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 .

.

FUNDAMENTAL CONCEPTS 5 1 Fundamental Concepts 0 1.1. a ) 0 is a null operator.) 1. that is.2 A spin 1 system is known to be in an eigenstate of S n with 2 eigenvalue h=2.1 Consider a ket space spanned by the eigenkets fja ig of a Hermitian operator A. . (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. What is the probability of getting +h=2? (b) Evaluate the dispersion in Sx. ~ ^ 1. (a) Prove that Y (A . h(Sx . 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) What is the signi cance of Y (A . hSx i)2i: (For your own peace of mind check your answers for the special cases = 0. =2. a ) ? a =a 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. There is no degeneracy.3 (a) The simplest way to derive the Schwarz inequality goes as follows. and .

hxi) hx j i = (2 d 0 0 0 .4 (a) Let x and px be the coordinate and linear momentum in one dimension. (x . 0 0 h( x)2i h( p)2i = h : 2 q 1. in agreement with (b). Evaluate the commutator (c) Using the result obtained in (b). (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 . Evaluate the classical Poisson bracket x F (px)]classical : (b) Let x and px be the corresponding quantum-mechanical operators this time. 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.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. prove that x exp iph a jx i 0 x x exp iph a : (xjx i = x jx i) 0 0 0 .

What is the corresponding eigenvalue? 1. QUANTUM DYNAMICS 7 is an eigenstate of the coordinate operator x. mc Sz = !Sz write the Heisenberg equations of motion for the time-dependent operators Sx(t). Solve them to obtain Sx y z as functions of time.1 Consider the spin-procession problem discussed in section 2. Sy (t).1 in Jackson. and Sz (t). 2. exp ix h 2 Quantum Dynamics 2. Using the Hamiltonian eB H = .2.5 (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. (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. Evaluate x(t) x(0)] : .2 Let x(t) be the coordinate operator for a free particle in one dimension in the Heisenberg picture. It can also be solved in the Heisenberg picture.

2 x 0 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. Evaluate the correlation function explicitly for the ground state of a one-dimensional simple harmonic oscillator. Under what condition would this happen? 2. 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 = 0.4 (a) Write down the wave function (in coordinate space) for the state ipa exp . .8 2. Does this probability change for t > 0? 2.5 Consider a function. .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.h j0i: You may use 2 !3 x 25 1 hx j0i = 1=4x0 1=2 exp 4. de ned by C (t) = hx(t)x(0)i where x(t) is the position operator in the Heisenberg picture.

(b) Prove the minimum uncertainty relation for such a state. 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. Find the most probable value of n. in general. a complex number. 2. (a) Construct a linear combination of j0i and j1i such that hxi is as large as possible.6 Consider again a one-dimensional simple harmonic oscillator. that is. QUANTUM DYNAMICS 9 2. (a) Prove that j i = e =2e a j0i is a normalized coherent state. (c) Evaluate h( x)2i as a function of time using either picture. and l is the displacement distance) to the ground state. n=0 . Do the following algebraically.2.j j 2 y 1 Show that the distribution of jf (n)j2 with respect to n is of the Poisson form. . hence of E . without using wave functions.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. (b) Suppose the oscillator is in the state constructed in (a) at t = 0. (c) Write j i as X j i = f (n)jni: . (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.

and (x) the Dirac delta function. r etc.9 Prove the relation d (x) = (x) dx 0 dteiEt=hK (x y t 0) = A X sin(nrx) sin(nry) r E . ).) 2.) 2. 2.11 The Lagrangian of the single harmonic oscillator is 1 _ L = 2 mx2 . 1 m!2x2 2 . 2. moving in a potential is given by: K (x y E ) = Z 1 where A is a constant.8 The quntum mechanical propagator. 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 . for a particle with mass m. (a) What is the potential? (b) Determine the constant A in terms of the parameters describing the system (such as m. (Hint: study the e ect on testfunctions.10 (e) Show that the coherent state j i remains coherent under timeevolution and calculate the time-evolved state j (t)i. (Hint: directly apply the time-evolution operator.10 Derive the following expression i m! h Scl = 2 sin(!T ) (x2 + 2x2 ) cos(!T ) .

xcl(t) be the new integration variable. det . By expanding j+1 in minors show the following recursion formula for the pj : pj+1 = (2 .1) 0 0 0 0 . (Hint: Let y(t) = x(t) . 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 . "2!2)pj .1 7 where n = 6 .nT n) Z dN ye nT n . (c) Show that Z dy1 : : : dyN exp(. Write the symmetric 4 . 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) . =p N=2 Hint: Diagonalize by an orhogonal matrix.nT n) G = Nlim 2 ih" 1 N 2 3 6 y.2.) (b) Show that G can be written as m (N2+1) Z dy : : : dy exp(. De ne j j matrices j that consist of the rst j rows and j columns of N and whose determinants are pj . pj 1 j = 1 ::: N (2. yj ) . 7 and nT is its transpose. 5 yN !1 matrix .] ih" (d) Let 2m N det det N pN .

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. ~ ~x ~ p~ (d) Show that r %(~ ) = h1 (S ~)% where p is the momentum oper~ ator. 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 . Indicate how we may nd V (r). ~ (c) Show that S is a vector operator.4 Consider a particle with an intrinsic angular momentum (or spin) of one unit of h. what is the l-value? If ~ not. 3. Using the ladder operator method express all .3. Determine S ! ~ ~ (b) Show that L and S commute. 2 " ~ ~ ~ u~ = 1 .5 We are to add angular momenta j1 = 1 and j2 = 1 to form j = 2 1 and 0 states. (One example of such a particle is the %meson). Quantum-mechanically. (a) Show explicitly that in nitesimal rotations of %i(~ ) are obtained x by acting with the operator ~ i^ ~ ~ where L = h r r. 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 ) " (3.1) 3.

Explicitly write T (1)0 in 1 terms of Ux y z and Vx y z . 3. . You may. y2) as components of a spherical (irreducible) tensor of rank 2.2) 3. 1 1 jj = 1 m = 1i = p j+ 0i . p j0 +i : : : (3. (b) Prove. and (x2 . for instance. j) jd(mm ( )j2m 0 1 for any j (integer or half-integer) then check your answer for j = 2 .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 . Write down explicitly T (2) 1 0 in terms of Ux y z 2 and Vx y z .7 (a) Evaluate j X m= j . 1): 2 2 0 0 Hint: This can be proved in many ways. j X m= j . xz. Write your answer as 2 2 where + and 0 stand for m1 2 = 1 0 respectively.8 (a) Write xy. examine the rotational properties of Jz2 using the spherical (irreducible) tensor language.16 (nine) j m eigenkets in terms of jj1j2 m1m2i. (j m2jdm)m( )j2 = 1 j (j + 1) sin + m 2 + 1 (3 cos2 . (b) Construct a spherical tensor of rank 2 out of two di erent ~ ~ vectors U and V .] 3.

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

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

5.1 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. 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. Compare with the perturbation results obtained in q (b). APPROXIMATION METHODS 19 Solve this problem exactly to nd the normalized energy eigenstates and eigenvalues. (c) Solve the H0 + V problem exactly. 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 . 5.) Is this Hamiltonian invariant under time reversal? How do the normalized eigenstates you obtained transform under time reversal? 5 Approximation Methods 5. (A spin-dependent Hamiltonian of this kind actually appears in crystal physics.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 .

Finally. . use the second-order degenerate perturbation theory. Find. For t 0 it is subjected to a time-dependent but spatially uniform force (not potential!) in the x-direction. Show that the t ! 1 ( nite) limit of your expression is independent of time.4 Consider a composite system made up of two spin 2 objects. 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 + . The quantities a and b are to be regarded as perturbations that are of the same order and are small compared with E2 . obtain the probability of nding the oscillator in its rst excited state for t > 0).i for t 0. 5. 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 . E1. j . Compare the three results obtained.3 A one-dimensional harmonic oscillator is in its ground state for t < 0. F (t) = F0e . t= (a) Using time-dependent perturbation theory to rst order. as a function of time. the probability for being found in each of the following states j + +i.20 where E2 > E1. Use the second-order nondegenerate perturbation theory to calculate the perturbed eigenvalues. (Is this procedure correct?) Then diagonalize the matrix to nd the exact eigenvalues.i.i: (a) By solving the problem exactly. j + . 1 5. +i. for t < 0. j .

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

22 .

Part II Solutions 23 .

.

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

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

9) The eigenvalues and eigenfuncions of this operator can be found if one solves the secular equation ! h=2 cos .h=2 cos . FUNDAMENTAL CONCEPTS 27 (b) Evaluate the dispersion in Sx.4.7) : = 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. cos = a 2 sin cos = a tan 2 : (1. h sin2 = 0 ) 2 .(S-1. h=2 sin ~ ^ det(S n .j) cos + h (j+ih. j. cos (1.ih+j) sin (1.36).3. it can be written in the following way n = ez cos + ex sin ^ ^ ^ (1. b cos = b 2 sin .j + j. I ) = 0 ) det h=2 sin .6) So ~ ^ S n = Sz cos + Sx sin = # (S-1. h = 0 ) = h : (1. =2. = 0 ) 2 2 2 .18)] " # " h (j+ih+j .7) we have that ! sin : ~ n = h cos S ^ 2 sin . h(Sx .8) From (1.10) 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 . that is. h cos2 + 2 . and .ih.) Since the unit vector n makes an angle with the positive z-axis and is ^ lying in the xz-plane. cos 2 b 2 sin2 2 b = a 1 . hSx i)2i: (For your own peace of mind check your answers for the special cases = 0.1.11) sin 2 2 .

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)

.32) is 1 (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: (1.32) The equality sign in this relation holds according to (1.31) (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.h j i=h j i the previous relation will be 31 0 ) h j i.35) Thus the equality sign in the uncertainty relation holds when Ajai = B jai (1.33) On the other hand the right-hand side of (1.30) when Ajai = B jai: (1.1.36) with purely imaginary.30) Notice that the equality sign in the last relation holds when jci = j i + j i = 0 ) j i = . (b) The uncertainty relation is To prove this relation we use the Schwarz inequality (1.29) (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. FUNDAMENTAL CONCEPTS If we now choose = . j i that is if j i and j i are colinear.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. h .

