Quantum Physics

1.
BASIC P R I N C I P L E S A N D O N E - D I M E N S I O N A L MOTIONS
1001 Quantum phenomena are often negligible in the "macroscopic" world. Show this numerically for the following cases: (a) The amplitude of the zero-point oscillation for a pendulum of length I = 1 m and mass m = 1 kg. (b) The tunneling probability for a marble of mass m = 5 g moving at a speed of 10 cm/sec against a rigid obstacle of height H = 5 cm and width w = 1 cm. (c) The diffraction of a tennis ball of mass m = 0.1 kg moving at a speed v = 0.5 m/sec by a window of size 1 x 1.5 m 2 . (Wisconsin) Solution: (a) The theory of the harmonic oscillator gives the average kinetic energy as V = \E, i.e., \mu)2A2 = \hui, where w = y/gjl and A is the root-meansquare amplitude of the zero-point oscillation. Hence A = \l 0.41 x 1 0 - 1 7 m .
2mu) Thus the zero-point oscillation of a macroscopic pendulum is negligible. (b) If we regard the width and height of the rigid obstacle as the width and height of a gravity potential barrier, the tunneling probability is exp = where 2mw ^2gH - v2 0.9 x 10 30 . Hence - - j - \ / 2 m ( mgH -
e x p ( - ^ V ^ ^ )
Te--9xlo30
That is, the tunneling probability for the marble is essentially zero, (c) The de Broglie wavelength of the tennis ball is \ = h/p= h/mv = 1.3 x 10~ 30 cm,
1
Problems and Solutions
on
Electromagnetism
and the diffraction angles in the horizontal and the vertical directions are respectively 01 w X/D = 1.3 x 10" 3 2 rad, 02 A/L = 9 x 10~ 33 rad.
Thus there is no diffraction in any direction.
1002 Express each of the following quantities in terms of h, e, c, m =electron mass, M =proton mass. Also give a rough estimate of numerical size for each. (a) Bohr radius (cm). (b) Binding energy of hydrogen (eV). (c) Bohr magneton (choosing your own unit). (d) Compton wavelength of an electron (cm). (e) Classical electron radius (cm). (f) Electron rest energy (MeV). (g) Proton rest energy (MeV). (h) Fine structure constant. (i) Typical hydrogen fine-structure splitting (eV). (Berkeley) Solution: (a) a = h2/me2 = 5.29 x 1(T 9 cm. (b) E = me4/2h2 = 13.6 eV. (c) HB = eh/2mc = 9.27 x 1 0 - 2 1 erg G s - 1 . (d) A = 2Tvh/mc = 2.43 x 1(T 10 cm. (e) re = e2/mc2 = 2.82 x 10" 1 3 cm. (f) Ee=mc2 = 0.511 MeV. (g) Ep = Mc2 = 938 MeV. (h) a = e2/hc = 7.30 x 10~ 3 w 1/137. (i) AE = e 8 mc 2 /8/i 2 c 4 = a4mc2 = 1.8 x 10~ 4 eV.
1003 Derive, estimate, guess or remember numerical values for the following, to within one order of magnitude:
Basic Principles
and One-Dimensional
Motions
(a) The electron Compton wavelength. (b) The electron Thomson cross section. (c) The Bohr radius of hydrogen. (d) The ionization potential for atomic hydrogen. (e) The hyperfine splitting of the ground-state energy level in atomic hydrogen. (f) The magnetic dipole moment of 3 Li 7 (Z = 3) nucleus. (g) The proton-neutron mass difference, (h) The lifetime of free neutron. (i) The binding energy of a helium-4 nucleus, (j) The radius of the largest stable nucleus, (k) The lifetime of a TT meason. (1) The lifetime of a fx~ meason. (Berkeley) Solution: (a) Ae = h/mec = 2.43 x 10" 2 A. (b) a = ^ r 2 = 6.56 x 10~ 31 m 2 .
(c) = S* = -53 A "

AEf
(d) I = = 13.6 eV. (e) The splitting of the ground-state energy level is = 13.6 x (-L\ 10-4eV.
The hyperfine splitting of the ground-state energy level is AEhf w A E / / 1 0 3 w 1(T 7 eV.
(f) /x= 1.67 x 10- 2 6 J T-1. (g) Am = mp-m = - 2 . 3 x 1 0 - 3 0 kg. (h) rn w 15 min = 9 x 102 s. ( i ) = 4 x 7 MeV = 28 MeV. (j) The radius r corresponds to a region of space in which nuclear force is effective. Thus T 1.4 A$ = 1.4 x (100)5
=
6.5 fm.
(k) r = 8.28 x 10" 1 7 s. (1) The decay of fi~ is by weak interaction, and so r = 2.2 x 1 0 - 6 s.
Problems and Solutions on
Electromagnetism
1004 Explain what was learned about quantization of radiation or mechanical system from two of the following experiments: (a) (b) (c) (d) (e) Photoelectric effect. Black body radiation spectrum. Franck-Hertz experiment. Davisson-Germer experiment. Compton scattering.
Describe the experiments selected in detail, indicate which of the measured effects were non-classical and why, and explain how they can be understood as quantum phenomena. Give equations if appropriate. (Wisconsin) Solution: (a) P h o t o e l e c t r i c Effect This refers to the emission of electrons observed when one irradiates a metal under vacuum with ultraviolet light. It was found that the magnitude of the electric current thus produced is proportional to the intensity of the striking radiation provided that the frequency of the light is greater than a minimum value characteristic of the metal, while the speed of the electrons does not depend on the light intensity, but on its frequency. These results could not be explained by classical physics. Einstein in 1905 explained these results by assuming light, in its interaction with matter, consisted of corpuscles of energy hv, called photons. When a photon encounters an electron of the metal it is entirely absorbed, and the electon, after receiving the energy hv, spends an amount of work W equal to its binding energy in the metal, and leaves with a kinetic energy 1 , - mv = hv W. 2 This quantitative theory of photoelectricity has been completely verified by experiment, thus establishing the corpuscular nature of light. (b) Black B o d y Radiation A black body is one which absorbs all the radiation falling on it. The spectral distribution of the radiation emitted by a black body can be derived from the general laws of interaction between matter and radiation. The
Basic Principles
and One-Dimensional
Motions
expressions deduced from the classical theory are known as Wien's law and Rayleigh's law. The former is in good agreement with experiment in the short wavelength end of the spectrum only, while the latter is in good agreement with the long wavelength results but leads to divergency in total energy. Planck in 1900 succeeded in removing the difficulties encountered by classical physics in black body radiation by postulating that energy exchanges between matter and radiation do not take place in a continuous manner but by discrete and indivisible quantities, or quanta, of energy. He showed that by assuming that the quantum of energy was proportional to the frequency, e hv, he was able to obtain an expression for the spectrum which is in complete agreement with experiment: _ Snhjy3 1 A c e'er -I where h is a universal constant, now known as Planck's constant. Planck's hypothesis has been confirmed by a whole array of elementary processes and it directly reveals the existence of discontinuities of physical processes on the microscopic scale, namely quantum phenomena. (c) FranckHertz Experiment The experiment of Franck and Hertz consisted of bombarding atoms with monoenergetic electrons and measuring the kinetic energy of the scattered electrons, from which one deduced by subtraction the quantity of energy absorbed in the collisions by the atoms. Suppose Eo,E\,E2, . . . are the sequence of quantized energy levels of the atoms and T is the kinetic energy of the incident electrons. As long as T is below A = E\ EQ, the atoms cannot absorb the energy and all collisions are elastic. As soon as T > Ei EQ , inelastic collisions occur and some atoms go into their first excited states. Similarly, atoms can be excited into the second excited state as soon as T > E2 EQ, etc. This was exactly what was found experimentally. Thus the Franck-Hertz experiment established the quantization of atomic energy levels. (d) DavissonGermer Experiment L. de Broglie, seeking to establish the basis of a unified theory of matter and radiation, postulated that matter, as well as light, exhibited both wave and corpuscular aspects. The first diffraction experiments with matter
on
Electromagnetism
waves were performed with electrons by Davisson and Germer (1927). The incident beam was obtained by accelerating electrons through an electrical potential. Knowing the parameters of the crystal lattice it was possible to deduce an experimental value for the electron wavelength and the results were in perfect accord with the de Broglie relation A = h/p, where h is Planck's constant and p is the momentum of the electrons. Similar experiments were later performed by others with beams of helium atoms and hydrogen molecules, showing that the wavelike structure was not peculiar to electrons. (e) Compton Scattering Compton observed the scattering of X-rays by free (or weakly bound) electrons and found the wavelength of the scattered radiation exceeded that of the incident radiation. The difference AA varied as a function of the angle 6 between the incident and scattered directions: AA = 2 sin2 - , mc 2 where h is Planck's constant and m is the rest mass of the electron. Furthermore, AA is independent of the incident wavelength. The Compton effect cannot be explained by any classical wave theory of light and is therefore a confirmation of the photon theory of light.
1005 In the days before Quantum Mechanics, a big theoretical problem was to "stop" an atom from emitting light. Explain. After Quantum Mechanics, a big theoretical problem was to make atoms in excited states emit light. Explain. What does make excited atoms emit light? (Wisconsin) Solution: In the days before Quantum Mechanics, according to the Rutherford atomic model electrons move around the nucleus in elliptical orbits. Classical electrodynamics requires radiation to be emitted when a charged particle accelerates. Thus the atom must emit light. This means that the electrons would lose energy continuously and ultimately be captured by the nucleus. Whereas, in actual fact the electrons do not fall towards the nucleus and
Basic Principles
and One-Dimensional
Motions
atoms in ground state are stable and do not emit light. The problem then was to invent a mechanism which could prevent the atom from emitting light. All such attempts ended in failure. A basic principle of Quantum Mechanics is that, without external interaction, the Hamiltonian of an atom is time-independent. This means that an atom in an excited state (still a stationary state) would stay on and not emit light spontaneously. In reality, however, spontaneous transition of an excited atoms does occur and light is emitted. According to Quantum Electrodynamics, the interaction of the radiation field and the electrons in an atom, which form two quantum systems, contains a term of the single-photon creation operator a+, which does not vanish even if there is no photon initially. It is this term that makes atoms in excited states emit light, causing spontaneous transition.
1006 Consider an experiment in which a beam of electrons is directed at a plate containing two slits, labelled A and B. Beyond the plate is a screen equipped with an array of detectors which enables one to determine where the electrons hit the screen. For each of the following cases draw a rough graph of the relative number of incident electrons as a function of position along the screen and give a brief explanation. (a) Slit A open, slit B closed. (b) Slit B open, slit A closed. (c) Both slits open. (d) "Stern-Gerlach" apparatus attached to the slits in such a manner that only electrons with sz = h/2 can pass through A and only electrons with sz = h/2 can pass through B. (e) Only electrons with sz = h/2 can pass through A and only electrons with sz = h/2 can pass through B. What is the effect of making the beam intensity so low that only one electron is passing through the apparatus at any time? (Columbia) Solution: (a) The probability detected at the screen is that of the electrons passing
t h r o u g h slit A:
on
Blectromagnetism
a)
:+
B
>~>-
b)
c) B
Y/////X
d)
T7777X B
el B
Fig. 1.1
h =
IA(X)
(b) The probability detected at the screen is that of the electrons passing through slit B:
I2 = IB(X).
(c) Ic = Iu(x) = I\ + I2 + interference term ^ I\ + h(d) The eigenstate of the electrons passing through slit A is different from that of the electrons passing through slit B, and so there is no interference term. The intensity on the screen is just the sum of the intensities of the single-slit cases:
Basic Principles
and One-Dimensional
Motions
Id = h + h(e) Similar to (c), but the intensity is half that in (c): Ie = Ic/2 . Because of the self-interference of the wave functions of the electrons, the answers above remain valid even when the incident electron beam intensity is so low that only one electron passes through at a time.
1007 A particle of mass m is subjected to a force F(r) = VV(r) such that the wave function <p(p, t) satisfies the momentum-space Schrodinger equation (p 2 /2m - aV2p) <p(p, t) = id<p(p, t)/dt, where ft 1, a is some real constant and
vi = dydpi + d2/dPi + d*/dpi.

Find the force F (r). (Wisconsin) Solution: The coordinate and momentum representations of a wave function are related by
^(r,i)=(^) 2 yV(k,i)e* k - r rfk, V(k,t)=(j-y where k = f. Thus (withft,= 1) pV(P,0->-V^(r,t), V^(p,t)-*-rV(r,t), |^(r,i)e-ikrdr,
10
on
Electromagnetism
and the Schrodinger equation becomes, in coordinate space, A-V2 ,\ , . .d<p(r,t)
Hence the potential is V(r)=ar2, and the force is F ( r ) = - V V ( r ) = - - -^- V(r) = - 2 o r .
1008 Consider the one-dimensional time-independent Schrodinger equation for some arbitrary potential V(a;). Prove that if a solution ip(x) has the property that ip(x) 0 as x oo, then the solution must be nondegen> > erate and therefore real, apart from a possible overall phase factor. Hint: Show that the contrary assumption leads to a contradiction. {Berkeley) Solution: Suppose that there exists another function <f>(x) which satisfies the same Schrodinger equation with the same energy E as V and is such that lim^ôo <j>(x) = 0. Then rP"/iP = 070= and hence ip"(j> - 4>"i> = o, or tp'(j> (j/ip = constant. The boundary conditions at x oo then give -2m(E-V)/h2, -2m{E-V)/h2,
V>V - </>V = o ,
Basic Principles
and One-Dimensional
Motions
11
or 4, <p '
Integrating we have In ip =ln <p + constant, or ip = constant x <p. Therefore, ip and <p> represent the same state according to a statistical interpretation of wave function. That is, the solution is nondegenerate. When V(x) is a real function, ip* and ip satisfy the same equation with the same energy and the same boundary condition lim^-,.,*, ip* = 0. Hence ip* = cip, or ip = c*ip*, from which we have |c| 2 = 1, or c = exp (iS), where S is a real number. If we choose 5 = 0, then c = 1 and ip is a real function.
1009 Consider a one-dimensional bound particle, (a) Show that d_ dt /

oo
-oo
ip*(x, t) ip(x, t) dx = 0.
oo
(tp need not be a stationary state). (b) Show that, if the particle is in a stationary state at a given time, then it will always remain in a stationary state. (c) If at t = 0 the wave function is constant in the region a<x<a and zero elsewhere, express the complete wave function at a subsequent time in terms of the eigenstates of the system. (Wisconsin) Solution: (a) Consider the Schrodinger equation and its complex conjugate in dip/dt =- -ihdip*/dt h2 V2V> + Vip, + Vip*. (1) (2)
2m h2 = -V2ip* 2m
Taking ip* x (1) - ip x (2) we obtain
ih ^ (r ip) = - | - V (V*W - < W )
12
Electromagnetism
For the one dimension case we have, integrating over all space,
d_ (
dx \
d_^djT
dx dx
dx
If ijj is a bound state, then t/j (x 00) = 0 and hence d f / ip*{x, t)-ip(x, t)dx = 0. at J_ 0 0 (b) Supposing the particle is in a stationary state with energy E at t = to, w e have
iJVO*;. to) = V(a;, to),

where H does not depend on t explicitly. At any later time t, the Schrodinger equation ihdipix, t)/dt = Hip{x, t) applies. As H does not depend on t explicitly, the Schrodinger equation has the formal solution tf)(x, t) = exp[iH(t to)/h]ip(x, to)
Multiplying both sides by H from the left and noting the commutability between H and exp [~i (t to )H/h], we find Hip(x, t) = exp ~iH(t-t0) h -iH(t-t0 Hij){x, t0)
= E exp Etp(x,
h
t).
ip(x, t0)
Hence ip(x, t) represents a stationary state at any later time t. (c) The wave function given for t = 0 can be written as ip{x,0) = C, 0, | x | < o, otherwise,
Basic Principles
and One-Dimensional
Motions
13
where C is a constant.
Normalization J_a ip*ipdx 1 requires that =
c = (&)*
Suppose the eigenfunction of the bound state is (x \ n) and H\n) En\n). Then l = |n)(n|, and | ^ ( x , 0)) = | n ) < n | V ( z , 0)),
n
\i;(x,t))=J2\n)(n\Tp(:<:>Q))
n
ex
P (
^
_ i
X*)
'
Hence V>(z, <) = 5 3 o-nipnix) exp ( - i - r 1 * ] , with

OO
an = (n\ip
oo
1010 0 (a;, t) is a solution of the Schrodinger equation for a free particle of mass m in one dimension, and ip(x, 0) = A exp (-x2/a2).
(a) At time t 0 find the probability amplitude in momentum space. (b) Find V(z, *) (Berkeley)
14
Electromagnetism
Solution: (a) At time t = 0 the probability amplitude in momentum space is 1 i> (p, 0) = = r /
oo V2TTH
e~
ipx/h
V (x, 0) dx
/ exp
exp (x2/a2 ipx/h) dx (-a2p2/4h2).
-oo
Aa
V2h
(b) The Schrodinger equation in momentum space for a free particle,
ihdip(P, t)/dt = Hip{P, t) = j ^ V ( P , t), gives ip(p, t) = B exp At time t = 0, we have B = ip(p, 0). Hence a2p2 Ah2 ip2t 2mh -ipH 2mh
W > *) = ~^z 2ft
ex
Hx t) =
f exp
ipx
'
j 2 ^ JAa
2
ip(p, t)dp
a +
2iht m
exp az +
2iht m
We can also expand the wave function as a linear superposition of plane waves and get
Basic Principles and One-Dimensional 1 pea

