Solutions to Problems in Quantum Mechanics

P. Saltsidis, additions by B. Brinne

1995,1999

0

Most of the problems presented here are taken from the book Sakurai, J. J., Modern Quantum Mechanics, Reading, MA: Addison-Wesley, 1985.

Contents
I Problems
1 2 3 4 5 Fundamental Concepts . . . . . . . Quantum Dynamics . . . . . . . . . Theory of Angular Momentum . . . Symmetry in Quantum Mechanics . Approximation Methods . . . . . . Fundamental Concepts . . . . . . . Quantum Dynamics . . . . . . . . . Theory of Angular Momentum . . . Symmetry in Quantum Mechanics . Approximation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 7 14 17 19

3

II Solutions
1 2 3 4 5

. 25 . 36 . 75 . 94 . 103

23

1

2 CONTENTS .

Part I Problems 3 .

.

3 (a) The simplest way to derive the Schwarz inequality goes as follows. FUNDAMENTAL CONCEPTS 5 1 Fundamental Concepts 0 1. that is. (a) Suppose Sx is measured. where n is a unit vector lying in the xz-plane that ^ makes an angle with the positive z-axis. hSx i)2i: (For your own peace of mind check your answers for the special cases = 0. There is no degeneracy. . (b) What is the signi cance of Y (A . a 00 00 a 0 6 0 0 00 1 (c) Illustrate (a) and (b) using A set equal to Sz of a spin 2 system.2 A spin 1 system is known to be in an eigenstate of S n with 2 eigenvalue h=2.1. h(Sx . a ) 0 is a null operator. ~ ^ 1. (a) Prove that Y (A . a ) ? a =a a . =2.) 1. First observe (h j + h j) (j i + j i) 0 for any complex number then choose in such a way that the preceding inequality reduces to the Schwarz inequility. and . What is the probability of getting +h=2? (b) Evaluate the dispersion in Sx.1 Consider a ket space spanned by the eigenkets fja ig of a Hermitian operator A.

in agreement with (b). h 4d2 satis es the uncertainty relation q Prove that the requirement hx j xj i = (imaginary number)hx j pj i is indeed satis ed for such a Gaussian wave packet. Evaluate the classical Poisson bracket x F (px)]classical : (b) Let x and px be the corresponding quantum-mechanical operators this time. (x . (c) Explicit calculations using the usual rules of wave mechanics show that the wave function for a Gaussian wave packet given by " 2# 2 ) 1=4 exp ihpix . prove that x exp iph a jx i 0 x x exp iph a : (xjx i = x jx i) 0 0 0 . Evaluate the commutator (c) Using the result obtained in (b). hxi) hx j i = (2 d 0 0 0 .6 (b) Show that the equility sign in the generalized uncertainty relation holds if the state in question satis es Aj i = B j i with purely imaginary. 0 0 h( x)2i h( p)2i = h : 2 q 1.4 (a) Let x and px be the coordinate and linear momentum in one dimension.

exp ix h 2 Quantum Dynamics 2.1 in Jackson.1 Consider the spin-procession problem discussed in section 2. Sy (t). 2.2. It can also be solved in the Heisenberg picture.5 (a) Prove the following: @ (i) hp jxj i = ih @p hp j i Z @ (ii) h jxj i = dp (p )ih @p (p ) where (p ) = hp j i and (p ) = hp j i are momentum-space wave functions. Solve them to obtain Sx y z as functions of time. (b) What is the physical signi cance of 0 0 0 0 0 0 0 0 0 0 0 where x is the position operator and is some number with the dimension of momentum? Justify your answer. What is the corresponding eigenvalue? 1. mc Sz = !Sz write the Heisenberg equations of motion for the time-dependent operators Sx(t). QUANTUM DYNAMICS 7 is an eigenstate of the coordinate operator x.2 Let x(t) be the coordinate operator for a free particle in one dimension in the Heisenberg picture. and Sz (t). Using the Hamiltonian eB H = . Evaluate x(t) x(0)] : .

h~ rV i: x ~ dt x p m To identify the preceding relation with the quantum-mechanical analogue of the virial theorem it is essential that the left-hand side vanish.4 (a) Write down the wave function (in coordinate space) for the state ipa exp .3 Consider a particle in three dimensions whose Hamiltonian is given by 2 By calculating ~ p H ] obtain x ~ ~ H = 2pm + V (~ ): x * + d h~ ~i = p2 . known as the correlation function. 0 0 @x0 h m! !1=21 A: (b) Obtain a simple expression that the probability that the state is found in the ground state at t = 0.8 2. de ned by C (t) = hx(t)x(0)i where x(t) is the position operator in the Heisenberg picture. Does this probability change for t > 0? 2. 2 x 0 0 .h j0i: You may use 2 !3 x 25 1 hx j0i = 1=4x0 1=2 exp 4. Evaluate the correlation function explicitly for the ground state of a one-dimensional simple harmonic oscillator. . . Under what condition would this happen? 2.5 Consider a function.

Do the following algebraically. (b) Suppose the oscillator is in the state constructed in (a) at t = 0. 2. What is the state vector for t > 0 in the Schrodinger picture? Evaluate the expectation value hxi as a function of time for t > 0 using (i) the Schrodinger picture and (ii) the Heisenberg picture. (a) Construct a linear combination of j0i and j1i such that hxi is as large as possible. without using wave functions. in general. a complex number. that is. hence of E . . n=0 .2. (d) Show that a coherent state can also be obtained by applying the translation ( nite-displacement) operator e ipl=h (where p is the momentum operator. (b) Prove the minimum uncertainty relation for such a state. Find the most probable value of n. (c) Evaluate h( x)2i as a function of time using either picture.j j 2 y 1 Show that the distribution of jf (n)j2 with respect to n is of the Poisson form.6 Consider again a one-dimensional simple harmonic oscillator. QUANTUM DYNAMICS 9 2. (a) Prove that j i = e =2e a j0i is a normalized coherent state. (c) Write j i as X j i = f (n)jni: .7 A coherent state of a one-dimensional simple harmonic oscillator is de ned to be an eigenstate of the (non-Hermitian) annihilation operator a: aj i = j i where is. and l is the displacement distance) to the ground state.

and (x) the Dirac delta function. (Hint: directly apply the time-evolution operator.10 Derive the following expression i m! h Scl = 2 sin(!T ) (x2 + 2x2 ) cos(!T ) . (a) What is the potential? (b) Determine the constant A in terms of the parameters describing the system (such as m. r etc. h22m2 n2 n where (x) is the (unit) step function. x0xT 0 T for the classical action for a harmonic oscillator moving from the point x0 at t = 0 to the point xT at t = T . ).9 Prove the relation d (x) = (x) dx 0 dteiEt=hK (x y t 0) = A X sin(nrx) sin(nry) r E . 2. (Hint: study the e ect on testfunctions. 2.8 The quntum mechanical propagator.) 2.10 (e) Show that the coherent state j i remains coherent under timeevolution and calculate the time-evolved state j (t)i. 1 m!2x2 2 .11 The Lagrangian of the single harmonic oscillator is 1 _ L = 2 mx2 . moving in a potential is given by: K (x y E ) = Z 1 where A is a constant. for a particle with mass m.) 2.

1) 0 0 0 0 . (c) Show that Z dy1 : : : dyN exp(.] ih" (d) Let 2m N det det N pN . "2!2)pj . (Hint: Let y(t) = x(t) .1 7 where n = 6 . QUANTUM DYNAMICS 11 (a) Show that where Scl is the action along the classical path xcl from (xa ta) to (xb tb) and G is G(0 tb 0 ta) = Z lim dy1 : : : dyN 2 mh" N i !1 . xcl(t) being the solution of the Euler-Lagrange equation. 2 j =0 b ta where " = (tN +1) .nT n) G = Nlim 2 ih" 1 N 2 3 6 y. De ne j j matrices j that consist of the rst j rows and j columns of N and whose determinants are pj . xcl(t) be the new integration variable.2. By expanding j+1 in minors show the following recursion formula for the pj : pj+1 = (2 . hxbtbjxatai = exp iScl G(0 tb 0 ta) h (N +1) 2 8 N 9 <i X m 1 "m!2y2 = 2 exp : h j 2" (yj+1 .) (b) Show that G can be written as m (N2+1) Z dy : : : dy exp(. 5 yN !1 matrix . pj 1 j = 1 ::: N (2. Write the symmetric 4 .nT n) Z dN ye nT n . yj ) . =p N=2 Hint: Diagonalize by an orhogonal matrix. det . 7 and nT is its transpose.

12

(e) Let (t) = "pj for t = ta + j" and show that (2.1) implies that in the limit " ! 0 (t) satis es the equation
d2 = ;!2 (t) dt2 = with initial conditions (t = ta) = 0 d (tdt ta) = 1.

(f) Show that

( ) im! (x2 + x2) cos(!T ) ; 2x x ] m! exp hxbtbjxatai = 2 ih sin(!T ) a b 2h sin(!T ) b a where T = tb ; ta. s

2.12 Show the composition property
Z
dx1Kf (x2 t2 x1 t1)Kf (x1 t1 x0 t0) = Kf (x2 t2 x0 t0)

where Kf (x1 t1 x0 t0) is the free propagator (Sakurai 2.5.16), by explicitly performing the integral (i.e. do not use completeness). 2.13 (a) Verify the relation
i

! ihe " B j] = c ijk k

~ ~ x p where ~ m dt = ~ ; ecA and the relation

" !# x ~ ~ x d2~ = d~ = e E + 1 d~ B ; B d~ : x ~ m dt2 dt 2c dt dt
@ +r ~ =0 ~ j @t
0

(b) Verify the continuity equation

2. QUANTUM DYNAMICS

13

with ~ given by j

! h =( r ) ; e Aj j2: ~ ~ ~= j m mc
0

2.14 An electron moves in the presence of a uniform magnetic eld ~ in the z-direction (B = B z). ^ (a) Evaluate where
x x y]

(b) By comparing the Hamiltonian and the commutation relation obtained in (a) with those of the one-dimensional oscillator problem show how we can immediately write the energy eigenvalues as ! h2k2 + jeB jh n + 1 E =
2 where hk is the continuous eigenvalue of the pz operator and n is a
kn

px ; eAx c

y

py ; eAy : c

2m

mc

nonnegative integer including zero.

2.15 Consider a particle of mass m and charge q in an impenetrable cylinder with radius R and height a. Along the axis of the cylinder runs a thin, impenetrable solenoid carrying a magnetic ux . Calculate the ground state energy and wavefunction. 2.16 A particle in one dimension (;1 < x < 1) is subjected to a constant force derivable from
V= x
( > 0):

14

(a) Is the energy spectrum continuous or discrete? Write down an approximate expression for the energy eigenfunction speci ed by E. (b) Discuss brie y what changes are needed if V is replaced be V = jxj:

3 Theory of Angular Momentum
e i( + )=2 cos 2 ei( )=2 sin 2
; ;

3.1 Consider a sequence of Euler rotations represented by ! ;i 3 exp ;i 2 exp ;i 3 (1=2)( D ) = exp
2 =

Because of the group properties of rotations, we expect that this sequence of operations is equivalent to a single rotation about some axis by an angle . Find . 3.2 An angular-momentum eigenstate jj m = mmax = j i is rotated by an in nitesimal angle " about the y-axis. Without using the j explicit form of the d(m)m function, obtain an expression for the probability for the new rotated state to be found in the original state up to terms of order "2.
0

;e i( )=2 sin 2 ei( + )=2 cos 2
; ;

2

2 !

:

3.3 The wave function of a patricle subjected to a spherically symmetrical potential V (r) is given by
(~ ) = (x + y + 3z)f (r): x

THEORY OF ANGULAR MOMENTUM 15 ~ (a) Is an eigenfunction of L? If so. what is the l-value? If ~ not.3. what are the possible values of l we may obtain when L2 is measured? (b) What are the probabilities for the particle to be found in various ml states? (c) Suppose it is known somehow that (~ ) is an energy eigenfuncx tion with eigenvalue E .1) 3. 3. Indicate how we may nd V (r). 2 " ~ ~ ~ u~ = 1 .4 Consider a particle with an intrinsic angular momentum (or spin) of one unit of h. i h (L + S ) " (3. (a) Show explicitly that in nitesimal rotations of %i(~ ) are obtained x by acting with the operator ~ i^ ~ ~ where L = h r r. (One example of such a particle is the %meson). Determine S ! ~ ~ (b) Show that L and S commute. Quantum-mechanically. ~ (c) Show that S is a vector operator.5 We are to add angular momenta j1 = 1 and j2 = 1 to form j = 2 1 and 0 states. Using the ladder operator method express all . such a particle is described by a ketvector j%i or in ~ representation a wave function x %i(~ ) = h~ ij%i x x where j~ ii correspond to a particle at ~ with spin in the i:th dix x rection. ~ ~x ~ p~ (d) Show that r %(~ ) = h1 (S ~)% where p is the momentum oper~ ator.

1): 2 2 0 0 Hint: This can be proved in many ways. examine the rotational properties of Jz2 using the spherical (irreducible) tensor language. (j m2jdm)m( )j2 = 1 j (j + 1) sin + m 2 + 1 (3 cos2 . for instance. 3. (b) Prove. You may. xz. 1 1 jj = 1 m = 1i = p j+ 0i .6 (a) Construct a spherical tensor of rank 1 out of two di erent ~ ~ vectors U = (Ux Uy Uz ) and V = (Vx Vy Vz ). for any j .] 3. Explicitly write T (1)0 in 1 terms of Ux y z and Vx y z . y2) as components of a spherical (irreducible) tensor of rank 2.8 (a) Write xy. Write down explicitly T (2) 1 0 in terms of Ux y z 2 and Vx y z . p j0 +i : : : (3.7 (a) Evaluate j X m= j .2) 3. and (x2 . (b) Construct a spherical tensor of rank 2 out of two di erent ~ ~ vectors U and V . j) jd(mm ( )j2m 0 1 for any j (integer or half-integer) then check your answer for j = 2 . Write your answer as 2 2 where + and 0 stand for m1 2 = 1 0 respectively. .16 (nine) j m eigenkets in terms of jj1j2 m1m2i. j X m= j .

p that is.~ )? p p p p Justify your answer. (~ ). r2)j j m = j i is known as the quadrupole moment.1 (a) Assuming that the Hamiltonian is invariant under time reversal. 2 : : : )in terms of Q and appropriate ClebschGordan coe cients.~ ). (. y2)j j m = j i (where m = j j . 0 0 4. (~ ) = h~ j i.3 Read section 4. ika n k (x + a) = e n k (x) . prove that the wave function for a spinless nondegenerate system at any given instant of time can always be chosen to be real.2 Let (~ ) be the momentum-space wave function for state j i. You know that the energybands in solids are described by the so called Bloch functions n k full lling. Evaluate eh j m j(x2 . (b) The wave function for a plane-wave state at t = 0 is given by px a complex function ei~ ~=h.4. 1 j . or (.Is the momentum-space wave function for the p p time-reversed state j i given by (~ . Why does this not violate time-reversal invariance? 4. SYMMETRY IN QUANTUM MECHANICS 17 (b) The expectation value Q eh j m = j j(3z2 . 0 0 0 0 0 0 0 4 Symmetry in Quantum Mechanics 4.3 in Sakurai to refresh your knowledge of the quantum mechanics of periodic potentials.

5 The Hamiltonian for a spin 1 system is given by 2 H = ASz2 + B (Sx .4 Suppose a spinless particle is bound to a xed center by a potential V (~ ) so assymetrical that no energy level is degenerate. Prove that any Bloch function can be written as. Ri )e i where the sum is over all lattice vectors Ri. and the lattice momentum k is restricted to the Brillouin zone . x Using the time-reversal invariance prove ~ hLi = 0 for any energy eigenstate. The functions n are called Wannier functions. but the construction generalizes easily to three dimensions. 4. and are important in the tight-binding description of solids. Ri) n(x .18 where a is the lattice constant. Rj ) ij mn m Hint: Expand the n s in Bloch functions and use their orthonormality properties. =a =a]. Sy2): .). (In this simble one dimensional problem Ri = ia. Show that the Wannier functions are corresponding to di erent sites and/or di erent bands are orthogonal. X ikR n k (x) = n (x . n labels the band. i:e: prove Z dx ? (x . (This is known as quenching of orbital angular momemtum.) If the wave function of such a nondegenerate eigenstate is expanded as XX l m Ri Flm(r)Ylm ( ) what kind of phase restrictions do we obtain on Flm(r)? 4.

(A spin-dependent Hamiltonian of this kind actually appears in crystal physics. (c) Solve the H0 + V problem exactly.5. Find the zeroth-order energy eigenket and the corresponding energy to rst order that is the unperturbed energy obtained in (a) plus the rst-order energy shift] for each of the three lowest-lying states.1 Consider an isotropic harmonic oscillator in two dimensions. The Hamiltonian is given by (a) What are the energies of the three lowest-lying states? Is there any degeneracy? (b) We now apply a perturbation where is a dimensionless real number much smaller than unity. 5. APPROXIMATION METHODS 19 Solve this problem exactly to nd the normalized energy eigenstates and eigenvalues. p p + n ):] You may use hn jxjni = h=2m! ( n + 1 0 p2 + p2 + m!2 (x2 + y2) y x H0 = 2m 2m 2 V = m!2xy n n+1 0 n n 1 0 . Compare with the perturbation results obtained in q (b).) Is this Hamiltonian invariant under time reversal? How do the normalized eigenstates you obtained transform under time reversal? 5 Approximation Methods 5.2 A system that has three unperturbed states can be represented by the perturbed Hamiltonian matrix 0 1 E1 0 a B 0 E1 b C @ A a b E2 .

For t > 0. as a function of time. F (t) = F0e .i: (a) By solving the problem exactly. Use the second-order nondegenerate perturbation theory to calculate the perturbed eigenvalues. . j + . The quantities a and b are to be regarded as perturbations that are of the same order and are small compared with E2 . +i. j .4 Consider a composite system made up of two spin 2 objects. Compare the three results obtained. . Finally. for t < 0.i. use the second-order degenerate perturbation theory. 1 5. Find. (Is this procedure correct?) Then diagonalize the matrix to nd the exact eigenvalues.i for t 0. Show that the t ! 1 ( nite) limit of your expression is independent of time. 5. For t 0 it is subjected to a time-dependent but spatially uniform force (not potential!) in the x-direction. the Hamiltonian does not depend on spin and can be taken to be zero by suitably adjusting the energy scale.3 A one-dimensional harmonic oscillator is in its ground state for t < 0. j . obtain the probability of nding the oscillator in its rst excited state for t > 0). the probability for being found in each of the following states j + +i. the Hamiltonian is given by ~ ~ H = 4 2 S1 S2: h Suppose the system is in j + .20 where E2 > E1. E1. Is this reasonable or surprising? (b) Can we nd higher excited states? q You may use hn jxjni = h=2m! (pn + 1 n n+1 + pn n n 1 ):] 0 0 0 . t= (a) Using time-dependent perturbation theory to rst order.

l = 0) is subjected to a time-dependent potential as follows: V (~ t) = V0cos(kz . the nal wave function may be taken to be 1 x f (~ ) = 1 Z 2e x n=1 l=0 (~ ) = p a0 3 . obtain an expression for the transition rate at which the electron is emitted with momentum p. Under what condition does (b) give the correct results? 5. how you may compute the angular ~ distribution of the ejected electron (in terms of and de ned with respect to the z-axis). Discuss brie y the similarities and the di erences between this problem and the (more realistic) photoelectric e ect.5 The ground state of a hydrogen atom (n = 1. in particular. APPROXIMATION METHODS 21 (b) By solving the problem assuming the validity of rst-order time-dependent perturbation theory with H as a perturbation switched on at t = 0. Show. but you should be able to show that the observable e ects are independent of L. !t): x Using time-dependent perturbation theory.) L 3 2 px ei~ ~=h .5. Zr=a0 : with L very large. (note: For the initial wave function use If you have a normalization problem.

22 .

Part II Solutions 23 .