39) So substituting in (1. 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 .4 (a) Let x and px be the coordinate and linear momentum in one dimension. hxi) d 0 0 0 0 0 0 0 (1.i2d2 hx j pj i: hx j xj i = h (1.ih @x hx j i . hpihx j i h h = 2id2 (x .38) @ hx j i = hx j @x = hx j 0 0 0 " # @ ihpix .38) we have 0 0 h hx j pj i = hpihx j i + 2id2 (x . hpi)j i @ = .37) On the other hand hx j pj i 0 But hx j(p . hxi) hx j i = 2id2 hx j xj i ) . hxi)j i = x hx j i . hxihx j i = (x . hpihx j i 0 0 0 0 0 (1. Evaluate the classical Poisson bracket (b) Let x and px be the corresponding quantum-mechanical operators this time. hxi)2 i @x h 4d2 # " p i ihh i . (x . hxi) hx j i . hxi)hx j i: 0 0 0 0 0 0 (1. Evaluate the commutator (c) Using the result obtained in (b). 212 (x .40) 0 0 0 0 0 1.

1. What is the corresponding eigenvalue? (a) We have x F (px)]classical @x @F (px) . 0 0 0 0 0 0 0 0 0 0 0 0 . . ai = C exp iph a jx i where C is a constant which due to normalization can be taken to be 1. 1 . (c) We have now b x x x x exp iph a jx i (=) exp iph a xjx i . 1 . a) exp iph a jx i: (1. @x @F (px) @x @px @px @x = @F (px) : @p x 1 (1.41) (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 (.a) n (ih) h n=1 n! h n=1 k=0 n 1 X 1 x = . . a exp iph a jx i x x = x exp iph a jx i .a (n .42) h n=1 1 1 . So we can write x (1. 1)! ia px = .a exp iph a : (1.43) x So exp iph a jx i is an eigenstate of the operator x with eigenvalue x .44) jx . a. . a exp iph a jx i x = (x . FUNDAMENTAL CONCEPTS 33 is an eigenstate of the coordinate operator x. . . 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. (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 . 0 0 0 0 . 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. . 0 0 .45) and that h jp i = (p ) and hp j i = (p ).34 1. 0 0 0 0 .46) where we have used (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.

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 = (. 1 . 1)! h n=0 n! h = exp ix : (1.ih)xn 1 = n! ih n(.ih) ih = n=1 (n . 1 1 .1.ih)xn 1 n=1 n! h n=1 k=1 n 1 X 1 X 1 ix n i = xn 1 (. FUNDAMENTAL CONCEPTS 35 (b) The operator exp ix gives translation in momentum space. exp ix p jp i ) h h h ix ix jp i . .47) h 1 1 1 .49) h where A is a constant which due to normalization can be taken to be 1.48) p exp h h Thus we have that exp ix jp i Ajp + i (1. 1 . 0 0 0 0 0 0 0 0 0 0 0 0 0 . . 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 . p exp ix jp i ) exp h = p exp h h ix jp i = (p + ) exp ix jp i : (1.

4) x dt ih y ih y z = ih x dSz = 1 S H ] = 1 S !S ] (S 1:4:20) 0 ) S = constant: (2.3) y y dt ih x ih x z = ih dSy = 1 S H ] = 1 S !S ] (S 1:4:20) ! (ihS ) = !S (2.3) and (2.!2S ) S (t) = C cos !t + D sin !t ) S (0) = C: y y y dt2 dt = But on the other hand dSx = .B: (2.2) The Heisenberg equation of motion gives dSx = 1 S H ] = 1 S !S ] (S 1:4:20) ! (.1) Indeed we have h i AH BH ] = U AS U U BS U = U AS BS U .4) we get d2Sx = . .!S ) y dt . U BS AS U = U AS BS ] U = U CS U = CH : (2.1 in Jackson. eB Sz = !Sz mc write the Heisenberg equations of motion for the time-dependent operators Sx(t). (2. 2 Quantum Dynamics . D! sin !t ) A=D C = .36 2.!2S ) S (t) = A cos !t + B sin !t ) S (0) = A : x x x 2 dt dt d2Sy = ! dSx (2:3) .1 Consider the spin-procession problem discussed in section 2. Let us rst prove the following AS BS ] = CS ) AH BH ] = CH : (2.5) z dt ih z ih z z = Di erentiating once more eqs.A! sin !t + B! cos !t = . Solve them to obtain Sx y z as functions of time.!S (2. .6) y y y y y y .C! cos !t . Using the Hamiltonian H = .! dSy (2=4) .ihS ) = . It can also be solved in the Heisenberg picture. Sy (t). and Sz (t).

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) (2.2.13) 2.3 Consider a particle in three dimensions whose Hamiltonian is given by 2 p H = 2~m + V (~ ): x .9) Sx(t) = Sx(0) cos !t . Evaluate x(t) x(0)] : The Hamiltonian for a free particle in one dimension is given by 2 H = 2pm : (2. im : (2.12) Thus nally t ht t x(t) x(0)] = m p(0) + x(0) x(0) = m p(0) x(0)] = .8) (2.7) (2. QUANTUM DYNAMICS So. Sy (0) sin !t Sy (t) = Sy (0) cos !t + Sx(0) sin !t Sz (t) = Sz (0): 2.2 Let x(t) be the coordinate operator for a free particle in one dimension in the Heisenberg picture.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.

ih)xi @xi x ij m i = ih p2 . n=0 n n k=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.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~ . ih~ rV (~ ): x ~ x (2.ih annxn 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. X X X n1 @ X = an (.ih @x anxn i i i i n n n k=0 @ x = . . . .38 By calculating ~ p H ] obtain x ~ * + d h~ ~i = p2 . 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 .16) i The right-hand side of (2.ih @x V (~ ): (2.15) 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.ih)xn 1 = .14) = xi 2m pi + xi pi V (~ )] : i ij The rst commutator in (2.14) now becomes X ih X @ V (~ ) x p ~ ~ H] = ij pj pi + (.

Does this probability change for t > 0? (a) We have ipa j t = 0i = exp . aj0i 0 0 = .4 (a) Write down the wave function (in coordinate space) for the state ipa exp . a 25 : 1=4x 1=2 exp 4. 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. 2.18) where in the last step we used the fact that the state kets in the Heisenberg picture are independent of time. Indeed we have d hnj~ pjni = 1 hnj ~ ~ H ] jni = 1 (E hnj~ pjni .h j0i: You may use 2 !3 x 25 hx j0i = 1=4x0 1=2 exp 4. ~ rV (~ ) ) :17) p ~ x x p x p dt ih + m x * d h~ pi = p2 . 1 0 2 x0 0 . h 2 !3 x . h~ rV i x ~ dt x ~ m 39 (2. 0 (2.2. QUANTUM DYNAMICS The Heisenberg equation of motion gives d ~ ~ = 1 ~ ~ H ] (2= ~2 . 1 x 2 0 0 0 .19) . . 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 j0i (Pr:=:4:c) hx . The left-hand side of the last equation vanishes for a stationary state.h j0i ) 1 hx j t = 0i = hx exp .

1 x 2 + x 2 + a2 .ipa j0i hexp h 2h !3 Z x 25 1=4x 1=2 = dx 1=4x0 1=2 exp 4. 0 0 0 0 0 0 So 0 0 0 jh0j For t > 0 t = 0ij2 ! a2 : = exp .ipa j0ij2: h It is . a2 : (2.5 Consider a function.22) jh0j tij2 = jh0jU (t)j t = 0ij2 = jh0j exp . 2ax = dx 0 2 " 2x0 !# 1 Z dx exp . . Evaluate the correlation function explicitly for the ground state of a one-dimensional simple harmonic oscillator. known as the correlation function. a 25 exp 4.24) where x(t) is the position operator in the Heisenberg picture. .40 (b) This probability is given by the expression (2.23) . iHt j t = 0ij2 h 2 = e iE t=hh0j t = 0i = jh0j t = 0ij2: (2. . 1 x 2 0 " # Z 1=2x 1 exp . 2x a + a2 + a2 = p x 2x2 2 4 4 0 0 ! ! 2 a p 1 p x = exp . 1 x 0 2 0 2 !3 x . 2 x 2 .20) jh0j t = 0ij2 = jhexp . . de ned by C (t) = hx(t)x(0)i (2.ipa j0i = Z dx h0jx ihx j exp . . 4x2 0 x 4x2 0 0 0 0 0 . 2x2 0 (2. 0 2.21) = exp . 0 0 .

27) = 2ih 2 ih Di erentiating once more the equations (2.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) we get d2x(t) = 1 dp(t) (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 .!2x(t) ) x(t) = A cos !t + B sin !t ) x(0) = A :27) 2 dt m dt d2p(t) = 1 dx(t) (2= .30) .28) (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 (2.26) and (2.29) (2.!x(0) sin !t + B! cos !t = m m p(0) B = m! D = . QUANTUM DYNAMICS 41 But on the other hand from (2.m!2x(t): (2.26) we have dx(t) = p(t) ) dt m p(0) cos !t + D sin !t ) .!2p(t) ) p(t) = C cos !t + D sin !t ) p(0) = C: :26) 2 dt m dt (2.2.

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. 2.42 Since we are interested in the ground state the expectation values appearing in the last relation will be h h (2. The state j i should be normalized. (b) Suppose the oscillator is in the state constructed in (a) at t = 0. (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 .32) = . a)(a + a )j0i hp(0)x(0)i = i 2 2m! (2.31) 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 . (c) Evaluate h( x)2i as a function of time using either picture. Do the following algebraically.i h : 2 2 Thus h h h C (t) = 2m! cos !t .i h h0jaa j0i = . without using wave functions.34) We can write the constands c0 and c1 in the following form c0 = jc0jei 0 q :34) i 1 (2= ei 1 1 . i 2m! sin !t = 2m! e i!t: (2. jc j2 : c1 = jc1je 0 (2. that is. jc0j2: (2.35) . (a) Construct a linear combination of j0i and j1i such that hxi is as large as possible.6 Consider a one-dimensional simple harmonic oscillator.

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): (2.38) 1 0 1 0 @1 But for hxi maximum we want also @ 2hxi (2. @ hxi = 0 ) q1 .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 . q jc0j2 c) 1 1 . sin( . Thus 1 j i = p2 (j0i + j1i): (2. 0 that make the average hxi as large as possible. jc j2 .37) 2 @ hxi = 0 ) . What we need is to nd the values of jc0j and 1 . 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. ) = 0 ) = + n n 2 Z : (2. )jc jq1 . jc j2 = 0 0 = 0 0 0 @ jc0j 1 . jc0j2 1 ) jc0j = p (2. jc j2 = 2 2m! (2.36) 0 1 0 0 q h where we have used that x = 2m! (a + a ). jc j2 .40) We can always take 0 = 0.

. . hxi2 (2. j t0 ti = U (t t0 = 0)j t0i = e .43) = 2 e i!t 2m! 2 2m! 2m! (ii) In the Heisenberg picture we have from (2. iHt=hj 1 i = p2 e .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. .44) . iE0 t=hj0i + 1 pe 2 .44 (b) We have j t0i = j i. . So 1 1 = p e i!t=2j0i + e i!3t=2j1i = p e i!t=2 j0i + e i!tj1i :(2. .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. iE1 t=hj1i (c) It is known that h( x)2i = hx2i . .45) . .

