
$ ( T , X ) = (2KW2)~1^4 AA
exp (7.6.33)
\w2 )
7.7. TIME-DEPENDENT EXPECTATION VALUES 85
where
2
h2t
= ¿ 2 + £ í 27 22 (7-6-34)
Am L
and
"1 f ht \ htx2
A = — arctan 4- (7.6.35)
2 \ 4 mL2 J 8mL2w2
Jt(A)t = ^(V{t,x),[H,A]y{t,x)) .
Solution
In this case we have a free particle so
(7-7-38>
Also we are using the equation
~(A) = ^ , [ H , A } ^ ) . (7.7.39)
Since
[H,p] = 0 (7.7.40)
86 CHAPTER 7. PHYSICAL INTERPRETATION
jt(p) = 0 (7.7.41)
so that
= (7.7.43)
so that
0
= <£> = • <7 7 44
»
Therefore,
(P 2 )t = (p 2 )o - h 2 \ A \ 2 J s i n 2 ( 7 T x / a ) - j ^ (sin 2 (7rx/a)) dx
a
2n2
/
sin 2 (7rx/a)—Y {cos2 (nx/a) — sin 2 (7rx/a)) dx . (7.7.46)
-a ^
2 2
Now, we rewrite cos as 1 — sin and use the fact that
. 2\ 4ft27T2 2h7TT2 2
<A = — 2 IT-«HI 2 - (7-7.48)
<A = ^ ( 2 - 4 / 3 ) = ^ . (7.7.49)
We now find
4-px]
[H,xp + rurl==-2ih~
—9¿/„ . V2
(7.7.50)
2m
Therefore,
d
\ 2
/ 2x 2(2nh) 12
-{ l p + p , ) = - - & ) = - L _ f . (7.7.51)
7.8. EHRENFEST THEOREM
So, we get
. . 2(27T h)2
(xp + px)t = (xp + px)o + , t
3ma,2
We now compute
(xp + px) o
Therefore,
2(2tt/í) 2
(XP + Px)c — -3^2" ' •
Finally we find
iH
[H, x2] = x[H, x] + [H, x]x = — — (xp + px)
Hence,
d , 2, *' - i h , v 1 2(2tt/í) 2 4
- x ) = (xp + px) t = ^ f-i .
dt nm m 3mcr
Therefore,
/ 2\ _ / _ 2 \ 1 (27rfi)~
(x ) t _ (x >0 + 3 m 2 f l 2 j 2 .
Now,
7.8 Ehrenfest T h e o r e m
Consider a particle under the influence of a Hamiltonian
H = ^ + V(x)
so that
88 CHAPTER 7. PHYSICAL INTERPRETATION
Show that if (x) is the "centre of mass" of the wave packet and (p) the average
momentum of the particle then
d(x) _ (p)
dt m
and
m = <*•(«)> = _ < i ! £ > ) .
Solution
We have
n2
H = £ - + V{x). (7.8.60)
¿m
Therefore,
[H,p] = i h ~ . (7.8.62)
= !([«,«]) = * H » , ( £ ) = M (7 . 8 . 63)
and
Fw = f (m) + £ - (7.8.66)
7.9 Compatibility T h e o r e m
Prove the compatibility theorem which states:
The following statements are equivalent for any pair of observables A, B.
i) A and B are compatible.
ii) The operators A and B, representing the observables A and B, possess a
common eigenbasis.
iii) A and B commute, that is
[A, B] — AB - BA = 0 .
Solution
The compatibility theorem states that, given two observables A , B represented
respectively by the self-adjoint operators A , B the following statements are
equivalent.
i) A and B are compatible observables.
ii) A and B possess a common eigenbasis.
iii) [A, B]= 0 .
We begin by recalling that two operators A , B are compatible if a measurement
of A followed immediately by a measurement of B and another measurement of
A yields the same result for the value of A in both cases. To prove the theorem
we now show t h a t
i) ->• ii) - » iii) - y i).
i) - » ii)
Suppose the measurement of A yields a . Then, immediately after the measure-
ment the state of the system has to be an eigenstate of A say <j)a. To guarantee
that after the measurement of B when we again measure A we again get the
value q requires that after the measurement of B the state be 4>a¡p where
B<t>a,p = (7.9.68)
and
^a,/3 = • (7.9.69)
This means that when acting on an eigenbasis [A, B] = 0. Since the eigenbasis
forms a complete set this means that quite generally
[A,B] = 0 . (7.9.71)
iii) -> i)
Since [A, B] = 0, the two operators may be diagonalized simultaneously. Then,
90 CHAPTER 7. PHYSICAL INTERPRETATION
Solution
Since,
[H,A] = 0 (7.10.72)
it follows that
5<A> = ° (7.10.74)
as well as t h a t
(7.10.75)
±(AA)>=±{{A-{A))2) = 0 (7.10.76)
^ ( O , ^ ) = [27tL2] 1/i4
exp — (x/2L)2 .
Using the experience from problem 7.6 answer the following questions:
a) W h a t is (x) at any time?
b) W h a t is Ax as a function of time?
This is known as the spreading of a wave packet. If this packet corresponds to
an electron with 20 eV energy and a width of 1.0 Á at t = 0, what is its width
after travelling 100 m?
7.11. SPREADING OF A GAUSSIAN WAVEPACKET 91
Solution
We are given as initial condition
Now following the steps in problem 7.6 we first Fourier transform Vf(0,x).
and
hp2t/2m)
dp (7.11.79)
/ OO
i{px
F{p) e -
•OO
Thus, carrying out the integral, we see that the wavefunction after a time t is
again a shifted Gaussian of the form
-(k-p)2L2 -x2 + 4ikL2x + 4k2 L2
2\-i/4 2 exp . (7.11.80)
«(*,«) (27tL , [i + iht/2mL ] 2
4 (L + iht/2m)
From this we compute that
.x . hkt
( ) = t;— (7.11.81)
2m
and
h2t2 \
(Ax) 2 = (x2) - (x)2 = L2 I 1 + (7.11.82)
4 in2L4 •
10
For the problem under consideration we have L = 1.0 Á = 10 m . The
speed of the electron is given by
E = -mv~ (7.11.83)
2
Substituting 20 eV for the energy we find that v = 8.4 x 10® m / s . Therefore
the time to travel 100 m is
ht
Ax 69 m (7.11.85)
2mL
This means t h a t in going down a 100 m beam tube the wavepacket has spread to
macroscopic dimensions. The reason for this is that originally it was extremely
narrow and therefore had a very large spread in momentum. If instead the
wavepacket originally had a width of 1.0 m m the width after travelling down
the tube would have increased by a negligible 6.9 x 1 0 - 3 m m .
92 CHAPTER 7. PHYSICAL INTERPRETATION
[E,t] = ih
Ef(E) = Ef(E)
and
Show that the operator t so defined cannot be an observable, i.e. that it has
no self-adjoint extensions. This proves that if the energy has a lower bound, a
relationship such as
[E, t] = ih .
Solution
We first find the deficiency indices for the putative time operator —ihd/dE.
Thus, we look at
- i h ^ f ( E ) = ±iTf(E) . (7.12.86)
Only the first of these is square integrable on the interval (0,oo). Thus, the
deficiency indices are (0,1) and this operator has no self adjoint extensions.
Since for any physical system the energy must have a lower bound we can
always shift the minimum of energy to E — 0. This allows us to consider the
space £ 2 ( 0 , 0 0 ) as the appropriate space for this problem. Our result means
that the relationship
[E,t] = ih
is false.
Find the length of an interval around the origin such t h a t the probability of
finding the particle in this interval is 40 % .
7.14. SPHERE BOUNCING ON SPHERE 93
Solution
We have the unnormalized wavefunction
oo poo
/ 2 2
\f(x)\ dx = 2\A\ e~2axdx= 1 (7.13.89)
or -oo J0
A = V¿ . (7.13.90)
Now, the probability that the corresponding particle is found between — L <
x < L is given by
1 _ e - 2 « ¿ _ o.4 . (7.13.92)
1 = - ® ! ! . (7.13.93)
2a
Solution
Assume the ball falls in the y-direction and is deflected in the x-direction. The
ball actually falls a distance of 1 m. Its terminal momentum just before contact
is
3
Py = m\/2g(l — 2r) « 10 x \/2 x 10 x 1 kg m / s
-3
= 4.47 x 1 0 kg m / s . (7.14.94)
Suppose t h a t on contact the vertical makes an angle Atp with the line connecting
the two centres (see figure 7.1). Then, on bouncing, the angle increases to 2A<p.
After n bounces the angle will be nAip.
CHAPTER 7. PHYSICAL INTERPRETATION
Figure 7.1: Sphere bouncing on a sphere. The angle <p is the same as the angle
A f in the text.
Now, if the deflection is Aip, then the momentum in the x-direction after
the collision is
since Aip is a very small angle. The angular momentum of the upper ball with
respect to the point from which it was dropped is
On the other hand, for very small angles we have the uncertainty relation
Therefore,
h
(A r ) 2 « 1.1 x 1(T 3 2 . (7.14.99)
2(1 — 2r)m\/2g(l — 2 r)
So,
n = JL « « 1014 (7.14.102)
aip 10-16
7.15. CLOUD CHAMBER TRACKS 95
Solution
The momentum of the electron in the direction of the track (say x) is given by
px = V2mE
= y/2 x 9.1 X 10-31 X 1.6 X 10-19 X 1.0 x 10 3
-23
= 1.71 x 1 0 kg m / s . (7.15.103)
This is also the momentum for the classical trajectory. Quantum mechanically
there is also a transverse component (say py) of the momentum given by the
Heisenberg uncertainty principle.
ApyAy « h . (7.15.104)
=
Apy ~ -7— = ^ ^ ^ ^ 1-05 x 1 0 - 2 8 kg m / s . (7.15.105)
Ay 10-6
This means t h a t the fractional change in the momentum is of the order of
^ « 6 x 10" 6 (7.15.106)
Px
and hence, the trajectory is essentially the same as the classical.
Solution
a) Since the particle has spin h the outcome for the measurement of spin along
any direction must be one of h, 0, or — h.
b) Since we obtain the largest possible value, in the initial measurement, we
96 CHAPTER 7. PHYSICAL INTERPRETATION
obtain h. To compute the probabilities for the new measurement we call the di-
rection of the first measurement the z-direction. In that case the corresponding
eigenfunction is
ip = | 0 (7.16.107)
0
We let the new direction lie in the x-z plane. In that case the appropriate
angular momentum operator for this measurement is
( cos 0 0
72sin
Ln = Lz cos 9 + Lx sin 9 = h 0 (7.16.108)
72sin( 72sinf
0 ^ sin 9 — cos 9
ip (7.16.109)
| ( 1 + cos 9)
= ^ sin 6 (7.16.111)
| ( 1 — cos 9)
0 h:
a cos 9 + —= sin 9 b = 0
v2
—= sin 9 b — cos 9 c = 0 . (7.16.112)
y/2
-h:
and are
pi = i ( l + cos0) 2
4
P0 = ^ sin 2 9
0 otherwise
a) Find the probability P„(0) that at t = 0 a measurement of the energy yields
the value
h2n2 7T2
= 2
" 2m i '
Solution
The eigenvalues and eigenfunctions for a particle in a box of length L are
h2n7 7T2
and
un = \ j — sm(nnx/L) . (7.17.119)
v jl/
98 CHAPTER 7. PHYSICAL INTERPRETATION
r
Pn( 0 ) ik,$(x,o))|2
_2_
sm(mrx/L) dx
L2
Jo
r [l — cos(n7r)]¿
n
T&r — 1,3,5,...
(7.17.120)
0 n = 2,4,6,...
«„( x)e~iB*t'h
2V2
CL = ($(x,0),M¡(x)) 1 = 1,3,5,.. (7.17.122)
In
Then,
y t ci(un,ui)
1=1,3,5,...
2
Y] c;¿ní
/ = 1,3,5,...
n = 1,3,5,...
(7.17.123)
0 n = 2,4,6,...
as before.
2) A simper method is to realize that the evolution is unitary. This means that
both the eigenstates un and the total state $ evolve with the unitary operator
Pn(t) = |(l/(í)^(x,0),í/(í)«„(x))|2
= |K,^(x,0))|2 (7.17.125)
= (7.18.126)
at 2m
is form invariant under a translation
x —»• x' = x — vt
y -»• y' = y
z z'= 2 . (7.18.127)
That is, show that if the wavefunction in the new coordinate system is Í '(x1, y', z', t)
then it satisfies the Schrodinger equation
DV H2
i h ^ = ~ — W . (7.18.128)
dt 2m
Hint: The wavefunction transforms as a scalar with an additional phase trans-
formation.
b) Apply this result to a plane wave and discuss the results so obtained.
Solution
a) Using the hint we write the wavefunction in the new coordinate system as
^ ' ( x ' , y'> z ' , t ) and relate it to the wavefunction in the old coordinate system by
an additional phase factor. Thus,
* ( x , t ) = e-,'/(*'-t>tf'(x'It) . (7.18.129)
Since the transformation is only on the x coordinate, we need not consider the
y and z coordinates and so we choose / independent of y and z. Then,
dV dx' d^_.fdj_d¿_ df\ ¡
e~if
dt dx' dt dt \c?x' dt ^ dt )
dv dv .( df df\ /
e~i} (7.18.131)
-vd^ +
- d T - l
[ - v
d ^ +
m r
100 CHAPTER 7. PHYSICAL INTERPRETATION
Also,
DV_ d*' . df ;
= e -if j——w (7.18.132)
dx dx' dx'
and
D2V D2V •D2F , , D
F .2 df d*'
= e~i} 2 2 + ( ^ 7 ) 2
p ' " ^ (7.18.133)
2
dx dx' dx'
.ihv = 2 (7.18.135)
2m dx'
DF D F
J J. \ y (7.18.136)
2m dx'> dx'2
N V +
[ - D * M ) =
mv da mv2
(7.18.139)
dt 2h
Thus,
,,, mir
a(t) = 1 (7.18.140)
V;
2h
where we have set the arbitrary constant of integration equal to zero since it
would only yield an irrelevant constant phase factor. So, finally
This means that in the moving frame (primed system) the frequency and wavenum-
bers are given by
, mv 2
UJ = UI
2n
, < i
i +
x
~ x
=
BY
k'z = kz . (7.18.146)
then the wavefunction ^'(a:', t) in the new coordinate system satisfies a Schrodingei
equation of the form
d^'(x' t) h2 d2y/'(x' t)
102 CHAPTER 7. PHYSICAL INTERPRETATION
where
V(x') (7.19.151)
is the potential for a constant force t h a t produces a uniform acceleration a.
Hint: In transforming the wavefunction to the accelerated frame allow for a
phase factor so t h a t wavefunction in the stationary frame 9(x,t) and the wave-
function in the accelerated frame are related by
Solution
We begin with equation (7.19.152) and rewrite it using (7.19.149) as
Then,
09 d9' d9' dx'
P-'F + ]$/
dt dt dx' dt dx' dt dt
p-ij d9' d9'
+ at (7.19.154)
~dT dV
Also,
09_
= e~iJ 99' .df
I—— (7.19.155)
dx dx' dx'
and
d29 = e~iJ d29' df d9' f.d2f .df
- 2 i (7.19.156)
dx2 dx'2 dx' dx' dx'2 dx
Substituting this into the Schrodinger equation we get
d9' d9'
ih + at — i at 9'
dt dx' dx' dt
d29' . df 09' ,d2f
'' vvdx' / (7.19.157)
2m dx'2 dx' dx' dx'2
Therefore, to recover a Schrodinger equation, with the possibility of a potential,
in the new coordinate system we need
. d2 f
2m L
D ^
+ (ÉL)'
\dx' J
—h A T
df
D * +
Of
M
= V(x'). (7.19.159)
20. THE WIGNER FUNCTION 103
and
db ma2t2
(7.19.163)
dt ~ 2/t~
So,
b{t) = ma2t3
- ^ T (7.19.164)
+ ( 7 ,9,66)
where, as stated,
QX / H
W{X, P, t ) = ~ J d3q * ( p + q/2, t ) * * ( p - 9/2, t)
104 CHAPTER 7. PHYSICAL INTERPRETATION
b)
J d3pW(x,p,t) = \V{x,t)\2 .
c
)
Solution
We begin by writing out the Fourier transforms.
Therefore,
W(x, p, t)
= ~ J d3yd3q d3k e-'P-y/" ei^+y/2)/K e~ik\x-y ¡2) / h t)
= ~Jd3ke2i{P-h'>£/n9(2P-k,t)9*(k,t). (7.20.169)
We now let
p— k = q/2
so t h a t
(2nh)3 [ ,
= - J ^ - J d3y9(x + y/2, t)9*(x - y/2, t)S(y)
2
= 9(x, t)9*(x, t) = | $ ( x , i)| . (7.20.171)
7.21. PROPERTIES OF THE WIGNER FUNCTION 105
J d3p W(x, p, t)
gives the probability density for finding a particle at the point x at the time t.
c) Here, just as in part b) above we simply integrate over x. Thus,
now gives the probability density for finding a particle with momentum p at the
time t.
* ( « ) = Axe~r2'2a2
Solution
Using the suggested wavefunction we have
W{x,p,t)
Here,
P= (£> V, 0 •
106 CHAPTER 7. PHYSICAL INTERPRETATION
W{x, p, t)
2 2
= M ! J dtdT)dCeiP P/h (x2 - í 2 /4) e " 2 ^+i /4)/2a
2 2
e-2(yw/4)/2a e _2(, +c 2 /4)/2 a 2
= M! J die-*'*"1 (x*-e/4)e-**'a3e-?'4a3
2 2 2
x J dr, e-ip^'h e -»
a
/« a e -i a /4« a J d( e"- < p « c / f i e - ^ / « e - c / 4 a
So, for
2z 2 2a 2 p 2 ,
1 — < 1
a2 2
/i
we get negative values. This means that, in this case, W(x, p, <) cannot be a
true probability density. Notice the region for which we get negative values is
the interior of an ellipse in the x — px plane with semi-major and semi-minor
axes a/\/2 and fi/(\/2a). Inside this region we would have
\W(x,p,t)\2
(7.22.175)
m
u 2N 3
and states that a particle can not be localized in a cell of phase space smaller
than (A/2) 3 .
Hint: Use the Schwarz inequality.
BIBLIOGRAPHY 107
Solution
The Schwarz inequality states that for any two square integrable functions /
and g we have
where
( / , 9) = J f'(x)g(x)d3x .
We now choose
h
FW = ^{s-Y/2,T)E^yl
Then,
(/,<?) = +
= W(x,p,t) . (7.22.178)
So,
26
J d3(y/2)\V(x + y/2,t)f J d3(z/2)\V* (x - z/2, <)|2
he _
6
|J . (7.22.179)
Bibliography
[7.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 7.9, (2002) .
/ OO
f(x)S(x-a)dx — f(a) (8.1.1)
b) -OO
c)
f(x)S(x — a) = f(a)S(x — a) (8.1.3)
d)
OO
/
S(x - y)S(y - z) dy = S(x - z) (8.1.4)
-OO
Hint: Integrate both sides of the equation with a well-behaved function f(x) .
81. PROPERTIES OF THE DELTA FUNCTION 109
Solution
a) We begin with
(8 L6
m . ) = { * ¡:¡>: »
Therefore,
f(
lim f >X + a)
dx = lim f(a + 9e) = / ( a ) . (8.1.8)
e-t-0+ J-e 2e f >0+
-
b) In this case we consider
= -/(0) . (8.1.9)
a
Here we have changed variables from x to z = ax in going to the second line
and again invoked a mean value theorem to get to the third line where 9 is a
number bewtween 0 and 1. So, as required,
¿(ax) = 1 % ) . (8.1.10)
a
c) We evaluated S(x — a) in part a) so now we simply compute
This equals the result from part a) and therefore we have verified that
/ ( x ) á ( x — a) = f(a)6(x — a) . (8.1.12)
d) In this case we first write out the two functions whose limit we need to take.
he(x~y) = { h x-e<y<x + e
\ 0 otherwise
< z +< 13
*•(»-') = {t — • <«>
110 CHAPTER 8. DISTRIBUTIONS AND FOURIER TRANSFORMS
Therefore,
OO
/
f(x)hei (x - y)he2 (y - z) dx dy
•OO
1 1 /*z+e2 r%+e i
= lim f lim - — - — / dx fix) / dy
ei-*o+ 2->o+ 2ei 2e2 j z _ £ 2 z+e2 jx-ti
1 1 f
= lim lim - — - — 2 e i / dx fix)
E1-+0+ £2-V0+ 2tl 2f2 J2 —c2
= f(z) . (8.1.14)
r
Therefore, we have again verified that
Therefore,
/ OO
/ ( x ) / i f ( x 2 — a 2 ) / ( x ) dx
•OO
1 r r-y/a2-e /•V<i2+£
lim — / / ( x ) dx+ fix) d
e-yo+ 2e j-/^l
1 r /~(o-«/2|o|) /-(a+e/2|a|;
= x dx
liW, ^7 / /( ) + f(x) d x
£->0+ |_./_(o+£/2|a|) J(a-e/2|a|)
| [ /-f/2|a| /-i/2|a|
lim — / fix + a)dx + / / ( x - a) dx
e-f0+ L-/_e/2|al J- e /2|al
Solution
a) In this case we again consider
1 yOO pOO
1= lim ^ - / dx f ( x ) eikx-ek dk . (8.2.19)
c-to ¿ir j_00 J-oo
First, we evaluate the integral over k (by completing the square in the argument
of the exponential and changing variables).
f
-Í7
ikx — ek
dk e x p l - - (8.2.20)
J — C
í= (8.2.21)
»mo¿/„v^exp("í;) n
')jx•
Now, as e —> 0 the exponential in the integrand is non-zero only near x = 0 i.e.
within a distance 9c where 9 is some finite number. So we can use a mean value
theorem to write
(8.2.22)
b) For this part we proceed exactly as in part a) but we change the integration
variable from x to y = ax. The desired result then follows immediately from
part a).
c) This result is also obvious from the calculation in part a).
d) In this case we first consider the expression
W
2
I = lim lim _ dy dk dq e ik(x-y)-e1k Jq(x-y)-e2q2 (8.2.23)
¿i—^0+ é2—•o-j- (2
and integrate over k and q to get
w
(z - y)2 (y - z) 21
I = lim lim , dy\ exp .(8.2.24)
d->0+e 2 -+0+ (2 f 1^2 4ci 4í 2
If we now substitute this result back into (8.2.24) and integrate over y we get
I = —
{x
exp (8.2.26)
2tt y e i + £2 4(ei + e 2 )
^Jiauuiiujvj ai\U PUUHIER TRANSFORMS
In the limit as e —> 0 + the integrant vanishes everywhere except in the vicinity
of x2 = a2 which is the same as in the vicinity of x = ±a. So we can use a mean
value theorem and write
(x — a) 2 4a 2
I = lim 7 7 - ^ / (xa + 0e) / \/*f(x)dxexp
>0+ 2?r 1/ ' ' J V £"V ' 4e
J fa [f{a+0e)+f{ a+e (x + a ) 2 4 a 2 1
= ^+h - ^\f^]/
+ f(-a + 0c) x) dx exp
= ¿ ] [ ^ a ) + ^(-a)] • (8-2.28)
to yield ¿-function normalization is l/y/ir. You will have to use the continuity
of the wavefunction at x = ± a as well as the equation
Solution
We begin with the given solution.
{ cos(kx - S+(k))
A(k) cos(A'x)
c o s ( k x + S+(k))
x < a
—a<x<a
x > a
where throughout we have k > 0 and, from continuity of the wavefunction, that
(8.3.29)
cos ka
A(k) = ( +S+(k))
cos (Ka) (8.3.30)
8.3. NORMALIZATION OF SCATTERING SOLUTION 113
as well as
with
+ ^ (8.3.32)
or
[~a
I = / dx cos[^x — <?+(&)] cos[ga; — i+(g)]
J—oo
r
+ >l(fc)^4(g) / dx cos(Kx) cos(Qx)
J—a
+ poo
/ dx cos[fca: + Ó+ (k)] cos[gx + <5+ (g)]
Ja
poo
= 2 dx cos[&£ + i+(&)] cosf^rx + ¿+(9)]
Jo
+ 1A(k)A(q) / dx cos(Kx) c o s ( Q x )
J—a
I = v6{k-q). (8.3.35)
C = ~ (8.3.36)
o. I / U j. iiiDu IJL/JVJ /IIVL* R (JUKIKH THANSt (JRMS
k=oJ~°° ax
ii*(*)i<c fc (i+m
for some Ck and j depending on k. As a matter of fact every tempered distri-
bution can be written in this form. Symbolically one then writes
* dkFk(x)
r = £(-i)
dxk
k=0
This formula cannot be taken literally however since the Fk (x) need not be
differentiable. It arises from a formal integration by parts of the first equation
above.
Solution
Let / G S. Then,
dkf
^ £ 5 for all k . (8.4.37)
Each term on the right can be bounded by one of the semi-norms || / || r n and
hence, T ( f ) is a continuous functional on S.
Solution
Consider
(8.5.40)
/ OO
ip
e- *f(x)dx
-oo
Since the support of / , s u p p / , is bounded, the integral on the right converges
uniformly for all complex values of p. We may therefore differentiate under the
the integral sign with respect to the complex variable p to get
(8.5.41)
F = fT
Solution
We have t h a t a tempered distribution of fast decrease is of the form / T where
/ G S and T is a tempered distribution. Now, f ( x ) elkx is itself a test function
belonging to S and therefore
(8.6.42)
RU^N-tr
(8.6.43)
Unr i r 1 1 x
f i )
+ dx
[x — a + te x — a — ic x—a
£-+o+ 2 J_,
where the principle value integral is defined by
•f f /(«) r f(x).
x
f ( )
dx = lim J — C
x—a e^o+ Ja x —a
J —I
Furthermore show t h a t
e/ir
lim = ¿(z) .
e^+0+ x + e 2
2
Solution
To study
P
1 1
/ = lim — + f ( x ) dx (8.7.44)
e-»0 + ^x — a — i c x — a + ic
we consider the 1contour integral
I f .J-I 1 1
a
2 R \x — a — ic + x — a + te
f ( x ) dx
1 1
+ f ( z ) dz
a — ic z — a + ic
around the contour shown in figure 8.1. We could equally well have closed the
contour in the lower half plane. Next we consider the integral
-L rr R ñ ± i 1 m f /(*)
x + d i + dz
R x -a JA+e x a
jci 2 a
jcr z—a
Prmdx+r n±dz+i
R
dz (8.7.46)
where the contour isX shown in figure 8.2. To evaluate the integral around the
J-R ~ A JCR z -a JCI z - a
8.7. A USEFUL IDENTITY 117
S(X)
CR
a
. R »(*)
v
x a — te '
CR
rrs CI
R »(z)
z = a + e e,e . (8.7.47)
We then obtain
I Í
f(a + eeie
i e e'e dd = —inf(a) . (8.7.48)
Here we have already taken the limit of e —)• 0 under the integral sign since the
j cconverges
integral ^ - A jn
uniformly. This means we can conclude that
P ¡ R ñ * L J X = _ ¡ lü*+,•„/(„). (8.7.49)
J-R x - a JCr z-a
Next taking the limit e 0 + of equation (8.7.46) and comparing with (8.7.49)
we conclude that
1
ii,„ I f l — L + f(x)dx . (8.7.50)
J-R x — a c-fo+ 2 J_R \x — a it x — a + ic
After taking the lim/j-^oo we have the desired result.
o n / i r i M . o. JJ1S1 tilt!U11UNS AND FOURIER TRANSFORMS
For the next part we simply do a contour integration by closing the path of
integration with a large semicircle (at infinity) in either the upper or lower half
plane. In either case we get
dx =
lim
€—•o-f"/ oo 21
TTTTi fW
+ ' " '
~ 2 7 r i 42ie?
£-+o+ 7T
= /(°) • (8.7.51)
•oo
Now a simple partial fraction decomposition yields t h a t
e/n _ 1 1 1
2 2 (8.7.52)
x + e 2rr x — a — te x — a + ie
Combining this with our previous computation we again have the desired result.
1 roo rU
-2 wf_ c eikx dk = lim / e ' k ( x + ' e ) DK + / ) dk
c«fe(*-«e
>o+ 2tt
.J 0 J—oo
Prove that
f OO
dk = <J(x) .
/ OO
•oo eikr
Solution
The desired result follows immediately by writing out the integrals. Thus,
ik x ie
fHm+ -L Qf e ( + ) dk + J ««(-'O dkj
,. 1 / -1 1 e/n
= lim -—: I 1 lim (8.8.53)
«-»o+2ni \x + ie x — ie e - + 0 + X¿u +, 2e '
Solution
In this case we can proceed quite formally
j 8^\x)eikxdx = ( - 1 ) n J ó(x)£¿eikx dx
Show t h a t
{ 0
(—l) m m!i(a:)
if
if
n < m
n = m
(-l)mt^á(n- m )
(*) if
" > m -
Solution
Let f ( x ) be a test function. Then,
Now we consider the three cases separately and remember that / ( 0 ) is finite.
Thus, we find:
i) for n < m the righthand side vanishes. So,
zm¿(n>(x) = 0 T K m . (8.10.56)
iii) for n > m the righthand side yields (—1)" ( ^ r ¿ ^ y í / ' " _ m ' ( 0 ) . So,
*(«,*) = - 1 - |e í - É - 1| 2 (8.11.59)
1
U¿T '
that occurs in the derivation of Fermi's Golden Rule satisfies the following prop-
erties.
a)
OO
/
<t>(u),t) dui = 2n . (8.11.60)
-OO
b)
Solution
a) We begin by rewriting
^ (e*" - l ) (e-<¿U>t
1 — iujt j
2 [ ^ d x = 29 f —dx (8.11.64)
J — oo X J— oo X
Thus, to evaluate this integral we close the contour with a large semicircle C in
the upper half plane and a small semicircle Co of radius e around the origin as
shown in figure 8.3. Since there are no poles inside the contour the total contour
9(z)
Co
R 8(*)
The contribution from the large semicircle vanishes and after taking the limit
for the first two terms we see that
f°° — dx = - lim [ — dx
x
J-oo x <-*°+Jco
f° exp(ze e !
— lim I "—-iee,ed9
£->0 +
Jw eei6
r°
= — i / dO = ix . (8.11.66)
J IT
Therefore,
OO
/
</>(u,t)duj = 2tt . (8.11.67)
-OO
b) To prove that
we integrate the left hand side with a test function f(uj) before taking the limit.
Thus, we consider
oo »oo 2
/
<t>(w,t)f(u)du> = lim / ——(1 — cosuit)f(oj) du
• OO '-•"j-oo w '
f°° 2
= lim / — (1 — cos x)f(x/t) dx
'-"loo z2
f oo 2
=
/(0) I —r(l—cos x) dx
d-oo
= 2tt/(0) . (8.11.69)
Bibliography
[8.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., chapter 8, (2002) .
Chapter 9
Algebraic M e t h o d s
[A,eB] — ceB .
[A,Bn] =ncBn~1 .
Solution
We are told that
n=0
oo
= Y-^ncb"-1
„ nn!I
n=0
oo
= c y - B"-1
v 7
n=1
= ceB . (9.1.3)
9.2. EXPECTATION VALUES: SIMPLE HARMONIC OSCILLATOR 123
where UQ(X) is the ground state, U\(x) is the first excited state and u3(x) is the
third excited state of a S.H.O.:
Solution
The wavefunction at t = 0 is given by
H = ¿ + \k„' , (9.2.5)
so that
We want ( H ) , (p2/2m), (1/2 fea:2), (p), and (x) as functions of time. Therefore,
a2 V(t) = V&Cu-ie-7^1'2
aW(i) = B u \ e ~ 3 i w t / 2 _|_ 3Ci¿3 e ~ 7 i w t ¡ 2 . (9.2.11)
<jtiAfl'ER 9. ALGEBRAIC METHODS
Furthermore,
a a ' = a^a + 1 . (9.2.
So, we find
= [ s ' C V 6 e - 4 i w t / 2 + C*B\fáe4i"t/2 - 2 [ B * B + 3 C * C ) - l]
where
B*C= \B*C\ e'* .
Since
it follows that
<i*^>=<*>-<£>.
Next, we again use that
p = i —-—(at — a) as well as x = + a)
and find
(p) = —V2mfiw|j4*5| sin(w< + a )
/ ih
(i) = a / |A*B| cos(ait + a )
V mu
where
= |^*s|eia .
— (Yjm, LjY\mi)
where j = ar, y, z.
b) Show that they satisfy the cyclic commutation relation
x + iy x — iy
2
' '
Solution
a) We begin with
Therefore,
1 ü
0 0 0 (9.3.23)
0 0
Also
So we finally get
Lx - i ( l + + 1 _ ) = | 1 0 1 ) (9.3.26)
V2 0 1 0
0 -i 0
L ^ y ( I + - q = A | ¿ o -i i (9.3.27)
Thus,
H
Now,
i o 1 0 0
u
0 0 0 0 0 0 (9.3.32)
h
0 0 h2
0 0 1
and
0 0 0
l 0 1 0 (9.3.33)
h2
0 0 0
e't 0
0 (9.3.34)
6 X P
( ^ 0 0 e 7
since then
xm = ^ y l i m . (9.3.37)
®i = nei7
Xg — X()
x'_! = (9.3.38)
= 5Zexp ("T") Xm
'' (9.3.39)
H = E a) a + V(a + a*)
a positive constant.
Hint: Try to transform to operators
6 = ua + v , = u* a} + v*
where u and v are complex numbers, and recall the simple harmonic oscillator.
Solution
We are given
where
[a,a*] = 02 (9.4.41)
So we try as suggested
6 = ua + v , = u*a* + v* . (9.4.42)
Then,
and we find
a = (3(b-v) , a1 = - V*) . (9.4.45)
So,
Therefore, if we now choose v real and such t h a t (32Ev = ,6V then the terms
linear in b and 6* go away and we have just a simple harmonic oscillator Hamil-
tonian.
H = P2Eb^b-^. (9.4.47)
Jb
The energy eigenvalues are therefore given by
En = (n + 1 / 2 ) 0 2 E - — . (9.4.48)
Solution
In spherical coordinates we have 2 = r cos 9. So we have to evaluate the integral
1 = J sin 9 d9 dip Y[*m (6, ip) cos 9 Yr ¡m> (0, tp) . (9.5.49)
1/2
Y¡¡m(0,ip) = '21 + 1 ( / - m ) P Pi"1 (cos0) etmip (9.5.50)
4 7T (I + m)\
we see that the ^-integration yields ó m ¡ m >. After we make the substitution
u = cos 9 the integral becomes:
J = y/(2l+l)(2l'+ l)(l-m)\(l'-m)\
4tr (I + m)\(l' + m)!
u p¡Jl{u) =
2ÜTÍ[(/'" m + 1)p[r+1 + ( /
'+ • (9 5 52)
- -
9.6. SCALAR OPERATOR 1'29
Therefore we find t h a t
Calculate [A,L2].
Hint: First compute [A,Lz\-
Solution
We have
[^37) Ly\ Lz
Therefore,
[A,L2} = 0. (9.6.55)
% — y + 2 ixy 3z
V> = A + 2 2 2
+ 5 exp(—oí\J x + y + z )
«.2
-j.2 +I y 2 I y2
+ z 2
y J x 2_s r y2j s _
Solution
The wavefunction is
x2 - y2 + 2ixy 3z -ay/x^+yi+z' 2
r¡) = A 2 2 + 2 2 2
+ 5 (9.7.56)
X2
+ y + z ' y/x + j/ + z
In terms of spherical harmonics (see for example table (9.1) of [9.1]) we find
'XO'TT I t+jr.
tp = A - ¡ t - Y2 2 + 3 y — Yi 0 + bViivYo 0 (9.7.58)
15 V o
Thus, a measurement of L2, Lz will only yield the values (2,2), (1,0), (0,0). The
corresponding probabilities are
32tt 1 _ 32
^2,2 = 15 327r/15+9(4?r/3)+25(47r) —
1712
p _ 12ir(15) _ _180
1712ir ~ 171'
1712
0,0
1007t(15) 1500
(9.7.59)
1712ir 1712
Solution
a) We have
0 1 0
1 0 1
0 1 0
9.8. PROBABILITY TO OBTAIN I, m ALONG DIFFERENT AXIS 131
n °.
= —=\ I 0 -I
i o
1 0 0
= h | o o o (9.8.61)
0 0 - 1
*=5 ^ (9.8.62)
1 0
vi = i 0 , v'o = tf-i = I 0 (9.8.63)
0
P+i = - , Po = - , = - • (9.8.64)
ip = | 0 (9.8.65)
0
/ cos 9 4= sin f 0 \
>/2
L$ = Lz cos 9 + Lx sin 9 — h 0 cos 9 (9.8.66)
72sin(
0
72 sin V — cos ( /
The eigenvalues are, of course, still úzh, 0. The eigenstates are found in the
standard way to be (up to an arbitrary phase factor)
j / ± ( 1 ± cos 9) \/2sin 9
\/2sin 9 , ip o —2 cos 9 (9.8.67)
± ( 1 cos 9) —\/2sin 9
Solution
By definition
L —r x p . (9.9.69)
This means
Lx = VPz - zpy , Ly = zpx - xpz , Lz = xpy - ypx . (9.9.70)
Then we find
x, Lx] = 0
X, Ly] = z[x,px] = ihz
x,Lz] = -y[x,Pr] = —ihy
V) Lx] = -z[y,Py} = — ihz
V,Ly] = 0
y, LZ] = ~x[y,Py] = ihx
Zj Lx] = y[z,pA = ihy
Z
i Ly] = -x[z,pz] = —ihx
z,Lz] = 0. (9.9.71)
This means that r and L satisfy the same sort of cyclic commutation relations
as L and L. Similarly we find t h a t p and L satisfy
[Px, Ly] = ihpz etc. in cyclic fashion . (9.9.72)
Solution
By straightforward computation we have
Similarly,
and
l x 1py 71pz
I
k
L z Ipy I
~Wx
I
(9.10.79)
134 CHAPTER 9. ALGEBRAIC METHODS
and
fi ,
L x rp z = yV'y
L y 1 fi z = . (9-10.80)
Therefore,
=
—LyXpz .L2rl>y
I 4
= -[-2^]ipx
I I
2h2x¡>x , (9.10.81)
, fi, h ^ r ,
LJrpy
(
= -7 J¿ 1 ^ 2 +
" iI I
2 h2 Ipy
and (9.10.82)
9 fir / fir/
Z/ ipz — T LxYy ~7 Lylpx
1 1
2h 2 rp z . (9.10.83)
9.11 E x p e c t a t i o n Value of Lx
A system is in a state of angular momentum given by
$ = a y j i + fcy^o + c y i , - !
where
|a|2+|fe|2+|c|2= 1 .
may be useful.
b) Compute the expectation value of L2.