.

a )j i = (A . a ) 5 z }| { 4 4 ja iha j i a . a X = ha j i a a ja i = ha j ija i ) 000 0 00 000 000 000 00 X a =a 00 6 0 0 a 000 6 0 00 000 000 0 000 0 0 00 0 0 00 6 0 0 00 a: 0 (1. (a) Prove that Y (A . a 00 00 6 0 a 0 0 00 1 (c) Illustrate (a) and (b) using A set equal to Sz of a spin 2 system. a ) ? a =a a . a )ja i a =a 0: 00 0 00 00 ca 00 a 00 0 00 00 a 0 2f 0 g (1. a ) 0 0 is a null operator. a ) = ha j i ja i = a a =a a . a )ja i j } | 0 0 00 00 a 0 = X a 00 a 0 ca 00 Y a 0 a 00 (a . FUNDAMENTAL CONCEPTS 25 1 Fundamental Concepts 1.1 Consider a ket space spanned by the eigenkets fja ig of a Hermitian operator A.1. (b) What is the signi cance of Y (A . a ) 5 = ja iha j a =a a . 0 . a ) ja i ha {z i = ca (A . There is no degeneracy.a 00 00 00 00 000 000 a =a 00 6 0 0 a 2 3 4 Y (A . Then X Y Y Y X (A .1) (b) Again for an arbitrary state j i we will have 2 3 2 3 1 Y (A . (a) Assume that j i is an arbitrary state ket. a 000 Y (a .2) So it projects to the eigenket ja i.a j i = a . a ) 5 X Y (A .

j).4) 0 00 00 00 6 0 0 00 00 00 6 0 00 00 6 00 .ih.hj.hj.2 A spin 1 system is known to be in an eigenstate of S n with 2 eigenvalue h=2. j. where n is a unit vector lying in the xz-plane that ^ makes an angle with the positive z-axis. h (j+ih+j + j.2h=2 a =a a .i with eigenvalues h=2 and -h=2 respectively. ~ ^ 1.j) .j) + h (j+ih+j + j. a ) 2 a .j] hj+ih+j] = . j.ih. h 1 = = .j) = .j) = h 2 2 1 = h hj+ih+j = j+ih+j: (1. a " # 1 h (j+ih+j .h=2 . What is the probability of getting +h=2? .h 2 2 1 (.j = 1.ih.j) .ih. For a = h=2 we have Y (Sz . a ) Sz .ih. a 1) Sz + h 1 Y (Sz .ih.j) 2 2 0 0 0 0 Similarly for a = . (a) Suppose Sx is measured.5) 0 00 00 00 6 0 0 z }| { = .j: = .h=2 we have Y (Sz . j. a = a =h=2 h=2 .h2j. a ) = (Sz .h (1. a = h=2 + h=2 a =a " # 1 h (j+ih+j .i h.ih.j) = 2 2 " # h (j+ih+j .j+ih+j = 0 (1. a a = h=2 .3) where we have used that j+ih+j + j.ih. j.j.ih.ih. a 1) a "a # h (j+ih+j . This operator has eigenkets j+i and j. h (j+ih+j + j.ih.h=2 .ih. So Y Y (Sz .j) = j. a 1) Y (Sz .26 (c) It is Sz h=2(j+ih+j.j) + h (j+ih+j + j.ih.

it can be written in the following way n = ez cos + ex sin ^ ^ ^ (1. b cos = b 2 sin . FUNDAMENTAL CONCEPTS 27 (b) Evaluate the dispersion in Sx.7) we have that ! sin : ~ n = h cos S ^ 2 sin .ih.3. =2. = 0 ) 2 2 2 .4. cos = a 2 sin cos = a tan 2 : (1. h=2 sin ~ ^ det(S n . cos (1.6) So ~ ^ S n = Sz cos + Sx sin = # (S-1.ih+j) sin (1.9) The eigenvalues and eigenfuncions of this operator can be found if one solves the secular equation ! h=2 cos .h=2 cos .7) : = 2 2 ~ ^ Since the system is in an eigenstate of S n with eigenvalue h=2 it has to satisfay the following equation ~ ^~ ^ ~ ^ S njS n +i = h=2jS n +i: (1.18)] " # " h (j+ih+j . h(Sx .) Since the unit vector n makes an angle with the positive z-axis and is ^ lying in the xz-plane. and .j) cos + h (j+ih. that is.10) 4 4 4 2 ! a we will have that ~ ^ Since the system is in the eigenstate jS n +i b ! ! ! ( ) h cos sin a = h a ) a cos + b sin = a ) b a sin . hSx i)2i: (For your own peace of mind check your answers for the special cases = 0. h sin2 = 0 ) 2 . cos 2 b 2 sin2 2 b = a 1 .8) From (1. I ) = 0 ) det h=2 sin .1. j. h = 0 ) = h : (1. h cos2 + 2 .j + j.11) sin 2 2 .36).(S-1.

28

~ ^ But we want also the eigenstate jS n +i to be normalized, that is a2 + b2 = 1 ) a2 + a2 tan2 2 = 1 ) a2 cos2 2 a2 sin2 2 = cos2 2 r 2 = cos2 ) a = cos2 2 = cos 2 ) a (1.12) 2 where the real positive convention has been used in the last step. This means that the state in which the system is in, is given in terms of the eigenstates of the Sz operator by ~ ^ (1.13) jS n +i = cos 2 j+i + sin 2 j;i: (a) From (S-1.4.17) we know that 1 1 jSx +i = p2 j+i + p2 j;i: (1.14) So the propability of getting +h=2 when Sx is measured is given by ! 2 1 1 h+j + p h;j cos j+i + sin j;i 2 ~ ^ hSx +jS n +i = p2 2 2 2 2 1 1 = p cos 2 + p sin 2 2 2 1 cos2 + 1 sin2 + cos sin = 2 2 2 2 2 2 1 + 1 sin = 1 (1 + sin ): = 2 2 (1.15) 2
For = 0 which means that the system is in the jSz +i eigenstate we have
1 jhSx +jSz +ij2 = 2 (1) = 1 : 2

(1.16)

For = =2 which means that the system is in the jSx +i eigenstate we have

jhSx +jSx +ij2 = 1:
1 1 jhSx +jSz ;ij2 = 2 (1) = 2 :

(1.17) (1.18)

For = which means that the system is in the jSz ;i eigenstate we have

1. FUNDAMENTAL CONCEPTS
(b) We have that
2 h(Sx ; hSxi)2i = hSx i ; (hSx i)2:

29

(1.19)

As we know

Sx = h (j+ih;j + j;ih+j) ) 2 2 2 Sx = h (j+ih;j + j;ih+j) (j+ih;j + j;ih+j) ) 4 2 2 2 = h (j+ih+j + j;ih;j) = h : Sx {z } 4 4|
1

(1.20)

So

2 (hSxi)2 = h sin2 and 4 2 2 i = cos h+j + sin h;j h cos j+i + sin j;i hSx 2 2 4 2 2 h2 cos2 + sin2 ] = h2 : = 4 2 2 4 So substituting in (1.19) we will have

hSx i = cos 2 h+j + sin 2 h;j h (j+ih;j + j;ih+j) cos 2 j+i + sin 2 j;i 2 h cos sin + h sin cos = h sin ) = 2 2 2 2 2 2 2

(1.21)

h(Sx ; hSxi)2i = h (1 ; sin2 ) = h cos2 : 4 4
and nally

2

2

(1.22)

h( Sx)2i =0 Sz + = h 4 h( Sx)2i = =2 Sx + = 0 2 h( Sx)2i =0 Sz = h : 4
j i j i j ;i

2

(1.23) (1.24) (1.25)

30

1.3 (a) The simplest way to derive the Schwarz inequality goes as follows. First observe (h j + h j) (j i + j i) 0 for any complex number then choose in such a way that the preceding inequality reduces to the Schwarz inequility. (b) Show that the equility sign in the generalized uncertainty relation holds if the state in question satis es Aj i = B j i with purely imaginary. (c) Explicit calculations using the usual rules of wave mechanics show that the wave function for a Gaussian wave packet given by " 2# 2 ) 1=4 exp ihpix ; (x ; hxi) hx j i = (2 d
0 0 0 ;

h( h( p)2i = h : 2 Prove that the requirement hx j xj i = (imaginary number)hx j pj i is indeed satis ed for such a Gaussian wave packet, in agreement with (b).
x)2i
0 0

satis es the uncertainty relation
q

h

4d2

q

(a) We know that for an arbitrary state jci the following relation holds hcjci 0: (1.26) This means that if we choose jci = j i + j i where is a complex number, we will have (h j + h j) (j i + j i) 0 ) (1.27) 2 h j i 0: h j i+ h j i+ h j i+j j (1.28)

j i that is if j i and j i are colinear.36) with purely imaginary.29) (1.31) (1.1.h j i=h j i the previous relation will be 31 0 ) h j i.32) is 1 (1. + h j ih j i jh j ij2: j ih j j i h j ih j j i h i h i jh h j j ij i 2 (1. . (b) The uncertainty relation is To prove this relation we use the Schwarz inequality (1.33) On the other hand the right-hand side of (1.30) when Ajai = B jai: (1.30) Notice that the equality sign in the last relation holds when jci = j i + j i = 0 ) j i = . FUNDAMENTAL CONCEPTS If we now choose = .32) The equality sign in this relation holds according to (1.29) for the vectors j i = Ajai and j i = B jai which gives h( A)2ih( B )2i 1 jh A B ]ij2 : 4 h( A)2ih( B )2i jh A B ij2: (1.35) Thus the equality sign in the uncertainty relation holds when Ajai = B jai (1.34) jh A B ij2 = 4 jh A B ]ij2 + 1 jhf A B gij2 4 which means that the equality sign in the uncertainty relation (1.31) holds if 1 jhf A B gij2 = 0 ) hf A B gi = 0 4 :33) ) haj A B + B Ajai = 0 (1) haj( B )2jai + haj( B )2jai = 0 ) ( + )haj( B )2jai = 0: (1. h .

Evaluate the classical Poisson bracket (b) Let x and px be the corresponding quantum-mechanical operators this time.40) 0 0 0 0 0 1.39) So substituting in (1.4 (a) Let x and px be the coordinate and linear momentum in one dimension.i2d2 hx j pj i: hx j xj i = h (1. hxihx j i = (x . hpihx j i 0 0 0 0 0 (1.37) On the other hand hx j pj i 0 But hx j(p . hxi) hx j i .38) @ hx j i = hx j @x = hx j 0 0 0 " # @ ihpix . hxi) hx j i = 2id2 hx j xj i ) . hxi)2 i @x h 4d2 # " p i ihh i . prove that x exp iph a jx i 0 x F (px)]classical : x x exp iph a : (xjx i = x jx i) 0 0 0 .32 (c) We have hx j xj i 0 hx j(x . hxi)j i = x hx j i . hpihx j i h h = 2id2 (x . (x .38) we have 0 0 h hx j pj i = hpihx j i + 2id2 (x .ih @x hx j i . Evaluate the commutator (c) Using the result obtained in (b). hxi) d 0 0 0 0 0 0 0 (1. hpi)j i @ = . hxi)hx j i: 0 0 0 0 0 0 (1. 212 (x .

44) jx . FUNDAMENTAL CONCEPTS 33 is an eigenstate of the coordinate operator x. . @x @F (px) @x @px @px @x = @F (px) : @p x 1 (1. a exp iph a jx i x = (x .42) h n=1 1 1 . . a) exp iph a jx i: (1. . a exp iph a jx i x x = x exp iph a jx i . . So we can write x (1.a exp iph a : (1. ai = C exp iph a jx i where C is a constant which due to normalization can be taken to be 1.a) n (ih) h n=1 n! h n=1 k=0 n 1 X 1 x = . .43) x So exp iph a jx i is an eigenstate of the operator x with eigenvalue x . (c) We have now b x x x x exp iph a jx i (=) exp iph a xjx i . 0 0 0 0 0 0 0 0 0 0 0 0 . a. .41) (b) When x and px are treated as quantum-mechanical operators we have " # ipxa = x X (ia)n pn = X 1 (ia)n x pn] x x exp h n n x n=0 h n! n=0 n! h X X 1 (ia)n n 1 k px x px] pn k 1 = x n n=0 n! h k=0 X 1 (ia)n n 1 k n k 1 X n (ia)n 1 n 1 X = px px = n! n 1 px (. 1 . What is the corresponding eigenvalue? (a) We have x F (px)]classical @x @F (px) . 1)! ia px = .1. 1 . 1 .a (n .

0 0 .46) where we have used (1. 0 0 ip x h 0 0 hx j i 0 0 0 0 0 0 0 0 0 0 (1.5 (a) Prove the following: (i) hp jxj i = ih @ hp j i Z @p @ (ii) h jxj i = dp (p )ih @p 0 0 0 0 0 0 0 0 0 0 (p ) 0 where (p ) = hp j i and (p ) = hp j i are momentum-space wave functions.45) and that h jp i = (p ) and hp j i = (p ). . 0 0 0 0 . (a) We have (i) exp ix h hp jxj i = 0 = = = hp jxj i = 0 1 }| { Z hp jx dx jx ihx j i = dx hp jxjx ihx j i Z Z ip x dx x hp jx ihx j i (S 1:7:32) dx x Ae h hx j i = Z ip x @ Z @ A dx @p e h (ih)hx j i = ih @p dx Ae @ Z dx hp jx ihx j i = ih @ hp j i ) ih @p @p @ ih @p hp j i: 0 z Z 0 0 0 0 0 0 0 0 0 0 0 0 .34 1.45) (ii) @ h jxj i = dp h jp ihp jxj i = dp (p )ih @p (p ) 0 0 0 0 0 0 0 0 0 0 0 Z Z (1. (b) What is the physical signi cance of where x is the position operator and is some number with the dimension of momentum? Justify your answer. 0 0 0 0 .

49) h where A is a constant which due to normalization can be taken to be 1. This can h be justi ed by calculating the following operator " X # ix 1 ix n = X 1 i n p xn ] p exp h = p n! h n=0 n=0 n! h nX n X1 i n k k 1 = n! h k=1 x p x]x n=1 n n X 1 i nX X1 = (.ih)xn 1 = n! ih n(.48) p exp h h Thus we have that exp ix jp i Ajp + i (1. FUNDAMENTAL CONCEPTS 35 (b) The operator exp ix gives translation in momentum space. 1 .ih)xn 1 n=1 n! h n=1 k=1 n 1 X 1 X 1 ix n i = xn 1 (. 1 . 1 1 .1. p exp ix jp i ) exp h = p exp h h ix jp i = (p + ) exp ix jp i : (1. . So when this commutator acts on an eigenstate jp i of the momentum operator we will have p exp ix jp i = p exp ix jp i . .ih) ih = n=1 (n . 0 0 0 0 0 0 0 0 0 0 0 0 0 . 1)! h n=0 n! h = exp ix : (1.47) h 1 1 1 . exp ix p jp i ) h h h ix ix jp i .

4) x dt ih y ih y z = ih x dSz = 1 S H ] = 1 S !S ] (S 1:4:20) 0 ) S = constant: (2.!S (2.1 in Jackson. Using the Hamiltonian H = .6) y y y y y y .!2S ) S (t) = A cos !t + B sin !t ) S (0) = A : x x x 2 dt dt d2Sy = ! dSx (2:3) .! dSy (2=4) .3) and (2. Let us rst prove the following AS BS ] = CS ) AH BH ] = CH : (2. and Sz (t).2) The Heisenberg equation of motion gives dSx = 1 S H ] = 1 S !S ] (S 1:4:20) ! (.36 2. Sy (t). D! sin !t ) A=D C = .ihS ) = .!S ) y dt .A! sin !t + B! cos !t = .B: (2.4) we get d2Sx = .C! cos !t . 2 Quantum Dynamics .1 Consider the spin-procession problem discussed in section 2. Solve them to obtain Sx y z as functions of time.5) z dt ih z ih z z = Di erentiating once more eqs. eB Sz = !Sz mc write the Heisenberg equations of motion for the time-dependent operators Sx(t). (2.!2S ) S (t) = C cos !t + D sin !t ) S (0) = C: y y y dt2 dt = But on the other hand dSx = .3) y y dt ih x ih x z = ih dSy = 1 S H ] = 1 S !S ] (S 1:4:20) ! (ihS ) = !S (2. .1) Indeed we have h i AH BH ] = U AS U U BS U = U AS BS U . It can also be solved in the Heisenberg picture. U BS AS U = U AS BS ] U = U CS U = CH : (2. .

8) (2.2.12) Thus nally t ht t x(t) x(0)] = m p(0) + x(0) x(0) = m p(0) x(0)] = .2 Let x(t) be the coordinate operator for a free particle in one dimension in the Heisenberg picture.7) (2.13) 2. QUANTUM DYNAMICS So. Sy (0) sin !t Sy (t) = Sy (0) cos !t + Sx(0) sin !t Sz (t) = Sz (0): 2.10) This means that the Heisenberg equations of motion for the operators x and p will be " # @p(t) = 1 p(t) H (t)] = 1 p(t) p2(t) = 0 ) @t ih ih 2m p(t) = p(0) (2. im : (2.11) " 2 (t) # 1 (t :11) (0) @x(t) = 1 x H ] = 1 x(t) p = 2mih 2p(t)ih = pm) (2= pm ) @t ih ih 2m t x(t) = m p(0) + x(0): (2. nally 37 (2. Evaluate x(t) x(0)] : The Hamiltonian for a free particle in one dimension is given by 2 H = 2pm : (2.9) Sx(t) = Sx(0) cos !t .3 Consider a particle in three dimensions whose Hamiltonian is given by 2 p H = 2~m + V (~ ): x .

X X X n1 @ X = an (.16) i The right-hand side of (2. . h~ rV i: x ~ dt x p m To identify the preceding relation with the quantum-mechanical analogue of the virial theorem it is essential that the left-hand side vanish.38 By calculating ~ p H ] obtain x ~ * + d h~ ~i = p2 .15) m m The second commutator can be calculated if we Taylor expand the function V (~ ) in terms of xi which means that we take V (~ ) = Pn anxn with an x x i independent of xi.14) now becomes X ih X @ V (~ ) x p ~ ~ H] = ij pj pi + (.ih annxn 1 = .14) will give " 2# pj xi 2m = 21 xi p2] = 21 (pj xi pj ] + xi pj ]pj ) = 21 (pj ih ij + ih ij pj ) m j m m = 21 2ih ij pj = ih ij pj : (2.17) m~ . So " X # X X n1 X pi V (~ )] = pi anxn = an pi xn] = an xk pi xi] xn k 1 x i i i i 1 . ih~ rV (~ ): x ~ x (2. n=0 n n k=0 .ih)xn 1 = . . Under what condition would this happen? Let us rst calculate the commutator ~ ~ H ] x p 3 " # 2X 3 p ~2 + V (~ ) = 4 3 x p X p2 + V (~ )5 j x p ~ ~ H ] = ~ ~ 2m x p x x i i i=1 j =1 2m X " p2 # X j x (2. .14) = xi 2m pi + xi pi V (~ )] : i ij The rst commutator in (2.ih)xi @xi x ij m i = ih p2 .ih @x anxn i i i i n n n k=0 @ x = .ih @x V (~ ): (2. .

E hnj~ ~jni) = 0: n x p dt x ~ ih x p ih n x ~ So to have the quantum-mechanical analogue of the virial theorem we can take the expectation values with respect to a stationaru state. aj0i 0 0 = .2.h j0i ) 1 hx j t = 0i = hx exp .ipa j0i (Pr:=:4:c) hx . a 25 : 1=4x 1=2 exp 4. Indeed we have d hnj~ pjni = 1 hnj ~ ~ H ] jni = 1 (E hnj~ pjni . Does this probability change for t > 0? (a) We have ipa j t = 0i = exp . .h j0i: You may use 2 !3 x 25 hx j0i = 1=4x0 1=2 exp 4. 2.19) .18) where in the last step we used the fact that the state kets in the Heisenberg picture are independent of time. QUANTUM DYNAMICS The Heisenberg equation of motion gives d ~ ~ = 1 ~ ~ H ] (2= ~2 . h 2 !3 x . 0 (2. 1 x 2 0 0 0 .4 (a) Write down the wave function (in coordinate space) for the state ipa exp . ~ rV (~ ) ) :17) p ~ x x p x p dt ih + m x * d h~ pi = p2 . 1 0 2 x0 0 . h~ rV i x ~ dt x ~ m 39 (2. The left-hand side of the last equation vanishes for a stationary state. 0 @x0 h m! !1=21 A: (b) Obtain a simple expression that the probability that the state is found in the ground state at t = 0.

ipa j0ij2: h It is . de ned by C (t) = hx(t)x(0)i (2. 0 0 . Evaluate the correlation function explicitly for the ground state of a one-dimensional simple harmonic oscillator. iHt j t = 0ij2 h 2 = e iE t=hh0j t = 0i = jh0j t = 0ij2: (2.ipa j0i = Z dx h0jx ihx j exp . 1 x 0 2 0 2 !3 x . 0 2. 4x2 0 x 4x2 0 0 0 0 0 .ipa j0i hexp h 2h !3 Z x 25 1=4x 1=2 = dx 1=4x0 1=2 exp 4. 2x2 0 (2. 0 0 0 0 0 0 So 0 0 0 jh0j For t > 0 t = 0ij2 ! a2 : = exp .23) . . 2 x 2 . 1 x 2 + x 2 + a2 .22) jh0j tij2 = jh0jU (t)j t = 0ij2 = jh0j exp . a2 : (2. . . 1 x 2 0 " # Z 1=2x 1 exp .24) where x(t) is the position operator in the Heisenberg picture. a 25 exp 4. 2ax = dx 0 2 " 2x0 !# 1 Z dx exp . . 2x a + a2 + a2 = p x 2x2 2 4 4 0 0 ! ! 2 a p 1 p x = exp . known as the correlation function.21) = exp .40 (b) This probability is given by the expression (2. .20) jh0j t = 0ij2 = jhexp .5 Consider a function.