!3t=2)h1jaa i h 1 j1i + 2 h1ja aj1i 2m! (2.2.48) 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 . a a) sin2 !t m2s ! i 2 + m! hmh! (a + a )(a . 2m! cos2 !t = 2m! sin2 !t: (2. a) sin2 !t s 4m! i 2 + m! hmh! (a . 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 (2.47) h h :43) h h( x)2iS (2= 2m! .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 . aa . y y # h1 i(!t=2 2e . 2mh!2 (a2 + a 2 . h1 i!t=2 j0i + e i!3t=2j1i . i h h + 2 1 + 1 2m! = 2m! : 2 2 2 !t=2)h0jaa y j0i + 1 ei(!3t=2 2 .

2m! cos2 !t = 2m! sin2 !t: 2. a complex number.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 . a a) sin2 !t + 2m! (a 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) Prove that j i = e =2e a j0i is a normalized coherent state. a2) sin 2!t y y y y h ih . a2) sin 2!t y y y y (2. . in general. (b) Prove the minimum uncertainty relation for such a state.j j 2 y .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.50) (2. aa .51) So h h :44) h h( x)2iH (2= 2m! . 2m! (a2 + a 2 .46 h h h = 2m! (aa + a a) + 2m! a2 cos 2!t + 2m! a 2 cos 2!t ih + 2m! (a 2 .

n=0 f (n)jni: (a) We have aj i = e . 1)! n! ( a ) = e : n=0 n=1 y 1 y y 1 1 y . 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.j j 2 =2 h i a e a j0i y (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.) . 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. Find the most probable value of n.j j 2 e a j0i = e y . y y So from (2. (Hint: directly apply the time-evolution operator.52) since aj0i = 0.j j 2 h0je ae a j0i y . 1 1 y . If it is normalized. it should satisfy also h j i = 1.j j 2 =2ae a y j0i = e 1 .j j 2 =2 e a j0i = j i y (2.54) which means that j i is a coherent state. Indeed h j i = h0je ae .52) aj i = e . y y . (e) Show that the coherent state j i remains coherent under timeevolution and calculate the time-evolved state j (t)i. y . hence of E . and l is the displacement distance) to the ground state.53) = (n .2.

hxi)j i = 2m! (a .61) . a) we have s p s mh! h! pj i = i 2 (a .60) (2. p p .3) the state should satisfy the following relation xj i = c pj i (2. a)j i = i m2 (a .j j j j (b) According to problem (1.48 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 j . Since j i is a coherent state we have aj i = j i ) h ja = h j : y (2. h jpj i and c is a purely imaginary number.j j .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.j j 2 y y . )j i: q h! Similarly for the momentum p = i m2 (a . )j i y y y y s (2.55) = e .56) where x x .59) y y s and so h xj i = (x . h jxj i.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.

) (2.66) m! m!hnjmi = e n! m=0 2 n jf (n)j2 = (j nj! ) exp(.65) 1 1 i . 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 (2.63) (a .j j y 1 1 . )j i ) s (2.j j y . h jaj i) hpi = h jpj i = i 2 2 s h! = i m2 ( .i m2 pj i: h! y y y y s So using the last relation in (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.i) 2 pj i = xj i = 2m! mh! pj i (2. a)j i = i mh! (h ja j i .j j y 1 . QUANTUM DYNAMICS and 49 s mh! h j(a . (c) The coherent state can be expressed as a superposition of energy eigenstates X X j i = jnihnj i = f (n)jni: (2.j j . )j i = .60) s s h (. hpi)j i = i m2 (a . .2.62) and so s h! pj i = (p .j j2) (2.j j y .j j .64) | {z } purely imaginary and thus the minimum uncertainty condition is satis ed.

ipl=h) i j0i . .ipl=h) j0i is a coherent state with eigenvalue l m! .il h! pn 1 i m2 = n=1 n! h s !k=1 mh! . For the commutator in the last relation we have n h ( ipl=h)i X 1 .ilp !n 1 = i 2 h n=1 (n .j j ) ( 2 )n exp(.71) 2 So substituting (2. a] = i mh! : a p] = i 2 (2.70) = l 2h e( ipl=h) where we have used that s s mh! a a .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.ipl=h) j0i = a e( 1 . We have h a exp (. 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 .68) which means that the most probable value of n is j j2.69) we get r m! a exp (. y 1 . . (d) We should check if the state exp (.69) .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(.j j2 ) n! nihilation operator a. 1 . .ipl=h) j0i] (2. since aj0i = 0.il !n n X 1 . 1)! h r m! (2.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 . (2.j j . .il X 1 .70) in (2.72) which q means that the state exp (.ipl=h) j0i] = l 2h exp (.

i!t i (2. i!t n e .j j e i!t 2 =2 ( . h22m2 n2 n K (x y E ) = = Z Z0 1 Z0 dteiEt=hK (x . QUANTUM DYNAMICS = = = Thus 51 1 . . i!t=2j e 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 .j X 1 . X 1 n=0 e . We have X sin(nrx) sin(nry) r E . :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 .74) 2. dteiEt=hhxje iHt=hjyi 0 dteiEt=hhx tjy 0i n 0 . .8 The quntum mechanical propagator. . i!t i (2. . y t 0) . n=0 X e i!t=2 e n=0 1 e . (a) What is the potential? (b) Determine the constant A in terms of the parameters describing the system (such as m. ).j . j e :66) p jni (2= e n! . i!t e i!t=2j . 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. for a particle with mass m. r etc. 1 .73) aj (t)i = e i!t=2aj e i!ti = e = e i!t j (t)i: . iEn t=he .j j 2 1 =2 p n! n jni . i!t )n n! .2.

1 . 1 1 .78) (2. En ih =A . En +i")t=h # 1 0 (2. h22m2 n2 n n s 2 r2 (x) = iA sin(nrx) En = h m n2: (2.79) A = 2r ) A = i 2hr : ih d (x) = (x) dx 2.76) we get L = r ) L = r and ( 0 V = 1 for 0 < x < r otherwise while (2.75) X sin(nrx) sin(nry) ) r E .1 1 . En + i" ! " .ih ei(E n (x) n (y ) lim " 0 E . En Comparing with (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. (Hint: study the e ect on testfunctions.52 = = So X X n n ih : n (x) n (y ) E .9 Prove the relation where (x) is the (unit) step function. and (x) the Dirac delta function.) For an arbitrary test function f (x) we have Z + d (x) Z+ d Z+ dx f (x)dx = dx (x)f (x)] dx .1 (x (x) dfdx ) dx .77) n (x) = L L X n (x) n (y ) E .

The Lagrangian for the one dimensional harmonic oscillator is given by i m! h Scl = 2 sin(!T ) (x2 + 2x2 ) cos(!T ) . QUANTUM DYNAMICS = (x)f (x) ! 1 1 53 + 1 . f (x) = f (0) 0 Z ++ = (x)f (x)dx ) .2.83) which is the equation of motion for the system.1 d (x) = (x): dx Z + df (x) dx dx 0 + = x lim f (x) . 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 . This can be solved to give x(t) = A cos !t + B sin !t with boundary conditions x(t = 0) = x0 = A (2.80) 2.84) x(t = T ) = xT = x0 cos !T + B sin !T ) B sin !T = xT . d (mx) = 0 ) :81) @x dt @ x _ dt _ x + ! 2 x = 0: (2. 2 m!2x2: _ 1 _ 1 From the Lagrange equation we have @ L . x0!T !T : sin .1 (2.81) (2. x0 cos !T ) cos (2.85) B = xT . x0xT 0 T L(x x) = 2 mx2 .m!2x .82) (2. 1 1 .

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!T !T sin !t sin x0 cos !t sin !T + xT sin !t . t) ) sin ! x(t) = xT ! cos !t . 1 mxx . x(0)x(0)] x _ = m xT ! (xT cos !T . 2x x i : = 2 sin !T T 0 (2.54 So cos x(t) = x0 cos !t + xT . x x . x0!Tcos !(T .87) sin With these at hand we have ZT ZT 1 _ S = dtL(x x) = dt 2 mx2 . x0 cos !T ) 2 sin !T sin !T m! hx2 cos !T . x x + x2 cos !T i = 0 T 0 T 0 2 sin !T h T m! (x2 + x2) cos !T .11 The Lagrangian of the single harmonic oscillator is L = 1 mx2 . 1 m! 2 x2 = dt 2 dt _ 2 2 0 ZT T 1 _ = .88) 0 T 2. 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) = . t) : _ (2. x0! (xT . x0) . x0 cos !T sin !t = sin !T sin (2.86) = xT sin !t + x0!T !(T .

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

2x x ] hxbtbjxatai = 2 ih sin(!T ) a b 2h sin(!T ) b a where T = tb . . . . . .92) . ta. . . . . . .90) So !1 hxbtbjxatai = Nlim dx1dx2 : : : dxN hxb tbjxN tN ihxN tN jxN 1tN 1i : : : hxi+1ti+1jxitii : : : hx1t1jxatai: (2. it is legitimate to insert the identity operator written as Z dxjxtihxtj = 1 (2. . .91) .56 (f) Show that ( ) m! exp im! (x2 + x2) cos(!T ) . xi )2 i . . . 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. . . . .

yj ) . .2. . x )2 . 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. d @ L 1 my 2 . 57 (2.94) and this into (2. 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 . .92) we get . 2 j =0 .95) h (N +1) 2 9 8 N <i X m 1 "m!2y2 = : 2 exp : h j 2" (yj+1 . 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. 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) . 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 .

. . 7: . . . .101) .99) So we can diagonalize by an orthogonal matrix R.58 (b) For the argument of the exponential in the last relation we have N i X m (y . .1 5 6 0 0 0 ::: 1 0 7 4 4 5 0 0 0 : : : .. .1 : : : 0 0 7 6 0 1 0 ::: 0 0 7 6 6 7 6 0 .7 .7 6 6 7 6 0 0 0 : : : 2 . 1 "m!2y2 (y0= =0) j +1 j j h j=0 2" 2 N N i X m (y2 + y2 . Thus we have m (N2+1) Z dy : : : dy exp(..7 6 . . y y ) . Since is symmetric the following will hold = U DU ) T = U T D(U )T = U T D U = ) U = U : (2.. . . i"m! 2h i j=1 yi ij yj :(2. .7 .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 . . yi i+1 j yj ) . .98) . . = p N=2 i=1 i det (2. (2. . 7 2h 6 . 6. y y .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 . So = RT D R and det R = 1 (2.. .nT n) (2.1 2 . ..1 0 : : : 0 0 1 0 0 ::: 0 0 6 .7 .97) G = Nlim 2 ih" 1 N with 2 3 2 3 2 . . y )2 . . . yi i j+1yj .. 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 .96) i j =1 where the last step is written in such a form so that the matrix will be symmetric.1 2 0 0 0 ::: 0 1 !1 (c) We can diagonalize by a unitary matrix U . 2"ih (2yi ij yj .

