The Physics of Quantum Mechanics (Instructor Solution
The Physics of Quantum Mechanics (Instructor Solution
Problems
1.1 What physical phenomenon requires us to work with probability amplitudes rather than just
with probabilities, as in other fields of endeavour?
Soln: Quantum interference.
1.2 What properties cause complete sets of amplitudes to constitute the elements of a vector space?
Soln: Meaning can be given to (a) adding two sets of amplitudes and (b) multiplying a set by an
arbitrary complex number.
1.3 V ′ is the dual space of the vector space V . For a mathematician, what objects comprise V ′ ?
Soln: The linear functions on V .
1.4 In quantum mechanics, what objects are the members of the vector space V ? Give an example
for the case of quantum mechanics of a member of the dual space V ′ and explain how members of
V ′ enable us to predict the outcomes of experiments.
Soln: Any complete set of amplitudes is a member of V . Since a complete set of amplitudes
uniquely specifies the physical state of the system, we can consider that physical states are the
members of V . V ′ is inhabited by bras. A bra is a function that extracts from a ket the amplitude
for a specific event. For example hE2 | extracts from |ψi the amplitude hE2 |ψi that a measurement
of energy determines that the system is in its second excited state.
1.5 Given that |ψi = eiπ/5 |ai + eiπ/4 |bi, express hψ| as a linear combination of ha| and hb|.
Soln:
hψ| = e−iπ/5 ha| + e−iπ/4 hb|
1.6 What properties characterise the bra ha| that is associated with the ket |ai?
Soln: If we choose any orthonormal basis that includes |ai, then ha| is the linear function that
vanishes on every basis function except |ai.
1.7 An electron can be in one of two potential wells that are so close that it can ‘tunnel’ from one
to the other (see §5.2 for a description of quantum-mechanical tunnelling). Its state vector can be
written
|ψi = a|Ai + b|Bi, (1.1)
where |Ai is the state of being in the first well and |Bi is the state of being in the second well and all
kets are correctly normalised. What is the probability
√ of finding the particle in the first well given
that: (a) a = i/2; (b) b = eiπ ; (c) b = 13 + i/ 2?
Soln: (a) P = 14 ; (b) P = 0; (c) P = 18 7
.
1.8 An electron can ‘tunnel’ between potential wells that form a chain, so its state vector can be
written
X∞
|ψi = an |ni, (1.2a)
−∞
where |ni is the state of being in the nth well, where n increases from left to right. Let
|n|/2
1 −i
an = √ einπ . (1.2b)
2 3
a. What is the probability of finding the electron in the nth well?
b. What is the probability P
of finding the electron in well 0 or anywhere to the right of it?
∞
Soln: (a) P = 21 3−|n| . (b) n=1 3−n = 12 and P0 = 12 so P (n ≥ 0) = 34 .
2 Problems
Problems
2.1 How is a wavefunction ψ(x) written in Dirac’s notation? What’s the physical significance of
the complex number ψ(x) for given x?
Soln: ψ(x) ≡ hx|ψi. ψ(x) is the amplitude to be found at x.
2.2 Let Q be an operator. Under what circumstances is the complex number ha|Q|bi equal to the
complex number (hb|Q|ai)∗ for any states |ai and |bi?
Soln: When Q is Hermitian.
2.3 Let Q be the operator of an observable and let |ψi be the state of our system.
a. What are the physical interpretations of hψ|Q|ψi and |hqn |ψi|2 , where |qn i is the nth eigenket
of the observable Q andPqn is the corresponding eigenvalue?
b. What is P the operator n |qn ihqn |, where the sum is over all eigenkets of Q? What is the
operator n qn |qn ihqn |?
c. If un (x) is the wavefunction of the state |qn i, write down an integral that evaluates to hqn |ψi.
Soln: (a) hψ|Q|ψi is the expectation
P value of Q. |hqn |ψi|2 is the probability ofRobtaining qn on
measuring the observable Q. (b) n |qn ihqn | is the identity operator. (c) hqn |ψi = dx u∗n (x)ψ(x).
2.4 What does it mean to say that two operators commute? What is the significance of two
observables having mutually commuting operators?
Given that the commutator [P, Q] 6= 0 for some observables P and Q, does it follow that for all
|ψi 6= 0 we have [P, Q]|ψi =
6 0?
Soln: Commutation implies you always get the same result no matter in which order you use the
operators. If observables have commuting operators, there is a complete set of states in which both
observables have well-defined values.
No, [P, Q] 6= 0 implies only that there is some state |φi such that P Q|φi 6= QP |φi; one may
readily construct an example of operators such that P Q|ii = QP |ii for basis vectors |ii with i = 2, ∞
but P Q|0i 6= QP |0i and P Q|1i 6= QP |1i.
2.5 Let ψ(x, t) be the correctly normalised wavefunction of a particle of mass m and potential
energy V (x). Write down expressions for the expectation values of (a) x; (b) x2 ; (c) the momentum
px ; (d) p2x ; (e) the energy.
What is the probability
R∞ that the particleR will be found in the interval (x1 , x2 )?
∞
Soln: (a) hxi = −∞ dx x|ψ|2 ; (b) hx2 i = −∞ dx x2 |ψ|2 ;
Z ∞ Z ∞ Z ∞ 2 2
dψ d2 ψ −h̄ d ψ
(c) hpi = −ih̄ dx ψ ∗ ; (d) hp2 i = −h̄2 dx ψ ∗ 2 ; (e) hEi = dx ψ ∗ + V (x)ψ .
−∞ dx −∞ dx −∞ 2m dx2
Rx
P (x1 , x2 ) = x12 dx |ψ|2 .
2.6 Write down the time-independent (tise) and the time-dependent (tdse) Schrödinger equa-
tions. Is it necessary for the wavefunction of a system to satisfy the tdse? Under what circumstances
does the wavefunction of a system satisfy the tise?
Soln:
∂|ψi
tise : H|En i = En |En i; tdse : ih̄ = H|ψi.
∂t
The state of every system must satisfy the tdse, whereas the tise is satisfied only if the state has
well defined energy (which for real systems is never the case).
2.7 Why is the tdse first-order in time, rather than second-order like Newton’s equations of
motion?
Soln: Because |ψi is a complete set of information, so for physics to be possible it must contain
complete initial conditions for solution of the time-evolution equation. If the tdse were second-order
in t, this condition would be violated.
2.8 A particle is confined in a potential well such that its allowed energies are En = n2 E, where
n = 1, 2, . . . is an integer and E a positive constant. The corresponding energy eigenstates are |1i, |2i,
. . . , |ni, . . . At t = 0 the particle is in the state
|ψ(0)i = 0.2|1i + 0.3|2i + 0.4|3i + 0.843|4i. (2.1)
a. What is the probability, if the energy is measured at t = 0, of finding a number smaller than
6E?
Problems 3
b. What is the mean value and what is the rms deviation of the energy of the particle in the state
|ψ(0)i?
c. Calculate the state vector |ψi at time t. Do the results found in (a) and (b) for time t remain
valid for arbitrary time t?
d. When the energy is measured it turns out to be 16E. After the measurement, what is the state
of the system? What result is obtained if the energy is measured again?
Soln: (a) 0.13; (b) the probabilities are P1 = 0.04, P2 = 0.09, P3 = 0.16, P4 = 0.71 so hEi = 13.2E,
hE 2 i = 196E 2 , and σE = 4.67E. (c) |ψ(t)i = 0.2e−iEt/h̄ |1i + 0.3e−i4Et/h̄ |2i + 0.4e−i9Et/h̄ |3i +
0.843e−i16Et/h̄|4i. (d) Thereafter |ψi = |4i and E = 16E with certainty.
2.9 A system has a time-independent Hamiltonian that has spectrum {En }. Prove that the prob-
ability Pk that a measurement of energy will yield the value Ek is is time-independent. Hint: you
can do this either from Ehrenfest’s theorem, or by differentiating hEk , t|ψi w.r.t. t and using the
tdse.
Soln: This follows from Ehrenfest’s theorem: the expectation value of any function of H is con-
stant, so the probability distribution defined by the Pk is time-independent. For a more direct
answer, consider the amplitude ak (t) to get Ek : ak = hEk , t|ψi, so using the tdse and its adjoint
∂ak ∂hEk , t| ∂|ψi 1
= |ψi + hEk , t| = (−hEk , t|H|ψi + hEk , t|H|ψi) = 0.
∂t ∂t ∂t ih̄
2.10 Let ψ(x) be a properly normalised wavefunction and Q an operator on wavefunctions. Let
{qr } be the spectrum of Q and {ur (x)} be the corresponding correctly normalised eigenfunctions.
WritePdown an expression for the probability that a measurement of Q will
R ∞ yield the value qr . Show
that r P (qr |ψ) = 1. Show further that the expectation of Q is hQi ≡ −∞ ψ ∗ Q̂ψ dx.1
Soln: Z 2
Pi = dx u∗i (x)ψ(x) where Qui (x) = qi ui (x).
!
Z Z X X X Z X X
1= dx |ψ|2 = dx a∗i u∗i aj u j = a∗i aj dx u∗i uj = |ai |2 = Pi
i j ij i i
Similarly
Z Z ! Z
X X X
dx ψ ∗ Qψ = dx a∗i u∗i Q aj u j = a∗i aj dxu∗i Quj
i j ij
X Z X X
= a∗i aj qj dx u∗i uj = |ai |2 qi = Pi qi
ij i i
which is by definition the expectation value of Q.
2.11 Find the energy of neutron, electron and electromagnetic waves of wavelength 0.1 nm.
Soln:
p2 (h/λ)2 mn
En = = = 1.32 × 10−20 J = 0.0821 eV; Ee = = 1837En = 150.8 eV;
2m 2m me En
Eγ = hν = hc/λ = 1.988 × 10−15 J = 12.407 keV.
2.12 Neutrons are emitted from an atomic pile with a Maxwellian distribution of velocities for
temperature 400 K. Find the most probable de Broglie wavelength in the beam.
Soln: The rate at which a neutron hits the wall of the pile is proportional to its momentum p, so the
2 2
density of neutrons hitting the wall is proportional to p3 e−p /2mkT = (2mkT )3/2 x3 e−x ,
probability √
2
wherepp = 2mkT x.√ The probability density is extremised when (3x2 − 2x4 )e−x = 0,√i.e., when
x = 3/2 and p = 3mkT . Thus the most probable de Broglie wavelength is λ = h/ 3mkT =
0.126 nm.
1 In an elegant formulation of quantum mechanics, this last result is the basic postulate of the theory, and one
derives other rules for the physical interpretation of the qn , an , etc., from it – see J. von Neumann, Mathematical
Foundations of Quantum Mechanics, Princeton University Press.
4 Problems
2.13 A beam of neutrons with energy E runs horizontally into a crystal. The crystal transmits
half the neutrons and deflects the other half vertically upwards. After climbing to height H these
neutrons are deflected through 90◦ onto a horizontal path parallel to the originally transmitted beam.
The two horizontal beams now move a distance L down the laboratory, one distance H above the
other. After going distance L, the lower beam is deflected vertically upwards and is finally deflected
into the path of the upper beam such that the two beams are co-spatial as they enter the detector.
Given that particles in both the lower and upper beams are in states of well-defined momentum,
show that the wavenumbers k, k ′ of the lower and upper beams are related by
mn gH
k′ ≃ k 1 − . (2.2)
2E
In an actual experiment (R. Colella et al., Phys. Rev. Let., 34, 1472, 1975) E = 0.042 eV and
LH ∼ 10−3 m2 (the actual geometry was slightly different). Determine the phase difference between
the two beams at the detector. Sketch the intensity in the detector as a function of H.
Soln: We take the zero of gravitational potential to be at the level of the lower beam. Then
p2 p′2 2m2 gH mgH
E= = mgH + ⇒ p′2 = p2 1 − = p 2
1 − .
2m 2m p2 E
Square rooting each side and simplifying with the binomial theorem, we find
′ mgH mgHLk m2 gHL
k ≃k 1− and ∆φ = (k − k ′ )L ≃ = ≃ 56.4.
2E 2E h̄2 k
The intensity in the detector will be proportional to |1 + ei∆φ |2 = 4 cos2 (∆φ/2), so the plot is a
sinusoid in H with wavenumber m2 gL/h̄2 k.
2.14 A particle moves in the potential V (x) and is known to have energy En . (a) Can it have well-
defined momentum for some particular V (x)? (b) Can the particle simultaneously have well-defined
energy and position?
Soln: (a) A state of well defined momentum is eip·x/h̄ . This is an eigenfunction of H only if
V = constant. (b) A state of well defined position is δ(x), which is not a plane wave so not a state
of well defined momentum.
2.15 The states {|1i, |2i} form a complete orthonormal set of states for a two-state system. With
respect to these basis states the operator σy has matrix
0 −i
σy = . (2.3)
i 0
Could σ be an observable? What are its eigenvalues and eigenvectors in the {|1i, |2i} basis? Deter-
mine the result of operating with σy on the state
1
|ψi = √ (|1i − |2i). (2.4)
2
Soln: It is Hermitian so could be the operator of an observable. Its eigenvalues are 1 and −1 and
the associated eigenkets are 2−1/2 (i, −1) and 2−1/2 (i, 1), respectively.
0 −i 1 1 i
σy |ψi = 2−1/2 =√ .
i 0 −1 2 i
2.16 A three-state system has a complete orthonormal set of states |1i, |2i, |3i. With respect to
this basis the operators H and B have matrices
1 0 0 1 0 0
H = h̄ω 0 −1 0 B = b0 0 1, (2.5)
0 0 −1 0 1 0
Problems 5
2.20 Given that for any two operators (AB)† = B † A† , show that
(ABCD)† = D† C † B † A† . (2.7)
Soln: We group the four factors into two pairs and apply what’s given twice:
(ABCD)† = (CD)† (AB)† = D† C † B † A† .
[ABC, D] = [(AB)C, D] = AB[C, D] + [(AB), D]C = AB[C, D] + A[B, D]C + [A, D]BC.
Analogously
d df dg dh
(f gh) = gh + f h + f g .
dx dx dx dx
2.22 Show that for any three operators A, B and C, the Jacobi identity holds:
[A, [B, C]] + [B, [C, A]] + [C, [A, B]] = 0. (2.9)
Soln: We expand the inner commutator and then use the rule for expanding the commutator of a
product:
[A, [B, C]] + [B, [C, A]] + [C, [A, B]] = [A, BC] − [A, CB] + [B, CA] − [B, AC] + [C, AB] − [C, BA]
= [A, B]C + B[A, C] − [A, C]B − C[A, B] + [B, C]A + C[B, A] − [B, A]C
− A[B, C] + [C, A]B + A[C, B] − [C, B]A − B[C, A]
= 2 ([A, B]C − C[A, B] − [A, C]B + B[A, C] + [B, C]A − A[B, C])
= 2 ABC − BAC − CAB + CBA − ACB + CAB + BAC − BCA + BCA
− CBA − ABC + ACB = 0
2.23 Show that a classical harmonic oscillator satisfies the virial equation 2hKEi = αhPEi and
determine the relevant value of α.