1
Motions
15
lp(x, t)
(2TTH)
dp
~(27rfi)V2
/ ^
n
exp P 2 V dp . 2mfi.
2 2 " a"p Ah2
aV
4/i 2
x exp 1 I - x \i Vfi Aa 2H^ Aa a +

2
exp
ip 2 i 2mh
ipi h
dp
2i7ii
exp
a2 + m J
2iht\
which agrees with the previous result.
1011 A particle of mass m is confined to a one-dimensional region 0 < x < a as shown in Fig. 1.2. At t = 0 its normalized wave function is ip (x, t 0) = y/8/5a (a) What (b) What (c) What the box (i.e., Solution: The time-independent Schrodinger equation for 0 < x < a is 1 + cos ( J sin {irx/a).
is the wave function at a later time t = to? is the average energy of the system at t = 0 and at t = to? is the probability that the particle is found in the left half of in the region 0 < x < a/2) at t = to? {MIT)
2 ^
^ = -
It has solution tp (x) = A sin kx, where A is given by fc2 = 2j^-, satis; fying V(0) = 0. The boundary condition ip(a) = 0 then requires ka = nn. Hence the normalized eigenfunctions are
16

QD
on
00
Electromagnetism
Vix)
->
0
Fig. 1.2
ITITTX \
Wn = \ ~ sin I ) V a V a J and the energy eigenvalues are n2Tr2h2 En = n = l , 2, 3, . . . . r , 2maz Any wave function tl>(x, t) can be expanded in ipn: ip(x, t) = ^ with An(t)=An(0)exV(As , / \ / 8 /, irx\ . irx w (x, 0) = \ 1 + cos I sin V 5o V a ) a 8 . TTX / 2 27ra: ' sin 1 \/ sin 5a a V 5a a we have Ai(0) = -^=, A 2 (0) = -^=, An(Q) = 0 for n ^ 1,2. iEnt An{t)tpn{x, 0)
Basic Principles
and One-Dimensional
Motions
17
(a) T h u s iir2hto\ 2ma2 i2ir2hto sin -i-K2hto 5o + exp exp 2ma2 irx cos
TTX
tp(x,
t0)
exp
. irx sin
2TTX
+ \l exp 5a
-i2-K2hto\ ma*
(b) T h e average energy of the system is
= An(o)2
= 3 S i + - Ei o o
_ 47T2ft2
5ma2 (c) T h e probability of finding t h e particle in 0 < x < | at t = to is

ra/2
P(p<x<^)
=J
5a J 0
\iP(x,tQ)\2dx
Va / 1 + cos 2 a
7r:r + 2 cos cos a 1 J^_ 2 + 15^ C S l
/37r ftto 5V 2ma^ (3ir2ht0 ~W~
dx
18
Electromagnetism
1012 A particle of mass m moves in a one-dimensional box of length / with the potential V = oo, x < 0, V = 0, 0<x < /, V oo, x > I. At a certain instant, say t = 0, the wave function of this particle is known to have the form tp = A / 3 0 / / 5 x{l-x), 0<x<l, 0 = 0, otherwise. Write down an expression for tj){x,t > 0) as a series, and expressions for the coefficients in the series. (Wisconsin) Solution: The eigenfunctions and the corresponding energy eigenvalues are
,2
iPn(x) = f1 Thus
sin(^n),
n =
A-(^
n = 1, 2, 3,
IV>> = \n)(n | V)
where /"' /2~
(-KX
(n | V(* = 0 =
\1
sm
\ T ) ' y ( ~~x'
/30
.,
= 4 v l 5 I 1 (1 cos Tin) \mr J = 4\/l5[l-(-l)"](l/n7r)3, and hence ip(x, t) = ^2(n\

n=l
/ E tp(t = 0))ipn(x) exp f - i -^^
\ t)
'
Basic Principles
and One-Dimensional
Motions
19
1013 A rigid body with moment of inertia of Iz rotates freely in the x-y plane. Let <f> be the angle between the ar-axis and the rotator axis. (a) Find the energy eigenvalues and the corresponding eigenfunctions. (b) At time t = 0 the rotator is described by a wave packet ip(0) = A sin 2 <j>. Find tp(t) for t > 0. (Wisconsin) Solution: (a) The Hamiltonian of a plane rotator is H=-(h2/2Iz)d2/d<t>2 and so the Schrodinger equation is -(H2/2IZ) Setting a2 = 2IzE/h2, d2ip/d<t>2 = .Ety-
we write the solution as i> = A e i a * + B e ~ i a 0 ,
where A, B are arbitrary constants. For the wave function to be singlevalued, i.e. tp(<f>) = xp((f> + 2ir), we require a = m = 0, The eigenvalues of energy are then Em = m2h2/2Iz, m = 0, 1, ... . 1 , 2, . . . .
and the corresponding eigenfunctions are
^ m (0) = _ L e i m * ,
yj
lit
m = 0,l, . . . ,
after normalization | 0 * \pmipmd(j) = 1. (b) At t = 0 tp(0) =A sin 2 <f> = (1 - cos 20) = A/2-(J2* + e^2*),
which corresponds to m = 0 and m = 2. The angular speed is given by Em = \ hi<2, or 4 = -y^- Hence we have for time
20 ,0(t)

=
on
Electromagnetism
+ e-i2(0+Rt//z)i
d _ d [e*2(*-fit//)
1014 An electron is confined in the ground state in a one-dimensional box of width 10- 1 0 m. Its energy is 38 eV. Calculate: (a) The energy of the electron in its first excited state. (b) The average force on the walls of the box when the electron is in the ground state. (Wisconsin) Solution: (a) An electron confined to a one-dimensional box can have energy levels (Problem 1011) En = h2TT2n2/2ma2, n = l, 2, 3, . . . .
Thus for the first excited state (n = 2), the energy is E% = AE\ = 152 eV. (b) The average force on the walls of the box is F = -(dH/da).
Differentiating the equation of a stationary state (H En )ipn = 0, we have (dH dEn\ C^n
| ^ - ^ r j ^ + (H-En)
and hence
=0,
Integrating the left-hand side of the above, we have
Jrn{H-En)^dx
=J
^(H-En)*rndx,
which is zero since H is real. Integrating the right-hand side of the equation then gives (dH/da) = dEn/da.
Basic Principles
and One-Dimensional
Motions
21
Hence F = For the ground state, n = 1 and F = 2Ei/a = 7.6 x 109 eV/cra = 1.22 x 1(T 2 dyne. -dEn/da.
1015 Give the energy levels EM) of the one-dimensional potential in n ' Fig. 1.3(a) as well as the energy levels En of the potential in Fig. 1.3(b) (Wisconsin)
2a (a) Fig. 1.3 (b)
T i
Solution: (a) Use coordinate system as shown in Fig. 1.4. The Schrodinger equation is
V \ a
-a
0
K .
Fig. 1.4
22
on
Electromagnetism
h2 d2ip 2mdx*+VÊ^
where V = 0 for x > a (region I ) , V = VQ for a < x < a (region II), V = 0 for x < a (region III).
For bound states we require Vo < E < 0. Let
_ 2m( + Vb)
'
,2 _
2mE
~ v
The Schrodinger equation becomes d2ib 2 dx2 + k ip = 0 and -^ - k'2tp = 0 for region I and III, dx1 which have solutions rp = A sin kx + B cos kx for a < x < a , ip = Ce~h x + Dek x for x < a and x > a . The requirement that i / ) - > 0 a s i - > oo demands that ip = Ce~k ip = De
k x x
for region I I ,
for x > a (region I ) , for x < a (region III).
The boundary conditions that ip and ip' be continuous at x = a then give A sin ka + B cos ka Ce~k , A sin ka + B cos ka = De~k
a
,
a
Ak cos ka - Bk sin ka = -Cfc' e~k

ka
Afc cos ka + Bk sin fca = Dk' e~ ' ; or
Basic Principles
and One-Dimensional
Motions
23
2A sin ka = (C - D) e"-k'a 2 S cos ka = (C + D) e -k'a 2Ak cos ka = -(C-D) k' e~k'a ,
2Bk sin ka = (C + D)k' e~k'a .
For solutions for which not all A, B, C, D vanish, we must have either A = 0, C D giving ktanka = k', or B = 0, C = -D giving k cot ka = k'. Thus two classes of solutions are possible, giving rise to bound states. Let = ka, r\ = k'a. Class 1: j tan = 77, U where 7 2 = fcV + k'2a2 2
+ r?2=72,
^ <
Since and r\ are restricted to positive values, the energy levels are found from the intersections in the first quadrant of the circle of radius 7 with the curve of tan plotted against , as shown in Fig. 1.5. The number of discrete levels depends on VQ and a, which determine 7. For small 7 only one solution is possible. T|
Fig. 1.5
24
on
Electromagnetism
Class 2: cot :
2 _,_^2
-v,
-T
2
+V
VQO?
A similar construction is shown in Fig. 1.6. Here the smallest value of gives no solution while the larger two give one solution each.
Fig. 1.6
Note that = 0,77 = 0 is a solution of tan = 7 and so no matter how 7 small 7 is, there is always a class 1 solution, whereas 7 has to be above a minimum for a class 2 solution to exist, given by cot = 0 which has a 2 2 H minimum solution = ^, i.e. 7 = | or VQOL2 - n8m " (b) Use coordinates as shown in Fig. 1.7.
1L
v.. v 0
E
III
II
Fig. 1.7
Basic Principles
and One-Dimensional
Motions
25
The Schrodinger equation has solutions ip = A sin kx + B cos kx ip = Ce~k x ip = 0 for 0 < x < a , for x > a, for x < 0,
satisfying the requirement ip > 0 as x oo. The boundary conditions at > x = 0 and a; = a then give B = 0, A sin ka = Ce~k'a , Ak cos fca = -Ck'e~ka and finally f cot = -77, 2 + rf. = 7 2 > as for the class 2 solutions above. ;
1016 Consider the one-dimensional problem of a particle of mass TO in a potential (Fig. 1.8)
V = 00, x < 0,
V = 0, 0<z<a, V = V0, x>a. (a) Show that the bound state energies (E < Vb) are given by the equation V2mEa tan h V V0 - E ' (b) Without solving any further, sketch the ground state wave function. (Buffalo) Solution: (a) The Schrodinger equations for the two regions are ip" + 2mEip/K2 = 0, 0 < x < a, 2 xp" - 2m (Vb - E)xp/h = 0, x > a ,
26
Electromagnetism
with respective boundary conditions ip = 0 for x = 0 and ip 0 for > a; +oo. The solutions for " < Vb are then V = sin(y/2rnEx/h), < lP = Ae-V*(Vo-E)x/Ht 0 < x < a;
x > a t
where A is a constant. The requirement that ip and -^ are continuous at x = a gives tan (V2mEa/h) = -[E/(V0 - E)}1'2 . (b) The ground-state wave function is as shown in Fig. 1.8. V(x)
Ki
a
Fig. 1.8
1017 The dynamics of a particle moving one-dimensionally in a potential V{x) is governed by the Hamiltonian H0 = p2/2m + V(x), where p = ih d/dx is the momentum operator. Let En, n = 1, 2, 3, . . . , be the eigenvalues of Ho. Now consider a new Hamiltonian H = Ho + Xp/m, where A is a given parameter. Given A, m and E , find the eigenvalues of H. (Princeton) Solution: The new Hamiltonian is H = H0 + Xp/m = p2/2m + Xp/m + V{x) = (p + A) 2 /2m + V(x) X2/2m,
Basic Principles
and One-Dimensional
Motions
27
or
where H' = H + ^ , p' = p + A. The eigenfunctions and eigenvalues of H' are respectively En and ^ . As the wave number is k' = ^ = ^ {p + A), the new eigenfunctions are
V = V(0) e-^/' 1 , .
and the corresponding eigenvalues are
En=EnV-\2/2m.
1018
Consider the one-dimensional wave function 1>(x) = A{x/x0)n e~x'xa,
where A, n and Xo are constants. (a) Using Schrodinger's equation, find the potential V(x) and energy E for which this wave function is an eigenfunction. (Assume that as x oo, > V(x) -> 0). (b) What connection do you see between this potential and the effective radial potential for a hydrogenic state of orbital angular momentum /? (Wisconsin) Solution: (a) Differentiating the given wave function, ax 2 V(x =A ax2
d2 ii \
xQ
\XQJ
\XOJ
XQJ
A n(n-l)
( x\n~
xi
\xoJ
_x/xn x Xa
'
, n-l
( J
\XQJ
e~x/x + A ()
X\ \X0J
e~x/xo
n{n - 1)
_n
tp{x),
28
on
Electromagnetism
and substituting it in the time-independent Schrodinger equation h2 2 2 d /dx +V{x)\ 2m we have i>(x) = Etp(x): 2n 1 + ~2
E-V(x)
Asy(i)
-5L
h2 2m
n(n 1)
2m 0 when x oo, we have E = h2/2mxQ and hence > V(aO fn(n-l) 2n/xox
(b) The effective radial potential for a hydrogen atom is e2/r 1(1 + 1) H /2mr2. Comparing this with V(x) we see that the 1/r2 term is formally identical with the 1/x2 term with the angular momentum I taking the place of n. The ^ term of V(x) depends on n I, while the (Coulomb) term in the effective potential for the hydrogen atom is independent of the orbital angular momentum /. This is the difference between the two potentials.
2
1019 Consider the following one-dimensional potential wells: V[x\
-*- X
Fig. 1.9
V(x)
M
~L
Fig. 1.10
Basic Principles
and One-Dimensional
Motions
29
(a) Can each well support a bound state for an arbitrarily small depth Vi (i = 1, 2)? Explain qualitatively. (b) For Vi = V2, what is the relationship between the energies of the bound states of the two wells? (c) For continuum states of a given energy, how many independent solutions can each well have? (d) Explain qualitatively how it is possible to have bound states for which the particle is more likely to be outside the well than inside. (Wisconsin) Solution: (a) For bound states, we must have V < E < 0. Let
2
_ 2m{E + V) h2 '
_ 2mE 2 ft '
where V = V\, V2 for the two cases, and set = ka, r] = k'a, 7 = *2 a. The discussion in Problem 1015 shows that for the potential in Fig. 1.9, the solutions are given by cot = -77, 2 + rf = 7 2 .
The energy levels are given by the intersection of the curve cot = 7 7 with a circle of radius 7 with center at the origin (Fig. 1.6) in the first quadrant. As the figure shows, 7 must be greater than the value of for which cot = 0, i.e. > f. Hence for a bound state to exist, we require For the potential shown in Fig. 1.10, two classes of solutions are possible. One class are the same as those for the case of Fig. 1.6 and are not possible for arbitrarily small V2. The other class of solutions are given by
As the curve of tan = 7 starts from the origin, 7 may be arbitrarily 7 small and yet an intersection with the curve exists. However small V2 is, there is always a bound state. (b) For Vi = V2, the bound states of the potential of Fig. 1.9 are also bound states of the potential of Fig. 1.10.
30
Problems and Solutions on Electromagnetism
(c) For continuum states of a given energy, there is only one independent solution for well 1, which is a stationary-wave solution with tp = 0 at x 0; there are two independent solutions corresponding to traveling waves in +x and x directions for well 2. (d) Let pi, p2 denote respectively the probabilities that the particle is inside and outside the well. Consider, for example, the odd-parity solution = A sin kx V = Ce~k x >
I/J
for for
" ~
0< x < a, a<x,
where A = * ^ p ^ v" = ' "h (i 1, 2), h
*' = ^ P ^ , for which h A2 k'a C2 e~2k'a ( [} sin 2ka 2ka
Pi p~2
Jo A2 sin2 kx dx /a C2e~2k'x dx
The continuity of ip at x = a gives A C e-k'a sin ka
Setting, as before, rj = k'a, = ka, we have Pi V

2
(\
sin 2^\
2
2?7
l
(^
2
sin 2
2
P2
sin C V"
cos
V"
The odd-parity solutions are given by j cot = -77, U 2 + r / 2 = 72, where 7 2 = ^ ^ (i = 1, 2). An analytic solution is possible if 7 (n + | ) 7r, or > , (n+i) 2 7T 2 /i 2 2 VJa -+? , n = 0, 1, 2, . . . 2m for which the solution is (n + ^)n, r) - 0, and >
P2
The particle is then more likely outside the well than inside.
Basic Principles
and One-Dimensional
Motions
31
1020 Obtain the binding energy of a particle of mass m in one dimension due to the following short-range potential: V{x) = -V0S{x). (Wisconsin) Solution: The Schrodinger equation d2iP/dx2 + ?^[E-V nr on setting k = y/2m\E\/h, can be written as ip"{x) - k2ip(x) + U0 5(x) ip{x) = 0. Integrating both sides of the above equation over x from e to e, where is an arbitrarily small positive number, we get ip'(e)-ip'{-e)-k2 which becomes, by letting e -> 0, f i> dx + U0rl>(0) = 0, U0 = 2mV0/h2 , (x)] $ = 0, (E<0),
^'(o+) - v'(cr) + c/0v(o) = o.