. . . . "pj 1 2 (ta + j") . . . "2!2)"pj . 6 .1 2 0 det N pN : 0 0 0 59 0 1 0 . . 0 0 0 0 1 . "2!2 (ta + j") . . 0 0 39 ::: 0 0 > : : : 0 0 7> 7> > 7= : : : 0 0 7> 7 = . " .1 2 . . . 7 6 6 0 2!2 0 ::: 2 . . " . "2!2 : : : 7 0 0 0 7 6 7 6 0 7 6 . . . . . . (t .. " From the above it is obvious that 0 det j +1 pj+1 0 = (2 .2.1 7 >4 4 5 > : 0 0 0 : : : . (ta + (j . . (d) From (2.7 6 >6 60 >6 0 0 0 : : : 2 . . 7> : : : 1 0 7> 5> ::: 0 1 (2. 7: . "2!2 (t) .. ..1 : : : 0 0 0 7 6 . 1)") 2 (t) .1 2 . .. .103) with p0 = 1 and p1 = 2 . 6 . "): (2. (2. "2!2 . det j 1 ) = (2 . . pj 1 for j = 2 3 : : : N 0 0 . QUANTUM DYNAMICS where ai are the diagonal elements of the matrix D .1 0 : : : 0 0 3 >6 .7 6 . .1 7 2! 2 0 0 ::: 0 .1 2 .7 .. 7> . . So 2 2 .102) We de ne j j matrices j that consist of the rst j rows and j columns of N . 7> .. (e) We have (t) (ta + j") ) (ta + (j + 1)") = = ) (t + ") = "pj "pj+1 = (2 .1 : : : 0 0 0 3 6 . 6 . .1 : : : 0 0 7 60 >6 7 6 >6 7 6 >6 0 .98) we have ! 2ih" N det = m 82 21 > 2 . . "2!2)pj . 7 det j+1 = det 6 . "2!2.1 2 : : : 0 0 7 <6 7 .104) .1 0 7 7 6 6 0 2!2 5 4 0 : : : . "2!2) det j .1 2 . "2!2 6 0 6 det >6 . .

"2!2 (t) ) = .108) (2. 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) " . 1 ! 1: The general solution to (2.105) is (t) = A sin(!t + ) and from the boundary conditions (2. .107) (2.107) we have (ta) = 0 ) A sin(!ta + ) = 0 ) = .110) (2.!2 (t) ) 2 lim (t) . (t) = (t) . " (t .60 So (t + ") . 0 ! (2.!ta + n n 2 Z which gives that (t) = A sin !(t .111) t (t) = sin !(! .106) and (2. ") = .109) (2. (t . p 1 0 dt a " " = 2 . " (t) (t ") " .106) (ta) = "p0 ! 0 Thus d (t ) = (ta + ") . (ta) = "(p1 . p0) = p .!2 (t) ) d 2 = . while d = A! cos(t . ") . "2!2 . ta). t ) ) (t ) = A! (2) :107) a a dt 1 A! = 1 ) A = ! 0 and (2. .105) (2.

1 2" 1 1 0 # Z im x2 + im x2 .2. x0) 2 ih(t1 . im 2x x . t )(t . 1=2 (2. do not use completeness). t0 s 2 2 1 = 2 mh (t . x1)2 = dx1 2 ih(t . im 2x x dx1 exp 2h(t .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 ) exp 2h(imx2 t ) exp 2h(imx0 t ) i t2 . t ) 1 2h(t . 1 t2 .112) So from (a) hxbtbjxatai = exp iScl G(0 tb 0 ta) s h (2:88) = m! exp im! h(x2 + x2 ) cos !T . t ) 1 2h(t . QUANTUM DYNAMICS = (d) 61 .e.5. by explicitly performing the integral (i. t ) 1 0 2 1 2 1 2 1 2 1 . We have Z dx1Kf (x2 t2 x1 t1)Kf (x1 t1 x0 t0) " # Z s m im(x2 .16). = (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 . !1 . 2x x i : b a 2 ih sin(!T ) 2h sin !T b a 2. t ) 1 2 2h(t . t ) exp 2h(t . t ) 2 1 2 1 " s 2# m (x ) exp imh(t1 . t0) 2 1 .

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

e Ai pj ] c c c c! = ihe @Aj . eAj = . @Ai c @xi c @xj c @xi @xj ! ihe " B : = c ijk k (2. ecA and the relation " !# d2~ = d~ = e E + 1 d~ B .13 (a) Verify the relation i ! ihe " B j] = c ijk k ~ x p ~ where ~ m dt = ~ . e pi Aj ] . QUANTUM DYNAMICS 63 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 . eAi pj .2. 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 ) . e Aj j2: ~ ~ ~= j m mc (a) We have i j] = pi . ihe @Ai = ihe @Aj .

c hx j t0 ti + e (~ )hx j t0 ti x m " # 1 .ihr . ec + e j t0 ti m 2 3 2 3 ~ (~ ) 5 4 ~ eA(~ ) 5 ~x ~ = 21 4.h2r 2 + e ih r A + 2ih e A r + e2 A2 + e : (2.ihr .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 ~ ) x ~ x ~ eEi + 2c dt dt i " !# i ~ ~ ~ ~ x x ~ e E + 21c dt B .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 .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 . e @ ) 2m2c "ijk dt k ! ikj k dt !m @xi # e ~ B .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 . eAi 2m2ih c ijk k j c ijk j k ihm i c e (.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 . eAcx . B dt : (2. " B xj ." B + " B ) + e p ] 2m2c ikj k j ijk j k ihm i e m " xj B .

e Aj j2 = 0 ) ~ ~ ~ @t m mc @ +r ~ =0 ~ j (2.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 @ . r i + e ih r A j j2 + e ihA (r j j2) ~ ~ ~ ~ ~ ~ . 2hm r 2 . e ih r A j j2 . QUANTUM DYNAMICS " 1 .14 An electron moves in the presence of a uniform magnetic eld ~ in the z-direction (B = B z).2. mc Aj j2: and = j j2 j h 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2. r 2 e e ~ ~ ~ ~ ~ + mc ih r A j j2 + mc ihA ( r + r ) ! @ + @ = ih @t @t ) 2 ~ h r . ^ (a) Evaluate where x x y] px . 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 .h2 r 2 . h r h=( r )i + e r hAj j2i ) ~ ~ ~ ~ @t m" mc # @ j j2 + r h =( r ) . eAy : c . eAx c y py . 2hm r mc mc = ih @ j j2 ) @t @ j j2 = .ih @t = " 1 .117) @t e ~ ~ with ~ = m =( r ) .

y x c c 2c y 2c = .120) and H2 21 p2. By Ay = Bx Az = 0: (2. eBx 118) x y ] = px .122) 2 .121) H1 H2] = 4m2 x 2c y z 2c there exists a set of simultaneous eigenstates jk ni of the operators H1 and H2.118) 2 2 Thus we have eAx 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. Since m z " # 1 p + eBy 2 + p . eAy (2:= p + eBy p . 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. ~ ~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 . eB px x] + eB y py ] = iheB + iheB 2c 2c 2c 2c eB : = ih c (2.

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. Along the axis of the cylinder runs a thin.15 Consider a particle of mass m and charge q in an impenetrable cylinder with radius R and height a. impenetrable solenoid carrying a magnetic ux . h2 r2 ] = 2E i ) m h h2 @ 2 + 1 @ + 1 @ 2 + @ 2 .126) 2 mc j j 2. Calculate the ground state energy and wavefunction. From (a) we have eB ) xc (2. m @ 2 @ 2 @ 2 @z2 (~ ) = 2E (~ ) x x (2. ~ In the case where B = 0 the Schrodinger equation of motion in the cylindrical coordinates is .127) .123) m with x p] = ih.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 (2.125) = 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: (2.2. 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.

e ilz = C sin lz Z (a) = 0 ) C sin la = 0 ) la = n ) l = ln = n a n = 1 2 : : : If we write ( . So 1 d2Z = .129) 1 1 Z (z) dz2 dz2 with Z (0) = 0 ) A1 + B1 = 0 ) Z (z) = A1 eilz . l2 = 0 2 ( )d 2 R( ) d 2 R( ) d 2 2 1 2 ) R( ) d R + R( ) dR + ( ) d 2 + 2(k2 .m2 ) ( ) = e im : d (2. m + (k . . l ) = 0 R( ) d . l2) = 0 d2 d d 2 (2.128) 2 ( )d 2 R( ) d 2 R( ) d Z (z) dz2 with initial conditions ( a z) = (R z) = ( 0) = ( a) = 0.132) 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 ) d2Z + l2Z (z) = 0 ) Z (z) = A eilz + B e ilz (2.68 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 Z (z) = C sin lnz Now we will have (2.130) 1 d2R + 1 dR + 1 d2 + k2 .

l2 ) (2. eA(~ ) 5 (~ ) = E (~ ) ~ ~ x x 2m c c " !# h2 ^ @ + z @ + ^ 1 @ . l2 ) + 1 .133) In the case at hand in which a ! 0 we should take B3 = 0 since Nm ! 1 when ! 0.ihr . m2 R( ) = 0 ) d2 2 d # " d2R 1 dR m2 ) d(pk2 . l2) = 0 ) R k2 . l2 = R2 ) 2mE . ~ Now suppose that B = B z.138) p p . l2 ) + B3Nm( k2 . l2 d(pk2 .136) nm (~ ) = Ac Jm ( R with n = 1 2 : : : and m 2 Z .134) where m is the -th zero of the m-th order Bessel function Jm.2. l2 )2 + pk2 .ihr . QUANTUM DYNAMICS 69 " # d2R + 1 dR + (k2 . This means that the energy eigenstates are given by the equation 2 2 2 p m m = R k2 . l2 = m (2. l2) 2 R( ) = 0 p p ) R( ) = A3Jm( k2 . hc 2 ^@ (~ ) = E (~ ): x x (2. We can then write ^ ! ! B 2 ^= a ^: ~ (2. From the other boundary condition we get R(R) = 0 ) A3Jm(R k2 . l2) .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. 2m @ ^ @z @ !# 2 hc " ie ^@@ + z @z + ^ 1 @@ . (k2 .135) ) E = 2m R2 a while the corresponding eigenfunctions are given by m x )eim sin n az (2. l2 ) k2 . ie ) . n a = R2 m h2 " # h2 2 + n 2 m (2. eA(~ ) 5 4.