QUANTUM DYNAMICS 41 But on the other hand from (2.26) = 22hm p(t) = pm) i # " dp(t) = 1 p(t) H ] = 1 p(t) p2 (t) + 1 m!2x2(t) dt ih ih 2m 2 m!2 hp(t) x2(t)i = m!2 .29) (2.2.!2p(t) ) p(t) = C cos !t + D sin !t ) p(0) = C: :26) 2 dt m dt (2.27) we get d2x(t) = 1 dp(t) (2= .2ihx(t)] = .25) 2 So the Heisenberg equations of motion will give " # dx(t) = 1 x(t) H ] = 1 x(t) p2(t) + 1 m!2x2(t) dt ih ih 2m 2 i 1 h i 1 h = 2mih x(t) p2(t) + 2 m!2 i1 x(t) x2(t) h (t ih (2.30) .26) we have dx(t) = p(t) ) dt m p(0) cos !t + D sin !t ) .m!x(0): So x(t) = x(0) cos !t + p(0) sin !t m! and the correlation function will be 1 :29) C (t) = hx(t)x(0)i (2= hx2(0)i cos !t + hp(0)x(0)i m! sin !t: The Hamiltonian for a one-dimensional harmonic oscillator is given by 2 (t H = p2m) + 1 m!2x2(t): (2.27) = 2ih 2 ih Di erentiating once more the equations (2.!2x(t) ) x(t) = A cos !t + B sin !t ) x(0) = A :27) 2 dt m dt d2p(t) = 1 dx(t) (2= .26) and (2.m!2x(t): (2.28) (2.!x(0) sin !t + B! cos !t = m m p(0) B = m! D = .

i h h0jaa j0i = .6 Consider a one-dimensional simple harmonic oscillator.31) h hx2(0)i = h0j 2m! (a + a )(a + a )j0i = 2m! h0jaa j0i = 2m! s s mh! h h0j(a .33) y y y y y y .34) We can write the constands c0 and c1 in the following form c0 = jc0jei 0 q :34) i 1 (2= ei 1 1 . Do the following algebraically. (a) We want to nd a state j i = c0j0i + c1j1i such that hxi is as large as possible. This means q jc0j2 + jc1j2 = 1 ) jc1j = 1 . (a) Construct a linear combination of j0i and j1i such that hxi is as large as possible. What is the state vector for t > 0 in the Schrodinger picture? Evaluate the expectation value hxi as a function of time for t > 0 using (i) the Schrodinger picture and (ii) the Heisenberg picture. The state j i should be normalized.42 Since we are interested in the ground state the expectation values appearing in the last relation will be h h (2. (b) Suppose the oscillator is in the state constructed in (a) at t = 0. that is.32) = . jc0j2: (2. a)(a + a )j0i hp(0)x(0)i = i 2 2m! (2. jc j2 : c1 = jc1je 0 (2. i 2m! sin !t = 2m! e i!t: (2.i h : 2 2 Thus h h h C (t) = 2m! cos !t .35) . without using wave functions. (c) Evaluate h( x)2i as a function of time using either picture. 2.

jc j2 = 0 0 = 0 0 0 @ jc0j 1 .41) s y y y j j6 hxi = h jxj i = (c0h0j + c1h1j) x (c0j0i + c1j1i) = jc0j2h0jxj0i + c0c1h0jxj1i + c1csh1jxj0i + jc1j2h1jxj1i 0 s h h = jc0j2 2m! h0ja + a j0i + c0c1 2m! h0ja + a j1i y y . jc j2 . QUANTUM DYNAMICS by 43 The average hxi in a one-dimensional simple harmonic oscillator is given s h h1ja + a j0i + jc j2 h h1ja + a j1i +c1c0 2m! 1 2m! s s h h = 2m! (c0c1 + c1c0) = 2 2m! <(c0c1) s h cos( .2.40) We can always take 0 = 0. )jc jq1 . jc j2 . ) = 0 ) = + n n 2 Z : (2.37) 2 @ hxi = 0 ) .39) 2 @ 1 1= 1max < 0 ) n = 2k k 2 Z : So we can write that 1 1 1 j i = ei 0 p2 j0i + ei( 0+2k ) p2 j1i = ei 0 p2 (j0i + j1i): (2. What we need is to nd the values of jc0j and 1 . jc0j2 1 ) jc0j = p (2.36) 0 1 0 0 q h where we have used that x = 2m! (a + a ). sin( . jc j2 = 2 2m! (2. Thus 1 j i = p2 (j0i + j1i): (2. @ hxi = 0 ) q1 . 0 that make the average hxi as large as possible. q jc0j2 c) 1 1 .38) 1 0 1 0 @1 But for hxi maximum we want also @ 2hxi (2.

.29) that xH (t) = x(0) cos !t + p(0) sin !t: m! So hxiH = h jxH j i !" " # # 1 1 h0j + p h1j x(0) cos !t + p(0) sin !t p j0i + p j1i 1 1 = p m! 2 2 2 2 1 1 = 2 cos !th0jxj1i + 1 cos !th1jxj0i + 1 m! sin !th0jpj1i 2 2 1 1 + 2 m! sin !th1jpj0i s s s h cos !t + 1 h cos !t + 1 sin !t(. . . .i) mh! 1 = 2 2m! 2 2m! 2m! 2 s h! 1 + 2m! sin !ti m2 s h = 2m! cos !t: (2.43) = 2 e i!t 2m! 2 2m! 2m! (ii) In the Heisenberg picture we have from (2. . j t0 ti = U (t t0 = 0)j t0i = e .44) . iE1 t=hj1i (c) It is known that h( x)2i = hx2i .45) .42) 2 2 (i) In the Schrodinger picture hxiS = h t0 tjxS j t0 tiS " # " # 1 ei!t=2h0j + ei!3t=2h1j x p e i!t=2j0i + e i!3t=2j1i 1 = p 2 2 1 = 2 ei(!t=2 !3t=2)h0jxj1i + 1 ei(!3t=2 !t=2)h1jxj0i s s s2 h + 1 ei!t h = h cos !t: 1 (2. . iE0 t=hj0i + 1 pe 2 . . hxi2 (2. So 1 1 = p e i!t=2j0i + e i!3t=2j1i = p e i!t=2 j0i + e i!tj1i :(2. . iHt=hj 1 i = p2 e .44 (b) We have j t0i = j i.

2mh!2 (a2 + a 2 .47) h h :43) h h( x)2iS (2= 2m! .2. a)(a + a ) sin2 !t 4m! h (a2 + a 2 + aa + a a) cos2 !t = 2m! y y y y y y y y y y y y y . QUANTUM DYNAMICS In the Schodinger picture we have 32 2s h (a + a )5 = h (a2 + a 2 + aa + a a) 2 = 4 x 2m! 2m! y y y y 45 (2. 2m! cos2 !t = 2m! sin2 !t: (2.48) In the Heisenberg picture " # p(0) sin !t 2 x2 (t) = x(0) cos !t + m! H p2 = x2(0) cos2 !t + m2(0) sin2 !t !2 + x(0)p(0) cos !t sin !t + p(0)x(0) cos !t sin !t m! m! h = 2m! (a2 + a 2 + aa + a a) cos2 !t . aa . a a) sin2 !t m2s ! i 2 + m! hmh! (a + a )(a . y y # h1 i(!t=2 2e . i h h + 2 1 + 1 2m! = 2m! : 2 2 2 !t=2)h0jaa y j0i + 1 ei(!3t=2 2 . a) sin2 !t s 4m! i 2 + m! hmh! (a .46) which means that hxi2 = h t0 tjx2j t0 tiS S " # " 1 ei!t=2h0j + ei!3t=2h1j x2 p e 1 = p 2 2 = = So . h1 i!t=2 j0i + e i!3t=2j1i . !3t=2)h1jaa i h 1 j1i + 2 h1ja aj1i 2m! (2.

. a2) sin 2!t y y y y (2. a2) sin 2!t y y i h h = 4m! h0jaa j0i + h1jaa j1i + h1ja aj1i h h = 4m! 1 + 2 + 1] = m! : y y y (2. a complex number. in general. 2m! cos2 !t = 2m! sin2 !t: 2. (b) Prove the minimum uncertainty relation for such a state. a2) sin 2!t y y y y h ih .50) (2.7 A coherent state of a one-dimensional simple harmonic oscillator is de ned to be an eigenstate of the (non-Hermitian) annihilation operator a: aj i = j i where is. 2m! (a2 + a 2 . a a) sin2 !t + 2m! (a 2 .49) which means that hx2 iH = h jx2 " iH H Hj # 1 h p h0j + p h1j 1 = 2m! 2 2 " # 1 j0i + p j1i 1 p 2 2 y y h i aa + a a + a2 cos 2!t + a 2 cos 2!t + i(a 2 . aa .51) So h h :44) h h( x)2iH (2= 2m! .j j 2 y .46 h h h = 2m! (aa + a a) + 2m! a2 cos 2!t + 2m! a 2 cos 2!t ih + 2m! (a 2 . (a) Prove that j i = e =2e a j0i is a normalized coherent state.

52) aj i = e .j j 2 h0je ae a j0i y .j j 2 =2 e a j0i = j i y (2. Indeed h j i = h0je ae . QUANTUM DYNAMICS 47 (c) Write j i as j i= X 1 Show that the distribution of jf (n)j2 with respect to n is of the Poisson form. n=0 f (n)jni: (a) We have aj i = e .j j 2 =2 h i a e a j0i y (2. The commutator is # " X h ai 1 ( a )n = X 1 n ha (a )ni ae = a n! n=0 n! n=0 n n h i n k X 1 nX n 1 X 1 nX k 1 = (a ) a a (a ) = n! (a ) n=1 n! k=1 n=1 k=1 X1 X 1 n n1 n a (a ) = (2. 1 1 y . hence of E . y .53) = (n .54) which means that j i is a coherent state. (d) Show that a coherent state can also be obtained by applying the translation ( nite-displacement) operator e ipl=h (where p is the momentum operator. 1)! n! ( a ) = e : n=0 n=1 y 1 y y 1 1 y . y y . it should satisfy also h j i = 1. (e) Show that the coherent state j i remains coherent under timeevolution and calculate the time-evolved state j (t)i. Find the most probable value of n.52) since aj0i = 0. If it is normalized.) . y y So from (2.j j 2 =2ae a y j0i = e 1 . (Hint: directly apply the time-evolution operator.j j 2 e a j0i = e y .2. and l is the displacement distance) to the ground state.

3) the state should satisfy the following relation xj i = c pj i (2.j j .j j 2 y y . h jpj i and c is a purely imaginary number. )j i y y y y s (2. hxi)j i = 2m! (a . a)j i = i m2 (a . a) we have s p s mh! h! pj i = i 2 (a .61) .j j .56) where x x . Since j i is a coherent state we have aj i = j i ) h ja = h j : y (2.58) and s h h j(a + a )j i = h (h jaj i + h ja j i) hxi = h jxj i = 2m! 2m! s h = 2m! ( + ) (2. p p .57) Using this relation we can write s s h (a + a )j i = h ( + a )j i xj i = 2m! 2m! y y (2.59) y y s and so h xj i = (x .55) = e .j j j j (b) According to problem (1. h jxj i.48 p X 1 ( )n mh0jan j(a )m j0i (a )mj0i = m!jmi] n m n!m! p p n! m! ( )n mhnjmi = e 2 X 1 (j j2)n 2 X = e n m n! m! n n! 2 2 = e e = 1: (2. )j i: q h! Similarly for the momentum p = i m2 (a .60) (2.

2.i) 2 pj i = xj i = 2m! mh! pj i (2.j j2) (2.66) m! m!hnjmi = e n! m=0 2 n jf (n)j2 = (j nj! ) exp(. . )j i ) s (2. h jaj i) hpi = h jpj i = i 2 2 s h! = i m2 ( . )j i = .i m2 pj i: h! y y y y s So using the last relation in (2.j j y 1 1 . m! n=0 n=0 for the expansion coe cients f (n) we have f (n) = hnj i = hnje 2=2e a j0i = e 2=2hnje a j0i X 1 X 1 = e 2=2hnj m! ( a )mj0i = e 2=2 m! mhnj(a )mj0i m=0 m=0 p 2 =2 X 1 2 =2 1 m p n) = e (2.67) which means that the distribution of jf (n)j2 with respect to n is of the Poisson type about some mean value n = j j2. ) (2. a)j i = i mh! (h ja j i .j j y .j j y 1 . QUANTUM DYNAMICS and 49 s mh! h j(a . hpi)j i = i m2 (a .65) 1 1 i .j j y .60) s s h (.j j . (c) The coherent state can be expressed as a superposition of energy eigenstates X X j i = jnihnj i = f (n)jni: (2.62) and so s h! pj i = (p .j j .64) | {z } purely imaginary and thus the minimum uncertainty condition is satis ed.63) (a .

j j ) ( 2 )n exp(. . (d) We should check if the state exp (. We have h a exp (.il !n X k 1 ae = a p ] = n! h p a p]pn k n! h n=0 n=1 k=1 s !n X n X 1 .il !n n X 1 .il h! pn 1 i m2 = n=1 n! h s !k=1 mh! .50 The most probable value of n is given by the maximum of the distribution jf (n)j2 which can be found in the following way j j2 1 = n+1 (2. a] = i mh! : a p] = i 2 (2.68) which means that the most probable value of n is j j2. (2.il X 1 .ipl=h) j0i] = l 2h exp (.70) in (2.j j2 ) n! nihilation operator a. For the commutator in the last relation we have n h ( ipl=h)i X 1 . . .72) which q means that the state exp (. ipl=h) i j0i .69) . .69) we get r m! a exp (.71) 2 So substituting (2.ipl=h) j0i] (2. 2h (e) Using the hint we have X :66) j (t)i = U (t)j i = e iHt=hj i (2= e iHt=h e 2=2 p1n! n jni n=0 1 1 .ipl=h) j0i is a coherent state with eigenvalue l m! .j j . 1 .ipl=h) j0i is an eigenstate of the anjf (n + 1)j2 = jf (n)j2 j j j j ( 2 )n+1 2 (n+1)! exp(.ilp !n 1 = i 2 h n=1 (n . 1)! h r m! (2.ipl=h) j0i = a e( 1 . since aj0i = 0. y 1 .70) = l 2h e( ipl=h) where we have used that s s mh! a a .

i!t e i!t=2j . i!t i (2. .2.8 The quntum mechanical propagator. . X 1 n=0 e . . ).j X 1 . QUANTUM DYNAMICS = = = Thus 51 1 . h22m2 n2 n K (x y E ) = = Z Z0 1 Z0 dteiEt=hK (x . moving in a potential is given by: K (x y E ) = Z 1 0 dteiEt=hK (x y t 0) = A where A is a constant.73) aj (t)i = e i!t=2aj e i!ti = e = e i!t j (t)i: .j j e i!t 2 =2 ( . y t 0) .74) 2. We have X sin(nrx) sin(nry) r E . Z 1 1 X dteiEt=h hxje iHt=hjnihnjyi n Z0 iEt=h X e iEn t=hhxjnihnjy i dte = 0 nZ X ei(E En)t=hdt = n (x) n (y ) 1 1 . for a particle with mass m. i!t i (2. (a) What is the potential? (b) Determine the constant A in terms of the parameters describing the system (such as m. iEn t=he . :3 n jni (2=:9) X e hit h!(n+ 1 ) e 2 n! n=0 1 i!t=2e 2 =2 p n jni j 2 1 =2 p . j e :66) p jni (2= e n! .j . . n=0 X e i!t=2 e n=0 1 e . i!t )n n! . i!t n e . 1 . i!t=2j e e . r etc. dteiEt=hhxje iHt=hjyi 0 dteiEt=hhx tjy 0i n 0 .j j 2 1 =2 p n! n jni .

9 Prove the relation where (x) is the (unit) step function. En +i")t=h # 1 0 (2.76) we get L = r ) L = r and ( 0 V = 1 for 0 < x < r otherwise while (2. and (x) the Dirac delta function.1 1 .78) (2. h22m2 n2 n n s 2 r2 (x) = iA sin(nrx) En = h m n2: (2.ih ei(E n (x) n (y ) lim " 0 E . En + i" ! " . En ih =A . En Comparing with (2. 1 1 .77) n (x) = L L X n (x) n (y ) E .1 .1 (x (x) dfdx ) dx . (Hint: study the e ect on testfunctions.79) A = 2r ) A = i 2hr : ih d (x) = (x) dx 2.76) n h 2 For a one dimensional in nite square well potential with size L the energy eigenvalue En and eigenfunctions n(x) are given by s 2 2 sin n x h2 En = 2m L n2: (2.52 = = So X X n n ih : n (x) n (y ) E .) For an arbitrary test function f (x) we have Z + d (x) Z+ d Z+ dx f (x)dx = dx (x)f (x)] dx .75) X sin(nrx) sin(nry) ) r E .

84) x(t = T ) = xT = x0 cos !T + B sin !T ) B sin !T = xT .2. 1 1 . This can be solved to give x(t) = A cos !t + B sin !t with boundary conditions x(t = 0) = x0 = A (2.m!2x .1 d (x) = (x): dx Z + df (x) dx dx 0 + = x lim f (x) .81) (2.80) 2.1 (2. x0 cos !T ) cos (2.83) which is the equation of motion for the system. d @ L = 0 (2) .10 Derive the following expression for the classical action for a harmonic oscillator moving from the point x0 at t = 0 to the point xT at t = T .85) B = xT . QUANTUM DYNAMICS = (x)f (x) ! 1 1 53 + 1 . x0!T !T : sin . f (x) = f (0) 0 Z ++ = (x)f (x)dx ) . d (mx) = 0 ) :81) @x dt @ x _ dt _ x + ! 2 x = 0: (2. The Lagrangian for the one dimensional harmonic oscillator is given by i m! h Scl = 2 sin(!T ) (x2 + 2x2 ) cos(!T ) .82) (2. 2 m!2x2: _ 1 _ 1 From the Lagrange equation we have @ L . x0xT 0 T L(x x) = 2 mx2 .

x x + x2 cos !T i = 0 T 0 T 0 2 sin !T h T m! (x2 + x2) cos !T .87) sin With these at hand we have ZT ZT 1 _ S = dtL(x x) = dt 2 mx2 .54 So cos x(t) = x0 cos !t + xT . x0 cos !T ) 2 sin !T sin !T m! hx2 cos !T . 1 m! 2 x2 = dt 2 dt _ 2 2 0 ZT T 1 _ = . x0!Tcos !(T . 2x x i : = 2 sin !T T 0 (2.88) 0 T 2.86) = xT sin !t + x0!T !(T . x(0)x(0)] x _ = m xT ! (xT cos !T . 1 m!2x2 _ 2 0 0 # ZT " d 1 m (xx) . 2 m dtx x + !2x] + m xx 2 0 0 (2:82) m = 2 x(T )_ (T ) . x0! (xT . 1 mxx . x0 cos !T sin !t = sin !T sin (2. x0) .11 The Lagrangian of the single harmonic oscillator is L = 1 mx2 . x x . t) ) sin ! x(t) = xT ! cos !t . t) : _ (2. x0!T !T sin !t sin x0 cos !t sin !T + xT sin !t . 1 m!2x2 2 _ 2 (a) Show that hxb tbjxatai = exp iScl G(0 tb 0 ta) h where Scl is the action along the classical path xcl from (xa ta) to (xb tb) and G is G(0 tb 0 ta) = .