Soln: Classical harmonic motion is the solution x = A cos ωt to mẍ = −mω 2 x, where the amplitude
A is arbitrary. The time-averaged kinetic energy is 21 mhẋ2 i = 14 mω 2 A2 . The time-averaged potential
energy is 21 mω 2 hx2 i = 14 mω 2 A2 , so the virial theorem is satisfied for α = 2, as it should be for a
quadratic potential.
2.24 Given that the wavefunction is ψ = Aei(kz−ωt) + Be−i(kz+ωt) , where A and B are constants,
show that the probability current density is
J = v |A|2 − |B|2 ẑ, (2.10)
h̄kẑ
J=− (Aeikz + Be−ikz )(−A∗ e−ikz + B ∗ eikz ) − (A∗ e−ikz + B ∗ eikz )(Aeikz − Be−ikz ) .
2m
When we multiply out the brackets on the right, the terms involving both A and B cancel and
the other terms add up to give the required expression. Each wave represents a stream of particles
moving at speed v, that with coefficient A moving to increasing z and that with coefficient B moving
in the opposite direction.
Problems 7
Problems
3.1 After choosing units in which everything, including h̄ = 1, the Hamiltonian of a harmonic
oscillator may be written H = 21 (p2 + x2 ), where [x, p] = i. Show that if |ψi is a ket that satisfies
H|ψi = E|ψi, then
1 2 2
2 (p + x )(x ∓ ip)|ψi = (E ± 1)(x ∓ ip)|ψi. (3.1)
Explain how this algebra enables one to determine the energy eigenvalues of a harmonic oscillator.
Soln: Let A ≡ x + ip so
A† A = (x − ip)(x + ip) = p2 + x2 + i[x, p] = p2 + x2 − 1 = 2H − 1.
and
[A† , A] = [x − ip, x + ip] = 2i[x, p] = −2.
H|ψi = E|ψi ⇒ EA† |ψi = A† H|ψi = HA† |ψi + [A† , H]|ψi (3.2)
Now [A , H] = 2 [A† , A† A] = 21 A† [A† , A] = −A† . Substituting this into (3.2)
† 1
3.4 Show that the minimum value of E(p, x) ≡ p2 /2m + 12 mω 2 x2 with respect to the real numbers
p, x when they are constrained to satisfy xp = 12 h̄, is 21 h̄ω. Explain the physical significance of this
result.
8 Problems
3.5 How many nodes are there in the wavefunction hx|ni of the nth excited state of a harmonic
oscillator?
Soln: n nodes.
p
3.6 Show that in terms of a harmonic oscillator’s characteristic length ℓ ≡ h̄/2mω the ladder
operators can be written
x ∂ x ∂
A= +ℓ and A† = −ℓ . (3.4)
2ℓ ∂x 2ℓ ∂x
2 2
Hence show that the wavefunction of the second excited state is hx|2i = constant×(x2 /ℓ2 −1)e−x /4ℓ
and find the normalising constant.
Soln: In the position representation,
ℓ x ℓ x ∂ x ∂
A = (mωx + ip) = +i p = +ℓ and A† = −ℓ
h̄ 2ℓ h̄ 2ℓ ∂x 2ℓ ∂x
∂u0 x 2 2
Au0 = 0 ⇒ ℓ = − u0 ⇒ u0 ∝ e−x /4ℓ
∂x 2ℓ
R∞
From 1 = −∞ dx |u0 |2 we find that the normalizing constant is (2πℓ2 )1/4 .
1 ∂ 2
u1 = A† u0 = 2 1/4
1
2 y − e−y /4
(2πℓ ) ∂y
with y ≡ x/ℓ
1 2
−y /4
= ye
(2πℓ2 )1/4
√ 1 ∂ 2 1 2
2u2 = A† u1 = 2 1/4
1
2 y − ye−y /4 = 2 1/4
y 2 − 1 e−y /4
(2πℓ ) ∂y (2πℓ )
So 2
1 x 2 2
u2 (x) = − 1 e−x /4ℓ
(8πℓ2 )1/4 ℓ2
3.7 Explain why the wavefunction hx|ni of the oscillator’s nth stationary state must have the form
2
/4ℓ2
hx|ni = Hn (x/ℓ) e−x , (3.5)
th √
where Hn is √
an n -order (‘Hermite’) polynomial. By casting the equations A|ni = n|n − 1i and
A† |n − 1i = n|ni in the x-representation, show that
√ √ x
Hn′ (x/ℓ) = nHn−1 (x/ℓ) and ′
nHn (x/ℓ) = Hn−1 (x/ℓ) − Hn−1 (x/ℓ). (3.6)
ℓ
and thus that √ x √
nHn (x/ℓ) = Hn−1 (x/ℓ) − n − 1Hn−2 (x/ℓ). (3.7)
ℓ
Problems 9
Given that H0 = (2πℓ2 )−1/4 and H1 (y) = y/(2πℓ2 )1/4 , use this recurrence relation to reproduce the
plots of the wavefunctions hx|2i and hx|40i shown in Figure 3.4. Explain the physical significance of
the vertical arrows. Why is the amplitude of hx|40i largest near the right arrow?
2 2
Soln: We obtain hx|ni by applying A† to e−x /4ℓ and by Problem 3.6 each such application adds
a factor x/2ℓ either by multiplication or from differentiation of the exponential, so it is clear that
hx|ni will be a polynomial times the exponential. Now
√ √ −x2 /4ℓ2 x ∂ 2 2 2 2
nhx|n − 1i = nHn−1 (x/ℓ) e = +ℓ Hn (x/ℓ)e−x /4ℓ = Hn′ (x/ℓ)e−x /4ℓ
2ℓ ∂x
from which the first required relation follows. Similarly
√ √ −x2 /4ℓ2 x ∂ 2 2
nhx|ni = nHn (x/ℓ) e = −ℓ Hn−1 (x/ℓ)e−x /4ℓ
2ℓ ∂x
x 2 2
′
= Hn−1 (x/ℓ) − Hn−1 (x/ℓ) e−x /4ℓ
ℓ
′
from which the second required relation follows. Using the first relation to eliminate Hn−1 from the
second relation, we obtained the required √ recursion relation. The arrows are located at the end of
the classically-allowed domain xmax = 4n + 2 ℓ and the amplitude peaks there because classically
a particle lingers at its turning points. Here’s a routine to compute Hn (x/ℓ)
double Hermite(double x,int n){//returns normalised <x|n> less exponential
double tpi4=pow(2*acos(-1),.25), Hnm1=1/tpi4, Hn=x/tpi4;
for(int i=1; i<n; i++){
double Hsav=Hn;
Hn=(x*Hn-sqrt(i)*Hnm1)/sqrt(i+1);
Hnm1=Hsav;
}
return Hn;
}
3.8 Use r
h̄
x= (A + A† ) = ℓ(A + A† ) (3.8)
2mω
to show for a harmonic oscillator that in the energy representation the operator x is
√
√0 1 √0 0 ...
1 √0 2 √0
0 2 √0 3 ···
3 ...
... ... ... √ . . .
xjk = ℓ (3.9)
... √ 0 n−1 .√. .
n−1 √0 n √
n √ 0 n + 1 · · ·
n+1 0
··· ··· ··· ··· ···
10 Problems
3.9 Show that the momentum operator of a harmonic oscillator can be expressed in terms of the
creation and annihilation operators as
r
ih̄ † h̄
p = (A − A) where ℓ ≡ . (3.10)
2ℓ 2mω
Hence show that 2
2 h̄
h0|p |0i = . (3.11)
2ℓ
How does this result relate to the physics of a free particle discussed in §2.3.3?
Soln: Trivially from the definition of A
√
† 2mh̄ω
p = (A − A )
2i
and from this and the definition of ℓ the required expression for p immediately follows. Now
h̄2 h̄2
h0|p2 |0i = − 2 h0|((A† )2 + A2 − A† A − AA† )|0i = 2 .
4ℓ 4ℓ
1/2
We have h0|x2 |0i = ℓ2 , so we have proved that x2 p2 = 12 h̄, which is the classic formulation
of the uncertainty principle. The wavefunction hx|0i is a Gaussian, so we have recovered a central
result of §2.3.3 by a short-cut.
3.10 At t = 0 the state of a harmonic oscillator, mass m, frequency ω, is
1 1
|ψi = √ |N − 1i + √ |N i. (3.12)
2 2
Show that subsequently r
√ h̄
hxit = N ℓ cos(ωt) where ℓ ≡ . (3.13)
2mω
Interpret this result physically. What does this example teach us about the validity of classical
mechanics?
Show that a classical oscillator with energy (N + 21 )h̄ω has amplitude
q
xmax = 2 N + 12 ℓ. (3.14)
Problems 11
To explain the discrepancy between these results, consider the case in which initially
N +K−1
X
1
|ψi = √ |ki (3.15)
K
k=N
√
with N ≫ K ≫ 1. Show that then hxit ≃ 2 N ℓ cos(ωt) consistent with classical physics.
Soln:
hxit = ℓhψ|(A + A† )|ψi
= 21 ℓ hN − 1|ei(N −1/2)ωt + hN |ei(N +1/2)ωt (A + A† ) |N − 1ie−i(N −1/2)ωt + |N ie−i(N +1/2)ωt
√ √
= 12 ℓ hN − 1|ei(N −1/2)ωt + hN |ei(N +1/2)ωt N |N ie−i(N −1/2)ωt + N |N − 1ie−i(N +1/2)ωt
√ √
= 21 ℓ N eiωt + e−iωt = N ℓ cos(ωt)
Since the potential energy is given by E = 12 mω 2 x2 , a classical oscillator with energy (N + 21 )h̄ω
has amplitude r r q
2E (2N + 1)h̄ω
xmax = = = 2 N + 12 ℓ.
mω 2 mω 2
which is twice the amplitude we have obtained from quantum mechanics.
When the energy is significantly uncertain,
N +K−1 N +K−1
ℓ X i(k+1/2)ωt X
hxit = e hk|(A + A† ) |jie−i(j+1/2)ωt .
K
k=N j=N
After A has operated on each ket in the sum, we will have K kets from |N − 1i up to |N + K − 2i.
For all these kets except the first there is in the left-hand sum exactly one bra that will produce a
non-zero product
√ √ −iωt with it. So the final result will be the sum of K − 1 ≃ K terms, each of order
je−iωt ≃√ iωtN e in size. Similarly after A† has operated on the kets we will have ∼ K terms
of order N√e . Gathering the terms together √ to form cosines and multiplying through by the
prefactor ℓ/ K we conclude that hxit ≃ 2 N ℓ cos(ωt).
We learn from this example that while classical physics is a useful approximation when probabil-
ity distributions are sharply peaked around their expectation values (so K ≪ N ), to be an accurate
approximation we require also that the energy is significantly uncertain (K ≫ 1).
3.11∗ By expressing the annihilation operator A of the harmonic oscillator in the momentum
representation, obtain hp|0i. Check that your expression agrees with that obtained from the Fourier
transform of r
1 −x2 /4ℓ2 h̄
hx|0i = e , where ℓ ≡ . (3.16)
(2πℓ2 )1/4 2mω
Soln: In the momentum representation x = ih̄∂/∂p so [x, p] = ih̄∂p/∂p = ih̄. Thus from Problem
3.8
x ℓ ℓp h̄ ∂
A= +i p =i +
2ℓ h̄ h̄ 2ℓ ∂p
ℓp h̄ ∂u0 2 2 2
0 = Au0 ⇒ u0 = − ⇒ u0 (p) ∝ e−p ℓ /h̄
h̄ 2ℓ ∂p
Alternatively, transforming u0 (x):
Z Z ∞ 2 2
1 e−x /4ℓ
hp|0i = dxhp|xihx|0i = √ dx e−ipx/h̄
h −∞ (2πℓ2 )1/4
Z ∞ 2 ! √
1 x ipℓ −p2 ℓ2 /h̄2 2ℓ π 2 2 2
= 2 2 1/4
dx exp − + e = 2 2 1/4
e−p ℓ /h̄
(2πℓ h ) −∞ 2ℓ h̄ (2πℓ h )
3.12 Show that for any two N × N matrices A, B, trace([A, B]) = 0. Comment on this result in
the light of the results of Problem 3.8 and the canonical commutation relation [x, p] = ih̄.
12 Problems
Soln:
X X X X X
trace(AB) = (AB)jj = Ajk Bkj ; trace(BA) = (BA)kk = Bkj Ajk = Ajk Bkj
j jk k kj kj
These two expressions differ only in the order Pof the summations over j and k. We can reverse
one order and thus show them to be equal iff jk |Ajk Bkj |2 is finite, which it will be for finite-
dimensional matrices with finite elements. Hence x and p can only be represented by infinite matrices
with elements that are not square summable, (which the matrices of Problem 3.8 are not).
3.13∗ A Fermi oscillator has Hamiltonian H = f † f , where f is an operator that satisfies
f 2 = 0, f f † + f † f = 1. (3.17)
2
Show that H = H, and thus find the eigenvalues of H. If the ket |0i satisfies H|0i = 0 with
h0|0i = 1, what are the kets (a) |ai ≡ f |0i, and (b) |bi ≡ f † |0i?
In quantum field theory the vacuum is pictured as an assembly of oscillators, one for each
possible value of the momentum of each particle type. A boson is an excitation of a harmonic
oscillator, while a fermion in an excitation of a Fermi oscillator. Explain the connection between
the spectrum of f † f and the Pauli principle.
Soln:
H 2 = f † f f † f = f † (1 − f † f )f = f † f = H
Since eigenvalues have to satisfy any equations satisfied by their operators, the eigenvalues of H
must satisfy λ2 = λ, which restricts them to the numbers 0 and 1. The Fermi exclusion principle
says there can be no more than one particle in a single-particle state, so each such state is a Fermi
oscillator that is either excited once or not at all.
Moreover,
H|bi = f † f f † |0i = f † (1 − f † f )|0i = f † |0i = |bi
so |bi is the eigenket with eigenvalue 1.
3.14 In the time interval (t + δt, t) the Hamiltonian H of some system varies in such a way that
|H|ψi| remains finite. Show that under these circumstances |ψi is a continuous function of time.
A harmonic oscillator with frequency ω is in its ground state when the stiffness of the spring
is instantaneously reduced by a factor f 4 < 1, so its natural frequency becomes f 2 ω. What is the
probability that the oscillator is subsequently found to have energy 32 h̄f 2 ω? Discuss the classical
analogue of this problem.
Soln: From the tdse δ|ψi = −(i/h̄)H|ψiδt, so limδt→0 |δ|ψi| = 0, proving that |ψi changes con-
tinuously no matter how suddenly H changes.
Z 2
3 2
P (E = 2 f ω) = dx u′∗
1 (x)u0 (x) = 0 by parity.