140) = @ 2 hc 2 @ @ 2 @ Following the same procedure we used before (i.e.4A2 . hc 2 2.143) nm (~ ) = Ac Jm ( R but now m is not an integer.m2 ) d2 . A2) 2iA 2im = = i(A m) ) l= 2 2 which means that ( ) = C2ei(A m) : (2. A) m 2 Z : (2. A2 : (2. A2)el = 0 ) l2 . 2iA d + (m2 .141) But ( +2 )= ( ))A m=m m 2Z ) m = (m .144) E = 2m R2 a D2 0 0 0 0 0 @ @ ie e . hc 2 we get " # " # @ + z @ + ^ 1 D (~ ) = E (~ ) h2 ^ @ + z @ + ^ 1 D ^@ ^ @z x x . A2 = .70 ie Making now the transformation D @@ . 2iA @ .142) This means that the energy eigenfunctions will be m x )eim sin n az (2. 2iAlel + (m2 . Leting A = hc 2 we get . ( z) = R( ) ( )Z (z)) we will get the same equations with the exception of " 2 # @ . A2) = 0: @ 2 @ d 2 d The solution to this equation is of the form el . So l2el . 4(m2 . 2ie @ . As a result the energy of the ground state will be # " h2 2 + n 2 m (2. 2m @ 2 @ x (2. A2 = @ 2 . 2iAl + (m2 .139) 2 @z2 x where D2 = ! ! @ 2 . A2) q 2iA . 2m @ ^ @z " # h2 @ 2 + 1 @ + 1 D2 + @ 2 (~ ) = E (~ ) ) . 2iA @ .

A is not zero in general but it corresponds to m 2 Z such that 0 m . From the discussion on WKB approximation we had that for E > V (x) A iZ q (x) = E . V (x)]dx ! c 1 Z x q2m E .16 A particle in one dimension (. V (x)]1=4 sin h x 2 3 4 s 3=2 2m c 4. = E .146) . x 5 = E . A < 1. V (x)]1=4 exp h 2m E .145) e 0 0 0 0 2. 2 E . h 2m 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)]dx I iZ q B + E .4 0 0 2 1 = qc1=4 sin 3 q3=2 + 4 ] (2. = E . Notice also that if we require the ground state to be unchanged in the presence of B . V (x)]1=4 sin h 4 p x Z x =E= ! 1=2 c 2m E . we obtain ux quantization e m . V (x)]dx . V (x)]1=4 exp . V (x)]1=4 sin 3 h . A = 0 ) hc 2 = m ) = 2 m hc m 2 Z : (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. QUANTUM DYNAMICS 0 0 0 71 where now m = m . x dx .2.

Ep : ) E (p) = c exp h 6m (2.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 )dx II (x) = x . p2 dp ) (p) h 2m ! 3 . E ]1=4 exp .148) 0 0 We also have (E . 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) = . 6pm + c1 " !# i p3 .147) .i E . Ep : 6m (2. h2m x =E= c3 exp . p2 (p) ) dp h 2m ! d (p) = . E )d(2m x) = x .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 .q)2=3 : = (2. ) ln (p) = h i Ep .72 h i m 1=3 where q = E . 2 (.i E . x and = 2h2 . E ) 0 So Z jcj2 dp exp hi (E .150) . E ) ) (2. E )p jcj22 h (E . On the other hand when E < V (x) ! c2 1 Z x q2m( x . E ]1=4 exp " . h x =E= # c2 1 Z x q2m( x .

q)3=2 + 4 q < 0 Ai(q) (2. leading terms in the asymptotic series are as follows 2 p 1q1=4 exp . i E .153) 3 2 hE i m 1=3 and q = where = 2h2 . iuq (2. uq du (2.2. In terms of the Airy functions ! 1 Z + cos u3 . x p : = (2. 3 q3=2 q > 0 Ai(q) 2 2 1 p (.156) (2.1 1 .151) h 6m h 2 h Using now the substitution 3 3 p (2. x u(h2m )1=3 ) (x) = 2 h " 3 3 #h Z+ = p du exp iu . x .1 1 For large jqj. uq du = 0. 1 . 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 . So ! ! Z+ u3 . QUANTUM DYNAMICS 73 These are the Hamiltonian eigenstates in momentum space.152) u = (h2m )1=3 ) h p6m = u 3 we have " # (h2mp 1=3 Z + du exp iu3 . i E .157) . uq = p Z + cos u3 .154) Ai(q) = p 3 0 we will have (x) = p Ai(. uq du du cos 3 (x) = p 3 0 2 R since + sin u33 .1 1 .155) .q)1=4 sin 3 (.1 1 1 .q): (2.

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(.x) . x 3=2 E= = 2p2m 2 E 3=2 ) = .159) x2 The turning points are x1 = .155) we get p q11=4 sin 2 q3=2 + 4 for E > V (x) (q) 3 2 p (. 2 E . 3 (.160) 4 2m p 0 .q1)1=4 exp .2 2m . So in this case the energy eigenstates heve to satisfy the consistency relation Z x1 q 1 dx 2m E . So Z E= q Z E= q 1 h = n+ 2 2m E . jxj] = n + 2 h n = 0 1 2 : : : (2. (b) When V = jxj we have bound states and therefore the energy spectrum is discrete.158) (q) 2 as expected from the WKB approximation. x]dx dx 2m E . E and x2 = E .q)3=2 for E < V (x) (2. jxj] = 2 0 E= 1=2 p Z E= E = .74 Using these approximations in (2.

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

.h 2j 2j jj j i . 2i .h 2j (J+ . J : (3. 0 The rotated state is given by up to terms of order "2. . 1i (3. 2h + (3. iJhy " jj j i ^ " # iJy " + (.4) h 2h2 y and from (3. J ) + "2 (J .76 j explicit form of the d(m)m function.i)2"2 J 2 jj j i = 1. . 1)jj j .h 2j jj j . . . J )2jj j i = . 1i . J )jj j i = . We can write Jy in terms of the ladder operators ) J+ = Jx + iJy ) J = J+ . J )2 jj j i jj j iR = 1 . J )jj j .h 2j (J+jj j .9) q (J+ . iJy 2i Subtitution of this in (3. "2 2j jj j i + "2 2qj (2j . m)(j + m + 1)jj m + 1i (3.4). . 2i: = 1.8) So q (J+ . jj j iR = R(" y)jj j i = d(j)(")jj j i = exp . 1i) q q q = . 4 2 4 . 1i (3. 2i jj j iR = jj j i + 2 8 8 ! "2 j jj j i + " q2j jj j . 1)jj j .J jj j i = . 1i + "2 qj (2j . gives " # " (J . (3. . m + 1)jj m . obtain an expression for the probability for the new rotated state to be found in the original state up to terms of order "2.5) y J = Jx .6) " q2j jj j . 1)jj j . 1i q = . . 1i .6) 8h2 + We know that for the ladder operators the following relations hold q J+jj mi = h (j . 2(2j . J jj j .7) q J jj mi = h (j + m)(j . .

10) 4 2 3. what is the l-value? If ~ not. (a) We have (~ ) h~ j i = (x + y + 3z)f (r): x x So (3.12) " !# 2 ~ 2j i (S 3:6:15) .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. " j jhj j jj j iRj = 1 . 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 .14) ) . sin ) = . 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. y= (3.11) (3. rf (r) (cos + sin sin sin2 @ 2 x sin2 @ .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 .3. Indicate how we may nd V (r). " j + O("4): (3.13) (3.

.78 and ! 1 @ sin @ (~ ) = rf (r) @ h.17) .14) and (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 . cos2 + sin2 ) . 1)Y1+1 : (3.15) in (3. . sin2 )(3. i)jl = 1 m = 1i and if we want it normalized we will have jN j2(18 + 2 + 2) = 1 ) N = p1 : (3. . Y1+1 + iY1+1 + iY1 1 x s hp i = 23 rf (r) 3 2Y10 + (1 + i)Y1 1 + (i .1i + (1 . x (b) Since we already know that l = 1 we can try to write (~ ) in terms of x m ( ). 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 = . 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 = .15) : sin Substitution of (3.h2rf (r) . sin (cos + sin )(1 .16) ~ which means that (~ ) is en eigenfunction of L2 with eigenvalue l = 1.18) 22 . .3 sin2 + (cos + sin ) sin cos i = x sin @ sin @ @ i rf (r) h. 8 r = < x = r q23 Y1 1 .11) gives 1 h~ jL2j i = . 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. . .6 sin cos + (cos + sin )(cos2 .

22) 3. 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 . 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 (3.21) (c) If E E (~ ) " x 2 .3.20) (3.4 Consider a particle with an intrinsic angular momentum (or spin) of one unit of h.19) (3.1) = jhl = 1 m == 1j ij2 = 22 = 11 : 79 (3. Quantum-mechanically. 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.h Y m d2 rf (r)] + 2 d rf (r)] . (a) Show explicitly that in nitesimal rotations of %i(~ ) are obtained x by acting with the operator " ~ ~ ~ (3. 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)] . i h (L + S ) " . 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 = .23) u~ = 1 .h2 " @ 2 (~ ) + 2 @ (~ ) . (One example of such a particle is the %meson).

Determine S ! ~ ~ (b) Show that L and S commute. h ~ L%(~ ) % x i" ~~ x (3. ~ ~ ) ~ x ~ x ~x " x (3. 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. 0 .80 ~ ~ i^ ~ where L = h r r. ~ ~x ~ p~ (d) Show that r %(~ ) = h1 (S ~)% where ~ is the momentum operp ator.27) = ~(~ . ~ (c) Show that S is a vector operator. ~ ~ ): %x " x " ~~ " x On the other hand %x " x ~(~ . (3. ~ ~ ) = ~(~ ) . 0 %i (~ ) x 0 = i=1 (1) Dil (R)%l(R 1~ ) ) % (x) = R~(R 1~ ): x ~ ~ % x .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 . ~ (~ r)%(~ ) % x " x ~~ x % x " x ~ ~ x = ~(~ ) . % (~ ) = R( )%(R 1~ ) = R( )%(~ . ~ ~ ) + ~ %(x .28) .25) (3. (~ ~ ) r%(~ ) = ~(~ ) .26) Under an in nitesimal rotation we will have R( So 0 n)~ = ~ + ~ = ~ + (^ ~) = ~ + ~ ~: ^r r r r n r r " r .