At x ^ 0 (6(x) = 0) the Schrodinger equation has solutions i>(x) ~ exp (kx) ip(x) ~ exp (fca;) It follows from Eq. (1) that for x > 0, for x < 0 .
(i)
V>'(o + )-v'(<r) = -2fcV>(o).

A comparison of the two results gives k = Uo/2. Hence the binding energy is -E - h2k2/2m = mV02/2h2.
32
Electromagnetism
1021 Consider a particle of mass m in the one-dimensional 8 function potential V(x) = VQ6(x). Show that if Vo is negative there exists a bound state, and that the binding energy is TUVQ /2h2. (Columbia) Solution: In the Schrodinger equation d2ip/dx2 + 2m[E - V(x)\ ip/ti2 = 0, we set E < 0 for a bound state as well as k2 = 2m\E\/h2, and obtain d2ip/dx2 - fcV - U0 6(x)ip = 0. Integrating both sides over x from e to +e, where e is an arbitrarily small positive number, we obtain V>' (e) - V>'(-e) k2
U0 =
2mV0/h2,
i>dx-
U0ip(0) = 0.
With -> 0+, this becomes ip'{Q+) - tp'(0~) = U0ip(0). For x ^ 0 the Schrodinger equation has the formal solution ip(x) ~ exp (k \x\) with k positive, which gives \x\ x and hence V>'(0+) - V'(0~) = -2&^(0) = CZoV'(O) Thus fc = Uo/2, which requires VQ to be negative. The energy of the bound state is then E = ^ - = -TUVQ /2h2 and the binding energy is Eb = 0 E = mVo2/2/i2. The wave function of the bound state is V>(x) = A exp ( - | x | 1 = y/-mV0/h2 exp (mV0 \ x \/h2), ,, , f -ke~kx, [ke ,
kx
x > 0, x<0,
Basic Principles
and One-Dimensional
Motions
33
where the arbitrary constant A has been obtained by the normalization
/-, ^dx + C tfdx = l.

1022 A particle of mass m moves non-relativistically in one dimension in a potential given by V(x) = a5(x), where 5(x) is the usual Dirac delta function. The particle is bound. Find the value of xo such that the probability of finding the particle with | x | < XQ is exactly equal to 1/2. (Columbia) Solution: For bound states, E < 0. The Schrodinger equation
" At A ip(x) = Etp{x) ^ r~?r aoix) v ; 2m dx2 has for x ^ 0 the solutions finite at x = oo as follows,
ft2
Ms) = {
f A ekx
for x < 0, for a; > 0 ,
I A e-kx
where A = * | m g and A is an arbitrary constant. Applying lim_>.o+ ; / dx to the Schrodinger equation gives
<//(o+)-v/(o-) = - ^ V ( o )
since / ip(x)6(x)dx = tp(0), lim / ip(x)dx = 0 s J-e ^J-e for finite tp(0). Substitution of ip(x) in (1) gives ma Hence ma I a; tp (x) = A exp I tfi
(i)
34
Electromagnetism
On account of symmetry, the probabilities are P(|a;|<xo)=2|A|2

2
/ Jo
r x
e~
2kx
dx= -^- k \A\2
I A I2
(l-e-2kx),
r
Jo
P(-oo
<x<
oo) = 2
\A\
~2kx dx =
As it is given 1-e have

-2fcln
1 ft2 Xo = r In 2 = In 2. 2k 2ma
1023 A particle of mass m moving in one dimension is confined to the region 0 < x < L by an infinite square well potential. In addition, the particle experiences a delta function potential of strength A located at the center of the well (Fig. 1.11). The Schrodinger equation which describes this system is, within the well, _^_ 2m d2^{x) + \6(xdx2 Vlx) L/2)ip(x) = Eip(x), 0<x<L.
L/2 Fig. 1.11
Find a transcendental equation for the energy eigenvalues E in terms of the mass m, the potential strength A, and the size L of the system. (Columbia)
Basic Principles
and One-Dimensional
Motions
35
Solution: Applying lim_>o $L/2-E ^X t o k t n s ^ e s f t ' i e Schrodinger equation, we get V>' (L/2 + e) - V' (L/2 - e) = (2m\/h2) V (L/2), (1) since
/^%(*)*(*-|)fa = ^ ( ) , ^ T _ % ( * ) i x = 0.
Subject to the boundary conditions V() = VK-^) = 0, the Schrodinger equation has solutions for x ^ |f : f i4x sin (kx), ( A2 s i n [ * ( x - L ) ] , 0<x< L/2 - e
L/2 + < a ; < L ,
where fc = v ' 2 " f i and e is an arbitrarily small positive number. The continuity of the wave function at L/2 requires A\ sin(fcL/2) = A? sin(fcL/2), or A\ = Ai- Substituting the wave function in (1), we get A2k cos (fcL/2) -Aik cos (jfcL/2) = (2m\Ai/h2) sin (fcL/2), whence tan ^ = - ^ , or tan V 2 ^ L = - \ A f , which is the transcendental equation for the energy eigenvalue E.
1024 An infinitely deep one-dimensional square well potential confines a particle to the region 0 < x < L. Sketch the wave function for its lowest energy eigenstate. If a repulsive delta function potential, H' = X8(x L/2) (A > 0), is added at the center of the well, sketch the new wave function and state whether the energy increases or decreases. If it was originally EQ, what does it became when A oo? > (Wisconsin) Solution: For the square well potential the eigenfunction corresponding to the lowest energy state and its energy value are respectively 0o (x) = A / 2 / L sin(7ra;/L), E ^
36
on
Electromagnetism
A sketch of this wave function is shown in Fig. 1.12 With the addition of the delta potential H' = X5(x L/2), the Schrodinger equation becomes i>" + [k2 - aS(x - L/2)} 4> = 0, where k2 = 2mE/h2,a 2mX/h2. The boundary conditions are
iiK
V(0)=t/;(L) = 0,
(1)
(2)
V'
= V(L/2),
V >
Note that (2) arises from taking lime_o [I
2
1>
e
(3)
dx over both sides of the
Schrodinger equation and (3) arises from the continuity of ip(x) at x = \ . The solutions for x ^ ^ satisfying (1) are Ai sin(A;a;), 0 < x < L/2,
rj,-.
^ A2 sin[k(x - L)},
L/2<x<L.
Let k = ko for the ground state. Condition (3) requires that A\ = A2 = A, say, and the wave function for the ground state becomes A sin(kox), 0<x<L/2, V'o(z) -A sm[k0(x-L)}, L/2<x<L.
Basic Principles
and One-Dimensional
Motions
37
Condition (2) then shows that fc0 is the smallest root of the transcendental equation cot(kL/2) = - ^ .
As cot ( ^ ) is negative, TT/2 < k0L/2 < IT, or TT/L < k0 < 2ir/L. The new ground-state wave function is shown Fig. 1.13. The corresponding energy is E = ti2 k\j2m, > EQ = ^j? , since fc0 > . Thus the energy of the new ground state increases. Furthermore, if A -> +00, fco -) 2n/L and the new ground-state energy E -> 4E 0 .
*-x
1025 A nonrelativistic particle of mass m undergoes one-dimensional motion in the potential V (x) = -g [5 (x - a) + S (x + a)} where g > 0 is a constant and S(x) is the Dirac delta function. Find the ground-state energy eigenfunction and obtain an equation which relates the corresponding energy eigenvalue to the constant g. {Berkeley) Solution: Since V (x) = V (x), the energy eigenfunctions have definite parity. The ground state has even parity, ip(x) = ip(x). It is a bound state and its energy is negative, E < 0. For x > 0, the Schrodinger equation is [-(H2/2m) d2/dx2 -gS(xa)} ip{x) = Eip{x),
38
on
Electromagnetism
whose solutions for x ^ a are ip ~ exp (kx), where k = y/2mE/h. With the condition that the wave function remains finite as x oo and > has even parity, we obtain
{ {
D
Ae~kx,
x > a,
B cosh (kx), 0 < x < a. The continuity of ip at x = a requires that A = Beka cosh (ka). Thus Beka cosh (ka) e~kx, B cosh(fca;), x > a, 0 < x < a.
Normalization /0 ip2 dx + f^ ip2 dx = \ gives
, z2ka
2k
l+
2k
2kaY1/2
) '
At x = a, there is a discontinuity of the first differential of the wave function (cf Problem 1024): V>'(a+) - Tp'(a-) = -(2mg/h2) Substitution of tp gives k [1 + tanh(fca)] = 2mg/K2 , which is to be satisfied by k. By symmetry the wave function for the entire space is f Beka cosh (ka) e~k\x\ |x|>a, ip(a).
V-(*) = <
I B cosh(fc:c), | x | < a.
1026 An approximate model for the problem of an atom near a wall is to consider a particle moving under the influence of the one-dimensional potential given by V(x) = -V05(x), x>-d, V(x) = oo, x < d,
Basic Principles and One-Dimensional Motions
39
where S (x) is the so-called "delta function". (a) Find the modification of the bound-state energy caused by the wall when it is far away. Explain also how far is "far away". (b) What is the exact condition on Vo and d for the existence of at least one bound state? (Buffalo)
- X
Fig. 1.14
Solution: (a) The potential is as shown in Fig. 1.14. In the Schrodinger equation ip" + (2m/h2) let k = y/2mE/h, tp(x) [E + V05 (x)} tp = 0, x > -d,
where E < 0. This has the formal solutions aekx + be~kx e"kx for for - d < x < 0, x > 0,
as ip(x) is finite for x > oo. The continuity of the wave function and the discontinuity of its derivative at x = 0 (Eq. (1) of Problem 1020), as well as the requirement tp(x = d) = 0, give a + b= 1, -k - (a - b)k = -2mV0/h2 ae~
kd
+ be
kd
0.
40
on
Electromagnetism
Solving these we find

2kd
1 - e2kd mVo h2 (1
b=
-2kd
1-e
Ikd
The wall is "far away" from the particle if kd 3> 1, for which k mVo/ti2. A better approximation is k (mVo/h2)[l exp (2mVo d/h2)], which gives the bound-state energy as h2k2 E = 2m mV02 2h2 _tf_ ~2m (mVp \ H2 2mV0d\ 2mV0d\ 2mV0d
exp
1 2 exp
mVf
2h2
mVf
h2
exp
The second term in the last expression is the modification of energy caused by the wall. Thus for the modification of energy to be small we require d 1/fc = h2/mVo- This is the meaning of being "far away".
*. k
Fig. 1.15
(b) Figure 1.15 shows line 1 representing y = k and curve 2 representing y = ye [1exp (2kd)}, where ye = mV0/h2. The condition for the equation
k = mV0 [1 - exp (-2kd)]/h2
Basic Principles and One-Dimensional
Motions
41
to have a solution is that the slope of curve 2 at the origin is greater than that of line 1: dy_ = 2mV0d/h2 > 1. dk fc=0 Hence if V$d > ^ , there is one bound state.
1027 The wave function of the ground state of a harmonic oscillator of force constant k and mass m is V o (x) = (a/71-)1/4 e~ax2/2, > a = muja/h, LJI = k/m.
Obtain an expression for the probability of finding the particle outside the classical region. (Wisconsin) Solution: The particle is said to be outside the classical region if E < V(x). For the ground state, E = HUJO/2 and the nonclassical region is | hu>o <
5 muiQX
1 0 1
, i.e.,
x2> worn
x > = , a or<
x <
The probability of finding the particle in this nonclassical region is therefore
P= f
J\*\>y/i
ij)l{x)dx
dx
J-00
V Idx
Jy/lJ^ V 7 T
Jy/l/Z
V*
- f -J.
e-t di~16%.
42
Electromagnetism
1028 Consider a linear harmonic oscillator and let ipo and tpi be its real, normalized ground and first excited state energy eigenfunctions respectively. Let Aipo + Bip\ with A and B real numbers be the wave function of the oscillator at some instant of time. Show that the average value of x is in general different from zero. What values of A and B maximize (x) and what values minimize it? (Wisconsin) Solution: The orthonormal condition I"(Ail>0 + Bipi)2 dx = 1 gives A2 + B2 = 1. Generally A and B are not zero, so the average value of x, (x) = J x(Aip0 + Bipi)2 dx = 2AB (rp0 | x | Vi) is not equal to zero. Rewriting the above as (x) = [1 - (A2 +B2 + 2AB)} (Vo I x | Vi) = {l-(A-B)2]Wo\x\ip1)
and considering / = AB = A (1 A2) *, which has extremums at A = 4=, we see that if A = B = l/V%, (x) is maximized; if A = B = 1/v^, (x) is minimized.
1029 Show that the minimum energy of a simple harmonic oscillator is fku/2 if AxAp = ft/2, where (Ap) 2 = ((p - (p)) 2 ). (Wisconsin) Solution: For a harmonic oscillator, (x) = (p) = 0, and so (Ax)2 = (x2), (Ap) 2 = (p 2 ).
Basic Principles
and One-Dimensional
Motions
43
Then the Hamiltonian of a harmonic oscillator, H = p2/2m gives the average energy as (H) = (p2)/2m + mu2{x2)/2 = (Ap)2/2m
mw2x2/2,
+ mu2(Ax)2/2
As for a, b real positive we have {^/a Vb)2 > 0, or a + b > 2\/ab, (H)min = (Ap)(Ax)u> = huj/2.
1030 An electron is confined in the ground state of a one-dimensional harmonic oscillator such that y/{{x {x))2) = 1 0 - 1 0 m. Find the energy (in eV) required to excite it to its first excited state. [Hint: The virial theorem can help.] (Wisconsin) Solution: The virial theorem for a one-dimensional harmonic oscillator states that (T) = (V). Thus E 0 = {H) = (T) + (V) = 2(V) = m e w 2 (a;2), or, for the ground state, = meu}z giving h 2me (x 2 ) " As (x) = 0 for a harmonic oscillator, we have y/((x - (x))2) = v/(x2> - (x) 2 = J&2) = 10- 1 0 m . (xz),
The energy required to excite the electron to its first excited state is therefore AE = tkj = 2me (x 2 ) 2m e c 2 (x 2 )
(6.58 x 10- 1 6 ) 2 x (3 x 10 8 ) 2 = 3.8 eV. 2 x 0.51 x 10 -20
44
on
Electromagnetism
1031 The wave function at time t = 0 for a particle in a harmonic oscillator potential V = \ kx2, is of the form ip(x, 0) = A e - ( a x ) 2 / 2 . sin P TT a TT . cos p Ho (ax) H ^= #2 (ax) 2v2 and the Hermite polyno-
where /? and A are real constants, a 2 = Vmk/h, mials are normalized so that f
J-oo
e~a2x2 [^x(aar)] 2
dx=^-2nn\.
Oi
(a) Write an expression for ip{x, t). (b) What are the possible results of a measurement of the energy of the particle in this state, and what are the relative probabilities of getting these values? (c) What is (x) at t = 0? How does (x) change with time? (Wisconsin) Solution: (a) The Schrodinger equation for the system is ihdtip(x, t) = Hi}>(x, t),
where tp(x, t) takes the given value tp (x, 0) at t = 0. As H does not depend on t explicitly,
V>(x, t) =
^n(x)e-iE^h,
where tpn(x) is the energy eigenfunction satisfying Htpn(x) = Expanding ip(x, 0) in terms of ipn (x): ip(x, 0) = ^ a - 0 n ( x ) ,
n
Enxpn(x).
where an= i)*n{x)i>{x, 0)dx.
45
Thus ip(x, t) = Y2an'ipn{x,

n
t) = ^2ani>n(x)
n
e-iEnt/h
For a harmonic oscillator, ^jn(x)=Nne-a2x2/2Hn(ax), so that an = f Nne-a2x2'2Hn(ax) cos (3 Ho (ax) H As the functions exp( ^x ) Hn(x) a 0 = AiV0 a2 = AN2 J ^ Hence il>(x,t)=A J ^ + 2V2N2 [No cospô(x)e-iEotlh sinj3ip2(x)e-iE*t/h].
r . E0t . E2t 1
Ae^^l2
-= H2 (ax) dx. 2 V2
are orthogonal, all a n = 0 except cos (3,

2A
/2
sin /3.
J cos 0-00(a;) e J 1 +sin / 5'02(a;) e l " , as JVn are given by f[ipn(x)}2 dx = 1 to be iV0 = (*)i, 7V2 = ^ L = ( ^ ) l . (b) The observable energy values for this state are EQ = fkjj/2 and E2 = 5 hjj/2, and the relative probability of getting these values is P0 /p2 = cos2 P/ sin2 ,3 = cot 2 /?. (c) As ip (x, 0) is a linear combination of only ipo (%) and tp2 (%) which have even parity, i)(-x, 0 ) = V ( x , 0).
7r \ J
46
Electromagnetism
Hence for t = 0,
<) = / ^ M ) ^ o ) ^ = o. *
It follows that the average value of x does not change with time.
1032 (a) For a particle of mass m in a one-dimensional harmonic oscillator potential V m w V / 2 , write down the most general solution to the timedependent Schrodinger equation, tp(x, t), in terms of harmonic oscillator eigenstates <f>n(x). (b) Using (a) show that the expectation value of x, (i), as a function of time can be written as A cos wt + B sin u>t, where A and B are constants. (c) Using (a) show explicitly that the time average of the potential energy satisfies {V) = \{E) for a general ip(x, t). Note the equality frruJ 1 ln + 1 fn W -jT- X(j)n = W <pn+i + W - 4>n-l (Wisconsin) Solution: (a) From the time-dependent Schrodinger equation ih tp(x, t) = Hip(x, as H does not depend on time explicitly, we get fp(x, t)=e'iHt/h We can expand ip(x, 0) in terms of <f>n (x): 1p(x, 0) = ^ 0.n<S>n{x),
n
t),
ip(x,0).
where an = {K{x)\ip(x, 0)),
Basic Principles
and One-Dimensional
Motions
47
and 4>n (x) are the eigenfunctions of H<j>n(x) = En<pn(x), Hence

n
with
En=(n+-\
hu.
(b) Using the given equality we have (x) = I ip* (x, t)xip(x, = > > , e-^-'-W* t) dx J <t>*n{x)x<t>n,{x) dx
n,n'
/^.
V 2
\ nr
/ V mo;
= J2an I 0 "- 1 J| eiU "+ a +l Y