Lx9 = h$t ?
Lx = 2^+ + L-) .
9.11. EXPECTATION VAL UE OF Lx
Solution
The wavefunction is
i¡j = a v i j + 6 ^ , 0 + cYi,-i
where
|a|2+|6|2+|c|2=l
a) we use
and
l_v> = \ í 2 h a y \ f i + \ í 2 h b y \ ^ i .
So,
Therefore,
Lx tp = h ip .
= h(aYiti + 6 y l i 0 + cYx-i) .
Therefore,
= + y i +
^ r 7 5 ' ° •
136 CHAPTER 9. ALGEBRAIC METHODS
URHU^ = H
and using
, it „
U = exp{— n • L)
n
[H, Ñ • L] = 0 .
[H,L] = 0 .
Solution
We are given
U r ( 0 ) = exp (9.12.97)
1
UR(9) = exp -9N • L (9.12.98)
For 9 e we get
UR(0) w 1 + r f ñ - L (9.12.99)
n
Uñl(0)t* 1-^eñ L .
h (9.12.100)
Therefore, the equation
UFÍHUÑ1 = H
(9.12.101)
becomes to this approximation
1 + —EÑ • L I H ( 1 — - e ñ • L
J \ N J " ' (9.12.102)
2
or to order e
[ñ-L,H] = 0 . (9.12.103)
[L,H} = 0 . (9.12.104)
9.13. UNCERTAINTY RELATION FOR SHO 137
(Ax) 2 = <x2>
(Ap) 2 = <p2)
(Ax)(Ap) = (n + 1 / 2 ) h .
Solution
We have
X = ( a t + a )
\ / S j (9.13.105)
p=iJ^-(a^ - a) . (9.13.106)
Thus,
(Ax) 2 = (x 2 ) = (n\a^ 2
+ a 2 + a*a + aa* |ri)
2 raw
= (nla^a + aa^ln)
2 mui
( 2 + 1
= db " >' (9.13.110)
Similarly,
(Ap)" = (p 2 ) = 2
+ a 2 — a^a — aa^|n.)
TTlhuJ
= — — (2n + l) . (9.13.111)
138 CHAPTER 9. ALGEBRAIC METHODS
Therefore,
Ax Ap = h(n + 1/2) . (9.13.113)
[A,[A,B]] = [B,[A,B}] = 0
then
eA+B = eA eB e-l/2[A,B]
f ( x ) = exA exB
Integrate this equation and obtain the desired result. This result is a special
case of the Baker-Campbell-Hausdorff formula.
Solution
The following derivation is due to Glauber. We have
and
exA Be~xA = B -[B,A]x . (9.14.119)
Therefore,
P = (A + B + [A,B]x)f(x) (9.14.120)
ax
with / ( 0 ) = 1. Furthermore, since
[A + B,[A,B]] = 0 (9.14.121)
If we now set x = 1 and multiply from the right by e - 1 / 2 ' ' 4 " 0 ' we get the desired
result
eA+B _ eA eB e-i/2[A,B] (9.14.123)
[A, B] = \A
then
AeB = exeB A .
This formula can be used, for example, if the operator B is proportional to the
Hamiltonian / k j ( a t a + 1/2) and the operator A is proportional to either a or a*.
Solution
Given
[A, B] = XA (9.15.124)
Thus,
AeB = V l ( A + S ) " A
* J ni
nI
n
= ex+BA
= exeBA. (9.15.126)
CHAPTER 9. ALGEBRAIC METHODS
9.16 Uncertainty in Lz
A particle is in a state described by the wavefunction
Solution
The particle is in the state given by
since the state is an eigenstate of L2 with this eigenvalue. The possible values
of Lz are:
lh , (l-\)h , (l-2)h
with the corresponding probabilities:
1/6 , 4 / 6 , 1/6 .
Thus, the most likely value to be obtained is (/ — 1 )ft.
b) ( A L z ) 2 is given by
(A Lz)2 = (L2Z)-({LZ))2
= J^2 . (9.16.129)
c) In this case the state is an eigenstate of Lz and thus we are certain to get
the eigenvalue namely mh. So, A L z = 0.
Solution
a) We begin with the step-up and step-down operators
L± = Lx±iLy . (9.17.130)
So,
Therefore,
1
Ll = -(Ll + Ll + L+L- + L.L+)
L2 = -1-{L\+L2_-L+L--L-L+) . (9.17.132)
Now,
Therefore,
But,
Therefore,
So,
ft2,
(£*) = (¿y) = — [ 2 / ( / + 1) - ra(ra+1)-ra(ra - 1)]
4
ft2,
= y['('+l)-"»2] . (9.17.137)
(Lz) = mft
Therefore
ALx — a ty =
= AL — - 7 = y / l { l + 1) - m 2 (9.17.140)
v2
and
ALZ = 0 . (9.17.141)
= 1(LZ) = y m . (9.17.142)
y [ / ( / + 1) - m 2 ] > y m . (9.17.143)
Therefore the inequality is satisfied and the results are consistent. Notice that
for a state with / = 0 the right hand side of (9.17.142) vanishes so that the
product of the uncertainties
ALxALy = 0 .
!«=-<£>=<"<*»•
Suppose that
Given that the system of interest is a simple harmonic oscillator and the wave
function at t = 0 is
Solution
As a first step we need to compute the wavefunction for all times. This is given
by
Now,
F((x)) = A(xy \ 2 . (9.18.146)
Thus, we need
W = + (9.18.147)
But,
sinuit . (9.18.148)
Thus,
FUx))
u
= A^U+ ^ sin2 cut . (9.18.149)
v
" 4 2 mu '
Similarly,
+
= 8 = (<^°»
imply that
[o,6]=l (9.19.154)
Solution
If we take matrix elements of (9.19.152) and (9.19.153) in the energy repre-
sentation where H is diagonal and the energies are labelled EQ, E\, j? 2 , • •. we
get
En Em = -hu (9.19.155)
En-Em = hui . (9.19.156)
En = Em±fuj . (9.19.157)
Ei — EQ + hui
E2 — Ei -(- fiu - EQ -|- 2 hu)
En = EQ + nhui . (9.19.158)
Furthermore, we see that the only non-zero matrix elements are: an,n+1 a l l ( l
. We now impose that the diagonal Hamiltonian should have the eigen-
values that we obtained. Thus, we get
H = ñ w ( 6 n i „ _ i a n _ i ) n + 1/2)
= hui (n + • (9.19.159)
Therefore,
= 1 . (9.19.161)
Thus,
[a,b] = l (9.19.162)
and a and b satisfy the commutation relations for an annihilation and creation
operator, respectively.
p= x = i[H,x] (9.20.164)
—x=p = i[H,p] (9.20.165)
imply that
where EQ is the ground state energy. You may assume that all the energy eigen-
values are positive and non-degenerate as well as that all the matrix elements
of x in the energy representation are real.
Hint: Work with matrix elements in the energy representation.
Solution
Since we are in the energy representation, H is diagonal and has eigenvalues
Eo, Ex, E 2 , . . . where EQ is the lowest eigenvalue and the eigenvalues are ordered
by increasing size. Taking matrix elements of (9.20.164) and (9.20.165) we get
or
(En-Em) = ±1 - (9.20.170)
So,
En = Eo + n . (9.20.171)
It also follows that the only non-zero matrix elements of x and p are of the form
I n n + 1 , x n +i,n and Pn,n+1 and p„+i,„. By hermiticity we also have t h a t
Xn,n + 1 — 2-n+l,n
Pn,n +1 = Pn+l,n • (9.20.172)
Also, from the Heisenberg equations we have
2 VXn>n + l X r l + l J n + x
n}n — lXn — l , n ~i~ Pn,n+lPn + l,n Pn,n— lPn — l , n )
X x
2 (^n.n + l n,n — 1 + •''n.n + l + n,n — l)
Eo + 2 -
—
x 2
2,3 + XL,2
„2
2,3
= EQ + 1
2 2
EQ + 3 - T x + X
2,3 X3 , 4
x
= 2
— 3,4
2
Eo + 4 —- T4 , 5 x + X
3,4 => X24 , 5 = i?o + 2
(9.20.176)
These results mean that
(n\[p, x]|n)
— Pn,n + l*En + l,n %n}n + lPn + l,n 4" Pn,n — l^n — l,n ^nyn — lPn — l,n
(<x • A)(a • B) = A • É + ia • (Á x B) .
Solution
We want to show t h a t if A and B are two vector operators that commute with
the Pauli matrices 5 then the following equation holds
(a A)(a • B) = A • B + ia • (A x B) .
• (9.21.180)
b) Use this result to show that if Sx, Sy, Sz are spin operators then
e»s, v/fi e-tS,<p/h _ ^ c o s y, _ sin ¡p
Solution
a) The proof follows simply by expanding the exponentials. Thus,
ÍXG
,AG r, (¿AG)2 (iXGf
eiXG A e " = [1 + i\G + v nl' + - — +
2! 3!
„ri _ (-¿AG) 2 (-¿AG)3 - .
x A[\ H ^ 1 t " ]
2! 3!
(¿A)
= A + iX[G,A]+^-[G'A-2GAG+AG¿¡
(¿al! 3
+ [G A + 3G¿4G2 - 3 G 2 A G - AG3} + • (9.22.182)
3!
But,
[G, [G, A]] = [G, GA - AG]
G2A-2GAG+ AG2 . (9.22.183)
Similarly,
[G, [G, [G, A]]] = [G, G2A - 2GAG + AG2]
= G3A + 3GAG2-ZG2AG-AG3 . (9.22.184)
Substituting these results back into 9.22.182 we have the desired verification,
b) Since Sx, Sy, Sz are spin operators we have that
[S2 , S r ] = ih Sy
[S„Sy] = -ihSx . (9.22.185)
Using this we have
Similarly, we get
+ ( ik)3Sx +
3! ( f t ) '
5
y2 'fi4 < p +\ ? +
+ Sx 5! ~
— <7y i — — + 1- •
~~ 2! 4!
= Sy cos tp + Sx sin tp . (9.22.187)
Show that
dZ_
= [A,Z] .
da
Use this result to derive the operator identity
2 3
Solution
To prove the first part we simply differentiate Z(a), keeping in mind t h a t the
operator order must be maintained. Then,
dZ r]eaA\ Í dp~aA
de
=
^ B e~oA ++ eaA B ' d£
6
di W V~da
= AeaABe~aA - eaA Be~aAA
= [A,Z] . (9.23.188)
(9.23.189)
n
dn+lz
n+1 = [A,[A,--[A,^]]]..:i
da
dnZ
= [A,[A,...[A,B)YIJ^ (9.23.191)
dan
as desired.
[A,B~l] = -B~1[A,B}B-1 .
[A,BB~1] = 0.
Solution
Using the hint we have that
Therefore,
So, acting with B~l on the left of this equation we obtain the desired result
= ^(a2al+ala2)
J
2 = y(aj,ai-a{a2)
= ^ajai-aía2)
and that
[J, A = 0
as well as that
J2 = j ( j + l)^2 •
Solution
The transformations given are due to Schwinger [9.3]. We first verify the angular
momentum commutation relations using t h a t
Then,
ih2
[•/i,J 2 ] = — [ a í a x + a j a 2 , a^a! - a{a 2 ]
= ([aía2,a2ai] _
[a2ai,aía2])
= ( [ a í , a 2 a i ] a 2 + a}[a 2 , a|,ai])
ih2 ( t t \
= — ^-a^a2 + alaij
= ihJ3 - (9.25.198)
Similarly we find
i^ 2 / t t t t \
= — [a'2ai + aia^ + a2a[ + a[a2 J
= ihJi . (9.25.199)
And finally
h2
[«^3, </i] = — [a{ai - a l a 2 , a \ a i + a j a 2 ]
= — a \ a i + a\a2 — a!,ai + a 2 a { )
= ihJ2 . (9.25.200)
152 CHAPTER 9. ALGEBRAIC METHODS
Next we compute J 2
n2
J2 = — [(<4)2a2 + (a{)2a2 + a^aia2al + a\a2a\a i
h2
= — (a\a\ + a\a2)2 — 4a\aia\a2
h2 1
= — (a\ai + a\a2)2 + 2(a\ai + a\a2)
= JÜ + 1) • (9.25.201)
This last equation also shows that
[ J 2
J ] = 0 • (9.25.202)
9.26 M i n i m u m Uncertainty in Jx
Given the eigenstates |j, m) of J 2 , Jz find the value of m t h a t minimizes the
uncertainty in Jx.
Solution
We first write
1
Jx — 2 (J+ + J-) (9.26.203)
where
m m
J±\j, ™) = ViU + !) - ( ± !)lh m i l ) . (9.26.204)
It thus follows that
( j , m\Jx\j, m) - 0 . (9.26.205)
Also,
m=—l
Solution
a) The addition theorem for the sperical harmonics states t h a t
i
£ * Y
v.* (0 (f)' w'l =
l*m{0,ip)Yim(0',<p')
= -^ ± l p ; ( C o s 7 ) (9.27.209)
47T
m1—
=——Ll
where
cos 7 = sin 0 sin 0' cos(</? — ip') + cos 0 cos 0' . (9.27.210)
Therefore,
¿ Yrm(9,v)Ylm(9,<p) = . (9.27.213)
m= — l
But, all values of m are equally probable. This means t h a t all the states
Rni{r)Y¡m(9, <p) are equally probable. So, we can take
154 CHAPTER 9. ALGEBRAIC METHODS
so t h a t all states
1
1pn,l,m (0 — :Rni(r)Yim(9,<p) (9.27.216)
V2T+T
are equally probable. Therefore, the probability density for finding the particle
at r is the sum over all m values of these states
i ' i
hMr)|2 = 2TTT £ 2rn"|/?"'(r)l*Y¿n(0,<P)Yim(0,<p)
m— — l
|Bn,(r)[22¿+l
2/ + 1 4tt
2
i^wi (9.27.217)
47T
and is a function in terms of r only.
/ 0 1 0
Jx — 1 0 1 (9.28.218)
V 0 1 0
( 0 —i
3y - — i 0 (9.28.219)
~ V2
\ 0 i
i 0 0
0 0 0 (9.28.220)
0 0 -1
Use this result to show that the trace of this expression determines the rota-
tion angle 9 and t h a t the antisymmetric part of this expression determines the
rotation axis ñ.
Solution
Again using the matrix representations (9.28.218) - (9.28.220) we find that
( nz
V2 ^ 0 \
J •ñ nx+in y nx — inv
0 (9.28.222)
\
VT
0
nx+iny "yr
~T7o~ —n.
9.29. ALGEBRA AND CONSTANTS OF THE MOTION 155
Also, recalling t h a t
(ñ) 2 = n 2 + n 2 + n 2
= l (9.28.223)
= h¿ ~^{nx + zn
y) 1 — n? (9.28.224)
Taking the trace of the right hand side of (9.28.221) we now obtain the result
that
Thus, the trace determines the rotation angle 0. Also, taking the three different
antisymmetric parts of the right hand side of (9.28.221) we get
The three components nx,ny,nz are now easily solved from these expressions.
Thus, the antisymmetric part determines the rotation axis ñ.
H = +
\Mu>2(x2 + y2)
and
Lz = xpy - ypx
a) Show that
[L„H] = 0
156 CHAPTER 9. ALGEBRAIC METHODS
b) Show that if
then
[[A,B],H] = 0
Solution
a) Since the given Hamiltonian H is invariant under rotations about the z-axis
it follows that
[LZ,H] = 0 . (9.29.227)
This result also follows, of course, from a brute force computation which yields
- = +
k J^{Px
^f(Px-Py)2+
+ Py? + | m u 2 ( i + y)2
^Muj2(x-y)2 (9.29.229)
and Q as
\px -py,x + y] = 0
\Px+Py,x-y} = 0 . (9.29.231)
So, we get
Similarly,
2
[(Pr-Py)(Px + Py),(x-y) ] = [(Pr - P y ) , {x ~ t / ) 2 ] ( P r + Py)
and
[{x-y)(x + y ) , ( p x - p y ) 2
] = [ ( x - y ) , { p z - p y ) 2
] { x + y)
[Q,H] = 0 . (9.29.236)
Then,
[[A,B],H] = [AB,H]-[BA,H]
= A[B,H] + [A,H]B-B[A,H]-[B,H]A
= 0 (9.29.238)
since each of the four commutators vanishes. Therefore, we also have that
[R,H] = [[L2,Q],H]
= 0 . (9.29.239)
R = [LZ,Q]
2 2
= [xp y - y p x ,^(p x -p ) + ^Mu2(x2 - y2)]
Proceeding in t h e s a m e spirit we c o m p u t e
= ~ih ( ¿ ( p 2
~ py) + \mu2(X2 ~ 2/ 2 )^
= —ihQ . (9.29.241)
CHAPTER 9. ALGEBRAIC METHODS
Also,
= ihuj2(xpy - ypx)
= ihLz . (9.29.242)
This shows that the algebra closes and no further constants of the motion can
be produced in this manner.
( z | : F ( a , A*) : | W ) = F ( z * , U;)(Z|UJ) .
Solution
For the coherent states we have
Since : F(a, a*) : is, as indicated by the colon, normal order we have that if F
is written out as a series that
Hence,
9.31 N o r m a l Ordering of xn
Suppose we have units such that
« = 4(«+«») -
9.31- NORMAL ORDERING OF x" 159
'•
x
" • = /»(*)
Solution
We start with the algebra of the annihilation and creation operators
(9.31.249)
: xn : = 2 - " / 2
£í " 1 (at)"-rar . (9.31.250)
r=0 ^ '
r=0
+r(a^)n-rar_1] . (9.31.251)
But,
£ ( " ) [(atr-ra'"+1+(at)"+1-'-ar]
r=0 ^ '
= (at)"+1+ar,+1
160 CHAPTER 9. ALGEBRAIC METHODS
r (n + 1)! ^ n\
+ £ ( a t^n
) +l-rar
n + l ( n + l - r)!r! (n — r)!r!
r=l
n!(r+n+l-r) t +1 .rqr
= (at)n+1 + a n + 1
¿
(n + 1 — r)!r!
r=1
n+ 1
= E " + 1
^(at)n+1-rar (9.31.252)
r=0
Also,
£ ( ; J r(«t)-'«-'
r=0
n—1
= S> (a tV
) »-l-(r-l)ar-l
r —1
r=l
n—1
n — 1 (a Tt )^ n - l - r a r
= » l (9.31.253)
r=0
Therefore,
:x : x =: x ^ : -\-n : x : . (9.31.254)
gn{x) = 2n f n ( x )
: xn : = 2~nHn(x) . (9.31.257)
Bibliography
[9.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 9.5, (2002) .
[9.5]P. Carruthers and M.M. Nieto - Phase and Angle Variables in Quantum
Mechanics, Rev. Mod. Phys. 4 0 , 411, (1968).
C h a p t e r 10
Central Force P r o b l e m s
Solution
We are given
h + k f q ifi = x 2 +
~ 2 m \ ' y2 > f=pl+p 2
y (10.1.1)
a) Cartesian coordinates
H = HX + Hy [Hx,Hy] = 0 . (10.1.2)
Eni,n2 — (^i 4" 1 / 2 ) h u -(- (712 1/2)hw — (iii ~1" ^2 ~l" l)ft¿^ (10.1.3)
where, as always, ui2 = k / M . Also the eigenstates are (here we are setting
a2 = Muj/h)
(1014)
•
b) Cylindrical coordinates: The Schródinger equation reads
h2 ( d2 I d I d 1
+
2 m \dr* +^ + f *r'J ^ = Exp
- (10 L5)
Now let
2 = ^ r 2 ) a = 2£ a n d fi(r) _ g(j/) ( 1 q a 8)
ñ /kj
Then,
r a g 0
t t + - t - - t - - » ' + ) = - (w.1.9)
d?/2 y dy y¿ J
G ~ e±y /2
for i/—» oo . (10.1.10)
± m
G~y for y->0. (10.1.11)
Therefore we set
G=yme-y*'2F(y) (10.1.12)
and find the differential equation satisfied by F. After some simple algebra we
find
0 + ( ^ _ 2 s ) ^ + ( A _ 2 m _ 2 ) f = 0. (10.1.13)
F = . (10.1.14)
n
2(n + m + 1) — A / m i ic\
0,1+2
~ (n + 2)(2m + 2 + n) ' ( • • )
If this series does not terminate it leads to behaviour like exp(j/ 2 ) for large y and
is therefore unacceptable. Therefore the series must terminate and we require
that
2E
A = — = 2{n + m+ 1) . (10.1.16)
Solution
The Schrodinger equation, for the s-wave (/ = 0), reads
(10.2.18)
We now set
2 mE 2mVo = /32
rip = R , • —K (10.2.19)
(10.2.23)
This is Bessels equation of order K/a. The boundary conditions derive from the
boundary conditions on R. Thus,
R(0) = 0 and R(r) —> 0 for r —>• oo . (10.2.24)
This means that
H = P - + + V(fi - r 2 )
2mi 2m2
reduces to the Hamiltonian
P2 p2
H = + — + K(r)
W
2M 2m
under the transformations
w + w i ( 1 0 3 28)
mi + m2
r = fi — f , (10.3.29)
M = mx + m 2 , (10.3.30)
and
mim2
m = . (10.3.31)
mi + m2
Solution
The Hamiltonian is
?
o "iñ + m2r2
it = — , M = mi + m 2 (10.3.33)
m
17111712
r = f i — r2 , m— (10.3.34)
mi + m 2
^ ( n , ^ ) = F(xi,j/i,21;x2,y2,z2) . (10.3.35)
Then,
dF__dF_dx_ w d ^ _ m i d F OF
dxi dX dx\ dx dxx M dX dx
and
Thus we find,
2 v f + v x v f + v f (10
= © * i ¿ r ' * * • -3-38)
Similarly we find
dF_ _ 5F 5X 5F ax m? 5F _ 0F not*»
ax2 ax dx2 dx dx2 M 8X dx
and
a2F (m2\2d2F 2m2 a2F a2F /in o ^a\
ax2 _
( M ) ax2 M dxdX +
dx2 ( • • )
So,
V 2 F =
-V-F + V
'F- (10-3-41)
Therefore,
H + + v
= m ^ ^ - < , 0 3 - 43 >
— Vo for r < a
V(r) = |
0 for r > a
Find the transcendental equation which yields the energy eigenvalue for the
state with angular m o m e n t u m I.
What is the minimum degeneracy of this state?
If a proton and neutron are bound in an I = 0 state with an energy of 2.2
MeV, determine Vo given that o « 2 x 1 0 - 1 3 cm.
Solution
The radial wave equation becomes
£
, fc- = > (10.4.45)
d?u o
- 7 - 7 + A u = 0 for r < a
dr¿
Thus, we get
tan Ka = . (10.4.51)
Let
A z
T = - (10.4.52)
CHAPTER 10. CENTRAL FORCE PROBLEMS
Then we have
tan kaz = —z .
z tan kaz
3.6 -11.71
3.8 -5.58
3.9 -4.40
3.95 -3.97
2
V - —2 — — — 2
dr r dr r2 dip2 <9z
Find the energy eigenvalues and eigenfunctions for this motion.
Solution
Since the particle has to move on the surface of the cylinder we have that r = R,
the radius of the cylinder. Thus, the Schrodinger equation reads
1 d2ip d2ip ,
+ -?r~ = -k2ip . (10.5.54)
R? dip2 dz 2
Since Lz commutes with the Hamiltonian for a free particle we can write,
where F satisfies
d2F / , m2\
2
+ U 2 - —2 L P = 0 . (10.5.57)
dz ' V RJ
10.6. EXPECTATION VALUES: ELECTRON IN A H-ATOM
Li
E> (10.5.58)
- 2MR2
Therefore,
m 2
2 (10.5.59)
k2 > —
k
- R2
and the eigenfunctions are given by
2 2 2 2 2
v> = eimip A e x p ( i z \ / k — m / R ) + B e x p ( — i z y / k — m / R ? ) . (10.5.60)
J [ 4 | 1 , 0 , 0 ) + 3 | 2 , 1 , 1 ) - |2,1,0> + \ / l 0 | 2 , 1 , - 1 ) ]
Solution
We are given
(H) = (¿¡,\H\1>)
1 1
2 2
1 6 9 + 1 1 0
y2+ ¿ ¿ + ¿ -mc a
36 2
1_ *) 2
mc" a (10.6.62)
24'
b)
¿ [16 - 0 + 9 • i + 1 • o + 1 0 • ( - i ) ] h = - ~ h . (10.6.64)
170 CHAPTER 10. CENTRAL FORCE PROBLEMS
9 -> 9' = tt — 9
tp tp' = <p + n .
Solution
Under a parity transformation we have that
Therefore,
Since 0 < 9 < n and 0 < 9' < 7r we therefore get that
(10.7.67)
Therefore,
<p' = tp + x . (10.7.69)
Solution
a) The electric current in the direction e v due to an electron of charge —e and
mass M is given by
*, = — • • (10.8.70)
* 2iM r s i n # dip dip
For the state \n,l,m) the wavefunction is of the form
V'n,/,m(r,0^) = ñ n , i ( r ) P ¡ m ( c o s 0 ) e ' ^ . (10.8.71)
Therefore carrying out the required differentiation in (10.8.70) we see that
( 1 0 8 72)
•
Solution
a) The Hamiltonian is
H = ^ - + V(r) . (10.9.78)
2m
The eigenvalue equation for the I = 0 bound states now becomes
d2u ,2 n
—-2 + k u = 0 r < a
dr
= = (10.9.80)
fi-1 ft*1
and
u = rip (10.9.81)
u = A sin kr r < a
u = B e~Kr r > a . (10.9.82)
k
k cot ka — —K or tan&a = — . (10.9.83)
K
This completes part a).
b) For a solution to exist requires that the equation
2j
should have a solution. This requires that
n2 h2
- t s s ? • < 10 - 9 - 86 >
This result is also clearly the minimum condition since using the equal sign, by
taking the limit from above, yields the eigenvalue E = 0.
10.10. BINDING ENERGY AND POTENTIAL 173
r
where A is some constant and A < (3 are both positive constants.
a) What is the angular momentum of this particle?
b) What is the binding energy of this state?
c) What is the potential that produced this state?
Solution
a) Since the wavefunction has no angular dependence it follows that
L2V> = 0 (10.10.87)
so that 1 = 0.
b) To find the binding energy we look at the asymptotic behaviour of this
wavefunction for large r. This must be of the form
for r oo . (10.10.88)
Therefore,
h2a2
E = (10.10.89)
2m
c) To find the potential that produced this state we simply stick the wavefunction
and the given energy into the Schrodinger equation and solve for the potential.
Thus, we find
h2 1 d2(r<p) h2 ,a2e ar
(32 e -Pr
= —z—A
2m r dr2 2m
= (E — V)(J> = A-
2 _2
/rev -ar _ p-pr
- V A- (10.10.90)
2m
Therefore,
h2 „ e~&r
v w = - _ ( ^ - ! ) _ £ _ (10.10.91)
17 4 CHAPTER 10. CENTRAL FORCE PROBLEMS
l { )=
" ' é¡l'~''í±ftt-dt
to derive the generating function
/(*.s) = £ e x p [
r?75 (10.11.92)
n=0
for these polynomials. The contour in the integral is a circle about the origin.
b) By differentiating (10.11.92) / times and using the definition
dl
Lln.l{x) = (-\)l—[Ln{x) (10.11.93)
^Jn(x) n 1 r —sx
£ «p1tt71 • (10.11.94)
Solution
a) By changing the integration variable from t to — xt in (10.11.92) we get
/(«,.) = ¿ / e - ¿ ( . l l í ) " £ . (10.11.95)
We now take |s| < 1 to insure that for sufficiently large t the series converges.
Thus,
1
tt \ I t 1 dt
f { x s )
' ~ 2;rife 1-s(t-x)/t t
dt
= JL/e«
2iri J t — s(t — x)
et
—T-—— / : jr, r • (10.11.96)
2ni
ml 1 —
— ss jJ tt +
+ sx/(l
sx/(l —
— s)
s)
The integrand has a simple pole at
SX
t = —
1—s
so we immediately get the desired result
n
f ( x , s) = J2 ^ r ~ s = j 1 - exP["r~~] • (10.11.97)
' ni 1—s 1—s
10.12. NORMALIZATION OF HYDROGEN WAVEFUNCTION 175
< i o i L 9 8 )
n—l
( i o n i o o )
n=l
n=0
Solution
a) We first rewrite the generating formula in the form
1
£ - t i w ' " - ' - = « " f c l <1012101>
n=( +1
and then use this given generating formula twice. This gives
£i ,V (»(+i /)\!(«' +
n,n'—l+1
+2 /)! Í
y „ r 2,+2 00
p
e
~'^ln-'-llx)ln#-l(*)
i - .sí
d x
, i - s ; v 1^7
i - < /J X
io - - P [
l - ( T T
(i-»)(i-f).
poo
= (1 - s ) p _ 2 i _ 1 ( l - i ) p - 2 i - 1 ( l - st)~ip+1) / zp e~z dz
Jo
= (1 — s ) p _ 2 ' _ 1 ( l — í ) p - 2 i - 1 ( l — s i ) ~ ( p + 1 ' p ! . (10.12.102)
r rífjtí I .hMS
= £ (n-^-l)!**0""1"1 (10.12.103)
n=l+1
= t • (10.12.104)
For this to equal the left hand side for all values of s and t we clearly need
n = n' so that the left hand side reduces to
00
I i°°
E [(re -f Z)!]2 J0 e
~X x2l+1 L
n-i-1 ( * ) £ * - / - ! ( * ) d* • (10.12.105)
Hence, we find
jo e-xx 2 l + l
l2j+^1(x)lyjl_1(x)dx=-^±^s n t n . . (10.12.106)
If we now put p = 21 + 2 and n = n' we find that the right hand side yields
JiffS = (2/ + 2)!(1 — s ) ( l — i ) ( l — st)~2l+3
+ + 2 ) !
= ( i - « - . + » 1 ) e ' " ! ' , c r
n\
n=0
( l _ s _ í + s/) • (10.12.107)
n=i+1
¿(st)"'-'-1)
+ Jn
(n' + i + l ) [ ( n ' + /)!]2 '
poo
/X • (10.12.108)
Jo
Now, the right hand side may be further rewritten as
RHS =
£ {f^TTT + («")-')V-
n,n'=l-1-1
l'zzl + 1 ^ ^ '
(n + l + 1)! n 1
1 )T!s^( s t Í„
° n1'„" ' + 1
(«' + / + ! ) !
t(stn ^ • ' v '0.12.109)
1)!
lU.iú. i\RU\IVIRJTLD~ ¡XCJL,J\11y
OO / \n-l-1 r°°
/»C
y - l£í) í e -x 21+2 L2l + l ,x)L2í+X í x ) d x
^ [ ( n + lVVJo n-í-lV ; n-l-lV ;
n=/+1
oo
(n + l + 1)! (n +i /)!
7\t 1
= E (n — l — 1)! (n — / — 2)!
(sí)*n —/ — 1
n=/+1
£
n=l +1
(n + 0 ¡ o„C„j\n-/-l
2n(st)
(n — / — 1)!
(10.12.110)
2Z \ (n — I — 1)
^ = \ ^
nao/ 2n[(n + /)!] 3
(10.12.112)
1(1+1) j_ 1 1
Ki + -Ki O h~ ñn; = 0
1
- (2p + l ) ^ " ) + £ [(2/ + l ) 2 - p 2 ] a 2 ^ " 2 ) = 0 p > 1
nz 4
(r) , ( r _ 1 ) , and (r 2 ) .
Solution
a) If we write the radial equation in the form
'/(/ + 1) 2 1 1
K , = -~Ki + Rnl (10.13.113)
1Y8 CHAPTER 10. CENTRAL FORCE PROBLEMS
and multiply it by rp+2Rn¡ and integrate over all r we get (after dropping the
subscripts on Rn¡)
roo roo
/ R"rp+2Rdr = -2 / R'rp+1Rdr + /(/ + l)(rp~2)
Jo Jo
+ (i
- ¿ 5 " • °' 13114 )
If we integrate the left hand side of this equation by parts we obtain
roo r°° r0 0
/ R"rp+2Rdr=-(p + 2)/ R'rp+1Rdr- R'rp+2R'dr. (10.13.115)
Jo Jo Jo
After another integration by parts we further find
pOO r OO
/ R"rp+2Rdr = (p + 2) R[(p + l)rpR + rp+1 R'] dr
Jo Jo
r OO
+ / R[(p + 2)rp+1R'+ rp+2R"] dr . (10.13.116)
Jo
Hence, it follows that
r^ 1 1
J^ R'rp+1Rdr = - ^ - ( r p
~ 2
} . (10.13.117)
We now multiply the radial equation (10.13.113) by rp+sR' and again integrate
over all r to get
= -2 í R'rp+2R'dr-l(l + l)^l(r p
- 2
)
Jo 2
+
• <1013-118>
Therefore,
rOO
/
p+2
R r R' dr
'
Jo
= l(l+l)P±l(r»-2)--P±l( r
p
- l ) + - 2^2 P ± ^ ( r p
) .(10.13.119)
p- 1 ap —1 n a p- 1
+ 2p+2 ! l_P±A/rP\
ap —1 n2a2 p — 1
= (P+i)(r"-2) + l(l+l)(rp-2)
- - ( r " - 1 ) + ~ ¿^¿( r p ) • (10.13.120)
a na
Simplifying this result yields the desired expression
(!0.13.122)
f
2 / + 1
,'9r\ ?
e 2r/n
" — ) {Ln-i-i(2rlna)) d(2r/na)
na J
[(" + *)']
(n — I — 1)! (10.13.124)
J0
Equation (10.13.123) is just the normalization of the wavefunction and states
that
(r°)=l (10.13.125)
1
( r ~ ) =z ~ . (10.13.126)
na
Now, using these results and (10.13.121) with p = 1 we get
2 2
^¿( r ) = 3 a - i [ ( 2 / + l ) - lJ ] a ^ . (10.13.127)
n 4 n2a
After simplification this reads
2(T) = ( f - V V ) . (10.14.131)
b) Use the result of part a) to show that for a potential of the form
V(r) = Vo rn
this yields
2(T) = n(V) .
Solution
a) The proof of the first part is similar to the proof used in classical mechanics.
For a stationary state, the expectation value of r • p, the radial component of
momentum, is independent of t. Therefore, using (10.14.132) we have that
^(r p) = 0 = . (10.14.133)
But,
Px + + Vz
[f p,H] = [xpx + ypy + zpz, ^ + v(f)]
ih . 2 „ 2 / dV 8V dV
= - ( p r + p y + p 1 ) - t f t ^ _ + y _ + z--
= ih2T-ihr- VV . (10.14.134)
Therefore,
2(T) = (r • V V ) (10.14.135)
10.15. EHRENFEST THEOREM FOR ANGULAR MOMENTUM 181
b) If
V(r) = V0rn (10.14.136)
then,
f• V V ( f ) = nV(r) . (10.14.137)
2 ( T ) = n(V) . (10.14.138)
1 + v(f)
the rate of change of the expectation value of the angular momentum operator
L = r x p
is given by
¿(í> = - ( - v v > .
b) Under what conditions is this equation exactly the classical equation of mo-
tion for L ?
Solution
a) The result follows immediately if we use the equation for the time evolution
of the expectation value of an operator A.
Thus,
But,
Therefore,
b) For this result to be the same as the classical equation requires that, if
F(r) = - V V
then,
( f x F ) = (r)xF((f)) . (10.15.143)
miXi+m2X2 _
A = — , X = X i - X 2
M
where
m\m2
M = mi + m2 and m -
M
are the total and reduced mass respectively.
L = X x P + x x p
where
P = - V x , p=-Vx
I I
are the centre of mass momentum and relative momentum respectively. Give
an interpretation of this result.
Solution
Comment: This calculation is the same as the corresponding classical one if we
maintain the order of the operators.
The total angular momentum is given by
L = xi x pi + x2 x p2 . (10.16.144)
But,
,7 mo _
xi = X +—x
M
x2 = X - ^ - x (10.16.145)
M
10.17. HULTHÉN POTENTIAL: GROUND STATE 183
and
P - P1+P2
P = j f P i - ^P2 • (10.16.146)
L = Xi x pi + X2 x pi
m
i ,7 ^ ,7 - mim2 _ - m2 _
= — X x P + X x p + — — i x P l - x x p
m M M
m2 - - - toim2 _ - mi ^
+ x P - I x p ——x x P -\——x x p
M M M
= X x P + x x p . (10.16.148)
Vne~ar
V(r) = - 1 _ e _ t t r • (10.17.149)
Solution
The ground state is an s-state (I = 0) and so setting the radial wavefunction
R(r) equal to 1 / r u(r) we get the Schródinger equation
h2 d2u Vo e~ar
2
u=Eu . (10.17.151)
2m dr 1 — e~ar
If we set
2mV 2 2 m E
ffi = 2
° , K = - 2
(10.17.152)
h h
IV. ^niVTKAL FORCE PROBLEMS
K u = 0 . (10.17.153)
d ^ ~ 1 — e~ ar
[7 2 - k 2 ] e ,r
+ [ - ( 7 + q ) 2 + K2 + /3 2 ] e ~ ' 7 + a ' r = 0 . (10.17.155)
Thus,
7 2 = K2 or 7 = K (10.17.156)
and
(7 + a ) 2 = K2 + f32 . (10.17.157)
P2 a
K (10.17.158)
2a 2
Therefore,
mVo a
77 == kK ==
i ^ - 2 <1017159>
and
E = h2K2 h2 (mVn a
- ^ - =-2^{T^-2) ' I101"60»
This is an exact solution of the 1 = 0 Schródinger equation and does not have
any nodes. Therefore, it is indeed the ground state solution.
Solution
The Hamiltonian for the hydrogenic atom is
10.18. HYDROGENIC ATOM IN TWO DIMENSIONS 185
1 tmtp
ip{r,(p) = R{r) m = 0, ± 1 , ± 2 , (10.18.163)
y/2ñ
h2 d2 R 1 dR m2 Ze2
R= ER (10.18.164)
2m [ dr 2
r dr r 2
J r
d2y ( m 2 — 1/4) Z
-7-^ 2 2 / + - j / + ey = 0 . (10.18.166)
¿ ¿
ax x x
The asymptotic behaviour near x = 0 yields the solutions
y^x±(m±l/2) as £ 0 . (10.18.167)
The only acceptable (square integrable) solution is therefore the one with the
asymptotic behaviour £ l m l + 1 / 2 .