"y %1 ez (3. Using this in (3. "z %2 ex + "z %1 ."x C B %2 C = A@ A @ A @ . ij km + jm ki) X = h2 ( jm ki . h ~ L%(~ ) ~ x % x i" ~~ x " ~ x i" ~~ x = ~(~ ) .27) we get % (~ ) = ~(~ ) .ih) jlm .ih 0 i 6" B 0 0 . THEORY OF ANGULAR MOMENTUM ~ ~ x h~ x i where r%(~ ) r%i (~ ) jii.ih C + " B 0 0 0 C + " B ih 0 0 C7 B %2 C . h ~ (L + S ) %(~ ) ) U~ = 1 .ih) jkl(.ih) ikl(. Thus we will have that i" ~ ~ i" ~ ~ ~ x % (~ ) = U~~(~ ) = 1 . h ~ L%(~ ) + ~ %(~ ) .1 0 % 6"x B 0 0 . h2 jkl iml X = h2 ( ij km .h 4 x @ A y@ A z@ A5 @ % A 0 ih 0 . only on the j~ i basis and S x ~ (c) S is a vector operator since X Si Sj ]km = SiSj .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 .1 0 0 0 0 0 %3 2 0 1 0 1 0 13 0 1 1 0 0 0 0 0 ih 0 .32) .ih) ilm] i Xh 2 = h ikl jml . im jk ."z "y B %2 C = B "z 0 . "x%1 ey + "x%2 . im jk ) X X X = h2 ijl kml = ih ijl(.ih kml) = ih ijl(Sl)km : (3.31) ~ 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. h ~ (L + S ): (3. h ~ S ~(~ ) with (S )kl = . Sj Si]km = (."y "x 0 %3 %3 2 0 1 0 1 0 13 0 1 1 0 0 0 0 0 1 0 .3.30) ^ ^ ^ or in matrix form 10 1 1 0 1 1 0 % % 0 .29) " ~ ~ % = "y %3 .ih 0 0 0 0 0 %3 which means that i " ~% x " ~ ~ % = . (.ih kl. h ~ L%(~ ) + ~ ~(~ ): % x i" ~~ x " % x 0 81 (3.

m = 2.35) j j mi = hj1j2 m1m2jj1j2 jmijj1j2 m1m2i m=m1 +m2 So setting j = 2. p2 j0 +i : : : (3. m = 2 in (3.5 We are to add angular momenta j1 = 1 and j2 = 1 to form j = 2 1 and 0 states.34) where + and 0 stand for m1 2 = 1 0 respectively.33) 3. j2j : : : j1 + j2 = 0 1 2 states. Using the ladder operator method express all (nine) j m eigenkets in terms of jj1j2 m1m2i. .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. 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. This state is related to jj1m1 j2m2i through the following equation X (3. Jqjj = 2 m = 2i = (J1 + J2 )j + +i ) h (j + m)(j . j + +i (3. Let us take rst the state j = 2. . We want to add the angular momenta j1 = 1 and j2 = 1 to form j = jj1 .37) .35) we get jj = 2 m = 2i = hj1j2 + + jj1j2 jmij + +i norm. m + 1)j0+i + h (j2 + m2)(j2 . .36) = If we apply the J operator on this statet we will get . Write your answer as 1 1 jj = 1 m = 1i = p2 j+ 0i .

+i 3 . . .i + 1 p2j .i + 2 p2j . 0i + 1 p2j0.i + p j . .i + p 2j0.i + p6 j . +i + j + . .i + p 2j . 0i (3. . 0i ) 3 6 6 2 p2j0.i + p (J1 + J2 )j .35).i ) s 1 1 (3.i + p (J1 + J2 )j .2i = p 2 2 jj = 2 m = . . +i ) 6 s 6 i 1p hp p p 1p 6jj = 2 m = .1i = 2 2j . .i: (3.2i = j .1i = p2 2 J jj = 2 m = 0i = . . 1 1 J jj = 2 m = .1i = 6 6 6 6 1 1 j0. . .i ) 2 2 p 6jj = 2 m = 0i = 2j00i + 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 . .40) Now let us return to equation (3. s . . 0i ) 2 2 p 1 p2j . .38) jj = 2 m = 0i = 2 j00i + p6 j + . . 2 (J + J )j00i 2 3 1 1 1 + p (J1 + J2 )j + .i + p p2j . 0i + 2j0.39) jj = 2 m = . 0i ) jj = 2 m = .b : (3. . If j = 1. p p . . +i + 2j00i + p 2j00i + 2j + .42) . So hj = 2 m = 1jj = 1 m = 1i = 0 ) 12 a + 12 b = 0 ) a + b = 0 ) a = .1i = p (J1 + J2 )j0. m = 1 we will have jj = 1 m = 1i = aj + 0i + bj0+i (3.i ) 1 4jj = 2 m = .3.41) This state should be orthogonal to all jj mi states and in particular to jj = 2 m = 1i.

p2 c3 ) c2 = c3: (3. . +i: (3.i . 0i: 1 jj = 1 m = . p j .i . p (J1 + J2 )j . +i + p2j00ii ) 1 2jj = 1 m = 0i = p 2 2 1 1 j + .ii . p j .1i = p2 (3. 1 1 J jj = 1 m = 0i = p (J1 + J2 )j + .48) .43) jj = 1 m = 1i = p2 j + 0i . p hp2j . p2 j0+i: Using the same procedure we used before 1 1 J jj = 1 m = 1i = p (J1 + J2 )j + 0i .i .i . p 2j .1i = p 2j0.35) we see that the state jj = 0 m = 0i can be written as jj = 0 m = 0i = c1j00i + c2j + . So s 1 1 hj = 2 m = 0jj = 0 m = 0i = 0 ) 2 c1 + p6 c2 + p6 c3 3 ) 2c1 + c2 + c3 = 0 (3. . . 0i ) 2 2 1 j0.45) 2 Returning back to (3. +i ) 2 2 p p p 1 1 2jj = 1 m = . .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.47) 1 1 hj = 1 m = 0jj = 0 m = 0i = 0 ) p2 c2 . . +i (3.84 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 . . . p (J1 + J2 )j0+i ) 2 2 p 1 hp2j00i + p2j + .44) jj = 1 m = 0i = p2 2 . . That is 1 1 (3. . 12 .

12 j . we get 2c1 + 2c2 = 0 ) c1 + c2 = 0 ) c1 = . +i 3 = 12 j0. . Thus p p 1 3 and 1 1 1 jj = 0 m = 0i = p3 j + . Explicitly write T (1)0 in 1 terms of Ux y z and Vx y z .i + p3 j . 0i = 13 j + .1i jj = 0 m = 0i = j + +i 1 = q2 j0+i + 12 j + 0i = 2 j00i + 16 j + . 13 j00i: p p p p p p p p p p p p p p p (3.i + 16 j .47). +i .i + 13 j .1i jj = 2 m = .3.c2: The state jj = 0 m = 0i should be normalized so 85 (3. 0i = j .51) jj = 2 m = 2i jj = 2 m = 1i jj = 2 m = 0i jj = 2 m = .50) By convention we take c2 to be real and positive so c2 = c3 = c1 = .49) hj = 0 m = 0jj = 0 m = 0i = 1 ) jc1j2 + jc2j2 + jc3j2 = 1 ) 3jc2j2 = 1 1 ) jc2j = p3 : (3. +i = 12 j0. .2i jj = 1 m = 1i jj = 1 m = 0i jj = 1 m = .i + 12 j . p3 j00i: So gathering all the previous results together (3.6 (a) Construct a spherical tensor of rank 1 out of two di erent ~ ~ vectors U = (Ux Uy Uz ) and V = (Vx Vy Vz ). 12 j0+i = 12 j + . 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 .i . 12 j . (b) Construct a spherical tensor of rank 2 out of two di erent ~ ~ vectors U and V .i = 12 j + 0i . +i . 13 .52) 3.i .

iUy ) V 1 = 12 (Vx .58) U+1 = . .59) q1 q 2 . . iUy ) ) 1 U 1 = p (Ux . . p (Ux + iUy ) (3.53) From the components of a vector operator we can construct a spherical tensor of rank 1 in the following way. iJy Uz ] J U0] (3= (3:53) = .56) 2 :54) p 2hU 1 = J Uz ] = Jx . iUy ) (3. p It is known (S-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 : (3.86 ~ ~ (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 : (3.55) p :54) J+ U0] (3= 2hU+1 = J+ Uz ] = Jx + iJy Uz ] (3:53) = .10. p p . i(ih)Ux = h(Ux . 12 (Vx + IVy ) (3.57) 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 = .ihUy + i(ih)Ux = .ihUy . iVy ) .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. 2 x y U 1 = 12 (Ux .h(Ux + iUy ) ) 1 U+1 = . p .54) It is :53) :54) Jz Uz ] (3= 0hUz (3) Uz = U0 (3.

p U0V+1 2 2 1 (U + iU )V + 1 U (V + iV ) = . p 1 (U + iU )(V . 2 (Ux .2 x x y y (3. 1j11 20iU+1 V 1 . .60) T0(1) = h11 00j11 10iU0 V0 + h11 . 1j11 10iU+1 V 1 1 1 (3:52) = . . U V + iU V + iU V ) = . 87 (3. .64) T0(2) = h11 00j11 20iU0 V0 + h11 . iUy Vx .61) 2 T (1) = h11 . 1 (Uz Vx + UxVz + iUz Vy + iUy Vz ) 2 . iV ) 1 = p 2 x y x y y x y 2 22 x 1 = p UxVx + iUxVy . (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 . iUy Vx + Uy Vy . . . iUy )Vz + 2 Uz (Vx .2 x y z 2 z x y . . (3. 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 . 1j11 11iU0 V 1 1 1 1 (3:52) = . Uy Vy ] 2 2 i = p (UxVy .1 + 1j11 10iU 1 V+1 +h11 +1 .10j11 11iU 1 V0 + h11 0 . .62) . iU )(V + iV ) . . Uy Vx) (3.63) 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 .1 + 1j11 20iU 1 V+1 +h11 +1 . p U 1V0 + p U0V 1 2 2 1 1 = . UxVx + iUxVy . iVy ): (3.3. p U 1V+1 + p U+1V 1 2 2 1 1 (U .

for any j . 2 Uy Vy . 1 p 1 (U + iU )(V . i U V + i U V 6 z z 2 x x 2 x y 2 y x i i 1 1 . iU )(V + iV ) . 1j11 2 . . 2 Uy Vx . iUy )(Vx . You may. U V . (j m2jdm)m( )j2 = 1 j (j + 1) sin + m 2 + 1 (3 cos2 . . 2 UxVx + 2 UxVy .10j11 2 + 1iU 1 V0 1 1 p U0V 1 + p U 1 V0 2 2 1 (U V + U V . iU V ) x z z y y z 2 z x . . iV ) y x y y x y z z 2 x 6 6 22 x s3 1 2U V . . examine the rotational properties of Jz2 using the spherical (irreducible) tensor language. . . j) jd(mm ( )j2m 0 1 for any j (integer or half-integer) then check your answer for j = 2 . (b) Prove. . for instance.1 . h11 0 . iUy Vx): (3.67) 2 3. .7 (a) Evaluate j X m= j . 1): 2 2 0 0 Hint: This can be proved in many ways. (3.] . T (2) = 1 (3:52) = = .66) :52) 1 T (2) = h11 .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 . . 1 Uy Vy 2 s 1 (2U V . 1iU0 V 1 + h11 .65) 6 z z x x y y . 1j11 2 . j X m= j . 1 1 (U . 1 U V . iU V . iUxVy . Uy Vy . 2iU 1 V 1 (3= U 1V 1 = 2 (Ux . iVy ) 2 = 1 (UxVx . U V ) (3.