Write the symmetric 4 . m 2 ih" 8 N 9 <i X m 1 "m!2y2 = 2 exp : h j 2" (yj+1 . 7 and nT is its transpose. yj ) . 2 j =0 b ta where " = (tN +1) . "2!2)pj .nT n) G = Nlim 2 ih" 1 N 2 3 6 y. xcl(t) being the solution of the Euler-Lagrange equation.89) implies that in the limit " ! 0 (t) satis es the equation d2 = . QUANTUM DYNAMICS N 55 (N +1) 2 lim Z !1 dy1 : : : dyN . (c) Show that Z dy1 : : : dyN exp(. xcl(t) be the new integration variable.2. . pj 1 j = 1 ::: N (2. Hint: Let y(t) = x(t) .] (b) Show that G can be written as m (N2+1) Z dy : : : dy exp(.1 7 where n = 6 .!2 (t) dt2 with initial conditions (t = ta) = 0 d (t=ta) dt = 1. By expanding j+1 in minors show the following recursion formula for the pj : pj+1 = (2 . De ne j j matrices j that consist of the rst j rows and j columns of N and whose determinants are pj .nT n) Z dN ne .] ih" (d) Let 2m N det det N pN .89) 0 0 0 0 . (e) Let (t) = "pj for t = ta + j" and show that (2. nT n =p det N=2 Hint: Diagonalize by an orthogonal matrix. 5 yN !1 matrix .

. .56 (f) Show that ( ) m! exp im! (x2 + x2) cos(!T ) . . . . . . . Z It is = e For the second term in this last equation we have Z " p2 " p2 i h 2m jx i = hxi+1 je dpi hxi+1je i h 2m jpiihpi jxii i Z " p2 i = dpi e i h 2m hxi+1jpi ihpi jxii Z "p2 = 2 1h dpi e i h 2im eipi(xi+1 xi )=h h " 1 Z dp e i 2mh p2 2pi m (xi+1 xi)+ m22 (xi+1 xi )2 m22 (xi+1 i " " " = 2 h i i" " 1 e 2mh m22 (xi+1 xi)2 Z dp e i 2mh pi pi m (xi+1 xi)]2 " " = 2 h i . . hxi+1ti+1jxitii = hxi+1 je = hxi+1 je = hxi+1 je iH (ti+1 ti )=hjx i = hx je iH"=hjx i i i+1 i " ( 1 mp2 + 1 m!2 x2 ) ih 2 2 jxii (since " is very small) 2 1 m!2 x2 " " i h 2pm e i h 2 jxii " 1 m!2 x2 " p2 ih 2 i hxi+1 je i h 2m jxi i: (2. s (a) Because at any given point the position kets in the Heisenberg picture form a complete set.92) . . xi )2 i . . . . . . . 2x x ] hxbtbjxatai = 2 ih sin(!T ) a b 2h sin(!T ) b a where T = tb . .91) . . it is legitimate to insert the identity operator written as Z dxjxtihxtj = 1 (2. . . . .90) So !1 hxbtbjxatai = Nlim dx1dx2 : : : dxN hxb tbjxN tN ihxN tN jxN 1tN 1i : : : hxi+1ti+1jxitii : : : hx1t1jxatai: (2. ta.

1 m! 2 y 2 dt: y+ = Scl + @ x _ 2 _ ta ta @x dt @ x xcl _ ta 2 So Z i i iZt h 1 hxb tbjxatai = Dy exp h Scl + h t b 1 my_ 2 .2. x )2 .92) we get . 2 m!2y2 dt a 2 = exp iScl G(0 tb 0 ta) (2. .93) m 1 e hi 2m" (xi+1 xi )2 1 "m!2 xi] 2 hxi+1ti+1jxitii = 2 ih" 2 (2. yj ) .95) h (N +1) 2 9 8 N <i X m 1 "m!2y2 = : 2 exp : h j 2" (yj+1 . d @ L 1 my 2 . . 57 (2. . 2 j =0 . QUANTUM DYNAMICS s i" 1 e 2mh m22 (xi+1 xi )2 2hm " = 2 h i" r m im = 2 hi" e 2h" (xi+1 xi)2 : Substituting this in (2.94) and this into (2. 1 "m!2x2 = : lim dx1 : : :dxN j : h j=0 2" j+1 j N 2 ih" 2 !1 with G(0 tb 0 ta) = Z lim dy1 : : : dyN 2 mh" N i !1 Let y(t) = x(t) . xcl(t) ) x(t) = y(t) + xcl(t) ) x(t) = y(t) + xcl(t) with _ _ _ boundary conditions y(ta) = y(tb) = 0: For this new integration variable we have Dx = Dy and Z tb S x] = S y + xcl] = L(y + xcl y + xcl)dt _ _ ta 3 Z tb 2 @ L y + @ L y + 1 @ 2L y 2 + 1 @ 2L y 25 4L(xcl xcl) + = _ 2 @ x2 @x xcl @ x xcl _ 2 @ 2x xcl _ _ xcl _ ta " !# Z tb h i @ L y tb + Z tb @ L .91): Z i hxbtbjxatai Dx exp h S x] = 8 N 9 Z m (N2+1) exp < i X m (x .

.98) .7 .1 2 . 2"ih (2yi ij yj . yi i+1 j yj ) . Thus we have m (N2+1) Z dy : : : dy exp(. i X 1 "m!2y y (yN +1 =0) = i ij j h j=0 2" j+1 j j+1 j j j+1 h i j=1 2 N N 2 X m X . y )2 . . .1 2 : : : 0 0 7 7 6 6 7 i"m!2 6 0 0 1 : : : 0 0 7 7 m 6 7+ 6 = 2"ih 6 . y y .7 6 .97) G = Nlim 2 ih" 1 N with 2 3 2 3 2 .7 . = p N=2 i=1 i det (2.. . . So = RT D R and det R = 1 (2. . 7: . . . .99) So we can diagonalize by an orthogonal matrix R.7 6 6 7 6 0 0 0 : : : 2 . y y ) .1 5 6 0 0 0 ::: 1 0 7 4 4 5 0 0 0 : : : .. . i"m! 2h i j=1 yi ij yj :(2. .96) i j =1 where the last step is written in such a form so that the matrix will be symmetric.101) .1 0 : : : 0 0 1 0 0 ::: 0 0 6 .1 : : : 0 0 7 6 0 1 0 ::: 0 0 7 6 6 7 6 0 . 7 2h 6 . 1 "m!2y2 (y0= =0) j +1 j j h j=0 2" 2 N N i X m (y2 + y2 . 6. . . . .. (2.. . yi i j+1yj . . .nT n) (2. .1 2 0 0 0 ::: 0 1 !1 (c) We can diagonalize by a unitary matrix U . Since is symmetric the following will hold = U DU ) T = U T D(U )T = U T D U = ) U = U : (2..100) whichZmeans that Z Z N ne nT n = N ne nT RT Rn Rn= d d = dN e T Z Z Z 2 2 2 1 a1 2 a2 : : : = d 1e d 2e d N e N aN s s s N=2 N=2 qQN = p = a a ::: a = det D 1 2 N a y y . .58 (b) For the argument of the exponential in the last relation we have N i X m (y . .7 .

"2!2)"pj . 7 det j+1 = det 6 . 7> .1 : : : 0 0 0 3 6 .1 0 7 7 6 6 0 2!2 5 4 0 : : : . (e) We have (t) (ta + j") ) (ta + (j + 1)") = = ) (t + ") = "pj "pj+1 = (2 . 6 .103) with p0 = 1 and p1 = 2 . . " . . 7> : : : 1 0 7> 5> ::: 0 1 (2. (ta + (j .1 2 : : : 0 0 7 <6 7 . 7> . "2!2 6 0 6 det >6 . . . . 6 . . . 6 . " From the above it is obvious that 0 det j +1 pj+1 0 = (2 . (2.1 2 0 det N pN : 0 0 0 59 0 1 0 . . "2!2 (ta + j") .1 7 >4 4 5 > : 0 0 0 : : : .. .1 7 2! 2 0 0 ::: 0 . . .. pj 1 for j = 2 3 : : : N 0 0 .7 .104) . . 0 0 39 ::: 0 0 > : : : 0 0 7> 7> > 7= : : : 0 0 7> 7 = .1 2 .1 2 .7 6 >6 60 >6 0 0 0 : : : 2 .102) We de ne j j matrices j that consist of the rst j rows and j columns of N . .1 2 .. So 2 2 .2. . "): (2. . . . .1 : : : 0 0 0 7 6 . 7: . (t .98) we have ! 2ih" N det = m 82 21 > 2 .. " . QUANTUM DYNAMICS where ai are the diagonal elements of the matrix D . . "2!2) det j .1 : : : 0 0 7 60 >6 7 6 >6 7 6 >6 0 . "2!2)pj . . "2!2 (t) . . . det j 1 ) = (2 . "pj 1 2 (ta + j") . .. "2!2 . "2!2.. 1)") 2 (t) . .7 6 . "2!2 : : : 7 0 0 0 7 6 7 6 0 7 6 . 0 0 0 0 1 . . (d) From (2.1 2 .1 0 : : : 0 0 3 >6 . . 7 6 6 0 2!2 0 ::: 2 . .

0 ! (2.60 So (t + ") .107) we have (ta) = 0 ) A sin(!ta + ) = 0 ) = .108) (2.!2 (t) ) 2 lim (t) . ta).!ta + n n 2 Z which gives that (t) = A sin !(t .!2 (t) ) d 2 = . " (t .111) t (t) = sin !(! . p 1 0 dt a " " = 2 .106) (ta) = "p0 ! 0 Thus d (t ) = (ta + ") .106) and (2. (t . ") .105) (2. . " (t) (t ") " . 1 ! 1: The general solution to (2.109) (2. "2!2 . (t) = (t) . ta) : (f) Gathering all the previous results together we get # " m (N +1) p N 1=2 G = Nlim 2 ih" det !1 .!2 (t): " 0 dt From (c) we have also that 0 (t+") (t) " . (ta) = "(p1 . p0) = p . t ) ) (t ) = A! (2) :107) a a dt 1 A! = 1 ) A = ! 0 and (2. .110) (2. ") = .107) (2. "2!2 (t) ) = .105) is (t) = A sin(!t + ) and from the boundary conditions (2. while d = A! cos(t .

1=2 (2. We have Z dx1Kf (x2 t2 x1 t1)Kf (x1 t1 x0 t0) " # Z s m im(x2 . x0) 2 ih(t1 .112) So from (a) hxbtbjxatai = exp iScl G(0 tb 0 ta) s h (2:88) = m! exp im! h(x2 + x2 ) cos !T . t0 s 2 2 1 = 2 mh (t . QUANTUM DYNAMICS = (d) 61 . 1 2" 1 1 0 # Z im x2 + im x2 .12 Show the composition property Z dx1Kf (x2 t2 x1 t1)Kf (x1 t1 x0 t0) = Kf (x2 t2 x0 t0) where Kf (x1 t1 x0 t0) is the free propagator (Sakurai 2. t )(t . t ) 1 2h(t . t ) exp 2h(imx2 t ) exp 2h(imx0 t ) i t2 . t ) exp 2h(t . im 2x x . t ) 1 2h(t . im 2x x dx1 exp 2h(t . = (2:111) = 2 3 1=2 ! m 1=2 4 lim " 2ih" N det 5 N 2 ih m m 1=2 lim "p 1=2 (=) m 1=2 (t )] e N b 2 ih s2 ih m! N 2 ih sin(!T ) : !1 . t ) 1 2 2h(t . by explicitly performing the integral (i. 1 t2 . t0) 2 1 .16). do not use completeness).5. x1)2 = dx1 2 ih(t .2. 2x x i : b a 2 ih sin(!T ) 2h sin !T b a 2. t ) 1 0 2 1 2 1 2 1 2 1 .e. t ) 2 1 2 1 " s 2# m (x ) exp imh(t1 . !1 .

2h (t . t0) exp 2h (t2 . t0) s m 2 ih(t2 . t 2h # t2 . t1)(t1 . t )(t . t1) (t .. t0) #) x2(t1 .t2 . t )(t . 2h x1 im (t . t1) (t1 " t0) . t1)(t1". t1) 2 2h 2 1)( 1 t0 ) 8 2Z ih (t2 . x2(t1 . t ) 2 ih (t2 . t0 im x x2(t1 . t ) (t . t0) m(t2 . t0) ( Z dx1 exp .tt2 . (t2 . t0) + x0(t2 . t ) 2 1 2 1 1 0 s ( 1 "0 2 2 #) 1 m exp im x2 + x0 2 ih (t2 ( t1)("1 . t1)]2 exp .113) . t ) (t2 . t ) x1 h 1 (t . t0) . 0) 2x2x02(t1 . x2(t2 .tt0. t1)(t1 . t1) 2 . t2 + t1) . t0) + x0(t2 .0t) +tx0. t1)(t1 . t1) 5 2ih (t2 . t1)(t2 . t0 + 4 dx1 exp : x1 . t0)(t2 . t0) .. t )(t . t1) ( " 2 t t + 2( exp im x2(t1 .m t2 . t ) s 2 0m 2 1 1" im0(x . x0)(t2 . t0) 2 1 1 0 " 2 2(t2 . t0) + x0 . t0) #) 2( 1 m t exp im x2(t(1t. 2x2x02(t1 . t 0) 2 ih(t . t0)(t2 . t0)(t2 . t0)2 . t ) 2 h 2 1)( 1 0" " ##) t2 . t )(t . t )(t . t0(t. tt1)(t2 . x )2 # exp 2h(t2 .im (t .1 t+)(0) . t ) (t . t0) 2 1 1 v0 s u 2ih(t . # #) Z im 1 + 1 x2 + x0 2 . t0)(t2 . t ) 2 0 Kf (x2 t2 x0 t0): 2 0 (2. t )(t . 2h #) 2 0 t2 t1 1 . t1) 2 0 (t2 . t ) h 1 s 2 1 1 0 ( " 2 (t2 . x0)(tt2 . t0 . t0) x2(t1 .62 = = = = = = = s ( " 2 #) 1 m im x2 + x2 0 2 ih (t2 ( t1)(t1 . im x dx1 exp 2h (t . t0) #" #293 = < . t ) ( im u 1 m 1 2 1 1 0 t exp 2h (t . t0) (t2 t0 ( ) im 1 x2(t1 . t1) (t1 . t1)2 . " .

ihe @Ai = ihe @Aj . e Ai pj ] c c c c! = ihe @Aj . e pi Aj ] . @Ai c @xi c @xj c @xi @xj ! ihe " B : = c ijk k (2.114) We have also that 2 3 2 3 dxi = 1 x H ] = 1 4x ~ 2 + e 5 = 1 4x ~ 2 5 dt ih i ih i 2m ih i 2m 1 = ih1m f xi j ] j + j xi j ]g = ih2m f xi pj ] 2 2ih = 2ihm j ij = mi ) j+ j xi pj ]g .13 (a) Verify the relation i ! ihe " B j] = c ijk k ~ x p ~ where ~ m dt = ~ . QUANTUM DYNAMICS 63 2. ecA and the relation " !# d2~ = d~ = e E + 1 d~ B . eAi pj . B d~ : x x ~ ~ x ~ m dt2 dt 2c dt dt @ +r ~ =0 ~ j @t 0 0 (b) Verify the continuity equation with ~ given by j ! h =( r ) .2. eAj = . e Aj j2: ~ ~ ~= j m mc (a) We have i j] = pi .

115) (b) The time-dependent Schrodinger equation is 0 0 0 0 1 ~ A2 @ p A ih @t hx j t0 ti = hx jH j t0 ti = hx j 21 @~ . B dt : (2. B ~ ) x ~ x ~ eEi + 2c dt dt i " !# i ~ ~ ~ ~ x x ~ e E + 21c dt B .h2r 2 + e ih r A + 2ih e A r + e2 A2 + e : (2. ec + e j t0 ti m 2 3 2 3 ~ (~ ) 5 4 ~ eA(~ ) 5 ~x ~ = 21 4.ihr . " B xj . eAi 2m2ih c ijk k j c ijk j k ihm i c e (.64 d2xi dt2 = = (2:114) = = = m ddtxi 2 2 2 = = x md ~ dt2 2 3 " # 1 dxi H = 1 4 ~ 2 + e 5 ih dt ihm i 2m 1 f e i j ] j + j i j ]g + i ] 2 ih2m " i #hm 1 ihe " B + ihe " B + e p . c hx j t0 ti + e (~ )hx j t0 ti x m " # 1 . e @ ) 2m2c "ijk dt k ! ikj k dt !m @xi # e ~ B .ihr .116) ~ ~ ~ ~ = 2m c c c2 Multiplying the last equation by we get @ ih @t = 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ." B + " B ) + e p ] 2m2c ikj k j ijk j k ihm i e m " xj B . eAcx .h2r 2 (~ t) + e ih r A (~ t) + e ihA(~ ) r (~ t) x m c c x # e A(~ ) r (~ t) + e2 A2(~ ) (~ t) + e (~ ) (~ t) + ih c ~ x ~ x x x c2 x x " # 1 .h2r r + e ihr A(~ ) + ih e A(~ ) r + e2 A2(~ ) (~ t) ~ ~ ~ ~x ~ ~ x = 2m c c x c2 x +e (~ ) (~ t) x x ~ ~ x ~ ~ x = 21 .

e Aj j2 = 0 ) ~ ~ ~ @t m mc @ +r ~ =0 ~ j (2. 2hm r 2 . e ih r A j j2 . 2hm r mc mc = ih @ j j2 ) @t @ j j2 = .14 An electron moves in the presence of a uniform magnetic eld ~ in the z-direction (B = B z).h2 r 2 . eAy : c . h r h=( r )i + e r hAj j2i ) ~ ~ ~ ~ @t m" mc # @ j j2 + r h =( r ) .117) @t e ~ ~ with ~ = m =( r ) . ^ (a) Evaluate where x x y] px . QUANTUM DYNAMICS " 1 . eAx c y py .h2 r 2 + e ih r A j j2 + 2ih e A ~ ~ ~ 2m c c 0 0 65 # e2 A2j j2 + e j j2: ~ r + c2 0 The complex conjugate of this eqution is @ .2. r i + e ih r A j j2 + e ihA (r j j2) ~ ~ ~ ~ ~ ~ . r 2 e e ~ ~ ~ ~ ~ + mc ih r A j j2 + mc ihA ( r + r ) ! @ + @ = ih @t @t ) 2 ~ h r . mc Aj j2: and = j j2 j h 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.ih @t = " 1 . 2ih e A r ~ ~ ~ ~ 2m c c 0 0 0 0 0 # e2 A2j j2 + e j j2: + c2 Thus subtracting the last two equations we get 2 h i .

121) H1 H2] = 4m2 x 2c y z 2c there exists a set of simultaneous eigenstates jk ni of the operators H1 and H2. eAy (2:= p + eBy p . By Ay = Bx Az = 0: (2. ~ ~x The magentic eld B = B z can be derived from a vector petential A(~ ) ^ of the form Ax = . So if hk is the continious eigegenvalue of the operator pz and jk ni its eigenstate we will have 2k2 2 z H2jk ni = 2pm jk ni = h m : (2.119) (b) The Hamiltonian for this system is given by 0 1 ~ 2 1 @p .122) 2 . eA A = 1 2 + 1 2 + 1 p2 = H + H H = 2m ~ c 1 2 2m x 2m y 2m z where H1 1 2 1 2 2m x + 2m y (2. eB px x] + eB y py ] = iheB + iheB 2c 2c 2c 2c eB : = ih c (2.120) and H2 21 p2. Since m z " # 1 p + eBy 2 + p . eBx 2 p2 = 0 (2.66 2 where hk is the continuous eigenvalue of the pz operator and n is a (b) By comparing the Hamiltonian and the commutation relation obtained in (a) with those of the one-dimensional oscillator problem show how we can immediately write the energy eigenvalues as ! h2k2 + jeB jh n + 1 E = kn 2m mc nonnegative integer including zero.118) 2 2 Thus we have eAx p . y x c c 2c y 2c = . eBx 118) x y ] = px .