Classically the effect of a sudden change in spring stiffness depends on the phase of the oscillation
when the spring slackens: if the change happens when the mass is at x = 0, the mass has all its
energy in kinetic form and there is no change in its energy; if the spring slackens when the mass is
stationary, all its energy is invested in the spring and the mass suddenly gets poorer. We expect
quantum mechanics to yield a probability distribution for the changes in E that is similar to the
average of the classical changes over phase. Of course classical mechanics averages probabilities for
states of even and odd parity because it doesn’t recognise the parity of a state.
3.15∗ P is the probability that at the end of the experiment described in Problem 3.14, the
oscillator is in its second excited state. Show that when f = 12 , P = 0.144 as follows. First show
that the annihilation operator of the original oscillator
A = 21 (f −1 + f )A′ + (f −1 − f )A′† , (3.18)
Problems 13
where A′ and A′† are the annihilation and creation operators Pof the final oscillator. Then writing
the ground-state ket of the original oscillator as a sum |0i = n cn |n′ i over the energy eigenkets of
the final oscillator, show that the condition A|0i = 0 yields the recurrence relation
r
f −1 − f n
cn+1 = − −1 cn−1 . (3.19)
f +f n+1
Finally using the normalisation of |0i, show numerically that c2 ≃ 0.3795. What value do you get
for the probability of the oscillator remaining in the ground state?
Show that at the end of the experiment the expectation value of the energy is 0.2656h̄ω. Explain
physically why this is less than the original ground-state energy 12 h̄ω.
This example contains the physics behind the inflationary origin of the universe: gravity explo-
sively enlarges the vacuum, which is an infinite collection of harmonic oscillators (Problem 3.13).
Excitations of these oscillators correspond to elementary particles. Before inflation the vacuum is
unexcited so every oscillator is in its ground state. At the end of inflation, there is non-negligible
probability of many oscillators being excited and each excitation implies the existence of a newly
created particle.
Soln: From Problem 3.6 we have
mωx + ip mf 2 ωx + ip
A≡ √ A′ ≡ p
2mh̄ω 2mh̄f 2 ω
x iℓ fx iℓ
= + p = + p
2ℓ h̄ 2ℓ f h̄
Hence
f 2iℓ 1 f
A′ + A′† = x A′ − A′† = f p so A = (A′ + A′† ) + (A′ − A′† )
ℓ f h̄ 2f 2
X
0 = A|0i = 21 (f −1 + f )ck A′ |k ′ i + (f −1 − f )ck A′† |k ′ i
k
Xn √ √ o
−1
= 1
2 (f + f ) kck |k − 1′ i + (f −1 − f ) k + 1ck |k + 1′ i
k
Multiply through by hn′ |:
√ √
0 = (f −1 + f ) n + 1cn+1 + (f −1 − f ) ncn−1 ,
which is a recurrence relation from which all non-zero cn can be determined in√ terms of c0 . Put
c0 = 1 and solve for the cn . Then evaluate S ≡ |cn |2 and renormalise: cn →Pcn / S.
The probability of remaining in the ground state is |c0 |2 = 0.8. hEi = n |cn |2 (n + 21 )h̄f 2 ω. It
is less than the original energy because of the chance that energy is in the spring when the stiffness
is reduced.
3.16∗ In terms of the usual ladder operators A, A† , a Hamiltonian can be written
H = µA† A + λ(A + A† ). (3.20)
What restrictions on the values of the numbers µ and λ follow from the requirement for H to be
Hermitian?
Show that for a suitably chosen operator B, H can be rewritten
H = µB † B + constant, (3.21)
†
where [B, B ] = 1. Hence determine the spectrum of H.
Soln: Hermiticity requires µ and λ to be real. Defining B = A + a with a a number, we have
[B, B † ] = 1 and
H = µ(B † − a∗ )(B − a) + λ(B − a + B † − a∗ ) = µB † B + (λ − µa∗ )B + (λ − µa)B † + (|a|2 µ − λ(a + a∗ )).
We dispose of the terms linear in B by setting a = λ/µ, a real number. Then H = µB † B − λ2 /µ.
From the theory of the harmonic oscillator we know that the spectrum of B † B is 0, 1, . . ., so the
spectrum of H is nµ − λ2 /µ.
14 Problems
3.17∗ Numerically calculate the spectrum of the anharmonic oscillator shown in Figure 3.2. From
it estimate the period at a sequence of energies. Compare your quantum results with the equivalent
classical results.
Soln:
3.18∗ Let B = cA + sA† , where c ≡ cosh θ, s ≡ sinh θ with θ a real constant and A, A† are the
usual ladder operators. Show that [B, B † ] = 1.
Consider the Hamiltonian
H = ǫA† A + 21 λ(A† A† + AA), (3.22)
where ǫ and λ are real and such that ǫ > λ > 0. Show that when
ǫc − λs = Ec, λc − ǫs = Es (3.23)
with E a constant, [B, H] = EB. Hence determine the spectrum of H in terms of ǫ and λ.
Soln:
[B, B † ] = [cA + sA† , cA† + sA] = (c2 − s2 )[A, A† ] = 1
[B, H] = [cA + sA† , ǫA† A + 12 λ(A† A† + AA)] = c[A, ǫA† A + 12 λA† A† ] + s[A† , ǫA† A + 21 λAA]
= c(ǫA + λA† ) − s(ǫA† + λA) = cEA + sEA† = EB
as required. Let H|E0 i = E0 |E0 i. Then multiplying through by B
E0 B|E0 i = BH|E0 i = (HB + [B, H])|E0 i = (HB + EB)|E0 i
So H(B|E0 i) = (E0 − E)(B|E0 i), which says the B|E0 i is an eigenket for eigenvalue E0 − E.
We assume that the sequence of eigenvalues E0 , E0 −E, E0 −2E, . . . terminates because B|Emin i =
0. Mod-squaring this equation we have
0 = hEmin |B † B|Emin i = hEmin |(cA† + sA)(cA + sA† )|Emin i
= hEmin |{(c2 + s2 )A† A + s2 + cs(A† A† + AA)}|Emin i
= cshEmin |{(c/s + s/c)A† A + s/c + (A† A† + AA)}|Emin i
But eliminating E from the given equations, we find λ(c/s + s/c) = 2ǫ. Putting this into the last
equation
2ǫ †
0 = hEmin | A A + s/c + (A† A† + AA) |Emin i
λ
Multiplying through by λ/2 this becomes
0 = hEmin |{H + sλ/2c}|Emini
so Emin = −sλ/2c. Finally, x = s/c satisfies the quadratic
r
2 ǫ ǫ ǫ2
x −2 x+1=0 ⇒ x= ± − 1.
λ λ λ2
Also from the above E = ǫ − λx so the general eigenenergy is
p
En = Emin + nE = − 21 λx + nǫ − nλx = nǫ − (n + 21 )λx = nǫ − (n + 12 ) ǫ ± ǫ2 − λ2
p
= − 21 ǫ ∓ (n + 21 ) ǫ2 − λ2
We have to choose the plus sign in order to achieve consistency with our previously established value
of Emin ; thus finally p
En = − 12 ǫ + (n + 21 ) ǫ2 − λ2
3.19 This problem is all classical electromagnetism, but it gives physical insight into quantum
physics. It is hard to do without a command of Cartesian tensor notation (Appendix B). A point
Problems 15
charge Q is placed at the origin in the magnetic field generated by a spatially confined current
distribution. Given that
Q r
E= (3.24)
4πǫ0 r3
and B = ∇ × A with ∇ · A = 0, show that the field’s momentum
Z
P ≡ ǫ0 d3 x E × B = QA(0). (3.25)
Write down the relation between the particle’s position and momentum and interpret this relation
physically in light of the result you have just obtained.
Hint: write E = −(Q/4πǫ0 )∇r−1 and B = ∇ × A, expand the vector triple product and
integrate each of the resulting terms by parts so as to exploit in one ∇R· A = 0 and in
H the other
∇2 r−1 = −4πδ 3 (r). The tensor form of Gauss’s theorem states that d3 x ∇i T = d2 Si T no
matter how many indices the tensor T may carry.
Soln: In tensor notation with ∂i ≡ ∂/∂xi
Z
Q
P=− d3 x ∇r−1 × (∇ × A)
4π
becomes Z
Q X
Pi = − d3 x ǫijk ∂j r−1 ǫklm ∂l Am
4π
jklm
Z X
Q
=− d3 x ǫkij ǫklm ∂j r−1 ∂l Am
4π
jklm
Z X
Q
=− d3 x (δil δjm − δim δjl )∂j r−1 ∂l Am
4π
jlm
Z X
Q
=− d3 x (∂j r−1 ∂i Aj − ∂j r−1 ∂j Ai )
4π j
In the first integral we use the divergence theorem to move ∂j from r−1 to ∂i Aj and in the second
integral we move the ∂j from Aj to ∂j r−1 . We argue that the surface integrals over some bounding
sphere of radius R that arise in this process vanish as R → ∞ because their integrands have explicit
R−2 scaling, so they will tend to zero as R → ∞ so long as A → 0 no matter how slowly. After this
has been done we have Z Z
Q
Pi = − − d3 x (r−1 ∂j ∂i Aj + d3 x ∂j ∂j r−1 Ai
4π
The first integral vanishes because ∂j Aj = ∇ · A = 0 and in the second integral we note that
∂j ∂j r−1 = ∇2 r−1 = −4πδ 3 (x), so evaluation of the integral is trivial and yields the required expres-
sion.
The physical interpretation is that when we accelerate a charge, we have to inject momentum
into the emag field that moves with the charge. So the conserved momentum p associated with the
particle’s motion includes both the momentum in the particle and that in the field, which we have
just shown to be QA(x), where x is the particle’s location. Thus p = mẋ + QA. The particle’s
kinetic energy is H = 21 mẋ2 = (p − QA)2 /2m.
3.20 From equation (9.32) show that the normalised wavefunction of a particle of mass m that is
in the nth Landau level of a uniform magnetic field B is
2 2
rn e−r /4rB
e −inφ
hx|ni = √ n+1
, (3.26)
2(n+1)/2 n! π rB
p
where rB = h̄/QB. Hence show that the expectation of the particle’s gyration radius is
√ (n + 21 )!
hrin ≡ hn|r|ni = 2 rB . (3.27)
n!
16 Problems
A = A′ + ∇Λ. Thus the required wavefunction is the one just given multiplied by
eiQΛ/h̄ = eiQ(B×a)·x/2h̄
as required.
3.22 A particle of charge Q is confined to move in the xy plane, with electrostatic potential φ = 0
and vector potential A satisfying
∇ × A = (0, 0, B). (3.30)
Consider the operators ρx , ρy , Rx and Ry , defined by
1
ρ= êz × (p − QA) and R = r − ρ, (3.31)
QB
where r and p are the usual position and momentum operators, and êz is the unit vector along B.
Show that the only non-zero commutators formed from the x and y components of these are
2 2
[ρx , ρy ] = irB and [Rx , Ry ] = −irB , (3.32)
2
where rB= h̄/QB.
The operators a, a† , b and b† are defined via
1 1
a= √ (ρx + iρy ) and b= √ (Ry + iRx ). (3.33)
2 rB 2 rB
Problems 17
Evaluate [a, a†] and [b, b† ]. Show that for suitably defined ω, the Hamiltonian can be written
H = h̄ω a† a + 12 . (3.34)
Given that there exists a unique state |ψi satisfying
a|ψi = b|ψi = 0, (3.35)
what conclusions can be drawn about the allowed energies of the Hamiltonian and their degeneracies?
What is the physical interpretation of these results?
Soln: We have
1 1
ρx = − (py − QAy ) ; ρy = (px − QAx )
QB QB
so
1 1
[ρx , ρy ] = − [(py − QAy ), (px − QAx )] = ([py , Ax ] + [Ay , px ])
(QB)2 QB 2
−ih̄ ∂Ax ∂Ay ih̄ 2
= − = = irB
QB 2 ∂y ∂x QB
Also
1 1 ih̄ 2
[ρx , x] = − [py , x] = 0 ; [ρx , y] = − [py , y] = = irB
QB QB QB
1 2
[ρy , x] = [px , x] = −irB ; [ρy , y] = 0
QB
So
2
[ρx , Rx ] = [ρx , x − ρx ] = 0 ; [ρx , Ry ] = [ρx , y − ρy ] = irB (1 − 1) = 0
and similarly [ρy , Ri ] = 0. Finally
2
[Rx , Ry ] = [x − ρx , y − ρy ] = −[ρx , y] − [x, ρy ] + [ρx , ρy ] = −irB
Now
1 1
[a, a† ] = 2 [ρx + iρy , ρx − iρy ] = − 2 2i[ρx , ρy ] = 1
2rB 2rB
1 1
[b, b† ] = 2 [Ry + iRx , Ry − iRx ] = 2 2i[Rx , Ry ] = 1
2rB 2rB
1 1
a† a = 2 (ρx − iρy )(ρx + iρy ) = 2 (ρ2x + ρ2y + i[ρx , ρy ])
2rB 2rB
1 |p − QA|2 2 1 |p − QA|2
= 2 2
− rB = − 12
2rB (QB) 2h̄ QB
But
|p − QA|2 QB
H= = (a† a + 21 )h̄ = (a† a + 21 )h̄ω,
2m m
where ω ≡ QB/m.
Given a|ψi = 0, we have H|ψi = (a† a + 12 )h̄ω|ψi = 12 h̄ω|ψi so |ψi is a stationary state of energy
1 †
2 h̄ω. Given any stationary state |Ei of energy |Ei we have that a |Ei is a stationary state of energy
E + h̄ω:
Ha† |Ei = h̄ωa† (aa† + 21 )|Ei = h̄ωa† (a† a + [a, a† ] + 21 )|Ei = a† (H + h̄ω)|Ei = (E + h̄ω)a† |Ei,
so the energies are h̄ω × ( 21 , 32 , 52 , . . .).
Since [ρi , Rj ] = 0 we have [a, b] = [a, b† ] = 0 etc, and it follows that [b† , H] = 0. Hence b† |Ei is
a stationary state of energy E.
To show that it is distinct from |Ei, we define the number operator N ≡ b† b. N |ψi = 0 because
b|ψi = 0. Given that N |ni = n|ni, we have that
N (b† |ni) = b† (bb† |ni) = b† (b† b + [b, b† ])|ni = b† (N + 1)|ni = (n + 1)b† |ni
18 Problems
Thus by repeatedly applying b† to |ψi we can create distinct states of energy 12 h̄ω, and by applying
it to a† |ψi we can make distinct states of energy 32 h̄ω, etc. Hence all energy levels are infinitely
degenerate.
Physically the energy levels are degenerate because there are an infinite number of possible
locations of the gyrocentre of a particle that is gyrating with a given energy.
3.23 Using cylindrical polar coordinates (R, φ, z), show that the probability current density asso-
ciated with the wavefunction (3.26) of the nth Landau level is
2 2
h̄R2n−1 e−R /2rB R2
J(R) = − n+1 2n+2 n + 2 êφ , (3.36)
2 πn!mrB 2rB
p
where rB ≡ h̄/QB. Plot J as a function of R and interpret your plot physically.