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

44) that ! cos 2 . m2hjm je 0 . m2jhjm je 0 . j m2jd(m)m( )j2 0 = = = = m= j j X . iJy =hjjmij2 m= j j X .90 For j = 1=2 we know from (S-3. while for m = . 2 cos2 2 + 2 sin2 2 0 = . sin 2 : (1=2) dmm ( ) = sin cos 2 2 So for m = 1=2 1=2 X (j) jdm1=2( )j2m = . iJy =hm2jjmihjmjeiJy =hjjm i (3. j 1 1 jd(m)1=2( )j2m = . 1 sin2 2 + 1 cos2 2 2 2 0 (3. 0 . 1 cos = m cos : 2 (3.2. 0 . iJy =hjjmi hjm je 0 . 0 .1=2 1=2 X 0 = 1 2 cos = m cos 0 (3.74) (b) We have j X m= j j X .73) m= 1=2 .72) 0 m= 1=2 .75) . 0 0 y 0 m= j hjm je 0 . m2hjm je 0 . 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: ^ ^ . iJy =hjjmi 0 m= j j X .

78) ^y z ^y i: 3 h2 h3 3 We know that for a spherical tensor (S-3.80) 0 0 . q= 2 0 Dq(2)( ey)hjm jTq(2)jjm i: ^ 0 0 0 0 0 (3.76) where T0(2) is the 0-component of a second rank tensor. So 1 Jz2 = 36 T0(2) + 3 J 2 (3. But we know from the Wigner-Eckart theorem that hjm jTq(2)0jjm i = 0.22b) k (k) D (R) = X D(k) (R)T (k) D(R)Tq (3.3. 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 . THEORY OF ANGULAR MOMENTUM From (3.65) we know that 91 T0(2) = s 1 (3J 2 .10. 1 j (j + 1) = 3 2 00 3 0 0 0 0 0 j) m2jd(mm ( )j2 0 . 0 p 0 0 0 y 0 y q= k 0 0 0 . J 2) 6 z (3. So = 0 6 j X m= j . 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 (3.79) qq q y y . 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 .

81) = 2 2 0 0 0 where we have used d(2)( ) = P2(cos ) = 1 (3 cos2 . y2) . x . (2) T+2 . y ) = 6 (3z .92 1 1 = 3 j (j + 1) + 2 (3 cos2 . T+1 : (3.(xz + izy) (3. y2)j j m = j i (where m = j j . (2) T (2) . 2 : : : )in terms of Q and appropriate ClebschGordan coe cients. Evaluate eh j m j(x2 . 1 j .82) q 1 2 2 2 q 1 2 2 T 1 = xz .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 . So from the above we have x2 . 2 6 3 2 1 j (j + 1) sin2 + m 2 1 (3 cos2 . Q = eh j m = j j(3z2 . 0 0 ~ (a) Using the relations (3.63-3.83) 1 xz = 2 . r2)j j m = j i . 1). y2 (b) We have (2) = T+2 + T (2) 2 . xz.8 (a) Write xy. izy (2) = T0 6 (2z . 1) m 2 . 00 2 3. y2) as components of a spherical (irreducible) tensor of rank 2. and (x2 . 1) (3. r2)j j m = j i is known as the quadrupole moment. r ) . 1) = . y2) + ixy T (2) = 2 (x2 . (b) The expectation value Q eh j m = j j(3z2 . . T (2) 2 xy = 2i . 1 j (j + 1) 3 1 j (j + 1) cos2 + 1 j (j + 1) + 1 j (j + 1) + m 2 (3 cos2 . ixy 2 2 (2) (2) T+1 = .

2i = p hj 2 j 0jj 2 j j i m j 2: (3.84) e h j m j(x2 . 2jj 2 jj . 0 .2jj 2 j j . 0 0 .3. . = 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:) (3.85) 6 0 0 . 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 . 2i h j k2j +k1 j i (3:84) Q hj 2 j . THEORY OF ANGULAR MOMENTUM (3:82) 93 .

h2 r (x) = E (x) ~ n 2m 2m n px px ) n (x) = Aei~ ~=h + Be i~ ~=h (4. p ~ ) x n (~ . 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 .2 Let (~ ) be the momentum-space wave function for state j i. . (b) In the case of a free particle the Schrodinger equation is p2 jni = E jni ) . 0 4. pi. (~ ) = h~ j i.94 4. Why does this not violate time-reversal invariance? (a) Suppose that jni in a nondegenerate energy eigenstate. p that is. (b) The wave function for a plane-wave state at t = 0 is given by px a complex function ei~ ~=h.1 (a) Assuming that the Hamiltonian is invariant under time reversal.Is the momentum-space wave function for the p p 0 0 0 .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 . 2Ent the wave function will be h real. prove that the wave function for a spinless nondegenerate system at any given instant of time can always be chosen to be real. 4 Symmetry in Quantum Mechanics 2E t t) = ei( hn + ) n(~ t): x (4. So we cannot apply the previous result.1) So if we choose at any instant of time = . .

~ h ~j ~ p p p . ~ i p p 0 0 0 (4.~ ). p i = d3p h.~ j i = (.3 in Sakurai to refresh your knowledge of the quantum mechanics of periodic potentials.4.3 Read section 4. SYMMETRY IN QUANTUM MECHANICS 0 0 0 95 0 time-reversed state j i given by (~ ).5) up to a phase factor. . X ikR n k (x) = n (x .~ ): p p p p 0 0 0 0 0 0 0 0 0 0 (4. (4.h ~j~j ~ i ) p p 1 j ~ i ) ~ 1 = . where a is the lattice constant. Prove that any Bloch function can be written as. (.6) 4. You know that the energybands in solids are described by the so called Bloch functions n k full lling. n labels the band. Ri )e i Ri ika n k (x + a) = e n k (x) .3) For the momentum it is natural to require h j~j i = . (~ ).~ j i j~ i p ~ p p ) h~ j j i = ~(~ ) = h.4) 0 So pp = p p ~j~ i (4:4) . =a =a]. or (.~ j~ i ) j~ i = j . 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. and the lattice momentum k is restricted to the Brillouin zone . So nally Z Z j i = d3p h~ j i j .~ )? p p p p Justify your answer.

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

x Using the time-reversal invariance prove ~ hLi = 0 for any energy eigenstate. =a . Ri ) n(x . So Z =a X 2 dke ikRj n k (x) = n (x .4.11) where in the last step we used that Ri . (This is known as quenching of orbital angular momemtum. Rj )] = i(R . . SYMMETRY IN QUANTUM MECHANICS But 97 eik(Ri Rj ) =a = 2 sin =a(Ri . 0 0 0 .4 Suppose a spinless particle is bound to a xed center by a potential V (~ ) so assymetrical that no energy level is degenerate. . 0 0 0 . Ri) = 2a =a e ikRi n k (x)dk (4. Rj = na. 4.13) (2 ) .12) So using the orthonormality properties of the Bloch functions Z dx m(x . . 0 0 0 . Ri ) ij a =a R Z =a i ) n(x . . 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 . k )dkdk (2 )2 e Z =a 2 = a 2 mn eik(Ri Rj )dk = 2a mn ij : (4. R ) Ri . with n 2 Z . . Rj i j =a = ij 2a (4. . .) If the wave function of such a nondegenerate eigenstate is expanded as XX l m Flm(r)Ylm( ) .

19) . .1) m m l l = ei Flm(r) m m l l 0 0 0 . Hence jni jni = ei jni: ~ (4. m i m (r) = (.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)(.4.15) If there is no degeneracy jni and jni can di er at most by a phase factor.1) 0 .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. 0 0 .hnjLjni (4= .17) Z = d3xh~ jni j~ i (4= ei jni ) x x :16) ~ ~ ~ hx j jni = hx jni = ei hx jni: (4.16) For the angular-momentum operator we have from (S-4.1)m Yl m ( )d = ei Ylm Flm(r)Ylm ( )d ml X ml m X ) Flm(r)(. m (r)(.53) ~ ~ hnjLjni = . 0 0 ) Fl ml 0 .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 .hnjLjni ) ~ ~ ~ :16) ~ hnjLjni = 0 : We have (4.98 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 .

(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 . . 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. .4.20) 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)(2 + m) n m+1 + 2 h (1 + m)(2 . Sy2): Solve this problem exactly to nd the normalized energy eigenstates and eigenvalues. . SYMMETRY IN QUANTUM MECHANICS 99 4.5 The Hamiltonian for a spin 1 system is given by 2 H = ASz2 + B (Sx . . .) 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 = .1 0 +1.21) Sx = 4 @ A @ 1 A 1 0 1 2 0 2 2 2 . For the operator Sz we have Sz jl mi = hmjl mi ) hlnjSz jl mi = hmhnjmi ) (Sz )nm = hm nm (4.

2 . ) .4 0 C = h2 B 0 @ A @ 4 2 0 .c B ) a = 0 A ) . h B 0 . 0 B h2 C det(H . 2 0 0 1 0 1 2 2 .c B2 + cA = 0 c=0 A (4.26) .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 = . ) + (B h2)2 = 0 ) (Ah . Bh h 0 2Ah 2 i ) (Ah2 . 2 p 2 C ) 0 A @ 0 .100 In the same manner for the operator Sy = S+2iS we 1 0 0 p 2 p 0 p h Sx = 2i B . B h2)(Ah2 . .2 1 . 0 A=0 @ 2 2. nd 1 0 . + B h2) = 0 ) 1 = 0 2 = h2(A + B ) 3 = h2(A . I ) = 0 ) det B 0 . (B h2)2 = 0 ) (Ah2 . .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 (4.2 0 2 2 Sx = . )2(.23) To nd the energy eigenvalues we have to solve the secular equation 0 2 1 Ah . B ): (4.24) 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 (4.1 2 C 1 0 A: 0 1 2 (4.