= A cos uit + B sin wt, where
A=
mw
V^7,T/<
\"n-x\l + an+l
h x* . . i M /n + 1 J2 ia*n I a n _i W - - a n + i y V ma;
and we have used EW+i -E,, = fkv. (c) The time average of the potential energy can be considered as the time average of the ensemble average of the operator V on ip(x, t). It is sufficient to take time average over one period T = 2TT/LJ. Let (A) and A
48
on
Electromagnetism
denote the time average and ensemble average of an operator A respectively. As 1 mw 2 1 2 1 . Imw y-^ Imcj V ~T~ x Z ^ a " V ~T~ ^ /mw ^> / ln+1 .Eat * .
W .
/ ^ |
_ l ) )
c--(n+l/2)t
o ^ Tl=0 a 5Z
+
(n + l ) ( n + 2)
n + 2) -*%*
(n+|)|n>
^/r^|-2>
we have
v = (VI v IV) = 27^Sa"a" (n+2)

n=0 ^ '
+
2fiWâ"+2
n=0
/(n + l ) ( n + 2) 1l2a t 1. ^ xany ^ ^ i e 2 u" + - hw 2^ an (n + 1) (n + 2) x a n+2 -hw^2a*nan

n=0
-i2ujt
("+2)
^ '
I + 2 ^ 5 3 I a " + 2 a " I V ( n + l ) ( n + 2) cos (2wt + c5),

n=0
where Sn is the phase of a^+1a. term becomes zero, we get

1 rT
Averaging V over a period, as the second

, 0 0 / i\
Basic Principles
and One-Dimensional
Motions
49
On the other hand,
E=(iP\H\i>)=hLjJ2 <a n (n + i J ,
and (E)=E. Therefore (V) = {E)/2.
1033 Consider a particle of mass m in the one-dimensional potential V(x) =rmj2x2/2, V(x) = V0, \ x | > b; \x\<b,
where Vo ~> h2/mb2 Hu>, i.e. a harmonic oscillator potential with a high, thin, nearly impenetrable barrier at x = 0 (see Fig 1.16). V[x\
Fig. 1.16
(a) What is the low-lying energy spectrum under the approximation that the barrier is completely impenetrable? (b) Describe qualitatively the effect on the spectrum of the finite penetrability of the barrier. {MIT) Solution: (a) For the low-lying energy spectrum, as the barrier is completely impenetrable, the potential is equivalent to two separate halves of a harmonic oscillator potential and the low-lying eigenfunctions must satisfy the condition tp (x) = 0 at x = 0. The low-lying energy spectrum thus corresponds
50
on
Electromagnetism
to that of a normal harmonic oscillator with odd quantum numbers 2n + l, for which ipn (x) = 0, at x = 0 and En = (2n + 3/2)hw, n = 0, 1, 2, . . . with a degeneracy of 2. Thus only the odd-parity wave functions are allowed for the low-lying levels. (b) There will be a weak penetration of the barrier. Obviously the probability for the particle to be in \x\ < b, where the barrier exists, becomes less than that for the case of no potential barrier, while the probability outside the barrier becomes relatively larger. A small portion of the even-parity solutions is mixed into the particle states, while near the origin the probability distribution of even-parity states is greater than that of odd-parity states. Correspondingly, a small portion of the energy E'n = (2n+l/2)hu> is mixed into the energy for the case (a). Since (ip | barrier potential \xj)) > 0, the energy levels will shift upwards. The level shifts for the even-parity states are greater than for odd-parity states. Furthermore, the energy shift is smaller for greater energies for states of the same parity.
1034 The Hamiltonian for a harmonic oscillator can be written in dimensionless units (TO = h = CJ = 1) as H = a+a + 1/2, where a= (x + ip)/V2, a+ = (x-ip)/V2. (-x2/2). One unnormalized energy eigenfunction is V>0 = (2z 3 - 3x) exp
Find two other (unnormalized) eigenfunctions which are closest in energy to tpa. (MIT) Solution: In the Fock representation of harmonic oscillation, a and a+ are the annihilation and creation operators such that atpn = v ^ n - i , En= (n+-j a+ipn = \/n + lVn+i, fkv, n = 0, 1, 2, . . . . aa+ipn = (n + 1) ipn ,
Basic Principles
and One-Dimensional
Motions
51
As
= 1 (x+4-) (4xi-l2x2 2 \ ax J = A{2x3-3x)
+ 3)e-^
e~^ = (3 + l)V,
we have n = 3. Hence the eigenfunctions closest in energy to tpa have n = 2, 4, the unnormalized wave functions being
~ (2x2 - 1) e-^ 2 / 2 ,
^ = \^-^{x-Tx)^^^12
~ (4x 4 - 12x2 + 3) e " x 2 / 2 , where the unimportant constants have been omitted.
1035 At time t = 0 a particle in the potential V{x) = mu>2x2/2 is described by the wave function 1>(x, 0) = A ^ ( l / v / 2 ) " ^ ( x ) ,
n
where tpn (x) are eigenstates of the energy with eigenvalues En = [n + 1/2)fiw.You are given that (i/>n, tpni) = <5nn'. (a) Find the normalization constant A. (b) Write an expression for IJJ(X, t) for t > 0. (c) Show that | VK2^ *) | 2 i s a periodic function of time and indicate the longest period r . (d) Find the expectation value of the energy at t = 0. (Berkeley)
52
on
Electromagnetism
Solution: (a) The normalization condition
(V> (x, 0), V (x, 0)) = I A | 2 J^ (l/2)( + ">/ 2 (V, Vm)

m,n
gives A = l / \ / 2 , taking A as positive real. (b) The time-dependent wave function is 1>(x,t)=e-iAt/hil>(x,0)
=E
(^)"+1e-^-+*^(x).
(c) The probability density is
mn
Note that the time factor exp [iui(n m)t] is a function with period (n-m)uj' *^ e rnaximum period being 2IT/LJ. (d) The expectation value of energy is
m-\-n . i
H=(^(a;,0) ) ^(i I 0)) = 53 Q ) '

m,n ^ '
W>> ^ m )
= E (2) = E 2 ^ r (n
Noting
n=0
(m+2)
+
/Sw nm
\) *"
l a; 1'
E_ x" ~ x -
Motions
or, by differentiation,
n=0
-n it
xn+l -y.n+1
-1
(j. _ !)2
we have
n=0
Z ^ on+l 2
'
and hence H 3tuv/2.
1036 Consider the one-dimensional motion of a particle of mass fi in the potential V(x) = V0(x/a)2n, where n is a positive integer and Vo > 0. Discuss qualitatively the distribution of energy eigenvalues and the parities, if any, of the corresponding eigenfunctions. Use the uncertainty principle to get an order-of-magnitude estimate for the lowest energy eigenvalue. Specialize this estimate to the cases n = 1 and n oo. State what V(x) becomes in these cases and > compare the estimates with your previous experience.
(Buffalo)
Solution: Since the potential V(x) - > o o a 5 i - > oo, there is an infinite number of bound states in the potential and the energy eigenvalues are discrete. Also, the mth excited state should have m nodes in the region of E > V(x) given by kAx ~ (m + 1)TT. Ax increases slowly as ra increases. From the virial theorem 2T oc 2nV, we have k2 oc (Ax) 2 n oc [(m + l)7r/fc]2n , and so Eock2 (x(m+lfn/{n+1). Generally, as n increases, the difference between adjacent energy levels increases too. Since V(x) = V(x), the eigenstates have definite parities. The ground state and the second, fourth,... excited states have even parity while the other states have odd parity.
54
Electromagnetism
The energy of the particle can be estimated using the uncertainity principle px ~ h/2b, where
b= yJTKxf.
Thus E~(H/2b)2 + V0(b/a)2n.
For the lowest energy let dE/db = 0 and obtain b= Hence the lowest energy is E ~ [(n+ l)V0/a2n] (ftVn/8MnV0)n/(n+1). (h2a2n/8nnV0)1^n+1K
For n = 1, V{x) is the potential of a harmonic oscillator, V(x) = V0x21 a2 = nu2x2/2 .
In this case E equals HUJ/2, consistent with the result of a precise calculation. For n = oo, V(x) is an infinite square-well potential, and E = h2/Sfj.a2, to be compared with the accurate result h2TT2/2/j,a2.
1037 Consider a particle in one dimension with Hamiltonian H = p2/2m + V(x), where V(x) < 0 for all x, V(oo) = 0, and V is not everywhere zero. Show that there is at least one bound state. (One method is to use the Rayleigh-Ritz variational principle with a trial wave function V>(z) = {b/ir)1/4 exp{-bx2/2). (Columbia)
However, you may use any method you wish.)
Basic Principles
and One-Dimensional
Motions
55
Solution: Method 1: Assume a potential V(x) = f (x) as shown in Fig 1.17. We take a square-well potential V (x) in the potential V(x) such that V'(x) - -V0, V'(x) = 0, V'(x)>f(x) for \x\ < a, \x\>a, all x .
* x
Fig. 1.17
We know there is at least a bound state <p(x) in the well potential V (x) for which (<p(x) \H' | <p(x)) = (<p | p2/2m + V (x) | v > > = E0 < 0. We then have (<f\H\<p)=(<p\p2/2m < = + f(x)\<p)
(ip\p2/2m+V'{x)\<p) E0<0.
Let ipn-i (x), tpn{x), denote the eigenfunctions of H, and expand
56
As
(ip\H\ip) =
n
Y,\C-\2(^\H\i}n)<0,
there is at least an eigenfunction tpi (x) satisfying the inequality Wi\H\^)<0. Hence there exists at least one bound state in V(x). Method 2: Let the wave function be 1>(x) = (&/*")1/4 exp (-bx2/2),
where b is an undetermined parameter. We have
h2b/m + {V}
where / and thus
oo
V(x)
-OO
exp (-6a; 2 ) dx ,
d{H) _ rf_
db Am
1 I
/ h\ 1/2 '
.+ f+oo
giving
l<V>-<x>V>=0,
(V) (x2V) 4m
Substitution in the expression for (H) yields 2(x V) E=(H)

2
~ (V) 4m h2 2 (x V) -
h2
Basic Principles
and One-Dimensional
Motions
57
As V(x) < 0 for all x, V(oo) = 0, and V is not everywhere zero, we have (V) < 0, (x2 V) < 0 and hence E < 0, b > 0. In fact, under the condition that the total energy has a certain negative value (which must be greater than (V) to make (T) positive), whatever the form of V a particle in it cannot move to infinity and must stay in a bound state.
1038 The wave function for a particle of mass M in a one-dimensional potential V (x) is given by the expression ip(x,t) = ax exp(-px) = 0, exp(i'yt/h), x > 0, x < 0,
where a, (3 and 7 are all positive constants. (a) Is the particle bound? Explain. (b) What is the probability density p(E) for a measurement of the total energy E of the particle? (c) Find the lowest energy eigenvalue of V(x) in terms of the given quantities. (MIT) Solution: (a) The particle is in a bound state because the wave function ip(x, t) satisfies lim
XV OO
ip(x, t) = 0 , il>(x,t)= lim

X + OO
lim
X - + + OO
axe~<}xei'lt,h
= 0.
(b), (c) Substituting the wave function for x > 0 in the Schrodinger equation ih ip(x, t) = dt gives -lx = -^(p2x-2p) + V(x)x,
Wdx->+V{x)
ip(x, t)
58
Electromagnetism
whence the potential for x > 0:
V(x) =
-1+^(f32-2/3/x).
As the stationary wave function of the particle in V (x) satisfies h2 2M or j-2 by setting E' = E + 7 - P2h2/2M, e2 = 0h2/M,
*PE
( d2 ' \dx2
01 + 2(3/x\
i>E(x) = (E + 7) i>E(x),
(x > 0)
{x) + -jp- (E' + e2/x)
V E (X)
= 0
and ipE (x) >0, we see that the above equation is the same as that satisfied by the radial wave function of a hydrogen atom with Z = 0. The corresponding Bohr radius is a = H2/Me2 = 1/(3, while the energy levels are E'n = -Me4/2h2n2 = -01h2/2Mn2, n = 1, 2, . . . . Hence En = - 7 + (f32h2/2M) (1 - 1/n 2 ), n = 1, 2, . . . , and consequently the lowest energy eigenvalue is E\ 7 with the wave function ip{Xi t) = ax exp {fix) exp {i'yt/K) oc ipEi (x) exp {iE\t/K). The probability density p(E) = ip*ip =
V'EIV'BI
is therefore
( 1 for E = - 7 , P(E) = I 1 0 for E ^ - 7 .
1039 A particle of mass m is released at t = 0 in the one-dimensional double square well shown in Fig. 1.14 in such a way that its wave function at t = 0
Basic Principles
and One-Dimensional
Motions
59
is just one sinusoidal loop ("half a sine wave") with nodes just at the edges of the left half of the potential as shown.
Vix)
4>U,0) "
kî
Fig. 1.18
- x
(a) Find the average value of the energy at t = 0 (in terms of symbols defined above). (b) Will the average value of the energy be constant for times subsequent to the release of the particle? Why? (c) Is this a state of definite energy? (That is, will a measurement of the energy in this state always give the same value?) Why? (d) Will the wave function change with time from its value at t = 0? If "yes", explain how you would attempt to calculate the change in the wave function. If "no", explain why not. (e) Is it possible that the particle could escape from the potential well (from the whole potential well, from both halves)? Explain. (Wisconsin) Solution: (a) The normalized wave function at t = 0 is ip(x, 0) Thus
K2
I sin 22.
(KX\
{H)t=0
= -V0
h2TT
2_2
"'
'""
2m a J0
-Va.
sm
(1TX\
"
KT) dx2
Va )
dx
2ma
(b) (H) is a constant for t > 0 since d(H)/dt = 0. (c) It is not a state of definite energy, because the wave function of the initial state is the eigenfunction of an infinitely deep square well potential
60
Electromagnetism
with width a, and not of the given potential. It is a superposition state of the different energy eigenstates of the given potential. Therefore different measurements of the energy in this state will not give the same value, but a group of energies according to their probabilities. (d) The shape of the wave function is time dependent since the solution satisfying the given conditions is a superposition state: tp(x, 0) = Ji>{x,t) = n
sin ( J = ^2
n Y,Cni>n{x)e-iE"tlh.
cntpn{x),
The shape of t^(x, t) will change with time because En changes with n. (e) The particle can escape from the whole potential well if the following condition is satisfied: h2ir2/2ma > VQ. That is to say, if the width of the potential well is small enough (i.e., the kinetic energy of the particle is large enough), the depth is not very large (i.e., the value of Vo is not very large), and the energy of the particle is positive, the particle can escape from the whole potential well.
1040 A free particle of mass m moves in one dimension. At time t = 0 the normalized wave function of the particle is xl>{x, 0,o*) = ( 2 ^ ^ - 1 / 4 e x p ( - x 2 / 4 a 2 ) , where a2 = (x2). (a) Compute the momentum spread ap = \J(p2) (p) 2 associated with this wave function. (b) Show that at time t > 0 the probability density of the particle has the form \iP(x, t)\2 = mx, 0,a2x+a2pt2/m2)\2. (c) Interpret the results of parts (a) and (b) above in terms of the uncertainty principle. (Columbia)
Solution: (a) As
(p) = / ip* I ih ~\ipdx dx J-oo V J r 1 / X \ l o o ( 2 ^ ) 1 / 2 V 2a%)
rOO
(P )
J OO
V* >
= -ft2 x e
(-
*)
1
ipdx
/ : oc (21 ^ ) 1 / 2 dx == ft2/4^, ft
2CTX
l"2^|
+ 4^
z2
= V^> =
(b) By Fourier transform,
*<*>0)
(dljv* / e _ w " ^ 0 ) d*
ZTrftjVa y (27rft)1/2 7
e-ipx/h
(2^2)1/4 /V27rh}exp[-a2xp2/h2}
x exp (-a; 2 /4cr 2 ) dx = [(2^)1/4 Then
where
62
Electromagnetism
for a free particle. By inverse Fourier transformation ip[x, t)= I eipx/hip(p, t) dp
( 2 ^ ) 1 / 4 f pipx/n ___ / , P ipx/n -i-trt V2(nh) J e exp