QUANTUM DYNAMICS 67 On the other hand H1 is similar to the Hamiltonian of the one-dimensional oscillator problem which is given by 1 H = 21 p2 + 2 m!2x2 (2.127) .125) = 2m y 2 mc eB : eB In this form it is obvious that we can replace ! with mc to have ! h2k2 jk ni + jeB jh n + 1 jk ni H jk ni = H1jk ni + H2jk ni = 2m mc 2 " 2 2 ! # k 1 = h m + jeB jh n + 2 jk ni: (2. From (a) we have eB ) xc (2.15 Consider a particle of mass m and charge q in an impenetrable cylinder with radius R and height a. In order to use the eigenvalues of the harmonic oscillator 1 En = h! n + 2 we should have the same commutator between the squared operators in the Hamiltonian.124) x y ] = ih y ] = ih: c eB So H1 can be written in the following form 2 1 2+ 1 2= 1 2+ 1 x c jeB j2 H1 2m x 2m y 2m y 2m eB c2 !2 2 1 2 + 1 m jeB j xc (2. Calculate the ground state energy and wavefunction.126) 2 mc j j 2. impenetrable solenoid carrying a magnetic ux .2. m @ 2 @ 2 @ 2 @z2 (~ ) = 2E (~ ) x x (2. Along the axis of the cylinder runs a thin. h2 r2 ] = 2E i ) m h h2 @ 2 + 1 @ + 1 @ 2 + @ 2 .123) m with x p] = ih. ~ In the case where B = 0 the Schrodinger equation of motion in the cylindrical coordinates is .

129) 1 1 Z (z) dz2 dz2 with Z (0) = 0 ) A1 + B1 = 0 ) Z (z) = A1 eilz . So Z (z) = C sin lnz Now we will have (2. e ilz = C sin lz Z (a) = 0 ) C sin la = 0 ) la = n ) l = ln = n a n = 1 2 : : : If we write ( .130) 1 d2R + 1 dR + 1 d2 + k2 .l2 ) d2Z + l2Z (z) = 0 ) Z (z) = A eilz + B e ilz (2.m2 ) ( ) = e im : d (2.132) with ( + 2 ) = ( ) ) m 2 Z: So the Schrodinger equation is reduced to 2 d2 R dR 2 2 2 2 2 + R( ) d .131) ) (1 ) d 2 = . l2 = 0 2 ( )d 2 R( ) d 2 R( ) d 2 2 1 2 ) R( ) d R + R( ) dR + ( ) d 2 + 2(k2 .128) 2 ( )d 2 R( ) d 2 R( ) d Z (z) dz2 with initial conditions ( a z) = (R z) = ( 0) = ( a) = 0. m + (k .68 z) = ( )R( )Z (z) and k2 = 2mE we will have h2 2 2 ( )Z (z) d R + ( )Z (z) 1 dR + R( )Z (z) d 2 2 d2 d d 2Z +R( ) ( ) d 2 + k2R( ) ( )Z (z) = 0 ) dz 2R 1 d + 1 dR + 1 d2 + 1 d2Z + k2 = 0(2. So 1 d2Z = . . l2) = 0 d2 d d 2 (2. l ) = 0 R( ) d .

138) p p . hc 2 ^@ (~ ) = E (~ ): x x (2. l2 = m (2.2. m2 R( ) = 0 ) d2 2 d # " d2R 1 dR m2 ) d(pk2 . We can then write ^ ! ! B 2 ^= a ^: ~ (2. 2m @ ^ @z @ !# 2 hc " ie ^@@ + z @z + ^ 1 @@ . l2 ) (2.134) where m is the -th zero of the m-th order Bessel function Jm. l2 ) k2 .ihr . ~ Now suppose that B = B z. l2 )2 + pk2 .136) nm (~ ) = Ac Jm ( R with n = 1 2 : : : and m 2 Z . l2 ) + 1 .135) ) E = 2m R2 a while the corresponding eigenfunctions are given by m x )eim sin n az (2. eA(~ ) 5 (~ ) = E (~ ) ~ ~ x x 2m c c " !# h2 ^ @ + z @ + ^ 1 @ . n a = R2 m h2 " # h2 2 + n 2 m (2. eA(~ ) 5 4.137) A= 2 2 ~ The Schrodinger equation in the presence of the magnetic eld B can be written as follows 2 3 2 3 ~x ~x 1 4.133) In the case at hand in which a ! 0 we should take B3 = 0 since Nm ! 1 when ! 0. From the other boundary condition we get R(R) = 0 ) A3Jm(R k2 . (k2 . l2) 2 R( ) = 0 p p ) R( ) = A3Jm( k2 . l2 ) + B3Nm( k2 . l2) .ihr . l2 = R2 ) 2mE . ie ) . This means that the energy eigenstates are given by the equation 2 2 2 p m m = R k2 . QUANTUM DYNAMICS 69 " # d2R + 1 dR + (k2 . l2) = 0 ) R k2 . l2 d(pk2 .

139) 2 @z2 x where D2 = ! ! @ 2 . A2 = .141) But ( +2 )= ( ))A m=m m 2Z ) m = (m . 2m @ 2 @ x (2. hc 2 we get " # " # @ + z @ + ^ 1 D (~ ) = E (~ ) h2 ^ @ + z @ + ^ 1 D ^@ ^ @z x x . ( z) = R( ) ( )Z (z)) we will get the same equations with the exception of " 2 # @ . A) m 2 Z : (2. hc 2 2. As a result the energy of the ground state will be # " h2 2 + n 2 m (2. 2m @ ^ @z " # h2 @ 2 + 1 @ + 1 D2 + @ 2 (~ ) = E (~ ) ) .142) This means that the energy eigenfunctions will be m x )eim sin n az (2.m2 ) d2 . 2iAl + (m2 .144) E = 2m R2 a D2 0 0 0 0 0 @ @ ie e .4A2 .e. A2) q 2iA .143) nm (~ ) = Ac Jm ( R but now m is not an integer. 2iA @ . 2iAlel + (m2 . A2 : (2. A2 = @ 2 . 2iA d + (m2 . Leting A = hc 2 we get . A2) = 0: @ 2 @ d 2 d The solution to this equation is of the form el . A2) 2iA 2im = = i(A m) ) l= 2 2 which means that ( ) = C2ei(A m) : (2. 2iA @ . 4(m2 .70 ie Making now the transformation D @@ . 2ie @ . So l2el . A2)el = 0 ) l2 .140) = @ 2 hc 2 @ @ 2 @ Following the same procedure we used before (i.

V (x)]1=4 sin 3 h .16 A particle in one dimension (. QUANTUM DYNAMICS 0 0 0 71 where now m = m . Notice also that if we require the ground state to be unchanged in the presence of B . V (x)]dx ! c 1 Z x q2m E . V (x)]dx .145) e 0 0 0 0 2.1 < x < 1) is subjected to a constant force derivable from V= x ( > 0): (a) Is the energy spectrum continuous or discrete? Write down an approximate expression for the energy eigenfunction speci ed by E. (b) Discuss brie y what changes are needed if V is replaced be V = jxj: (a) In the case under construction there is only a continuous spectrum and the eigenfunctions are non degenerate. V (x)]1=4 sin h x 2 3 4 s 3=2 2m c 4. = E . we obtain ux quantization e m . 2 E .146) . h 2m E . = E . V (x)]1=4 exp h 2m E .2. V (x)]dx I iZ q B + E . A is not zero in general but it corresponds to m 2 Z such that 0 m . x 5 = E . V (x)]1=4 exp . From the discussion on WKB approximation we had that for E > V (x) A iZ q (x) = E . A < 1. A = 0 ) hc 2 = m ) = 2 m hc m 2 Z : (2. V (x)]1=4 sin h 4 p x Z x =E= ! 1=2 c 2m E .4 0 0 2 1 = qc1=4 sin 3 q3=2 + 4 ] (2. x dx .

6pm + c1 " !# i p3 . E ]1=4 exp " .147) .q]1=4 3 We can nd an exact solution for this problem so we can compare with the approximate solutions we got with the WKB method. E )d(2m x) = x . E )p jcj22 h (E . 2 (. We have H j i = E j i ) hpjH j i = hpjE j i 2 ) hpj 2pm + xj i = E hpj i 2 d ) 2pm (p) + ih dp (p) = E (p) ! d (p) = . Ep : 6m (2. E ]1=4 exp .q)2=3 : = (2. h x =E= # c2 1 Z x q2m( x .72 h i m 1=3 where q = E . p2 (p) ) dp h 2m ! d (p) = . h2m x =E= c3 exp . E )dx II (x) = x . E ) 0 So Z jcj2 dp exp hi (E . E ) ) (2. On the other hand when E < V (x) ! c2 1 Z x q2m( x . x and = 2h2 .i E .148) 0 0 We also have (E . p2 dp ) (p) h 2m ! 3 .i E .149) c = p1 : 2 h (2:148) = = hE jE i = dphE jpihpjE i = 0 0 0 Z Z E (p) E (p)dp 0 0 " 1 exp i E (p) = p h 2 h !# p3 .150) . Ep : ) E (p) = c exp h 6m (2. ) ln (p) = h i Ep .

153) 3 2 hE i m 1=3 and q = where = 2h2 . uq du (2.152) u = (h2m )1=3 ) h p6m = u 3 we have " # (h2mp 1=3 Z + du exp iu3 .1 1 . QUANTUM DYNAMICS 73 These are the Hamiltonian eigenstates in momentum space. leading terms in the asymptotic series are as follows 2 p 1q1=4 exp . x u(h2m )1=3 ) (x) = 2 h " 3 3 #h Z+ = p du exp iu . So ! ! Z+ u3 . uq du du cos 3 (x) = p 3 0 2 R since + sin u33 .155) . i E .151) h 6m h 2 h Using now the substitution 3 3 p (2. iuq (2.1 1 1 .156) (2.q): (2. i E . 1 .1 1 For large jqj.q)1=4 sin 3 (. For the eigenfunctions in coordinate space we have 3 Z 1p Z dpe ipx e hi 6pm Ep (2:150) h (x) = dphxjpihpjE i = " 2 h # Z 1p dp exp i p3 .1 1 . uq = p Z + cos u3 .154) Ai(q) = p 3 0 we will have (x) = p Ai(. In terms of the Airy functions ! 1 Z + cos u3 .2.157) .q)3=2 + 4 q < 0 Ai(q) (2. 3 q3=2 q > 0 Ai(q) 2 2 1 p (. uq du = 0. x . x p : = (2.

x) . E and x2 = E .155) we get p q11=4 sin 2 q3=2 + 4 for E > V (x) (q) 3 2 p (.q1)1=4 exp .74 Using these approximations in (2.160) 4 2m p 0 . 2 E . jxj] = 2 0 E= 1=2 p Z E= E = . So Z E= q Z E= q 1 h = n+ 2 2m E .q)3=2 for E < V (x) (2. x 3=2 E= = 2p2m 2 E 3=2 ) = . 3 (.159) x2 The turning points are x1 = . x]dx dx 2m E . jxj] = n + 2 h n = 0 1 2 : : : (2. So in this case the energy eigenstates heve to satisfy the consistency relation Z x1 q 1 dx 2m E .2 2m 3 3 0 1 +2 E 3=2 = 3 np 1 h ) E = 3 n + 2 h]2=3 ) 42=3(2m )1=3 4 2m 2 32=3 1 43 n + 2 h 5 : p En = (2. x d(. (b) When V = jxj we have bound states and therefore the energy spectrum is discrete.2 2m .158) (q) 2 as expected from the WKB approximation.

we expect that this sequence of operations is equivalent to a single rotation about some axis by an angle . .1) ei( + )=2 cos 2 2 while the same rotation will be represented by 1 0 (. 1 : ) = arccos 2 cos 2 (3. Without using the .e i( )=2 sin 2 ei( + )=2 cos 2 . Find . inz sin 2 ! ) cos 2 = cos ( + ) cos 2 2 ) cos = 2 cos2 2 cos2 ( + ) .inx + ny ) sin 2 cos 2 + inz sin 2 Since these two operators must have the same e ect.3. inz sin 2 : (3.2) D ^ = (. 2 Because of the group properties of rotations. 1 2 " # 2 cos2 ( + ) . In the case of Euler angles we have ! e i( + )=2 cos 2 . . .1 Consider a sequence of Euler rotations represented by . . .inx . each matrix element should be the same.3) 2 . . 2 2 ! : 3. THEORY OF ANGULAR MOMENTUM 75 3 Theory of Angular Momentum = 3. ny ) sin 2 A (1=2) ( n) (S 3:2:45) @ cos 2 . . That is ! ! i( + )=2 cos = cos e 2 2 .i 3 exp .2 An angular-momentum eigenstate jj m = mmax = j i is rotated by an in nitesimal angle " about the y-axis.i 2 ! exp . .i 3 D(1=2)( ) = exp e i( + )=2 cos 2 ei( )=2 sin 2 .e i( )=2 sin 2 (1=2) ( D ) = ei( )=2 sin (3.

h 2j (J+ . 2h + (3. iJhy " jj j i ^ " # iJy " + (. J jj j . gives " # " (J . 1)jj j . We can write Jy in terms of the ladder operators ) J+ = Jx + iJy ) J = J+ .h 2j (J+jj j .J jj j i = . 1)jj j . 1i (3.4).h 2j 2j jj j i . obtain an expression for the probability for the new rotated state to be found in the original state up to terms of order "2. "2 2j jj j i + "2 2qj (2j . . 1i + "2 qj (2j .6) 8h2 + We know that for the ladder operators the following relations hold q J+jj mi = h (j . . . .h 2j jj j . 2(2j . 1i q = . J )jj j . J : (3.5) y J = Jx .8) So q (J+ . m + 1)jj m . 0 The rotated state is given by up to terms of order "2. . 1i . 1)jj j . . 2i jj j iR = jj j i + 2 8 8 ! "2 j jj j i + " q2j jj j .4) h 2h2 y and from (3.76 j explicit form of the d(m)m function. 1i (3. jj j iR = R(" y)jj j i = d(j)(")jj j i = exp .i)2"2 J 2 jj j i = 1.6) " q2j jj j . J )jj j i = . 1i . m)(j + m + 1)jj m + 1i (3. J ) + "2 (J . . . 2i .9) q (J+ . . J )2 jj j i jj j iR = 1 .7) q J jj mi = h (j + m)(j . 4 2 4 . 2i: = 1. 1i) q q q = . J )2jj j i = . (3. iJy 2i Subtitution of this in (3.

3 The wave function of a particle subjected to a spherically symmetrical potential V (r) is given by (~ ) = (x + y + 3z)f (r): x ~ (a) Is an eigenfunction of L? If so. what are the possible values of l we may obtain when L2 is measured? (b) What are the probabilities for the particle to be found in various ml states? (c) Suppose it is known somehow that (~ ) is an energy eigenfuncx tion with eigenvalue E . y= (3. " j + O("4): (3.3. Indicate how we may nd V (r).h2 12 @ 2 + 1 @ sin @ h~ jL x = (~ ): x sin @ sin @ @ If we write (~ ) in terms of spherical coordinates (x = r sin cos x r sin sin z = r cos ) we will have (~ ) = rf (r) (sin cos + sin sin + 3 cos ) : x Then 1 @ 2 (~ ) = rf (r) sin @ (cos . what is the l-value? If ~ not.12) " !# 2 ~ 2j i (S 3:6:15) . " j jhj j jj j iRj = 1 .13) (3. (a) We have (~ ) h~ j i = (x + y + 3z)f (r): x x So (3. sin ) = .11) (3. rf (r) (cos + sin sin sin2 @ 2 x sin2 @ . THEORY OF ANGULAR MOMENTUM 77 Thus the probability for the rotated state to be found in the original state will be 2 !2 2 2 = 1.10) 4 2 3.14) ) .

.6 sin cos + (cos + sin )(cos2 . 6 cos x~ 1 = h2rf (r) sin 2 sin2 (cos + sin ) + 6 cos = 2h2rf (r) sin cos + sin sin + 3 cos ] = 2h2 (~ ) ) x 2 (~ ) = 2h2 (~ ) = 1(1 + 1)h2 (~ ) = l(l + l)h2 (~ ) L x x x x (3.15) in (3.3 sin2 + (cos + sin ) sin cos i = x sin @ sin @ @ i rf (r) h.1i + (1 . sin (cos + sin )(1 .18) 22 . But this means that the part of the state that depends on the values of m can be written in the following way hp i j im = N 3 2jl = 1 m = 0i + (1 + i)jl = 1 m = . Y1+1 + iY1+1 + iY1 1 x s hp i = 23 rf (r) 3 2Y10 + (1 + i)Y1 1 + (i .15) : sin Substitution of (3. 8 r = < x = r q23 Y1 1 . 1)Y1+1 : (3. Y1+1 q ): Y1 1 = 83 (x riy) y = ir 23 Y1 1 + Y1+1 So we can write s i hp (~ ) = r 23 f (r) 3 2Y10 + Y1 1 .14) and (3. i)jl = 1 m = 1i and if we want it normalized we will have jN j2(18 + 2 + 2) = 1 ) N = p1 : (3. .11) gives 1 h~ jL2j i = . We know that the spherical harmonics Y1 s s s 3 cos = 3 z ) z = r 4 Y 0 Y10 = 4 1 r 3 q q 3 (x+iy4) 9 8 Y1+1 = . cos2 + sin2 ) .16) ~ which means that (~ ) is en eigenfunction of L2 with eigenvalue l = 1.78 and ! 1 @ sin @ (~ ) = rf (r) @ h. .17) . . . sin2 )(3. .h2rf (r) . x (b) Since we already know that l = 1 we can try to write (~ ) in terms of x m ( ).

20) (3. Quantum-mechanically. 2 rf (r)]# + V (r)rf (r)Y m = ) 2m l dr2 l r dr r2 Erf (r)Ylm ) " # 1 h2 d f (r) + rf (r)] + 2 f (r) + rf (r)] .21) (c) If E E (~ ) " x 2 .22) 3. THEORY OF ANGULAR MOMENTUM So 9 P (m = 0) = jhl = 1 m = 0j ij2 = 9222 = 11 2 1 P (m = +1) = jhl = 1 m = +1j ij2 = 22 = 11 2 1 P (m = . (a) Show explicitly that in nitesimal rotations of %i(~ ) are obtained x by acting with the operator " ~ ~ ~ (3.3.1) = jhl = 1 m == 1j ij2 = 22 = 11 : 79 (3.h Y m d2 rf (r)] + 2 d rf (r)] .19) (3. L2 (~ )# + V (r) (~ ) = E x 2m @r2 E x r @r E x h2r2 E x 2 V (r) = E + rf1r) 2hm f (r) + f (r) + rf (r) + 2f (r)]] ) ( 2 rf (r) + 4f (r) V (r) = E + 2hm : rf (r) 0 0 00 0 00 0 (3. such a particle is described by a ketvector j%i or in ~ representation a wave function x %i(~ ) = h~ ij%i x x where j~ ii correspond to a particle at ~ with spin in the i:th dix x rection. i h (L + S ) " .h2 " @ 2 (~ ) + 2 @ (~ ) . (One example of such a particle is the %meson). 2 f (r) ) V (r) = E + rf (r) 2m dr r r 0 0 x E (~ ) is an energy eigenfunction then it solves the Schrodinger equation .23) u~ = 1 .4 Consider a particle with an intrinsic angular momentum (or spin) of one unit of h.

(3. 0 %i (~ ) x 0 = i=1 (1) Dil (R)%l(R 1~ ) ) % (x) = R~(R 1~ ): x ~ ~ % x . ~ (~ r)%(~ ) % x " x ~~ x % x " x ~ ~ x = ~(~ ) . ~ ~ ) = ~(~ ) . ~ ~x ~ p~ (d) Show that r %(~ ) = h1 (S ~)% where ~ is the momentum operp ator.26) Under an in nitesimal rotation we will have R( So 0 n)~ = ~ + ~ = ~ + (^ ~) = ~ + ~ ~: ^r r r r n r r " r .28) . 0 . ~ (c) Show that S is a vector operator. (~ ~ ) r%(~ ) = ~(~ ) . Determine S ! ~ ~ (b) Show that L and S commute.24) Under a rotation R we will have 3 XZ j% i = U (R)j%i = U (R) j~ i jii] %i(~ )d3x x x i=1 3 Z 3 XZ R X (1) jR~ i jiiDil (R)%l (~ )d3x det==1 x x j~ iiDil (R)%l (R 1~ )d3x x (1) x = i=1 i=1 3 XZ = j~ ii%i ~ )d3x ) x x 0 . 2 (a) We have j%i = 3 XZ i=1 j~ iih~ ij%i = x x 3 XZ i=1 j~ ii%i(~ )d3x: x x (3.25) (3.27) = ~(~ .80 ~ ~ i^ ~ where L = h r r. ~ ~ ) + ~ %(x . ~ ~ ): %x " x " ~~ " x On the other hand %x " x ~(~ . % (~ ) = R( )%(R 1~ ) = R( )%(~ . h ~ L%(~ ) % x i" ~~ x (3. ~ ~ ) ~ x ~ x ~x " x (3.