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 (4.30) 0 1 0 1 c norm: 1 B 0 C = p B 1 C) jnA+B i = @ A @ 0 A 2 . (4. B ) : aB + cA = c(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 (4.1i: .c ) a == 0 b So 1 8 > aA + cB = a(A . B ) a C)< bA > 0 = b(A .c 1 1 jnA B i = p2 j1 +1i . B ) B @ A@ A @ B 0 A c ( .1 .29) For = h2(A . B ) we have 0 10 1 0 A 0 B a B 0 0 0 C B b C = (A .1i: (4.28) 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 .31) .4. p2 j1 .

.38) .1)mjl . (4. Sy . 1 2 = A Sz2 1 + B ( Sx 1 . p j1 . .36) (4. Sy2) = H: . Sy2 1) = A Sz 1 Sz 1 + B ( Sx 1 Sx 2 = ASz2 + B (Sx .1i .34) (4. 1.37) (4:33) (4.1i 2 2 (4:33) (4.1i + p j1 +1i 2 2 = jnA B i: . .1i . .32) 1) To nd the transformation of the eigenstates under time reversal we use the relation (S-4.33) jn0i = :33) j10i (4= j10i = jn0i 1 1 jnA+B i = p j1 +1i + p j1 .mi: So (4.58) jl mi = (. .4.jnA+B i 1 1 = . 1 1 jnA B i = p2 j1 +1i . p2 j1 .102 Now we are going to check if the Hamiltonian is invariant under time reversal H . (4. p j1 . 1 Sy .35) 1 1 = . p j1 +1i 2 2 = . .

Compare with the perturbation results obtained in q (b). 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. 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.1) ax ax] = ay ay ] = 1: y y . ipy ): ay 2h m! From the fundamental commutation relations we can see that y y (5. (c) Solve the H0 + V problem exactly. De ne step operators: r m! ipx (x + m! ) ax 2h r m! ipx ax (x .1 Consider an isotropic harmonic oscillator in two dimensions. p p You may use hn jxjni = h=2m! ( n + 1 + n ):] 0 n n+1 0 n n 1 0 .5. m! ) 2h r m! ipy ay (y + m! ) 2h r m! (y . APPROXIMATION METHODS 103 5 Approximation Methods 5.

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. 1. 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 . 2 : : :) (5. energy eigenkets are also eigenkets of N : (5. 1 ) H0 = h!(N + 1): 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 (5.4) h j l0 i + j l1 i + : : : : i .e.104 De ning the number operators Nx ax ax y Ny ay ay y we nd H0 N Nx + Ny = h! . H0 ) j l0 i + j l1 i + : : : = (V .

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 2.5) and multiply it with another degenerate basis ket h m0 j E 0 . We need the matrix elements h 1 0 j V j 1 0 i. APPROXIMATION METHODS To 1'st order: (E 0 . H0) j l1 i = (V .5) h l0 j E 0 . 1h m0 j l0 i = h m0 j V j l0 i: 1 j l0 i ) (5. 1 j l0 i ) 1h l0 j l0 i = 1 = h l0 j V j l0 i (5. 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 . First excited state is degenerate j 1 0 i. j 0 1 i. h 0 1 j V j 0 1 i. Multiply with h l0 j to nd 1 ) j l0 i: 105 (5. Wind back to (5. 1 n i p ax j m n i = m + 1 j m + 1 n i etc: y p .6) In the degenerate case this does not work since we're not using the right basis kets. H0 j l1 i = 0 = h l0 j V . H0 j l1 i = 0 = h m0 j V .5. h 0 1 j V j 1 0 i.7) Now. 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. h 1 0 j V j 0 1 i.

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. However the only nonvanishing ones are: h! V11 20 = V20 11 = V11 02 = V02 11 = p (5.9) 3. . . j 1 0 i) E = h!(2 .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 . The V -matrix becomes . 1 0 1 0 = det B 1 . j 0 2 i. 1 C = . 2) @ A 0 1 . 1) + = (2 . so we need the corresponding 9 matrix elements. 2 ): (5. j 1 1 i. The second excited state is also degenerate j 2 0 i.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 . ( 2 .

)y 2]: m 2 p p So we get one oscillator with !x = ! 1 + and another with !y = ! 1 .5. y) 1 (p .12) 2 2 Note: x px] = y py ] = ih. y )2) + (x + y )2 . py ) are canonically conjugate. (x . p ): y p py p x y (5. px ) and (y . 2 j 1 1 i + j 0 2 i) . 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 . 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 . y )2) : (5. so that (x . p 1 j i+ = 2 ( j 2 0 i + 2 j 1 1 i + j 0 2 i) 1 j i0 = p2 (. (c) To solve the problem exactly we will make a variable change. j 2 0 i + j 0 2 i) p 1 j i = 2 ( j 2 0 i . APPROXIMATION METHODS which means that the eigenvalues are f0 as above we get the eigenvectors 107 h!g. ): . 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 . By the same method E+ = h!(3 + ) E0 = 3h! E = h!(3 . .

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

16) (b) Non degenerate perturbation theory to 2'nd order. Ej : 0 0 E1 2 2.15) 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 .17) . 1 2. E2 1 + 2 (jaj2 + jbj2)( E . E2 E1 . E2 2 + jaj2 + jbj2: E1 + E2 = 109 (5. Ek = 6 6 6 (5. E1 . E )2 + : : : = 2 2 1 2 2 + jbj2 = E1 + jaj .5. Ej : 2 2 E1 2 (5. E E1 2 2 + jb 2 E = E1 + 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. E2 k=2 E2 . 2 1. jaj . 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. APPROXIMATION METHODS s E1 . Ek 2 2 X jVk2 j jh 3 j V j 2 ij = jbj2 = 0 0 = E1 . 1 k=3 E3 . Ek 2 2 2 + jb 2 X jVk3 j2 = = E jaj E + E jbj E = . E2 (: : :) = E2 . 2 k=1 E1 .

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

) . 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 .E : . ab 0 = det a b jbj2 . = = (jaj2 . jaj2jbj2 = = 2 . 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 . )(jbj2 . E 0 E . 1 2 . 2 2 V23V31 = a b jV23j = jbj2 : M21 = E . E3 1. 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 . and take out a common factor E1 1 E2 ) ! jaj2 . 2 1 . E 0 E . E2). (jaj2 + jbj2) ) = 0 jaj2 + jbj2 2 2 (2) = 0 (2) = jaj + jbj ) 1 (5.21) 2 E . E 1 3 1 2 1 3 1 2 With this explicit expression for M .20) for j l1 i we get X h m0 j V j k0 ih k0 j V j l0 i = E 0 . E3 1. E M22 = E . 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. solve the eigenvalue equation (de ne = (2)(E1 .5.

eqn.112 From before we knew the non-degenerate energy shift.F0e t= 2m! ( m n m 1 + m + 1 n m+1): . For t 0 it is subjected to a time-dependent but spatially uniform force (not potential!) in the x-direction. m) h We need the matrix elements Vnm Vnm = h n j . 5. . (a) The problem is de ned by 2 2 H0 = 2pm + m!22x V (t) = . . (S. . We want to calculate X j t i = cn (t)e Ent=h j n i .F0e t= h n j x j m i = xe p h p = . @V ) @x At t = 0 the system is in its ground state j 0 i = j 0 i. t= . 0 En n (5.22) . 5.5. F0s t= j m i = . Em = !(n .15): X @ ih @t cn (t) = Vnm ei!nmtcm(t) m Vnm = h n j V j m i !nm = En . 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 . Show that the t ! 1 ( nite) limit of your expression is independent of time. = h!(n + 1 ) 2 where we get cn (t) from its di .3 A one-dimensional harmonic oscillator is in its ground state for t < 0. obtain the probability of nding the oscillator in its rst excited state for t > 0. (a) Using time-dependent perturbation theory to rst order. Everything is as we would have expected. and now we see that degenerate perturbation theory leads to the correct shifts for the other two levels. .F0xe t= (F = .

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

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

where (5. e4i!t .i . i t j 3 i + exp 3i t j 4 i = = p h h "2 1 = exp . e 4i!t) = 1 (1 . h Ht ( j 3 i + j 4 i) = 2 1 exp .6. + i) = 1 +exp h 2 h i 1 (e i!t + e3i!t) j + . + i = 2 . 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 .27) jii = j + . + j t ij2 = 1 (2 .17): c(0) = ni n Zt i!ni t (1) (t) = . . V given by (5. i + (e i!t + e3i!t) 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 . 4(!t) : : : > > : jh . (b) First order perturbation theory (use S. 5. + i)+ 2 # 3 i t p ( j + . j ti = 0 > > < 1 2 1 4i!t 4i!t 2 > jh + . i + j . j . cos4!t) ' 4(!t)2 : : : 4 2 ! . . .ih t p ( j + . APPROXIMATION METHODS the initial state is 1 j + .i cn h t0 dt e Vni (t ): Here we have (using the original basis) H0 = 0.5.23) 0 0 0 (5. j t ij = 4 (2 + e + e ) = 2 (1 + cos4!t) ' 1 .

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

. . . APPROXIMATION METHODS 119 If you have a normalization problem.5. !t) is turned on. . . .) To begin with the atom is in the n = 1 l = 0 state. but you should be able to show that the observable e ects are independent of L.30) L a0 y ! y .6. 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. the nal wave function may be taken to be 1 with L very large. At t = 0 the perturbation V = V0 cos(kz .5. .29) .44) gives us the transition rate 2 wi n = 2h Vni (En . (Ei + h!)) because the atom absorbs a photon h!. We want to nd the transition rate at which the electron is emitted with momentum ~f . 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. . y .

k~z 2 = (jkf jcos . k~z )2 k e 0 The transition rate is understood to be integrated over the density of states. 2kjkf jcos + jkf j2sin2 = 2 = kf + k2 .7. .39) 1 64 2 eikz ni = L3a5 h 1 i4 0 a2 + (~ f .31). k) ~z + jkf jsin (cos'~x + sin'~y )]2 = k e e e e = jkf j2cos2 + k2 . On the other hand we did not make any dipole approximation but performed the x-integral exactly.7) we note that since there is no polarization vector involved.120 So eikz ni is the 3D Fourier transform of the initial wave-function (and some constant) with ~ = ~ f . w has no dependence on the azimuthal angle . 2 64 2 L3 p2 1 wi ~f = 2h V40 L3a5 h 1 i4 (2 hf)3 d dpf : p 0 a2 + (~ f . Finally.31) In a comparison between this problem and the photoelectric e ect as discussed in (S. That can be extracted from (Sakurai problem q k e 5.5. k~z . We need to get that as a function of ~f = h~ f . As in (S. 2kjkf jcos : (5. The angular dependence is in the denominator: ~ f . 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. k~z )2 k e 0 n2dnd ! Note that the L's cancel. 5.

Sign up to vote on this title
UsefulNot useful