For i - + o o w e get that y satisfies
d2
y
-7—5- + ey « 0 as x oo . (10.18.168)
dxz
In this case the acceptable solution is of the form
y= x f(x) . (10.18.170)
fl
Substituting this into (10.18.166) we get the equation for /
dx2 +
2|m| + 1
— 2\/—e £ + i [Z - (2|m| + 1 ) > / = ! ] / = 0. (10.18.171)
186 CHAPTER 10. CENTRAL FORCE PROBLEMS
y = Ya>=xk (10.18.172)
k
Unless this series terminates it leads to exponential growth yielding a non square
integrable wavefunction. Therefore we need that the series terminates. This is
acomplished by setting the numerator equal to zero. The result is
^=5(wfi)TT I10'18175)
or
*= - f Í (10.18.176)
where
N = 2(\m\ + k) + l = 1, 3, 5 (10.18.177)
2 me 4 Z2e2 1
E=-—e = 10.18.178
ao (n + 1/2) 2
where
h2
do = J is the Bohr radius (10.18.179)
me
and
n = |m| + * = 1, 2, 3, . . . • (10.18.180)
axis of the orbit is a constant of the motion. This is called the Runge-Lenz
vector. In quantum mechanics the Runge-Lenz vector is given by
2
R= — (px L-L xp)-e - . (10.19.181)
2m r
Given the hydrogen Hamiltonian
-*2 2
H = ( 1 0 . 1 9 . 1 8 2 )
2m r
[R, H) = 0 .
Solution
1) It is convenient to work with individual components of the Runge-Lenz vector.
This is made even simpler by using the Levi-Civita antisymmetric tensor symbol
Now,
and
Similarly,
and
t
ñ = -í- \-(L x p) + (p x L)1 - e 2 - = R . (10.19.189)
2m L J
188 CHAPTER 10. CENTRAL FORCE PROBLEMS
"l nXi p2 e2
[Ri, H] = ijkiPjLk - LjPk) - e2y , — - -
bj.p2] = 0
[Lk,p2] = 0
[Lk,r] = 0 . (10.19.191)
Continuing, we find
6^ f 2 2 x¿
[Ri,H] = ih— {Ujk^XiXjPj - -Pi - 2ih^
2 €
2 x x
.2 txih 1
+ - P < - H ^ j P j ' + ^ j
= 0 . (10.19.192)
-* 1 —* —* V
R = -— (pxL — L x p) — e 2 - (10.20.193)
2m r
together with the hydrogen Hamiltonian
—2 e
2
H = ^ - - (10.20.194)
2m r
to prove the following steps.
1)
R • L = L • R —0 (10.20.195)
2)
— ih í i j k Lk (10.20.196)
3)
4)
2H
R2 _ e4 + (L2 + , (10.20.198)
= (10.20.199)
and
5) show that the components of K satisfy the commutation rules
as well as that
and verify that they both satisfy angular momentum commutation relations as
well as that they commute with each other. Finally,
6) use all these results to deduce the hydrogen atom energy levels
E
" = - 5 ? T - zi (10.20.203)
Ih- n
as well as the degeneracy of these levels.
Solution
1) In preparation for subsequent computations we list two properties of the
Levi-Civita symbol e¡jk as well as some commutators. Throughout we assume
the Einstein convention that repeated indices are to be summed over.
e i j k e i j m = 2S k m . (10.20.205)
Also,
2 2
\p¡, xjr] = -ihr'~3Xi . (10.20.207)
We are now ready to prove the first result. To do this we first compute
r • L = xi€ijkxjpk = 0 (10.20.208)
190 CHAPTER 10. CENTRAL FORCE PROBLEMS
as well as
L " V — ( iJ k Xj Pk J'i
— (¡j k ( Xj X1 l*k ihXjSik)
= 0 . (10.20.209)
Similarly,
L • p = CijkXjPkPi = 0 (10.20.210)
=
p •L £ijkPi%jPk
— ¿ij k ( Pi Pk ® hpk &ij )
= 0 . (10.20.211)
Therefore,
R L = (p x L — L x p) • L
2m
— 2tt2 ^ ^ ^ ^ ^
— 2 m ^*3 k ^1
— 7: €ijkLj{LiPk ih^kilPl)
Zm
ih
= —2sjiljpi = 0 (10.20.212)
etjkPjLkLi = p j L j = 0 . (10.20.213)
Similarly,
L •R — -^—£jjkL¡(pjLk LjPk)
= ~UjkLiPjLk = 0 . (10.20.214)
~ ^ {eilk^llk ~ l l p k
' 7"]
- W — , PiLm - L,pm]\ • (10.20.215)
10.20. RUNGE-LENZ VECTOR: HYDROGEN SPECTRUM 191
as well as
ibí.,2-] =
[2 + . ,10.20.217)
. f P2 2e2 |
[-fíj, Rjj — ih < Cijk^^Lk ^ijkLk |
9 tj
= ih eijk
J
Lk . (10.20.218)
m
3) The next result is
1 e2
[Ri, Lj] — - Cikl\Pk Ll LkPl i Lj] [Xj , Lj\
Zm r
— 2m^ikl{tljmPkLjn ~1~ CkjmPmLl
CkjmLmPl CljmLkPm}
_ ife..
í ít
' z?mc
r
= itx
2m ^im^jk)(PkLm LkPm)
(^ij&lm &im&jl) (pm LmPl) }
_ i fa.. e2^m
r
1 — f
X
— 2 m [ Pj Li + PiLj — LiPj + ¿jP¡] ijme ^ }
= iheijmRm . (10.20.219)
4) Now we compute
•R 2 = e
¡ f c í ( P f c ¿ / - ¿fcPi) - e2 — }
192 CHAPTER 10. CENTRAL FORCE PROBLEMS
{ 2m£'k'^Pk^' ~ ^kp'^
1
{^km^ln 3kn^ml){PkLl LkPl){P m Ln LmPn)
4m 2
2
c X' X'
-—eiki[{pkLi - LkPi)^ + y (PkLi - Lkpi)]
+ e4
C r/ ^ X¿ Xi
- 7r~€iki[(PkLi - Lkpi) h — (;VkU ~ LkPi)]
2m r r
+ A
e . (10.20.220)
But,
= p*L
Pm Ln L-m p„ = (Lnpm + i He
mn qPq) LmPn
— ihtmnqPqiPn
2
Lm nmrPr)
= - 2 típ
Pn LmPm Ln — 0
pnLmLmpn = (Lmpn + ihen
m qPq)L m Pn
— l m [ l m p n i h € n r n q P q s ) p n -j- Íhcnrnq(LmPq Ífl€qmrPr)Pr
= L2p2 + 2 h2p2 = P 2 L 2 + 2ft 2 p 2
LmPnPm Ln — LmPn{LnPm ~t" i^mnqPq) —0
LmPnLmPn — Lfn^L-mPn ih€nm qPq)Pn
= L2p2 = p2L2
LnPmPm Ln — L"p" ~ p L"
Also, n p m l m p n
l = 0 . (10.20.221)
1 1 Xk
£iklPk~L[Xi — £ikl{~Pk ~l~ ifo ó i h t i l m X m )
y» jiO
/I r *fe i fc2 xk%n
— £ikl\ ^%ipk ^n€ii m pkx m -f- a¿ í¿/m ^3
1 ñ2
= —L2 -f 2 i hp • r — 2 —
r r
1 r 2 o-fc- - ,
= - I +2ifir-p + 4—
r r
X' 1
tikiLkPi— = tikiLk ~{xipi — ihSu)
r r
— ~^iklLkXiPl
10.20. RUNGE-LENZ VECTOR: HYDROGEN SPECTRUM 193
--L2 =
r
%i j 1 r2
tikl—PkU - -L
r r
X' i1
€% kl — LkPl — ~~tiklxi{PlLk ihtlkmPm)
r v
- -~L2 + 2 i h r p . (10.20.222)
r
Combining these results we have
R2 = e
4
+ f ^ - 2 — ) (l2+ft2)
\mz mr J
K = J - ^ R . (10.20.224)
M=^(L + K) , N = ^ ( L - K ) . (10.20.227)
[Mi, Mj] = — {[Li, Lj] + [Li, Aj] + [A,-, Lj] + [A¿, Aj])
ih
— (Lk + Kk + a k + Lk)
= iheijkMk . (10.20.228)
Similarly,
[Ni, Nj] = i ([Li, Lj] - [Li, Kj] - [I<i, Lj] + [A,-, Kj])
ih
— ~^eijk (Lk — Kk — Kk + Lk)
= iheiikNk . (10.20.229)
CHAPTER 10. CENTRAL FORCE PROBLEMS
Also,
[Ni, Mj] = — ([¿i, Lj] + [Li, Kj] — [Kt, Lj] — [A,-, A'j])
777,6^
( 1 2 0
" = + ' ° -233'
But,
R L = L R=K L = L K = 0 . (10.20.235)
Therefore,
M2-N2 . (10.20.236)
H = -
2h2[4p(fi + 1) + 1]
(10.20.237)
2/i 2 (2/z + l ) 2 '
But,
2// + 1 = 1 , 2 , 3 , 4 , . . . . (10.20.238)
So, using more conventional notation we call 2p + l = n and find that the energy
levels for the hydrogen atom are given by the Bohr formula
The degeneracy is also easy to count since for a given value of n = 2 / i + 1 = 2v+1
we have
(2fi + l)(2u + 1) — n2
different states with the same eigenvalue. Therefore, including the spin, the
degeneracy is 2 n 2 .
It was by a technique of this sort that Pauli [10.5] first succeeded to use
matrix mechanics to find the spectrum of the hydrogen atom before Schródinger
came up with his result.
Bibliography
[10.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 10.6.1, (2002) .
[10.2] S. Pasternach, Proc. Natl. Acad. Sci. USA, 23, 91 (1937). ibid 23, 250,
(1937).
[10.4] B. Zaslow and M.E. Zander, Am. J. Phys. 35, 1118, (1967).
Transformation T h e o r y
/ OO
dx e~'pxu„(x) = in un(k)
-oo
where
w =
* " m
and Hn(x) is a Hermite polynomial. That is, compute the Fourier transform of
the Hermite functions.
Solution
Let {u n (ar)} be the set of normalized Hermite functions with k/(huj) = 1. So
that
( 1 U
"
and
OO
/
e~i'"un(x)dx . (11.1.2)
•OO
Since the Hermite functions damp very rapidly at ± o o we can integrate by parts
and use the following general results.
r OO *00 fip-ip* f00
/ e-'P*-f-dx=- f(x)— dx = ip e—"f(z)dx.( 11.1.3)
J— OO dx J-na dx J-OO
11.2. SCHRÓDINGER EQUATION IN MOMENTUM SPACE 197
Also,
oo J roo
/
e~i'"xf(x)dx = i-=- e-'r*f(x)dx. (11.1.4)
Thus, dp J-co
Vp,n
^/ñl \dp J tt1/4 J-oo
2
- h r ( ± r / í a - p v - l e ^
1 0
- u p v *1/4
= ( - i ) B «„(*>) - (11.1.5)
fi2
V2ip(r) + V(f)ip(r) = Etp(r) (11.2.6)
2m
to momentum space.
Solution
It is convenient to first rewrite the Schrodinger equation in the form
where
, u( ? , = 2Jnm . (11.2.8)
ip r
<¡>{p) = J e " "i>(r) d3r . (11.2.9)
m = J d s m ^ ' • (n.2.10)
J d3r(V2 + k2)iP(P)eif ?
= J d3rU(r}rp(f)eiP f
. (11.2.11)
To simplify the right hand side we replace ip(r) by its expression in terms of
<t>(p). Then,
J d3rU{r)xl>(r)eif f
= J d3rU{r)eip f
J d3q 4>(q)
=
(¿F / d3qHql / U{rleÍ{P'~ql ' r
=
( ¿ ) 3 / A ^ Í P - í X í ) (11.2.13)
where
»=¥-•
2m
If at t = 0 the particle is in the state |0, 0, 0) of a harmonic oscillator basis, find
(r2) as a function of t .
Solution
We are given the Hamiltonian
*=£• <il3i6>
The corresponding Heisenberg equations of motion are:
¡»f = k*] = o
p = -ihV (11.3.18)
11.4. DIRAC PICTURE FOR DISPLACED SHO 199
and
- P° 4 , - (11.3.19)
r —— t + x
m
where x is the position operator in the Schródinger picture. The wavefunction
in the Heisenberg picture is the same as the wavefunction in the Schródinger
picture at time t = 0. Thus, we have
where
k = (11.3.21)
Also,
Solution
We have the Hamiltonian in the Schrodinger picture
p2 1
and
/& =
' V 2 ^ ( 0 , + a) (11.4.30)
then
Next we compute
So we need to compute
[A, B] = XA (11.4.34)
then
AeB = ex eB A (11.4.35)
we find that
ae a g — iujt g — aa
(11.4.36)
and
x
o(t) = \j —[e~'ut a + eiwt a*]
V ¿muí
Ps
= xs cos ojt-\ sin cot . (11.4.38)
771LÜ
11.4. DIRAC PICTURE FOR DISPLACED SHO 201
An alternate way to solve this part is to use the equations of motion for
XD(t) and pD(t).
^ dXD r TT i I F I D X D
_ P D
ik =
~dT ix^Ho\= -PD or
~df ~ m
<
i t = \PD, Ho] = -ihkxo or ^-=-kxD. (11.4.39)
Therefore,
= }_d_PD = = _ J x d ( n 4 40)
dt2 m dt m
and
b) To find the state in the Dirac picture we first solve for the state in the
Schrodinger picture. The Hamiltonian is
<n 3)
«-Vsb "
Therefore the Hamiltonian may be rewritten as
where
6= a+ c , 6^ — a t _|_ c (11.4.45)
and satisfies
6|0) = 0 . (11.4.47)
This integral equation is of a type known as a Volterra equation and can always
be solved by simply iterating. Thus, we find
ip r
y¡d(<) = ^ s ( < ) - y J dtiXD(ti)yS{tl)
l
dt2XD(t2)iSf5(^2)
= + ±{^pj
+ j dtiXo(tl) J
*s(t) j\tlXD(h)
[tl rtn-1
x dt2xD(t2)...
dtnxD(tn)ips(tn) . (11.4.51)
Jo Jo
This is as far as one can go since this last expression cannot be simplified.
Solution
As in the previous problem, 11.4 above, we have
p2 1
dx p
(11.5.54)
dt m
and
dv
= [p, H] = —ihmui2x — ih/3 (11.5.55)
or
dp n
— = -mu x - p . (11.5.56)
11.6. HEISENBERG PICTURE: SHO AND CONSTANT FORCE 2U3
Also the Heisenberg state at any time t is the same as the Schródinger state at
time t — 0.
|* H (<)> = | * s ( 0 ) > = |0) • (11.5.58)
Solution
The Hamiltonian is
H = ^ - + \kx2 - Fx (11.6.59)
2m 2
The equations of motion are
dx p
dt m
(11.6.60)
k +F
t =~' •
The state in the Heisenberg picture is |0), the ground state of the simple har-
monic oscillator at t = 0 in the Schródinger picture. This state is annihilated by
the Schródinger picture annihilation operator which coincides with the Heisen-
berg picture annihilation operator at time t = 0. Also, in terms of the annihi-
lation and creation operators, the position and momentum are given by
\l (a^ + a)
V 2 mui
p = i (a _ a
) (11.6.61)
or
mujh ip
x +
muj
,t - mujh ip
(11.6.62)
moj
un/irj UK 11. TRANSFORMATION THEORY
The equation of motion for the annihilation operator, in the Heisenberg picture
is, by direct substitution
S= (u 6 63)
--
Therefore,
and
we find
<0|x|0) = 0 (11.6.67)
and
(0|p|0) = — ( 1 — c o s w i ) . (11.6.68)
id
Solution
Here, the Hamiltonian is given by
7)2
H = £ - - F z . (11.7.69)
¿m
The state is
^o(s) = . (11.7.70)
Therefore, we find
P = PO + Ft (11.7.72)
and
^ . - E l + L t (11.7.73)
dt m m
so that
2
x = x0 + ^ t + ^ t . (11.7.74)
m 2m
Since the Gaussian wavefunction is an even parity state we immediately find
that
So, we get
(4>o,prl>o) = Ft (11.7.76)
and
2
(V>o,*Vo ) = ^ - (11.7.77)
Solution
In this case we could try to first solve the time-dependent Schrodinger equation
99 h2 d29
ih
m = ~F'" • <1L8'78)
We would do this by first separating out the time
so that
h2 d2ip
Fx%¡) = Exp . (11.8.80)
2m dx2
This equation leads to Bessel functions of order ± 1 / 3 . We would then have to
expand the initial state, namely
2 2
m*) = ^e-* ' (11.8.81)
\ a
.ii. i n/\ IMS FORMATION THEORY
(p) = Ft . (11.8.83)
Therefore,
,—tV) = —t (11.8.84)
m m
and hence,
2
(x) = (x)\t=0 + ^ t . (11.8.85)
h ea
~ 2m (x^) +
with
A = — -x x B .
2
H
°=L+V(*>
and find the equations of motion for p and x in this picture. Write also the
Schrodinger equation for the wavefunction in the interaction picture.
Hint: Use the fact that the magnetic field is weak as well as, the result that you
will prove later (problem 17.8), that if B is a constant vector then
p•A + A •p = B • L
11.9. DIRAC PICTURE: CONSTANT MAGNETIC FIELD 207
Solution
In the Interaction picture the equation of motion for any dynamical variable is
given by
ih
~ 4 r - ~ I a D,HO] • (11.9.86)
at
Therefore using
Ho = ^ + V(x) (11.9.87)
2m
we get
Here we have used the fact that for any function of momentum
where
U0=e-iHot/n. (11.9.90)
where
Now for a weak magnetic field we can drop the term ( e 2 / 2 m ) A • A. Thus,
p • Á+ A • p= B •L (11.9.94)
H'd = —B • L . (11.9.95)
2m
So, the Schródinger equation now reads
J I T , e
ih!L\9
l D) = -J-B.L\*
_ D) . (11.9.96)
di' ~~ 2m
± riAivawRMAl'lUN THEORY
\z)=e~^'2eza'\0)
\z) = e[2at-2'al|0> .
Solution
We have that
as well as the result of problem 9.14 that if there are two operators A and B
such that
then
The operators za) and z*a satisfy this condition. Therefore we have that
And since
a|0) = 0 (11.10.101)
so that
It follows that
\z) = e * a t - i , a | 0 ) . (11.10.103)
Solution
We want to show that
where
°° 2"n
\*) = e_k|2/2 E 4=rln> • (11.11.105)
n = 0t\ vn!
YL 11 *
So, we have
2
(w\z) = e-(M +M2)/2 y * ™ (m|n)
(ma+ma)/a ^
El( w* z)
n=0
w z n
a S
e -(|z| +M -2««**)/2 (11.11.106)
Similarly,
2
(z|u>) = -2^')/2 . (11.11.107)
Thus,
| 2 ) = ( l - | 2 | 2 ) 1 / 4 e x p ( | a t 2 ) |0)
where
i*) = ( i - N 2 ) 1 / 4 E ' w > / ( 2 / ^ ) ! ! N
Solution
We have that
2 1 / 4 1)M
|z) = ( 1 - | z | ) y V ^ " |2n) . (11.12.109)
and
Also,
|2n)=-7=L=(at)2"|0). (11.12.112)
V(2«)!
Thus,
00
1 1
|2\ 1/4 ^ n
oo 1
= — (a»)'"|0)
n=0
= (1 - | z | 2 ) 1 / 4 e x p [ z / 2 ( a t ) 2 ] |0) (11.12.113)
as required.
where
S(r) = exp ^ ( a t 2 - a 2 ) ] .
Solution
From problem 9.15 we know that if
[A,B] = XA (11.13.114)
then
AeB=exeBA. (11.13.115)
(11.13.117)
[fl. ^ *- q 2
M = 22<jt = rat
and
Therefore,
Solution
We assume that the creation operator has a normalizable eigenstate |z) so
that
Next, we expand |z) in terms of the harmonic oscillator eigenkets |n) so that
c
l*> = E » - (11.14.122)
Therefore,
y/ñ c \frú
c
n — n —1 =
Co . (11.14.124)
Z z"
Hence,
(11.14.125)
212 CHAPTER 11. TRANSFORMATION THEORY
and
2
(z\z) = \ c o \ Y^- (11.14.126)
n
This sum clearly diverges for all values of z. Thus, there are no normalizable
eigenstates of the creation operator a f .
For the state |j, m) with j = 1 evaluate the spin coherent states corresponding
to
a) m = 1 .
b) m = 0 .
c) In both cases evaluate the overlap (6', ip'\9, ip) .
Hint: Use the result obtained in problem 9.3 that for any component Jk k =
x, y, z of the angular momentum operator corresponding to j = 1 you have
-ieJk/n = i - ^z + l
-^ s i n 9 c ¿o s d . (11.15.129)
n ft fi
Solution
For j = 1 we have t h a t
1 0
¿1 0 0 (11.15.130)
h
0 0
and
(11.15.131)
yfi
Therefore
1 ü 0
0 0 0 (11.15.132)
n?
o o i
11.15. SPIN COHERENT STATE: EULER ANGLES 213
and
jJ22 __ i 1 0 -1
Ji- 0 2 0 (11.15.133)
ff
-1 0 1
Now, using the hint we have that
0 0
— iipjz/h _ 0 0 0 (11.15.134)
0 0 é*
and
/ ^(l+cos0) —^sin# ^(1 — cos 9) ^
-iBJy/h sin 9 cos 9 —^ sin 9 (11.15.135)
^ | ( 1 — cos9) —^sin9 | ( 1 + cos#) j
a) If m = 1 then
IU> = (11.15.136)
and
—\ e (1 + cos 9)
|0, f ) = (11.15.137)
,72sind
h ,(p
e (1 — cos 9)
b) If m = 0 then
|1,0) = (11.15.138)
and
~72e ivsin0
|0> <fi) = cos 9 (11.15.139)
S 9
72^ ™
c) For m — 1 the overlap is
= — (1 + cos 9 cos 9') cos(<p — <p') — ¿(cos 9 + cos 9') sin(y? <p')
Solution
Supppose the operators of interest are A, and B. To derive the Heisenberg
uncertainty relation we first define
where the expectation values are taken with the state i/>. Then,
We now consider
\([A,B])\* = \([A',B'])\2
= ( < m ' b v ) - w*. s ' a ' v ' )
= (¿v,5'v>)-(b'v,¿v)
= (A'rp,B'rp)-(A'ip,B'i¡;y
= 2i^(A'ip,B'ip) . (11.16.145)
Therefore,
\{[A,B])\=\2Q(A'i,B'1p)\<2\(A'4>,B't)\ • (11.16.146)
11.16. MINIMUM UNCERTAINTY SPIN COHERENT STATES 215
AAAB>l-\^,[A,B]n • (11.16.148)
For a minimum uncertainty state this must be an equality and therefore both
of the inequalities (in (11.16.146) and (11.16.147)) must also be equalities. The
first of these requires that
1 B ' i 0 1 = B'ip)| (11.16.149)
J— — Jx iJy *
as well as
J - \ j , - j )= 0 . (11.17.159)
/ i n (l + z*z') 2 j
(1l17
' (i + i . - m i í | z t h • '160)
c) Verify
2j+lthe completeness
ft w , d2zrelation
= 1
> < (11.17.161)
7T J ( 1 + Z*Z')2
where
d z = dx dy , z = x + iy . (11.17.162)
Hint: To evaluate the integral it may be useful to use polar coordinates. Also
the following integral may prove useful.
r xn dx n\(m — n — 2)!
/ m • \rn= < n > —I , m > n + l . (11.17.163
m v
Jo (1 + x) (m - 1)!
Solution
a) To start we evaluate the action of J" on | j , j ) by making repeated use of
(11.17.158).
= y/lMlJ.i ~ 1)
Jl\j,j) = VH2j)V2(2j-l)\jJ-2)
Jt\j,j) = v/iMv/2(2j - l ) V m ~ 2 ) \ j j - 3)
and finally
Jlj+1\jj) = 0 • (11.17.165)
Expanding the exponential in the definition of |z) we get
i|2> = + í í m ü j í
2
= \j, j) + zVWU J ~ !) + \ZJW~1)1 li. i - )
2i
+ ••• + z \j, ~j) • (11.17.166)
= E(W2r
2J
xn ( 2j
n=0
= (l + | 2 | 2 ) 2 j . (11.17.167)
Therefore,
AT= (l + | r | 2 ) _ i (11.17.168)
b) We begin with the definition of \z). Then,
+ ••• + 2 2 j | i , - j ) • (11.17.170)
Therefore,
/,' } -| e (^j + )/(a n /2) Jz'j_)/(hV2)
+ ••• + (z*z')2j
2
• h*-*r( !)
= (l + z V ) 2 i
n=0 X
•
7
(11.17.171)
218 CHAPTER 11. TRANSFORMATION THEORY
So,
z
= — ~ J (1 + | 2 |2)2j + 2 ( i j ' j ) + v j \ j j ~ !) + i" z ' 2 j |j, - í ) )
I
d2zzn{z*)m
(1 + |z|2)2¿+2 '
If, in this integral, we let
x = r cos 9 , y = r sin 0 , dx dy = reír d9 (11.17.174)
so that
z = re'e _* -»0 (11.17.175)
then,
d2zz"(z*2)2m 2
/ (l + |z|)J+
,00 j . n + m + 1 ^ y2,r
-m)9 M
2 2 2 6
Jo (l + r ) ¿ + j >
yoo r „+m + l d r
2ttS, (11.17.176)
I (l + r 2)2j+2 "
n,m
So, only the diagonal terms survive. Jo
But these are now all of the form
«±12, ')!. °o
7T \
2j
n 2n + l
(l r + r 2)2j+2
d r I j,j ~ n)(j,j ~ 1
)j;
2j
= ( 2 j + 1) | j, j — n)(j, j — n\
( l + x ) 2 j + 2
2j \ nl(2j — n)!
= (2j + l ) \ j , j - n ) ( j , j - n \
n J (2.7 + 1)!
= \j,j-n)(j,j-n\. (11.17.177)
Here we have used the integral given in the "hint". Thus, finally we have
^/|í>WÜT??F =
2j + 1
E t í J - n W . } - »
n=0
j
m
= E Ii'm——j
l
= i . (11.17.178)
11.18. USEFUL COMMUTATOR 219
[fl , a 1 (fl t ) # ] = n ( « , ) B •
Solution
We want to prove, using induction, that
p o r n = 1 we have
[ota,at]=at[«,ot] = ot . (11.18.180)
v2 1
H(t) = —- + -muj2 x2 — xF(t) — pG(t)
2m 2
where both F(t) and G(t) vanish outside the interval 0 < t < T, calculate the
probability P(n) that the system is in the state of n quanta of the Hamiltonian
H(t) (t > T) if it was originally in the ground state of the Hamiltonian
H(t) (t < 0) .
Hint: Use the Heisenberg picture.
Solution
The Hamiltonian may be rewritten to read
Solution
The Heisenberg equations of motion are formally identical with the classical
equations. Furthermore, in the Heisenberg picture the state vector does not
change with time. Thus, the Heisenberg equations are
df p
dt m
The solutions of these equations are formally the same as the classical solutions.
u
Therefore, at t = T / 2 we have t h a t
where f is the position vector and are the components of r. For a system in
the state of definite angular momentum
/ (I + l ) 2 — m2 / I2 — m 2
<+1,m + ,-1,m
cos 9 Y l m = y (2/ + l)(2/ + 3) \ j (2/ - 1)(2Z + 1) "
Solution
We begin by writing the components xt in spherical coordinates and use the
fact that
J \R{r)\2r2dr= 1 (11.21.196)
to write
sin
/ V y*m (6, <p)Ylm (0, <p) dip
J0
/>2ir
= cos v? Y*m(9, <p)Y¡m(9, <p) dip
Jo
= 0 (11.21.199)
and hence t h a t
sin 2 if = - [ 1 — cos2yp]
we see that
l
= -(r2) J(1-Z cos2 9)YrmYlr¡ dd
(11.21.201)
Q z z
2 '
(l + l)2 — m2 i2 m
Qz + <r 2 ) . (11.21.203)
(2/ + l)(2/ + 3) (21 + 1)(2Z — 1)
21(21 + 1) - 6 m 2 22
Qz (r ) . (11.21.204)
(21 — 1)(2/ + 3)
As a direct check we also see immediately from the definitions for Qxx, Qyy and
Qzz that
J2, L2 and Jz — Lz + Sz .
224 CHAPTER 11. TRANSFORMATION THEORY
Solution
In terms of the Pauli spinors we have that
J = Ll + ^ha. (11.22.206)
/ + 2 0 \
• (11.22.207)
V o -i-fc ~ I /
The eigenfunctions of Jz corresponding to the eigenvalue mh are
/ ñi(r)fi(0)e1'(m-1/2)^ \
V = . (11.22.208)
\ R2{r) F2{6) J
Since we also want these to be eigenfunctions of L 2 we require that
L 2 V = /(/ + l ) / i 2 ^ (11.22.209)
Fl{0)ei(m-1/2)V = Y l i m _ 1 / 3 (e,p)
Then
1(1 + 1) + f + \LZ
J< = h (11.22.213)
\L+ 1(1 + 1) + 1 —
where
So, writing out the pair of equations resulting from the eigenvalue equation
we get
[(/ + 1/2) 2 - j{j + 1) + m]Ri + s/(l + 1/2) 2 - m?R2 = 0
y/(l + 1/2)2 - m? R\ + [(/ + 1/2)2 _ + J) _ = o .(11.22.216)
f o r a nontrivial solution we require that the determinant of the coefficients of
Rx and 7?2 must vanish. This yields
[(/ + 1/2) 2 — j ( j + l)] 2 —TO2— (/ + 1/2) 2 + m 2 = 0 . (11.22.217)
R , in(r) { V l + m+ 1/2y/,m-i/2
*,j=i+i/a.m = ~ ^ = f - I (11-22.219)
V ¿ 1 + 1
\ y/l-m + 1/2y,,m+i/2
j = l - 1/2, Z = l , 2 , 3,
R• Ir) ( — m + 1/2 i | , m - l / 2 \
. (11.22.220)
\ — -*// + m + 1 / 2 y ¡ j m + i / 2 /
Solution
The equation of motion (Heisenberg equation) for a time-dependent operator is
f = I [ A . * ] + £ . (11.23.221)
This shows that the operator A is independent of t. We can go even further and
obtain an explicit form for A by using the solutions for x(t) and p(t).
4^ P° • 4
x = xocoswr-l sinwr
mui
p = po cosut — mwxo sinwi . (11.23.224)
Here, XQ and po are the initial conditions at t = 0 for x and p respectively. They
are nothing other than the Schródinger picture operators for x and p. Then, we
see that
. . Po . .
A — mw Xo cos uit H sin u>t cos uit
mui
— [po cosut — rriLJXa sin uit] sinwi
= mujxo . (11.23.225)
[A,H]¿ 0 .
F(í) =
l^+^mw2(<)x2 "
a) Define the time-dependent annihilation and creation operators
a ¿
mult) x
. 1
() = \ nt +1 / =r^P
V 2h \Jlmhw(t)
ar(í) = 1
p . (11.24.226)
K
V 2h yj2mhui(t)
Use the Heisenberg equation, either directly or by first finding the equations of
motion for x and p, to obtain the equations of motion for a(t) and al(t).
b) Now rewrite H(t) in terms of a(t) and (t). Next use the fact t h a t the oper-
ators a(t) and a^(t) are related to the operators a(0) and a^(0) by a Bogoliubov
transformation to show that the expectation value of H(t) for a state |n), that
is an eigenstate of H(0),
tf(0)|n) = (n + l / 2 ) / k j ( 0 ) | n ) (11.24.227)
is of the form
Solution
The Heisenberg equations are obtained in the standard manner by using
«'(<) = • (11.24.231)
= H(t) . (11.24.233)
Íh
t=
=
huj(i)a{t) +
+
• (11.24.235)
228 CHAPTER 11. TRANSFORMATION THEORY
But, using the explicit form for a(t) and denoting differentiation with respect
to t by a dot the last term on the right hand side of (11.24.235) yields
da w r rnuiit) . 1 , 0 ; +
— = —f\/——x — i—. =p] = —a' . (11.24.236)
dt 2u> V 2h y^2mhuj(t) 2u
Thus,
— = -iu{t)a + • (11.24.237)
dt
In a similar manner or by taking the dagger of (11.24.237) we find that
rl n t , (jJ
í í f - = iiü(t)al + —a . (11.24.238)
dt Zio
One can check that these equations are correct by writing out the Heisenberg
equations for x and p
1 7 ~ ~ ' IT = - m u j 2 { t ) x • (11.24.239)
dt m dt
Now inverting equations (11.24.231) and substituting a(t) and a^(t) for x and p
we again get (11.24.237) and (11.24.238).
The operators a(t) and a^(t) must be related to the operators a(0) and a f (0)
by a Bogoliubov transformation. We parametrize the Bogoliubov transforma-
tion with the three real functions a(t), f3(t), and 7 ( t ) to get:
a(0)|n) = \/ñ\n — 1)
at(0)|n) = Vn + l|n + 1) . (11.24.242)
c) The expectation values of x and p vanish since both of these operators are
linear in a(0) and at(0). Thus,
h
( n | c o s h / ? s i n h ^ ( e - ' ' ( a - 7 ) + e i ( a - ^ ) [ a ( 0 ) a t ( 0 ) + a t (0)a(0)]
2 mu¡(t)
-f (cosh 2 (3 + sinh 2 /?)[a(0)a t (0) + a t (0)a(0)]|n)
Similarly,
^ cosh/? + /? sinh/?j e _ , a a ( 0 )
we get
This simplifies to
(• UJ .
(a — a;) cosh ¡3 + 2cj
— sin(a — 7) sinh ¡3 = 0
/ • ú .
( 7 + w) sinh /? — — sin(a — 7 ) cosh ¡3=0
lu>
P-^cos(a- 7 ) = 0 . (11.24.257)
Q
(0) = /?(0) = 7(0) = 0 .
We would have to solve these three equations, subject to these initial conditions,
if we wanted the explicit time dependence.
BIBLIOGRAPHY 231
Bibliography
[11.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., chapter 7, (2002) .
Non-degenerate
Perturbation Theory
00
i ' A *
=
In - r21 ^ ^ 2/tt ¿ ,
' /=0 ^ m— — l
where
V2V»(r) =6{r- r)
by
a) expanding in spherical harmonics,
b) realizing that ip(r) is the potential for a unit charge located at r", and com-
paring the two solutions.
Solution
We begin with Poisson's equation for the potential ip due to a unit point charge
V 2 ^ = i ( r — r1) . (12.1.1)
= • (12-1.2)
4tt r — PI
12.1. EXPANSION OF l / | n - r 2 | 233
Now, we also solve this differential equation in spherical coordinates. For this
purpose we write
V>(f) = £ a l t m ( r ) Y , , m ( 0 , p ) • (12.1.3)
1,171
where
Integrating this equation with rdr about r' from r' — e to r' + e we get in the
limit as e —> 0
Ida i ¡
¿2 [ra,¡m{r)]dr=
da¡,
^Yím{?)
1
.
= • (12.1.8)
dr r'-f e
dr
so we have that
Here we have already imposed the boundary conditions that the solution has
to be finite at r = 0 and has to vanish for r —>• oo. The solution also has to
be continuous at r = r' and to satisfy the condition of discontinuity in the first
derivative. Imposing these two conditions we get
or
r/ ( 2 Í + 1 )
bl^rn — a l ) t n (12.1.12)
^riAf i an 12. J V U N - D E G E N E R A T E PERTURBATION THEORY
as well as
Thus,
Al m (12.1.14)
' ~ 21 + 1 r'V+1)Yl,m^ '
So finally we get:
(0)
<m|">(r) =
„«>) p (0) [ ( 0 )
h # ' - ^l">(r_1) - 4 2 ) (0)
(m|n)(r-2)
En Em
(0)
- ^ (m|n)(r_3) - . ( ° ) { m | n ) ( 1 )
m^n (12.2.16)
(stf* - E{°>) - E ^ j
Solution
We start with the equation (12.2.16) for the r t h order correction to a state,
namely
(°><m|n)( r ) = 1
[<°)(m|jy' - Eln\n)^ - E<2> ^ h n ) ^ 2
)
En Em
- •E'^3) (0)
(m|n)^r_3) - . . . - (°)(rn|n) ( 1 'j m^n (12.2.18)
12.3. 1/2 Ax2 PERTURBATION OF SHO 235
r
and we set r — 2. Thus, we get
(0)(m|
(°)( m |„>(2) = ^' ~ g%'n)(1) - (12.2.19)
E(n0) -
Inserting, from equation (12.2.16) the expression for ("'(rain)' 1 ), namely
m
w»>(" = "vlff'-fll-)"»
we find, after writing everything out, the desired result, namely equation (12.2.17).