Soln: In cylindrical polars,
∂ 1 ∂ ∂
∇= , , ,
∂R R ∂φ ∂z
so we obtain the required expression for J of the Landau level when we take |ψ| and θ from (3.26)
and use A = 12 BRêφ .
The figure shows the current density as a function of radius for the first six Landau levels. The
ground state is
√ in a class by itself. For the excited states the characteristic radius R increases with
n roughly as n as predicted by classical physics, while the peak magnitude of the flow is almost
independent of n because J ∝ ω|ψ 2 | and ω ∼ constant and |ψ | 2 ≃ 1/R as expected from classical
mechanics.
3.24 Determine the probability current density associated with the nth Landau ground-state wave-
function (9.46) (which for n = 4 is shown in Figure 9.1). Use your result to explain in as much
detail as you can why this state can be interpreted as a superposition of states in which the electron
gyrates around different gyrocentres. Hint: adapt equation (3.36).
2
Why is the energy of a gyrating electron incremented if we multiply the wavefunction e−(mω/4h̄)r
by v n = (x − iy)n but not if we multiply it by un = (x + iy)n ?
Soln: The wavefunction of the nth state in the ground-state level differs from that of the first state
2
of the nth excited level only in having un instead of v n in front of e−uv/4rB . Consequently these
wavefunctions have the same amplitudes but where the phase of the nth excited state is θ = −nφ,
that of the nth ground state is θ = nφ. Since J ∝ (h̄∇θ − QA) it follows that the probability current
is given by equation (3.36) with a minus sign in front of the n:
2 2
h̄r2n−1 e−r /2rB r2
J(r) = − n+1 2n+2 2 − n êφ ,
2 πn!mrB 2rB
√
J now changes sign at r = 2nrB : at small r it flows in the opposite sense to the current of the nth
excited level (and the opposite sense to that of classical gyrations), but flows in the same sense at
large r. This behaviour
√ arises naturally if you arrange clockwise vortices of radius ∼ rB around a
circle of radius 2nrB . So whereas in the nth excited level the particle is moving counter-clockwise
Problems 19
√
around a path of radius√∼ nrB , in this state it is moving on a circle of radius ∼ rB that is offset
from the centre by ∼ nrB . The direction of the offset is completely uncertain, so we have to
imagine a series of offset circles. These clearly generate an anti-clockwise flow around the origin:
When the ground-state ψ(r) is multiplied by either un or v n , a factor rn e±inφ is added. This
factor moves outwards the radius rm at which |ψ|2 peaks in the same way regardless of whether
we multiply by un or v n . But our expressions for J(r) show that when we multiply by un we shift
rm out to the radius near which J = 0 so there is not much kinetic energy. Had we multiplied by
v n , J and the kinetic energy would have been large at rm . This asymmetry between the effects of
multiplying by einφ or e−inφ arises because A always points in the direction êφ .
Problems
P Verify that [J, x · x] P
4.1 = 0 and [J, x · p] = 0 by using the commutation relations [xi , Jj ] =
i k ǫijk xk and [pi , Jj ] = i k ǫijk pk .
Soln: It suffices to compute the commutator with the ith component of J
X X X X
[Ji , x · x] = [Ji , xj xj ] = [Ji , xj ]xj + xj [Ji , xj ] = i ǫijk (xk xj + xj xk ) = 0
j j j jk
because the term in brackets is symmetric in ij while ǫijk is antisymmetric in ij. Similarly
X
[Ji , x · p] = i ǫijk (xk pj + xj pk ) = 0
jk
for the same reason.
4.2∗ Show that the vector product a × b of two classical vectors transforms like a vector under
T
rotations. Hint: A rotation
P matrix R satisfies
P the relations R · R = I and det(R) = 1, which in
tensor notation read p Rip Rtp = δit and ijk ǫijk Rir Rjs Rkt = ǫrst .
Soln: Let the rotated vectors be a′ = Ra ′
Xand b = Rb. Then
′ ′
(a × b )i = ǫijk Rjl al Rkm bm
jklm
X
= δit ǫtjk Rjl Rkm al bm
tjklm
X
= Rip Rtp ǫtjk Rjl Rkm al bm
ptjklm
X
= Rip ǫplm al bm = (Ra × b)i .
plm
P
4.3∗ We have shown that [vi , Jj ] = i k ǫijk vk for any operator whose components vi form a
vector.
P The expectation value of this operator relation in any state |ψi is then hψ|[vi , Jj ]|ψi =
i k ǫijk hψ|vk |ψi. Check that with U (α) = e−iα·J this relation is consistent under a further rotation
|ψi → |ψ ′ i = U (α)|ψi by evaluating both sides separately.
Soln: Under the further rotation the LHS → hψ|U † [vi , Jj ]U |ψi. Now
U † [vi , Jj ]U = U † vi Jj U − U † Jj vi U = (U † vi U )(U † Jj U ) − (U † Jj U )(U † vi U )
X X
= [Rik vk , Rjl Jl ] = Rik Rjl [vk , Jl ].
kl kl
20 Problems
Similarly, we have
0 0 0
J x = i 0 0 −1
0 1 0
so
−1 i √0 0 0 0 −1 0 1
J x′ = 21 0 0 2 0 0 −i −i √0 −i
1 i 0 0 i 0 0 2 0
√
−1 i √0 0 0√ 0 0
√ 2 √0
= 12 0 0 2 0 −i 2 0 = 12 2 √0 2
1 i 0 1 0 1 0 2 0
These results show that the only difference between the generators of rotations of ordinary 3d
vectors and the spin-1 representations of the angular-momentum operators, is that for conventional
vectors we use a different coordinate system than we do for spin-1 amplitudes. Apart from this, the
three amplitudes for the spin of a spin-1 particle to point in various directions are equivalent to the
components of a vector, and they transform among themselves when the particle is reoriented for
the same reason that the rotation of a vector changes its Cartesian components.
4.5 Determine the commutator [J Jx′ , J z′ ] of the generators used in Problem 4.4. Show that it is
′ ′
equal to −iJ
Jy , where J y is identical with Sy in the set (4.3).
Soln: This is straightforward extension of the previous problem.
4.6∗ Show that if α and β are non-parallel vectors, α is not invariant under the combined rotation
R(α)R(β). Hence show that
RT (β)RT (α)R(β)R(α)
is not the identity operation. Explain the physical significance of this result.
Soln: R(α)α = α because a rotation leaves its axis invariant. But the only vectors that are
invariant under R(β) are multiples of the rotation axis β. So R(β)α is not parallel to α.
If RT (β)RT (α)R(β)R(α) were the identity, we would have
R (β)RT (α)R(β)R(α)α = α ⇒ R(β)R(α)α = R(α)R(β)α ⇒ R(β)α = R(α)(R(β)α)
T
which would imply that R(β)α is invariant under R(α). Consequently we would have R(β)α = α.
But this is true only if α is parallel to β. So our original hypothesis that RT (β)RT (α)R(β)R(α) = I
is wrong. This demonstrates that when you rotate about two non-parallel axes and then do the
reverse rotations in the same order, you always finish with a non-trivial rotation.
4.7∗ In this problem you derive the wavefunction
hx|pi = eip·x/h̄ (4.4)
of a state of well-defined momentum from the properties of the translation operator U (a). The state
|ki is one of well-defined momentum h̄k. How would you characterise the state |k′ i ≡ U (a)|ki? Show
that the wavefunctions of these states are related by uk′ (x) = e−ia·k uk (x) and uk′ (x) = uk (x − a).
Hence obtain equation (4.4).
Soln: U (a)|ki is the result of translating a state of well-defined momentum by k. Moving to the
position representation
uk′ (x) = hx|U (a)|ki = hk|U † (a)|xi∗ = hk|x − ai∗ = uk (x − a)
Also
hx|U (a)|ki = hx|e−ia·p/h̄ |ki = e−ia·k hx|ki = e−ia·k uk (x)
Putting these results together we have uk (x − a) = e−ia·kuk (x). Setting a = x we find uk (x) =
eik·x uk (0), as required.
4.8 By expanding the anticommutator on the left and then applying the third rule of the set (2.22),
show that any three operators satisfy the identity
[{A, B}, C] = {A, [B, C]} + {[A, C], B}. (4.5)
22 Problems
Soln:
[{A, B}, C] = [AB, C] + [BA, C] = [A, C]B + A[B, C] + [B, C]A + B[A, C]
= {A, [B, C]} + {[A, C], B}.
4.10 Let P be the parity operator and S an arbitrary scalar operator. Explain why P and S must
commute.
Soln: A scalar is something that is unaffected by orthogonal transformations of three-dimensional
space. These transformations are made up of the rotations and reflections. The parity operator
reflects states, so it is a member of the group of transformations under which a scalar is invariant.
Consequently S commutes with P .
4.11 In this problem we consider discrete transformations other than that associated with parity.
Let S be a linear transformation on ordinary three-dimensional space that effects a reflection in a
plane. Let S be the associated operator on kets. Explain the physical relationship between the kets
|ψi and |ψ ′ i ≡ S|ψi. Explain why we can write
Shψ|x|ψi = hψ|S † xS|ψi. (4.7)
What are the possible eigenvalues of S?
Given that S reflects in the plane through the origin with unit normal n̂, show, by means of a
diagram or otherwise, that its matrix is given by
Sij = δij − 2ni nj . (4.8)
√
Determine the form of this matrix in the case that n = (1, −1, 0)/ 2. Show that in this case
Sx = yS and give an alternative expression for Sy.
Show that a potential of the form
p
V (x) = f (R) + λxy, where R ≡ x2 + y 2 (4.9)
satisfies V (Sx) = V (x) and explain the geometrical significance of this equation. Show that [S, V ] =
0. Given that E is an eigenvalue of H = p2 /2m + V that has a unique eigenket |Ei, what equation
does |Ei satisfy in addition to H|Ei = E|Ei?
Soln: |ψ ′ i ≡ S|ψi is the state in which the amplitude to be at the reflected point x′ = Sx is equal
to the amplitude to be at x. The mathematical statement of this fact is
hx′ |ψ ′ i = hSx|Ŝ|ψi = hx|ψi, (4.10)
Problems 23
Equation (4.7) states that the expectation value of x for the reflected state is the reflection of the
expectation value of x for the original state. Since x = SSx for all x, S 2 = I and the eigenvalues of
S have to be ±1.
S reverses the component of its target vector that is perpendicular to the plane. Thus we get
x′ by taking from x twice (n̂ · x)n̂:
X X X X
x′ = Sx = x − 2(n̂ · x)n → Sij xj = δij xj − 2( nj xj )ni = (δij − 2ni nj )xj . (4.11)
j j j j
√
For the concrete example given, n̂ = (1, −1, 0)/ 2, so
1 0 0 1 −1 0 0 1 0
Sij = 0 1 0 − −1 1 0 = 1 0 0 (4.12)
0 0 1 0 0 0 0 0 0
where the last equality follows because y, being Hermitian, can be considered to act backwards. But
Similarly
hx, y, z|V S|ψi = V (x, y, z)hx, y, z|S|ψi = V (x, y, z)hy, x, z|ψi.
But V (x, y, z) = V (y, x, z) so for and |ψi these two expressions are equal, and it follows that
[V, S] = 0.
4.12 Show that the operator defined by hx, y|S|ψi = hy, x|ψi is Hermitian.
Soln: Z
(hφ|S|ψi)∗ = dxdy (hφ|x, yihx, y|S|ψi)∗
Z
= dxdy (hφ|x, yi)∗ (hy, x|ψi)∗
Z
= dxdy hψ|y, xihx, y|φi
Z
= dxdy hψ|y, xihy, x|S|φi
= hψ|S|φi.
24 Problems
Problems
5.1 A particle is confined by the potential well
n
V (x) = 0 for |x| < a (5.1)
∞ otherwise.
Explain (a) why we can assume that there is a complete set of stationary states with well-defined
parity and (b) why to find the stationary states we solve the tise subject to the boundary condition
ψ(±a) = 0.
Determine the particle’s energy spectrum and give the wavefunctions of the first two stationary
states.
Soln: (a) The Hamiltonian H = p2 /2m + V commutes with the parity operator, [H, P ] = 0,
because changing x to −x everywhere (including in the derivative) leaves H unchanged. So there is
a complete set of mutual eigenkets of H and P .
(b) In the realistic case of finite V , say V0 , outside the well, the wavefn is exponentially decaying
−1/2
with a lengthscale that shortens like V0 . Consequently just outside the well the curvature of the
wavefunction becomes larger and larger without limit as V0 → ∞ and the value of the wavefunction
at the wall becomes smaller and smaller without limit. In the limit of this problem ψ vanishes at
the wall and has a kink as it moves to zero value outside the wall.
We solve
h̄2 d2 ψ
− = Eψ
2m dx2
for |x| < a subject to the boundary condition ψ(±a) = 0. The solutions of well-defined parity are
sin(kx) and cos(kx) with k determined by the boundary condition. Hence
1 jπx
uj (x) = √ cos
j odd
h̄2 j 2 π 2 a 2a
Ej = with
8a2 m
1 jπx
uj (x) = √ sin j even
a 2a
5.3 Find the probability distribution of measuring momentum p for the particle described in Prob-
lem 5.2. Sketch and comment on your distribution. Hint: express hp|xi in the position representation.
Soln: Z ∞ Z a
e−ipx/h̄ 1
hp|ψi = dx hp|xihx|ψi = dx √ √
−∞ −a h 2a
h i r
1 1 a h̄ sin(pa/h̄)
=√ e−ipx/h̄ =
2ah −ip/h̄ −a πa p
Problems 25
5.5 Show that the energies of bound, odd-parity stationary states of the square potential well
0 for |x| < a
V (x) = (5.6)
V0 > 0 otherwise,
26 Problems
are governed by
s r
W2 2mV0 a2
cot(ka) = − − 1 where W ≡ and k 2 = 2mE/h̄2 . (5.7)
(ka)2 h̄2
Show that for a bound odd-parity state to exist, we require W > π/2.
Soln: Our trial wavefunction is ψ(x) = sin kx for |x| < a and ψ(x) = Ae−Kx for x > a, where
k 2 = 2mE/h̄2 and K 2 = 2m(V0 −E)/h̄2 . We equate the two expressions for ψ at x = a and similarly
their derivatives. Dividing the second of these equations into the first to eliminate A, we obtain
s
k ka W2
tan(ka) = − = − p ⇒ cot(ka) = − − 1.
K W 2 − (ka)2 (ka)2
In the plot the (dashed) curve of the right side terminates at ka < W and is always negative. If
W < π/2 this curve terminates before the (full) curve of the left side has crossed the x axis so the
curves do not intersect and there is no solution.
5.6 Show that the √ correctly normalised wavefunction of a particle trapped by the potential V (x) =
−Vδ δ(x) is ψ(x) = Ke−K|x| , where K = mVδ /h̄2 . Show that although this wavefunction makes
it certain that a measurement of x will find the particle outside the well where its kinetic energy is
negative, the expectation value of its kinetic energy hEK i = 12 mVδ2 /h̄2 is in fact positive. Reconcile
this apparent paradox as follows: (i) Show that for a narrow, deep potential well of depth V0 and
half-width a, with 2V0 a = Vδ , ka ≃ W ≡ (2mV0 a2 /h̄2 )1/2 , while Ka ≃ W 2 . (ii) Hence show that the
contribution from inside the well to hEK i is |ψ(0)|2 Vδ regardless of the value of a. Explain physically
what is happening as we send a → 0.