.
. IA .2 \ 11/4
-'-)*
x exp
( ')
1/2
\^(x,t)\2
*U+&-
exp
x" ~2
2*2
= |V(a; ) 0 ) ^ + o -^m-'')| 2 .
(c) Discussion: (i) The results indicate the width of the Gaussian wave packet at time t (which was originally ax at t = 0) is â2+a2t2/m^ where a2 = f^/Aa2.. (ii) As <JxCTp = h/2, the uncertainty principle is satisfied.
,2 , 2 + 2 ^ - 2 \ |2
1041 A particle of mass m moves in one dimension under the influence of a potential V(x). Suppose it is in an energy eigenstate ip(x) = (T 2 / 7 1 ") 1 ^ 4 exp (~7 2 x 2 /2) with energy E = h2ry2/2m. (a) Find the mean position of the particle. (b) Find the mean momentum of the particle. (c) Find V(x).
Basic Principles
and One-Dimensional
Motions
63
(d) Find the probability P(p) dp that the particle's momentum is between p and p + dp. (Wisconsin) Solution: (a) The mean position of the particle is f 7 (x) = / ij)*{x)xil)(x) dx = -j= (b) The mean momentum is f _ 2 2 / xe 7 X dx = 0.
(P) = J r(x) J ( ^ W)<*z

=J*
r e -7v /2 rf (e -,v/2 )rfI=0 ,
(c) The Schrodinger equation
can be written as
As
^ ^2 v2x2/2 ^ / 2 4 2\ --y2i2/2
we have E-V(x) = -~(-i2 2m or

4 2 + 1
x ),
2m ' ' 2m 2m (d) The Schrodinger equation in momentum representation is p2 2m fi474 d2 2m dp2) iP(p) = Ei/>(p)
V(a;) = ^- w 7 V ( w
2 7
) + *- = * - ^ .
64
Electromagnetism
Letting iP(p) = Ne-ap and substituting it into the above equation, we get
22 \ -ap (-2a + 4 a V ) e
2
2m or
2m
Ee~
4a 2 p 2 - 2a
1 ft 7
2 2
P -1 \ h2j2
As the parameter a is independent of p, the above relation can be satisfied by a = l/27i 2 7 2 . Hence ip(p) = N exp(-p 2 /2ft 2 7 2 ). This is the eigenfunction of the state with energy h2f2/2m in the momentum representation. Normalization gives N = (l/?i 2 7 2 7r) 1 / 4 . Thus the probability that the particle momentum is between p and p + dp is
1/2
P(p)dp=\^(p)\2dp
h 7 ir
2 2
exp
- fi272
dp.
Note that ip(p) can be obtained directly by the Fourier transform of tp(x):
2 \ 1/4
i>[P)
(2nhy/
-72x2/2
\ir)
2\ fll\ I / \ J
n
4
= Jf
dx
(27rft)V2
1/4
y/2hr/ ~ 7 1 /
exp
2ft 2 7 s
h2j2n
e-P/2?iV-
1042 In one dimension, a particle of mass m is in the ground state of a potential which confines the particle to a small region of space. At time
Basic Principles
and One-Dimensional
Motions
65
t 0, the potential suddenly disappears, so that the particle is free for time t > 0. Give a formula for the probability per unit time that the particle arrives at time t at an observer who is a distance L away. (Wisconsin) Solution: Let
TJJQ
{x) be the wave function at t = 0. Then ip(x, t) = (x\ exp ' -ipH 2mh V'oO)
(x\e-ipt^2mV\x')dx'{x'\ip0)
-oo
(x\e~ipt^2m^\x')^0(x')dx',
-oo
where x exp = f pH_\ 2mh)

f + OO
j _ " {x\p')dp'
+oo i
^ ^
(p' px
fi,
expf . px'
7i
2mh) . p2t
n
P ) dp
dp(p\x')
-00
2ra?l
2^/-l P{"*[V2mh
+
(x' x) y/m I2U
dp exp i{x' x)2 m 2ht
2-KK,
v-[^-^s/;;-'
/m xht exp
i{x>
dq
1i Thus ^(a;, t) 1i
2
x)2
im l l ! 2ht\
I m
V ?rfa
exp [i(a:' - x ) 2 j i/>o ( x ' ) cte'.
66
Electromagnetism
Represent the particle as a Gaussian wave packet of dimension a: ipo(x) =

(TO2)-1/4
exp(-a:2/2a2).
The last integral then gives ip(x> *) = (1-j) rfrT

.2
TTht 7 T 1 / 4 a l / 2
x exp <i i
mx 1 ~2ht r
+
^Hm + it*)
2
i
1
Ihi "" 2a
L_
J-OO
/
m +
m
i>* e x p
mOT
1 m+ i ^
ai/2^1/4 \
whence the current density h2xt V x exp m /

<Jira,5m2
I1+ 7 ^ 0
A ,
&2tM3/2
a2 l+f-M,) 2
1
Vma^/
By putting x L, we get the probability per unit time that the particle arrives at the observer a distance L away.
1043 A free particle of mass m moves in one dimension. The initial wave function of the particle is ip(x, 0). (a) Show that after a sufficiently long time t the wave function of the particle spreads to reach a unique limiting form given by ip(x, t) = yjmjht exp (i7r/4) exp {imx1 /2ht) ip(mx/ht),
Motions
67
where <p is the Fourier transform of the initial wave function: <p(k) =
(2TT)~ 1 / 2
/ ip(x, 0) exp (-ikx)
dx.
(b) Give a plausible physical interpretation of the limiting value of |V(s, t)\\ Hint: Note that when a > oo, exp (iau2) \J-KJa exp (i-ir/4) 5 (u). (Columbia) Solution: (a) The Schrodinger equation is [ihd/dt + (h2/2m)d2/dx2]xp{x, By Fourier transform, we can write 1 rP(k,t) = -7=
V27T
t) = 0.
f /
J-oo
dxe~ikx i>(x, t),
and the equation becames
Integration gives i>{k,t) = i>{k,0) exp ( - * ^ ^ J . where i> (k, 0) = - = = J Hence ip(k, t) = y(k) exp [-i-^ dxe~ikx V (*, 0) = <p (k).
68
on
Electromagnetism
giving i ip(x,t)=-j= = r / dkeikxip(k,t)
, / dkip(k) exp I ikx i I yfa J-oo \ 2m J 1

V2TT
exp
/ . mx i -
-) L
dk
x exp W i t h = k mx/ht, ^ ,
S('-T)1*
this becomes ) _ e x p ^
( j B
For a oo, >
Va
a n d so after a long time t (t oo), >
eXp _i
( l)*W'
exp and , . ^ ( x
-<")-\/*>*K)
:
. 1 / . mx'' t ) = - ^ e x p f .
^(* + ^)
exp I i exp
)p eX
27rm ^ ft*
H)
exp (
V^
exp I i
l2nm
r'iM^J
imx\
n\
/mx\
[m
K)
mx ~2hi <P
Basic Principles
and One-Dimensional
Motions
69
(b)
Because <p(k) is the Fourier transform of tp(x, 0), we have

+oo r+oo
/
-oo
\ip(k)\2
dk=
/
Joo
|^(a;, 0) | 2 da;. /mx\\2

/- + 0O
On the other hand, we have + r+ /

+0O
/ |^(fc)| 2 dfc= /
-oo J oo
|V(x, 0)| 2 da;,
which shows the conservation of total probability. For the limiting case of t oo, we have ^
|V0M)l a ->o- Mo)| 2 = o,
which indicates that the wave function of the particle will diffuse infinitely.
1044 The one-dimensional quantum mechanical potential energy of a particle of mass m is given by V(x) = VoS(x), V(x) = oo, a<x<oo, x < a,
as shown in Fig. 1.19. At time t = 0, the wave function of the particle is completely confined to the region a < x < 0. [Define the quantities k = V2mE/h and a = 2mV0/h2} (a) Write down the normalized lowest-energy wave function of the particle at time t = 0. (b) Give the boundary conditions which the energy eigenfunctions V f (x) = Vi (x) >c and Vfc (x) = V^1 (x)
must satisfy, where the region I is -a < x < 0 and the region II x > 0.
70
on
Electromagnetism
Vlx)
II
-a 0
Fig. 1.19 -> X
(c) Find the (real) solutions for the energy eigenfunctions in the two regions (up to an overall constant) which satisfy the boundary conditions. (d) The t 0 wave function can be expressed as an integral over energy eigenfunctions:
oo
f(k)ipk(x)dk.
-oo
Show how f(k) can be determined from the solutions ipk (x). (e) Give an expression for the time development of the wave function in terms of f{k). What values of k are expected to govern the time behavior at large times? (Wisconsin) Solution: (a) The required wave function tp (x) must satisfy the boundary conditions ip(a) = rp(0) = 0. A complete orthonormal set of wave functions defined in a < x < 0 and satisfying the Schrodinger equation consists of - sin I , \ a / / a 10. where n = 1, 2, . . . , with {4>n | H \(j>m) = En Smn, En = (h2/2m) {nir/a)2 .
2 . /TVKX\
4>n(x)
a < x < 0, outside [a, 0].
The normalized lowest-energy wave function is given by n = 1 as ip(x) = 0, sin I I, a \ a /

2 . / TTX \
a < x < 0, outside [a, 0],
Basic Principles
and One-Dimensional
Motions
71
(b) The Schrodinger equation for x > -a is
or with
ip" (x) + k2ip (x) = a5 (x) ip (x) k2 2mE a 2mV0
The boundary conditions and the discontinuity condition to be satisfied are tP1 (-a) = 0, ij)1 (0) = ipu (0), </>" (+oo) = finite,
ip11' (0) - ip1' (0) = aV 1 (0) The last equation is obtained by integrating the Schrodinger equation over a small interval [e, e] and letting e - 0 (see Problem 1020). (c) In both the regions I and II, the wave equation is ip"(x) + k2ip{x)=0,
whose real solutions are sinusoidal functions. The solutions that satisfy the boundary conditions are ip\ (x) Cc sin k(x + a), f ipk (a:) = " 0" (x) = Ck sin k(x + a) + Ak sin kx, .0, a < x < 0, x > 0, x < a.
The discontinuity and normalization conditions then give

A
C
kOL
sin ka, asm 2ka ^
Ck
=U
1 +
fa sin b V \~k~)
(d) Expand the wave function t/>(x) in terms of ipk (#)>

oo
/(fc)Vfc(x)dfc, /
-oo
72
on
Electromagnetism
and obtain / Vfc' ix) ip (x) dx = / / f (k) tpk(x) rpl> (x) dkdx
-f
or / (e) As / we have
+oo
f(k)6{k-k')dk
f{k'),
oo
^(x)^(x)dx.
-a oo
f{k)i>k{x)dk,
-oo
-oo
At time = 0, the particle is in the ground state of an infinitely deep square well potential of width a, it is a wave packet. When t > 0, since the (5 (x) potential barrier is penetrable, the wave packet will spread over to the region x > 0. Quantitatively, we compute first
[2 TXX
Ck sin k (x + a) \ - sin dx Va a \l / Ck s cos (fc ) x + ka\ V 2a J_a { IA a) J cos \{k-\ Ja; + A;a > dx 7r" 2 sin ka
and then V(x, t) =-\l~ J o V a J_x

,,
N
7T
/2
Z--00
ck 72 gy k2_
s i n A:a
^ sin k(x + a) + j sin fca sinfcxJ
Motions
73
where Ek = h2k2/2m. In the last expression the upper and lower rows are for regions I and II respectively. When t > oo, the oscillatory factor exp(iEkt/h) changes even more rapidly, while the other functions of the integrand behave quite normally (k = re J a is not a pole). Thus tp(x, t) tend to zero for any given x. When t is very large, component waves of small wave number k play the principal role. At that time the particle has practically escaped from the region
[-a, 0].
1045 The radioactive isotope 83Bi decays to siTl 2 0 8 by emitting an alpha particle with the energy E = 6.0 MeV. (a) In an attempt to calculate the lifetime, first consider the finite potential barrier shown in Fig. 1.20. Calculate the transition probability T for a particle of mass m incident from the left with energy E in the limit T<1. (b) Using the above result, obtain a rough numerical estimate for the lifetime of the nucleus Bi 212 . Choose sensible barrier parameters to approximate the true alpha particle potential. (CUS) Vlx)
212
-*
Fig. 1.20
Solution: (a) If T <C 1, the incident wave is reflected at x = 0 as if the potential barrier were infinitely thick. We thus have
ij}{x)
eifcx +
(tl _ j)
e-ifex;
<
Q)
V'(x) =
t\e -k'x
0 < x < b,
74
Electromagnetism
where t\ is the amplitude transmission coefficient and
*-l/ÊS. * - h^'
2
The continuity of xjj' (x) at a; = 0 gives i f c ( 2 - f l ) = -*'*!, or 2fc = ^ - ^ .
Consider the reflection at b. We have ip(x) = tie-k'b [e"*'^-*) + (t 2 - 1) efc'(x-fc>], %p(x) = tit 2 e- fc ' 6 e i f c ( x - 6 ), and so -fc' (2 - t2) = ikt2, or t2 = 2ik'/{k + ik'). 0 < x < b, a; > b,
Hence the transition probability is given by T = tit2 e-k'b
16fc2fc'2
2k,b
_ 16E(V0~E)
2k,b
1 1
~ (fc2 +
fc'2)2
v-02
(b) To estimate the rate of a-decay of 8 3 Bi 2 1 2 , we treat, in first approximation, the Coulomb potential experienced by the a-particle in the siTl nucleus as a rectangular potential barrier. As shown in Fig. 1.21, the width of the barrier ro can be taken to be r0= 2Ze2 E
1 f\0
2 ( 8 3 - 2 ) e 2 he 6 he MeV
1
= x x 6.58 x 1 0 - 2 2 x 3 x 10 10 6 137 = 3.9x 10~ 12 cm.
Basic Principles
and One-Dimensional
Motions
75
* r
The radius of the nucleus of Tl is R = 1 x 10" 1 3 x 2083 = 6 x 1 0 - 1 3 cm, corresponding to a Coulomb potential height of V = 2Ze2 R 39 MeV.
An a-particle, moving with speed v in the nucleus of Tl, makes ^ collisions per second with the walls. Hence the lifetime T of 83Bi212 is given by T i n 2R 1, or 2R -v\T\f Taking for the rectangular potential barrier a height VQ |(396)+6 = 22.5 MeV, b = r0-R 0.1c, we find 2k'b = 3 3 x l 0 ~ 1 3 cm (see Fig. 1.21), v = J ^ , = J ^ c ^
2y/2mc2(V0 - E) tic
2-y/2 x 940 x 16.5 x 33 x 10~ 13 b = - ! ^ r6.58 x 10- 2 2 x 3 x 10 10 -rr 7^^ : r^r^ = 59 . 22.5 2 59 16 x 6 x (22.5 - 6) x e
_ 2 x 6 x 10~ 13 3 x 109 = 5.4x 1 0 3 s .
76
Electromagnetism
1046 An electron with energy E = 1 eV is incident upon a rectangular barrier of potential energy Vo = 2 eV (see Fig. 1.22). About how wide must the barrier be so that the transmission probability is 10 - 3 ? (Wisconsin)
Fig. 1.22
Solution: The transmission probability is (Problem 1045) _

T ~
16E(V0 - E)
1
TTO2
Vo
exp
- ~
yj2m(V0 - E)
= 4 exp whence 2
ln
~ V2m(V0 - E) n
he y/2mc (VQ - E)
2
K^J 2
6.58 x I P " 1 6 x 3 x 101Q y/2 x 0.51 x 106
8.1 x 10~ 8 cm.
1047 Consider a one-dimensional square-well potential (see Fig 1.23) V(x) = 0 , V(x) = -V0, V(x) = 0 , x < 0, 0 < x < a, x > a,
Basic Principles
and One-Dimensional
Motions
77
where Vo is positive. If a particle with mass m is incident from the left with nonrelativistic kinetic energy E, what is its probability for transmission through the potential? For what values of E will this probability be unity? (Columbia) VM
Ii
-K)
Fig. 1.23
Solution: Let the wave function be tp(x) = eikx + Re~ik-\ ip(x) = Seikx, ik x ip{x) = Aeik'x + Be- ' , where kV2mE te' = y/2m(E + V0)
x<Q,
x > a, 0<x<a
The constants R, S, A, B are to be determined from the boundary conditions that ip(x) and ip'(x) are both continuous at x = 0 and x = a, which give
( 1 + R = A + B, k{l-R)
Aeik'a +
k'(A-B),
=
Be-ik'a
Seika,
L k'(Aeik'a Hence
- Be~ik'a)
= kSeika.
4kk'e-ika (k + k')2 e~ik'a - (k - k')2 eik'a
78
Electromagnetism
and the probability for transmission is P = ^ = |5|2

3i
4k2k'2 4(fcfc' cos k'a)2 + (k2 - k'2)2 sin2 (k'a) ' Resonance transmission occurs when k'a mr, i.e., when the kinetic energy E of the incident particle is E = n2n2h2/2ma2 - V0.
The probability for transmission, P, then becomes unity.
1048 Consider a one-dimensional square-well potential (see Fig. 1.24):
Vix)
n
* x
Fig. 1.24
V ( z ) = 00,
X < 0,
V(x) = -V0, V(x) = 0,
0<x<a, x>a.
(a) For E < 0, find the wave function of a particle bound in this potential. Write an equation which determines the allowed values of E. (b) Suppose a particle with energy E > 0 is incident upon this potential. Find the phase relation between the incident and the outgoing wave. (Columbia)
79
Solution: The Schrodinger equations for the different regions are 2m dx2 2m dx2 ip(x)=0, 0<x<a,
ip(x) = 0,
x > a.
(a) E < 0. (i) Consider first the case of VQ < E, for which the wave function is 0, ip(x) A sinh(fca;), I Be -k'x where k = 2m{-VQ-EJ
h?
k, =
x < 0, 0 < x < a, x > a, J2m{-E)