Solution
a) Here
ft2 = ^ , lj 2 — — , q 2 — — . (12.3.22)
m m m
The exact energy eigenvalues are given by
b) Perturbation Theory
We have that
E{n0) = (n+l/2)hu . (12.3.25)
£(!) = A <0><n|ff'|n)<0>
= ^(0)(n|x2|n)(0)
|(°)(n|tf'|s)(°)| 2
E ]
n X
~ 'Y1 (12.3.27)
E(0) E(0)
s¿n ~ rj >
For s n we have
Ah
\W(n\H'\s)W '°)(n[a 2 + at" + 2ata + l|s)' 0 ^
2 mw
Ah
\/s(s — 1)¿„J_2 + \ / ( s + l ) ( s + 2 ) ¿ n s + 2 (12.3.28)
2mu
So,
Then,
(12.3.32)
2m 2 k 2m 2
It is now clear t h a t
Therefore, the perturbation series is bounded by the perturbation series for H"
as long as |A| < k. On the other hand, the r t h order term for the perturbation
series for H" is
ATE^ = (A/Ar)r ( 0 )
H ¿ + ^x2|n)(r-1)
From this we see t h a t we can bound the perturbation series for H' by a conver-
gent result if |A| < k. In fact,
Solution
a) The Hamiltonian is
2
H = H0 + \H'= f - + h x +-Xx4 . (12.4.37)
2m 2 4
The unperturbed ground state energy is
E^ 0 ) = l-huj . (12.4.38)
a^1' = (o)
(0|íax4|0)w
= +«>>><«>
M.M.* M.M. -L X J I t. URATE PERTURBATION THEORY
=h
=
\ 2 mu J
4
-A
4
\ 2 raw
f—V
) (°'{0|a 2 a^" + aataatlO)^ 0 ^
(12.4.39)
A2 / h | (°)<0|(qt + a) 4 |n)(°>| 2
16 \2muí^ £
n?£0
nhui
1 / \4 (\z4!¿ n 4 + 6 \ / 2 j n 2) 2
16/iu; Ix 2raw y n
' n
_ 2 i j _ / fta y
(12.4.40)
8 huí \ 2mw y
<°>Hff'|0> ( O ) = 1 (0)
<m|(at + a) 4 |0)W
1
' ° ) ( m | a t 4 + a t 2 a a t + a t a a t 2 + aa t 3 |0)^°^
4 V 2 raw
1
>/24 ¿ m ,4 6 a/2 ¿im.2 (12.4.42)
4 V 2 mw
Therefore,
|0> = |0)<°> - - ( h
4 \2mui
= |0>(O) A ft
"l6 4 ^ [l2V2 !2><°) + 2V6 |4>(°)] . (12.4.43)
Solution
To first order the calculation is the same as for Rayleigh-Schródinger perturba-
tion theory. In second order the appropriate formula is
2
a ^ w<01^|0)(°)
A
(12.5.44)
- 4 L (0) •
h
0 0 ~ en
2 (o)
\ 2 p(2) (0l(at + a)4|0)(o)
A ¿
° - 4 \2muj) ¿- Eo_E(o)
A2 v4í¿n,4 + 2 + v2sn¡ 2
£0 - ^ 0 )
4 V 2raw
A2 ^4! 3>/2
(12.5.45)
4 K2mu) E0 - (4 + l/2)huj E0-{2 + l/2)hw
1 3 . / fi
EQ — —huí + —A I + A2 E ( 0 2) (12.5.46)
2 4 V 2mu>
After substituting this in the equation above and rearranging we get the follow-
ing equation for EQ.
E o - ^ f k j - ^a
2 4 2 raw
Vi! 3v/2
= A2 + (12.5.47)
2muj Eo - (4 + l/2)hu> Eo-(2 + l/2)hv
H = H0 + A H'
where
*•=(*„' l ) * ' = ( l í )
a) solve for the exact eigenvalues and eigenfunctions.
b) Solve for both eigenvalues and eigenfunctions to second order using Rayleigh-
Schródinger perturbation theory.
240 CHAPTER 12. NON-DEGENERATE PERTURBATION THEORY
Solution
The Hamiltonian is as always
H = H0 + XH' (12.6.48)
where
I) • = t) <12-6-49>
Exact solutions
The eigenvalues are given by
E+ = ±(E1+E2)+1-(E2-E1) + ^ 2
—
X2a2
— Ey "4" .
- E2-EX
E_ = \ { E x + E 2 ) - \ { E 2 - E x ) - - £ ^
X2a?
( 1 2 6 5 3 )
= E x - ~ ^ E [ " '
* = ( 0 ) (12.6.55)
where
so that
a± = — ^ f 3 ± . (12.6.57)
hi — h±
We can now write
—i\a
(12 6 58)
e 7 - E ± ) - -
where, with an arbitrary choice of phase
1
M i l - — — — (12.6.59)
1 ±!
OT
Now, to later compare with the perturbation theory we let A —>• 0 and find t h a t
2 2
Ei — E+ —(E 2 — Ei) — A a
E2 — Ei
(12 6 60)
E i - E - ^ ^~É[ ' '
A2 a 2
A+ —y (E2 — Ei) +
2(E2-EI)
A. -> Aa . (12.6.61)
So, with an arbitrary choice of phase we have
i\a
1
^ ^ j ^ j . (12.6.62)
Perturbation Theory
The Oth order eigenvalues and eigenvectors are
Ei(0) = Ei , E2(0) = E2 (12.6.63)
and
ipi -
1
iXa
E-2 — EI
o i\a
1p2 :
+
1 / E2 — Ei
iXa
E? — Ei (12.6.67)
1
To order A we now have agreement with the exact solutions.
The energy to second order is
(2) _
E{ ' p(0) (0)
1 "^1 E. £1 — E2
(2) _
= £ (12.6.68)
J?(Q) (0)
£"2 — Ei
n¿2 2
Therefore, the two energies to order A2 are
A2 a 2
£•_ = £"! -
E2 — Ei
2 2
E+ = E2 + A a (12.6.69)
£2 — Ei
These results again coincide with those obtained by expanding the exact solu-
tions in powers of A.
Also,
where
/
Jo
x2 In x dx —
4
a 3¡2 1 — - ( c + ln4a)
Solution
Using the first hint and writing, £is suggested,
where
u¡2 = k/m (12.7.72)
and |0) is the corresponding simple harmonic oscillator ground state. Thus, we
need to evaluate
/
/ = 4 -3/2
a~ax2x2 In |z| dx = —— [1 - l / 2 ( c + ln(4a))] (12.7.74)
Jo a
where c = 0.577 216 . . . = Euler's constant, we get
€
E0 - ^hui 1+ (12.7.75)
2v/2
Solution
The Hamiltonian H is
H
= 7T- + v(x, y) = + \kxx2 + + \k2y2 + A xy . (12.8.76)
2m 2m 2 2m 2
This can be rewritten as
where
wi2 = ki/m , u22 = k2/m , a 2
-\/m. (12.8.78)
a) Exact Solution
We now rotate the axes to remove the cross term.
x — X cos 9 + Y sin 9
y = - X s i n 0 + Ycos0 (12.8.79)
Then,
" é, é
So.
= + +
lmí>'x'+12ms>'y2 <12'8-83»
where
Í22 = w 2 cos 2 9 + w 2 sin 2 9 — a2 sin 29
= I (wi + | \J{"1 - )2 + 4 a 4
a4
—^ Wi -(- — ^ ~ as Of — 0 .
w2 -
fi2 = w 2 cos 2 9 + sin 2 9 + a2 sin 20
=
2 ( w i + wi>) + 2 \ / ( w 2 - ^ i ) 2 + 4 q 4
, a4
^ ^2 — —2 2 ^ ® ^0 • (12.8.84)
u>1 — w 2
The (exact) energy eigenvalues are
E
nitm = (m + l/2)/iQi + (n 2 + 1/2)htt 2 • (12.8.85)
So, to lowest order in A, for later comparison with perturbation theory, we have
=
£«i,ií2 ini + l/2)hui\ + (n2 + l/2)hu>2
h\- (n2 4- 1/2)^i - (ni + \¡2)u2 (|9 8 8fii
+
w ™(w2-w2) ' (12 8 86)
- '
12.8. TWO-DIMENSIONAL SHO 245
b) Perturbation Theory
The Hamiltonian H0 represents two uncoupled SHO's. The energy eigenvalues
are
£ ( ° ) n 2 = ( n i + 1 / 2 ) f k j 1 + (n2 + l / 2 ) f t u 2 . (12.8.87)
(°)(ni,n2\xy\ni,n2)'0) = 0 . (12.8.89)
E (12.8.90)
£j
ml9in1,m2?£n2 n1,n2 ^mum 2
Now,
n2\xy\mu m2)(0)
(0)
= ("i,n2| 7 T — 7 = = ( a I +
a
i)(a2 + a2)|mi,m2)(0)
Zm y/u)\u)2
= [ ( 0 ) ( n 1 , n 2 | \ / ( m i + l ) ( m 2 + l ) | m i + l , m 2 + 1) (0)
(0)
+ ( n i , n 2 | \ / ( m i + l ) m 2 | m i + l , m 2 - 1) (0)
(0)
+ ( " i , n2\\/mi(m2 + l ) | m i - 1, m 2 + 1) (0)
+ ^ 0 ) (ni, n 2 | 1 / m i m 2 | m 1 - l , m 2 - l)^°'j
a = —
\E\
and for harmonically bound electrons takes the form
e2 ^ fj
a =
An2m Ei/? Jj u -
2
'
3
j
Here, f j are dimensionless constants called the "oscillator strengths". In quan-
tum mechanics these are defined by
47rm „ ,
fj = 3^2^o|mi0|
47rm Ei — EQ , _
1 1 9
= sV "»
For N uncoupled electrons one then has the Kuhn-Thomas-Reiche sum rule
j
This polarizability can be used to describe the absorption of light which
carries an electron from its ground state |0) to an excited state |n) in an atom.
If, ^no = (n|r|0) and the Hamiltonian for the bound electron is
a) Show t h a t
(En - E o ^ n o = - — (n|plO) .
m
Hint: Use the commutator [H, r] and work component by component.
b) Use the commutators
2 m e
£ ' ^ » ' l k „ 0 p + |to0|>+,2„„n _ i
n
Solution
We start with the Hamiltonian
H = f - + V(r) . (12.9.93)
2m
Then,
2h
(12.9.94)
— (n\px\k) - (n|[x,//]|fc)
m
= (Ek - En)(n\x\k) . (12.9.95)
Hence, we have
ih
- < 0 | P x | * > = (E„ - £o)<0|x|*> (12.9.96)
771
and by complex conjugation
- £o)|x f c o | 2 = £ ( £ * - £o)<0|x|fc><*|s|0>
k k
= L (0|p r |fc)(fc|x|0)
—' m
k
= —(0|p x x|0) . (12.9.98)
m
Similarly,
_ j2—(0\ \k)(k\px\0)
z
—' mx
k
= -^(0|xpr|0) . (12.9.99)
Therefore,
Solution
The unperturbed energy for the ground state is
i 2 1 0 1 0 4
• ( - )
=
(12.10.108)
12.11. POSITRONIUM 249
12.11 Positronium
Positronium is a hydrogen-like system consisting of a bound state of an electron
and a positron (positive electron). The ground state consists of a singlet and
three triplet substates. The singlet state is the most stable lying about 8.2x 1 0 - 4
eV below the triplet levels which are degenerate. Field theoretic calculations
show that this is due to a spin-spin interaction of the form
A
h 0 = - t x s i • «2 •
Solution
a) The Hamiltonian for these four levels, in the absence of a magnetic field, is
as given
tfo = - ^ s i - s 2 • (12.11.109)
then, in terms of the spin states for the electron and positron, respectively, we
have
(12.11.110)
Also, since
si • s 2 = 1 [(si + s 2 ) 2 - « i - ® ! ] (12.11.111)
/ 1 0 0 0 \
_ a 0 1 0 0
0
(12.11.112)
~ 2 0 0 1 0
\ 0 0 0 -3 /
So, we have
A = 4.1 x 1 0 - 4 eV (12.11.114)
eB
= u>
2mc
we find that in the representation already used we get that
(00 0
0
0
2
0 \
0
—¡I • B = hw (12.11.116)
0 0 0 0
1 0 2 0 0 )
The eigenvalues of the total Hamiltonian are therefore (listed in the order in
which the states are listed in (12.11.110))
A/2 , A/2 + 2hu> , A/2 , A/2 - 2hui.
Solution
The unperturbed Hamiltonian is
L2
Ho = — . (12.12.117)
The interaction Hamiltonian, if we take the axis of the rotator parallel to the
z-axis, is
The effect of this perturbation is to produce (in lowest order) a change in energy
= . (i2.i2.i2i)
lm
A/ = 1 , Am = 0
I 4tt 1
cos# = \ — Yio and YQO
3
we get
So, finally
00 n
2 - a?
~ En)\dnm\2 = ^2 .
J
m j =1
Solution
For any operator A the matrix element of the time derivative of the operator is
given by
d„m = -TH[d,H]\m)
(n\dH\m) — (n\Hd\m)
(12.13.128)
J
j
So,
ih
~ En)\dnr 2 53 (d)mn (d)nm ' (p))mnj
m j
=
J
3
<ñ (12.13.129)
Mi
as required.
12.14. ANOTHER SUM RULE 253
1ri
Ee_i -iqrk
£ ( £ m - En) <n|£e«^| m) = N ML
2m
for a system of ./V interacting particles. Here |n) represents an energy eigenket
of the Hamiltonian H .
Solution
If we first write out the double commutator explicitly we find
a, £ ^
3 k
e 'q-rk
H - £ ^ ' f jH • £ e
~''4
j j k
= H ^2 lq r
e ' '> e ig r
~ ' '" - £ eq r J H
'~~ £e~ S-'irk
j k j k
e _ i f r% H e< e_,í r
~ £ £ * + £ " £ e'V' r"J H • (12.14.130)
k j k j
Next, we take the diagonal matrix elements (in the energy representation) of
this expression and insert a complete set of intermediate states.
(n | |n>
e
= En(n\^ J2 "q'fk + £ e iq fk
~ ~ £ ^ \ n
)
j k k j
- h e |m)(m| Y . z ~ i q f k i")
m
\ j k
iqrk
IIY.'-"" • £ « -
+e<i
= ^-E jk
áhq)2
- - 2 - Y e"q'fk ¿
2M
jk
= -2#" - (12.14.132)
2m
This proves the desired result.
It is worth noting t h a t if we make the dipole approximation by putting
e'^'rj fa 1 + iq • f j (12.14.133)
Solution
The unperturbed ground state of the two bosons is
with energy
Here
«•>=(&r-»(-£•') • <i2i5i35>
Also, the wavefunction ip(xi,x2) is already properly symmetrized. The energy
shift A E due to the perturbation is given to lowest order by
AE = (tp,vintip)
V (mu)Y/2 oo
/
e_i-(xj+x»)e_a(xi_a;3)2 d x i d x 2 (i2.i5.i36)
We now change variables to •oo
d *1+2:2
R =
~~2
r = x!-x2 • (12.15.137)
a) Find the ground state energy correct to first order in Vo for the case of the
singlet spin state.
b) Find the ground state energy correct to first order in Vo for the case of the
triplet spin state.
Solution
a) In the singlet case, the spin wavefunction is antisymmetric and the space
wavefunction is therefore symmetric in the interchange of the two coordinates.
Thus, this case is identical to the case of two bosons discussed in problem 12.15.
—- i vyj y xtlüUHY
b) In the triplet case the spin wavefunction is symmetric so the space wave-
function must be antisymmetric. This forces one of the particles to be in the
first excited state. Therefore, the unperturbed spatial wavefunction for the
ground state of the two fermions is
with energy
1 3
eQ — ~ huí -(- ~ hijj - 2 huí
Here
, „ /mui\1/4 ( mu ,\
e,p r
*><*> = (s») ra )
<l>i{x) = ^fap-x<t>o(x) . (12.16.140)
The energy shift A E due to the perturbation is now given to lowest order by
a x x
AE = Í e- ( i- 2) [tj)l(xi)<f>21(x2) + <j>\(xi)(j)l(x-i)
¿ J — OO
- 2<l>o(xi)<f>i(x2)<j>i(xi)<j)o{x2)] dx\dxi . (12.16.141)
We now again change variables to
D XX+X2
r =
- y -
r = Xl -x2 . (12.16.142)
The Jacobian of this transformation is again 1. Therefore,
. „ vo / m w \ 1 / 2 2 m w f°° ,
ae exp
= t(s») —J_x' (—ir(2B + r / 2 ) ) r dRd
'
\2ivh) \2mu J 27t \ m w + 2ah J
A £ = ~^aS2
Solution
If we take the electric field pointing in the x-direction then the perturbation is
V = -eSx. (12.17.144)
The problem is now, for all practical purposes, one-dimensional. The ground
state wavefunction is
^j+ 0 Q = (afcc) -1 ^ 2 cos(7rx/2a) cos(ny/2b) cos(nx/2c) . (12.17.145)
= + • (12.17.146)
« 0 , 0 . ^V>o+o,o) = ° . (12.17.147)
where we have included only the non-vanishing matrix elements and the super-
script " refers to parity = — 1.
as well as
h2ir2 (An2 1 + 1 1 2 I 7 I 5 0 )
= • <
Also,
4 o , o - ^ o , o ^ 2( l - V ) • (12.17.152)
8 ma
Combining these results we obtain
258 CHAPTER 12. NON-DEGENERATE PERTURBATION THEORY
After equating this to the expression for the energy shift in terms of the polar-
izability we find that the polarizability is given by
4096 e 2 ma 4 n2
a =
~ ? ^ (4n — l) 5 •
2
(12.17.154)
n= 1 '
V(r) = - ~ •
r
If we assume that the charge density in the nucleus is constant then we have
instead the potential energy
( Z¿_ [ r2 _ 3
/ { 22
v ( ) = {J
r R
\ l[22R 22j r
~ r • (12.18.155)
I r > R
a) Use perturbation theory to calculate the isotope shift, that is the depen-
dence on A of the K-electron (Is state) for an atom with Z protons and atomic
number A.
b) Use this result to compute the energy splitting for the K-electron between
the heaviest lead (Z — 82) isotope A — 214 and the lightest A — 195.
Neglect the presence of the other electrons.
Solution
a) The unperturbed Hamiltonian is
p2 Ze2
— - (12-18-156)
The perturbation is
f \j±. _ 3 , «] < D
= < r w 2 + r j r
s « (12.18.157)
i 0 r> R
The unperturbed ground state energy of the K-electron is
E 0)
° = - ¿ $ (12-18158)
12.18. ATOMIC ISOTOPE EFFECT 259
(12.18.159)
Then,
(i) 1 Ze2 r _x x2 3 a
r2 dx
K 6
2 R Jo 2a2 _
2 +
x
1 Ze 2 2 f 12 „ 12 12 „
3 + q —x 4 1-3 (12.18.163)
2 ^ z a r ~
If we now make the dependence on the atomic number A explicit by writing
a = _ 7j4i/3 (12.18.164)
a
we have the desired dependence on A.
(i) _ 1 Ze2 2 .-1/3 12 ^ - 2 / 3 _ 3 + 7j4i/3
E
° ~ 2 a/Z t l72
_ e - v i " 3 ( 1 | ¿ - 2 / 3 + h a - i / 3 + 3n (12.18.165)
\7 7 >
b) If we take Z = 82 and >1 = 195 we get that a = 0.238. Substituting these
values we find that
Use the last term as a perturbation to calculate the first order correction to the
energy levels of a hydrogenic atom.
Hint: The following expectation values for a hydrogenic atom may be useful
_1 1_
(1/r) =
n2 a/Z
2
(l/r
K
) 1
= — (12.19.170)
' n3(l + 1/2) ( a / Z ) 2 '
Solution
To save labour we use the fact that
(12.19.171)
2m 2 (a/Z) n 2
i ,2n 2 >
(P)
E
n \
n ,1 2 mc 2 2m
1 1 Ze2 1 Ze2Y
2
2 mc2 2 (a/Z) n r J
1 1 Z2e4 1 Z2e4 1
(l/r) + Z2e2(l/r2)
2 mc2 2 4
.4 (a/Z) ^ ~ (a/Z)
Z2eA 1 /zn 2
1 Z\2 1 1
2 mc2 4 v7 a ,/ n^ 4 +' V a J n3(l+ 1/2)
z 22e„ 4 1
(12.19.172)
mc' \ a 8 n4 n3(2l + 1)
where R is the separation of the centres of the two hydrogen atoms and f\ =
( * 1 , 2 / 1 , 2 1 ) and r*2 = ( # 2 , 2 / 2 , 22) are respectively the coordinates of the electrons
associated with atom 1 and 2. Using perturbation theory calculate the interac-
tion energy of two widely separated hydrogen atoms. As an approximation for
the unperturbed energies use
Ei - En « Ei .
Solution
The total Hamiltonian for the two hydrogen atoms is
where
n-2 7e2
Hoi = — ¿ = 1,2 . (12.20.174)
2m r,-
The unperturbed energies for HQÍ are
e2 1
En = =• (12.20.175)
ao n¿
with the corresponding eigenstates |n¿,/¿,m¡). We are interested in the energy
shift of the ground state
ew = (^(1,2)1^(1,2))
= ^ [ ( l | « 1 | l ) ( 2 | « a | 2 ) + ( l | W | l ) ( 2 | i f t | 2 ) + (1|zi|1)(2| 2 2 |2)]
= 0 . (12.20.177)
e(2) _ y- 1(^(1,2)|V>1,/i,mi)|n2,Z2,m2)|2
ni,n«il 2Ei- Eni- Eni
= ¿-<V(1,2)|K2|V(1,2)> . (12.20.178)
Here we have used the indicated approximation to go to the second line and the
completeness relation to obtain the last line. These matrix elements are now
262 CHAPTER 12. NON-DEGENERATE PERTURBATION THEORY
easy to evaluate using the symmetry of the ground state wavefunction. In fact
the only non-zero matrix elements are
But,
4 r°°
<í|r?|¿> = "3
a
/ e~*r'a°r4dr
o Jo
= 3a20 . (12.20.180)
,,1 3 e4<Zo
£<" = • (12.20.181)
Bibliography
[12.1] R.Jackiw, Phys.Rev. 157, 1220, (1967) .
[12.2] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., chapter 12, (2002) .
C h a p t e r 13
Degenerate Perturbation
Theory
Solution
We first write out the n = 2 levels for the hydrogen atom in units such that
length is scaled by the Bohr radius.
^2,0,0 = -^L(l-r/2)e"r/ 2
1 r 2
^2,1,0 = — re / cos#
VÍÜñ
1 r/ 2
^2,1,1 — re ' sin 9 e
\/647r
1
^2, i , - i — r e r / 2 sin 9 e~ (13.1.1)
\/647r
The matrix elements of the perturbation in this degenerate subspace are
(ip2,i,m\eE r cos9\rp2,i',m') •
The only non-zero elements are between ^2,0,0 and ip2,i,o- Thus, the matrix to
be diagonalized is
/ 0 -3 0 0 \
-3 0 0 0
eSa
0 0 0 0
\ 0 0 0 0 /
264 CHAPTER 13. DEGENERATE PERTURBATION THEORY
Here we have restored the usual units so t h a t the Bohr radius a is no longer of
unit length. The eigenvalues of this matrix are +e£a, —e£a, 0, 0. Therefore
the degeneracy of two of the states is lifted in first order and the new energies
are + e£a, E^ - e£a, E2°\ and E^ where
2
E{20) = ~ • (13.1.2)
V' = Xxy
is applied, find the change in the energy of the first excited state to first non-
trivial order.
Solution
The eigenfunctions for the particle in the box are given by
2 . nnx . knu
4>n,k = - s i n sin 0 <x,y<a. (13.2.3)
a a a
The corresponding energies are:
e
h2 7T2 ,
+ t )2
" ' * = w < " - i132-4'
This shows that the ground state with n = k = 1 is the only nondegenerate state.
All other states are clearly degenerate. The first excited state is two-fold de-
generate. The two degenerate states are |1,2)(°> and |2,1)( 0 '. The perturbation
Hamiltonian in the degenerate subspace has matrix elements ^(i,j\xy\i,j)^°\
When this is written out we get
/ a 16a \
9
H ' = ( xL f ) • (13.2.5)
\ 97t2 2 /
The corresponding eigenvalues and eigenvectors are respectively
1 _16_ 1 _
(13.2.6)
2 + 9tt 2 2 9tt 2
and
and
|(±|xy|n,Ar)(°)| 2
p(2) _ v \2 (13.2.10)
±
+k¿3 (h2n2)/(2ma2)(b — n2 — k2)
Now,
. irx . 2ny
(±\xy\n,k)(0) = [ dx dy sin — sin
a
Jo a a
. 2nx . iry nnx . kivy
± sin sin — xy sin sin (13.2.11)
a a a a
But,
2 fa . nnx . mn.
—2 / xdx sin sin
a J0 a a
J_ f xdx .cos (n — m)nx cos
(n + m)nx.
a2 Jo
1
f x r a . (N — m) TVX
d, sin
= —
a bn m r~
Jo ( ~ )*"
a , . (n + rn)irx.
- d sin
(n + m)n a
1 f . (n — m)7tx
7 i— / Sln dx
[
(n - m ) 7 r a J0
a
1 . (n + m)nx ,
sin dx
(n + m)na J0
1 1
[(-i)»-™ - 1 ] [(-1) n + m 1]
(n — m)2n2 (n + m) 2 7T2
1 Anm
= [(-i)"+m - 1 ] 2 2 (13.2.12)
7T (n — m 2 ) 2
[(-l)"+2-l] [(-l)*+i-1]
(13.2.13)
(n 2 — 4) 2 (k 2 — l ) 2
Substituting this in the sum for the perturbation series we have the desired
result.
266 CHAPTER 13. DEGENERATE PERTURBATION THEORy
Ho = ftw(a|ai + a\a2)
on the second excited states and to first order on the third excited states. What
are the effects on the ground state and first excited states?
Solution
In this case we have
and
This shows that all states except the ground state are degenerate. For the
second excited state we have three degenerate states:
All have the same unperturbed energy 2hw. In this degenerate subspace, the
perturbation Hamiltonian is
/ 0 0 2A \
H' = 0 0 0 . (13.3.17)
V 2A 0 0 /
2A , 0 , —2A .
—= 0 (13.3.18)
V2 v2 \ j
l3 4. TWO-DIMENSIONAL SHO WITH OFF-DIAGONAL TERM 267
= 71 | 2 , 0 ) (o)
-|0,2)(o) (13.3.19)
ni+n3ji2 \ l~r
— Ihoj + r2A + 0 to second order in A . (13.3.20)
/ o o o \
H' o o o Vñ\ (13.3.21)
v/l2A 0 0 0
V 0 VÍ2X 0 0 /
The eigenvalues are:
Not all the degeneracies are lifted. For this reason we only compute to first
order in A. To this order the eigenvalues are
Ai = a j cos 9 + a 2 sin 9
and choosing 9 appropriately. Do this and compare with the perturbation result.
268 CHAPTER 13. DEGENERATE PERTURBATION THEORY
Solution
a) Again we have
We first solve for the eigenvalues to second order in A, then we solve the problem
exactly and compare. For n\ + n2 = 2 (second excited state) we have the three
degenerate states:
|2,0>(°) , |1,1>(°) , and |0,2><°).
The perturbation Hamiltonian in this degenerate subspace is
0 V2X 0
H' =\ a/2A 0 v/2A (13.4.24)
0 y/2X 0
This yields:
Since H' does not change the sum over n\ + n2 there is no second order contri-
bution. The third excited states are:
|3,0>(°\ |2,1>(°), |1, 2)<°>, and|0,3)<°> .
0/ s/3X 0 0 \
H' = \/3A 0 2A 0
(13.4.28)
0 2 X 0 y/3X
\ 0 0 y/3X 0 j
13.4. TWO-DIMENSIONAL SHO WITH OFF-DIAGONAL TERM 269
The eigenvalues for this matrix are: 3A , -3A -A. The corresponding
eigenvectors are
( 1/v2 \ ( 1/V5 \
v/3/2 -%/3/2
V3/2 V3/2
V 1/V5 ) V -1/V2 /
( v/3/2 \ ( \/3/2 \
l/v/2 -1/v2
(13.4.29)
-1/V2 -1/72
\ -v3/2 V V3/2 /
We could now again compute the new basis vectors and go on to second order,
but because [ i í o , # ' ] = 0 there is no second order contribution. Thus, correct
to second order the energies are
E2 = 3 / k j ± 3A , 3TKJ ± A . (13.4.30)
b) We next examine the exact solution. For this purpose we make the principal
axis transformation
A = ai cos 9 + a2 sin 9
B — — ai sin 9 4- 02 cos 9 . (13.4.31)
Ho = huj(a\ai + a\a2)
= HU(A^A + B^B) (13.4.32)
and
Find to order A2 the energy of the ground state and first excited state. Compare
your answer with the exact solution obtained in problem 12.8c.
Solution
In this case ki « k2 and all levels except the ground state are almost degenerate.
Thus, for the first two excited states we need to use degenerate perturbation
theory.
Ground state: n\ = n2 = 0
£o?o = | A ( w i + w a ) (13.5.36)
and
4!o = 0 (13.5.37)
while
f( 2)_A2 y- l(0M™i)(%K)|2
E
°'°~X ^ —h(oJimi + ui2m2) (13 5 38)
-'
V 1 1
m1+m2¿0
where we have
Thus, we get
(2) A 2 /i 2 1 1
K,o = Z—2 -r, r • 13.5.40
UI\U>2 H[UJ\-\-UI2)
2 9 ^1 — k 2
UJ2 - U J 2 = y 13.5.41
M
H
+ ^k2y2 + Xxy . (13.5.42)
Our perturbation is
H' = A x y = (aj + a ¡ ) ( a 2 + a\) . (13.5.43)
13.5. NON-DIAGONAL TWO-DIMENSIONAL SHO 271
/ I rr'l / /\ ^ ( 0 1 (13.5.44)
h\
Ei1' = ±- (13.5.45)
fa =
^(±i (13.5.46)
HX
e ± — 2 huí i (13.5.47)
2RRIY/U¡\U>2
To compare with the exact solution obtained in problem 12.8a we must use
the fact that \kx — k2\ < 2A and put uii — ui 2 . It is then easy to see that the
results agree to this order.
We can now do higher order perturbation theory using that linear combi-
nation of 10,1) and | 1 , 0 ) which diagonalizes H'. In other words, to find the
perturbation to second order of the levels | 0 , 1 ) and |1, 0) we use as a basis the
set:
|0, 0) , | ^ + ) , 1^-) , | « i , n2) where nx + n2 > 2.
Thus, we get
E{2) - V (13.5.48)
" i , n ' 2 ~ n n^l
/ , E
n'lin'2 ~ E„1}n2
1+
But,
A ft 1
(ip±\H'\ni,n2) = (0, l | ( a | + ai)(<4 + o,2)\nx,n2)
2rriy/ujiu 2
± (l,0|(a| + ai)(aí + a2)|ni,n2) (13.5.49)
Aft
(ip±\H'\n1,n2) (13.5.51)
2M^/LÜ\ÍÜ2
A< ^ gruirían i J. DEGENERATE PERTURBATION THEORY
f( 2 ) _ \2h 1
^0,1 — (13.5.52)
2
4M ÜJIU)2 UI I + U>2 2wi
and
2
£•(2) _ Ah 1
+ (13.5.53)
1uir
—
w
1,0
4M2U>I<JJ2 1 + w2
Again, if we set uj\ = w 2 , we find that the results agree to order A2 with the
exact results obtained in problem 12.8a.
Solution
For a box with sides (a, a, b) the energy eigenstates for a particle in this box are
,FOL ¿\J¿
2^2 . ¡nTTX\
, , . ,tmny\ . /r7rz\
,m r
' ~ ^ a s m sin
(~ít)sin (~r) (13.6.54)
Since 6 < < a the ground state energy is given by E[°¡ : and the first excited
states are degenerate and their energy is given by E[°] : = E^\ Thus, the
ground state perturbation is given by taking the expectation value of the per-
turbing potential which is given by
, e£
—e<p = —xy . (13.6.56)
a
The result is
£•(0)
^-(1,1,11^11,1,1)
xdx
1,1,1
= e£a
~4~
sin2
( t ) l s1 2
" ( ? ) y d y
(13.6.57)
13.7. UNUSUAL PARTICLE ON INTERVAL 273
For the first excited states we need to diagonalize the perturbation in the de-
generate subspace. Thus, we have to evaluate
—<1,2,11*011,2,1)= —(2,1,11^12,1,1)
A A
si 2
= —4 f "> (—) ¡ ' ^ ( — ) »<ly
a a2 J0 Va ) J0 \ a J
eSa
(13.6.58)
~T~
as well as
—(1,2,11^12,1,1)= —<2,1,1m1,2,1)
a a
4 [a . /nx\ . (2nx\ fa . (iry\
7ry\ . flny
eS —2 I sin — sin ] x ax sin — I s i n ydy
a Jo \ a / \ a J J0 \ a / v a
a
256
— e£a (13.6.59)
81 7T4
Next we have to diagonalize the matrix
( 2Í6
I
817T 4
8
256
V
4
4
(13.6.60)
fn(x) = ~ei" i n + 1 / 2 ) x , a
n = 0,±1,±2,... (13.7.62)
v2 a
with the corresponding energy eigenvalues
En = E _ i n + 1 ) = £ ^ ( n + 1/2)2 . (13.7.63)
^nAf iUH 13. DEGENERATE PERTURBATION THEORY
Solution
T h e eigenfunctions of the unperturbed Hamiltonian are, as in [13.1], given by
This shows that every eigenvalue is two-fold degenerate. We thus need to diag-
onalize the perturbation Hamiltonian in the degenerate subspaces. This means
we have to evaluate the matrix elements
i f n , x f n ) = ( / _ ( „ + ! ) , ¡r/_(„+!)) = 0 (13.7.66)
as well as
(/ni x / - ( n + l)) =
(f-{n + l)t x f n ) ) — 7r(2n + 1) ' (13.7.67)
Aa ( 0 1 1
, ,
' (13.7.68)
(2n + 1) 7t \ 1 0
T h e eigenvalues are
Aa
(13.7.69)
(2 n + 1) 7r
Thus, all the degeneracies are lifted and the energies corresponding to En and
E-(n+i) are
TT2h2 9 Aa
h0 = — .
21
H' - yg — Lxx
Mc
is applied. Find the splitting of the first excited states.
13.8. RIGID ROTATOR IN MAGNETIC FIELD 275
Solution
We have
The eigenfunctions of L2 are Y¡¡m with eigenvalue /(/ + l)h2. Therefore, we have
l{l+
H0Yl¡m = 2 j)h2Y,,m. (13.8.72)
9
Wc \{L+ +
= 0 (13 8J4)
-
since
For the first excited states we have to diagonalize the matrix with matrix
elements
(Y.,m'} LxYitm)
h
^ [1/2 - m(m + l)¿m',m+i + \J2 - m(m - l)<5 m ', m _ij . (13.8.76)
2
Written out as a matrix this looks like
0 1 0
~ I 1 0011 I . (13.8.77)
0 1 0
The eigenvalues are: 0 , ± 1 . Thus, the energies correct to second order are
given by
% + y
, £ , 5 l _ y, f £ . v
(13.8.78)
/ Mc ' I ' I Mc '
Incidentally, since Lx c o m m u t e s with L2 and is known to have the eigenvalues
mh with m = —I, —(I— 1), . . . , (I— 1), I we can write down the exact eigenvalues
of the total Hamiltonian. These are
H = H0 + A H'
where
and
H' = xpy .
Can this problem be solved exactly?
Solution
T h e energy levels of the second excited state are |2, 0) , | 1 , 1 ) , |0, 2). They all
have the same energy 3hu>. We now diagonalize the perturbation part of the
Hamiltonian in the degenerate subspace. To do this we use
h
(aí+ai) , Py = » y ~ ~ í r ~ ( a 2 ~~ a 2 ) • (13.9.80)
2 mw
Then,
1 0
Ah
I— i —1 0 1 (13.9.82)
0
T h e eigenvalues and corresponding eigenvectors are \h/2 , 0 , —\h/2 and
-t= 2i 0 2i (13.9.83)
V6 ' V2 ' 76
Thus, the new basis states we use for starting the perturbation computation are
l^-> = ^=(|2,0>-2í|l,l>-|0,2» .
(13.9.84)
13.10. PASCHEN-BACK EFFECT 277
We now have to compute the terms H'\ip+) , H'\ipo) , and which do not
contain terms |n, m) with n + m — 2. Then we have (after omitting the terms
with n + m = 2)
A(n, m\H'\±)
—Z —
— '~~/= 3^71,1 — ¿n,l¿m,3) ± 2í(2¿„ i 2¿m,2 ¿n,0<5m,0) (13.9.85)
and
so that
p( 2 ) _ (Aft) 2 _2 2 16 4
+ + +
2,±
16ftw . ( 3 - 5 ) (3-5 3-5 3- 1
2
4A ft (13.9.87)
3w
Similarly we find
(2) A2 ft
E.2 , 0 (13.9.88)
Yes, this problem can be solved exactly since the total Hamiltonian is quadratic
in the annihilation and creation operators.
Solution
We take the magnetic field to point in the z-direction. In that case the Hamil-
tonian for the valence electron is
+ (13.10.90)
ÍJ I (J
^tiAf iKK 13. DEGENERATE PERTURBATION THEORY
we have the weak field or Zeeman effect. In that case it is convenient to choose
states of good j,rrij and this situation is solved in general in problem 17.10.
On the other hand if
( n , ¿ S > , í ) <
f e ¥ ' —(n,l\Lz + 2Sz\n,l) (13.10.91)
we have the Paschen-Back effect. In this case we want states of good m¡ , ms.
For this we use the results of problem 17.7 where we find
|3/2,3/2> = |l,l)|l/2,l/2)
|3/2,l/2> = y | | l , 0)11/2,1/2)+ - ^ | l , l ) | l / 2 , - l / 2 )
13/2,-3/2) = |l,-l)|l/2,-l/2)
11/2,-1/2) = y | | l , - 1 ) 1 1 / 2 , 1 / 2 ) - - ^ | l , 0 ) | l / 2 , - l / 2 ) . (13.10.92)
|l,l)|l/2,l/2) = |3/2,3/2)
|l,l)|l/2,-l/2) = -^|3/2,l/2)-y||l/2,l/2)
|l,0)|l/2,-l/2) = -L|3/2,-l/2)-y||l/2,-1/2)
fc2 / ,, 1 1 dvt
2mc 2 r dr
eB
= w (13.10.95)
2 mc
10. PASCHEN-BACK EFFECT 279
the result is
|Z,m/)|l/2,m5) E
|1, l ) | l / 2 , l / 2 ) £„,i+2hu
|1,0)|l/2,l/2) e n ,i + hu
|l,l)|l/2,-l/2) 0
|l,-l)|l/2,l/2) 0
11,0)|l/2,-l/2) i-fiw
|1,-1)11/2,-1/2) e „ , i - 2 hu
|0,0)|l/2,l/2) c„,i + ftw
|0,0)|l/2,-l/2) e B ,i + /Sw . (13.10.96)
2
1 - 5 = I P - P - S ] . (13.10.97)
h2
= — . (13.10.98)
= y[l/2(l/2+l)-0-3/4] = 0 . (13.10.99)
This leaves only four states to consider. These split into two pairs: those with
rrij = 1 / 2 and those with mj = —1/2. Thus, we can consider them separately.