Soln: AtR x > 0 the wave function is Ae−Kx , where RK is given by eqn. (5.14). The value of A follows
∞ a
from 12 = 0 dx A2 e−2Kx = A2 /2K. We have Vδ = −a dx V (x) = 2V0 a, so W ≡ (2mV0 a2 /h̄2 )1/2 =
(2mVδ a/h̄2 )1/2 tends to zero as a → 0. From Figure 5.2 it follows that ka ≪ 1, so in eqn. (5.8) we
can adopt tan(ka) ≃ ka to see that (ka)2 is a solution of the quadratic equation
p
x2 + x − W 2 = 0 ⇒ x = 12 (−1 ± 1 + 4W 2 ) ≃ − 21 [1 ∓ (1 + 2W 2 )].
The only viable solution is (ka)2 = W 2 , so k = W/a = (2mVδ /h̄2 a)1/2 . It follows that
(Ka)2 = (ka)2 tan2 (ka) ≃ (ka)4 ≃ W 4 .
In the well the particle has kinetic energy EK = h̄2 k 2 /2m = Vδ /2a. Consequently, the contribution
from the inside of the well to hEK i is |ψ(0)|2 × 2a × Vδ /2a as required.
Physically, as the well becomes narrower the kinetic energy of the particle in the well is increasing
at exactly the rate required to keep constant the contribution to the mean KE. In the limit of a
δ-function well, the kinetic energy is infinite so the vanishingly small amount of time spent in the
well still dominates the negative contributions to the KE coming from outside the well. Outside the
well the (negative) KE −h̄2 K 2 /2m ∼ W 4 /a2 is independent of a.
5.7 Reproduce the plots shown in Figure 5.1 of the wavefunctions of particles that are scattered
by a square barrier and a square potential well. Give physical interpretations of as many features
of the plots as you can.
Problems 27
Figure 5.1 The real part of the wavefunction when a free particle of energy E is scattered by a classically forbidden
square barrier barrier (top) and a potential well (bottom). The upper panel is for a barrier of height V0 = E/0.7 and
half-width a such that 2mEa2 /h̄2 = 1. The lower panel is for a well of depth V0 = E/0.2 and half-width a such that
2mEa2 /h̄2 = 9. In both panels (2mE/h̄2 )1/2 = 40.
Soln: For the top plot one has to plot the real part of
B sin(−kx + φ) − B ′ sin(−kx + φ′ ) for x < −a
ψ(x) = cos(Kx) + A sin(Kx) for |x| < a
B sin(kx + φ) + B ′ sin(kx + φ′ ) for x > a
p
where k = 40, a = 1/k, K = k 1/0.7 − 1, φ = arctan((k/K) coth(Ka))−ka, φ′ = arctan((k/K) tanh(Ka))−
′
ka, B = cosh(Ka)/ sin(ka+φ), B ′ = −Bei(φ −φ) andp A = B ′ sin(ka+φ′ )/ sinh(Ka). The correspond-
ing equations for the lower panel are a = 3/k, K = k 1 + 1/0.2, φ = − arctan((k/K) cot(Ka))−ka,
′
φ′ = arctan((k/K) tan(Ka)) − ka, B = cos(Ka)/ sin(ka + φ), B ′ = −Bei(φ −φ) and A = B ′ sin(ka +
φ′ )/ sin(Ka).
The dotted lines mark the discontinuities in the potential energy. In the barrier we see the
exponential decay of the wavefunction through the barrier, which flows into rather a small-amplitude
transmitted wave on the right. The profile is close to an exponential because the wave reflected by
the right-hand discontinuity is of very low amplitude compared to that transmitted by the left-
hand discontinuity. This results in A ≃ 1. In the lower panel the shorter wavelength in the well is
apparent, indicating that he particle speeds up as it enters the well. Its high speed is also responsible
to the wave having lower amplitude in the well than on the right, although the amplitude reduction
is modest because in the well particles move in both directions and the flux carried away on the
right is only the difference in the fluxes carried each way in the well. In both panels the wave on the
left represents particles moving in both senses, but this fact is not apparent from the real part of ψ
alone.
5.8 Give an example of a potential in which there is a complete set of bound stationary states of
well-defined parity, and an alternative complete set of bound stationary states that are not eigenkets
of the parity operator. Hint: modify the potential discussed apropos NH3 .
Soln: Two identical potential wells situated either side of the origin with an infinitely high barrier
between them. Each eigenenergy has an eigenket of even parity and one of odd parity. Linear
combinations of these (such as the ket in which the particle is certain to be in one well) are not
eigenkets of P .
5.9 A free particle of energy E approaches a square, one-dimensional potential well of depth V0
and width 2a. Show that the probability of being reflected by the well vanishes when Ka = nπ/2,
where n is an integer and K = (2m(E + V0 )/h̄2 )1/2 . Explain this phenomenon in physical terms.
28 Problems
Soln: The probability of reflection is cos2 (φ′ − φ) where the phases of the odd- and even-parity
solutions of the tise are given by
tan(ka + φ) = −(k/K) cot(Ka) and tan(ka + φ′ ) = (k/K) tan(Ka).
When Ka = nπ/2 with n odd, the cotangent vanishes and the tangent is infinite, so ka + φ = mπ
and ka + φ′ = (2r + 1)π/2 for integer m, r. Hence φ′ − φ = [2(m − r) − 1]π/2 and the reflection
probability vanishes. When n is even, the cotangent is infinite and the tangent vanishes, and φ′ − φ
is again an odd multiple of π/2.
There are contributions to the amplitude for reflection from reflections at x = −a and at x = a.
When Ka = nπ/2 the well is an integer number of half de Broglie wavelengths across, so the phase
of a wave advances by an integer multiple of 2π after crossing and recrossing the well. The reflection
at the near edge involves a phase change by π because the medium in the well is “more dense” than
that outside (in the sense that the wavelength is shorter). Consequently the reflection from the far
side of the well is π out of phase with the direct reflection when it interferes with the latter, and the
two reflections cancel.
5.10 Show that the phase shifts φ (for the even-parity stationary state) and φ′ (for the odd-parity
state) that are associated with scattering by a classically allowed region of potential V0 and width
2a, satisfy
tan(ka + φ) = −(k/K) cot(Ka) and tan(ka + φ′ ) = (k/K) tan(Ka),
where k and K are, respectively, the wavenumbers at infinity and in the scattering potential. Show
that
(K/k − k/K)2 sin2 (2Ka)
Prefl = cos2 (φ′ − φ) = . (5.8)
(K/k + k/K)2 sin2 (2Ka) + 4 cos2 (2Ka)
Hint: apply the cosine rule for an angle in a triangle in terms of the lengths of the triangle’s sides
to the top triangle in Figure 5.2.
Soln: The first pair of equations is derived at the top of §5.3.2. Using the notation t ≡ tan(ka + φ)
and t′ ≡ tan(ka + φ′ ) and calling the acute angle in the top triangle of the figure ∆, Pythagoras’
theorem and the cosine rule together yield
(1 + t2 ) + (1 + t′2 ) − (t′ − t)2 1 + t′ t
cos ∆ = √ √ = √ √
2 1 + t2 1 + t′2 1 + t2 1 + t′2
Substituting for t and t′ from the given expressions, we have
1 − (k/K)2
cos ∆ = p
1 + (k/K)4 + (k/K)2 (cot2 (Ka) + tan2 (Ka))
K/k − k/K
= p
(K/k)2 + (k/K)2 + cot2 (Ka) + tan2 (Ka)
(K/k − k/K) sin(Ka) cos(Ka)
= q
((K/k)2 + (k/K)2 ) sin2 (Ka) cos2 (Ka) + cos4 (Ka) + sin4 (Ka)
(K/k − k/K) 21 sin(2Ka)
= q
(K/k + k/K)2 sin2 (Ka) cos2 (Ka) + (cos2 (Ka) − sin2 (Ka))2
(K/k − k/K) sin(2Ka)
= q
(K/k + k/K)2 ) sin2 (2Ka) + 4 cos2 (2Ka)
Problems 29
5.11 A particle of energy E approaches from x < 0 a barrier in which the potential energy is
V (x) = Vδ δ(x). Show that the probability of its passing the barrier is
r
1 2mE 2mVδ
Ptun = where k = 2 , K= . (5.9)
1 + (K/2k) 2
h̄ h̄2
Soln: Integrating the tise over an infinitesimal interval around the origin, we find that at the
origin dψ/dx increases by Kψ. The wavefunction is
ikx
ψ(x) = e ikx + Re−ikx for x < 0
Be otherwise.
At x = 0 continuity of ψ and the required change in its gradient require
B =1+R B =1+R 1
⇒ ⇒ B=
ikB − KB = ik(1 − R) B(1 − K/ik) = 1 − R 1 + iK/2k
and we obtain the required result as |B|2 .
5.12 An electron moves along an infinite chain of potential wells. For sufficiently low energies
we can assume that the set {|ni} is complete, where |ni is the state of definitely being in the nth
well. By analogy with our analysis of the NH3 molecule we assume that for all n the only non-
vanishing matrix elements of the Hamiltonian are E ≡ hn|H|ni and A ≡ hn ± 1|H|ni. Give physical
interpretations of the numbers A and E.
Explain why we can write
X∞
H= E|nihn| + A (|nihn + 1| + |n + 1ihn|) . (5.10)
n=−∞
P
Writing an energy eigenket |Ei = n an |ni show that
am (E − E) − A (am+1 + am−1 ) = 0. (5.11)
ikm
Obtain solutions of these equations in which am ∝ e and thus find the corresponding energies
Ek . Why is there an upper limit on the values of k that need be considered?
Initially the electron is in the state
1
|ψi = √ (|Ek i + |Ek+∆ i) , (5.12)
2
where 0 < k ≪ 1 and 0 < ∆ ≪ k. Describe the electron’s subsequent motion in as much detail as
you can.
Soln: E is the energy of an electron in a well. A is the amplitude for tunnelling through to a
neighbouring well. Since the set {|ni} may be considered complete, any operator can be written
P
mn Qmn |mihn|, where Qmn = hm|Q|ni are the operator’s matrix elements. In the case of H most
of the matrix elements vanish and we are left with the given expression. When |Ei is expanded in
the |ni basis, the tise is
X ∞ ∞
X X
H|Ei = E|Ei = E an |ni = {E|nihn| + A (|nihn + 1| + |n + 1ihn|)} am |mi
n=−∞ n=−∞ m
X∞
= Ean |ni + A (an+1 |ni + an |n + 1i)
n=−∞
The required recurrence relation is obtained when this eqn is multiplied on the left by hm|. Substi-
tuting in the trial solution am = Keikm gives
0 = eikm (Ek − E) − A eik(m+1) + eik(m−1)
30 Problems
barrier’s typical height. Outline numerical work that would lead to an improved estimate of the
structure of the barrier.
Soln: The time between impacts with the wall is ∼ 10−14 /108 s, so in 6.4 Gyr ≃ 2 × 1017 s the α
makes ∼ 2 × 1039 escape attempts and the probability of any one attempt succeeding is ∼ 5 × 10−40 .
This probability is sin2 (∆φ) ≃ (4k/K)2 e−4W . Taking 4k/K to be order unity, we conclude that
W = 14 ln(1039 ) = 22.5.
The distance 2a from the nucleus at which the α becomes classically allowed is that at which
its energy of electrostatic repulsion from the thorium nucleus (Z = 92 − 2 = 90) is 4 MeV: 2a =
2 × 90e2 /(4πǫ0 4 × 106 e) = 6.5 × 10−14 m. If the barrier were square rather than wedge shaped, it
would not be so thick, so let’s adopt 2a = 3×10−14 m. Then V0 = (22.5h̄/a)2 /8mp ≃ 1.6×10−12 J ≃
10 MeV.
This calculation has two major problems: (a) the assumption of a square barrier instead of
a skew and sharply peaked one, and (b) the use of a one-dimensional rather than spherical model
given that α becomes allowed at a radius that is significantly greater than the radius of the nucleus.
Problem (a) is probably the most serious and could be addressed by numerically solving the one-d
tise as in Problem 5.15. To address problem (b) we would solve the radial wavefunction for a
particle with vanishing angular momentum.
5.15∗ Particles of mass m and momentum h̄k at x < −a move in the potential
(
0 for x < −a
V (x) = V0 12 [1 + sin(πx/2a)] for |x| < a (5.17)
1 for x > a,
where V0 < h̄2 k 2 /2m. Numerically reproduce the reflection probabilities plotted in Figure 5.20 as
follows. Let ψi ≡ ψ(xj ) be the value of the wavefunction at xj = j∆, where ∆ is a small increment
in the x coordinate. From the tise show that
ψj ≃ (2 − ∆2 k 2 )ψj+1 − ψj+2 , (5.18)
p
where k ≡ 2m(E − V )/h̄. Determine ψj at the two grid points with the largest values of x from
a suitable boundary condition, and use the recurrence relation (5.18) to determine ψj at all other
grid points. By matching the values of ψ at the points with the smallest values of x to a sum of
sinusoidal waves, determine the probabilities required for the figure. Be sure to check the accuracy
of your code when V0 = 0, and in the general case explicitly check that your results are consistent
with equal fluxes of particles towards and away from the origin.
Equation (12.40) gives an analytical approximation for ψ in the case that there is negligible
reflection. Compute this approximate form of ψ and compare it with your numerical results for
larger values of a.
Soln:
We discretise the tise
h̄2 d2 ψ h̄2 ψj+1 + ψj−1 − 2ψj
− 2
+ V ψ = Eψ by − + Vj ψj = Eψj
2m dx 2m ∆2
which readily yields the required recurrence relation. At the right-hand boundary we require a pure
outgoing wave, so ψj = exp(ijK∆) gives ψ at the two last grid points. From the recurrence relation
we obtain ψ elsewhere. At the left boundary we solve for A+ and A− the equations
A+ exp(i0k∆) + A− exp(−i0k∆) = ψ0
A+ exp(i1k∆) + A− exp(−i1k∆) = ψ1
Problems 33
The transmission probability is (K/k)/|A+ |2 . The code must reproduce the result of Problem 5.4
in the appropriate limit.
5.16∗ In this problem we obtain an analytic estimate of the energy difference between the even-
and odd-parity states of a double square well. Show that for large θ, coth θ − tanh θ ≃ 4e−2θ . Next
letting δk be the difference between the k values that solve
s p
W 2 coth W 2 − (ka)2 even parity
tan [rπ − k(b − a)] −1= p (5.19a)
(ka)2 tanh W 2 − (ka)2 odd parity,
where r
2mV0 a2
W ≡ (5.19b)
h̄2
for given r in the odd- and even-parity cases, deduce that
" 1/2 −1/2 # −1
W2 W2 1 (ka)2
−1 + −1 (b − a) + 1− δk
(ka)2 (ka)2 k W2 (5.20)
h p i
≃ −4 exp −2 W 2 − (ka)2 .