1 0 0 0 0 0 %3 2 0 1 0 1 0 13 0 1 1 0 0 0 0 0 ih 0 . (."z "y B %2 C = B "z 0 .ih 0 i 6" B 0 0 . im jk ) X X X = h2 ijl kml = ih ijl(.ih kl.27) we get % (~ ) = ~(~ ) .h 4 x @ A y@ A z@ A5 @ % A 0 ih 0 .3.ih kml) = ih ijl(Sl)km : (3. h ~ (L + S ): (3.ih C + " B 0 0 0 C + " B ih 0 0 C7 B %2 C . h2 jkl iml X = h2 ( ij km . h ~ L%(~ ) ~ x % x i" ~~ x " ~ x i" ~~ x = ~(~ ) . "x%1 ey + "x%2 .ih 0 0 0 0 0 %3 which means that i " ~% x " ~ ~ % = .ih) jlm . h ~ S ~(~ ) with (S )kl = . THEORY OF ANGULAR MOMENTUM ~ ~ x h~ x i where r%(~ ) r%i (~ ) jii.ih) ilm] i Xh 2 = h ikl jml . h ~ L%(~ ) + ~ %(~ ) .ih) jkl(. Using this in (3. h ~ L%(~ ) + ~ ~(~ ): % x i" ~~ x " % x 0 81 (3.1 0 % 6"x B 0 0 . im jk ."x C B %2 C = A@ A @ A @ . Thus we will have that i" ~ ~ i" ~ ~ ~ x % (~ ) = U~~(~ ) = 1 ."y "x 0 %3 %3 2 0 1 0 1 0 13 0 1 1 0 0 0 0 0 1 0 . ij km + jm ki) X = h2 ( jm ki .32) .ih) ikl(.31) ~ x " "% x 0 0 0 0 But ~ ~ ~ (b) From their de nition it is obvious that L and S commute since L acts ~ only on jii. only on the j~ i basis and S x ~ (c) S is a vector operator since X Si Sj ]km = SiSj . "z %2 ex + "z %1 . Sj Si]km = (.30) ^ ^ ^ or in matrix form 10 1 1 0 1 1 0 % % 0 . "y %1 ez (3. h ~ (L + S ) %(~ ) ) U~ = 1 .1 C + "y B 0 0 0 C + "z B 1 0 0 C7 B %2 C = 4 @ A @ A @ A5 @ A 0 1 0 .29) " ~ ~ % = "y %3 .

m + 1)j0+i + h (j2 + m2)(j2 . Let us take rst the state j = 2. . j2j : : : j1 + j2 = 0 1 2 states.35) we get jj = 2 m = 2i = hj1j2 + + jj1j2 jmij + +i norm.36) = If we apply the J operator on this statet we will get .37) . m = 2. Using the ladder operator method express all (nine) j m eigenkets in terms of jj1j2 m1m2i. m + 1)jj = 2 m = 1i = q q h (j + m )(j . . m2 + 1)j + 0i p 1 1 1 1 p p ) 4jj = 2 m = 1i = 2j0+i + 2j + 0i 1 1 ) jj = 2 m = 1i = p2 j0+i + p2 j + 0i: (3. .82 (d) It is ip ~ x i 1 ~ ~x r %(~ ) = h ~ %(~ ) = h i lp %l(~ )jii = h2 (S )lm p %m jii x ~ p~ = 12 (S ~)%: h (3.5 We are to add angular momenta j1 = 1 and j2 = 1 to form j = 2 1 and 0 states. Write your answer as 1 1 jj = 1 m = 1i = p2 j+ 0i . This state is related to jj1m1 j2m2i through the following equation X (3. We want to add the angular momenta j1 = 1 and j2 = 1 to form j = jj1 .35) j j mi = hj1j2 m1m2jj1j2 jmijj1j2 m1m2i m=m1 +m2 So setting j = 2. Jqjj = 2 m = 2i = (J1 + J2 )j + +i ) h (j + m)(j . m = 2 in (3. j + +i (3.33) 3. p2 j0 +i : : : (3.34) where + and 0 stand for m1 2 = 1 0 respectively.

.i: (3.i + p 2j .2i = j .2i = p 2 2 jj = 2 m = . +i 3 . . So hj = 2 m = 1jj = 1 m = 1i = 0 ) 12 a + 12 b = 0 ) a + b = 0 ) a = . If j = 1.i + p6 j .1i = p (J1 + J2 )j0.39) jj = 2 m = . .3. . 0i (3. .40) Now let us return to equation (3.i ) 2 2 p 6jj = 2 m = 0i = 2j00i + j .i + p p2j . . 2 (J + J )j00i 2 3 1 1 1 + p (J1 + J2 )j + . THEORY OF ANGULAR MOMENTUM 83 In the same way we have 1 1 J jj = 2 m = 1i = p (J1 + J2 )j0+i + p (J1 + J2 )j + 0i ) 2 2 i hp p i 1 hp p p 1 6jj = 2 m = 0i = p 2j .41) This state should be orthogonal to all jj mi states and in particular to jj = 2 m = 1i. p p . . .i + p (J1 + J2 )j .i + 2 p2j . . .35). .1i = p2 2 J jj = 2 m = 0i = .1i = 6 6 6 6 1 1 j0. m = 1 we will have jj = 1 m = 1i = aj + 0i + bj0+i (3.i + p j . +i + 2j00i + p 2j00i + 2j + . 0i ) 3 6 6 2 p2j0.i + p 2j0.i ) 1 4jj = 2 m = . . . 0i ) 2 2 p 1 p2j . 0i + 1 p2j0.42) . 1 1 J jj = 2 m = . 0i + 2j0.i + p (J1 + J2 )j .38) jj = 2 m = 0i = 2 j00i + p6 j + . 0i ) jj = 2 m = . .i + 1 p2j . .b : (3. +i ) 6 s 6 i 1p hp p p 1p 6jj = 2 m = .i ) s 1 1 (3. s .1i = 2 2j . . +i + j + .

p2 j0+i: Using the same procedure we used before 1 1 J jj = 1 m = 1i = p (J1 + J2 )j + 0i .i .1i = p 2j0. p j . p hp2j .48) . 12 . .1i = p2 (3. +i ) 2 2 p p p 1 1 2jj = 1 m = . .84 In addition the state jj = 1 m = 1i should be normalized so 1 :42) hj = 1 m = 1jj = 1 m = 1i = 1 ) jaj2 + jbj2 = 1 (3) 2jaj2 = 1 ) jaj = p2 : p p By convention we take a to be real and positive so a = 12 and b = .i + c3j .i . p (J1 + J2 )j . .44) jj = 1 m = 0i = p2 2 .i .47) 1 1 hj = 1 m = 0jj = 0 m = 0i = 0 ) p2 c2 .45) 2 Returning back to (3.46) This state should be orthogonal to all states jj mi and in particulat to jj = 2 m = 0i and to j = 1 m = 0i. p (J1 + J2 )j0+i ) 2 2 p 1 hp2j00i + p2j + . . 1 1 J jj = 1 m = 0i = p (J1 + J2 )j + .ii . So s 1 1 hj = 2 m = 0jj = 0 m = 0i = 0 ) 2 c1 + p6 c2 + p6 c3 3 ) 2c1 + c2 + c3 = 0 (3. .i . 0i ) 2 2 1 j0. +i (3.35) we see that the state jj = 0 m = 0i can be written as jj = 0 m = 0i = c1j00i + c2j + . +i: (3. . p j . . +i + p2j00ii ) 1 2jj = 1 m = 0i = p 2 2 1 1 j + .43) jj = 1 m = 1i = p2 j + 0i . p2 c3 ) c2 = c3: (3. . . 0i: 1 jj = 1 m = . p 2j . That is 1 1 (3.

52) 3. 12 j0+i = 12 j + . Thus p p 1 3 and 1 1 1 jj = 0 m = 0i = p3 j + . .6 (a) Construct a spherical tensor of rank 1 out of two di erent ~ ~ vectors U = (Ux Uy Uz ) and V = (Vx Vy Vz ).i + 12 j .c2: The state jj = 0 m = 0i should be normalized so 85 (3. Explicitly write T (1)0 in 1 terms of Ux y z and Vx y z . 12 j . 0i = j . 13 j00i: p p p p p p p p p p p p p p p (3. +i .50) By convention we take c2 to be real and positive so c2 = c3 = c1 = . . 0i = 13 j + .i = 12 j + 0i .49) hj = 0 m = 0jj = 0 m = 0i = 1 ) jc1j2 + jc2j2 + jc3j2 = 1 ) 3jc2j2 = 1 1 ) jc2j = p3 : (3.i .51) jj = 2 m = 2i jj = 2 m = 1i jj = 2 m = 0i jj = 2 m = . +i 3 = 12 j0.i . 13 .47). THEORY OF ANGULAR MOMENTUM Using the last relation in (3. Write down explicitly T (2) 1 0 in terms of Ux y z 2 and Vx y z .1i jj = 0 m = 0i = j + +i 1 = q2 j0+i + 12 j + 0i = 2 j00i + 16 j + .1i jj = 2 m = . (b) Construct a spherical tensor of rank 2 out of two di erent ~ ~ vectors U and V . we get 2c1 + 2c2 = 0 ) c1 + c2 = 0 ) c1 = . +i .i + p3 j . 12 j .i + 13 j . +i = 12 j0.i + 16 j .2i jj = 1 m = 1i jj = 1 m = 0i jj = 1 m = . p3 j00i: So gathering all the previous results together (3.3.

ihUy .ihUy + i(ih)Ux = . . iUy ) (3.27) that if Xq(1k1 ) and Zq(2k2) are irreducible spherical tensors of rank k1 and k2 respectively then we can construct a spherical tensor of rank k X Tq(k) = hk1 k2 q1q2jk1k2 kqiXq(1k1)Zq(2k2) (3. iJy Uz ] J U0] (3= (3:53) = . p p . iVy ) . iUy ) ) 1 U 1 = p (Ux .53) From the components of a vector operator we can construct a spherical tensor of rank 1 in the following way. .54) It is :53) :54) Jz Uz ] (3= 0hUz (3) Uz = U0 (3. i(ih)Ux = h(Ux .86 ~ ~ (a) Since U and V are vector operators they will satisfy the following commutation relations Ui Jj ] = ih"ijk Uk Vi Jj ] = ih"ijk Vk : (3.56) 2 :54) p 2hU 1 = J Uz ] = Jx . 12 (Vx + IVy ) (3.55) p :54) J+ U0] (3= 2hU+1 = J+ Uz ] = Jx + iJy Uz ] (3:53) = . iUy ) V 1 = 12 (Vx .h(Ux + iUy ) ) 1 U+1 = . p (Ux + iUy ) (3. The de ning properties of a spherical tensor of rank 1 are the following q Jz Uq(1)] = hqUq(1) J Uq(1)] = h (1 q)(2 q)Uq(1)1 : (3.10.57) 2 ~ ~ So from the vector operators U and V we can construct spherical tensors with components U0 = Uz V0 = Vz 1 (U + iU ) V+1 = . p . . p It is known (S-3. 2 x y U 1 = 12 (Ux .59) q1 q 2 .58) U+1 = .

64) T0(2) = h11 00j11 20iU0 V0 + h11 . U V + iU V + iU V ) = . . iUy Vx .10j11 11iU 1 V0 + h11 0 . (3.1 + 1j11 10iU 1 V+1 +h11 +1 .61) 2 T (1) = h11 . .62) .3. 1j11 11iU0 V 1 1 1 1 (3:52) = . Uy Vy ] 2 2 i = p (UxVy . . p U0V+1 2 2 1 (U + iU )V + 1 U (V + iV ) = . . . . THEORY OF ANGULAR MOMENTUM In this case we have (1) T+1 = h11 +10j11 11iU+1 V0 + h11 0 + 1j11 11iU0 V+1 1 (3:52) 1 = p U+1V0 . 2 (Ux .2 x y z 2 z x y . . . 87 (3. iU )(V + iV ) . p U 1V+1 + p U+1V 1 2 2 1 1 (U . iUy )Vz + 2 Uz (Vx .2 x x y y (3. (b) In the same manner we will have :52) (2) T+2 = h11 +1 + 1j11 2 + 2iU+1 V+1 (3= U+1 V+1 = 1 (Ux + iUy )(Vx + iVy ) 2 1 (U V . p U 1V0 + p U0V 1 2 2 1 1 = . 1j11 20iU+1 V 1 . p 1 (U + iU )(V .60) T0(1) = h11 00j11 10iU0 V0 + h11 . UxVx + iUxVy .63) x y y x (2) T+1 = (3:52) = = h11 0 + 1j11 2 + 1iU0 V+1 + h11 +10j11 2 + 1iU+1 V0 1 1 p U0V+1 + p U+1V0 2 2 . 1j11 10iU+1 V 1 1 1 (3:52) = . Uy Vx) (3. iUy Vx + Uy Vy .1 + 1j11 20iU 1 V+1 +h11 +1 . iV ) 1 = p 2 x y x y y x y 2 22 x 1 = p UxVx + iUxVy . iVy ): (3. 1 (Uz Vx + UxVz + iUz Vy + iUy Vz ) 2 .

] . 1): 2 2 0 0 Hint: This can be proved in many ways. 1iU0 V 1 + h11 . iUy Vx): (3. 2 Uy Vy . iU )(V + iV ) . 2iU 1 V 1 (3= U 1V 1 = 2 (Ux .67) 2 3. 1 Uy Vy 2 s 1 (2U V . examine the rotational properties of Jz2 using the spherical (irreducible) tensor language. T (2) = 1 (3:52) = = . for instance. Uy Vy . 1 U V . h11 0 . i U V + i U V 6 z z 2 x x 2 x y 2 y x i i 1 1 . You may. iVy ) 2 = 1 (UxVx . for any j . 1j11 2 . U V ) (3. 1 p 1 (U + iU )(V . iU V ) x z z y y z 2 z x . iUxVy . . 1j11 2 .66) :52) 1 T (2) = h11 . (j m2jdm)m( )j2 = 1 j (j + 1) sin + m 2 + 1 (3 cos2 . iUy )(Vx . 2 UxVx + 2 UxVy . iV ) y x y y x y z z 2 x 6 6 22 x s3 1 2U V . j) jd(mm ( )j2m 0 1 for any j (integer or half-integer) then check your answer for j = 2 . . 1 1 (U .10j11 2 + 1iU 1 V0 1 1 p U0V 1 + p U 1 V0 2 2 1 (U V + U V . . .1 . (3. . . .65) 6 z z x x y y . j X m= j . U V . (b) Prove. iU V . . . 2 Uy Vx .88 (3:52) = s = = = s s 2U V + 1U V + 1U V 3 0 0 s 6 1 +1 6 +1 1 s s 1 2 U V .7 (a) Evaluate j X m= j .

2 j 3 1 hjm jeiJy =hJ 4 X jjmihjmj5 e iJy =hjjm i = h z m= j 1 hjm jeiJy =hJ e iJy =hjjm i = h z 1 ^ ^ (3. iJy =hjjm ij2 0 0 .69) D ( ey)Jz D( ey ) = Rzj ( ey )Jj : ^ ^ . iJy =hjjm ihjm jeiJy =hjjmi . mhjmje 0 . m= j j X . sin hjm j J+ + J jjm i + hm cos 2 = m cos : (3. sin hjm jJxjjm i + cos hjm jJz jjm i] m= j 1 = h . sin 0 cos So j X (j) 1 jdmm ( )j2m = h . m= j j X .71) 0 j 0 0 0 0 . iJy =hjjm i 0 m= j j X . 0 0 0 m= j hjm jeiJy =hmjjmihjmje iJy =hjjm i 0 On the other hand we know (S-3. mjhjmje mhjmje . 0 0 0 . 0 . 0 .5b) that 0 1 cos 0 sin R( ey ) = B 0 1 0 C : ^ @ (3. THEORY OF ANGULAR MOMENTUM (a) We have j X j) jd(mm ( )j2m 0 89 = = = = m= j j X . 0 .10.70) A .3.1.3) we will have that X ^ (3. 0 . iJy =hjjm i 0 hjmje 0 .68) = h hjm jD ( ey )Jz D( ey )jjm i: ~ But the momentum J is a vector operator so from (S-3.

iJy =hm2jjmihjmjeiJy =hjjm i (3. m2hjm je 0 . 0 . iJy =hjjmi 0 m= j j X . 1 cos = m cos : 2 (3. j m2jd(m)m( )j2 0 = = = = m= j j X . 0 0 y 0 m= j hjm je 0 . j 1 1 jd(m)1=2( )j2m = . 0 . sin 2 : (1=2) dmm ( ) = sin cos 2 2 So for m = 1=2 1=2 X (j) jdm1=2( )j2m = . 2 cos2 2 + 2 sin2 2 0 = .75) . 1 sin2 2 + 1 cos2 2 2 2 0 (3. iJy =hjjmihjmjeiJy =hjjm i 0 2 j 3 1 hjm je iJy =hJ 2 4 X jjmihjmj5 eiJy =hjjm i = 2 z h m= j = 12 hjm je iJy =hJz2eiJy =hjjm i h 1 = h hjm jD( ey )Jz2D ( ey)jjm i: ^ ^ . m2hjm je 0 . 0 . iJy =hjjmij2 m= j j X . m2jhjm je 0 .2.73) m= 1=2 .74) (b) We have j X m= j j X . iJy =hjjmi hjm je 0 .90 For j = 1=2 we know from (S-3.44) that ! cos 2 . while for m = .1=2 1=2 X 0 = 1 2 cos = m cos 0 (3.72) 0 m= 1=2 .

80) 0 0 .10. 1 j (j + 1) = 3 2 00 3 0 0 0 0 0 j) m2jd(mm ( )j2 0 . 1 J 2jjm i 3 2 00 3 1 j (j + 1) + 1 d(2)( ) m 2 .77) and since D(R)J 2D (R) = J 2D(R)D (R) = J 2 we will have Pj 2 (j ) 2 m= j m jdmm ( )j = s 1 1 hjm jJ 2jjm i + 2 1 hjm jD( e )J 2D ( e )jjm (3. So = 0 6 j X m= j . So 1 Jz2 = 36 T0(2) + 3 J 2 (3.79) qq q y y . J 2) 6 z (3.65) we know that 91 T0(2) = s 1 (3J 2 . which means in our case that hjm jD( ey )Jz2D ( ey )jjm i = hjm j ^ ^ 0 y 0 0 2 X = 2 X q= 2 0 (2) Tq(2)Dq 0( ey )jjm i ^ 0 0 0 .78) ^y z ^y i: 3 h2 h3 3 We know that for a spherical tensor (S-3. s 1 h2j (j + 1) + 1 2 D(2)( e )hjm jT (2)jjm i ^y = 2 0 3h h2 3 00 = 1 j (j + 1) + 1 d(2)( )hjm jJz2 . 0 p 0 0 0 y 0 y q= k 0 0 0 .3. But we know from the Wigner-Eckart theorem that hjm jTq(2)0jjm i = 0. q= 2 0 Dq(2)( ey)hjm jTq(2)jjm i: ^ 0 0 0 0 0 (3. THEORY OF ANGULAR MOMENTUM From (3.76) where T0(2) is the 0-component of a second rank tensor.22b) k (k) D (R) = X D(k) (R)T (k) D(R)Tq (3.

81) = 2 2 0 0 0 where we have used d(2)( ) = P2(cos ) = 1 (3 cos2 . r2)j j m = j i . y2) + ixy T (2) = 2 (x2 . Q = eh j m = j j(3z2 . r ) . 1 j .82) q 1 2 2 2 q 1 2 2 T 1 = xz . y2) .63-3. (b) The expectation value Q eh j m = j j(3z2 .67) we can nd that in the case where U = ~ x V = ~ the components of a spherical tensor of rank 2 will be (2) 1 T+2 = 1 (x2 . 1). So from the above we have x2 . 1 j (j + 1) 3 1 j (j + 1) cos2 + 1 j (j + 1) + 1 j (j + 1) + m 2 (3 cos2 . T (2) 2 xy = 2i .(xz + izy) (3. 1) m 2 . izy (2) = T0 6 (2z . .92 1 1 = 3 j (j + 1) + 2 (3 cos2 . y2 (b) We have (2) = T+2 + T (2) 2 . y2) as components of a spherical (irreducible) tensor of rank 2. 2 6 3 2 1 j (j + 1) sin2 + m 2 1 (3 cos2 . r2)j j m = j i is known as the quadrupole moment. Evaluate eh j m j(x2 . 2 : : : )in terms of Q and appropriate ClebschGordan coe cients. 0 0 ~ (a) Using the relations (3. xz. x . ixy 2 2 (2) (2) T+1 = . 1) = . y2)j j m = j i (where m = j j .8 (a) Write xy. (2) T (2) .83) 1 xz = 2 . and (x2 . T+1 : (3. 1) (3. 00 2 3. (2) T+2 . y ) = 6 (3z .