However, L • S is not diagonal in this representation. In fact, we find that
£ • § = - ( Ltz l
7 ^ . (13.10.100)
2 y L.j. —L
we get
tnl Í
^
0 yfi. • (13.10.101)
2 v
CHAPTER 13. DEGENERATE PERTURBATION THEORY
£n¡_ Í 0 \/2
(13.10.103)
2 v yfi- 1
with the eigenvalues
(I310105)
WV5 wf) "
T h e eigenvalues are
€
2 ~ m / 2 ± \ J h 2 u 2 + c n ihjj + 9 / 4 e2nl
If we now let
e„i/(ñw) -> 0
we recover the strong field (Paschen-Back) case. T h e energy shifts then approach
This shows that the term with the + sign corresponds to the case m¡ = 0 , ms —
1/2 and the term with the — sign corresponds to ra¡ = 1 , m, = —1/2 .
Similarly, for m j = - 1 / 2 we find the eigenvalues
- + Cni/2 ± h 2 ui 2 + Ztn\hu) + 9 / 4 f 2 ¡^ .
Again, letting
fni/(^) 0
we recover the strong field (Paschen-Back) case. We then get
Solution
We have the weak perturbation
V { f ) = e£z (13.11.108)
where we have chosen the electric field to point in the z-direction. Also, from
problem 11.22 we have that the unperturbed normalized hydrogen a t o m wave-
functions are
L M 1 2Y
R,B j - l / j ( r ) ( V + + / ',M-L/2
**n,l,j=l-\-l/2,m
VW+I
\ f l — m + 1 / 2 y,, m + i/2
(13.11.109)
m 1
Rn,j+ !/2(r) / V ' - + / 2 V i / 2
® n , I, j=l —1/2, m
^ * \ — \/l + m + 1/2 Y¡,171+1/2
(13.11.110)
^n,l,j=l+l/2,m = 1p 1
^n,l,i=l-l/2,m — V>2 • (13.11.111)
282 CHAPTER 13. DEGENERATE PERTURBATION THEORY
The energy levels in the unperturbed hydrogen atom are labelled by n , j. Under
the perturbation the orbital angular m o m e n t u m no longer commutes with the
Hamiltonian and thus, I is no longer a good quantum number. However, the z-
component of the total angular momentum J 2 , namely m, continues to be a good
quantum number. Thus, the matrix elements of the perturbation Hamiltonian
V between states with different m vanish. The diagonal matrix elements of V
also vanish
m c o s
x v u + ) ( j - m + 1) / ^j-l/2,m-l/2^J + l/2,m-l/2 # dd
- V Ü - m ) ( j + m + 1) / — l / 2 , m + l / 2 ^ J — l / 2 , m + l / 2 COS 6 dQ j .
(13.11.113)
axr ,Q s. (I + m+ 1)(/ — m+ 1)
cos = J { 2 l + 1 ) { 2 l + i)
and may be evaluated using the generating function as in problem 10.11 to yield
--ny/n2 - (j + 1/2) 2 .
BIBLIOGRAPHY 283
(13.11.117)
V = ( ¿ "¿' ) . (13.11.118)
3 " \ / n 2 — ( j + 1/2)2m
= ±Ki2 = ± t — ., , —e£ . 13.11.119
4 ju + 1
Finally we see that for a fixed value of n all terms of the fine structure, except
the term with j = n — 1/2, are split into 2 j + 1 equidistant levels corresponding
to m = — j , . . . , j- The term with j — n — 1 / 2 is not split at all since / has a
fixed value I = j — 1 / 2 = n — 1 and is not degenerate as regards the quantum
number /.
Bibliography
[13.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 6.8, (2002) .
C h a p t e r 14
Further Approximation
Methods
iP(x) = Ae~aW
Solution
T h e trial wavefunction is
Normalization yields,
A = -Y/A . (14.1.2)
Since this wavefunction has a discontinuity in its first derivative, something that
is rather unphysical, we expect that it will yield a rather poor estimate of the
energy.
T h e Hamiltonian is
+ —muix (14.1.3)
Thus,
h2a2
(T) = • (14.1.5)
2m
Similarly,
I1416)
j2
W = Sr- 4q2
+ ( 1 4 L 7
= ^ >
_0 (14.1.8)
m laA
a2 = ± mw (14.1.9)
y/2 h
Only the + sign is acceptable (since ip has to be square integrable) and it yields
a minimum for the energy expectation value given by
As expected this is larger than the exact ground state energy of 1 / 2 hut by about
41 %.
73
V-(ri,r 2 ) = — 3 e - z ( r i + r ^ ' a °
7T a0
B = f + f - 2 e * ( ± + l ) + ¿ .
2m 2m \r i r2 J ri2
286 CHAPTER 14. FURTHER APPROXIMATION METHODS
Solution
T h e Hamiltonian is
rr pi2 2e 2 p2 2e 2 e2 „ rr e2
H = h 1 — Hi + H2 H (14.2.11)
2m ri 2m r2 r 12 rj2
where
2
2e 2
= ¿=1,2 (14.2.12)
2m r,-
z r
i>(ri,?2) ~ — E ( i+r=)/a = (j)(ri)<j)(r2) (14.2.13)
7t a
where
h2
(14.2.14)
H2 1 D2
Z2 2Z
(14.2.15)
2m cr ar
The same result holds for T2<t>(r2). Next using that h2/m = e2a we find that
(Ti) = (t2)
„2 7 3 /-oo
Z2e2
= ^ " 1 l^-2ZRA]^"^R (14.2.16)
2a
Also,
2e 2 = 2e 2
= — 1e2^— 47T I
2Ze2
7t a°
Jo
f e~2Zvlardr
(14.2.17)
14.3. SQUARE BARRIER: WKB APPROXIMATION 287
Thus,
5Ze
/il) = ~ (14.2.19)
V127 8a "
9 27 (14.2.20)
(H) = - Z2 - —Z
4-m=o = 2 z - % . ("-2-21)
oz 8
Then,
z = 27 (14.2.22)
16
So, clearly Z < 2 and we see that the electrons do indeed scr?611 e a c h other.
The best estimate for the ground state or "ionization" energy is n o w
\16y a a
This is clearly lower than the perturbation result
_llf! = _2.00- . (14-224)
2 a a
Solution
a) In this case we have the potential
Í 0 x <0
V(x) = I Vo 0 < x < a . (14.3.25)
288 CHAPTER 14. FURTHER APPROXIMATION METHODS
Also, in this case we have special boundaries so we do not have to use the
connection formulae. Calling
2 mE _ 2m(V 0 - E)
k2 = K
2
= (14.3.26)
2
h ' " h2
we can write our solutions as follows for a wave incoming from the left
and use the fact that the solutions are valid right up to the turning points, we
can impose the boundary conditions that both ip and dip/dx are continuous at
x = 0 and x = a. Thus, we get
£-i,r/4 [1 + fl] = A + B
in/4
ike~ [1 - R] = K[A-B]
E-IN/4T _ AEKA + BE-KA
Ka
ike-iit/4t _ K[Ae - Be~Ka] . (14.3.29)
Í „-i>/4
[{1 + ik/K)+ R{1-ik/K)]
i v
B = le- ' 4 [ ( l - i k / K ) + R(l + ik/K)] . (14.3.30)
and substitute for A and B to get, after a little algebra, the result that
T = . (14.3.34)
(k2 — A' 2 ) sinh Ka + 2ikI\ cosh Ka
Except for an irrelevant phase factor these results agree with the exact results
obtained in [14.1].
b) T h e solutions are exactly those given by (14.3.27). However, this time we
have to use the connection formulae. To do so we first rewrite ipj and ipm as
follows.
S = e~Ka (14.3.36)
we get
^ = ^(1 + fl)
B = i(l-R) (14.3.37)
as well as
T = 2 B S ' 1 = 2 i ( l - R)S~1
Clearly, the results from part a) and part b) are very different. On the other
hand, to the approximation used, we still have conservation of probability since
p — 4Ka
\T\2 + | i l | 2 = e~2Ka + 1 - e~2Ka + — — w 1 . (14.3.40)
V(r) = -V0e~ar2
zau CHAPTER 14. FURTHER APPROXIMATION METHODS
Solution
a) The potential
looks like a simple harmonic oscillator near the bottom of the well. So a rea
rea-
sonable choice of wavefunction for the ground state seems to be
a = 2(p3l*)l,a . (14.4.43)
Therefore,
rc
f
M W,H^) = \A\'< dr(/32r2 — 3 (3)e -PR*
. 2 m J0
So,
3/3 2m V0 /? 3 / 2
M = (14.4.46)
2m . 2 h? (a + /?)3/2
K =
2mVÓ
(14.4.47)
h2a
and get
3 j A'o/? 1 / 2 ^ A'a/? 3 / 2
:0 . (14.4.48)
2 ~ (/? + a ) 3 / 2 T
(/? + a ) 5 / 2
After setting /? = a x , to make everything dimensionless, we find
(1 + x ) 5 / 2 = Kx1/2 . (14.4.49)
14.4. VARIATIONAL GROUND STATE IN GAUSSIAN POTENTIAL 291
x fa \f~K rs 30 .
To find x we therefore make a table of values for the left side and the right side.
Thus, we conclude that, to the accuracy of the data given, x — 30.4. Putting
•A (1 + z ) 5 / 2 Kx1'2
30.0 5.35 x 10 3 5.48 x 10:
30.2 5.44 x 10 3 5.50 x 10:
30.4 5.52 x 10 3 5.51 x 10:
30.5 5.57 x 10 3 5.52 x 10:
this back into the expression for the energy we find E ft¡ —15.4 eV. Thus, the
ground state energy is about E « —15.4 eV .
b) If we Taylor expand the potential we get
This is just a simple harmonic oscillator potential and the ground state energy
is given by
E K - V o + ^Hlj (14.4.51)
where
(14.4.52)
so that
E = -V0 + ~ V 0 . (14.4.53)
VK
i? ss —18.1 eV . (14.4.54)
This shows that the two calculations give similar results and agree to within
about 20% . Although the variational calculation gives a larger result it should
be considered the more reliable since the approximate potential is wider in the
region where the wavefunction is appreciable and thus gives a lower kinetic
energy.
292 CHAPTER 14. FURTHER APPROXIMATION METHODS
t / ( r ) = A r 2 ( r 2 — a2)
where
a - 2.00 x 1 0 - 1 nm
Use Rayleigh Ritz to estimate the ground state energy. A numerical answer is
required.
Solution
We have
H
= k+vl-r) (14.5.55)
where
Since we are looking for the ground state we have I = 0 and the kinetic energy
is
1
T ! ^ ^ I L\ (14.5.57)
=
~2MRDR^
1 d2 a4 A 2 2 '
HX!> = € ---T-r(ar^) H x2(x2 - \)xp (14.5.58)
x dx¿ e
with
h2
e = ^ ~¿ = 0.95 eV (14.5.59)
¿ma
and
a4 A
= 2.00 . (14.5.60)
since near the b o t t o m of the potential well we have something that looks like
a simple harmonic oscillator and we want as few nodes as possible. To proceed
we need the following integrals.
(«,,« = j f f - . (14.5.62)
Also,
,-7/ M (v'.ffv')
£(
"<6> = "fimt
2 n 2 + 6n + 3 4n2 + 16ra+15 2n + 3
= 6
+ 2 (14.5.64)
4n + 2 2b b
2n2 + q n :]
+ h3 + ^2n + 3)6 _ ^ 4n 2 + l g n + 15) _ o. (14.5.65)
4n + 2
^ 6 3 + 36 - 15 = 0 . (14.5.66)
n=l
3
y6 + 5 6 - 35 = 0 . (14.5.68)
Since this result is greater than the previous one we conclude that the ground
state enerev for this electron is E < .*?_18 pV
294 CHAPTER 14. FURTHER APPROXIMATION METHODS
v _ í rngz z >0
{| oo z =0
This corresponds to a perfectly elastic ball bouncing on a floor.
a) Find the W K B solution for all the energy levels.
b) Find the W K B solution for the ground state wavefunction of this particle.
Solution
a) We have the potential
(14.6.70)
This means that the point z = 0 is a special boundary and the wavefunction
must vanish at z = 0. The other turning point is given by
E = mgzo or zo - (14.6.71)
mg
Thus, the regions are as follows
(14.6.73)
(14.6.74)
This connects onto the solution xp¡ in the classically allowed region if ipi is given
by
(14.6.75)
(14.6.76)
14.7. GROUND STATE OF f i j 295
Therefore,
rz o
[ ° k(z') dz' - j = (n + 1/2)7T n = 0,1, 2,... (14.6.77)
4
Jo
So we find,
[¿O
m / \/zo — z dz = (n + 3/4)irh . (14.6.78)
Jo
Integrating this expression and substituting for z0 in terms of the energy En we
get
1/3
(3 n + 9)27r2_1
En [MG2H2)^ (14.6.79)
128
In particular the ground state energy, which corresponds to the classical situa-
tion of the ball at rest on the floor, is given by
1/3
81tt
E0 = (:mg'h2)^3 « 1.84(my 2 /i 2 ) 1 / ' 3 . (14.6.80)
128
hl/2
ip = 2B 0 < z < zo
[2m2g(z0 - z)]1/4
ft1/2 2m
rp = 2B 1/4
exp z > z 0 . (14.6.81)
2
[2m g(z - z 0 )]
Here,
1/3 1/3
•Eo "81tt 2 ' r h2 i
20 (14.6.82)
mg m?g
oo
ÍO
r •1
I—
TA
R
B
A B
Solution
In this case the Hamiltonian is (see figure 14.1)
p2 e2 e2 e2
H = f + - 14.7.84
2m ta rg R
where ta and rg are the distances between the electron and the protons at A
and B respectively and R is the distance between the two protons. As a trial
wavefunction we choose a superposition of a wavefunction centered at A and at
B. Thus, we use
r A a r B a
^A = - ^ = e - ' , rl>B = - L = e - ' (14.7.85)
JV = ( ^ + V > B , ^ + l M = 2 [ l + 5 ] (14.7.86)
S = ¿ (14.7.87)
where tp> is the angle of rotation about the line joining the two protons. The
volume element in these coordinates is [14.2]
d3
d3rA = -¿-(£2 - r]2)d£dr)d<p (14.7.89)
o
with the range of integration
Thus,
d v
s =
¿ ( » R)
1h
1 (R
k + (—
S ~ -R/a
(14.7.91)
a 3 \ a
The energy is given by
where
Thus,
HAA + HAB
E(R) = (14.7.95)
1+ 5
p2 e2 „2
HAA = [IPA, IPA) ~ (*PA, + ( v m , -FTLPA
2m ta
t„2
2
e-22 ( e2
= + I IpA, IpA (14.7.96)
2a R V rs
The last term is again evaluated by using elliptical coordinates and yields
2
— d3r
y;
IpA, IpA
R
B J\*. A\
RB
e2R2
t e - r v a d £ , J X e - r " / a dr¡
2a3
L
POO
-Rt/ -Rri/a
+ Tj e df]
e"
~R
e
1 — e -2 R/a i+5 a
(14.7.97)
The final integral is also evaluated using elliptical coordinates and yields
L
p.2 — v2 r2lt
di> DIP
(R/2a)(p — v)
(*•£*») = &Í J!
1
1 + - I e~R'a . (14.7.99)
R
x = — (14.7.100)
a
that
2x x
1 1 ( l + x)e — x ( l + a;)e -
E(x) = e2 2 x (14.7.101)
2 ^ x ar(l + x + x /3)e~ +x
A graph of this function shows that E(x) has a minimum at x w 2.16. This
yields a value of about 1.76 eV below the energy for x = oo, i.e. when one of
the protons is removed to infinity.
{ oo
vo
oo
x < —a
—a < x < a
x >a
Hint: Pay close attention to the boundary conditions that the trial wavefunction
has to satisfy.
Solution
Since the wavefunction must vanish at x = ± a and must be a positive parity
solution with no nodes we choose
ip = A[\a\a-\x\a] . (14.8.102)
h2
Tip = —-—A[—a{a — l)xa~2} . (14.8.104)
2m
14.9. HYDROGEN ATOM: VARIATIONAL TECHNIQUE 299
Hence we find
roo
(T) = 2\A\2J ( | a r - 1 * 1 ° ) — a r ( a - i)xa~2
h
= ' 0»"+!)("+!) . (14.8.105)
Ama2 (2a — 1)
Also
(V) = V0 . (14.8.106)
Therefore,
—(H) = 0 = ( 2 a - l ) ( 4 a + 3) - 2 ( 2 a 2 + 3 a + 1) . (14.8.108)
da
1 +
a = 2 ^ « 1.7247 . (14.8.109)
This agrees with the exact result to within better than 0.3 % since we had a
very good choice of trial function.
3 / 4
, (2 A\ *
= e
* { t )
Solution
The Hamiltonian for the hydrogen atom is
p' e'
H = (14.9.111)
2m ~ 7 '
Using the trial wavefunction we get
(14.9.112)
This expectation value breaks up into two parts, the kinetic energy T plus the
potential energy V. For the kinetic energy we get
. ,'2a\3/23 /it
2M
2M ) 8 V 2a
h2a
(14.9.113)
2M
For the potential energy we find
3/2
<k> = -e 2 47r
vfi' - e
r
r2dr
-e 2 47r
ífjf
2ar
e reír
3/2
e247r ( — ) —
7r / 4a
—2e^ (14.9.114)
Therefore,
£<„) = 3 ^ - 2 e = (14.9.115)
dE(a)
3— e' — = 0 . (14.9.116)
da 2M 7T a
Solving for a we get
8 M2e4
a= (14.9.117)
97T h4
14.10. BOUND STATE IN ONE DIMENSION 301
After substituting this into the expression for E(a) we get the best estimate for
the ground state energy
£ = . (14.9.118)
37t 2 n-
This is to be compared with the exact ground state energy
1 Me4
Eo = - ^ • (14.9.119)
It is also to be noticed that the exact energy is lower than the estimate. This
is in keeping with the variational principle.
Solution
What we need to show is that for a potential V(x) that is everywhere attractive
Í2a\1'4
rP=[-1 -ax (14.10.120)
7t J e
+ (141(U2I
" = >
Using this wavefunction the expectation value of the energy is, by the variational
principle such that
where E is the energy of the possible bound state. Evaluating E(a) we get
To obtain the best estimate for the energy we now minimize E(a) with respect
to the variational parameter a.
dE(a) h2
da
12a
2 x 2 e-2"*2 V(x)dx . (14.10.124)
302 CHAPTER 14. FURTHER APPROXIMATION METHODS
dE(a) _ Q
da
we find
[2a
E < \/ — J {\/2+2ax¿)e~¿ax V(x)dx< 0 . (14.10.126)
\/i \- Í ~u x
< 0
— | _e£x x > 0
T(E) = exp
4}
where
c 4V2m\E\3'2
S c
~ 3 ~h '
Solution
The transmission probability is given by | S | 2 where, in W K B approximation,
2
S = exp j^- J K(X') dx' (14.11.127)
14.12. DEUTERON: VARIATIONAL PRINCIPLE 303
and the integral runs over the classically forbidden region. Also,
k(X) = — E) . (14.11.128)
Therefore,
2
T(E) \ =
- |IQI
Sf =- epyn
x p {í - ^ \ (14.11.130)
as desired.
V(r) = -V0e~r'a
where
i(a) = Ae~ar'a
Solution
T h e Hamiltonian of this system in the center of mass is
H = ^ - + V(r)
w
(14.12.131)
2m
where m is the reduced mass of the proton-neutron system. T h e energy is
estimated from
_ (14.12.132)
1
> WA)M<*)) '
i r,K 14. FURTHER APPROXIMATION METHODS
T = (V'(a).^V'(a))
[°° e-°r/a i i ^ r e - a r / a \ ^ d j ,
2m Jo \r dr2 J
AW i°° e~2ar/a ( * ± - ^ dr
2m J0 \ a2 a J
/¿2|j4|2 a
(14.12.134)
2m 4a
= M<*), V(r)ip(<*))
roo
2
= -\A\ V0 / e-2arlae-r'ar2dr
Jo
= ~w2v'J^TW • ( 1 4 1 2 1 3 5 )
Therefore,
h2
= ( 1 4 1 2 1 3 6 )
•
To use the variational principle we must minimize the energy with respect to
the parameter a. So we compute a from
0 a V a2
~~T^ - = — J ~ °~7 /0\4 • (14.12.137)
da ma2 2(q+1/2)4
After inserting the appropriate numbers we get the following equation for a.
(a + 1 / 2 ) ' = 3 w = 2 80 (14.12.138)
2
a 2h
A few straightforward numerical attempts yield the value
a = 0.67. (14.12.139)
After substituting this value back into the equation for the energy we find
2
E = 8.52 a -32 f ) MeV = - 2 . 2 MeV . (14.12.140)
va+1/2 j
14.13. BOUNCING BALL: VARIATIONAL CALCULATION 305
Solution
We choose the potential energy to be zero when the ball is in contact with the
surface at say x = 0. Furthermore, since the ball cannot penetrate into the
region below the surface we must have that the wavefunction must vanish at
x = 0. Thus, we choose as a trial function
A = 2a3'2 . (14.13.142)
P
H = - 1- MGX . (14.13.143)
2m
Then we have, using this trial function, that the expectation value for the energy
is
rOO h2 d2
e + mgx x e dx
2m dx2
E(a) = = 4 a 3 H2A2x 2
{IP, HTP)hra — 2 ax
4a 3 x + mgx3 dx
M 2m
h2a2 3mg
Jo+ Lm (14.13.144)
2m 2a
The best value is obtained by minimizing E(a) with respect to the variational
parameter a. Thus, we compute a from
dE h2 a 3mg
0 (14.13.145)
m 2 a2
This yields the value
2 1/3
3m g
2 (14.13.146)
2h
Substituting this back into the expression for E we get
£
= (I) (-v»2)"3
1/3
This result should be compared with the result obtained from the W K B ap-
proximation (problem 14.6) which yielded
14.14 B e t a D e c a y of Tritium
A tritium atom (H 3 ) in its ground state /^-decays to form a singly ionized helium
a t o m (He 3 ). W h a t is the probability that this new a t o m will be found in its
ground state? Assume that both nuclei have infinite mass and that there is no
interaction between the beta-decay electron and the rest of the system.
Solution
W i t h the stated assumptions we can neglect recoil and need only compute the
overlap between the initial and final states. For the ground state of a hydrogenic
at om with nuclear charge Z we have
1 /2z\3^2
(1414149)
(*'•*>) 3
state is therefore given by
_ g3/2 í
0,
>oo
e~xx2dx
= j'- "-
JO
8 3/2
- g4 ~ 0.559 . (14.14.152)
V = Ax 4
use the variational m e t h o d to find the energies of
a) the lowest energy level and
b) the first excited state.
The corresponding energies as found by numerical integration are respectively:
2 3
T2 \ 2/3 / T2 \ /
1.060 -?-) A1/3 and 3.800 ( Í - ) A1/3
2m j 2m j
Solution
a) For the ground state we choose a trial wavefunction similar to the ground
state wavefunction for the harmonic oscillator.
ipo(x) (14.15.154)
Then,
r OO
2 2 ft2 (i2
(H) = E0(a) = -= / e-"2* / + a
4 -a2x2/2
J —CO 2^5^ *
_^_/6mA\1/3 3A f h2 \ 2 / 3
En 2
4m \ h J "^4 \6mA J
= ^ y ' v 3 . , 0 8 2 x ( i i y , a a - (14.15.158)
2m J 2m /
This is within about 2 % of the value obtained by numerical integration,
b) In this case we need a trial wavefunction orthogonal t o the wavefunction
308 CHAPTER 14. FURTHER APPROXIMATION METHODS
for the ground state. For this reason we choose a wavefunction similar to the
wavefunction for the first excited state of the simple harmonic oscillator.
Then,
2a3 t" 2 2
E
' w
= x e -a x /2 . * L ^ L +
2
Xx*
2 mdx
2a3 f Í ft / 4 4 2 2o
= —¡= / - — (q x - 3 a x ) + Ax 6 e~a x dx
V* J-oo L 2 m
ft2 3 a 2 15 A
+ 4
(14.15.160)
2 ^ ¡ ~ y a '
Again the best value for a is obtained from
DEI 15A
= 0 (14.15.161)
da 2m 2 2a5
Thus, we find
1/6
10mA / 10mA\
2 or (14.15.162)
ft Q =
Hence, our best estimate for the energy of the first excited state is
. 2/3 / t2 \ 2/3
-i/d
EX 847 x í - ) a1/3 . (14.15.163)
2m J 2m y
This is within better than 1.5% of the value obtained by numerical integration.
HO = ^ — i - - m u ; 2 * 2
2m 2
and 9 is a variational p a r a m e t e r , to calculate t h e ground s t a t e energy of the
Hamiltonian
H = H0 + \\x4
4
Compare the result with that obtained by first order perturbation theory. (Prob-
lem 12.4 a).
14.16. NONLINEAR SHO: VARIATIONAL CALCULATION 309
Solution
The trial wavefunction is already normalized. So we need only compute
and minimize this result. To this end we need the matrix elements (0|Jio|0) ,
(2|tfo|2) , and ( 0 | t f o | 2 ) = (2\Ho\0), as well as ( 0 | x 4 | 0 > , ( 2 | x 4 | 2 ) , and <0|x 4 |2) =
4
( 2 | x | 0 ) . Now, we have
<0|tfo|0> =
h 4
x^ -
= I V ( a t + a )
2 mui
= | — | (a 4 + a 3 a f + a 2 a^a + a 2 a t _ + a a f a 2
a|n) = \ / ñ \ n — 1)
a^n) = V n + l|n + l) (14.16.167)
we get
as well as
Finally,
F H \ O
4
( 2mu ) ^2|x |2) = (2\cra}~ + aa)aa} + aa}~a + a)aa)a + a^"a 2 |2)
fjtiAfl'KH 14. FURTHER APPROXIMATION METHODS
L A f h
E(6) = cos 2 — hu) -|- 3— [
2 4 I 2 rwu
5. ( h
+ sin -hu + 39— (
2 4 \ 2 mu
A ( h \4
+ 2 sin 9 cos 9 6\/2— ( ] (14.16.171)
4 \ 2 raw J
Using that
cos 0 = -[l+cos(20)]
we find
21A / h
+ -hu + (14.16.173)
4 V 2 mu
To minimize this energy we differentiate with respect to 9 and set the result
equal to 0 to get
Therefore,
3\/2
tan(20) = — (14.16.175)
9 + 8(hu>/X)(mu/h)2 '
3a/2A
9 rj - (14.16.176)
16fiw
If we now substitute this result into the expression (14.16.171) for the energy we
get the approximate ground state energy. To simplify the writing we introduce
the parameter
A h2
(14.16.177)
4 m2u2
14.17. WKB SOLUTION AND PARITY 311
Then,
1 3 9 72 , 45 f _ 9 72
0
~ 2 4^ 16 huj 16 hu 2 hu>
2
1 3 9 7
= —ñu H—t . (14.16.178)
2 4 4fiw
A comparison with the perturbation result (Problem 12.4 a) shows that these
results agree to order A and differ by about 14 % in the A2 term.
V(x) = V(-x)
and an energy E such that there are four turning points, say at
x = ±a , x = ±6 b>a> 0
Solution
We begin by writing the solutions in the four "external" regions
We do not need to consider the "central" region —a < x < a . Thus, with
2 777,
«(*) = —5-[V(z) — E] E<V, x <—b or x>b
¿
N
k(x) = 2m
[E — K(x)] E > V , —b<x<—a or a < x < b
~W
(14.17.179)
we have
^/§^yexP [~ lb K(x') dx
'] x
>b
312 CHAPTER 14. FURTHER APPROXIMATION METHODS
Here, we have already imposed the conditions that the solutions must be ex-
ponentially damped for |x| —> oo as well as the matching conditions at x = — 6
and x — b. Letting x —T — x and remembering that both K(X) and k(x) are even
functions of x the solutions become
^ y e x p [-¡:*K(x')dx> x > b
ip(-x) = (14.17.181)
2B
c o s -b < x < —a
y f e
exp •FB X
K(X')DX' x < —b
Vk(x)
We now impose the conditions that these solutions should be eigenstates of the
parity operator. Thus, we want
as well as
From (14.17.183) we get by changing the integration variable on the right hand
side from x to — x that
Similarly by changing the integration variable on the right hand side of (14.17.184)
we get
But, both A and B have to be constants independent of x. So, the only possi-
bility is that
A = B = 0 . (14.17.187)
In this case we have no solution at all. Thus, these solutions can never be parity
eigenstates.
BIBLIOGRAPHY 313
Bibliography
14.1 A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 4.6, (2002) .
ibid, section 14.7.
ibid, section 14.8.
Time-Dependent
Perturbation Theory
where
y _ í 0 x < -a , x > a
~ {1 — Vo —a<x<a
Find the transition probability per unit time to a state of energy Ek > 0, due
to a perturbation
x
H'(t) = ve Ia sinujt
Solution
We need both the normalized ground state solution and the normalized contin-
uum solution. The normalized ground state solution is the even parity solution
given in section 4.5 of [15.1].
If we define
2m\E\
~hT~
2m(\E\ + V0
K.2 (15.1.1)
K2
15.1. TRANSITION PROBABILITY: BOUND STATE TO FREE 315
{ A eK°x
B cos
A e~K°x
The equation for the binding energy is
KX
x < —a
|x| < a
x >a
(15.1.2)
NO
tan na = — . (15.1.3)
K
'i
and
-2K0a
\A?e~ a+ (15.1.5)
«o
Therefore, we get,
K0 K0
A= B = (15.1.6)
1+ KOA 1 KQÜ
The normalized continuum solution is obtained from the solution given in prob-
lem 8.1, as well as
dE dK and dK = — dn . (15.1.9)
m L
^ •
Therefore, the density of final states is
i?/
(15.1.10)
316 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
H' = ^ e x p [ - x 2 / a 2 ] 2 s i n w i . (15.1.11)
^ ~[exp[_<,<A + K) /2]
v I «ó 1 a-vA' r r / T.r \2/(VI
= 2V
+ exp[—q(A' — k)2/2]] . (15.1.12)
Putting all this together we finally get w, the transition probability per unit
time.
w = —\M\2p(E) (15.1.13)
or more explicitly
w = mv2a2 «o a
3
32fi K 1 + kqü
x [exp[—q(A' + k ) 2 / 2 ] + exp[—a(A' — k ) 2 / 2 ] ] . (15.1.14)
15.2 Photo-disintegration of D e u t e r o n
The deuteron is an s-wave (I = 0) bound state of a proton and neutron with
a binding energy of 2.226 MeV. It is well approximated as a bound state in a
square well of depth Vo = 36.2 MeV and width a = 2.02 x 1 0 - 1 3 cm . Using
these data, c o m p u t e the probability for photo-disintegration of the deuteron.
A s s u m e the incident photon can be approximated by a perturbation
TT __ f e^4 • r s i n w í t > 0
0 t < 0
where A is a constant vector of m a g n i t u d e about 1 x 10 3 V / c m . Use whatever
other approximations seem reasonable.
Solution
Since, Vó is the depth of the well, we define
2
i» = - ^ > 0 , K ' -- -^lf S > 0 . (15.2.15)
id2 „
-~r-^(rR)
r dr2 + K R = 0 for r < a
1^2
— -j-^(rR) — k2 R = 0 for r>a. (15.2.16)
• dr2
15.2. PHOTO-DISINTEGRATION OF DEUTERON 317
T h e solutions are
„ sin K r _
R — A for r <a
r
R = B - for r > a . (15.2.17)
The solution of this transcendental equation yields the bound state energy which
for the values given is stated to be 2.226 MeV. So the wavefunction is (up to
normalization)
f ¿si"Kr r < a
R(r) = \ .r . (15.2.20)
1 A sin A a r>a
V R
K « 9.66 x 10 1 2 c m - 1 . (15.2.21)
This means that we can neglect the part of the wavefunction that is inside the
well and simply write
„-k(r-a) -kr
R(r) « B = B' . (15.2.22)
r r
The normalization of this wavefunction yields
\B'\ 2 = ¿ (15.2.23)
so that
= í : (15 2 24
*» v Í t - - ' »
This specifies the initial state with
2 m
] f ^ w 2.32 x 10 1 2 c m " 1 . (15.2.25)
h¿
For the outgoing state we take a plane wave
«w= (1^-26)
where nx, ny, nz are integers. Now we can count the levels. The number of
modes lying between qx and qx + dqx, qy and qy + dqy, as well as qz and qz + dqz
is
Eq = Eb + hj . (15.2.29)
e y/k - -
= (15.2.31)
where
r oo /»7t
1= rdre~kr I f eiqr cosB
sin 9 d9 . (15.2.32)
Jo Jo
By symmetry, this integral points in a direction parallel to q so
I = ql or I = ~ q l . (15.2.33)
Thus,
/ = i r°° f1
J rdre~kr J qrue'qru du
4 ik2
q2(q2 + k2)2' (15.2.34)
So,
4z k 2
1 q
^ q*{q2 + k2)2 (15.2.35)
e2
. ,2 & (q A)216k4
' ' ~ T (2tt) 2 qs(k2 + q2y ' (
^
15.3. EXCITATION OF SHO 319
H = hui a^a .
At t = 0 a perturbation
V' = h£l(a^ + a)
is turned on. Find the transition probability to any excited state of the system
for t > 0. W h a t is the probability that the a t o m remains in its ground state for
t >0?
Solution
Here we have an example of the sudden approximation. So for í < 0 we have
where
a|0, a) = 0 . (15.3.39)
H = fiw(6t-c*)(6-c) + ftO(fct+6-c*-c)
q2
= Hb = hwbH - h— . (15.3.41)
UJ
This is easily done if we realize that for any two operators A, B such that
[A, B] = c-number we have
U = ec (15.3.45)
Then,
i/t = e-(ca-c*at)
(15.3.46)
So,
(15.3.48)
u
yields the desired transformation. It then follows that
n2
En = (nb + l/2)hu> - h— . (15.3.51)
U!
is
Pn(t) = |(n,6,*|*(i))|2
= ¿|(0,6|6"C/t|0,6)|2
= l|(0,6|6"|0,a>|2. (15.3.53)
Therefore,
is turned on at t = 0.
a) Find a model Hamiltonian for this system.
b) If the atom was originally in its ground state, estimate the probability P(T)
that it is in its excited state at time t.
Solution
a) The model Hamiltonian is H = HQ + H' where
(15.5.63)
b) Let
(15.5.65)
— = — id a — iQi sinwi 6
dA 2 nt
—iQi sinwí e ' B
dt
dB
—iCli sin uit e 2'nt A (15.5.69)
dt
15.6. EXCITATION OF H-ATOM 323
and
B(t) « 1 . (15.5.71)
15.6 Excitation of H - a t o m
A hydrogen atom in an excited state \n,l,m) is perturbed by a uniform electric
field. If the interaction can be written
eE-f2s'muit 0 <t<T
V(t)
0 t < 0 , t > T
Find an expression for the transition probabilities to a lower level \n', m'). Do
not attempt to evaluate the radial integrals. T h i s is how intensities of spectral
lines can be computed. You may use the results of problem 9.54. See also [15.2],
Solution
T h e state | n , l , m ) is described in configuration space by the wavefunction
T h e perturbation is
eS r(sin 9 cos ip, sin 9 sin <p, cos 9) sin uit 0 < t < to
V(t) = | (15.6.75)
0 t < 0 , t > to
P = N',V
gl(w+w„„,)to _ x
(15.6.76)
2(ui -f uinni)
324 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
En — E„i 1 1
^nn' — 2
(15.6.77)
2Hao \n n'2
N o w , we have
I (I — m + \ ) ( l + 2 — m )
= <5rn,m'
, +1 S, ,¡+i
(2/ + 3 ) ( 2 / + 1)
I (I + m)(l + m — 1)
+ (21 - 1 ) ( 2 / + 1) I-1 (15.6.78)
I (/ — m)(lm — 1)
+ SI\ / - i (15.6.79)
(21 — 1)(2/ + 1)
2(1 — m — 1)(/ + m ) !
(9, <p)
= <5, SI',I+I
(2/ + 3)(2/ + 1)(/ — m)\
2(1 + M + 2)(L + m ) !
+ °L',I-1 (15.6.80)
(21 — 1) (2/ + 1)(/ — m ) !
2
{(4^) w+(- + m — 1x 2
K I (15.6.82)
21 - 1
15.7. EXPANDING BOX 325
15.7 Expanding B o x
A particle is in the ground state in a one-dimensional box of length L. Sud-
denly the box expands (symmetrically) to a length 2 L leaving the wavefunction
undisturbed. Calculate the probability that the particle will be found in the
ground state of the expanded box.
Solution
For a box extending from -L/2 to L/2 the ground state wavefunction is given
by
Also for a box extending from — L to L the ground state wavefunction is given
by
At t = 0, where this is the time at which the box expands, the state is the same
as that of the unexpanded box. Thereafter the state evolves according to the
Hamiltonian described by the box of length 2L. Thus, for t > 0 the state is
described by
005
*«,*) = \/5 (fir)cos (x) ^ { i \2LJ
( 7VX \
—J (15.7.85)
where
„ = . (15.7.86)
This is the sudden approximation. The probability for finding the particle in the
gound state is now given by the overlap of this wavefunction with the ground
state wavefunction, namely
e,W c o s x
$0(<,x)=i V L (Y) 2 ^ ^ 2 . (15.7.87)
0 |z| > f
V' = Ax
is turned on at time t = 0, find the probability that the particle will be in the
new ground state for times t > 0.
Solution
For t < 0 we have a simple harmonic oscillator with
b= a+ c , tf = ¿+c (15.8.91)
where
C2 =
2 ^ <15-8'92)
then, the Hamiltonian for t > 0 is again diagonalized and becomes
A2
En,6 = (n + l/2)ftw - J • (15.8.94)
Then the ground state |0, a) annihilated by a is a coherent state |c) and is given
by
| C ) := e
~ C
' 2
Y , - n |n'6>' (15.8.95)
n=»v«!
At t — 0 the state is given by the first excited state of 11a and is
But,
OO
C
+ c)|c) = e ~ ° 2 t 2 y;_=[c|n,&)+v/ir+T|n+l,6)]. (15.8.97)
v"!
15.9. SUDDEN PERTURBATION OF TWO-LEVEL ATOM 327
w
rn
9
X Y -^=[c\n, b) e~inwt + V^Tl\n + 1,6) e — ^ + 1 ^ t ] . (15.8.98)
nr r=f0, VN\
Therefore the probability amplitude for finding the particle in the new ground
state is (0,6|\P) and the probability is given by
| ( 0 , 6 | ^ ) | 2 = c 2 e~°2 . (15.8.99)
where
1 0
0"3 = 0 -1
0 1
H' = V .