Hence show that when W ≫ 1 the fractional difference between the energies of the ground and first
excited states is √
δE −8a
≃ e−2W 1−E/V0 . (5.21)
E W (b − a)
Soln: First
eθ + e−θ eθ − e−θ 1 + e−2θ 1 − e−2θ
coth θ − tanh θ = θ − = − ≃ (1 + 2e−2θ ) − (1 − 2e−2θ ) = 4e−2θ
e − e−θ eθ + e−θ 1 − e−2θ 1 + e−2θ
So when W ≫ 1 the difference in the right side of the equations for k in the cases of even and odd
parity is small and we may estimate the difference in the left side by its derivative w.r.t. k times the
difference δk in the solutions. sThat is
W2 −W 2 /(ka)2 δk/k √
−2 W 2 −(ka)2
− s2 [rπ − k(b − a)](b − a)δk − 1 + tan[rπ − k(b − a)] q ≃ 4e
(ka)2 W2
(ka)2 − 1
In the case of interest the right side of the original equation is close to unity, so we can simplify the
last equation by using s
W2
tan [rπ − k(b − a)] −1≃1
(ka)2
With the help of the identity s2 θ = 1 + tan2 θ we obtain the required relation. We now approximate
the left side for W ≫ ka. This yields
W √ 2
(b − a)δk ≃ −4e−2W 1−(ka/W ) ($)
ka
2 2
Since E = h̄ k /2m, δE/E = 2δk/k and
2mEa2 h̄2
(ka/W )2 = × = E/V0 .
h̄2 2mV0 a2
The required relation follows when we use these relations in ($).
5.17 We consider the scattering of free particles of mass m that move in one dimension in the
potential V (x) = −W δ(x), with W > 0. (a) For a well of finite depth V0 and width 2a the condition
on the phases φ and φ′ of the even- and odd-parity wavefunctions ψ ∝ sin(kx + φ), etc., for free
particles are
k k
tan(ka + φ) = − cot(Ka), tan(ka + φ′ ) = − tan(Ka).
K K
34 Problems
Show that in the limit a → 0, V0 = W/2a → ∞ we have tan φ → −h̄2 k/mW and φ′ → 0. Hence
obtain the scattering cross-section by the δ-function potential
2
σ= . (5.22)
1 + (h̄k/mW )2
(b) Re-derive the equation above for φ by requiring that ψ = sin(k|x| + φ) satisfy the tise. Convince
yourself that ψ = sin(kx) is also consistent with the tise.
Soln: (a) The wavenumber k is constant as we send a → 0 so ka → 0. Within the well the
wavefunction is either cos(Kx) or sin(Kx), where
s
2m(E + V0 )
K= .
h̄2
Thus in the limit a → 0, Ka → 0 and the conditions on the phases become
k k kh̄2 kh̄2 k
tan φ = − =− 2 =− =− ; tan φ′ = − (Ka) = −ka → 0.
K tan(Ka) K a 2mV0 a mW K
The scattering cross section is σ = 2 cos2 (φ − φ′ ) = 2 cos2 φ. We evaluate the cosine by drawing a
right-angle triangle with sides of length h̄2 k and mW adjacent to the right angle.
(b) By integrating
h̄2 d2 ψ
− = (W δ(x) + E)ψ
2m ∂x2
from 0 − ǫ to 0 + ǫ we show that at the origin dψ/dx has a discontinuity
dψ 2mW
= − 2 ψ(0).
dx h̄
For the given even-parity trial wavefunction [dψ/dx] = 2k cos φ and ψ(0) = sin φ and the required
result follows. The odd-parity solution vanishes at the origin, so our condition on [dψ/dx] implies
that the gradient is, in fact, continuous.
Problems
6.1 A system AB consists of two non-interacting parts A and B. The dynamical state of A is
described by |ai, and that of B by |bi, so |ai satisfies the tdse for A and similarly for |bi. What
is the ket describing the dynamical state of AB? In terms of the Hamiltonians HA and HB of the
subsystems, write down the tdse for the evolution of this ket and show that it is automatically
satisfied. Do HA and HB commute? How is the tdse changed when the subsystems are coupled
by a small dynamical interaction Hint ? If A and B are harmonic oscillators, write down HA , HB .
The oscillating particles are connected by a weak spring. Write down the appropriate form of the
interaction Hamiltonian Hint . Does HA commute with Hint ? Explain the physical significance of
your answer.
Soln: State AB represented by |ai|bi.
∂|ai|bi ∂|ai ∂|bi
ih̄ = (HA + HB )|ai|bi ⇔ 0 = ih̄ − HA |ai |bi + |ai ih̄ − HB |bi ,
∂t ∂t ∂t
which holds by virtue of the individual tdses.
[HA , HB ] = 0 because operators belonging to different systems always commute.
For harmonic oscillators
p2 2 2 p2 2 2
HA = A + 12 mωA xA ; HB = B + 12 mωB xB ; Hint = 12 k(xA − xB )2 .
2m 2m
Hint commutes with neither HA (which contains pA ) nor HB (which contains pB ).
6.2 Explain what is implied by the statement that “the physical state of system A is correlated with
the state of system B.” Illustrate your answer by considering the momenta of cars on (i) London’s
Problems 35
circular motorway (the M25) at rush-hour, and (ii) the road over the Nullarbor Plain in southern
Australia in the dead of night.
Explain why the states of A and B must be uncorrelated if it is possible to write the state of
AB as a ket |AB; ψi = |A; ψ1 i|B; ψ2 i that is a product of states of A and B. Given a complete set
of states for A, {|A; ii}, and a corresponding complete set of states for B, {|B; ii}, write down an
expression for a state of AB in which B is possibly correlated with A.
Soln: A and B are correlated if the probability of measuring some value for A depends on the
result of a measurement we have just made on B. If the momentum of car A on the fast lane of the
clockwise carriageway of the M25 is found to vanish, we are very likely to find that the momentum
of the next car, B, vanishes too, while if the momentum of A is consistent with A moving at the
speed limit, the momentum of B is likely to be in this range too. Thus the momenta of nearby cars
on a busy highway are strongly correlated. By contrast, on a little-frequented highway cars don’t
have to worry about bumping into other cars (it suffices to keep to your side of the road), so finding
that car A is proceeding West at near the speed limit does not change the probability distribution of
speed for the nearest car, B, which will remain peaked at ± a bit over the speed limit with a bump
at 0 to take care of rest stops.
If |AB; ψi = |A; ψ1 i|B; ψ2 i, when we measure some observable Q of A the state of the system
collapses to |AB; φi = |A; qn i|B; ψ2 i, where the state of A is an eigenstate of Q. Since the way
state of B appears in the ket for AB is unchanged by the measurement on A, the probability of
any measurement on B yielding any given result is unchanged by the measurement on A, so in
direct-product states of AB the state of B is uncorrelated with that of A.
6.3 Given that the state |ABi of a compound system can be Pwritten as a product |Ai|Bi of states
of the individual systems, show that when |ABi is written as ij cij |A; ii|B; ji in terms of arbitrary
basis vectors for the
P subsystems, every column of the matrix cij is a multiple of the leftmost column.
Soln: If |Ai = i ai |A; ii and similarly for |Bi, cij = ai bj , so the ratio of corresponding terms in
the j th and k th column is cij /cik = bj /bk is independent of i as required.
6.4 Consider a system of two particles of mass m that each move in one dimension along a given
rod. Let |1; xi be the state of the first particle when it’s at x and |2; yi be the state of the second
particle when it’s at y. A complete set of states of the pair of particles is {|xyi} = {|1; xi|2; yi}.
Write down the Hamiltonian of this system given that the particles attract one another with a force
that’s equal to C times their separation.
Suppose the particles experience an additional potential
V (x, y) = 12 C(x + y)2 . (6.1)
Show that the dynamics of the two particles is now identical with the dynamics of a single particle
that moves in two dimensions in a particular potential Φ(x, y), and give the form of Φ.
Soln:
p2 p2
H= 1 + + 1 C(x − y)2 .
2m 2m 2
With the additional potential H becomes
p2 + p22 p2 + p22
H= 1 + 12 C{(x − y)2 + (x + y)2 } = 1 + C(x2 + y 2 ),
2m 2m
which is the Hamiltonian of a particle moving in 2d under the potential Φ = C(x2 + y 2 ).
6.5 In §6.1.4 we derived Bell’s inequality by considering measurements by Alice and Bob on an
entangled electron–positron pair. Bob measures the component of spin along an axis that is inclined
by angle θ to that used by Alice. Given the expression
|−, bi = cos(θ/2) eiφ/2 |−i − sin(θ/2) e−iφ/2 |+i, (6.2)
1
for the state of a spin-half particle in which it has spin − 2 along the direction b with polar angles
(θ, φ), with |±i the states in which there is spin ± 21 along the z-axis, calculate the amplitude
AB (−|A+) that Bob finds the positron’s spin to be − 21 given that Alice has found + 21 for the
electron’s spin. Hence show that PB (−|A+) = cos2 (θ/2).
Soln: After Alice mas measured + 21 for the electron on her chosen axis (with which we align our
z axis, we know that the positron is in the state |−i. The amplitude we require is
AB (−|A+) = h−, b|−i = cos(θ/2)e−iφ/2 .
36 Problems
6.9 In this problem we consider an alternative interpretation of the density operator. Any quantum
state can be expanded in the energy basis as
XN
√
|ψ; φi ≡ pn eiφn |ni, (6.6)
n=1
where φn is real and pn is the probability that a measurement of energy will return En . Suppose
we know the values of the pn but not the values of the phases φn . Then the density operator is
Z 2π N
d φ
ρ= |ψ; φihψ; φ|. (6.7)
(2π)N
P 0
Show that this expression reduces to n pn |nihn|. Contrast the physical assumptions made in this
derivation of ρ with those made in §6.3.
Clearly |ψ; φi can be expanded in some other basis {|qr i} as
Xp
|ψ; φi ≡ Pr eiηr |qr i, (6.8)
r
where Pr is the probability of obtaining qr on a measurement of the observable Q and the ηr (φ)
are unknown phases. Why does this second expansion not lead to the erroneous conclusion that ρ
is necessarily diagonal in the {|qr i} representation?
Soln: In §6.3 we assumed that the system was in an (unknown) eigenstate |ni, which is in practice
usually taken to be a stationary state. This is an improbable assumption. Here we assume that
the system is in a generic and non-stationary state. So this physical interpretation of the density
operator is superior to our earlier one.
In equation (6.7) we assume that all values of the phases φn are equally likely. Assuming that
all values of the phases ηr are equally likely will lead to ρ being diagonal in the {|qr i} representaton.
By substituting X
|ni = |hqr |ni| eiχnr |qr i
r
into (6.6), we can show that
!
1 X r pn
ηr (φ) = ln |hqr |ni| ei(φn +χnr ) .
i n
Pr
Thus the dependence of ηr on the φn is highly non-linear, and a uniform distribution in the φn will
in general not correspond to a uniform distribution in the ηr .
6.10 Show that the equation of motion of the density operator ρ is solved by
ρt = U (t)ρ0 U † (t), (6.9)
where U (t) ≡ e−iHt/h̄ is the time-evolution operator introduced in §4.3.
Soln: We simply differentiate the proposed solution and use U̇ = −(i/h̄)HU :
dρt dU dU †
= ρ0 U † + U ρ0
dt dt dt
i i
= − HU ρ0 U † + U ρ0 U † H = − [H, ρt ].
h̄ h̄
6.11∗ Show that when the density operator takes the form ρ = |ψihψ|, the expression Q = Tr Qρ for
the expectation value of an observable can be reduced to hψ|Q|ψi. Explain the physical significance
of this result. For the given form of the density operator, show that the equation of motion of ρ
yields
∂|ψi
|φihψ| = |ψihφ| where |φi ≡ ih̄ − H|ψi. (6.10)
∂t
38 Problems
Show from this equation that |φi = a|ψi, where a is real. Hence determine the time evolution of |ψi
given the at t = 0, |ψi = |Ei is an eigenket of H. Explain why ρ does not depend on the phase of
|ψi and relate this fact to the presence of a in your solution for |ψ, ti.
Soln: X
Tr(Qρ) = hn|Q|ψihψ|ni
n
We choose a basis that |ψi is a member. Then there is only one non-vanishing term in the sum,
when |ni = |ψi, and the right side reduces to hψ|Q|ψi as required. This result shows that density
operators recover standard experimental predictions when the system is in a pure state.
Differentiating the given ρ we have
dρ ∂|ψi ∂hψ| 1
= hψ| + |ψi = (H|ψihψ| − |ψihψ|H)
dt ∂t ∂t ih̄
Gathering the terms proportional to hψ| on the left and those proportional to |ψi on the right we
obtain the required expression. Now
|φihψ| = |ψihφ| ⇒ |φihψ|φi = |ψihφ|φi,
which establishes that |φi ∝ |ψi. We define a as the constant of proportionality. Using |φi = a|ψi
in |φihψ| = |ψihφ| we learn that a = a∗ so a is real.
Returning to the definition of |φi we now have
∂|ψi
ih̄ = (H − a)|ψi.
∂t
This differs from the tdse in having the term in a. If |ψi is an eigenfunction of H, we find that its
time dependence is |ψ, ti = |ψ, 0ie−i(E−a)t/h̄ rather than the expected result |ψ, ti = |ψ, 0ie−iEt/h̄ .
We cannot determine a from the density-matrix formalism because ρ is invariant under the trans-
formation |ψi → e−iχ |ψi, where χ is any real number.
P
6.12 The density operator is defined to be ρ = α pα |αihα|, where pα is the probability
P that the
system is in the state α. Given an arbitrary basis {|ii} and the expansions |αi = i aαi |ii, calculate
the matrix elements ρij = hi|ρ|ji of ρ. Show that the diagonal elements ρii are non-negative real
numbers and interpret them as probabilities.
Soln: X X
ρ= pα aαi a∗αj ρij |iihj| ⇒ ρij = pα aαi a∗αj
αij α
P 2
The diagonal element ρii = α pα |aαi | is the product of two non-negative real numbers so it’s real
non-negative itself.
Consider the observable Q = |iihi| which is diagonal in the given basis and has eigenvalues 1
and 0: you measure unity if the system is in the state |ii and zero otherwise. ρii = Tr(Qρ) = Q is
the expectation value of Q. Consequently it’s the probability that if you measure Q you force the
system into the state |ii. Loosely speaking we can consider ρii to be the probability of finding the
system in the state |ii.