.84) e h j m j(x2 . THEORY OF ANGULAR MOMENTUM (3:82) 93 .3. 2i = p hj 2 j 0jj 2 j j i m j 2: (3. 2jj 2 jj .2jj 2 j j . 0 . 2i h j k2j +k1 j i (3:84) Q hj 2 j . y2)j j m = j i (3:83) (2) = eh j m jT+2 j j m = j i + eh j m jT (2)j j m = j i 2 0 z }| { (2) (2) pT pT = e hj 2 j 2jj 2 jm i h j k2j +k1 j i + e m j 2hj 2 j .85) 6 0 0 . = p 6eh j m = j jT0(2)j Q ) h j kT (2)k j i = p6e hj 2 j20jjj+ 1jj i : 2 So 0 h jpT (2) k j i p6e j m = j i = hj 2 j 0jj 2 jj i k2j + 1 p (W: E:) (3. 0 0 .

pi.94 4. 2Ent the wave function will be h real.2 Let (~ ) be the momentum-space wave function for state j i. Then H jni = H jni = En jni ) jni = ei jni ) jn t0 = 0 ti = e itH=hjni = e itEn=hjni = 2E t En t n eitEZ =h jni = ei( h + ) jni = ei( hn + )jn tZ = 0 ti 0 2 En t ) d3xj~ ih~ j jn t0 = 0 ti = ei( h + ) d3xj~ ih~ j jn t0 = 0 ti x x x x Z Z x x ) d3xh~ jn t0 = 0 ti j~ i = d3xei( 2Ehn t + )h~ jn t0 = 0 tij~ i x x . . (~ ) = h~ j i.Is the momentum-space wave function for the p p 0 0 0 . prove that the wave function for a spinless nondegenerate system at any given instant of time can always be chosen to be real.1) So if we choose at any instant of time = . (b) The wave function for a plane-wave state at t = 0 is given by px a complex function ei~ ~=h. p ~ ) x n (~ . 4 Symmetry in Quantum Mechanics 2E t t) = ei( hn + ) n(~ t): x (4.2) px px The wave functions n(2x) = e i~ ~=h and n (x) = ei~ ~=h correspond to the p and so there is degeneracy since these correspond same eigenvalue E = 2m to di erent state kets j~i and j .1 (a) Assuming that the Hamiltonian is invariant under time reversal. . h2 r (x) = E (x) ~ n 2m 2m n px px ) n (x) = Aei~ ~=h + Be i~ ~=h (4. So we cannot apply the previous result. Why does this not violate time-reversal invariance? (a) Suppose that jni in a nondegenerate energy eigenstate. p that is. 0 4. (b) In the case of a free particle the Schrodinger equation is p2 jni = E jni ) .

(. Ri )e i Ri ika n k (x + a) = e n k (x) . SYMMETRY IN QUANTUM MECHANICS 0 0 0 95 0 time-reversed state j i given by (~ ).~ ).~ h ~j ~ p p p . or (. (4.4) 0 So pp = p p ~j~ i (4:4) . where a is the lattice constant.~ j i = (. Prove that any Bloch function can be written as. n labels the band. =a =a].~ j i j~ i p ~ p p ) h~ j j i = ~(~ ) = h. You know that the energybands in solids are described by the so called Bloch functions n k full lling.6) 4. ~ i p p 0 0 0 (4.5) up to a phase factor.h ~j~j ~ i ) p p 1 j ~ i ) ~ 1 = . .~ ): p p p p 0 0 0 0 0 0 0 0 0 0 (4. p i = d3p h.3) For the momentum it is natural to require h j~j i = .4. So nally Z Z j i = d3p h~ j i j . (~ ).3 Read section 4.~ )? p p p p Justify your answer. X ikR n k (x) = n (x . and the lattice momentum k is restricted to the Brillouin zone . In the momentum space we have Z Z j i = d3p h~ j ij~ i ) j i = d3p (~ )j~ i p p p p Z Z p p ) j i = d3p h~ j ij~ i] = d3p h~ j i j~ i: p p 0 0 0 0 0 0 0 0 0 0 0 0 (4.~ j~ i ) j~ i = j .3 in Sakurai to refresh your knowledge of the quantum mechanics of periodic potentials.

Ri . i:e: prove Z dx ? (x . Rj )e j . a)]e i e Ri Ri Ri a=Rj ika X ikR = e (4. Show that the Wannier functions are corresponding to di erent sites and/or di erent bands are orthogonal. Ri)eikRi satisfy the same relation X X ikR ik(R a) ika n (x + a .96 where the sum is over all lattice vectors Ri. .). The functions n are called Wannier functions.10) n . n k (x): Rj which means that it is a Bloch function X ikR n k (x) = n (x . but the construction generalizes easily to three dimensions. Ri )e i = n x . Ri )e i : Ri (4. =a . The de ning property of a Bloch function n k (x + a) = e ika n k (x) is (4. Ri )e i Ri Z =a Z =a X ) dke ikRj n k (x) = (x .9) The last relation gives the Bloch functions in terms of Wannier functions. To nd the expansion of a Wannier function in terms of Bloch functions we multiply this relation by e ikRj and integrate over k. (Ri . . and are important in the tight-binding description of solids. . Ri) n(x . Rj ) m ij mn Hint: Expand the n s in Bloch functions and use their orthonormality properties. X ikR n k (x) = n (x . (In this simble one dimensional problem Ri = ia.7) We can show that the functions PRi n(x .8) n (x . =a . Ri) eik(Ri Rj )dk (4.

. So Z =a X 2 dke ikRj n k (x) = n (x . =a .4. Rj i j =a = ij 2a (4. . 0 0 0 . R ) Ri . k )dkdk (2 )2 e Z =a 2 = a 2 mn eik(Ri Rj )dk = 2a mn ij : (4.12) So using the orthonormality properties of the Bloch functions Z dx m(x .11) where in the last step we used that Ri . Z =a ik(R R ) e i j dk =a . Rj ) Z Z Z a2 ikRi ik Rj = n k (x)dkdk dx m k (x)e 2e (2 ) Z Z a2 Z ikRi ik Rj = m k (x) n k (x)dxdkdk (2 )2 e Z Z a2 ikRi ik Rj = mn (k . with n 2 Z . . Ri) = 2a =a e ikRi n k (x)dk (4. 0 0 0 . x Using the time-reversal invariance prove ~ hLi = 0 for any energy eigenstate. . Rj )] = i(R . 0 0 0 . Ri ) ij a =a R Z =a i ) n(x . Ri ) n(x . SYMMETRY IN QUANTUM MECHANICS But 97 eik(Ri Rj ) =a = 2 sin =a(Ri . .4 Suppose a spinless particle is bound to a xed center by a potential V (~ ) so assymetrical that no energy level is degenerate. . Rj = na. (This is known as quenching of orbital angular momemtum.) If the wave function of such a nondegenerate eigenstate is expanded as XX l m Flm(r)Ylm( ) . 4. .13) (2 ) .

.hnjLjni ) ~ ~ ~ :16) ~ hnjLjni = 0 : We have (4.hnjLjni (4= .19) .1) Fl m (r)e 0 0 0 0 : (4.14) So if jni is an energy eigenstate with eigenvalue En we will have H jni = H jni = En jni: (4.1) 0 .98 what kind of phase restrictions do we obtain on Flm(r)? Since the Hamiltonian is invariant under time reversal H = H: (4. 0 0 jni = Z d3xj~ ih~ jni = x x Z d3xh~ jni j~ i x x . 0 0 ) Fl ml 0 .18) x So if we use h~ jni = Pl Pm Flm(r)Ylm( ) X X Flm(r)Ylm ( ) = ei Flm(r)Ylm( ) ml ml (S 4:4:57) X m Y m ( ) = ei X F (r)Y m ( ) ) Flm(r)(.17) Z = d3xh~ jni j~ i (4= ei jni ) x x :16) ~ ~ ~ hx j jni = hx jni = ei hx jni: (4. m (r)(. m i m (r) = (.1) m m l l = ei Flm(r) m m l l 0 0 0 . Hence jni jni = ei jni: ~ (4.1)m Yl m ( )d = ei Ylm Flm(r)Ylm ( )d ml X ml m X ) Flm(r)(.16) For the angular-momentum operator we have from (S-4.1) l lm l ml ml Z Z X X ) Ylm Flm(r)(. m 0 = ei Fl m (r) ) Fl 0 0 ml 0 0 0 .53) ~ ~ hnjLjni = .15) If there is no degeneracy jni and jni can di er at most by a phase factor. 0 0 .4.

) Is this Hamiltonian invariant under time reversal? How do the normalized eigenstates you obtained transform under time reversal? For a spin 1 system l = 1 and m = . SYMMETRY IN QUANTUM MECHANICS 99 4. m)(2 + m) n m+1 + 2 h (1 + m)(2 . For the operator Sz we have Sz jl mi = hmjl mi ) hlnjSz jl mi = hmhnjmi ) (Sz )nm = hm nm (4.1 0 +1. . Sy2): Solve this problem exactly to nd the normalized energy eigenstates and eigenvalues. .4. .21) Sx = 4 @ A @ 1 A 1 0 1 2 0 2 2 2 . (A spin-dependent Hamiltonian of this kind actually appears in crystal physics.1 0 0 1 For the operator Sx we have S+ + S j1 mi = 1 S j1 mi + 1 S j1 mi ! Sxjl mi = 2 + 2 2 1 h1 njS j1 mi + 1 h1 njS j1 mi h1 njSxj1 mi = 2 + 2 q (S 3:5:39) 1 q 1 = h (1 .20) So 0 1 0 1 1 0 0 1 0 0 Sz = h B 0 0 0 C ) Sz2 = h2 B 0 0 0 C @ A @ A 0 0 . m) n m 1: 2 So 1 0 0 0 p 1 0 2 p C h p 0 0 0 Sx = h B 0 0 2 A + 2 B 2 p 0 C 0 @ A 2@ 0 0 0 0 2 0 p 0 1 0 0 p 2 p C = hB 2 p 0 2 A) @ 2 0 2 0 0 1 01 1 1 2 0 2 2 2 h2 B 0 4 0 C = h2 B 0 0 0 C : 2 (4. .5 The Hamiltonian for a spin 1 system is given by 2 H = ASz2 + B (Sx . .

I ) = 0 ) det B 0 . 2 p 2 C ) 0 A @ 0 . B h2)(Ah2 .26) . h B 0 . nd 1 0 .2 .c B2 + cA = 0 c=0 A (4. .22) Thus the Hamiltonian can be represented by the matrix 0 1 A 0 B H = h2 B 0 0 0 C : @ A B 0 A (4.25) =0 0 10 1 ( A 0 B B 0 0 0 C B a C = 0 ) aA + cB = 0 h2 @ A@ b A aB + cA = 0 c B 0 A ( ( a = . )2(. 0 A=0 @ 2 2.1 2 C 1 0 A: 0 1 2 (4. B ): (4.100 In the same manner for the operator Sy = S+2iS we 1 0 0 p 2 p 0 p h Sx = 2i B .24) To nd the eigenstate jn c i that corresponds to the eigenvalue solve the following equation 0 10 1 0 1 A 0 B CB a C a 2 B 0 0 0 A@ b A = B b C: h @ c@ A B 0 A c c For 1 c we have to (4. (B h2)2 = 0 ) (Ah2 .c B ) a = 0 A ) .23) To nd the energy eigenvalues we have to solve the secular equation 0 2 1 Ah . ) . 0 B h2 C det(H .4 0 C = h2 B 0 @ A @ 4 2 0 . ) + (B h2)2 = 0 ) (Ah .2 0 2 2 Sx = .2 1 . Bh h 0 2Ah 2 i ) (Ah2 . . 2 0 0 1 0 1 2 2 . + B h2) = 0 ) 1 = 0 2 = h2(A + B ) 3 = h2(A .

30) 0 1 0 1 c norm: 1 B 0 C = p B 1 C) jnA+B i = @ A @ 0 A 2 . B ) : aB + cA = c(A . B ) we have 0 10 1 0 A 0 B a B 0 0 0 C B b C = (A .28) 0 1 0 1 c norm: 1 1 jnA+B i = B 0 C = p2 B 0 C ) @ A @ A c 1 1 1 jnA+B i = p2 j1 +1i + p2 j1 .1 .29) For = h2(A . (4.1i: . SYMMETRY IN QUANTUM MECHANICS So 101 0 1 0 1 0 norm: 0 jn0i = B b C = B 1 C ) @ A @ A 0 0 jn0i = j10i: = h2(A + B ) 1 0 aC b A = (A + B ) B @ c (4.4.31) .1i: (4.c 1 1 jnA B i = p2 j1 +1i . B ) B @ A@ A @ B 0 A c ( . p2 j1 .c ) a == 0 b So 1 8 > aA + cB = a(A .27) In the same way for 0 10 A 0 B CB B 0 0 0 A@ @ B 0 A ( =c ) a= 0 b So 1 8 a C > aA + cB = a(A + B ) < b A)> 0 = b(A + B ) : aB + cA = c(A + B ) c (4. B ) c (4. B ) a C)< bA > 0 = b(A .

.34) (4.4. p j1 . p j1 +1i 2 2 = . . .37) (4:33) (4.1i . Sy2 1) = A Sz 1 Sz 1 + B ( Sx 1 Sx 2 = ASz2 + B (Sx .32) 1) To nd the transformation of the eigenstates under time reversal we use the relation (S-4. Sy .1)mjl .1i .33) jn0i = :33) j10i (4= j10i = jn0i 1 1 jnA+B i = p j1 +1i + p j1 .35) 1 1 = .1i 2 2 (4:33) (4. 1 2 = A Sz2 1 + B ( Sx 1 . Sy2) = H: . . p j1 . .jnA+B i 1 1 = .mi: So (4. (4. 1.58) jl mi = (. 1 Sy . 1 1 jnA B i = p2 j1 +1i . p2 j1 .1i + p j1 +1i 2 2 = jnA B i: .36) (4. .38) . (4.102 Now we are going to check if the Hamiltonian is invariant under time reversal H .

APPROXIMATION METHODS 103 5 Approximation Methods 5. m! ) 2h r m! ipy ay (y + m! ) 2h r m! (y . (c) Solve the H0 + V problem exactly. Compare with the perturbation results obtained in q (b). De ne step operators: r m! ipx (x + m! ) ax 2h r m! ipx ax (x . p p You may use hn jxjni = h=2m! ( n + 1 + n ):] 0 n n+1 0 n n 1 0 . Find the zeroth-order energy eigenket and the corresponding energy to rst order that is the unperturbed energy obtained in (a) plus the rst-order energy shift] for each of the three lowest-lying states.5.1 Consider an isotropic harmonic oscillator in two dimensions.1) ax ax] = ay ay ] = 1: y y . The Hamiltonian is given by (a) What are the energies of the three lowest-lying states? Is there any degeneracy? (b) We now apply a perturbation V = m!2xy p2 + p2 + m!2 (x2 + y2) y x H0 = 2m 2m 2 where is a dimensionless real number much smaller than unity. ipy ): ay 2h m! From the fundamental commutation relations we can see that y y (5.

104 De ning the number operators Nx ax ax y Ny ay ay y we nd H0 N Nx + Ny = h! .4) h j l0 i + j l1 i + : : : : i .2) Nx j m n i = m j m n i Ny j m n i = n j m n i ) N j m n i = (m + n) j m n i so that (5. Full problem: (H0 + V ) j l i = E j l i Unperturbed problem: H0 j l0 i = E 0 j l0 i Expand the energy levels and the eigenkets as E = E0 + 1 + 2 + : : : j l i = j l0 i + j l1 i + : : : so that the full problem becomes h i (E 0 . 1 ) H0 = h!(N + 1): I. H0 ) j l0 i + j l1 i + : : : = (V .3) H0 j m n i = Em n j m n i = h!(m + n + 1) j m n i: (a) The lowest lying states are state degeneracy E0 0 = h! 1 E1 0 = E0 1 = 2h! 2 E2 0 = E0 2 = E1 1 = 3h! 3 (b) Apply the perturbation V = m!2xy. 2 : : :) (5. energy eigenkets are also eigenkets of N : (5. 1.e.

h m0 j l0 i is not necessarily kl since only states corresponding to di erent eigenvalues have to be orthogonal! Insert a 1: X 0 h m j V j k0 ih k0 j l0 i = 1h m0 j l0 i: This is the eigenvalue equation which gives the correct zeroth order eigenvectors! Let us use all this: 1. The ground state is non-degenerate ) 1 = h0 00 k D 2 0 j V j 0 0 i = m!2h 0 0 j xy j 0 0 i h 0 0 j (ax +ax)(ay +ay ) j 0 0 i = 0 y y 2. h V = m!2xy = m!2 2m! (ax+ax)(ay +ay ) = h! (axay +axay +axay +axay ) 2 and y y y y y y ax j m n i = m j m . APPROXIMATION METHODS To 1'st order: (E 0 . h 0 1 j V j 1 0 i. Wind back to (5. 1 j l0 i ) 1h l0 j l0 i = 1 = h l0 j V j l0 i (5.7) Now. j 0 1 i. We need the matrix elements h 1 0 j V j 1 0 i. h 0 1 j V j 0 1 i. h 1 0 j V j 0 1 i. Multiply with h l0 j to nd 1 ) j l0 i: 105 (5. 1h m0 j l0 i = h m0 j V j l0 i: 1 j l0 i ) (5.6) In the degenerate case this does not work since we're not using the right basis kets.5. First excited state is degenerate j 1 0 i. H0 j l1 i = 0 = h l0 j V .5) and multiply it with another degenerate basis ket h m0 j E 0 . H0) j l1 i = (V . 1 n i p ax j m n i = m + 1 j m + 1 n i etc: y p . H0 j l1 i = 0 = h m0 j V .5) h l0 j E 0 .

j 0 2 i. 2) @ A 0 1 . j 1 0 i) E = h!(2 . .10) 2 p (where the 2 came from going from level 1 to 2 in either of the oscillators) and thus to get the eigenvalues we evaluate 0 . 1 0 1 0 = det B 1 . so we need the corresponding 9 matrix elements. 1 C = . j 1 1 i. . 1) + = (2 .106 Together this gives V10 10 = V01 01 = 0 V10 01 = h! h 1 0 j ax ay j 0 1 i = 2 h! h 1 0 j a a j 0 1 i = V01 10 = 2 x y y y h! 2 h! : 2 (5. The V -matrix becomes . The second excited state is also degenerate j 2 0 i. ( 2 . 2 ): (5.8) ! h! 0 1 2 1 0 and so the eigenvalues (= 1) are h! : 1= 2 To get the eigenvectors we solve ! ! ! 0 1 x = x 1 0 y y and get 1 E+ = h!(2 + 2 ) j i+ = p2 ( j 0 1 i + j 1 0 i) 1 j i = p2 ( j 0 1 i . However the only nonvanishing ones are: h! V11 20 = V20 11 = V11 02 = V02 11 = p (5.9) 3.

y )2) : (5. )y 2]: m 2 p p So we get one oscillator with !x = ! 1 + and another with !y = ! 1 . so that (x . y )2) + (x + y )2 . ): . py ) are canonically conjugate. By the same method E+ = h!(3 + ) E0 = 3h! E = h!(3 .5. j 2 0 i + j 0 2 i) p 1 j i = 2 ( j 2 0 i . 0 0 0 0 0 0 0 0 0 0 0 0 x 0 1 p (x + y) px 0 1 p (px + py ) 0 0 In these new variables the problem takes the form 2 H = 21 (px2 + py2) + m! (1 + )x 2 + (1 . The energy levels are: 0 0 0 0 0 0 E0 0 = h! p E1 0 = h! + h!x = h!(1 + 1 + ) = 1 = h!(1 + 1 + 1 + : : :) = h!(2 + 2 ) + O( 2) 2 0 . p ): y p py p x y (5. (c) To solve the problem exactly we will make a variable change. The potential is h1 i m!2 2 (x2 + y2) + xy = # " 2 1 ((x + y )2 + (x .11) = m! 4 4 Now it is natural to introduce 2 2 1 (x . APPROXIMATION METHODS which means that the eigenvalues are f0 as above we get the eigenvectors 107 h!g. y) 1 (p . (x .12) 2 2 Note: x px] = y py ] = ih. p 1 j i+ = 2 ( j 2 0 i + 2 j 1 1 i + j 0 2 i) 1 j i0 = p2 (. px ) and (y . 2 j 1 1 i + j 0 2 i) . .