1 0
is turned on, find the probability t h a t this atom is still in its ground state at
some later time t.
Solution
If we make the replacements
the Hamiltonian is
and
or
1 0
H = hQo ( J ^ ) + «I! ( J J ) = > • (15.9.103)
The eigenstates of Ho are, as before,
with
E± = ±KLo . (15.9.105)
(15 9 107
* ( nt ) ( » ) = <' ( I ) ' -- '
The eigenvalues are:
£± = ± f t ^ í 2 g + í2j . (15.9.108)
Calling
(15.9.109)
we have
e± = ± h Q . (15.9.110)
We then get:
V± ± 0
~ v ^ Q ( 0 - fi0) ( - ) (15.9.111)
If at t = 0 the system is in the state u+ = |+), then for t > 0 the system will
be in the state
tf(f) = b+v+e~int + fe_w_eint (15.9.112)
where b± are determined by the initial condition
u+ = b+v+ + 6_t;_ (15.9.113)
or
b±=(v±,u+). (15.9.114)
Thus,
6± = . Ql (15.9.115)
v
^ ( q í f l o )
Combining these results we get
(>«»•)
This wavefunction is an exact solution of the Hamiltonian (15.9.101) and
(15.9.102) with the initial condition *£(0) = u+. The approximation was made
in writing the Hamiltonian (15.9.101) and (15.9.102) in the first place.
X t/ • J- vy • j-rj-jx. u i i iJ ±
15.10 Berry's P h a s e
where
so t h a t
-L
where ^J-/J,BO is just the dynamical phase
27t/uj
E(t) dt
and
—7r(l — cos 9) - — A i l / 2
is the geometric phase (Berry's phase).
a) Use the formula
7„(C) = £(n,B\VB\n,B)dB
d-li .( dv± \
= L V ±
-DF { ' ^ R )
(9(t = 2 j r / w ) | t f ( * = 0)> .
— v j- j i r i J -iyr,riiwBJVT PERTURBATION THEORY
Solution
We are given
H f f
(t) = - ^ B ( t ) (15.10.117)
where
so that
- 7 r ( l ±COS0) = ( 1 5 . 1 0 . 1 2 2 )
1^(0) = ex
P - j( E(t')dt' p!'Tn(í)| ,B(t)). (15.10.123)
Now we apply the theory developed for Berry's phase (see [15.3]). To illustrate
the theory we solve the problem by three different methods.
a) Thus,
S ( 0 ) = B(2TT/U>) . (15.10.125)
H(t) = - ^ f f - B ( t ) . (15.10.126)
£ ± = ± ^ | i . (15.10.128)
if - íjf * TSIDV
2
= . (15.10.130)
hid
The geometrical phase (Berry's) phase is given by
where
B(t) = 5o(sin 9 cos cot ex + sin 9 sin cot ey + cos 9 ez) . (15.10.132)
VBv±= ( ¿ - . - ¿ - 4 . - 5 - ( 1 5 . 1 0 . 1 3 3 )
\ a B o #o aw B 0 sin y awi /
and
Now, we are finally ready to compute the geometrical phase using (15.10.131)
b) Another more direct way to compute the geometrical phases is to use the
equations
í a = , ' ( „ ± , ^ ) . (15.10.139)
^ = -wcos2(0/2) . (15.10.140)
Then,
/j 2n/u)
7+(C) = uit cos 2 ( —)
¿
0
2
= —27t cos (0/2)
= — 7r(l + c o s 0 ) . (15.10.141)
Similarly,
DF.
= —to sin 2 (0/2) (15.10.142)
dt
7-(C) = - w i sin 2 ( —)
c) Finally, we can also simply just solve the time-dependent Schrodinger equa-
tion exactly. To do this we look for a solution of the form
/ p-iwt/2 \
*(*) = ,„.W2
beiwt/2
K * (15.10.144)
with
M2+|6|2 = l .
¿/I— = #(*)*(*)
15.11- NEUTRON IN ROTATING MAGNETIC FIELD 333
now reads
ft(w/2 - A ) a e - W 2 - A ) '
-h(u/2 + A)6 e ' ( w / 2 + A ) i
hb0 -iwt ae-'(W2-A)t
cos Í7 sin c/ e (15.10.145)
Oicot f)ei(w/2+\)t
2 V sintfe""' —cos
/ /xbo \ w
— cos 0 +
2/5 V 2h ) 2
II is (15.10.147)
v 2h 2
Here we have used the approximation that
u « fib0 •
¥(0) = (15.10.148)
\2tt
($(27r/w), >í(0)) = exp i—A — in
U>
= EXP ITT n _l a\
( 1 ± c o s i )
(15.10.149)
±2^7-
Solution
In this case the Hamiltonian consists of
where a is the angular frequency of rotation of the magnetic field b(t) and
eB eb
= U = i
^2mc ' 2 ^ ' (15.11.151)
(15.11.152)
a_ = 1 (15.11.154)
and
Therefore,
2í2 -|- GL j
is given by
A, .2
p + sin2
~ 10o _l \2¿ ( 2 q + a )T • (15.11.158)
+ a)
4w 2
(2fi + a) 2
y/ÑR
Solution
£ 2
V(t) = -ex-^e~<-t/T)2 . (15.12.159)
Y/TTT
The probability amplitude for a transition from the state |0) to the state |m) is
1 f°°
an,o = ¿n,o + — Yt / dt (n\V(t)\m)amfie,u'nmt (15.12.160)
tfl
M J~°°
where
Unm = {En — Em)/fr= (n - m)uj . (15.12.161)
e_(t/T)2
= + Vm + l i „ , m + i ) . (15.12.162)
Therefore,
a
1 CS I TX
n,0 = ¿«,0 - {VmSn,m-l + Vm + l¿n,m+l)
/ oo
dt e_(t/,T'2 e
iu n m t
(~ )
-oo
= ¿n,0 - (V™&n,m-1 + V™ + l¿n,m+l)
v
in J-KT V 2 ma; '
m
xam,0
%
=
<5n,0 + y/2mtvJe ( v / n T T o n + 1,0 + l / " a
n-l,o) • (15.12.163)
336 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
This equation is still exact. To use perturbation theory we simply iterate this
equation. Then,
= w T
«n,0 ¿n,0 + ^ ^ e t4 (y/n + 1 + i o + y/ñ CIrí-l,o)
•
This equation also shows that, to lowest order, a transition can only occur to
the state with n = 1. The resultant probability is
|ai,o|2= e_w2r2/2
• (15.12.165)
Zmriuj
15.13 N e u t r o n Magnetic M o m e n t
A neutron has a magnetic moment fi n . A free neutron is placed in a magnetic
field B pointing in the z-direction and at time t = 0 the x-component of the
spin of the neutron is measured to be h/2. If this measurement is repeated at
some later time i, calculate the probability of obtaining the same result.
Solution
The Hamiltonian describing the evolution of the spin of the neutron is given by
H = —pn • B (15.13.166)
H = tkj<T3 (15.13.167)
where, as always,
<r 3 = ( J ^ ) . (15.13.16K)
V(< = 0) = - ^ ( J ) . (15.13.169)
ih^=hw<T3iP . (15.13.170)
Therefore, the probability of obtaining the value h/2 for a measurement of spin
along the x direction is given by
Solution
If we take the magnetic field pointing in the z-direction and the velocity in the
¡/-direction then we can take the spin pointing in the ^-direction. In this case
the problem is like problem 15.13 above and the probability of obtaining the
spin to point in the same direction after a time t as it was pointing at time t = 0
is given by
cos 2 uit
where
¡XB
w =
x '
In this case the time t is just the time that the electron spends in the magnetic
field and is given by
t = L/v .
Therefore, the probability that the electron's spin points in the same direction
after emerging from the magnetic field is
P = cos2(uiL/v) . (15.14.174)
Solution
Before the spring constant changes the state of the system is described by the
wavefunction
(15.15.175)
where
U) = Y/K/R (15.15.176)
M X ) eXP
={-^) H
x
'
(15.15.177)
For t > 0 the wavefunction evolves according to the new Hamiltonian and is
given by
ü n
$(x,t) = Y (15.15.178)
where (f>n is the wavefunction for the nth excited state corresponding to the
energy (2n + 1 )hu¡ of the new Hamiltonian. At t = 0 this wavefunction must
coincide with ip. So,
(15.15.179)
The probability that the system winds up in the new ground state is now given
by
r
2
m =
a/2 mu> 3 raw
exp dx
7T h 2h '
J —c
Y/2MUI 2IVH
7t h 3 muí
^ = 0.9428 (15.15.180)
15.16 C o u l o m b Excitation
A particle with charge Q and initial momentum
P -- Mvéz
is incident from infinity on an atom with Z electrons and energy levels En such
that the impact parameter is b.
15.16. COULOMB EXCITATION 339
is given by
iOe [e^ •
A}i =
~~hvbJ (xij sin9 + zíj cos 6) d9 (15.16.181)
where
hiofi — En< En
2) The dimensions of the target atom are much smaller than R(t) = |R(t)\ •
3) A dipole approximation is used.
For details on the validity of these approximations and further discussions see
[15.4],
Solution
Let Ho be the unperturbed Hamiltonian so t h a t
where r¡ are the coordinates of the electrons in the target atom. Since the
dimensions of the target atom are much smaller than R(t) we can write
1 oo I
r'
f 0 w ) ' (15.16.186)
| R(t)
Since we are using a dipole approximation we may truncate this series after
1 = 1. However, for 1 = 0 the perturbation is just a constant and since the
340 CHAPTER 15. TIME-DEPENDENT PERTURBATION THEORY
initial and final states are different and therefore orthogonal we see that the
I = 0 term does not contribute. Thus, the perturbation reduces to
" -
Z _Q
= sin
DU\-> + Z> c o s
(01 • (15.16.187)
Here we have used the fact that the trajectory of the incoming particle lies in the
x — z plane so that $,-(<) is the instantaneous angle between the z-axis and R(t),
the vector from electron i to the instantaneous position of the incoming particle.
Again, since |r¿| < < R(t) we see that all angles #¿ as well as coordinates x¡, Zi
are the same and we may as well call them 9 and x, z. Therefore,
—OZe
V= 9 [xsin0(¿) + zcosfl(<)] . (15.16.188)
R(ty
Thus, using first order perturbation theory we have t h a t the transition ampli-
tude Aif from |¿) to | / ) is given by
1 r°° •
Afi « jr e-^(i|V| f)dt
in
J-OO
A =
" — V V /
etu>fit (f\xsin9
+ zcos0\i) d9 . (15.16.191)
•7r
Here, 9(b) is the asymptotic value of 9(t) for t —>• oo .
Bibliography
[15.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd. (2002), section 15.9.1.
[15.3] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd. (2002), section 15.11 .
341
BIBLIOGRAPHY
P a r t i c l e in a U n i f o r m
M a g n e t i c Field
Solution
We want to estimate the two terms
e
* = Ú-csz ' =
For this purpose we take
B-L = hB and \A\ = ^\r\\B\ (16.1.2)
with r = |r| = typical atomic distance = 1 0 - 9 m.
Then we find
2mc 2 8 x 9.1 x 1 0 - 2 8 x (3 x 10 1 0 ) 2
17
= 3.16 x 10~ erg = 2 x 10~ 5 eV . (16.1.4)
16.2. RADII OF LANDAU LEVELS 343
»o = + 2/o
where x0 and y0 are given by
n = p+-A. (16.2.7)
c
Here, A is the vector potential for a uniform magnetic field and u = eB/Mc is
the Larmor frequency. Interpret the meaning of this result.
Hint: Use the symmetric gauge
A = y(-2/,x,0) •
Solution
We want to solve the eigenvalue equation
r
o f i — al fi (16.2.8)
where
r l = x l + yl . (16.2.9)
xpy - ypx
Therefore,
' (16.2.13)
[l„ht] = 0 .
2h h
r = + 1/>2 =
° ~ ~mZ^H ~ m +
^ " (16.2.15)
.o HE FI
A 1 6 2 1 6
= 73 = MZ I ' )
r2 = A 2 ( / + l ) . (16.2.17)
A2
ax0ayo > y . (16.2.18)
p = -e '*$
golution
The proof is a straightforward use of the Schrodinger equation and its complex
conjugate
ih^ ~ [ V + ie/(hc)Á}2*-e<W
ot 2m
-ihd-^~ = -^-[V-ie/{hc)Á]2V*-e<t>V . (16.3.19)
ot ¿m
If we multiply the first of these equations by and the second by \f> and
subtract the two equations from each other we get
F) f 1£ —*
IH—(V*<A) = 1 + — [«'a-
^ = V-J. (16.3.21)
After multiplying by the electric charge — e we have the electric current and
charge densities given by
2im i J
p = (16.3.22)
Solution
We again start with the time-dependent Schrodinger equation in the presence
of an electromagnetic field (<¡>, A)
ifi
^ =¿(f-F)2,1,+«^ (16 4 23)
'-
Now define the gauge transformations
and
A' A + VA
- <T>--D4- (16.4.25)
c dt
Then, inverting these definitions and substituting in the Schrodinger equation
ih
we find
i
1
exp | - ^ A | » ' P~ ~{A' — VA)j' exp ill A
2m he
i , 1<9A\
exp ( - j p - A ] \P' (16.4.26)
Á -+ÁA = Á+ VA
= (*A,o(p,ÁA,<i>A)*A).
The left side may be rewritten as
(tf,0(p,Í,<¿)<5) = (*A,oA(p,Á,<t>)<¡>A),
where
0 a ( p , a <T>) = e , '(9/' ic ) a 0 ( p , a , FT e - ' ( 9 / f t c ) a
is the unitary transform of the operator 0(p,A,<j)) t h a t maps O —> 0A. Show
that the resulting necessary and sufficient condition, namely
0A{p,Á,<p) = 0(p,ÁA,<j>A)
Solution
Clearly the condition
-» $ a = e*(9/ftc)a ^ (16.5.32)
we have that
Therefore,
0 ( p , Í A , <¡!>A)tf A = e i ^ ' h c 1 A 0 ( p + - A , <t>)V (16.5.34)
c
and hence, with
we have
0(p,A,<f>)=p (16.5.37)
and we need
p$A = e ^ / ^ V l r . (16.5.38)
mv = p+ (q/c)A (16.5.39)
is.
Solution
The Hamiltonian for this problem is
i(0) = ( I ) • (16.6.42)
^0 ,
where
w= ^ . (16.6.44)
A
16.7- SPIN 1/2 IN MAGNETIC tlULu: ^
T h e solution is
(sx) = 0
(sy) = 0
(s2) = J. (16.6.46)
b) In this case we t a k e
B = (B, 0 , 0 ) . (16.6.47)
T h e solution is
(sr) = 0
. . H .
(sy) = -sm2ui
Solution
The Hamiltonian for this problem is as before
H = —¡1 • B = —[/.a • B . (16.7.51)
a) In this case we again take
5 = (0,0,b) , s = - a . (16.7.52)
M = (J (16.7.53)
Íh< = (16.7.54)
^t ^
and when written out in detail become
dsx
dt - 2 UlSy
DSY
^ = -2 usx
dt
ds2
= 0 (16.7.55)
dt
where, as before,
fiB
w = (16.7.56)
x '
The solutions are
s
y (') sy (0) cos 2 + 'y W sin 2u>t
1U>
Sz(t) = S*(0) (16.7.57)
<M0)> = o
(s y (0)> = 0
h
< M 0 ) > (16.7.58)
2
The two operators ¿ r ( 0 ) and sy(0) are determined by the initial conditions that
may be read off the Heisenberg equations.
sx(0) = 2ws y (0)
s =
y(0) —2ws j; (0) . (16.7.59)
16.7. spin 1/2 in magnetic field: hmijuivunnu zu¿uj\±iviyj ool
Thus,
(¿,(0)) = 2w(s y (0)) = 0
<¿y(0)> = —2w(s r (0)) = 0 . (16.7.60)
Therefore, we obtain for the time-dependent expectation values
(s x (t)) = 0
0
<«»(<)> =
(Sz(t)) = ^ • (16.7.61)
(s2(0)) = \ . (16.7.65)
(sy) = — sin2wi
/ \ H
\sz) - — cos 2uit . (16.7.67)
These are, of course, the same results as obtained by using the Schrodinger
equation.
352 CHAPTER 16. PARTICLE IN A UNIFORM MAGNETIC FIELD
Solution
The Hamiltonian is
H = H0 - na • B = ( p ~ c / r - 4 ) +ev . (16.8.68)
2 777
We attempt to find a solution of the form
íh
~^x + íh
^^t =
(h°^ x
~ • (16.8.70)
dib
i h = H0ip (16.8.71)
and
, ( cosae-'^ \
^ u n a e ' - » ) •
a) Calculate the expectation values of the x and y-components of the spin.
b) Find the direction in space such that at any time t the 2-component of the
spin has a well-defined value along this direction.
16.9. SPIN 1 /2 IN TIM K-DEPENDENT MAGNETIC FIELD 353
Solution
If we write
= ( b(t) (16.9.73)
h .
sin(2a) cos HJ uj(t') dt' + ¡3 (16.9.77)
2
Similarly,
Hence,
Solution
The Schrodinger equation for the wavefunction
(16.10.82)
is
-ITOT
ih- d_ f a(t) \ =_ —^B cos a sin o e a(t)
(16.10.83)
dt b(t) sin f e icut COS ( b(t)
If we write
d2a . da / i i \ r\
+ lüJ (16.10.86)
dp ~dt + ^P~ q
~ + Uq a
^ =
'
If we now call
fii = + \ / w 2 / 4 + p2 + q2 + u>q
a(0)
M =
6(0)
a
l + a
2
(16.10.91)
, • a „ - h f cos 6 sin 0
v(sin 9, 0, cos 0) • s = - . . „ (16.10.92)
' ' 2 V sin 0 — cos 9
After a short computation we get that, up to normalization
0,2 - q~
ai = A tan 0/2 +
P
Sli-q
a2 = A tan 9/2 — (16.10.93)
p
The normalization yields t h a t
|2 p2 cos 2 0 / 2
km 2 = (16.10.94)
(n2-«i)2
ooo itlAflUH 16. f ARTICLE IN A UNIFORM MAGNETIC FIELD
(sinOcosu!t,sm0sinu>t,cos9)s= — i . cos
. t) ,, sin 9 e~-iuit \
v ;
2 V sin 6 e J (16.10.95)
— cos t
corresponding to the eigenvalue -h/2. Again, a short computation yields
sin( 2
V (16.10.96)
W) - y -cos9/2eiut
Then, the probability t h a t at time t the particle will be found to have the spin
value -h/2 aligned along the magnetic field is given by
p(t) = !(</>(<),v(*))|2 .
Thus,
p cos 9/2 ft2 - q
P = sin 9/2 — tan 9/2 +
2
(ft2-fti)
CLI-Q
+ sin 9/2 tan 9/2 —
fti - Q ft2 - q
— cos 9/2 p — tan 9/2 + -iflit
P
^2 -q fti - q jn2t
— cos 9/2 tan 9/2 —
p
ill - q
2
sin 9/2 — cos 9/2
(ft2-«i)
\
sin 9/2 — ——-cos9/2J [1 — cos(ili — Q2)<] (16.10.97)
2p 2
P = sin 0/2 i cos 9/2) x
w 2 + 4(p 2 + q2 + uq) P
^2 - q
sin 0/2 — - c o s 9 / 2 j cos 2 (^JUJ2/A + p2 + q2 + uiq .
(16.10.98)
Here,
Bibliography
[16.1] D.H. Kobe, Am. J. Phys. 54, 77, (1986).
C h a p t e r 17
Angular M o m e n t u m , Etc.
T = A\j2J\- — j\ J2-]
T\j,j,ji,h) = \j ~ 1, j ~ l j u h )
for j = ji + j 2 -
Solution
We consider the "unnormalized" operator
T={hJi-hh)/h. (17.1.1)
Now we have t h a t
Furthermore,
J2T\j\ + Í 2 , j l + J 2 ) JI1J2)
= [ji (¿1 + 1) + h (¿2 + 1) + 2JU2
— 2(ji + +Í2, jl + J2, Jl J J2)
= (¿1 + 32 - l ) ( i i + j 2 ) / j 2 r | j ' i + 32, ji + 32,ji,32) • (17.1.6)
Solution
The state \l + 1/2, m, I) is obtained by applying
J_=L_+S_. (17.2.7)
So we find
/C \ _ ( / + 1/2 + m)! (2l+l-2m)
[ z)
(2/ + 1 ) ! ( / + 1 / 2 -m)\
(/ + l / 2 + m ) ! fc —(2/+1 —2ro1
(2/ + 1)!(/ + 1/2 — m)!
where we have used the fact t h a t more than one application of S_ to the state
j 1 / 2 , 1 / 2 ) annihilates it. Thus, what we have to evaluate are quantities like
(/, / | ¿ ' + - ( m - 1 / 2 ) L ( _ - ( m - l / 2 ) |Z) ^ a n d ^ Z |i/-(m+l/2)¿l-(m+l/2)
11,1).
(17.2.9)
Hence we get t h a t
(/ + 1/2) 2 -m2 + 1
( 5 2 ) = /i (17.2.12)
( 2 / + 1 ) ( 2 / + 1 - m) '
For the states |/ — 1 / 2 , m, I) we proceed in exactly the same manner starting
with the equation
\l- 1/2,1- 1/2)
1
= \l,l - l ) | l / 2 , 1 / 2 ) - y Z | T | Z , / ) | l / 2 , - 1 / 2 ) . (17.2.13)
V2T+
Then,
(/ + m + 1/2)! 1
(5.) = (1,1- 1|(1/2,1/2|
(2/ + l ) ! ( / - m + l / 2 ) ! 2Z + 1
21
J jl-m+1/2
'1(1/2, — 1/21 +
21+ 1
X 5,j'_m+1/2 v^m-1)11/2,1/2)
2 ^ - l í , 011/2,-1/2) (17.2.14)
COUPLING OF SPIN 1 TO SPIN 1/2 361
This simplifies to
h [ l m + : ) + m ( m _ 2 ) + ? / 4 ] ( 1 7
(Sz> = ~ ' '2' 1 5 )
and
Jz = J lz + J2z
Solution
The possible eigenstates of J2 and JZ are
and
|l/2,l/2> , |l/2,-l/2).
13/2,1/2) = ^ j | l , 0 ) | l / 2 , 1 / 2 ) + - 1 | 1 , l ) | l / 2 , - l / 2 ) . (17.3.18)
+ ^|l,0)|l/2,-l/2) (17.3.19)
so that
v2,
| 3 / 2 , - l / 2 ) = ^=11, - l ) | l / 2 , l / 2 ) + ^ | l , 0 ) | l / 2 , - l / 2 ) . (17.3.20)
362 CHAPTER 17. ANGULAR MOMENTUM, ETC.
Finally,
|3/2,-3/2)= | l , - l ) | l / 2 , - l / 2 ) . (17.3.21)
The remaining two states must be linear combinations of
|l,0)|l/2,l/2) , |l,l)|l/2,-l/2)
and
|l,-l)|l/2,l/2) , |l,0)|l/2,-l/2)
and must be orthogonal to all the states above. Thus we find (up to an arbitrary
phase factor)
Another way to proceed for these two states would be to use the operator T of
problem 17.1. Then we find that, up to normalization,
11/2,1/2) = A[S-- 1/2 L _ ] | l , l ) | l / 2 , l / 2 )
which agrees with our previous result. The state 11/2,—1/2) is obtained by
applying J_ to this result.
Solution
We begin by taking
( a d e \
f b g . (17.4.27)
h k c )
Then using
[J*,T 0 ] = 0 (17.4.28)
we find
( a 0 0
To = 0 6 0 | . (17.4.29)
\ 0 0 c
Next we use
and find
/ 0 6—a 0 \
Ti = 0 b c-b . (17.4.31)
\ 0 0 0 /
that
/ 0 0 0 \
T_i = j a — b b 0 J (17.4.33)
\ 0 b— c 0 J
The condition
T l = (—l)mT_m (17.4.34)
simply shows that the three numbers a,b,c are all real. If we now use
a — 26 + c = 0 . (17.4.36)
Thus, we get
6= . (17.4.37)
364 CHAPTER 17. ANGULAR MOMENTUM, ETC.
Therefore,
„ c-ol ,
T\ - —-ZZT—J+
2Y/2 H
m c—a 1 T c+ a
T o = J Z
2 ~ H ~ ~ 2 ~
711 =
* i17"4'38)
RNITP) =
and
o) (l
b) If j = 1 repeat part a) but consider the states
1\ / o x / 0
0 , 1 and 0
0 / V 0 / V 1
Solution
a) We start with
ñ = (sin a cos /?, sin a sin/3, cos a ) . (17.5.39)
Then for j = 1/2
^ or2) (17.5.40)
and
ñ •J _ 1 / cosa sinae-1^
(17.5.41)
h 2 \ sin a e8'3 cos a
2
/ -\
I ñ • J\ l / l 0 \ 1
L (17 5 42)
n o i r ? ' '
Thus,
2n
¿ 1 (17.5.43)
17.5. rotations f o r spin 1/2 and spin 1 365
f ^ —A 2n + l
ñ •J \ 1 fi • J
= (17.514)
y~~h~J 22"+1 ~~h~ •
i sin(<p/2) p-'/ 3
Rn(f) (17.5.47)
cos(<p/2) — ¿sin(y>/2) c o s a
we have
^ sin a e 0
ñ • J
4 ^ sin a e'1* 0 cos o ; 17.5.49)
0 _l ... j b
\ — C O S (V
Thus,
2n + 1 2n
n •J Tl • J ñ •J n • J
(17.5.51)
where
H•J
Solution
The states of total possible angular momentum j are given by
To get the state |1,1) we use the result of problem 17.1 that, u p to a normal-
ization factor A we have
A
| ( | h-1) - | ( | - 1)13/2,1/2)11/2, l/2>
2
+ l) 1)|3/2
- Y V ~ ~ ' 3/2)|l/2, -1/2)
= | [\/3|3/2, l / 2 ) | l / 2 , 1 / 2 ) - 3 | 3 / 2 , 3 / 2 ) 1 1 / 2 , - 1 / 2 ) ] . (17.6.55)
Normalizing we find
|1,1) = i | 3 / 2 , 1 / 2 ) 1 1 / 2 , 1 / 2 ) - ^ | 3 / 2 , 3 / 2 ) 1 1 / 2 , - 1 / 2 ) . (17.6.56)
¡1,0) = a7_|l,l)
= A[\3/2,-1/2)11/2,1/2) -|3/2,l/2)|l/2,-l/2)] (17.6.57)
so that
Finally,
|1,-1) = AJ.\L,0)
= ^ [ > / 3 | 3 / 2 , - 3 / 2 ) | l / 2 , 1 / 2 ) - |3/2, - l / 2 ) | l / 2 , - 1 / 2 ) ] .
(17.6.59)
So
|1- - 1 ) = ^ 1 3 / 2 , - 3 / 2 ) | l / 2 , 1 / 2 ) - | | 3 / 2 , - l / 2 ) [ l / 2 , - 1 / 2 ) . (17.6.60)
Solution
Starting with the state of highest weight we have
Applying
J- — L- + S- (17.7.62)
368 CHAPTER 17. ANGULAR MOMENTUM, ETC.
Solution
We are given that
A = ~ r x B (17.8.68)
pÁ+Á-p = — - p - ( f x B ) — -(rxB)-p
= — (r x B) • p
= (B x r) • p . (17.8.71)
Next, using the fact t h a t the dot and cross product may be interchanged as long
as the order of the vectors is maintained we get
p Á+Á-p= B (rxp) = B L . (17.8.72)
j7.9. THE STATE \n,j,m,l) 369
Solution
I n,j,m,l)
We want = J w +tom show
to use induction
)!
that (17-9-73)
( 2 j ) \ ( j - m)!
We start with m = j. In this case the (17.9.73) is clearly correct. For m = j - 1
we have
J-\n,j,j,l) = \ / j { j + 1) - j ( j - l)H\n,j,j - 1,1) . (17.9.74)
So solving for \n,j,j — 1,1) we again see t h a t (17.9.73) holds. Now assume t h a t
(17.9.73) holds for m ~ j — M. Then, using
J-\n,j,j - M,l)
= VIU + 1) - U - M)(j - M - l)h\n,j,j — M — 1,1) (17.9.75)
so that
In,j,j- M -1,1)
1
h lJ^\n,j,j — M,l)
VjU + 1) - a - M)(j - M - 1)
l(2j - M)\ 1
H-W+VJ"+1\N,J,J,L)
(2j)!M! sj(2j — M)(M + 1)
(17 97<i
= - »
as required.
and
c) hence evaluate the m a t r i x element for the weak-field Zeeman effect
Solution
T h e first p a r t is simply a tedious c o m p u t a t i o n . We therefore first c o m p u t e
[J2, [J2,AX\[ . T h e n , by s y m m e t r y , t h e result also holds for Ay and Az. As a
first step we c o m p u t e
Similarly we find
Then,
= —h"[jyJxAy — JyAX + Jy Ay JX ~ Jy AX Jy
+ JyAy
Jy Jy Ax Jy + Ay JxJy AXJ~
+ JZJXAZ — JZAX + JZAZJX — JZAXJZ
+ JXAZJZ — JZAXJZ + AXJXJZ — Ax J~]
2 2 2 2 2
= h [(J + J )AX + AX(J + J )]
h [Jy(2AyJx + ihAz) — (AxJy — ihAz)J, y
Jy(Jy ""t~ ihAZ) -f- (AyJX + ihAZ)Jy + Ay(jyjx + ÍHJZ)
+ JZ(2AZJX — ihAy) — (AxJ)z + ihAy)Jz — JZ(JZAX — ihAy)
+ [AZJX - ihAy)Jz + AZ{JZJx - ihjy)] . (17.10.79)
[Jr,,4x] = 0 . (17.10.80)
17.10. LANDÉ g-FACTOR 371
or
0 = 4h4J(J + \){JM\SZ\JM) — 4h3M(JM\S • J\JM) . (17.10.85)
Thus, we have the desired result
This means that the eigenvalue of s ñ is h/2. The angular momentum wavefunc-
tion of this particle is given by the wavefunction for total angular momentum j
derived in section 17.7 of [17.1], namely
Find the relationship between the angles (a, /?) and (6, tp).
Solution
The eigenfunction of s • ñ corresponding to the eigenvalue h/2 is given by
2 (
h cos a sin a e ,IS H
(17.11.92)
sin a e l P — cos a 2
and
1/3
Y = cot a/2 e . (17.11.95)
B
(17.11.96)
P = <P
17.12
CLEBSCH-GORDON
value fflj = 1 / 2 ?
Solution
By s t a r t i n g f r o m
and applying
J_=L_+S_ (17.12.99)
0 0 < < 7T
V(v) =
-VO IT < tp, 2n
Solution
T h e H a m i l t o n i a n for this problem m a y be written
H = £ + V(<p) (17.13.101)
If, using the fact that we restrict ourselves to positive energies, we write
2 IE 21(E + V0)
K2 =
2
(17.13.103)
h2 ' h2
the eigenvalue equation for the energy becomes
2
:d ip
J-~2 + k2ip = 0 0 < ip < TT
dip
d2i>
= 0 7t < if < 27t . (17.13.104)
dipr¡2¿
The solutions are
we get
A eik2n + B e~ik2w = A + B
id is2
Ce * + fle" * = C+D (17.13.108)
Writing all this out we find a set of four homogeneous equations in four un-
knowns. For a nontrivial solution to exist requires t h a t the determinant of
the coefficients must vanish. This is the desired transcendental equation. The
required determinant is
0iirk e-ink
ilrk
kr e - k :e e~ i n k —K E
D = i2nk (17.13.109)
-1] [e -i2nk - i ] 0 0
i2 -Í2TTK
0 o [e
0 Í2HK™ - 1] -i]
This determinant when evaluated yields the value
K + k —n n = l , 2 , 3, . . . . (17.13.111)
17.14
SPIN DEPENDENT OPERATORS
Particles
If we have two particles with spin operators
h h
e - Q ~
O\ - 20"1 ' '"'2 =
20"2
separated by a distance r = fir show t h a t any positive integral power of either
of the o p e r a t o r s
A = 5\ • <72
or
Solution
We show t h e result s t e p by step. Recall t h a t for the Pauli m a t r i c e s we have
and
A2 = ai • <72<71 • <?2
= (a\x<^2x + ""1 y&2y + V\z&2z) (&IX&2X + &\y&2y + &1 z&2z)
=
3 — 2 ( a i x a 2 x + Vly&2y + (Tlz^z)
= 3 - 2 A. (17.14.115)
Next we c o m p u t e
Therefore,
(<?i • <?2)(<?i • ñ)(<?2 • ñ) = [ ñ + i(a 1 x ñ)][ñ + i(a2 x ñ)]
= 1 — (<?i x ñ)(<x2 x ñ)
= l-ai + -ñ)(a2 ñ) . (17.14.119)
376 CHAPTER 17. ANGULAR MOMENTUM, ETC.
So, we find
Finally,
si2^4 = 5 i 2 . (17.14.122)
Bibliography
[17.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., chapter 17, (2002) .
C h a p t e r 18
Scattering - Time
Dependent
Solution
We use the equation
AN = J N 0 A O . (18.1.1)
ail
Now we have
9
I = 5 x 10 A = incident p r o t o n current
3.12 x 10 1 0 p a r t i c l e s / s (18.1.2)
1.6 x 10~ 1 9 C / p a r t i c l e
J = I / A (18.1.4)
AN = 10protons/s (18.1.6)
da 1 AN A I AN I AN
(18.1.10)
dfi J No AÍ2 / Ap Aft Ip Aft '
Therefore we have
= 1.34 x 10 - 25
cm2/atom/steradian . (18.1.11)
aft
g t m = - l e - i h o t / n e,ts (18.2.12)
H
and evaluate the matrix elements {p\G^ (<)|£) where |p), |£) are free particle
states of momentum p and k respectively.
Solution
We begin with
_I ( MOT
G U t ) = T ^ e x p ( - i - J - ) 8{±t) . (18.2.14)
H0\p)=^\P). (18.2.15)
Then,
(p\G±\k) = T faexp(-i?jf\\k)0(±t)
f ( z ) = J dt eitz F(t)
where F(t) = 0 for t < 0. Show that / ( z ) is an analytic function for Si(z) > 0.
b) Consider a function f ( z ) which is analytic in the upper half of the complex
z-plane as well as on the real axis and vanishes rapidly for |z| —> oo. In that
case we may write
(18 317
= é ¡ f - '
where the contour runs along the real axis from — R to R and is closed by a
semicircle in the upper half-plane. Let R —>• oo and let z —> the real axis from
above and use this result to express 5f(/) in terms of 9 ( / ) and vice versa.
Hint: Recall that
lim —^— = P— ¿7rá(a;) . (18.3.18)
e-*0 + X ± It X
c) Verify the result obtained by applying it to the function
Solution
a) The proof that / ( z ) is an analytic function for 9 ( z ) > 0 is straightforward
since by assumption the integral
p 1 d
= h jZ¥ * *' •
So carrying out the delta function integration and simplifying we find
P
c¡[f(x)] = ~ l J • (18-3-24)
This is the required result. The pair of equations (18.3.24) are known as dis-
persion relations. These relations may be combined with a statement of causal-
ity, such as is assumed in part a), to yield important conditions for scattering
amplitudes. In optics the resulting equations are known as "Kramers-Kronig
relations".
c) If the given function is
• (18.3.26)
To verify the dispersion relations for this example we simply write out the
integrals
/
G%(t-t')VG±{t')dtl (18.4.29)
•OO
± ±
to obtain an equation for the Fourier transform G (Ü;) of G ( I ) in terms of the
Fourier transforms of G^(<). Write a formal solution for G ± (a>).
18.5. BORN APPROXIMATION 3»i
Solution
We begin with the integral equation for G ± ( ¿ ) .
/ O O
Now define
/ O O
eiuiG^{t)dt (18.4.31)
-OO
/ O O
eiutG±(t)dt . (18.4.32)
-OO
Then,
/ O O r OO
dt / e i w ( t - t ' ) e i w f ' G Í ( í - í ' ) l / G ± ( < ' ) di'.(18.4.33)
- OO J — OO
G ± (w) = G f ( w ) + G Í ( w ) K G ± ( w ) . (18.4.34)
[ l - G j ( w ) K ] G ± ( a ; ) = Gf(w) (18.4.35)
and
e~^r
V = -V0
Solution
The T - m a t r i x on shell is defined by
where
TFK2 o /i 2 9 2
EJ —
Thus,
_i R E-V-R .r _
T i = , q r V o (18
' j2ñyiJ ~ ~ -541)
So,
P = |*-«1 (18.5.43)
Tfi = . (18.5.44)
271"" \k — q]2 + p.2
Solution
Comment: The first thing to realize is that the calculations involved here are
completely classical.
In the laboratory frame a particle of mass mi, velocity vx is incident along the
¿-axis on a particle of mass rn-¿ at rest so that its velocity is v? = 0. This gives
rise to a differential cross section dcr^^L, ipi,)/dS2L for the incident particle.