P
6.13 Consider the density operator ρ = ij ρij |iihj| of a system that is in a pure state. Show
that every row of the matrix ρij is a multiple of the first row and every column is a multiple of the
first column. Given that these relations between the rows and columns of a density matrix hold,
show that the system is in a pure state. Hint: exploit the real, non-negativity of ρ11 established in
Problem 6.12 and the Hermiticity of ρ. P
Soln: Since the system is in a pure state |ψi = i ai |ii, ρ = |ψihψ|. So ρij = ai a∗j . The ratio of
the values of the j th elements in the ith and k th rows is
ρij ai
=
ρkj ak
which is independent of j as required. Similarly, the ratio of the ith elements of the j th and k th
columns, a∗j /a∗k is independent of i.
Problems 39
Conversely, given these relations we recall that the diagonal elements of ρ are real and non-
√
negative so we may define a1 = ρ11 . For j > 1 we define aj ≡ (ρ1j /a1 )∗ . Then from the condition
that each row is a multiple of the top row, there must be numbers bj such that
|a1 |2 a1 a∗2 a1 a∗3 ...
ρ = b2 |a1 |2 b2 a1 a∗2 b2 a1 a∗3 . . .
b3 |a1 |2 b3 a1 a∗2 b3 a1 a∗3 . . .
But ρ is Hermitian, so
b2 |a1 |2 = b∗1 a∗1 a2 ⇒ b2 = a2 /a1
b3 |a1 |2 = b∗1 a∗1 a3 ⇒ b3 = a3 /a1
P
This completes the proof that ρij = ai a∗j , from which it follows that ρ = |ψihψ| with |ψi = i ai |ii.
6.14 Consider the rate of change of the expectation of the observable Q when the system is in an
impure state. This is
dQ X d
= pn hn|Q|ni, (6.11)
dt n
dt
where pn is the probability that the system is in the state |ni. By using Ehrenfest’s theorem to
evaluate the derivative on the right of (6.11), derive the equation of motion ih̄dQ/dt = Tr(ρ[Q, H]).
Soln: Using Ehrenfest to replace each derivative of hn|Q|ni we get
dQ 1 X
= pn hn|[Q, H]|ni, (6.12)
dt ih̄ n
But X X
Tr(ρ[Q, H]) = hn|iipi hi|[Q, H]|ni = pn hn|[Q, H]|ni.
n,i n
The result follows on substituting for hn|[Q, H]|ni in the previous equation.
6.15 Find the probability distribution (p1 , . . . , pn ) for n possible outcomes that maximises the
Shannon entropy. Hint: use a Lagrange multiplier.
Soln: Bearing in mind the requirement for the pi to be normalised, we set to zero each derivative
∂/∂pj of !
X X
S= pi ln pi + λ pi − 1
i i
and obtain
0 = ln pj − 1 + λ ⇒ pj = e1−λ .
Thus the pi are all equal, and from the normalisation condition they must equal 1/n.
6.16 Use Lagrange multipliers λ and β to P extremise the Shannon
P entropy of the probability dis-
tribution {pi } subject to the constraints (i) i pi = 1 and (ii) i pi Ei = U . Explain the physical
significance of your result.
Soln: We set to zero each derivative ∂/∂pj of
! !
X X X
S= pi ln pi + λ pi − 1 + β pi Ei − U
i i i
and obtain
0 = ln pj − 1 + λ + βEj ⇒ pj = e1−λ e−βEj .
P
We determine λ from the condition i pi = 1, so eλ−1 is the partition function Z. We determine
β from the condition that the expectation of the energy is U : β is the inverse temperature 1/kB T
because the smaller it is the higher U is. This calculation establishes that the Gibbs distribution
maximises the Shannon entropy of the system subject to the given constraints.
6.17 Explain why if at t = 0 the density operator of a system is given by the Gibbs distribution,
it remains so at later times.
40 Problems
Soln: The time derivative of ρ is proportional to [H, ρ]. But the Gibbs distribution makes ρ =
e−βH /Z a function of H, so it commutes with H.
6.18 A composite system is formed from uncorrelated subsystem A and subsystem B, both in
impure states. The numbers {pAi } are the probabilities of the members of the complete set of states
{|A; ii} for subsystem A, while the numbers {pBi } are the probabilities of the complete set of states
{|B; ii} for subsystem B. Show that the Shannon entropy of the composite system is the sum of the
Shannon entropies of its subsystems. What is the relevance of this result for thermodynamics?
Soln: The state |AB; iji of the composite system has probability pij = pAi pBj so the Shannon
entropy of the whole system is
X X X X X
s= pij ln(pij ) = pAi pBj ln(pAi ) + pAi pBj ln(pBj ) = pAi ln(pAi ) + pBj ln(pBj )
ij ij ij i j
= sA + sB
P
where we’ve used the normalisation conditions i pAi = 1, etc. This result establishes that entropy
is an “extensive” thermodynamic quantity: the entropy of two litres of water at 20◦ C is twice the
entropy of two one-litre bottles of water at the same temperature.
6.19 The |0i state of a qubit has energy 0, while the |1i state has energy ǫ. Show that when the
qubit is in thermodynamic equilibrium at temperature T = 1/(kB β) the internal energy of the qubit
is ǫ
U = βǫ . (6.13)
e +1
Show that when βǫ ≪ 1, U ≃ 21 ǫ, while for βǫ ≫ 1, U ≃ ǫe−βǫ . Interpret these results physically
and sketch the specific heat C = ∂U/∂T as a function of T .
Soln: The partition function is
Z = 1 + e−βǫ
so the internal energy is
∂ ln Z ǫe−βǫ ǫ
U =− = = βǫ .
∂β 1 + e−βǫ e +1
When βǫ ≪ 1 we can approximate eβǫ ≃ 1 + βǫ and then binomial-expand the denominator in U :
ǫ
U≃ ≃ 21 ǫ(1 − βǫ + · · ·)
2 + βǫ
When βǫ ≫ 1 we neglect unity in the denominator obtaining U ≃ ǫe−βǫ .
When βǫ ≪ 1 the system is hot and the qubit is nearly as likely to be in either of its two states,
so the expectation value of the energy is the average of 0 and ǫ. When βǫ ≫ 1, the system is cold
and the probability of being in the excited state is ∼ e−βǫ ≪ 1 so the expectation value of the energy
is this probability times ǫ.
6.20 Show that the time-evolution of the density operator leaves the Shannon entropy s = − Tr ρ log ρ
invariant.
Soln: Equation (6.91) shows that s depends only on the probabilities pi and the equation of
motion of ρ was derived on the assumption that the pi are constant and ρ evolves only because its
eigenvectors change.
6.21 Show that the partition function of a harmonic oscillator of natural frequency ω is
e−βh̄ω/2
Zho = . (6.14)
1 − e−βh̄ω
Hence show that when the oscillator is at temperature T = 1/(kB β) the oscillator’s internal energy
is
1 1
Uho = h̄ω 2 + βh̄ω . (6.15)
e −1
Interpret the factor (eβh̄ω − 1)−1 physically. Show that the specific heat C = ∂U/∂T is
eβh̄ω
C = kB βh̄ω (βh̄ω)2 . (6.16)
(e − 1)2
Problems 41
Show that lim T→0 C = 0 and obtain a simple expression for C when kB T ≫ h̄ω.
Soln: The partition function is
X∞ X∞
−(n+1/2)βh̄ω −βh̄ω/2 1
Z= e =e z n = e−βh̄ω/2
n=0 n=0
1 − z
where z ≡ e−βh̄ω . Differentiating we obtain
∂ ln Z ∂ 1 h̄ωz
U =− =− − 2 βh̄ω − ln(1 − z) = 21 h̄ω +
∂β ∂β 1−z
and the required result follows on dividing through by z.
The factor (eβh̄ω − 1)−1 is the mean number of excitations that the oscillator has. As β becomes
small (high T ) this number becomes large as expected. Differentiating this factor w.r.t. T we obtain
the required expression for C.
As T → 0, β → ∞ and we can neglect the unity in the denominator for C, so C/kB →
e−βh̄ω (βh̄ω)2 which goes to zero on account of the exponential. When kB T ≫ h̄ω, βh̄ω ≪ 1 and we
can Taylor expand the exponentials in C, finding
C 1 + βh̄ω + · · ·
≃ ≃1
kB (1 + βh̄ω/2 + · · ·)2
This is in accordance with the classical principle of equipartition under which the oscillator will have
1
2 kB T of both kinetic and potential energy.
6.22 A classical ideal monatomic gas has internal energy U = 23 N kB T and pressure P = N kB T /V,
where N is the number of molecules and V is the volume they occupy. From these relations, and
assuming that the entropy vanishes at zero temperature and volume, show that in general the entropy
is
S(T, V) = N kB ( 32 ln T + ln V). (6.17)
A removable wall divides a cylinder into equal parts of volume V. Initially the wall is in place and
each half contains N molecules of ideal monatomic gas at temperature T . The wall is removed. Show
that equation (6.17) implies that the entropy of the entire body of fluid increases by 2 ln 2 N kB . Can
this result be squared with the principle that dS = dQ/T , where dQ is the heat absorbed when the
change is made reversibly? What conclusion do you draw from this thought experiment?
Soln: Heating the gas up at constant volume (so T dS = dU = 23 N kB dT ) we conclude that
S(T, V ) = 23 N kB ln T + f (V ), where f is an arbitrary function. Expanding our system at constant
temperature (and therefore constant U ) we have 0 = T dS − P dV which implies dS = N kB dV /V ⇒
S = N kB ln V + g(T ). Using our earlier result to identify g, we obtain the required result.
Initially the total entropy is Si = 2N kB ( 32 ln T + ln V ). Since only V changes, it is finally
Sf = 2N kB [ 23 ln T + ln(2V )], and we obtain the required increase. This result is clearly incompatible
with dS = dQrev /T because no heat enters the cylinder as the wall is withdrawn. If some molecules
could be identified as being initially in the right end of the cylinder, and others as being in the left
end, there would be a loss of information on withdrawing the wall because with the wall out you
would be more uncertain where any particular molecule was. Quantum mechanics requires us to
treat all molecules as identical and denies the possibility of labelling them as right- or left-molecules.
In a classical universe, in which an ideal gas was possible and molecules could be labelled,
entropy would be associated with missing information, but not heat flow. In the real quantum
universe an ideal gas satisfies neither U = 32 N kB T nor P V = N kB T , molecules cannot be labelled
and entropy is both associated with heat flow and missing information.
6.23 Consider a ‘gas’ of M non-interacting, monatomic molecules of mass m that move in a
one-dimensional potential well V = 0 for |x| < a and ∞ otherwise. Assume that at sufficiently
low temperatures all molecules are either in the ground or first-excited states. Show that in this
approximation the partition function is given by
π 2 h̄2
ln Z = −M βE0 + e−3βE0 − e−3(M+1)βE0 where E0 ≡ . (6.18)
8ma2
42 Problems
Show that for M large the internal energy, pressure and specific heat of this gas are given by
2E0 9E02 −3βE0
U = E0 (M + 3e−3βE0 ) ; P = M + 3e−3βE0 ; CV = e . (6.19)
a kB T 2
In what respects do these results for a quantum ideal gas differ from the properties of a classical
ideal gas? Explain these differences physically.
Soln: The ground-state energy is the kinetic energy of a particle that has de Broglie wavelength
= 4a so it can have nodes at x = ±a. Hence k = 2π/4a = π/2a and E0 = (h̄k)2 /2m = π 2 h̄2 /8ma.
The energy of the first excited state is four times the energy of the ground state because for this
state k is twice as large. The first excited state may contain 0, 1, . . . , M molecules, so
Z = e−MβE0 +e−(M−1+4)βE0 +e−(M−2+8)βE0 +· · ·+e−4MβE0 = e−MβE0 1 + e−3βE0 + · · · + e−3MβE0
Summing this geometrical progression we have
1 − e−3(M+1)βE0
Z = e−MβE0
1 − e−3βE0
Taking logs and using ln(1 + x) ≃ x yields the required result. We use M ≫ 1 to simplify Z. Then
the internal energy
∂ ln Z
U =− = M E0 + 3E0 e−3βE0 .
∂β
The first term on the right is clearly the energy when every molecule is in the ground state, and the
second term must give the correction arising from molecules in the first excited state. Remarkably
this is not proportional to the number of molecules. If we counted states classically, we would
have Z = e−MβE0 + M e−(M+3)βE0 + · · · because there would be a different state of the gas for
each molecule we chose to promote to the excited state, and the correction would be proportional
to M . Differentiating U w.r.t. T using ∂/∂T = −(β/T )∂/∂β we obtain the specific heat given.
The Helmholtz free energy is F = −β −1 ln Z ≃ M E0 − β −1 e−3βE0 , so P = −∂F/∂a = −(M +
3e−3βE0 )(∂E0 /∂a), from which the required expression follows trivially.
The pressure does not go to zero with T , as does that of a classical gas, on account of zero-point
energy. The specific heat vanishes with T , as Nernst’s law requires, rather than being independent
of temperature, on account of the finite energy required to reach an excited state.
Problems
7.1 Show that hj, j|Jx |j, ji = hj, j|Jy |j, ji = 0 and that hj, j|(Jx2 + Jy2 )|j, ji = j. Discuss the
implications of these results for the uncertainty in the orientation of the classical angular momentum
vector J for both small and large values of j.
Soln: We have
hj, j|Jx |j, ji = hj, j| 12 (J+ + J− )|j, ji.
p
Also J+ |j, ji = 0 and J− |j, ji = j(j + 1) − j(j − 1)|j, j−1i. From this it follows that hj, j|Jx |j, ji =
0 and similarly for Jy = (J+ − J− )/2i. Further
hj, j|Jx2 |j, ji = 14 hj, j|(J+ J− + J− J+ )|j, ji
p p
= 41 hj, j|J+ J− |j, ji = 14 j(j + 1) − (j − 1)j j(j + 1) − j(j − 1) = 12 j
1/2
and similarly for Jy2 . Hence Jx2 + Jy2 = j 1/2 , so even when the angular momentum is maximally
1/2
aligned with the z axis, j of angular momentum lies in the xy plane; but we don’t know the
direction of this component of J.
1/2 √
The angle between J and the z axis is θ = arctan Jx2 + Jy2 /Jz = arctan(1/ j). Thus
in the maximally aligned state J is√misaligned with the z axis by an angle θ that decreases from
55◦ ≃ 1 radian when j = 12 to ∼ 1/ j radians for large j.
7.2 In the rotation spectrum of 12 C16 O the line arising from the transition l = 4 → 3 is at
461.04077 GHz, while that arising from l = 36 → 35 is at 4115.6055 GHz. Show from these data that
in a non-rotating CO molecule the intra-nuclear distance is s ≃ 0.113 nm, and that the electrons
provide a spring between the nuclei that has force constant ∼ 1904 N m−1 . Hence show that the
Problems 43
vibrational frequency of CO should lie near 6.47 × 1013 Hz (measured value is 6.43 × 1013 Hz). Hint:
show from classical mechanics that the distance of O from the centre of mass is 37 s and that the
molecule’s moment of inertia is 48 2
7 mp s . Recall also the classical relation L = Iω.