V h2
The continuity conditions of the wave function give Asmh(ka) = Be~k'a, -Bk'e-k'a
Ak cosh(ka) = and hence
k coth(fca) = k'. As coth x > 0 for x > 0, there is no solution for this case. (ii) For V0 > E, ik -> k, k = y/2m(Vo + E)/h, and the equation determining the energy becomes k cot (ka) = k'. The wave function is '0, ip (x) = < A sin (kx), I Be -k'x x <0, 0 < x < a, x > a.
From the continuity and normalization of the wave function we get

1/2
A =
p- sin (ka) + a ^
sin (2ka)
80

1/2 fc'a efc a sin (fca).
B =
jg7 sin 2 (ka) + a, j sin (2fca)
(b) E > 0. The wave function is '0, ip(x) = < yl sin (fca;), x < 0, 0 < a; < a
. ? sin (fc'a; + ip), x > a, where
As d\nip/dlnx
is continuous at a; = a, (ka) cot (fca) = (fc'a) cot (fc'a + <p),
whence if = arccot I cot (fca) I fc'a. For x > a, '

D
_ ib(x) = e ik'xiip YK
2i
f^_ Akfx+i<p
T3
2i
where
Wnc
(i)e-itv,v,
tpwA{x)<Keik'x+i^.
Hence the phase shift of the outgoing wave in relation to the incident wave is
5 = 2<f = 2 arccot I cot (ka) 1 k'a
1049 Consider a one-dimensional system with potential energy (see Fig. 1.25) V(x) = V0, V(x) = 0, x>0, x < 0,
Basic Principles
and One-Dimensional
Motions
81
where Vo is a positive constant. If a beam of particles with energy E is incident from the left (i.e., from x = - c o ) , what fraction of the beam is transmitted and what fraction reflected? Consider all possible values of E. (Columbia) V[x)
h
Vo
* x
Fig. 1.25
Solution: For x < 0, the Schrodinger equation is d2 ^ whose solution has the form ip(x) = eikx + where k =
For x > 0, the equation is
, +
2mE ^
= 0,
re~ikx,
2mE h2
d2
2m(E-V0)
(i) If E < Vo, write the above as d2 dx*11' 2m(V0-E) h2 ^

= 0
"
As ip{x) must be finite for x ) oo, the solution has the form 4>{x)=te-k'x,
82
Electromagnetism
where
k,_ l2m(V0-E)
H2
The continuity conditions then give l + r = t, ik -^ ikr = tk', whence r = (k' + ik)/(ik - k') = (1 - ik'/k)/{\ + ik'/k). Therefore the fraction reflected is R = jref/jinc = \r\2 = 1, the fraction transmitted is T=1-R = 0. (ii) E > V0. For x > 0, we have d2 dx2 where 2m(E-V0) h2 ip(x) = 0 .
(,)-*". = s j ^ ^
Noting that there are only outgoing waves for x oo, we have 1 +r = t, > ik ikr = ik't, and thus r (k' k)/(k' + k). Hence the fraction reflected is R = [(k' - k)/(k' + k)]2, the fraction transmitted is T = 1 - R = 4kk'/(k + k')2.
1050 A particle of mass m and momentum p is incident from the left on the potential step shown in Fig. 1.26. Calculate the probability that the particle is scattered backward by the potential if (a) p2/2m < V0, (b) p2/2m > V0. (Columbia)
83
V(x)
^0
x
Fig. 1.26
-> X
Solution: The Schrodinger equations are ( d2 2mE\ , , . (dx->+1^)^X)=0 d2 2m . T r , dxl + J ? ^ - ^ (a) If E < V0, we have
eifc(x-xo) + re-ik(x-x0)^ x < a.0j
,
fOT
x < XQ,
ip(x) = 0 for x > XQ.
ip(x) = where 2m fi2 ' *' = 2m(V0-E) h2
the condition that tp(x) is finite for x oo having been made use of. > The continuity conditions give 1 + r = t, ik ikr = k't, whence r = (k' + ik)/(ik k'). The probability of reflection is R = jr/ji = | r | 2 = 1. (b) If E > V0. We have
eik(x-x0) + re-ik(x-x0)
X<XQJ
ip(x)
te ifc'(x-xo) )
X>X0,
84
Electromagnetism
where k = k' = 2mE 2m(E-V0) h2
noting that there is only outgoing wave for x > XQ. The continuity conditions give 1 + r = t, ik ikr = ik't, and hence r = (k k')/(k + k'). The /\1 probability of reflection is then R=\r\2 = [(k- k')/(k + k')] 2
1051 Find the reflection and transmission coefficients for the one-dimensional potential step shown in Fig. 1.27 if the particles are incident from the right. (Wisconsin)
Fig. 1.27
Solution: As the particles are incident from the right we must have E > VQ. And there are both incident and reflected waves in the region x > 0. The Schrodinger equation for x > 0, i>"{x) + k\ r/>(x) = 0, where fci = y/2m(E Vo)/h, has solutions of the form tp = exp (ik\x) + R exp (ikix).
Basic Principles
and One-Dimensional
Motions
85
There are only transmitted waves in the region x < 0, where the Schrodinger equation is 1>"(x) + k%rl>(x) = 0 with &2 = \/2mE/h, and has the solution ip(x) = S exp (ik2x). Using the continuity conditions of the wave function at x = 0, we get 1 + R = S. From the continuity of the first derivative of the wave function, we get ki(l R) = &2 S. Hence R = (&i k2)/(k\ + k2), giving the reflection coefficient V2 k\ - k2 lô \R\ k\ + k2 (VE + y/E^Vo)4 ' and the transmission coefficient \S\2 = l-\R\2
V = l-= f (VE+^/E^Voj*
1052 Consider, quantum mechanically, a stream of particles of mass m, each moving in the positive x direction with kinetic energy E toward a potential jump located at x = 0. The potential is zero for x < 0 and 3.E/4 for x > 0. What fraction of the particles are reflected at x = 0? (Buffalo) Solution: The Schrodinger equations are ty" + k2tp = 0
2
for
x < 0, x > 0,
-0" + (k/2) i) = 0 for
where k \/2mE/h. As for x < 0 there will also be reflected waves, the solutions are of the form ip = exp (ikx) + r exp (ikx), tp = t exp (ikx/2), x < 0, x > 0.
86
Electromagnetism
Prom the continuity conditions of the wave function at x = 0, we obtain 1 + r = t, fc(l r) = kt/2, and hence r = 1/3. Thus one-ninth of the particles are reflected at x = 0.
1053 Consider a particle beam approximated by a plane wave directed along the a;-axis from the left and incident upon a potential V(x) = jS(x), 7 > 0, S(x) is the Dirac delta function. (a) Give the form of the wave function for x < 0. (b) Give the form of the wave function for x > 0. (c) Give the conditions on the wave function at the boundary between the regions. (d) Calculate the probability of transmission. (Berkeley) Solution: (a) For x < 0, there are incident waves of the form exp (ikx) and reflected waves of the form R exp (ikx). Thus tp(x) = exp (ikx) + R exp (ikx), x < 0. (b) For x > 0, there only exist transmitted waves of the form S exp (ikx). Thus ip(x) = S exp (ikx), x > 0. (c) The Schrodinger equation is h2 d2 ~2m dx2 ^ ^ + 7 < ^ ^ ^ and its solutions satisfy (Problem 1020)
= E x
^)
V/(o+)-v'(o-) = - V ( o ) . ^
As the wave function is continuous at x = 0, I/J(0+) = X/J(0~). (d) From (a), (b) and (c) we have 1+R = S, ikS-ik(l-R) = 2mjS/h2, giving 5 = 1/(1 + irwy/fftk). Hence the transmission coefficient is
where E =
K2k2/2m.
Basic Principles
and One-Dimensional
Motions
87
1054 Consider a one-dimensional problem of a particle of mass upon a potential of a shape shown in Fig. 1.28. Assume that E at x -> oo is greater than Vb, where Vb is the asymptotic potential as x oo. > Show that the sum of reflected and transmitted intensities the incident intensity is one. m incident the energy value of the divided by {Princeton)
Fig. 1.28
Solution: As E > Vb we may assume the asymptotic forms ip elkx + re~%kx > ip -> tet0x for for x oo, > x -> +00,
where r, t, k, (3 are constants. The incident intensity is defined as the number of particles incident per unit time: / = hk/m. Similarly, the reflected and transmitted intensities are respectively R=\r \2hk/m, T=\t \2h/3/m.
Multiplying the Schrodinger equation by ijj*,
f v V + ip*Vip = Eip*i>, 2m
and the conjugate Schrodinger equation by tp, h2
- VVV* + rp*ViP = Eip*ip,
88
Electromagnetism
and taking the difference of the two equations, we have
V>*v2v> - v>v2v>* = v {ip*Vip - V'VV'*) = o.

This means that / (x) = ip*dtp/dx ipdip*/dx is a constant. Then equating / (+00) and / (00), we find k(l-\r\2) Multiplying both sides by ^ gives I = R + T. = 0\t\2.
1055 A Schrodinger equation in one dimension reads (-d2/dx2 (h= l , m = 1/2). (a) Show that exp(ifcx)(tanh x 4- const) is a solution for a particular value of the constant. Calculate the S-matrix (transmission and reflection coefficients) for this problem. (b) The wave function sech x happens to satisfy the Schrodinger equation. Calculate the energy of the corresponding bound state and give a simple argument that it must be the ground state of the potential. (c) Outline how you might have proceeded to estimate the ground-state energy if you did not know the wave function. (Buffalo) Solution: (a) Letting the constant in the given solution tp be K and substituting ijj in the Schrodinger equation, we obtain fc2 (tanh x + K) - 2(ik + K) sech2 x = s (tanh x + K). - 2 sech2 x) ip = ej>
89
This equation is satisfied if we set K = -ik and e k2. Hence i>(x) = eikx (tanh x - ik) is a solution of the equation and the corresponding energy is k2. Then as tanh x > 1 for x oo and tanh (x) = tanh x we have > ip = (1 ik) elkx 0 = (1 + ik)e
lkx
as as
x > oo, x > oo.
Since V(x) < 0,e > 0, the transmission coefficient is T = 1 and the reflection coefficient is i? = 0 as the particle travels through V(x). So the S-matrix is / 1 -(l-tfc)/(l+tk)\
\-(l-ik)/(l
+ ik)
J"
(b) Letting ip =sech x in the Schrodinger equation we have ip = etp. Hence e = 1. Because sech a; is a non-node bound state in the whole coordinate space, it must be the ground state. (c) We might proceed by assuming a non-node bound even function with a parameter and obtain an approximate value of the ground state energy by the variational method.
1056 A monoenergetic parallel beam of nonrelativistic neutrons of energy E is incident onto the plane surface of a plate of matter of thickness t. In the matter, the neutrons move in a uniform attractive potential V. The incident beam makes an angle 6 with respect to the normal to the plane surface as shown in Fig. 1.29. (a) What fraction of the incident beam is reflected if t is infinite? (b) What fraction of the incident beam is reflected if V is repulsive and V = El Consider t finite. (CUS)
90
Electromagnetism
Fig. 1.29
Solution: (a) Let ko be the wave number of an incident neutron, given by fcjj ^f-- For x < 0, the wave function is 01 (a; v) = eikx
cos e+ikoy sin 6 =
r> ikox cos 6+ikoy sin 8
With t infinite and the potential negative, for x > 0 the Schrodinger equation is 2m Assuming a solution iP2(x,y) = Teik*x+ikvy V)/h2.
and substituting it the equation, we obtain k\ + k2 = 2m(E + The boundary conditions at x = 0 i M 0 , y ) = V2(0,y), dipi dx then give
gifcoy sin 8 , Dgifcoy sin 8 _ rpeikyy
sin
x=0
9^2 dx
x=0
ik0 cos 6eikoy
sin
* - flifco cos 9eikoy
* = Tikx eiky .,
91
As the potential does not vary with y, ky = ko sin 0 and the above become 1 + R = T, k0 (1 - R) cos 9 = kxT, which give R = (ko cos 0kx)/(ko cos Q+kx). The probability of reflection is then P = \ R | 2 = (k0 cos 0 - kx)2/(k0 cos 0 + kx)2, with kl = 2m(E + V)/h2 - k2 sin2 0, kl = 2mE/h2. (b) For x < 0 the wave function has the same form as that in (a). For 0 < x <t, E V = 0, and the Schrodinger equation is -{h2/2m) V 2 V = 0.
As the potential is uniform in y we assume ip = exp(ik'y) exp(fcx), where k' = ko sin 0. Substitution gives k'2 + k2 = 0, or k = k'. Hence the wave function for 0 < x < t is il>2(x,y) = (aek'x +be~k'x) eik'v.
Writing ip(x, y) = 4>(x) elk v, we have for the three regions x<0, 0 < x < t, x>t, with kx = J The boundary conditions cos 0, fc' = \j2mE/h2 sin 0. &(x) = eik*x + re~ik*x', <p2{x) = aek'x + be~k'x, 03 (a;) =ceikxX,
0i(O) = fc(O), fc (*) = &(*)>

x=0
d02 dx
d<p2
a;=0
d(f>3
x=t
dx
dx
give
1 + r = a + 6, ifcx (1 - r) = fc' (a - b), c exp (ikxt) = a exp (fc't) + 6 exp (k't), ikxc exp(ikxt) = k'a exp (k't) k'b exp (-k't),
92
Electromagnetism
whose solution is ikx (a/b + 1

i/b-1
-1
ikx
k'
-Ik''t
fa/b+1
\a/b-l
a/b = [(k' + ikx)/(k' - ikx)} e

Hence
e2fc'* - 1
1 - /32e2fc't
with
and the fraction of neutrons reflected is |i?l 2 = | r | : Alternative Solution: The solution can also be obtained by superposition of infinite amplitudes, similar to the case of a Fabry-Perot interferometer in optics (see Fig. 1.30).
e2k't e2k't + + &-2k't 2
e -2fc't _
C o s A6 '
>x
Fig. 1.30 We need only consider the x-component of the waves. Let T12, R12 denote the coefficients of amplitude transmission and reflection as a wave goes from medium 1 to medium 2, respectively. Let T21, R21 denote coefficients of amplitude transmission and reflection from medium 2 to
Basic Principles
and One-Dimensional
Motions
93
medium 1. Take the amplitude of incident wave as 1. Then the amplitude of the wave that is transmitted to medium 2 is the sum T = T12e~s T21 + T12e~s -R2ie-2SR21 + T12e~
5 6
T21
(R21e- R21)
2 21 l
25
T 21 + + (R&-25)2 + }
= T12e- T21 [1 + R = T12T21e~s
e'
2S
1-R221e-V
In the above exp(5) is the attenuation coefficient of a wave in medium 2, where 6 = k't with k' = s/2mE(l-cos26)/h = V2mE sin 0/h.
From (a) we have the coefficients of transmission and reflection -R12 = (fax - k2x)/(klx T 12 = 1 + R12 = 2k1/(klx As k\x = v2mE cos 0/h, sin 0/h, + k2x), + k2x).
k2x = ik' = i v2mE we find T12 = 2klx/(klx T21 = 2k2x/(klx + k2x) =
2 COS0 . Q
= 2 cos 6e-ie, 0e-ie,

e "",
cos 0 + i sin 0 2 sin 0 + k2x) = , . = 2i sin cos 0 + i sin 9

+ k2x)
r. /, T x /,, -R21 = (^2x - klx)/(klx and hence
N i sin = . .
0 cos 0 = i sin 0 + cos #
_,,-,
r=
rr>
rri
**
-ii2-i2ie
1 - iJlje- 2 *
2ld
4i cos 0 sin Oe
I _
fc
e-4i0e-2fc't
2isin26'e- 2 i e e- f c ' t
J __ e - 4 i 0 e - 2 f c ' t '
94
on
Electromagnetism
The transmissivity is therefore ,,o

1 1
4 sin226>e-2fc'*
( i _ e -2fc't
cos
4 5,) 2
( e -2fc't
sin
4 6 i)2
_ 4 sin2 26e-2k>t ~ 1 + e-4fc't _ 2 e - 2 f e '' cos 40 ~ e2fc't + ~ and the reflectivity is 1-|T|2 = 1 4 sin2 20
e2k't + e -2fc't e -2fc't
4 sin2 20 e-2fc't _ 2 cos 40 '
_ _
2 2
COS 4 0
e 2fc't + e 2fc't +
e -2fc't
COS 4 0 '
where k' = \/2mE
sin 6/h.
1057 Find the wave function for a particle moving in one dimension in a constant imaginary potential iV where V <S E. Calculate the probability current and show that an imaginary potential represents absorption of particles. Find an expression for the absorption coefficient in terms of V. (Wisconsin) Solution: The Schrodinger equation is ihdip/dt = (p2/2m - iV) ip. Supposing V = exp(iEt/h) iV/E). As V < E, exp(ikx), we have k2 = (2mE/h2)(l +
2mE
. V
Basic Principles
and One-Dimensional
Motions
95
and hence ip(x, t) = exp 2mE h2 V 2E'
exp I i
2mE
~W
exp
iEt
[-T
where ip+ and V- refer to the exponentially attenuated right- and lefttraveling waves respectively. The probability current is
j =Re ( V * 2. A = Re U* i/> ) = Re >