18.6. SCATTERING IN CM AND LABORATORY t namn,
p - mv
p = mi(vi — V) = mvi
—p = 7772(0 — V) = —mv1
p = — mv
P1 TTL\ VY
Pi = mxvi 9L,<PL
PI = Pi Pi
Laboratory Frame
The situation in the two cases is as depicted in figure 18.1. In the centre of
mass frame a "fictitious" particle with the reduced mass
m\ + m2 M
is incident on a fixed centre of force. The centre of mass moves with a constant
speed
m i
1/
V (18.6.46)
~ T7V1 •
M
We still need the relation between 61, tpL and 9, (p. This is obtained from the
relation between the velocity v[ of the scattered particle in the laboratory frame
where it has components
7^ = i>i(sin 9L cos ipL, sin 9 i s i n ^ L , cos 9 i ) (18.6.47)
and the centre of mass velocity V and the velocity v1 of the scattered particle
in the centre of mass frame. The speed v' of the scattered particle in the centre
of mass frame is given by
The components of the velocity of the scattered particle in the centre of mass
frame are therefore
FTL 2
m2
•a • • a s i•n
Sln
ui sm 0 l sin ifiL — V
m
2 „ mi
üicos9L = —1>1 cos9 +—VI . (18.6.51)
M M '
Thus, we have the desired relation
sin9
tan 9l =
MI/M,2 + cos 9
<PL = <P • (18.6.52)
The same number of scattered particles must be observed in the centre of mass
frame. Therefore, we have
d a i d<7
^ ^ sin 9Ld9Ld<pL = ^' ^ sin9d9dp> (18.6.54)
aí]¿ DIÍ,
Hence,
dcriifii, ¡pi) da(9, ip) sin# d9 dip
(18.6.55)
dQi dQ sin 6i d9i dpi
Now, from (18.6.52) we get
|Tii/m2 + cos 9\
18.7. PROPAGATOR FOR A FREE PARTICLE 385
= {(*!">}
Solution
a) Formally the differential equation (18.7.59) with its accompanying initial
condition is solved by
n,n'
n,n'
= (18.7.63)
n
i-OO r1
(2ttm\3/2 J\ |£ - y\2 \
exp m
{-m) r ^r|
Substituting this back into the expression for the propagator we get the desired
result
0) = »p{i'» ! £
2sr} (18.7.70)
Solution
The result of problem 18.7 shows t h a t the retarded propagator is given by
2n roo 2
H„(x) = —f=
71
/ (x + ip)n e p
dp (18.8.74)
V " j-oo
we get
G+(x,y,t, 0) = - ! , ( , ) z
H — ' V 7T J 7T71!
oo
/
(ax + ip)na 2e~p 2 dp(ay
3 + iq)n e~q dq e~a ^ +y
' e~'(n+i/2)wt
-OO
e - (® +y ) E-IWT/2
h 7T^/2
dpdq
/ oo 1
— [2(aa; + ip)(ay + ¿9) e~'wt]n e~( p +q
n
-00
1 „ a - 2
— 0(f) r-<x (*2+y2) --iut/2
~ fa *•' jj-3/2 e e
/ OO
exp {2(ax + ip){ay + iq) e~'ut — p2 — q2} dpdq . (18.8.75)
•OO integration variables from p, q to r, a where
If we now change
p = [ T + ] = [ T 1
^ " » ^ - '
(18.8.76)
the argument of the exponential in the integral becomes, after a little algebra,
-icot
i a(x + y) e
- {l + e~iwt) iu
V2 1 + e~
i a(x — y) e~
- (1 — e~iwt) O• +
V^2 1 - e~iut
g — ildt
I _ e-2Út [ q 2 ( ® 2 + y 2 ) e _ , w t - 2a2xy] . (18.8.77)
388 CHAPTER 18. SCATTERING - TIME DEPENDENT
This now leads to a pair of standard Gaussian integrals. Thus, we can combine
these results to get
G+(x,y]t, 0)
—z x OL A' / 2 ,,1+e-^" 4xye
= -T 2 iuit
-iiot/2
2 g —2iujt
(18.8.78)
G*(x,y,t.0) = i S L u t
Bibliography
[17.1] For a totally different way of computing propagators see
E.P. Feynman and A.R.Hibbs, Quantum Mechanics and Path Integrals,
New York: McGraw-Hill, (1965).
r
C h a p t e r 19
Scattering - Time
Independent
X U 2
^ ( X ) = J ^ £ ^ E ( U + L ) ' D U
a
l roo e'
to show that
1
a) Zi-i(x) + Zi+i(x) = Zi(x) for / > 1
x
C) ^ [x_iZ;(i;)] = -x~'zt+1(x) .
Here Zi(x) may be any one of the four spherical Bessel functions.
Solution
We demonstrate these formulae explicitly for ji(x). The calculations for the
other two Bessel functions is exactly the same.
390 CHAPTER 19. SCATTERING - TIME INDEPENDENT
a)
jl = (
"2 + 1 }
' d u
• (19 L2
)
Now,
So,
=
1 . . .
Ly xi+1
x '+l
+
1
1 /•'
(1
+ ")<'" +
i 2 t + 1 (/ + 1 ) ! 2f / , < " +
1) «"»•'«+a á - . w
1 a;'+ 2 1 í' r
= -2iJ' M +
. 2 H . ( / + l ) ! 2 i /_.(«I + l ) ' + ' « ' « < ¡ » + 2 í A - , W
= + + • (19.1.5)
Thus finally,
2/4-1 T
— y - i i i x ) = 2y [ Á + i ( * ) + ( 1 9 . 1 . 6 )
and
21 + 1
— - — j ¡ ( z ) = j¡+i(z)+.?'/-i(x) • (19.1.7)
Then,
2/ + 1 . . . (2/ + 1 W - 1 r Tn , , w ,
d
—*<«> = i2'W[! J ' <-+'> "
(2Z+ l ) z —2 /•• 2
J (u + l)ldexu . (19.1.9)
¿ 2 ,+1 Z!
19.1. EQUATIONS FOR SPHERICAL BESSEL FUNCTIONS ayi
So,
and
dji(x) _ Ix1-1 r
'
l+1 - f e ™ (u 2 + 1)' du
dx i2 l\
J
+
T¥*1i/.«"V+ !)'«<<»
= (19.1.12)
X
i l + 1
jl+i(x) = ji(x) - j , _ i ( x ) . (19.1.13)
X
Then,
dji(x) . , . , f l 2/+l\
= j , + 1 ( * ) + ( ~ ^ j j , ( * )
= (19.1.14)
d + i )
^'+i
= T ¥ ^ . r h L ' " ^ '
—a;
- Í exu (u2 + 1)' + 1 ¿ií
i 2<+2 ( / + i )) !• J — i
= -x~l ji+1{x) . (19.1.16)
As stated at the beginning, the calculations for hj 1 ^ and hj ' follow exactly
the same lines.
392 CHAPTER 19. SCATTERING - TIME INDEPENDENT
«(*>=*' i r l ú ) ' z
"
dx \ x '
Solution
To get the desired expression we start with the results of problem 19.1 a)
dZi(x) 21+1 I+ 1
- f a - = —¿-Zi(x)-Z,+1(x)-—Z,(x)
= -Zl(x)-Zl+1(x) . (19.2.18)
X
d Z
Z,(x) = — Zi-i(g)- '-l(x) . (19.2.19)
x dx
Therefore,
19.3
WRONSKIAN FOR SPHERICAL
W = ji{x)n[(x) - 7i;(x)j,'(:c)
™ = -2-W.
dx x
Solve this equation and use the behaviour of ji(x), ni(x) for small x to fix the
constant of integration to get
w = i .
Hint: Start with the differential equations for ji(x) and n¡(x) .
Solution
We start with the differential equation for both and n/.
d~ 2 d_ /(/+!)
j,(x) = 0
dx2 x dx x¿
1(1 + 1)
d~ 2 d_ : _
ni(x) = 0 . (19.3.22)
dx2 x dx
Now we multiply the first by n¡(x) and the second by ji(x) and subtract to get
d
-^ = -2-w.
(19.3.23)
dx xx
Therefore,
(19.3.24)
x- '
We can evaluate the constant C for any value of x. So we choose x near 0
and use the asymptotic form of the spherical Bessel functions as well as the
definition of W. Then we find C = 1.
V(r) = f
JA
M
DN<R(N)
e~"r
R
(19.4.25)
394 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Solution
In Born approximation the solution of the Schrodinger equation has the asymp-
totic form
9m 1 p'kr r
— J e^V{r')d\' (19.4.26)
where
q = it-It' (19.4.27)
e í (
/(m) = -f?¿/ "~£' ) r
"v(r')dv
wL •
2m f°° a(n)
Here, 1 9 4 2 8
< ' >
2 2 2
q = 4k sin (0/2) . (19.4.29)
v(r).= v 0 e - ^ 2 .
Solution
a) We have
V
V{r) = V 0 e . (19.5.30)
where
g = 2fcsin(0/2) . (19.5.32)
19.6. BORN APPROXIMATION FOR SQUARE WELL 395
'
Therefore, we get
(19.5.34)
00
1
/ ( M ) = i X](2/ + 1)e'á' siná
' p '( c o s é l ) • (19.5.35)
;=o
1 mV
HO0 • X °
—e sm óo = — 2 exp ( sin 2 (0/2) ) sin 9 d9 . (19.5.37)
2fi p
Therefore we get
1
iS0 • r 2mV 0 [ ¥ / fe2/mu>
(19.5.38)
2mVn [ k
ío = ¿ l n 1- 2 2 - (i_e-
fc2 mu
/ ) (19.5.39)
HK V ^
Solution
This time we have
in \ f ~V° r < a
V
^ = 1 0 r > a (19.6.40)
f
2mVo
r sin qrdr
qh2
2mVo<z sin qa
cos qa (19.6.42)
Jo qa
a) Hence,
da 2 mV 0 2 a 2
|JV cos qa — sin qa
n (19.6.43)
dQ ' qa q4h4
b) Proceeding as for part b) in question 19.6 above we find
Í
2mVo a cos(2ak sin 9/2) sin(2a&sin 9/2)
sin I (19.6.45)
h2 (2k sin 9/2Y- (2A: sin 9/2)3
Thus, letting
x = 2ak Jo
sin 9/2 (19.6.46)
we can rewrite the integral as
.3 /-2 ka
piSo0 sin jq = — 2mVoa
i e' cos x sin x dx (19.6.47)
u
k h2 Jo x x-
The integrand is a perfect differential
cos x sin x d (sin x
2 (19.6.48)
x x dx
So, we immediately get
1 3 /
id0 • c 2mV0a ^ sin 2ka
r ° =
W~ (19.6.49)
2 ka
Then finally, using the fact t h a t the phase shift is a small angle, we find
nd
. 2mVoa2(ka) ( sin2&a
19.7. PHASE SHIFTS FOR DELTA-FUNCTION POTENTIAL 397
Solution
With the potential
Bi ji(kr) r<a
Ri(r) = (19.7.52)
Ai[cosS¡ji(kr) — sinn¡(A:r)] r > a
and
h2 dRi dRi
= — Vqü R¡(a) . (19.7.54)
2m dr dr
a+0 a —0_
Writing out this last equation and substituting for B¡ from the first equation
we find with
2mVoa~
Q (19.7.55)
h2(ka)
that
Another way to solve this problem is to use the partial wave Lippmann-
Schwinger equations
where
Since we are interested in the asymptotic region we have r > a and with
we find that
Putting r = a we get
that
Thus,
and
, ika2mY§ha ji(ka)i l
(2lV + l)ji(ka)
i'(2l + 1 )ie's' sin J, = V
^ — - . 19.7.68)
1 + ika2m^a ji(ka)h¡ (ka)
From this it follows t h a t
e 2 it,_1 ka2mVn°ljl(ka) 2
(19.7.69)
2i
1 + ika2m^r2 ji(ka)h{l1\ka)
19.8. PHASE SHIFTS FOR YUKAWA POTENTIAL ¿y»
If we now write
e 2IS, _ j 2
t.ná, = x (19.7.71)
we see that the result coincides with our previous result (19.7.58).
Solution
For a high energy particle, the Born approximation should be valid. Further-
more, for the Yukawa potential
e-ar
V{r) = V0- (19.8.72)
r
we have the scattering amplitude (in Born approximation) as given, for instance,
in [19.2],
where
(19.8.76)
400 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Now projecting out the I = 0 partial wave (by integrating with Po(cos0) = 1)
we get
2mVo i n si d0
—2 iS0(k) =
2 2 2
h J 0 4k sm (0/2) + a2
2mV 0 2sin(fl/2)cos(fl/2)d0
JQ 4k 2 _:„2
A 1.1
sin (0/2) + a 2
2 mVo 4k2'
In 1 + (19.8.77)
' h2k2
Hence we find
2mVb 4/fc2
S0(k) = In 1 + (19.8.78)
~H^K~
Similarly for the 1 = 1 partial wave we get by integrating with Pi (cos 0) = cos f
I sin 0 cos 0 d0
2« , ( * ) =
2k o 4k2 sin2(0/2) + a2
2mV 0 f
I
2MV0 [* 2sin(0/2)cos(0/2)[l - 2sin 2 (0/2)]i
a 2
4k2 sin" (0/2) 4- a 2
, 4/fc2
+ ( i ) , n 2~ (19.8.79)
w r - f q-
Therefore,
, 4arl
}• (19.8.80)
a¿
fo(k) = - Co (19.9.81)
r
+ icok — 1/2 cokr"o
as well as in the form
1
fo(k) = (19.9.82)
k cot ó — ik
where c 0 is the scattering length and ro is the effective range [19.3]. Also verify
that both versions of the amplitude satisfy the optical theorem.
Solution
In terms of the phase shift the scattering amplitude is
AW = ¿(1 (19.9.83)
19.9. LOW ENERGY S-WAVE AMPLITUDE 401
fo(k) = ¿ ( l-e2iá)
e's
= —— sin S
k
sin <5 sin <5
'lS k cos S — ik sin
sinSó
1
(19.9.84)
k cot S — ik
On the other hand for low energies the phase shift is given in terms of the
scattering length c 0 and effective range r 0 by
(19 9 88)
ywo(*)] = x n r • - -
a
=J
On the other hand the total cross-section is given by
da
dQ
dQ
\f(k)\2dQ = 4n\f(k)\2
- I
4tt . 2
sin" S . (19.9.89)
P
=1
a = d n
/ DN
\F(K)\2DN = 4*\F(K)F
AttcI
(19.9.91)
(1 - l/2c 0 A:rg) 2 + cgfc2
Thus, in this case we have also verified the optical theorem.
Solution
The general scattering solution for the Ith partial wave can be written
Í
ji(kr) r < a
Solution
We have t h a t near x = 0
sin x , , cos x 1 , „ „ ,
= 1 and = . 19.11.107
X X X
Thus, they have the correct asymptotic behaviour to be jo(x) and no(x) respec-
tively. Furthermore we find
and
D2 s m i
s.nx_2cos£ »m£ (19.11.109)
+ 2
2
dx V x
Therefore,
0
•dx¿£ + —
x ax+ > )J (V —x 1 = ' (1911110>
Similarly
and
= (19.11.112)
dx¿ V x ' x x1 xó
Therefore,
( £z + l ± + l ) ( ™ ) = 0. (19.11.113)
\dx x dx J \ x /
w r ) = / -^o r<a
k
' \ 0 r> a
find the effective range and the scattering length for the s-wave (1 = 0).
19.13. EFFECTIVE RANGE, ETC.: YUKAWA POTENTIAL 405
Solution
As a first step we need to find the s-wave phase shift. To a good approximation
this is given by equation [19.2]
f
2mVo\ sin2(A:r)
sin ¿o r'dr
h2 J (kr)2
mvo 1
ka sin 2ka (19.12.114)
'WK*2 2
Jo
For k —» 0 this yields
l k "T^ 3
2 mvoa 2
sin ¿o
H2
1 1
= —A- ka — -(ka)3 (19.12.115)
5
2mVoa2
= (19.12.116)
° h2 '
The scattering length and effective range are defined for k —• 0 by
1 1 (19.12.117)
k cot do ' + ^2RO
Co 2
Using our result for sin ¿o we find
V =
406 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Solution
Again we start with the approximate equation for sin Sq
3ín¿o «
r
I ¡ dr -
2mV\
2
e /ir
sin2(kr)
(19.13.121)
Jo h J fir (Arr)2
2
In 1 +
H FIK
2 mV k_ _ k3
h2 fi2
k2
1 - ""a
K
-P - (19.13.122)
¡i 2 j
a*
where again we have introduced a dimensionless parameter
2 mV
(19.13.123)
So,
P
c0 - r 0 = (/? - 1//?) (19.13.125)
H FT
Solution
From the previous two problems we get:
For the Yukawa potential
P
Co = - r 0 = (/? - 1//3) (19.14.126)
r
» 19.15. PHASE SHIFTS FOR HARD SPHERE 407
where
c0 = i a a , r 0 = ( a / 3 - 6/(5a)) a (19.14.128)
o
where
2 mvoa 2
a - (19.14.129)
h2
Equating the two scattering lengths we find
2 mV 1 2 mVoa2
a . (19.14.130)
ftV 3 h2
So that
V =^(fia)3V0 . (19.14.131)
P=\a(na). (19.14.132)
( 1 9 1 4 1 3 3 )
< < " • » •
fta = y j | . (19.14.134)
So, finally
1 /ci\3/'2
V =
Z \ 2 j ^ ' (19.14.135)
k oo .
Solution
If we now define
^2=2m_e (19.15.137)
n¿
d2 equation
the radial 2 d becomes
1(1 + 1) 2
2 Ri(r) = 0 r>a. (19.15.138)
d r ^ r dr r2
The solutions of this equations satisfying the condition that Ri(r) corresponds
to a fixed incoming flux for large values of r is
Thus,
4/ A
9/2
or
So,
a result that is quite obvious since that is the cross-sectional area that intercepts
the incoming beam.
The result for low energy a = 47ra2 (19.15.145) is clearly 4 times the classical
cross-section of na2. The reason for this is that for k —• 0 the wavelength A of
the particle gets infinitely large and the particle scatters off the whole surface
410 CHAPTER 19. SCATTERING - TIME INDEPENDENT
of area 47ra2 of the sphere rather than just off the cross-sectional area na2 per-
pendicular to the incident beam.
Therefore,
(T « ^ Í (21 + 1) sin 2 S¡ dl
* Jo
-J J0
_8_ r
k2 J-ka
(21 -(- 1) sin 2 (&a — lx/2) dl
ka(it/2— 1)
Ka{
— (x + ka) + 1 sin 2 x dx
— •
- 27ra (19.15.153)
In this case, since we are approaching a classical limit, we would expect the
cross-section to be just the classical cross-section of na2 and not 27ra2 that we
found. The factor of 2 is due to the way the quantum mechanical cross-section is
defined in terms of the scattered wave. Thus, one defines the total wavefunction
as
ip = 0
Figure 19.2: Shadow scattering off a hard sphere by high energy particles: The
vertical lines depict wavefronts.
19.16. RESONANCE FOR SQ UARE WELL 411
In the case of the hard sphere we have for Ar —> o o that in the shadow of the
sphere the wavefunction vanishes identically (See figure 19.2). This means that,
in this region, one defines
and so one obtains a cross-section that is twice the classical cross-section. This
effect is known as shadow scattering.
Solution
The s-wave phase shift is given by
l t u 6 1 5 6
< >
k jo(ha)n'0(ka) — A j'0(Ka)no(ha)
Using the explicit form in terms of trigonometric functions for j0 and no we find
that
. k tan(A'a) - Atan(fca)
tan¿0 = „ , ,f ,, w /„ ' • (19.16.157)
A + Artan(A:a) t a n ( A a )
This can be rewritten as
1 (19.16.158)
tan So - tan tan •( j7 tan(A'a) ) — ka
Ka = — . (19.16.159)
2
Writing out the definition of K we get
(Ka)2 = — = 2 m
^ 2
+ V
°^ • (19.16.160)
e o = £ ¿ - v 0 . (19.16.161)
This is the energy for a particle in a box whose zero of energy is at — Vo-
412 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Solution
Using the assumptions stated in the problem we may take the scattering poten-
tial to be
Vo¿(z) for \(d - a) < |z| < ¡¿(d + a) , \y\ <
V (19.17.162)
0 otherwise
Notice that Vo has the dimensions of energy xlength. Also, here we have chosen
the slits to lie in the x — y plane and thus we choose the incident beam along
the z-axis. Since the scattering is weak we may use the Born approximation.
Defining
2 mV(r)
U(Ñ = (19.17.163)
h2
we find that
Ü(q) = J eiqr~U(r) d3
(d+a)/2 rh/2
dz ¿(z)e iqr
UO -(d-a)/2 Ma+a)/¿ t-n/¿ /-c
/ (d+a)/2 (d-a)/2 -h/2
dx+ dx dy
sin•(d+a)/2
(qyh/2) J(d—a)/2 J—h/2 J—
WO [sin(g x (d + a)/2) - s m ( q x ( d - a)/2)]
QXLY
sm(qyh/2)
8t/ 0 s i n ( g r a / 2 ) cos(qxd/2) (19.17.164)
QxQy
Here
q = k — k' . (19.17.165)
We also have that
£ = (0, 0, k) , k' = (k sin 6, 0, k cos 9) (19.17.166)
so that
q = (k sin 0, 0, fc(l — cos 9)) = (¿sin 9, 0, 2k sin 2 (9/2)) . (19.17.167)
In this case we find that the differential cross-section is given by
2
da 1 t2sin (qyh/2) sin2(qxa/2)
r64 U ; \qxd/2)
DQ 167T- Q^
íHy Qx
sin(«/2) < ¿ •
Solution
The scattering amplitude is given in Born approximation by
(19.18.169)
where
q = 2k\ sin 9/2 \ .
The differential cross-section is then
• (19.18.170)
J dCl
n
rrr f ~
= 2n-—/ 2 4 \V(q)\2 s\i\ 9 d9
47T /i Jo
-.2 /-t
[ \V(2ksm9/2)\72sm9/2cos9/2d9 . (19.18.171)
2tt h 4 Jo
Therefore, letting q = 2ksm9/2 we find that
2
m '[•¿K
2k
a = ( 1 9 1 8 I 7 2 )
Now, we are told t h a t the potential V has a finite range a. This means that
the Fourier transform V(q) is appreciably different from zero only for q < 1 /a.
But, from the definition of q this means that
2A*sin 9/2 < l/a . (19.18.173)
Rewriting this we have t h a t for appreciable scattering to occur we need
sin 9/2 < — . (19.18.174)
2 ka
v u / i r i jC/A I y. SUA'1 "1ERING - TIME INDEPENDENT
(r|V|r") = V { f y ) .
If
V ( f y ) = V(r)0(f-f)
(f\V\ip) = J d3r'{r\V\ff){r'\%l>)
Solve the scattering problem for a separable potential for an incoming plane
wave
^(0) =
(2^ e ¿ f c
" " (19.19.177)
Solution
The Schrodinger equation for this case is
h2 „ r
Introducing
a 2 = 2 ^ = (19.19.179)
we get
where C is a constant defined by the integral. To solve this equation for incoming
boundary conditions we require the incoming Green's function given by
G(+)(r-r-) = _ _ _ _ . (19.19.182)
c = { k )
W f T ^ ' - i / • (19.19.184)
we finally get
(2-K)3/2g(k)
C =
I+ A (19.19.186)
where
Ei - 2(m¿ — m 0 )c 2 .
The outgoing particles, of type i, interact via a potential V¿¿. The coupling
between different types of particles (channels) is due to potentials
Vij = Vj i .
where
T, = | ¿ (19.20.190)
for the wavefunction l^o) by eliminating the wavefunction \-tpi) - The resultant
"potential" V in this equation is known as a generalized optical potential. Note
that this potential is complex and allows for absorption or emission of particles
of type i = 0.
Solution
The Schrodinger equation for this case is
—
(To + Voo — E + Eo)\ipo) = Voil^i) (19.20.191)
{Ti + Vn - E + E i W Ú = -Violto) • (19.20.192)
\ipi) =—(Ti + E —E i i e V \ i ) 1
Violto) • (19.20.193)
V f c l m M . p ) = (Lx - iLyy-m(Px + i P y ) l
^ j ^ .
19.21. FREE PARTICLE EIGENFUNCI l(JI\b lit
Solution
For a free particle the states of definite angular momentum satisfy
h2k2
2M
L2Í>klm = 1(1 + 1 )ti21pklm
Lztpkim = rnhipkim . (19.21.196)
Here,
H
=%M • (19.21.197)
ipkoo = kr • (19.21.198)
We now introduce
P+ = Px + ipy
L± = Lx±iLy. (19.21.199)
and act on it from the left with p+ and use the results of (19.21.201) we get
But,
L+1>ku = 0 . (19.21.204)
Thus, we find
Therefore,
Thus, the effect of p+ is to raise both I and m in ipkim by one unit. On the
other hand we know t h a t L_ lowers the value of m in ipkim by one unit, but
leaves the value oil unchanged. So,
^ = fco. (19.21.210)
i = m = 0 .
so that, up to normalization,
L l I m ip l ku — ipkim • (19.21.213)
V(r) = V , V o > 0 .
19.23. NEUTRON-PROTON SCATTERING: SPIN FLIP 419
Solution
The Schródinger equation for the radial function reads
d2 2d 1(1+1) b2 2
R i ( r ) = 0 . (19.22.214)
dr2 r dr r2 r2
Here we have introduced the parameters
,, 2mE , ,9 2mVoa2
K A N D B
= - J ¿ T ~ = — • (19.22.215)
The solution finite at the origin is j\(kr) where A is the positive root of
so that,
2
A = + y/(l + 1/2) + b2 . (19.22.217)
^ s m ( t r - / , / 2 + ¿,) (19.22.219)
kr
Therefore,
Solution
The spin wavefunction of the neutron proton system before the collision is
I t , " ) l i,P) •
This may be rewritten in terms of the triplet and singlet spin wavefunctions as
1
|t,")U,P> = ~ A \ t , " ) l I.P> + I -I-.")I t , p ) )
+ ^ ( 1 t , n ) U , P ) - I 4-.")I t , P ) ) (19.23.221)
420 CHAPTER 19. SCATTERING - TIME INDEPENDENT
1
^ = ( 1 t,rc)l i,p) + 14-,«)l t , p ) )
V2
fi
1(/3-/I)/212 = I/3-/1I2
2 2 (19.23.223)
|(/ 3 + /i)/2| + | ( / 3 - / I ) / 2 | 2(|/ 3 | 2 + |/ 1 | 2 ) •
h2 A2 Ae~Xx
V(x) = -
m (Ae"Al + i f
h2 d2<j)
2 + [V{x) - E]<t> = 0
2m dx
where
and
h2k2
E
2m
b) Use these solutions to show t h a t there is no reflection from this potential.
c) Find the energy of the bound states, if any.
Solution
a) By direct differentiation we find t h a t
C
K ^ A + l
+ 2/c ] —
AA
(A+l)2
AA
ka
( ?ÍT-2'<) +
+ D = 0 . (19.24.230)
(A+l)2.
422 CHAPTER 19. SCATTERING - TIME INDEPENDENT
For a nontrivial solution the determinant of coefficients must vanish. This yields
1 2A,4
-4 k + A + = 0 . (19.24.231)
^A + LJ ' ( a + 1) 2
The energy is now given by
HHc2 K2 [ A 2 Q 4 - 1 ) 2 + 2\Á\
E= — (19.24.232)
2m 8 m(A + l)2
Solution
a) There are two different expressions for the scattering amplitude as a matrix.
1) If we write it for states of total spin namely the states
( i 0 0 0 \
0 1 0 1
(l/2,mi;l/2,m2|5M) = 72 72 (19.25.234)
0 i 0 1
"v2
K 0 0 1 0
19.26. PHASE SHIFT, SCATTERING LENGTH, ETC. 423
.RRIJ
b) The differential cross sections for the reactions listed may now be written
down by inspection from equations (19.25.234) and (19.25.235).
2) n ( + ) + p ( - ) - » n ( + ) + p ( - ) : — = -|at-t-as|2
3) n ( + ) + p ( — ) - » • n ( — ) + p ( + ) : — = -|at-as|2
c) For unpolarized neutrons and protons the differential cross section is obtained
by averaging (19.25.233) over the initial spin states and summing over the final
spin states. So, using the orthogonality of the Clebsch-Gordon coefficients we
get
%=\ £
= ^ [3|a(| 2 + |a s | 2 ] • (19.25.238)
Solution
a) In the asymptotic region r —> oo the wavefunction must be of the form
u(r) Asin(kr + J) . (19.26.239)
Thus, we take the asymptotic form of the given wavefunction. To do this we
use that
Then,
b2 — a2 kcos(kr) — bsin(kr)
u(r) Si {kr) +
" b2 + k2 b+ a
• ? ^ , o V ( k 2 + ab) 2 + fc2(¿ - a) 2 x
6*= +
k2 + ab
s'm(kr)
v / ( F + a6) 2 + P ( 6 - a ) 2
k(b — a)
+ 2
cos(Arr) (19.26.241)
\J(k + a6) 2 + &2(6 — a) 2
This is of the desired form (19.26.239) if we define the phase shift J by
k2 + ab
cos S :
2
V(k + afc)2 + k2(b — a) 2
&(6 — a)
sin <5 = (19.26.242)
i / (A; + ab)2 + k2(b — a) 2
2
b) To get the scattering length cq and the effective range r*o we use the result
that for k —>• 0
1 1,9
k cot o —y 1—k ro . (19.26.243)
c0 2
But,
(k2 + ab)
k cot S = (19.26.244)
b— a
Thus, we find that
1 1 9 k2 ab
-~ + -k2r 0 = (19.26.245)
cq 2 b— a
So,
6 —a _ 1 1
co (19.26.246)
ab b a
and
i-o (19.26.247)
19.27. SCATTERING OFF A DIATOMIC MOLECULE 425
Solution
With reference to figure 19.3 we see t h a t the electron arrives at atom A earlier
\
f
> r\
( «>íY
)
\,A_y
than at atom B. So, its phase at atom A relative to the centre of the molecule
is
eikR/2
e-¿fc fi/2
But,
where hk' is the momentum of the scattered electron. We define the "momentum
transfer"
q = k-k' (19.27.250)
and in the denominator of (19.27.248) we replace TA and RG by r. Then,
So the scattering amplitude for the molecule in this particular orientation (fixed
value of R) is
£iqR/2 _|_ e-iq R/2
F(M) = M
= 4 | / ( f ) | 2 c o s 2 ( g ^ R/2) . (19.27.253)
However, in any given experiment, the direction of the vector R is random. So,
writing
q-RzzqRcosa (19.27.254)
we must average over all values of a . The result is
j i r
— = 4|/(0)| 2 — / cos2(-qRcos a) da . (19.27.255)
as I it Jq 1
If we now change integration variables from a to
x = -qRcos a (19.27.256)
we get
L
da 4|/(0)|2 fqR/2 cos 2 x dx
DCL
qR/2 2
2
= 2\f(9)\ [1 + J0(qR)] . (19.27.257)
Solution
If we replace the radial function R¡ (r) by
u¡(r) = rR¡(r) (19.28.259)
the Schrodinger equation for 1 = 0 becomes
dru 2. .
— + g (r)u = 0 (19.28.260)
where
R0(r)
= -u(r)
r
to compare it with the general solution given by (19.28.258). This asymptotic
behaviour is
A
Ro(r) — sin [ q(r')dr'
kr Jo
A .
— sin [ {q(r') — k} dr'+ kr (19.28.264)
kr .Jo
Comparing this with (19.28.258) and taking the limit r —> oo we get
r OO
¿o =
= I [<7(r) — k] dr
Jo
poo
/ [y/k2 — U(r) — k] dr (19.28.265)
Jo
Solution
From problem 19.28 we get that the s-wave phase shift is given by
k2_ 2mVAr)_k
-/
Jo
h2
dr (19.29.267)
2 mV(r)
z — \lk2 — (19.29.268)
ft2 '
Therefore,
2mV 0 1
22 = k 2 + (19.29.269)
h2 ear - 1
If we call
p l22 = 2 mvo
(19.29.270)
/i 2
we get that
/? 2
ear = 1 + (19.29.271)
z -fc2
2
'
So, we find that
2 /?2 r dz
dr = —- (19.29.272)
a (z — k )(z — k2 + /?2)
2 2 2
Hence,
= Ml [ a
2 /*oo
JK
(z + fc)(2 2
: dz
— k2 + (32)
(19.29.273)
^ In ^ ^ arctan \ / P 2 — k2 f3 > k
8 = (19.29.274)
2k , aA 2 -/? 2 In k-y/k*-P 0< k
f lnf + K + ^/K*-0\
rR(r) = u(r)
becomes
d2u Lo . 1(1 + 1)"
+ kz - U(r) - u = 0 .
dr2
19.30. WKB APPROXIMATION FOR PHASE SHIFTS 429
Here,
U r
( )= and =
'
kr = ex
^ + Q2{x) v = 0
where
Here, the variable x runs over the region —oo < x < oo. Assume that V(r) goes
to zero rapidly at r —> oo and is no more singular than 1 / r near r = 0. Also
assume t h a t there is only one point xo such t h a t
Q J (xo) = 0 .
Clearly, for x —> large positive, Q 2 ( x ) > 0 and for x —> large negative, Q2(x) < 0.
Thus, such a turning point always exists.
Use the WKB approximation to solve the equation for v(x) and then convert
back to the variables r and u(r) and look at the asymptotic region r —^ oo to
obtain the phase shifts.
Solution
For the equation for v(x) the W K B solution in the asymptotically forbidden
region x < XQ must be exponentially damped (since u(r) vanishes at r = 0) and
is
A
i)(a;) : exp - \Q(x')\dx' (19.30.275)
V\Q(x)\
v(x) = l=[cexp-i{£<3(*v*'-i
Vol
¡{/
D exp (19.30.276)
430 CHAPTER 19. SCATTERING - TIME INDEPENDENT
The appropriate matching condition is the one that joins the exponentially
damped solution onto the corresponding solution in the classically allowed re-
gion. The solution in the classically allowed region corresponding to the solution
(19.30.275) is
2A í r 7t i
°{') =
7 m C ° S
\ L Q {
' ' ) d
" ' ~ * l • (19.30.2T7)
We also define
kr0 = ex° . (19.30.281)
Then,
I z
u(r) - 2A\
k — U (r) — (/ + l / 2 ) 2 / r 2
2
u (r) = 2A\
k2 - U(r) - (/ + l / 2 ) 2 / r 2
<5, = (/ + 1 / 2 ) | - kr0
dr1 . (19.30.285)
+
L : V k2 - U(r') - (/ + 1 / 2 ) 2 / r ' 2 - k
Although this equation is quite different from the one obtained earlier (equation
(19.28.265)) they yield similar results.
19.31. ZERO-RANGE POTENTIAL 431
lim
1 d
, n = a constant independent of energy . (19.31.286)
j—>-0
h2a2
e (19 3i 287
' = - 1 í t • ' - >
a) Find the bound state wavefunction ipo(r) as well as the scattering solutions
IK(R) •
d3r \ipo(r)\2 = 1
/'
/' D3RIP0[R)IPJ;{R) = 0
/ ' d rip~{f)^{f)
3
= S(k — q) . (19.31.288)
/•d k e
3 ikr
Solution
a) If we set
u = rip (19.31.291)
then the function u, for the bound state, satisfies
d2u 2mEo i
d^ = —vr- u = a u
• (19.31.292)
u = Ae-ar . (19.31.293)
(19.31.295)
A =
The scattering solution has to satisfy the free Schrodinger equation and must
have only an / = 0 outgoing spherical wave. Therefore, it must be of the form
pikr
1
v» s (f) 3 2
e ,fc f + 7f - (19.31.296)
(2tt) / r
We now impose the boundary condition (19.31.286) and get
lim
r-+ 0 [4¿(r,As) (27T) 3 / 2
+ ikf (19.31.297)
Therefore,
1
/ = - (a + ik) (19.31.298)
(2tt)3/2
and
Akr
eik f _ 1
^(f) = 3 2 (19.31.299)
(2tr) / a + ik r
b) The normalization of the bound state solutions has already been demon-
strated so we next consider the orthogonality between the bound state and
scattering solutions.
/ d3rrp0{f]ip^(f)
Q
I J3
°~
:
aT
-e -
e»*-r _ 1 0 — ar-\-ikr
(19.31.300)
(2tt) /2 y -—
3
2TV / d r a + ik
1 y.ó 1. zUIÍU-RANGE POTENI ¡AL
To carry out the integrals we use spherical coordinates and line up the z-axis
with the vector k. In that case the integration over ip yields simply a factor 2n
and the integration over 9, after setting cos 9 = t, becomes
Therefore,
J d3r rp0(r)ip¿(r)
VA
27T ik a — ik
VE/I a — ik a + ik A2 + K2 }
27T \ ik
= 0 (19.31.302)
as required. Next, we consider the (delta-function) normalization of the scat-
tering solutions
(r)
J d3rrpt(f)ip?i
d 3 r -t'fcr gtq f
(27 r
_ j _ [fle^e-r k r + _i !1 _ [Í d—e*q
a — ik J r a + iq a —— ikik J,J r2r2 J
where
o- r°°
h = r—————-7- I dr e~ikr (e,qr — e~lqr)
iq(a - ik) J0
2?r
- f dr (e~'(k~q',r — e~i<-k+q)r\ (19.31.304)
q(k + ia) Jo ^ >
H = (19.31.305)
• U R M L
and
H =
47T
(k + ia)(q — ia-f
¡O) JO
dre —i(k — q)r (19.31.306)
Hence we get
^1+^2 + I3
27r (fc-g)r _ i(k+q)r
(k + ia)(q — ia) )
-sr2r*('- 2 ^ 0 ° d r e - i ( f c - 9 ) r j . (19.31.307)
(k-q)r _ —i(fc+9): •j +
434 CHAPTER 19. SCATTERING - TIME INDEPENDENT
= ia(rl)CJre~"k~')r
+ Í5L R DRE^Y - R DRE-^+I)r
kr O OJo rOO
q Jo
+ / DRE^K+^R + / DRE-^+I^
Jo Jo
1 1\ —i ia i ia —i
q kJ k — q — ie k k + q + ie q k + q — ie
i —i
+ +
k + q + ie k + q — ie
a a
« 0,1
___+2,i(i+?)
0 (19.31.308)
since both k > 0 and q > 0. Thus, we have shown t h a t
J d3k
H + H
"J
d3k ikr „ — ik FF _ 1 / D3K „ — ikr' ikr
e e
a + ik r' J a — ik e e
oo g»fc(r-r') _ g i k ( r + r ' ) eik(r-r') _ g-ik(r+r')
k dk
fc - zq k + ia
2tt r ik(r+r')
k2 2tt R e
J_c
rr' Jo Jk(r-r')
dk k dk-
k2 + a 2 +
rr' k — ia
4tr f c 2
K Jk(r-r')
rr' Jo dk 2
K + A2
I
4tt
rr' Jo
r CO ik(r+r')
ik( r+r
+ dk e - "> - ia dk-
k — ia
43(5 CHAPTER 19. SCATTERING - TIME INDEPENDENT
Solution
a) To obtain the desired form we simply write sinfc(r — r') = sin kr cos kr' —
cos kr sin kr' in the integral equation to get
i R
= sinfcr f — / dr' cos kr' U (r')ip(r')
k Jo
1 fr 1
+ cos kr — / dr' sin kr' U(r')ip(r')
k Jo
= C(r) sin kr + S(r) cos kr . (19.32.319)
So,
1 F
C(r) = 1+ — dr cos kr' U(r')xj)(r')
* Jo
1 R
S(r) — — / dr sin kr' U(r')ip(r') . (19.32.320)
k Jo
b) For r —y oo we have the general result that
where A is a constant. Also for U(r) —> 0 sufficiently rapidly for r —> oo we
have that
Therefore,
f SYoo) .
tan óo = — — - = lim t(r) . (19.32.324)
G(oo) r~y oo
where
dt U_
sin kr + — cos kr
dr ' k o
U ;
= — —- [sinfer + t(r) cosfcr]' (19.32.328)
k
d) If |¿(r)| < < 1 we may may expand the right hand side of (19.32.328) and
drop the term involving t 2 to get
u
— fs — — [sin 2 kr + 2<(r) sin kr cos krl
d r k
= — y [sin 2 kr + t(r) sin 2kr\ . (19.32.329)
K
So,
dt
U • , U . 2,
——I—— sin ¿kr i = —— sin kr . (19.32.330)
dr k k
This is a linear equation with integration factor
U(R')
exp / dr sin 2kr' .