(E1 . E ) . 2 ) + O( 2) h! + 2h!x = : : : = h!(3 + ) + O( 2) h! + h!x + h!y = : : : = 3h! + O( 2) h! + 2h!y = : : : = h!(3 . E 2 .2 A system that has three unperturbed states can be represented by the perturbed Hamiltonian matrix 0 1 E1 0 a C B 0 E1 b A @ a b E2 where E2 > E1. Use the second-order nondegenerate perturbation theory to calculate the perturbed eigenvalues.13) So rst order perturbation theory worked! 5. Finally. E b = a E2 . nd the exact result by diagonalizing the Hamiltonian: E1 . (jbj2 + jaj2) = 0 i. use the second-order degenerate perturbation theory. E )] = = (E1 . E1. (a) First. E )2(E2 . (jaj2 + jbj2) = 0 ) + E1 + E2 2 .14) So. Compare the three results obtained. (Is this procedure correct?) Then diagonalize the matrix to nd the exact eigenvalues. The quantities a and b are to be regarded as per- turbations that are of the same order and are small compared with E2 . E )(E2 . E )(E2 .108 E0 1 E2 0 E1 1 E0 2 = = = = 1 h! + h!y = : : : = h!(2 . E )(jbj2 + jaj2): (5. E h b i = (E1 . E ) . ) + O( 2): 0 0 0 0 0 (5. E 0 a 0 = 0 E1 . a (E1 . E ) (E1 . E = E1 or (E1 . E ) . E E + jaj2 + jbj2 = E = E1 + E2 1 2 2 2 .e. (E1 + E2)Es E1E2 . jbj2 + a 0 .

E2 (: : :) = E2 . 2 k=1 E1 . jaj . E )2 + : : : = 2 2 1 2 2 + jbj2 = E1 + jaj . E E1 2 2 + jb 2 E = E1 + E2 . Ej : 0 0 E1 2 2. E2 1 + 2 (jaj2 + jbj2)( E .15) 2 2 Since jaj2 + jbj2 is small we can expand the square root and write the three energy levels as: E = E1 2 1 E = E1 + E2 + E1 . 1 k=3 E3 . 1 2. E2 k=2 E2 . But the 2'nd order k shifts are (2) 1 (2) 2 (2) 3 2 X jVk1 j2 3 1 2 = jh Ej V jE ij = E jaj E 0 0 1.17) . Ej : 2 2 E1 2 (5. E2 E1 . Ek 2 2 2 + jb 2 X jVk3 j2 = = E jaj E + E jbj E = . Ek = 6 6 6 (5. APPROXIMATION METHODS s E1 . jaj . The basis we use is 0 1 0 1 0 1 1C 0C 0 j1i = B 0 A j2i = B 1 A j3i = B 0 C: @ @ @ A 0 0 1 0 1 0 0 aC The matrix elements of the perturbation V = B 0 0 b A are @ a b 0 h 1 j V j 3 i = a h 2 j V j 3 i = b h 1 j V j 2 i = h k j V j k i = 0: Since (1) = h k j V j k i = 0 1'st order gives nothing. 2 1.5. E2 2 + jaj2 + jbj2: E1 + E2 = 109 (5. Ek 2 2 X jVk2 j jh 3 j V j 2 ij = jbj2 = 0 0 = E1 . E1 .16) (b) Non degenerate perturbation theory to 2'nd order.

Using nondegenerate perturbation theory we expect only the correction to E2 (i.18) leaves us with h i (E 0 . H0) j l2 i = V j l1 i . so 1'st order pert. (2)) 3 to give the correct result.e. Inserting the expansions in (5. (2)h m0 j l0 i: k=D 6 k . Note that the superscript index in a bra or ket denotes which order it has in the perturbation expansion. The problem we want to solve is (H0 + V ) j l i = E j l i using the expansion j l i = j l0 i + j l1 i + : : : E = E 0 + (1) + (2) + : : : (5.thy. will again give nothing. (2) j l0 i: To get rid of the left hand side. (2) : : :) j l0 i + j l1 i + : : : : (5. multiply with a degenerate bra h m0 j (H0 j m0 i = E 0 j m0 i) h m0 j E 0 . H0 j l1 i = h k0 j V j l0 i X 0 ih 0 V 0 (5.18) where H0 j l0 i = E 0 j l0 i. H0) j l1 i = (V . and indeed this turns out to be the case.110 The unperturbed problem has two (degenerate) states j 1 i and j 2 i with energy E1. and one (non-degenerate) state j 3 i with energy E2. Multiply this from the left with a bra h k0 j from outside the deg.19) At rst order in the perturbation this says: (E 0 . H0 j l2 i = 0 = h m0 j V j l1 i . We have subspace is 0 0 to go to 2'nd order.20) ) j l1 i = j k E 0k. H0 ) j l0 i + j l1 i + : : : = h i (V . (1)) j l0 i where of course (1) = 0 as noted above.jE j l i : This expression for j l1 i we will use in the 2'nd order equation from (5. (1) . The V -matrix for the degenerate ! 0 0 . (c) To nd the correct energy shifts for the two degenerate states we have to use degenerate perturbation theory.19) (E 0 . Since the (sub-) problem we are now solving is 2-dimensional we expect to nd two solutions corresponding to l = 1 2. subspace h k0 j E 0 . Di erent solutions to the full problem are denoted by di erent l's.

E 1 3 1 2 1 3 1 2 With this explicit expression for M . = = (jaj2 . (jaj2 + jbj2) ) = 0 jaj2 + jbj2 2 2 (2) = 0 (2) = jaj + jbj ) 1 (5. solve the eigenvalue equation (de ne = (2)(E1 . ) . ab 0 = det a b jbj2 . E2). E3 1. )(jbj2 . 2 2 V23V31 = a b jV23j = jbj2 : M21 = E .E : . E 0 E . E 0 E . E M22 = E . Ek n D k=D 2 6 Maybe it looks more familiar in matrix form X Mmn xn = (2)xm n D 2 where Mmn = xm X h m0 j V j k0 ih k0 j V j n0 i E 0 .21) 2 E . and take out a common factor E1 1 E2 ) ! jaj2 . Ek k=D 6 111 (2)h m0 j l0 i: To make this look like an eigenvalue equation we have to insert a 1: X X h m0 j V j k0 ih k0 j V j n0 i 0 0 h n j l i = (2)h m0 j l0 i: E 0 . 2 1 .20) for j l1 i we get X h m0 j V j k0 ih k0 j V j l0 i = E 0 .5. jaj2jbj2 = = 2 . Evaluate M in the degenerate subspace basis D = f j 1 i j 2 ig 2 M12 = EV13V32 0 = E ab E M11 = EV13V31 0 = E jaj E 1 . 1 2 . E3 1. APPROXIMATION METHODS Inserting the expression (5. Ek k=D = 0 h m j l0 i 6 are expressed in the basis de ned by j l0 i.

5. Everything is as we would have expected. . m) h We need the matrix elements Vnm Vnm = h n j . and now we see that degenerate perturbation theory leads to the correct shifts for the other two levels.112 From before we knew the non-degenerate energy shift. Show that the t ! 1 ( nite) limit of your expression is independent of time.3 A one-dimensional harmonic oscillator is in its ground state for t < 0. Em = !(n . 5. = h!(n + 1 ) 2 where we get cn (t) from its di . obtain the probability of nding the oscillator in its rst excited state for t > 0.22) .F0xe t= (F = . We want to calculate X j t i = cn (t)e Ent=h j n i . @V ) @x At t = 0 the system is in its ground state j 0 i = j 0 i. . For t 0 it is subjected to a time-dependent but spatially uniform force (not potential!) in the x-direction. Is this reasonable or surprising? (b) Can we nd higher excited states? q You may use hn jxjni = h=2m! (pn + 1 n n+1 + pn n n 1 ):] 0 0 0 F (t) = F0e .F0e t= 2m! ( m n m 1 + m + 1 n m+1): . F0s t= j m i = . eqn. 5.15): X @ ih @t cn (t) = Vnm ei!nmtcm(t) m Vnm = h n j V j m i !nm = En . (a) The problem is de ned by 2 2 H0 = 2pm + m!22x V (t) = . (S. . 0 En n (5.F0e t= h n j x j m i = xe p h p = . (a) Using time-dependent perturbation theory to rst order. . t= .

5. APPROXIMATION METHODS
Put it back into (5.22)

113

s @ c (t) = ;F e t= h pn + 1e i!tc (t) + pnei!tc (t) : ih @t n 0 n+1 n 1 2m! Perturbation theory means expanding cn(t) = c(0) + c(1) + : : :, and to zeroth n n order this is @ c(0)(t) = 0 ) c(0) = n0 n @t n To rst order we get Zt X c(1)(t) = i1 dt Vnm (t )ei!nmt c(0) = n m h 0 m s p p h Zt 0 = ; Fh 2m! 0 dt e t= n + 1e i!t c(0) (t) + nei!t c(0)1 (t) n+1 n i We get one non-vanishing term for n = 1, i.e. at rst order in perturbation theory with the H.O. in the ground state at t = 0 there is just one non-zero expansion coe cient s F0 h Z t dt ei!t t = p1 c(1)(t) = ; ih 2m! 1 10 = 1 0 s # " F0 h 1 e(i! 1 )t t = ; ih 2m! i! ; 1 0 s h 0 = Fh 2m! 1 1 1 ; e(i! 1 )t i i! ; and X h h j t i = c(1)(t)e iEnt j n i = c(1)(t)e iE1t j 1 i: 1 n
; ; ; 0 0
0

0

;

;

0

0

;

0

0

;

0

;

;

0

;

;

;

The probability of nding the H.O. in j 1 i is As t ! 1

n

jh 1 j t ij2 = jc(1)(t)j2: 1

s F0 h 1 = const: c(1) ! ih 2m! i! ; 1 1 This is of course reasonable since applying a static force means that the system asymptotically nds a new equilibrium.

114 (b) As remarked earlier there are no other non-vanishing cn's at rst order, so no higher excited states can be found. However, going to higher order in perturbation theory such states will be excited.
1 5.4 Consider a composite system made up of two spin 2 objects. for t < 0, the Hamiltonian does not depend on spin and can be taken to be zero by suitably adjusting the energy scale. For t > 0, the Hamiltonian is given by

~ ~ H = 4 2 S1 S2: h Suppose the system is in j + ;i for t 0. Find, as a function of

time, the probability for being found in each of the following states j + +i, j + ;i, j ; +i, j ; ;i: (a) By solving the problem exactly. (b) By solving the problem assuming the validity of rst-order time-dependent perturbation theory with H as a perturbation switched on at t = 0. Under what condition does (b) give the correct results?
(a) The basis we are using is of course j S1z S2z i. Expand the interaction potential in this basis: ~ ~ S1 S2 = S1x" 2x + S1y S2y + S1z S2z = fin this basisg S 2 = h ( j + ih ; j + j ; ih + j )1 ( j + ih ; j + j ; ih + j )2 + 4 + i2(; j + ih ; j + j ; ih + j )1 (; j + ih ; j + j ; ih + j )2 + # + ( j + ih + j ; j ; ih ; j )1 ( j + ih + j ; j ; ih ; j )2 = " h2 j + + ih ; ; j + j + ; ih ; + j + = 4 + j ; + ih + ; j + j ; ; ih + + j + 2 ( j + + ih ; ; j ; j + ; ih ; + j + + i ; j ; + ih + ; j + j ; ; ih + + j ) +

5. APPROXIMATION METHODS
+ j + + ih + + j ; j + ; ih + ; j + # ; j ; + ih ; + j + j ; ; ih ; ; j =

115

In matrix form this is (using j 1 i = j + + i j 2 i = j + ; i j 3 i = j ; + i j 4 i = j ; ; i) 01 0 0 01 B C H = B 0 ;1 ;1 0 C : B0 2 2 0C (5.23) @ A 0 0 0 1 This basis is nice to use, since even though the problem is 4-dimensional we get a 2-dimensional matrix to diagonalize. Lucky us! (Of course this luck is due to the rotational invariance of the problem.) Now diagonalize the 2 2 matrix to nd the eigenvalues and eigenkets ! ;1 ; 2 2 2 0 = det 2 ;1 ; = (;1 ; ) ; 4 = + 2 ; 3 =1:

) = 1 ;3
! ! x = x y y 1 ) ;x + 2y = x ) x = y = p2

;1 2 2 ;1

!

! ! x = ;3 x y y 1 ) ;x + 2y = ;3x ) x = ;y = p2 So, the complete spectrum is: 8 > j + + i j ; ; i 12 ( j + ; i + j ; + i with energy < > 1 (j + ;i ; j ; +i : with energy ; 3

= ;3 :

;1 2 2 ;1

!

p

p

2

116 This was a cumbersome but straightforward way to calculate the spectrum. we know that two spin 2 systems add up to one triplet (spin 1) and one singlet (spin 0). i j 3 i = p2 ( j + . t0) : The initial state was j + . and 1 ( j + .3 : E(spin=0) = 2 4 h From Clebsch-Gordan o decomposition we know that j + + i j . i . + i) are spin 1. i . ~ ~ 1 S = 1 (3 states) ) S1 S2 = 2 (h21(1 + 1) . i. In the new basis H is diagonal and time-independent. + i) p p n (5. j . i + j . so we can use the simple form of tthe time-evolution operator: i U (t t0) = exp . i. j . 4 h2 ~ 2 2 S = 0 (1 state) ) S ~ ~ Since H = 4h2 S1 S2 we get h2 E(spin=1) = 4 2 14 = h 4 . i 1 ( j + .24) 3 ~1 S2 = 1 (.e.3h2 = . S2 1 ~2 ~2 We know that S1 = S2 = h2 2 1 2 +1 1 2 = 3h2 so 4 ~ ~ S1 S2 = S2 . 3h2 ) = 1 h2 2 4 : (5. . h H (t . In the new basis n 1 j 1 i = j + + i j 2 i = j . ~ ~ ~ A smarter way would have been to use S = S1 + S2 to nd ~ ~2 ~2 ~ ~ ~ ~ 1 ~ ~2 ~2 S 2 = S 2 = S1 + S2 + 2S1 S2 ) S1 S2 = 2 S 2 . + i) o 1 j 4 i = p2 ( j + . 3h2 2 ! 1 Also. S1 . . i + j . 3h2 ) = .25) . + i) is spin 0! 2 2 Let's get back on track and nd the dynamics.

. + i = 2 . j ti = 0 > > < 1 2 1 4i!t 4i!t 2 > jh + .i . e 4i!t) = 1 (1 . V given by (5. h Ht ( j 3 i + j 4 i) = 2 1 exp . + j t ij2 = 1 (2 . e4i!t .27) jii = j + . i .i cn h t0 dt e Vni (t ): Here we have (using the original basis) H0 = 0. where (5.6. . 5. i + j .26) h: The probability to nd the system in the state j i is as usual jh j t ij2 8 > h+ + j ti = h. i + (e i!t + e3i!t) j . + i) = 1 +exp h 2 h i 1 (e i!t + e3i!t) j + . i t j 3 i + exp 3i t j 4 i = = p h h "2 1 = exp . + i)+ 2 # 3 i t p ( j + . cos4!t) ' 4(!t)2 : : : 4 2 ! .5. j t ij = 4 (2 + e + e ) = 2 (1 + cos4!t) ' 1 .ih t p ( j + . j . i = p2 ( j 3 i + j 4 i): 117 Acting with U (t 0) on that we get 1 i j t i = p exp . APPROXIMATION METHODS the initial state is 1 j + .17): c(0) = ni n Zt i!ni t (1) (t) = . (b) First order perturbation theory (use S. .23) 0 0 0 (5. 4(!t) : : : > > : jh .

5 The ground state of a hydrogen atom (n = 1. how you may compute the angular p distribution of the ejected electron (in terms of and de ned with respect to the z-axis).i j. in correspondence with the exact result. !t): x Using time-dependent perturbation theory. so for a given t: h: !t 1 ) t 5. i) = 1 j . The probability of nding the system in j .l = 0) is subjected to a time-dependent potential as follows: V (~ t) = V0 cos(kz . . i jf i = j . (note: For the initial wave function use 2 1 Z 3e x n=1 l=0 (~ ) = p a0 .2i!t: as the only non-vanishing coe cients up to rst order.27) yields c(0) = c(0) = 1 i + Z (1) = c(1) = .118 Vfi = 2 Vni = 0 n 6= f: Inserting this into (5. +i !ni = En . Show.28) + h 0 dt2 = . obtain an expression for the transition rate at which the electron is emitted with momentum ~. + i) = jc(1)(t) + c(2)(t) + : : : j2 = j2i!tj2 = 4(!t)2 f f to rst order. i or j + + i is thus obviously zero. in particular. The approximation breaks down when !t 1 is no longer valid. Ei = fEn = 0g = 0 h P ( j . whereas for the other two states P ( j + . Discuss brie y the similarities and the di erences between this problem and the (more realistic) photoelectric e ect. i t cf (5. Zr=a0 : .

The initial wave-function is p 1 1 3=2e r=a0 x i (~ ) = p a0 and the nal wave-function is 1 ei~ ~=h: px x f (~ ) = L3=2 The perturbation is h i V = V0 ei(kz !t) + e i(kz !t) = V ei!t + V e i!t: (5. . .5. but you should be able to show that the observable e ects are independent of L. . .29) . We want to nd the transition rate at which the electron is emitted with momentum ~f .6. !t) is turned on. (Ei + h!)) because the atom absorbs a photon h!.30) L a0 y ! y . At t = 0 the perturbation V = V0 cos(kz . . y .5.44) gives us the transition rate 2 wi n = 2h Vni (En . the nal wave function may be taken to be 1 with L very large. . x f (~ ) = 2 L3 px ei~ ~=h Time-dependent perturbation theory (S. The matrix element is 2 Vni 2 = V40 eikz ni 2 and Z ikz ~ f j eikz j n = 1 l = 0 i = d3xh ~ f j eikz j x ih x j n = 1 l = 0 i = e ni = h k k Z e i~f ~ k x 1 1 3=2 = d3x L3=2 eikx3 p a e r=a0 = 0 Z 1 k x = 3=2p 3=2 d3xe i(~f ~ kx3 ) r=a0 : (5. . APPROXIMATION METHODS 119 If you have a normalization problem.) To begin with the atom is in the n = 1 l = 0 state. .

Finally.7.31) In a comparison between this problem and the photoelectric e ect as discussed in (S. k~z )2 k e 0 n2dnd ! Note that the L's cancel. The angular dependence is in the denominator: ~ f .5.120 So eikz ni is the 3D Fourier transform of the initial wave-function (and some constant) with ~ = ~ f . k~z . 2kjkf jcos : (5. k~z 2 = (jkf jcos . On the other hand we did not make any dipole approximation but performed the x-integral exactly. w has no dependence on the azimuthal angle .7) we note that since there is no polarization vector involved.31). We need to get that as a function of ~f = h~ f . k) ~z + jkf jsin (cos'~x + sin'~y )]2 = k e e e e = jkf j2cos2 + k2 . 2 64 2 L3 p2 1 wi ~f = 2h V40 L3a5 h 1 i4 (2 hf)3 d dpf : p 0 a2 + (~ f . As in (S. the volume p k element is dn n2dnd = n2d dp dpf : f Using 2 2 2 2 p kf = f2 = n (22 ) L h we get dn = 1 2L2pf = 2 h L2pf = L dp 2n (2 h)2 Lp (2 h)2 2 h which leaves f f 2 L3kf L3p2 = (2 )3h d dpf = (2 hf)3 d dpf and this is the sought density. . 5. 2kjkf jcos + jkf j2sin2 = 2 = kf + k2 . That can be extracted from (Sakurai problem q k e 5.39) 1 64 2 eikz ni = L3a5 h 1 i4 0 a2 + (~ f . k~z )2 k e 0 The transition rate is understood to be integrated over the density of states.

Sign up to vote on this title
UsefulNot useful