2mE h2 exp 2mE V h2 E'
2mE V h2 E
q=
These are the exponentially attenuated currents in the respective directions. The absorption coefficient is then 1 dj_ j dx dh\j dx 2mE V h2 ~E
The imaginary potential iV is responsible for the absorption of the particle, since the exponent in j would be imaginary. Hence there would be no absorption if V were real.
1058 Let the solution to the one-dimensional free-particle time-dependent Schrodinger equation of definite wavelength A b e ^ i ( i , t ) as described by some observer O in a frame with coordinates (x, t). Now consider the same particle as described by wave function ip1{x', t') according to observer O' with coordinates (x', t') related to (x, t) by the Galilean transformation x' = x vt, t' = t. (a) Do tp(x, t), tp' (x', t') describe waves of the same wavelength? (b) What is the relationship between tp(x, t) and ip' (x', t') if both satisfy the Schrodinger equation in their respective coordinates? {Berkeley)
96
on
Electromagnetism
Solution: (a) The one-dimensional time-dependent Schrodinger equation for a free particle ihdtip(x, t) = {-h2/2m)dlxp{x, t) has a solution corresponding to a definite wavelength A ip\(x, t) = exp [i (kx ut)] with A = 27r/fc = 2irh/p, u = Kk2/2m.
As the particle momentum p is different in the two reference frames, the wavelength A is also different. (b) Applying the Galilean transformation and making use of the Schrodinger equation in the (x', t') frame we find ihdtip' (x', t') = ihdtijj' (x - vt, t) = = ih[d'ti>'(x',t')-vd'xrP'(x',t')] -~d'x2iP'(x',t')-ihvd'xi>'(x',t') t) - ihvdxip'(x - vt, t). (1)
h2 = - d%ip'(x-vt, Considering
ihdt[e^kx-ut)xl)'{x',t')}
and *Ld2[Skx-t)^{x^tl)] = - j - e^kx-^ (-fc2V' + 2ikdxiP' + d2xi)>')

2
Basic Principles
and One-Dimensional
Motions
97
making use of Eq. (1) and the definitions of k and u>, we see that -d2x [eimvx'h e-imv'tw^
eim
(x _
vt>
t)]
= ihdt [eimvx/h
" a * / 2 V [x - vt, t)]. Hence,
This is just the Schrodinger equation that ip(x, t) satisfies. accurate to a phase factor, we have the relation tp(x, t) = ip' (x vt, t) exp mv mvx
1059 A particle of mass m bound in a one-dimensional harmonic oscillator potential of frequency LJ and in the ground state is subjected to an impulsive force p5(t). Find the probability it remains in its ground state. (Wisconsin) Solution: The particle receives an instantaneous momentum p at t = 0 and its velocity changes to p/m instantaneously. The duration of the impulse is, however, too short for the wave function to change. Hence, in the view of a frame K' moving with the particle, the latter is still in the ground state of the harmonic oscillator %J)Q(X'). But in the view of a stationary frame K, it is in the state tpo{x') exp(ipx/h). We may reasonably treat the position of the particle as constant during the process, so that at the end of the impulse the coordinate of the particle is the same for both K and K'. Hence the initial wave function in K is 00 = Vo (x) exp " (-ipx/h).
Thus, the probability that the particle remains in its ground state after the impulse is
98
Problems and Solutions on 2
Electromagnetism
Po =
(-00 | exp [j-i j
| ip0)
-i r
exp
-"^-^ds a ip V 2a 2 ft/
- 2a27i2
exp
-or [x +
dx
= exp \2mojhJ where a = y/mcj/h.
1060 An idealized ping pong ball of mass m is bouncing in its ground state on a recoilless table in a one-dimensional world with only a vertical direction. (a) Prove that the energy depends on m, g, h according to: e = Kmg (m2g/h2)a and determine a. (b) By a variational calculation estimate the constant K and evaluate e for m = 1 gram in ergs. {Princeton) Solution: (a) By the method of dimensional analysis, if we have
=
U+2<*[5]l+ [h]2 a
or
[L?
[T]
2
\L]l - 3 a [T]2
then a = j . Thus, provided a = ^, the expression gives the energy of the ball. (b) Take the x coordinate in the vertical up direction with origin at the table. The Hamiltonian is H
2m
PL + mgx
2m dx2
+ mgx,
Basic Principles
and One-Dimensional
Motions
99
taking the table surface as the reference point of gravitational potential. Try a ground state wave function of the form ijj = x exp (Xx2/2), where A is to be determined. Consider
(H)
- ^;Hf
_ J0 xe-^l* (-
2
+Tngs)
Ax2
ae
-^/dx
J0 x e~ 3fi2 , 4m 2mp v/Âi/2 '
dx
To minimize (if), take ^ = 0 and obtain A = (j^j* The ground state energy is then (H) giving
K = 3{3/4TT)1/3.
=3(3/4*)1/3mg(m2g/ti1)-1'3.
Numerically
e=
3k2 3m ( Anm2g V \ J
3 x 980 x
, 3 x 1.0542 x l ( r 5 4 V 4?r x 980 /
= 1.9 x 10 _ 1 6 erg.
1061 The following theorem concerns the energy eigenvalues En (E\ < E^ < E$ < . . . ) of the Schrodinger equation in one dimension: Theorem: If the potential Vi (x) gives the eigenvalues E\n and the potential Vi (x) gives the eigenvalues Ein and V\ (x) < V2 (x) for all x, then Eln < E2n(a) Prove this theorem.
100
Electromagnetism
Hint: Consider a potential V(X, x), where V(0, x) = V\ (x) and V(l, x) V2(x) and OV/dX > 0 (for all x), and calculate dEn/dX. (b) Now consider the potential (Fig. 1.31) U(x) U(x) kx2/2, ka2/2, < a, > a.
Uix)
- *
Fig. 1.31
We want to determine the number of bound states that this potential can hold. Assume this number N is 3> 1. It may be helpful to draw a qualitative picture of the wave function for the highest bound state. Choose a solvable comparison potential and use the theorem above to determine either a rigorous upper bound to N or a rigorous lower bound to N. (Both can be done but you are asked for only one.) (Berkeley) Solution: (a) Define V(X, x) = XV2(x) + (1 - X)V1(x). Obviously V(0, x) = Vi (x), V(l, x) = V2 (x), dV/dX = V2 (x) - Vi (x) > 0. The Hamiltonian is then H(X)=p2/2m and the eigenequation is
H(X)\n,X)=En(X)\n,X),
V(X,x),
where En (A) = (n, A | H(X) | n, A). As
Basic Principles
and One-Dimensional
Motions
101
dEn(X)/d\
= [(n, A|tf(A)|n, A)] = (nX | dV(X)/dX \ nX)
= J ^
\*n(x,
X)\2dx>0,
we have E\n = En(0) < En(l) = E-m, and the theorem is proved. Note that we have used (nX \ nX) = 1. (b) Let V(x) = kx2/2. Then V(x) > U{x). If En is an energy level for the potential U(x), then En < (n + 1/2) fiw, where w = y/k/m. For a bound state, En < ka2/2. Solving (N + 1/2) Hw < ka2/2, we find N < ~
muja 1
2h
muja? 2h \ '
where [A] indicates the maximum integer that is less than A. We now choose for V(x) a square well of finite depth, V[x) = ka2/2, V(x)=0, | x | > a, |x|<a.
The number of bound states of U(x) is less than that of V(x), which for the latter is [2muja2/nh] + 1. We can take the upper bound to the number of bound states of U(x) as [2mu>a2/irH] as for ./V 3> 1 the term 1 can be neglected. Taken together, we get that the number of bound states is between [muta2/2h] and [2mLja2/Trh].
1062 For electronic states in a one-dimensional system, a simple model Hamiltonian is

N n=l N n=l
H = 52^|n><n| + ^W{|n)<n+lH-|n+l)<n|}, where \n) are an orthonormal basis, (n | n') = 5nn>; EQ and W are parameters. Assume periodic boundary conditions so that | N + j) = \j). Calculate the energy levels and wave functions. (Wisconsin)
102
on
Electromagnetism
Solution: From the fact that | n) form a complete set of orthonormal functions, where n = 1, 2, 3, . . . , iV, and
N N
E
H = J2
n=l
o I n) (n | + J2 W{ \ n) (n + 11 + | n + 1) (n \ } ,
n=l
or H = E0 + W(A + A+), with
A=Y,\n)(n + l\,A+ = Y,\n

n=l n=l
+ 1
)^\'
and A\n) = | n - l ) , A + | n ) = | n + l ) , AA+ = A+A = 1, or A+ = A~x,
we know that H,A and A+ have the same eigenvectors. Hence we only need to find the eigenvectors and eigenvalues of the operator A+ to solve the problem. As
Aklk = {k'\A\k) =(Sfc', fe _i,
We have /0 0 0 1 0 0\ 0 1 - 0 0 0 1 0 1 0/NxN
A =
. . . . M 0 0
Motions
103
and so
-XI
/-A 0 0
1 -A 0
0 1 -A
1 -A
V 1
i.e.,
= (-1)N
-A/
NxN
det {A - XI) = (-X)N
+ (-l)N+1
(XN - 1) = 0,
giving Xj = e*', 0j = ^ j , j = 0, 1, 2, . . . ,N - 1. If |i5) are the same eigenvectors of the operators A and A+, i.e. A\Ej) then H\Ej) = E0 + W [Xj + = Xj\Ej), A+\Ej) = -\Ej),
AJ .
\Ej)
= {E0 + 2W cos 6j)

Hence the eigenvalues of H are
O-TT
\Ej).
Ej = E0 + 2W cos 6j;,
with
Oj = j ,
(j = 0, 1, 2, . . . , N - 1).
The corresponding eigenfunctions can be obtained from the matrix equations (A-Xj)\Ej)=0. Thus / \Ei)
l
ei2e* y/N {eHit-ito/
104
Electromagnetism
or
\Ej) =
*
^=JZei{n~l)9i\n)n=l
1063 Give a brief discussion of why there are energy bands in a crystalline solid. Use the ideas of quantum mechanics but do not attempt to carry out any complicated calculations. You should assume that anyone reading your discussion understands quantum mechanics but does not understand anything about the theory of solids. (Wisconsin) Solution: A crystal may be regarded as an infinite, periodic array of potential wells, such as the lattice structure given in Problem 1065. Bloch's theorem states that the solution to the Schrodinger equation then has the form u(x)ex.p(iKx), where K is a constant and u(x) is periodic with the periodicity of the lattice. The continuity conditions of u(x) and du(x)/dx at the well boundaries limit the energy of the propagating particle to certain ranges of values, i.e., energy bands. An example is given in detail in Problem 1065.
1064 A particle of mass m moves in one dimension in a periodic potential of infinite extent. The potential is zero at most places, but in narrow regions of width b separated by spaces of length a (b <C a) the potential is Vo, where Vo is a large positive potential. [One may think of the potential as a sum of Dirac delta functions:
oo
V(x)= Y,
noo
V0b5{x na).
Alternatively one can arrive at the same answer in a somewhat more messy way by treating the intervals as finite and then going to the limit. ] (a) What are the appropriate boundary conditions to apply to the wave function, and why?
Basic Principles
and One-Dimensional
Motions
105
(b) Let the lowest energy of a wave that can propagate through this potential be EQ = T?k\j2m (this defines fco). Write down a transcendental equation (not a differential equation) that can be solved to givefcoand thus E0. (c) Write down the wave function at energy EQ valid in the region 0 < x < a (For uniformity, let us choose normalization and phase such that tp(x 0) = 1). What happens to the wave function between x = a and x = a 4- 6? (d) Show that there are ranges of values of E, greater than Eo, for which there is no eigenfunction. Find (exactly) the energy at which the first such gap begins. (Berkeley) Solution: (a) The Schrodinger equation is
ft2 d2
2m dx2 +
V^
T^X/
y .
Vobo(x na)
ip(x) = Etl>(x).
Integrating it from x = a e to x = a + and letting e 0, we get ip' (a + ) - V' (a~) = 2Jty(a), where SI = mVob/h2. This and the other boundary condition
V>(a+)-V(a") =0
apply to the wave function at x = na, where n = - c o , . . . , 1, 0, 1, . . . ,
+ 00.
(b) For x 7^ na, there are two fundamental solutions to the Schrodinger equation: ui(x) = eihx, u2(x) = e~ikx, the corresponding energy being E = Let V (i) = Aeikx + Be-ikx, > 0 < x < a. h2k2/2m.
106
on
Electromagnetism
According to Bloch's Theorem, in the region a < x < 2a ip(x) = eiKa [Aeik{x~a) + Be-ik<-x~a)],
where K is the Bloch wave number. The boundary conditions give

eiKa
(A
+ B)=
Aeika
Be~ika^
ikeiKa (A-B)
= ik (Aeika + 2n{Aeika
Be~ika) + Be~ika).
For nonzero solutions of A and B we require

piKa pika piKa p ika
ikelKa or
- {ik + 2ft) elka
-ikelKa
0 + (ik - 20.) e ika
= o,
ft cos ka + sin ka = cos Ka, k which determines the Bloch wave number K. Consequently, the allowed values of k are limited to the range given by ft cos ka + sin ka < 1, k or ft cos ka + sin ka ) < 1.
fco is the minimum of k that satisfy this inequality, (c) For E = Eo, ip(x)=Aelk0X + Be ikox 0 < x < a,
/2mEo where fco = y- h Normalization ip (x = 0) = 1 gives
ip (x) = 2i A sin k0x + e~ikoX, The boundary conditions at x = a give eiKa =2iAsmk0a +
0 < x < a.
e-ikoa,
Basic Principles
and One-Dimensional
Motions
107
2iA = {eiKa - e- i f e o )/sin So Mx) = (eiKa - e-ikoa) $

s i n KQCL
k0a. 0 < x < a.
+ e~ikx,
For x [o, a + b], the wave function has the form exp (fci, x), where fei = y/2m(V0 - Eo)/tL (d) For ka mr + 6, where 5 is a small positive number, we have cos ka + sin ka k
n
cos (mr + 5) + sin (mr + 5)
i
5
< l.
When S is quite small, the left side 1 + fW/A; > 1. Therefore in a certain region of k > nn/a, there is no eigenfunction. On the other hand, ka = mr corresponds to eigenvalues. So the energy at which the first energy gap begins satisfies the relation ka = ir, or E = Tr2h2/2ma2.
1065 We wish to study particle-wave propagation in a one-dimensional periodic potential constructed by iterating a "single-potential" V(x) at intervals of length I. V(x) vanishes for | x \ > 1/2 and is symmetric in x (i.e., V{x) = V{x)). The scattering properties of V(x) can be summarized as follows: If a wave is incident from the left, ip+(x) = exp(ifcx) for x < 1/2, it produces a transmitted wave ip+(x) = exp (ikx) for x > 1/2 and a reflected wave ip-{x) = exp(ifcx) for x < 1/2. Transmitted and reflected coefficients are given by
W _
V+(-|^|)"2
'
^ 2 '
>
R(x) - ^ - H ^ l ) ~ I e f c M \e2iSe K{X} [e -TP+{-\x\)~2e
o2i60-\
Z 2'
108
on
Electromagnetism
and Se and SQ are the phase shifts due to the potential V(x). Take these results as given. Do not derive them. Now consider an infinite periodic potential V ^ z ) constructed by iterating the potential V(x) with centers separated by a distance I (Fig. 1.32). Call the points at which Voo(a;) = 0 "interpotential points". We shall attempt to construct waves propagating in the potential Vô(a;) as superpositions of left- and right-moving waves <j>+ and <j>-.
Vnlx)
1
-1/2
''2
Fig. 1.32
(a) Write recursion relations which relate the amplitudes of the rightand left-moving waves at the nth interpotential point, <p, to the amplitudes at the (n - l)th and (n + l)th interpotential points, 0 _ 1 and <+1. (b) Obtain a recursion relation for <_ or <fr+ alone by eliminating the other from part (a). (c) Obtain an expression for the ratio of amplitudes of <> to 4>- at /+ successive interpotential points. (d) Find the condition on k, 5e and <o such that traveling waves are 5 allowed. (e) Use this result to explain why it is "normal" for conduction by electrons in metals to be allowed only for bands of values of energy. (MIT) Solution: For the wave incident from the left, the potential being V(x), let i>t+=t<]>+eikx,
Vv- = r<f>-e~ikx,
t=(ei2S<+ei25);
r = ]- {ei25< - e i2 *).
Basic Principles
and One-Dimensional
Motions
109
For the wave incident from the right, let the transmission and reflection coefficients be t' and r' respectively. It can be shown that t' = t, r' = r*t/t*. In the periodic potential, the transmission and reflection coefficients at adjacent interpotential points have relations tn = tn-i and Tn = rn-\ exp (i2kl). So the transmission coefficient can be denoted by a single notation t. (a) The waves at adjacent interpotential points are as shown in Fig. 1.33. Obviously, only the reflection term of 0 and the transmission term of <A"-1 contribute to $:
4>l=r'n.14>n_+t<f>l-1.
Similarly, ^=r Thus we have <j>l=ei2klr'n<f>n-+t<f>l'\
4>1 = rn<S>n+ + t(j>n-+l
n
(1)
^+i'C
+ 1
(2)
(3)
(4)
Fig. 1.33
(b) With n replaced by n + 1, Eq. (1) gives
r++1 = r'nr-+1+tcf>n+.
Equations (3), (4), (5) then give
(5)
1 + t2ei2kl - rnr'nei2kl
'
110
on
Electromagnetism
Let r0 = r. Then rn = r exp(i2nM). Assume r'n r*ntjt*. Then r'n = r' exp (i2nkl). Hence
*^+
1 + f2 ei2H _
rri
ei2kl
'
Similarly, 1 + t 2 e i2W rr' ei2kl '
(c) As the period of the potential is I, if ip(x) is the wave function in the region [xn-\, xn], then ip(x I) exp (iS) is the wave function in the region [xn, xn+i]. Thus
( (pl+1 = e ^ - f c ' ) ^ ,
(7) Let c n = 0 /</>" From (4) and (5) we obtain respectively 1 = rncn +1 ,
.n+l ^n+1
<t>-+1
(8)
^ = ^ ^ r - + ^ Using (7), (9) can be written as
(9)
An 0!!
J.n
"r C n j
or, using (8),