Jo
Therefore, we have t h a t
1 i [r U(r') ,1
tlr) = —— exp — / dr'—-—sin 2 kr
k I Jo *
Bibliography
[19.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 19.6, (2002) .
1Í8 <JH AFTER 19. SCATTERING - TIME INDEPENDENT
C h a p t e r 20
S y s t e m s of I d e n t i c a l
Particles
Solution
If we neglect the interaction between the electrons then each electron experiences
simply a central Coulomb potential and all of them have the same energy levels
given by
7 p1 1
En = J. (20.1.1)
a n¿
If we neglect the spin, all of these energy levels except n = 1 are degenerate.
Including spin, the degeneracy of the rath energy level is given by
n —1
(20.1.2)
1=0
Solution
a) If I = 0 the spatial wavefunction is even under an interchange of spatial
coordinates. Since spin 1/2 particles are fermions, this requires that the spinor
wave function be odd under an interchange of particles. The only possibility is
|«,m„ 1/2,1/2) = | 0 , 0 , l / 2 , l / 2 )
= ± = [|i/2, l / 2 ) | l / 2 , - 1 / 2 ) - | l / 2 , - l / 2 ) | l / 2 , 1 / 2 ) ] . (20.2.3)
11,1,1/2,1/2) = |l/2,l/2)|l/2,l/2)
Solution
In the center of mass system the Hamiltonian is
P2 P2
H = — + f— + V(r) (20.3.6)
4m m
where
The first factor (describing the center of mass motion) is clearly symmetric under
the interchange of particle labels. Thus, since the spin is 0 and we are dealing
20.4. IDENTICAL 5 = 1 / 2 FAKl'lolí» íív oísív / n c IVIADS 111
- L + V(-r)]
1 eilr + e-ik-f ^ _
7i
+ [f{dj + f { n 9 i p +
(20.3.11)
Now, the detector cannot distinguish between the two particles. Therefore, the
differential cross-section is the sum of the cross-sections for particles 1 and 2.
However, they have identical scattering amplitudes. Thus,
Solution
We proceed as in the problem above. However, this time the total wavefunction,
in the center of mass system, is given by
ñ
Ȓ(R, r, S, m s ) = e'^ tp(r)x(S,m,) , (20.4.13)
where x{S, rns) is the wavefunction for the total spin of the two particles,
a) If the scattering occurs in a singlet (s = 0) state then x{S> m ' ) antisym-
metric under the interchange of the two particles. In this case we need that
the total wavefunction must be antisymmetric under the interchange of the two
2U. SYSTEMS OF IDENTICAL PARTICLES
particles and hence, that the spatial part of the wavefunction should be sym-
metric just like in the spin 0 case considered above. Thus, for the singlet case
we get that the appropriate scattering amplitude is
Here, we have deliberately dropped the factors of 2 and 1 / \ / 2 that cancel each
other.
b) If the total spin is S = 1 the spin wave function is symmetric under the
interchange of particles. Thus, we need that the scattering amplitude should be
antisymmetric. The result is
a
(0,<p) = |/(»=i)(0,¥>)| 2 = \f(0,<p) - f ( n ~ 0,<P+ t ) | 2 • (20.4.17)
Again we have deliberately dropped the factors of 2 and l/y/2 that cancel each
other.
h = ^ 2 fw k a\ak •
k=o
where
OO
^»(®) = •
k=0
Hint: Expand the second exponential and commute ak through showing that
Solution
The field Hamiltonian is
H=^hLjka\ak. (20.5.18)
k=0
So, the field operator in the Heisenberg picture is related to the field operator
in the Schrodinger picture by
V's(i) = ^ { x \ k ) a k . (20.5.20)
k=o
To evaluate exp{¿Hí/h} ak exp{—iWt/h) we first consider
v a —1
^k (^n^n) — (^n^n)
5—1
— (^/cn + ^k (^n^n)
5-2
— (&kn "i" ^k^n^n (®n®n)
= {Skn + aÍan)2 ak (a^an)3 . (20.5.21)
So, after n steps, we have
ak ( 4 a t ) ) ' = ( 4 n + alan)' ak . (20.5.22)
Hence,
Therefore,
ex
ip(t,x) = exp{¿Hf/ft} ips{x) p{ — ¿Hi/ft}
OO
Solution
The solution for fermions starts as in the problem above. However, now it is
even simpler since ak commutes with aj,a n for k ^ n and for k = n we find
exp(¿Hí/ñ) ak exp(—iHt/h)
= exp it ak exp -it an)
= e~iWkt ak . (20.6.27)
= j 2 ( s \ k ) e~iwkt ^ • (20.6.28)
k=0
20.7 T w o - b o d y Interaction
Obtain the equation
V=^ J d3xd3y^(x)ip^y)V(x,y)rp(y)ip(x)
1 n
v(xi,x2,...,xN) = ~Y^y(xj,xk).
Solution
For any configuration space operator F its representation in occupation number
space is given by [20.1]
Now we replace F by
1 JV N
V{xux2,...,xff) = = ^2v(xj,xk) . (20.7.30)
j<k
(no,ni,...|V|mo,mi,...)
=
N(N — 1) I d3Xjd3Xkip^ (£j)
j<k
= (n0,ni,...\^ J d3xd3yi^x)ip](y)V{x,y)rp(y)xp{x)\rn0,Tril,...) .
(20.7.31)
Therefore,
as required.
and choose the constants uk and vk appropriately. What happens if ak, a[ are
fermi operators? These are simple examples of Bogoliubov transformations.
Solution
We want the transformations
bk = uka{ + vkak
Vk = e"*k cosh 6k
To get the (off-diagonal) terms involving ój.6], as well as those involving bkbk to
cancel we require that
1 at
— Ek e ( +'3t) sinh 6k cosh 6k
+ Xk e2iak cosh 2 6k + X*k e2i/3k sinh 2 9k = 0 . (20.8.39)
Therefore, writing
A/c = | Ajt |eitTfc (20.8.40)
we see that if we choose
afe = - y , = y (20-8.41)
So, we take
Hk =
~JWk — 4|Afc|2 ^ ~ 4|Afc|2cos<r] [b\bk + 0 - ^ . (20.8.45)
Solution
We prove the result in two steps. First we prove by induction that
BA
" = ¿ ( " ) Ar[B,A]„-r (20.9.47)
where
[B,A]0 = B , [B,A]n+1 = [[B,A]n,A]. (20.9.48)
Clearly the result is true for n = 0. We now assume the result holds for n — 1.
Thus,
n-1 •
BA"'1 = ( n\ — 1 J Ar[B, A] n _ x _ r . (20.9.49)
r=0
Then,
n—1
BA n 1
" = E ( r )Ar[B,A]n^rA
r=0 ^ '
n 1
X) ( ~ ) X ] „ - i - r + [S, A]„_ r } . (20.9.50)
448 CHAPTER 20. SYSTEMS OF IDENTICAL PARTICLES
(20.9.51)
71—1
(20.9.52)
0
Also for s = 0 there is only one term, namely
BAn
=Í2( " )Ar[B,A]n-r. (20.9.54)
r=0
Now consider
n
°° An °° i
= = b.A].. (20.9.55)
v
n=0 n=05=0 '
D A _ An~' [B,A],
2^Z^(n-s\!(n — s)! «!s!
5—0n=s
oo oo
Ar [B,A],
= E E (r)\ s!
a=0 r=0
[B,A],
= e
E
5= 0
(20.9.56)
A
Multiplying by e on the left we obtain the desired result.
with
|U|2+|t,|2=l , UV*~U*V = 0.
U = exp | c a j a 2 — c* a 2 a i j
b\ = Ua\U] , bi = UaiU]
b\ = Ua\Ut , b2 = Ua2Uf .
Solution
a) We first verify the anticommutation relations.
Similarly
M i = M2 - 0 (20.10.59)
Now,
a
[a.j, A] 0 = j • (20.10.62)
utiaftisk -ju. systems of identical particles
Then,
[ai,A]1 = [al,A] = ca\ , [a2,A]l = [a2,A] = ca\
[ai,A]2 = [ca\,A] = - | c | 2 a i , [a2, A}2 = [ca\, A] =-\C\2CL2
K,^]3 = -[|c| 2 «i,-4] = - c | c | 2 a í , [a.2,A)7i=[\c\'2a2,A] = c\c\'2a\
[ai,^] 4 = -c[\c\2a\,A] = \c\Aax , [a 2 , A]4 = c[|c| 2 aj, A] = |c| 4 a 2
etc.
Thus,
= Sk,k' , [afc,a A ] = 0 .
20.11. DIAGONALIZATION OF BOSE HAMILTONIAN 451
Solution
We have
H= + (AKAK + 4 4 ) (20.11.66)
k k
where ui(k) and V{k) are given functions of k and
4 = u b
k l + v'kbk ' a
k = u*kbk + vkbl • (20.11.71)
Inserting this into the Hamiltonian we get
cos
j(k) sinh2x(fc) + ^ h ^x(k) (ȒȒ+'i'O
k
J
(k) g (cosh 2 x ( k ) - 1) + sinh %x{k) (20.11.73)
wJinriün ¿\). O 1 S I MS Ut IDEN TICAL PARTICLES
For the Hamiltonian to be in diagonal form requires that the terms involving
fclfcl
k k
+ Kbrbr
K vanish. This means that
So we find
u(k)
cosh2x(fc) =
y/u[k)2 - V(k)2/4
sinh2x ( k ) = _ (20.11.75)
^fu;{k)2 -V{k)2/4
Then we have that
.(k)2-VCk)2/2 « U{k)>/2 + vM/4 „(*) ( 2 o n 7 6 )
yjuj(k)2 -V(k)2/4 k
\/u(k)2 - V(k)2/4 2
H +
= óX ^ * lfl)( a I a fc + alfcO-ik) + + afca_fc)] .
fc> 0
Solution
The Bogoliubov transformation may be written
ak = 6/. cosh Xk ~ &Lfc sinh Xfc
a_ fc = 6_fc cosh x-fc - ¿¡sinh x~k
a
l = ¿fcCoshx/t - 6-fcsinhx/t
a}_k = 6Lfc cosh x-fc - sinh x-fe (20.12.77)
where bj., bk etc. are again Bose creation and annihilation operators. The
Hamiltonian, in terms of these new operators, becomes
H 6
= \ X IM(-Ea + If I) (cosh2 Xk + sinh 2 X-fc)
k> 0
- 2|¿f| cosh Xk sinh x~k
+ &L k b- k [(E k + |^|) cosh2 x-fc +sinh 2 xfc) - 2\g\cosh X-fc sinh Xfc]
We now choose
(20.12.79)
X-k = Xk
and
l e k + | <71
cosh(2xfc) =
Ek
sinh(2xic) = (20.12.81)
+ ttEk +
M)[cosh(2X*) - 1] - \g\sinh(2xfc)] . (20.12.82)
k> 0
After we replace cosh(2x/c) and sinh(2x/c) by their values we get
Define
bk = cosh A « t + sinh A a\
b\ = cosh A a [ + sinh A ak •
vjv <*k VN = bk .
c) Show that
lim (\C|W|$) = 0
N-too
Solution
a) Straightforward computation shows that
ÍTn a "
= \ £ (a* - aI2) (20.13.88)
a= 1
¿u.u. BOGOLIUBOV THAiysrunmAiiuiy: aus>& ur&nAi uns 4óó
A N
iT N = - S 2 A k . (20.13.89)
k=1
¿ í ^ ( 0 | («2-at T | 0 ) . (20.13.91)
n=0
For the typical mth order term, in this sum, all contributions come from terms
having an even value of m or m = 2n and of the form (—l)"a 2 .. .a*" with an
equal number of a 2 and a*** and the at 2 to the right of the a 2 . Using this we
find that
and
and generally
Thus,
Now, for limjv^-oo the number of such terms tends to oo and we get a term
bounded by
lim [e /4]-'v r
x irrelevant finite factors .
N—too
J I R / J I I X.L/. ÍJ ± KJ JL JLJIVIID V_/R 1JL/JC/Í\ 1 I O A L I V L I Í L IGLJFC/S
Solution
a) The expectation value of A i s given by
= J y * ( x i , q j ) A ( r ) y ( x i , q j ) d x i .. ,dxndqi .. ,dqN
But,
(x|a< r >|x)
Therefore,
= Tr[A( r V] • (20.14.98)
c) To get the equation of motion for the reduced density matrix we expand it
in terms of the eigenfunctions of the Hamiltonian H ^ for the reduced
system. Thus, the time-dependent density matrix is
S = f l + g(ñ • (?)
l/l 2 + m 2 = i •
Solution
Since the incident beam is unpolarized, it is described by the density matrix
Pi = l-l . (20.15.103)
tüO <^ru\r inn. zu. sxSTEMS Ut IDENTICAL PARTICLES
pj = SpiS^
= iss'
= ¿1 + »(/'s)(S • s) . (20.15.104)
Therefore, the polarization after the scattering is given by
where <?i and <72 are the Pauli matrices associated with the spins of particles 1
and 2 respectively and where
vi = \i
r < a
V2 = ( —U r <a
0 r > a " \ 0 r >a
Here, V and U are positive constants. If this system is known to have a single
zero energy 1 = 0 bound state and a single zero energy 1 = 1 bound state, find
V and U.
Solution
For the 1 = 0 state we have a symmetric spatial wavefunction in the relative
coordinates and therefore the spin state must be antisymmetric or s = 0. For
the / = 1 state the spatial wavefunction is antisymmetric and therefore the spin
wavefunction corresponds to s = 1. Now,
s2 = (si + s 2 ) 2 (20.16.106)
so that
V W
V(r) = V i - 3 V 2 = { - + H I . (20.16.110)
The Schrodinger equation for an I = 0 bound state with energy E < 0 becomes
H2 1 d2(rip)
—» t~5 (-(—V + 3 U — E)ip = 0 r<a
zm r dr¿
h2 1 d2(rip)
-2¿>rd¿r~E* = 0 r > 0
- (2016U1)
The solution is
J j4sin(Arr) r < a ¿ 2 = 2m(v-3t/+£)
r,> = (2 16 112
{ Be~" r>a «' = # 1 " ° ' »
kcot(ka) = -a . (20.16.113)
or
h2
ka = (n + l/2)7r =>• V — 3C/ = j ( n + l/2) 2 7r 2 . (20.16.115)
The first case V — 3Í7 = 0 is not possible since this would mean that for this
situation there is no potential at all and hence there is no bound state. So,
.2
V-3U = . (20.16.116)
8 ma¿
We now repeat the solution for the I = 1 case. Here the Schrodinger equation
reads
h2 Id2(rip) 2h2
~2mr~dri ^ 2mr^ + (—V — U — E)i> = 0 r < a
Now, letting l^l -4 0 we find that the right hand side goes to —1 so that the
equation for the energy becomes
(A' 2 a 2 + 1) sin(Aa) — A'acos(Aa) = —A'acos(A'a)+sin(A'a) .(20.16.120)
Hence, we find
A' 2 a 2 sin(A'a) = 0 (20.16.121)
so that Ka = 0, which is impossible since both U and V are positive, or else
h2
Ka = pn =>• U + V = 2ma2p2x2 * (20.16.122)
So,
U + V = § ¿ . (20.16.123)
Combining this with the result for the singlet case (20.16.115) we get
u r 120 16 1241
= =IS •
20.17 B o s e N u m b e r Operator: Constant
of M o t i o n
Show that for a system of bosons with the Hamiltonian
E
H = (k)alak + A
£ a}alaíafi^+kSr+ñV0~ «)
k jklñ
the number operator
a
n = p h
Solution
We first compute
[atdp-, ata g ] = [atap-, at]a¿ + at[atap-, a £ ]
= — alap^f,k
= 0 . (20.17.125)
Next we compute
[alap, ata^.a^a^] = aí[ap, ata^-jajja.ís] + ata^-[at, aña^ñ]af
a
= UaU6P¿+jp,-fcada"a-"
- a\aKk{a_ñS?¡ii+añ8Pi-ñ)ap . (20.17.126)
20.18. BOSE OPERATOR: MORE CONSTANTS OF MOTION 461
— ^ ^ 1 2 A Í A ^ J . A ^ A _ ^ V(—k — TÍ)
kñ
= 0. (20.17.127)
Combining these results we have that
[N,H] = 0 (20.17.128)
which was to be shown.
show that
a) the operator
N = a¡¡ao + at + a^-a_¡¡
b) the operator
D = a*a-k - a[.a_'k
[a 0 ,a^-] = 0 . (20.18.129)
as well as
[ a o a o; ( a o) 2 ] = -2(ao) 2
[aoao> ( a o) 2 ] = 2(aJ) 2 . (20.18.131)
It then follows that
[aja 0 , H] = - 2 g | ( a j ) 2 a ¿ a _ ¿ - a l a ^ . ( a 0 ) 2 } (20.18.132)
as well as that
[aUa-k> H
l = 9 {(«í) 2 «fc a -fc - 4 a - f c ( a ° ) 2 } • (20.18.134)
So, after combining these results we immediately see that
[N,H] = 0 . (20.18.135)
b) Furthermore, it is also immediately clear that
[D, H] = 0 . (20.18.136)
20.19 T h e Pauli P r o b l e m
a) Using the equation
E
m = —
cr
calculate the total mass of one gram of electrons confined to a cube 1.0 cm on
a side.
b) Repeat the calculation assuming that instead of electrons the particles are
bosons with the same mass as the electrons.
Solution
a) This problem shows the dramatic effect of the Pauli exclusion principle. In
one gram of electrons there are
=
N
i ^ fn n = 1.1 x 1027 electrons .
-28
9.1 x 1 0
The energy levels for a particle in a cube of side a are
E„ = £Lk> (20.10.13T)
where
k7 = kl + k2y + k] (20.19.138)
and each of the kx, ky, kz is an integer.
As a first step we need to evaluate the degeneracy of each level k. This means
we need the number of different combinations of kx, ky, k2 such that (20.19.138)
is satisfied for a given k. In general, with a few exceptions, this degeneracy is
3. Since the number of electrons is very large, we can take the degeneracy to
be 3. This means that, according to the Pauli exclusion principle, we can place
in each energy level
2x3 = 6 electrons .
The factor of 2 is due to the two possible spin values. To get the total energy
of the N electrons we have to sum all the energies from k = 1 up to k = Ar/6.
Thus, the total energy is
h2*2 ^
e
= 5 = ? e * "
k=1
h27T2 (N/6)[{N/6) + 1]
2
2 ma 2
2 2 2
h 7T N
* h w ' < 20 1 9 139
»
E
m = —
c-
we convert this to a mass and find
10 26
= 105 gm = 100 kg! . (20.19.141)
9 x 1020
b) If the particles were bosons they would all be in the ground state and the
total energy would be simply
E = ^ - N ^ 14 e r g s . (20.19.142)
2 ma-
in this case the additional mass would be a negligible
14
= 1.6 x 10~ 20 gm . (20.19.143)
9 x 1020
464 CHAPTER 20. SYSTEMS OF IDENTICAL PARTICLES
r 0 = 1.2 x 10" 1 3 cm
and A is the atomic number of the nucleus. Thus, the potential energy experi-
enced by an electron near a nucleus is not simply
V(r) = •
r
If we assume that the charge density in the nucleus is constant then we have
instead the potential energy
Ze*
R
H r <R
V{r) = <! \2R2 2
J ' ~ " . (20.20.144)
I r >R
a) Use perturbation theory to calculate the isotope shift (that is the dependence
on A of the K-electron (Is state) for an atom with Z protons and atomic number
A.
b) Use this result to compute the energy splitting for the K-electron between
the heaviest lead (Z = 82) isotope A = 214 and the lightest A = 195.
Neglect the presence of the other electrons.
Solution
a) The unperturbed Hamiltonian is
v2 Ze2
Ho=£ • (20.20.145)
v
2m r
The perturbation is
E
*) =
~ \ % (20.20.147)
z a, I ZJ
(1)= i -2Zr/a r2 _ 3
f '
Jo r~ dr (20.20.150)
2ft 2 ~ 2 + 7
0
We now let 2 Va J R
2ZR
a - x - (20.20.151)
2Zr
Then,
1 Ze 2 x2 3 a
I?(!)
c/0
_
— 2 fi Í ' - + x 2 i/x
Jo 2a2 ~ 2 x
1 z e 2 2 r 12 ,12 12 „
3+ a —e [ —— -| 1- 3 (20.20.152)
2 a / 2 a [a2 Ot
If we now make the dependence on the atomic number A explicit by writing
)(A =
- 9214)
1 A\ -= I £ l _ x 1.08 X 10-2 (20.20.156)
2 a/Z
Thus, recalling that
1 e2
= 13.6 eV
a2
the energy difference between the two isotopes is
1 7 f>2
AE = x 9.4 x 10- 4 = 86 eV . (20.20.157)
2 a/Z
466 CHAPTER 20. SYSTEMS OF IDENTICAL PARTICLES
Bibliography
[20.1] A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd edition, World Sci-
entific Publishing Co. Pte. Ltd., section 20.10, (2002) .
C h a p t e r 21
Q u a n t u m Statistical
Mechanics
Solution
a) The microcanonical ensemble is given by
(21.1.1)
E<En<E+A
Let
Then we have
N7
(21.1.3)
n=N,
So
i N,+I-N,
huj rjV2(Ar2 + 1) _ (Ni - l)iVi N2 + 1 - Ni
N2 + 1 - Nx 2 2 2
468 CHAPTER 21. QUANTUM STATISTICAL MECHANICS
ñu {N2 + N!){N3 - Ni + 1)
1+
~2 N2 + 1 - JVi
zi = E 1 ,-/3fi(n +1/2)
n=0
e~ptUo/2
~ j _ e-phu
1
(21.1.6)
2sinh(/?/iu;/2)
Next we have that
= [ZX)N = [2 sinh(/3ñw/2)]_'ÍV (21.1.7)
Then,
U = ((E)) = -~\nZN
p\
= N— ln[2sinh(/?/k<;/2)]
OP
= N^-coth(/3hw/2) . (21.1.8)
U =
=
fruí z coth(/3 hw / 2)
(21.1.10)
2 2 sinh(/?fiw/2) — z
But we also have that
N
~ zjrlnZG(P>z) = o - utnl /<)\ (21.1.11)
dz 2 sinh(/?/kj/2) — z
21.2. pROpERTIES OF THE DENSITY MATRIX 469
Therefore,
U = N^-coth(phoj/2) (21.1.12)
P2 <P •
This implies p > 0.
b)
Tr([p,A]) = 0 .
To see that this is not trivial, consider Tr([x,p]) and discuss a necessary condi-
tion on the operator A for this to hold.
Solution
a) To prove
p2<p (21.2.13)
P =^\n)Pn(n\ (21.2.14)
n
with
Y2Pn = 1 Pn
- 0
' (21.2.15)
n
b) For this part we simply write out the commutator for a general matrix element
and sum over the diagonal elements
Thus, as required,
t r b , ¿ ] = 0. (21.2.18)
4YU CHAPTER 21. QUANTUM STATISTICAL MECHANICS
This result relied heavily on the fact that the density matrix is a bounded op-
erator, that is, its eigenvalues have a finite bound. Clearly the commutator of
the two unbounded operators x and p is ih and its trace blows up. This is the
case for all unbounded operators since their eigenvalues are not bounded and
therefore neither is the sum of their eigenvalues. So, a necessary condition is
that one of the operators be bounded.
Solution
a) Using the given information we have that
Therefore, we find
is maximized when
Tr(Ai) = 1 • (21.3.23)
p' = a ¡3
(21.3.24)
¡3* 1 — a
21.4. EXPECTATION VALUE FOR NUMBER OF PARTICLES 471
where
is maximized when
Tr(pV 2 ) = 1 . (21.3.27)
>>' = ( r l - a ) P13'28)
where
Tr
((«*» = (Ps¿) = P^ + -p)Tr(p'<r 3 ) (21.3.30)
is maximized when
Tr(pV 3 ) = 1 . (21.3.31)
£ ) (21.3.32)
and
21.4
P = P
I (1
7" ) • I21'3 33»
E x p e c t a t i o n Value for N u m b e r of Particles
Verify the equation
JV = « N » = z ^ l n Z G
Solution
To verify
we begin with
N = ^ (21.4.35)
where
p = zNe~pH (21.4.36)
and
ZG=rTrp. (21.4.37)
Then,
Z lnZa = 2 ln TT zN e P
lL ^ [ ( ~")}
TT(UZn e~0H)
t ^ j
TV(Np) (21.4.38)
Tr(p)
as required.
Solution
a) The most general pure spin 1/2 state is of the form
P = (X\A\X) • (21.5.40)
- (J)
P -
_ (JiJk + JkJi) 2
~ j ( j + l)^ 2 ~ 3 <fe
'
Hint: Show that the unit matrix 1, the polarization vector p, and the quadrupole
polarization tensor Qik form a basis for expanding any 3 x 3 matrix.
Solution
We first need to show that the unit matrix 1, the polarization vector p, and
the quadrupole polarization tensor Qik form a basis for the expansion of the
density matrix. To do this we need 9 linearly independent matrices. We start
by considering the 9 matrices ( l / h 2 ) J i J j i,j = 1, 2, 3 where
0 -i
í - i í i or ii ül i v/2
J3
0 0 (21.6.45)
h
0 0
V^UÍT.1» X UÍW j l / l l i O l I U A L M£/OiÍAiViOi!)
1 1 4 2 i
^2 (JiJj + JjJi) — + J +
2 3 '¿h2 e'jk k
=
3^ ,J 2K¿ e i i k ^ k ' (21.6.46)
Notice that all of these matrices, except the unit matrix, are traceless. Also,
only five of the Q¡j can be independent since
To show that these 8 traceless matrices plus the unit matrix form a basis for
3 x 3 matrices we need to show that if
then,
ao = 0, a = 0 , a,j = 0 .
To complete this proof we simply evaluate (by multiplying out the various ma-
trices) the following traces.
Tr (J?) = 2h2
Tr (Ql) = 2/3
Tr (Qfj) = 1/2
Tr ( J i J j ) = 0 i
Tr(JfcQij) = 0
Tr(QkiQij) = 0 i (21.6.49)
We now take the trace of (!
:
3ao 0 .
Next we multiply (21.6.48) from the left by Jk and take the trace to get
2ak = 0 .
Finally we mutliply (21.6.48) from the left by Qk¡ and take the trace to get
a,j = 0 .
This shows that these 9 matrices are linearly independent. Thus, we can write
1 1 ^ 1 j"2 , *^3 ,
p — ao 1 + ai ——b a>2 ——h 03 ——h
n n n
+ «11q11 + CI12Q12 + 113Q13 + 0.22Q22 + 0.23Q23 • (21.6.50)
I 21.7. POLARIZATION VECTOR FOR SPIN j 475
- (j)
P=\k (21.6.51)
then
Also the remaining five constants a n , , a 2 3 are expressed in terms of the five
linearly independent quantities
(Qij) • (21.6.53)
( / )
*
P—
jh
and are given a "hamiltonian"
H = -7/• B
Í Í2 = 0 .
dt
Solution
This is a straightforward computation.
i k f = L t h l i , .-{•><•}
at jn dt
- • ^ T r { / ( / • B)p - Jp(J • / ) }
-j-hTx{[J{J-B)-J{J-J)]p}
jj^ihTr | ( J x 5 ) p |
ihyp x B . (21.7.54)
476 CHAPTER 21. QUANTUM STATISTICAL MECHANICS
Therefore,
^ = 7 PxB. (21.7.55)
b)
Tr p = 1
c
)
P2 <P
as well as the equations
d)
=
Pm]m' Pm,n-,m',n (21.8.58)
n
and
= " (21.8.59)
m
Solution
The density matrix p is defined by
(1)
= 0 * />(2)*
b)
Tr p — ^ ^ pm,n]m,n
mn
= = l • (21.8.62)
mn
c
)
=
(P)m,n;k,l ^ ^ Pm,n\m' }n' Pm' ,n'\k }l
m'n'
= v 0 ( 1 ) p (2) 19(1) 0 (2)
/ v rm,m' rn ,n' Pm' tkPn' ,1
m'n'
= (p(1))m,fc(/9(2))2,i • (21.8.63)
Therefore,
(P)2 = (P(1))V2))2 < P(1V(2) = P• (21.8.64)
d)
, (i) r . p )
,n — Pm,m' / ,Pn.n
= Prn,m' • (21-8.65)
Similarly,
Y^P^m.n' = Pn}n-
m \ m /
= p™n. • (21-8.66)
and
Tr<r(i> = Trr ( i ) = 1 .
1 + 6 = fa p + S = p 2
satisfies
=
Pm]m' P m
, n
, m
' , n
(21.9.67)
n
and
and is a possible density matrix for the composite system. This establishes the
necessity of the condition that p' 1 ' and p^> correspond to pure states in order
that
give a unique solution for a density matrix for the composite system.
Solution
We have that
= «1 Vm,m- + PlTSm>
= PÍ]m" (21-9.71)
a a T a Ó T
X / V » ; - . » ' = n ^ > + P n , l + 1 n ! n > + n %
m
= ( a + 7 ) ^ ' + (^ + í ) t " s »
(2) , a (2)
= <X2*n,n' + foTn,i,'
= til • (21-9.72)
Z1.10. TWO ENERGY LUVULS BUS je, Ü/IO <i/y
Solution
The Hamiltonian is
H = E(a\a2 - a j a j ) = E(N2 - Ni) . (21.10.73)
a) To compute the canonical partition function we use the fact that the total
number of particles is fixed. So,
N = N1 + N2 or NX=N-N2 (21.10.74)
and
H = E(2N2 — N) . (21.10.75)
So,
N
/3(2N2-N)
ZN = Tre-"" = Y , e~
N2=0
1 _ e-/32£(iV+l)
E0EN
1 — e~p2e
smh[(3E(N + 1)]
(21.10.76)
sinh /3E
For bosons we have
= no-»-'*)-1
k=1
1 1
= (1 - z e - ^ y {l-ze^y
= (l — 2z cosh (3E + 2 r 2 ) 1
. (21.10.77)
tou ON/URI^RI ¿±. U¿UA¡\ 1 UM S1A11SUUAL MECHANICS
V Z N
( ,P(n+I) _ ,-p(N+iy
ih((3E) ^¿7-1
2 sinh(/?£) V
1 / e?e e~pe
2sinh(/?¿?) — zePE 1 — ze~PE
1 / 1 1
2sinh(/?Z?) \e~PE — z e^E — z
E
1 eP _ e-0E
= (1 - 2 z c o s h ( / ? £ ) + z 2 ) - 1 . (21.10.78)
—2z£"sinh (3E
(21.10.79)
1 — 2z cosh ¡3E + z 2
N = m ) = z~ In ZG((3,z)
2 z cosh (3 E — 2 z 2
(21.10.80)
1 — 2z cosh (3E + z 2
Therefore,
u = -AT - i E ^ J E . (21.10.81)
2 cosh j3E — z
We can further solve for z in terms of N using the equation for N. This is
simply a quadratic and yields
2 =
fiT+2 { ( ^ + ! ) c o s h p E ± y / { N + 1)2 cosh 2 (3E-N{N + 2) | . (21.10.82)
Now
Therefore we see that for E —y 0 we have to choose the + sign in the above
expression.
21
-U. DENSITY MATRIX: PARTICLES COUPLED BY SPRING 481
í, = + + í ( i i ! ) !
¿ ¿ 5 '- • <2111-84>
u = y/k/m , m = —— . (21.11.87)
mi+m2
Finally compute the reduced density matrix for particle 1 and show that it is
not idempotent; i.e. that correlations between the two particles persist.
Solution
In the centre of mass system defined by
_ m1xi+m2x2
M
x = x\ — x2 (21.11.88)
with
M•= mi + m2
17111712
m = (21.11.89)
mi + m 2
the Hamiltonian becomes
The eigenfunctions for the nth excited state written in the centre of mass system
are
where
d
~§m cfo 2 ^ +
\ r m j j 2 x 2 u ^ i x ) = (" + 1/2 )hu un(x) . (21.11.93)
Pi(x i,«i)
iK
1
2
(2*)
exp (~jfi~(xi ~ xii) J dz un(z)un(z + (x i - a r i ) ) . (21.11.97)
the density matrix p\ is not idempotent and we see that correlations persist.
To convince ourself of the fact that correlations persist we check that p\ is not
idempotent.
e x iK _ x
J PÁxux'^p^x'^x'Ddx'x = ^ ? P "))]
= Pi(xi,x") . (21.11.99)
21.12. PARTICLES: DISSIMILAR, BOSE, AND FERMI 483
Solution
For the non-interacting particles the energies are just the sums of the individual
energies for a particle in a box and so are given by
(21.12.100)
where
a) Distinguishable Particles
In this case the unperturbed wavefunctions are
^ñuñ2{ri,r2) = uñl(fy)uñi(r2) • (21.12.102)
To simplify the computation we work with one coordinate at a time. The change
in energy in first order perturbation theory is given by
dx sm2(n\xnx/L) sin2(n2l:7ra;/L)
l
2 1 f
dx [1 — cos(2nij;7ra;/L)][l — cos(2ri2Xnx/L)]
^(2 + ¿n li n 2 x) • (21.12.104)
484 UHAr i UK 21. QUANTUM STATISTICAL MECHANICS
4 S i l , = (ifo..*». • (21.12.107)
After carrying out the integrations over the delta functions we get exactly double
the integral in the previous case for each direction. Thus,
=
"g"(2 + ^ni x n 3 l )(2 + íni v n 3¡) )(2 + ¿nij,n3B) • (21.12.108)
c) In the case of fermions of spin 1/2 we have two possibilities.
1) If the particles are in a singlet state (s = 0) then the spatial wavefunction is
as for the bose case above and we get
=
^1^2 ~g"(2 + ¿r»i I n 3l )(2 + ¿n lB n 2s )(2 + in ly n 3 y) • (21.12.109)
2) If the particles are in a triplet state (s = 1) the spatial wavefunction is
antisymmetric. In this case carrying out the delta function integrations causes
each integrand to vanish and we get
4 ' u = 0 • (21.12.110)
^laser — t •
h
Assume that the only radiation present is the pumping frequency
_ E3 — Ei
h
and use the Einstein Coefficients A¡j, D t j (see problem 1.8) to determine the
ratio of N3/N2 as well as the ratio N-2/N\.
Hint: Write out the equation for the rate of change of the three populations and
impose the equilibrium conditions.
21.13. A THREE-LEVEL LASER 485
Solution
The rate of change of the population in level 3 is entirely due to emission of
radiation. Thus,
dN3
— = -N3Bl3p(u) - N3(A23 + A13) . (21.13.111)
The first term on the right describes the induced transitions from level 3 to level
1, while the second term describes the spontaneous transitions from level 3 to
level 2 as well as from level 3 to level 1. For level 2 we have only spontaneous
transitions from level 3 to level 2 and from level 2 to level 1. Thus,
dN2
^ - = N l B 3 1 p(v) . (21.13.113)
Thus, if A2\ > A32 we have population inversion. Also, at equilibrium, the rate
of change of population 3 equals the rate of change of population 1. Therefore,
N3 _ B3ip(v)
(21.13.117)
Ni B\3p(v) + (^23 + ^13)
Now, writing
N2 _ N3 N2 (21.13.118)
~Ñ[ ~ ~Ñ~X~Ñ~3
we find that
N2 B3ip(v)A32 (21.13.119)
N\ A2\[B\3p(v) + (A23 + A\3)\
lOU \jnj\r run zi. ljuawium STATISTICAL MECHANICS
,±lh)
= L ? ± t - <2114120>
a) Show that I+(k) can be expressed in terms of I~(k).
b) Evaluate I±(k) in terms of the Riemann zeta function [21.1]
00 1
= • (21.14.121)
n = 1
Solution
a) To express I+(k) in terms of I-(k) we use the identity
1 1 2_
(21.14.122)
ex + 1~ ex - 1 e2x -
Then,
, r xkdx n r°° xkdx
'*{t) =
L — i - 2
1 (2i 14123)
'
The first integral is simply I~(k) and in the second integral we replace the
integration variable x by t — 2x. Then this integral becomes 2 - f c /_(fe). Thus,
we have that
/+(*) = ( l - 2 - * ) / _ ( / f c ) . (21.14.124)
b) To express I-(k) in terms of the Riemann zeta function we expand the
/_(*) in
denominator = the/f integrand
x* dx in a binomial series.
Jor°° e T.k- , 1
^L
n=00 .©o
* e—' dx
1
=
n=1 J o n = l
OO
= r l'tHEj •
1
r(A+l)C(Jfc+l) (21.14.125)
Bibliography
[21.1] E.T.Whittaker and G.N.Watson, A Course of Modern Analysis - ^th
Edition, Chapter 13, Cambridge, (1963).
[21.2] Kerson Huang, Statistical Mechanics, John Wiley and Sons, Inc., (1963).
References
There are several books that deal with problems and solutions in quantum
mechanics. The following is a fairly exhaustive list of such books presently
available in English.
7. L.P. Kok and J. Visser, Quantum Mechanics: Problems and their Solutions,
Coulomb Press Leyden, (1987).
On quantum mechanics itself there are very many excellent texts on the
market. A small selection of representative texts is listed below.
12. A.Z. Capri, Nonrelativistic Quantum Mechanics 3rd Edition, World Scien-
tific, (2002).
15. E.H. Dicke and J.P. Wittke, Introduction to Quantum Mechanics, Addison-
Wesley Publishing Co., (1960).
18. S. Gasiorowicz, Quantum Physics, John Wiley and Sons, Inc., (1974).
19. L.D. Landau and E.M. Lifshitz, Quantum Mechanics, Pergamon Press,
(1985).
21. E. Merzbacher, Quantum Mechanics, John Wiley and Sons, Inc., (1970).
25. L.I. Schiff, Quantum Mechanics, McGraw-Hill Book Co. Inc., (1968).