Cne yl fn^n)
=
^nU
^n^n)
~r C n ;
r nC 2 + (tV2*< - r < e i 2 W - 1) c n + r-^ 2 *' = 0. Solving for c n we have _ (1 + rr'ei2kl cn - <V 2fc ') VE , 2r n
Motions
111
where A = (tV 2 f c i - rnr'nei2kl = (t2ei2kl - rr'ei2kl - l)2 - l)2 4rnr'nei2kl 4rr'ei2kl.
(d) The necessary condition for a stable wave to exist in the infinite periodic field is
where Si is real and independent of n. If this were not so, when n oo one of 0" and 4>+ would be infinite. From (7), we see that 5\ = 5 kl Prom (6), we obtain
1
+ t 2 e i2fci
r r / e i2fci
= t[ei{kl~s)
+ei(-kl+s)].
Substituting r' = r*t/t* in the above equation and using rr* +tt* = 1, we obtain 2tt* cos S, teikl + t*e~ikl which means telM + t*e 2tt*
-ikl
<i.
or, using the definition of t, cos (2<5e + kl) + cos (20 + kl) <1, l+cos[2(Je-<50)]
i.e.. COS (i5e +
SQ
+ kl)
cos(Se - S0)
< 1.
In general, only some of the values of A satisfy the above inequality, ; i.e., only energy values in certain regions are allowed while the others are forbidden. Thus we obtain the band structure of energy levels. (e) In metals, the distribution of positive ions is regular and so the conduction electrons move in a periodic potential, (d) Shows that the
112
Electromagnetism
electron waves can only have certain k values, corresponding to bands of electron energies.
1066 You are given a real operator A satisfying the quadratic equation A2 - 3A + 2 = 0. This is the lowest-order equation that A obeys. (a) What are the eigenvalues of A! (b) What are the eigenstates of Al (c) Prove that A is an observable. (Buffalo) Solution: (a) As A satisfies a quadratic equation it can be represented by a 2x2 matrix. Its eigenvalues are the roots of the quadratic equation A 3A + 2 = 0, Ai = 1, A2 = 2. (b) A is represented by the matrix
HIDThe eigenvalue equation
G!) CM:)
then gives a 1, b = 0 for A = 1 and a = 0, b = 1 for A = 2. Hence the eigenstates of A are (J) and ( 1 ). (c) Since A = A+, A is Hermitian and hence an observable.
1067 If | -0q) is Y eigenstate of the electric charge operator Q corresponding to eigenvalue q, that is to say,
Q\^q)=<l\^q),
an
Basic Principles
and One-Dimensional
Motions
113
the "charge conjugation" operator C applied to \tpq) leads to an eigenstate | ip-q) of Q corresponding to eigenvalue q:
c | v , > = IV'-,>.
(a) Find the eigenvalues of the operator CQ + QC. (b) Can a state simultaneously be an eigenstate of C and of Q? (Chicago) Solution: (a) Let
Then
(CQ + QC) \1>q)=qC\1>q)+Q\1>-q)
= q\il>-q)-q\il>-q)
= 0.
Thus the eigenvalue of the operator CQ + QC is zero. (b) As C is the charge conjugation transformation, CQC~X = Q, or CQ + QC 0, i.e., C and Q do not commute (they anticommute) they cannot have common eigenstates. (Unless q = 0, in which case it is quite meaningless to introduce charge conjugation.)
1068 A quantum-mechanical system is known to possess only two energy eigenstates denoted |1) and |2). The system also includes three other observables (besides the energy), known as P, Q and R. The states |1) and 12) are normalized but they are not necessarily eigenstates of P, Q or R. Determine as many of the eigenvalues of P, Q and R as possible on the basis of the following sets of "experimental data". [Warning: one data set is unphysical.] ( a ) ( l | P | l ) = l/2, ( l | P 2 | l ) = l/4. (b) (11 Q 11> = 1/2, ( l | Q 2 | l ) = l / 6 . ( c ) < l | f l | l ) = l, ( l | f l 2 | l ) = 5/4, ( l | f l 3 | l ) = 7 / 4 . (MIT)
114
Electromagnetism
Solution: We are given three observables P, Q, R, which satisfy the Hermiticity (n\P\m) = (m\P\n)*, and that the mechanical system has a complete set of energy eigenstates |1) and |2). (a) The completeness of the two states and the "experimental data", give P | l ) = i | l ) + a|2), where a is a constant to be determined. The orthogonality of the eigenstates and the Hermiticity of P give ( l | P | 2 ) = i ( l | 2 ) + a 2 ( 2 | 2 ) = a*. So we have P|2) = a*|l)+/?|2), where (3 is to be determined. Then P 2 |1) = P ( P |1)) = i P |1) + aP |2) = (1/4 + a*a) |1) + (a/2 + a/3) |2>. As P 2 | l ) = \ according to experiment, a*a = 0 and hence a = 0. Therefore,
i.e., at least one of the eigenvalues of P is 1/2. (b) Let
Q|l) = i | l ) + 7 | 2 ) ,
where 7 is to be determined. By a similar procedure, we get 7*7 = 1/6 1/4 < 0. So this data set is unphysical and the eigenvalue of Q could not be determined. (c) As (1 | i l | l ) = 1, we can write P|l) = |l)+A|2),
Basic Principles
and One-Dimensional
Motions
115
where A is to be determined. Then <1|J2|2) = <1|2> + A*<2|2) = A ' , showing that fl|2) = A*|l)+7,|2>. where 7 is to be determined. Consider 7 /? 2 |1) = .R|1) + A#|2) = |1) + A|2)+AA*|1) + Ar?|2). Then ( l | f l 2 | l ) = l + AA* = ^ we have AA* = 4 and so A= and exp(i5),
R\2) =
\e-iS\\)+r1\2),
,R 2 |1> = 111> -H- I (1 -h 77) e i 5 |2>. It follows also i?3|l)=^|l) +l-(l+T1)ei5R\2)
= | | l ) + |e|2) + i(l+r7)|l> + \{\+ri)r)eii\2).
116
Electromagnetism
Experimentally (11 R3 |1) = \. Thus f + \ (1 + v) = f, giving 7 = 1. 7 Hence on the bases |1) and |2) the matrix of R is 1 ie To find the eigenvalues of R, solve 1 - A \ e~iS = 0, \ei& 1-A i.e., (1 A)2 7 = (1 A \) (1 A + h) = 0, and obtain the eigenvalues of i ? - 2> 2|e1
j
1069 For a charged particle in a magnetic field, find the commutation rules for the operators corresponding to the components of the velocity. (Berkeley) Solution: Suppose the magnetic field arises from a vector potential A. Then the velocity components of the particle are Hi = Pi/mHence
[Vi, Vj = =
qAi/mc.
p.-I
A-
P--Ac
m2c ihq
/
k=i
,jjkBk,
where ijk is the Levi-Civita density, use having been made of the correspondence rule pi 4 ^ .
Basic Principles
and One-Dimensional
Motions
117
1070 Using the coordinate-momentum commutation relation prove that y ^ (En - E0) \{n | x 10)2 = constant,
where En is the energy corresponding to the eigenstate |n). Obtain the value of the constant. The Hamiltonian has the form H = p2/2M + V {x). (Berkeley) Solution: As H = p2/2M we have [H, X) = JM b 2 . x] = and so [[H,x},x] Hence h2 (m\ [[H, x], x) 1-) = - ^ On the other hand, {m\\[H, x], x] \m) = (m\ Hx2 - 2xHx + x2H \m) = 2Em {m \x2 | m) - 2 (m | xHx \ m) = 2EmYJ\{m\x\n)\2
n n
+ V(x),
-ihp/M,
= -^[p,x}
-h2/M.
-2YiEn\(m\x\n)\2
2Y/(Em-En)\(m\x\n)\2.
In the above we have used
118
on
Electromagnetism
H \m) = Em
n
\m),
(m | x2 | m) = Y J (m | x | n) (n | x \ m)
= |<m|x|n)|2
n
{mIxifx
Im) = /J(7Ti|xH
n
\n){n\x\m) m)
2_. &n (m\x\n)(n\x\

n
=
n
Y,EnVjn\x\n)\*.
Equating the two results and setting m = 0, we obtain Y/(En-E0)\(n\x\0)\2

n
= h2/2M.
1071 (a) Given a Hermitian operator A with eigenvalues an and eigenfunctions un(x) [n = 1, 2, . . . , N; 0 < x < L], show that the operator exp(iA) is unitary. (b) Conversely, given the matrix Umn of a unitary operator, construct the matrix of a Hermitian operator in terms of Umn. (c) Given a second Hermitian operator B with eigenvalues bm and eigenfunctions vm(x), construct a representation of the unitary operator V that transforms the eigenvectors of B into those of A. (Chicago) Solution: (a) As A+ = A, A being Hermitian, {exp(iA)}+ = exp(-iA+) exp(-iA) = {exp (iA)}-1.
Hence exp(iA) is unitary. (b) Let

^ r a n Umn + ^nm
=
Umn ~r [U
)mn i
Basic Principles I.e.,
and One-Dimensional
Motions
119
c = u + u+.
As U++ = U,C+ = C. Therefore Cmn = Umn + U*m is the matrix representation of a Hermitian operator. (c) The eigenkets of a Hermitian operator form a complete and orthonormal set. Thus any | um) can be expanded in the complete set \vn): \um) - ^
fe fc
I W> {Vk I Um) = 5 3 \vk) Vkm , f e
w h i c h defines VfcTO,
Vkm = f Jo Similarly, \Vn) = 5 3

3
v*k (x) um (x) dx.
\uj)(uj\vn)
= 5 3 IUJ) (Vn IUJ)*

3
Hence
) = E E i>) Kk v = E K> <w.

3 k j
or
i.e.,
V+V
= l,
v+ = v~\
120
on
Electromagnetism
showing that V is unitary. Thus V is a unitary operator transforming the eigenvectors of B into those of A
1072 Consider a one-dimensional oscillator with the Hamiltonian H = p 2 /2m + mw2x2/2.
(a) Find the time dependence of the expectation values of the "initial position" and "initial momentum" operators XQ = x cos wt {p/muj) sin wt, Po = P cos wt + mux sin wt. (b) Do these operators commute with the Hamiltonian? (c) Do you find your results for (a) and (b) to be compatible? Discuss. (d) What are the motion equations of the operators in the Heisenberg picture? (e) Compute the commutator [po, XQ\. What is its significance for measurement theory? (Princeton) Solution: (a) Making use of the relation
7t-m[LH]+m'
we have i- = [(a;) cos wt, H] LJ{X) sin ut at in, 1 (p) ^ - sin wt, H cos wt ih mw m [x, H] cos wt w(x) sin wt ih [p, if] sin wt mw ih m (p) cos wt = 0 ,
Basic Principles
and One-Dimensional
Motions
121
d(po) dt
1 [(p) cos wt, H) {p)tJ sin ujt ih 4- \muj(x) sin u)t, H] + mJ2{x)
in
cos uit
1 = [p, H] cos ut - w(p) sin wi

in
+ mu [x, H] sin wt + muj (x) cos u>t = 0. -in Thus the expectation values of these operators are independent of time. (b) Consider \xo, H] = \x, H] cos ojt '- sin ujt mo)
ihp = cos ut + injjx sin ut, m [po, H\ = [p, H] cos uit + mcj [x, H] sin ujt = ihmu>2x cos wt + ihusp sin u>t. Thus the operators XQ,PQ do not commute with H. (c) The results of (a) and (b) are still compatible. For while the expressions for XQ and po contain t explicitly, their non-commutation with H does not exclude their being conserved. In fact dx0 dpo l
r
dx0
1 r
, , dpQ
showing that they are actually conserved. (d) In the Heisenberg picture, the motion equation of an operator is 1
dA/dt=rn[A,H] +
BA - .
Thus the motion equations of xo and po are respectively dxo/dt = 0, dpo/dt 0.
122
on
Electromagnetism
(e) Using the expressions for XQ and po, we have [Po , %o] = [p c o s ut + mux sin uit, = [p,
XQ] XQ]
cos uit + [x, muj
XQ]
mui sin uit cos uit
p, x cos uit
sin uit
sin uit mui sin uit. mui = [p, x] cos2 u)t \x, p] sin2 uit
x, x cos uit
= - [x, p] = -ih, as [a;, x] = [p, p] = 0, [a;, p] = ih. In general, if two observables A and B satisfy the equation [A, B] = ih, then their root-mean-square deviations AA, AB, when they are measured simultaneously, must satisfy the uncertainty principle AA AB > ^ . In the present case, the simultaneous measurements of position and momentum in the same direction must result in Ax Ap> The relation shows h
^jA~4s[A~p~l>h/2.
It is a relation between possible upper limits to the precision of the two quantities when we measure them simultaneously.

Quantum Physics

Caricato da

Informazioni sul documento

Descrizione originale:

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Quantum Physics

Caricato da

Copyright:

Formati disponibili

1.

Problems and Solutions

Thus there is no diffraction in any direction.

(c) = S* = -53 A "

Problems and Solutions on

Problems and Solutions

Problems and Solutions

vi = dydpi + d2/dPi + d*/dpi.

Problems and Solutions

and the Schrodinger equation becomes, in coordinate space, A-V2 ,\ , . .d<p(r,t)

Hence the potential is V(r)=ar2, and the force is F ( r ) = - V V ( r ) = - - -^- V(r) = - 2 o r .

1009 Consider a one-dimensional bound particle, (a) Show that d_ dt /

Taking ip* x (1) - ip x (2) we obtain

ih ^ (r ip) = - | - V (V*W - < W )

Problems and Solutions on

iJVO*;. to) = V(a;, to),

Normalization J_a ip*ipdx 1 requires that =

Hence V>(z, <) = 5 3 o-nipnix) exp ( - i - r 1 * ] , with

Problems and Solutions on

exp (x2/a2 ipx/h) dx (-a2p2/4h2).

(b) The Schrodinger equation in momentum space for a free particle,

W > *) = ~^z 2ft

Basic Principles and One-Dimensional 1 pea

x exp 1 I - x \i Vfi Aa 2H^ Aa a +

which agrees with the previous result.

Problems and Solutions

(b) T h e average energy of the system is

5ma2 (c) T h e probability of finding t h e particle in 0 < x < | at t = to is

7r:r + 2 cos cos a 1 J^_ 2 + 15^ C S l

/37r ftto 5V 2ma^ (3ir2ht0 ~W~

Problems and Solutions on

= 4 v l 5 I 1 (1 cos Tin) \mr J = 4\/l5[l-(-l)"](l/n7r)3, and hence ip(x, t) = ^2(n\

/ E tp(t = 0))ipn(x) exp f - i -^^

we write the solution as i> = A e i a * + B e ~ i a 0 ,

and the corresponding eigenfunctions are

Problems and Solutions

Integrating the left-hand side of the above, we have

2a (a) Fig. 1.3 (b)

Problems and Solutions

For bound states we require Vo < E < 0. Let

for x > a (region I ) , for x < a (region III).

Ak cos ka - Bk sin ka = -Cfc' e~k

Afc cos ka + Bk sin fca = Dk' e~ ' ; or

2A sin ka = (C - D) e"-k'a 2 S cos ka = (C + D) e -k'a 2Ak cos ka = -(C-D) k' e~k'a ,

2Bk sin ka = (C + D)k' e~k'a .

Problems and Solutions

Problems and Solutions on

Problems and Solutions

1019 Consider the following one-dimensional potential wells: V[x\

Problems and Solutions on Electromagnetism

0< x < a, a<x,

where A = * ^ p ^ v" = ' "h (i 1, 2), h

*' = ^ P ^ , for which h A2 k'a C2 e~2k'a ( [} sin 2ka 2ka

The continuity of ip at x = a gives A C e-k'a sin ka

Setting, as before, rj = k'a, = ka, we have Pi V

^'(o+) - v'(cr) + c/0v(o) = o.

V>'(o + )-v'(<r) = -2fcV>(o).

Problems and Solutions on

where the arbitrary constant A has been obtained by the normalization

/-, ^dx + C tfdx = l.

for x < 0, for a; > 0 ,

Problems and Solutions on

On account of symmetry, the probabilities are P(|a;|<xo)=2|A|2