Soln: The energy levels of a diatomic molecule are El = l(l + 1)h̄2 /2I, where I is the moment of
inertia about the centre of mass. Since l(l + 1) − (l − 1)l = 2l, the energy associated with l → l − 1
is ∆El = lh̄2 /I and I = lh̄2 /∆El = lh̄/(2πνl ), where νl is the measured frequency. From classical
2 2
physics I = mp (16rO +12rC ), where rO is the distance of O from the centre of mass, so 16rO = 12rC .
Consequently, rO = 7 s, where s is the interatomic distance, and I = 16mp 73 s2 . We have finally
3
7
s2 = 48 lh̄/(2πmp νl ), which yields s = 0.112678 nm from l = 4 and 0.113139 nm from l = 36.
The force that has caused this increase in length is the change from l = 4 to l = 36 in
l(l + 1)h̄2 l(l + 1)h̄2
F = MO rO ω 2 = MO rO = ,
I 2 16mp 73 s3
where the classical relation L = Iω has been used. Numerically
F36 − F4 = (8.9182 − 0.135559) × 10−10 N.
The spring constant is
δF
k= = 1.904 kN m−1
δs
During an oscillation, the displacement of the O is only 37 δs, so the frequency of its motion is given
by (2πν)2 = 73 k/mO , which yields ν = 6.4750 × 1013 Hz.
7.3 Show that Li commutes with x · p and thus also with scalar functions of x and p.
Soln: X X X
[Li , x · p] = [Li , xj pj ] = ([Li , xj ]pj + xj [Li , pj ]) = i ǫijk (xk pj + xj pk )
j j jk
P
Now for any symmetric matrix Sik = Ski , ik ǫijk Sik = 0 because when we swap ik on both ǫ and
S we obtain X X
ǫijk Sik = − ǫkji Ski
ik ik
P P P
but by renaming k i and i k we can show that ik ǫkji Ski = ik ǫijk Sik . Hence ik ǫijk Sik is equal
to minus itself so must vanish. Applying this result with Sjk = (xk pj + xj pk ) completes the proof
that Li commutes with the given dot product. It is easy to see that this demonstration extends to
the dot product of any two vector operators.
7.4 Write down the expression for the commutator [σi , σj ] of two Pauli matrices. Show that the
anticommutator of two Pauli matrices is
{σi , σj } = 2δij . (7.1)
Soln: Since σi = 2si is twice the spin operator of a spin-half particle, from the angular-momentum
commutation relations we have [σi , σj ] = 2iǫijk σk .
Calculating explicitly
0 1 0 1 1 0
σx2 = =
1 0 1 0 0 1
and similarly σy2 = σz2 = I so {σi , σi } = 2I as required. Further
0 1 0 −i 0 −i 0 1 i 0 −i 0
{σx , σy } = + = + =0
1 0 i 0 i 0 1 0 0 −i 0 i
and similarly for {σy , σz } and {σx , σz }.
7.5 Let n be any unit vector and σ = (σx , σy , σz ) be the vector whose components are the Pauli
matrices. Why is it physically necessary that n · σ satisfy (n · σ)2 = I, where I is the 2 × 2 identity
matrix? Let m be a unit vector such that m · n = 0. Why do we require that the commutator
[m · σ, n · σ] = 2i(m × n) · σ? Prove that these relations follow from the algebraic properties of the
44 Problems
Pauli matrices. You should be able to show that [m · σ, n · σ] = 2i(m × n) · σ for any two vectors n
and m.
Soln: n · σ is twice the spin operator for the direction n. Since all directions are equivalent, it
must satisfy the same equations that σz does, and σz2 = I. If m is perpendicular to n, we could take
the x-axis to lie along m, the y axis to lie along n and then the z-axis would lie along m × n. In
this system, [m · σ, n · σ] = [σx , σy ] = 2iσz = 2i(m × n) · σ.
Given arbitrary vectors m and n
X X X
[m · σ, n · σ] = mi nj [σi , σj ] = mi nj 2iǫijk σk = 2i (m × n)k σk = 2i(m × n) · σ.
ij ijk k
7.6 Let n be the unit vector in the direction with polar coordinates (θ, φ). Write down the matrix n·
σ and find its eigenvectors. Hence show that the state of a spin-half particle in which a measurement
of the component of spin along n is certain to yield 21 h̄ is
|+, ni = sin(θ/2) eiφ/2 |−i + cos(θ/2) e−iφ/2 |+i, (7.2)
where |±i are the states in which ± 21 is obtained when sz is measured. Obtain the corresponding
expression for |−, ni. Explain physically why the amplitudes in (7.2) have modulus 2−1/2 when
θ = π/2 and why one of the amplitudes vanishes when θ = π.
Soln:
n · σ = sin θ cos φ σx + sin θ sin φ σy + cos θ σz
cos θ sin θe−iφ
=
sin θeiφ − cos θ
We know it’s eigenvalues are λ = ±1. The eigenvector for λ = 1 satisfies
cos θ sin θe−iφ a a
= ⇒ b sin θe−iφ = a(1 − cos θ)
sin θeiφ − cos θ b b
Using 1 − cos θ = 2 sin2 θ/2 and sin θ = 2 sin(θ/2) cos(θ/2)
b cos(θ/2)e−iφ = a sin(θ/2)
It’s easy to see that this equation and the normalisation condition a2 + b2 = 1 is satisfied by
cos(θ/2)e−iφ/2
|+i = = cos(θ/2)e−iφ/2 |+i + sin(θ/2)eiφ/2 |−i
sin(θ/2)eiφ/2
This is the state in which a measurement of sz is guaranteed to return + 21 . Similarly we find that
the eigenvector for λ = −1
sin(θ/2)e−iφ/2
|−i = = sin(θ/2)e−iφ/2 |+i − cos(θ/2)eiφ/2 |−i
− cos(θ/2)eiφ/2
When θ = π/2 n lies in the xy plane, so by symmetry when in this state the probabilities of
measuring either 12 or − 12 must be equal. Hence the amplitudes in its expansion in terms of |±i
have to have modulus 2−1/2 . When θ = π, n = −êz so |+, ni = |−i and the coefficient of |+i has
to vanish.
7.7 For a spin-half particle at rest, the rotation operator J is equal to the spin operator S. Use
the result of Problem 7.4 to show that in this case the rotation operator U (α) ≡ exp(−iα · J) is
α α
U (α) = I cos − iα̂ · σ sin , (7.3)
2 2
where α̂ is the unit vector parallel to α. Comment on the value this gives for U (α) when α = 2π.
Soln:
(α · σ/2)2 (α · σ/2)3 (α · σ/2)4
U (α) = exp(− 21 iα · σ) = I − iα · σ/2 − +i + + ···
2! 3! 4!
Problems 45
Using α = αα̂ and (α̂ · σ)2n = I, and gathering real and imaginary parts, we have
(α/2)2 (α/2)4 (α/2)3
U (α) = 1 − + + · · · I − i (α/2) − + · · · (α̂ · σ).
2! 4! 3!
The required result now follows from the power series for cos θ and sin θ. For α = 2π we have rotated
the system right around but the value of the wavefunction has changed sign because U = −I.
7.8 Write down the 3 × 3 matrix that represents Sx for a spin-one system in the basis in which Sz
is diagonal (i.e., the basis states are |0i and |±i with Sz |+i = |+i, etc.)
A beam of spin-one particles emerges from an oven and enters a Stern–Gerlach filter that passes
only particles with Jz = h̄. On exiting this filter, the beam enters a second filter that passes only
particles with Jx = h̄, and then finally it encounters a filter that passes only particles with Jz = −h̄.
What fraction of the particles stagger right through?
Soln: Using Sx = 12 (S+ + S− ) we can readily show that
h+|Sx |+i h+|Sx |0i h+|Sx |−i 0 1 0
h0|Sx |+i h0|Sx |0i h0|Sx |−i = √1 1 0 1 .
2
h−|Sx |+i h−|Sx |0i h−|Sx |−i 0 1 0
Since all quantum states have equal a priori probability, an atom emerging from the oven is equally
likely to be in any one of the three eigenstates of Jz , so a third of the particles pass the first filter.
Those that do are in the state | + zi. The amplitude that they pass the second filter is h+x| + zi.
This number is the complex conjugate of the three components a ≡ h+z| + xi, b ≡ h0z| + xi and
c ≡ h−z| + xi of | + xi in the basis that makes Sz diagonal. We write the eigenvalue equation
Sx | + xi = | + xi in this basis
0 1 0 a a √
1
√ 1 0 1 b = b ⇒ a = c = b/ 2
2
0 1 0 c c
But a2 + b2 + c2 = 1, so a = 12 . Hence a2 = 41 of the particles that passed the first filter get past the
second filter. By symmetry a quarter of these pass the third filter. So the fraction of particles that
get straight through is 31 × 41 × 14 = 1/48.
7.9∗ Repeat the analysis of Problem 7.8 for spin-one particles coming on filters aligned successively
along +z, 45◦ from z towards x [i.e. along (1,0,1)], and along x.
Use classical electromagnetic theory to determine the outcome in the case that the spin-one
particles were photons and the filters were Polaroid. Why do you get a different answer?
Soln: We√adapt the calculation of Problem 7.8 by replacing the matrix for Jx by that for n · J =
(Jx + Jz )/ 2. So if now (a, b, c) is | + ni in the usual basis, we have
−1/2 1 b
2 0 a a
a = 2 − √2
2
12 0 1
2
b = b ⇒
1 −1/2
c= b
0 2 −2 c c √
2+ 2
√
The normalisation yields b = 12 , so a = 21 /(2 − 2) and the required probability is the square of this,
√
0.25/(6 − 4 2) ≃ 0.73. So the probability of getting through all three filters is 31 × (0.73)2 ≃ 0.177.
In electromagnetism just one of two polarisations gets through the first filter, so we must say
that a photon has a probability of half of passing the first filter. Then we resolve
√ its E field along the
direction of the second filter and find that the amplitude of E falls by 1/ 2 on passing the second
filter, so half the energy and therefore photons that pass the first filter pass the second. Of these
just a half pass the third filter. Hence in total 81 = 0.125 of the photons get right through.
Although photons are spin-one particles, there are two major difference between the two cases.
Most obviously, polaroid selects for linear polarisation rather than circular polarisation, and a photon
with well-defined angular momentum is circularly polarised. The other difference is that a photon
can be in the state | + zi or | − zi but not the state |0zi, where the z-axis is parallel to the photon’s
motion. This fact arises because emag waves are transverse so they do not drive motion in the
direction of propagation k; an angular momentum vector perpendicular to k would require motion
46 Problems
along k. Our theory does not allow for this case because it is non-relativistic, whereas a photon,
having zero rest mass, is an inherently relativistic object; we cannot transform to a frame in which
a photon is at rest so all three directions would be equivalent.
√
7.10 A system that has spin momentum 6h̄ is rotated through an angle φ around the z-axis.
Write down the 5 × 5 matrix that updates the amplitudes am that Sz will take the value m.
Soln: We need the matrix of U (φ) = exp(−iφSz ) when squeezed between the states |2, mi, i.e.,
−2iφ
e 0 0 0 0
−iφ
0 e 0 0 0
U (φ) = 0 0 1 0 0
iφ
0 0 0 e 0
2iφ
0 0 0 0 e
7.11 Justify physically the claim that the Hamiltonian of a particle that precesses in a magnetic
field B can be written
H = −2µs · B. (7.4)
In a coordinate system oriented such that the z-axis is parallel to B, a proton is initially in the
eigenstate |+, xi of sx . Obtain expressions for the expectation values of sx and sy at later times.
Explain the physical content of your expressions.
Bearing in mind that a rotating magnetic field must be a source of radiation, do you expect
your expressions to remain valid to arbitrarily late times? What really happens in the long run?
Soln: Eq. (7.4) is just the classical expression for the energy of a magnetic moment µ = 2µs in a
magnetic field – the minus sign ensures that the lowest energy is achieved when the dipole is aligned
with B.
With B = Bez the stationary states of H are clearly the eigenkets |±i of sz and they have
energies E± = ∓µB = ∓h̄ω, where ω = µB/h̄, so the time evolution of an arbitrary state is given
by
|ψ, ti = ae−iωt |−i + beiωt |+i,
where a, b are set by the initial conditions. Since |ψ, 0i is an eigenket of sx , we have
1 0 1 a a
sx |ψ, 0i = 2 |ψ, 0i ⇒ =
1 0 b b
√ 1
from which we easily show that a = b = 1/ 2. Writing sx = 2 (s+ + s− ) in terms of the ladder
operators, the required expectation value is
hsx i = √12 eiωt h−| + e−iωt h+| 12 (s+ + s− ) √12 e−iωt |−i + eiωt |+i
= 41 eiωt h−| + e−iωt h+| e−iωt |+i + eiωt |−i
1
= 2 cos(2ωt)
Similarly 1
hsy i = √1
2 eiωt h−| + e−iωt h+| 2i (s+ − s− ) √12 e−iωt |−i + eiωt |+i
= 1
4i eiωt h−| + e−iωt h+| e−iωt |+i − eiωt |−i
= − 12 sin(2ωt)
These expectation values are what we would expect if the spin vector rotated in the xy plane with
angular velocity −2ω.
The rotating magnetic dipole will radiate e.m. waves at frequency 2ω and thus lose energy. This
loss of energy must cause the amplitude b to be in the lower -energy state to increase, so eventually
the proton will settle to the state |+i.
7.12 Show that a classical top with spin angular momentum S which is subject to a torque G =
µS × B/|S| precesses at angular velocity ω = µB/|S|. Explain the relevance of this calculation to
magnetic resonance imaging in general and equation (7.70b) in particular.
Soln: Since torque is equal to rate of change of angular momentum, we have
dS
= µS × B/|S|.
dt
Problems 47
When a vector S is rotated with angular velocity ω its rate of change is (Box 4.3)
dS
= ω × S.
dt
Comparing these equations we see that the torque rotates S at a rate ω = −µB/|S|.
The relevance to NMR is that classically the magnetic field would apply a torque µp S × B/|S|
to the proton, and for the proton |S| = 21 h̄, so classical physics predicts precession of the proton’s
spin at the angular frequency given by equation (7.70b).
7.13∗ Write a computer program that determines the amplitudes am in
s
X
|n; s, si = am |s, mi
m=−s
where n = (sin θ, 0, cos θ) with θ any angle and |n; s, si is the ket that solves the equation (n ·
S)|n; s, si = s|n; s, si. Explain physically the nature of this state.
Use your am to evaluate the expectation values hSx i and Sx2 for this state and hence show
p
that the rms fluctuation in measurements of Sx will be s/2 cos θ.
Soln: We use a routine tridiag() that computes the e-values and e-kets of a real symmetric
tri-diagonal matrix – the routine tqli() in Numerical Recipies by Press et al. is suitable.
#define J 100
#define NT 3
double tridiag(double*,double*,int,double**)// evaluates & ekets of real,
// symmetric tridiagonal matrix
double alphap(int j,int m){
if(m>=j)return 0;
return sqrt((double)(j*(j+1)-m*(m+1)));
}
double alpham(int j,int m){
if(m<=-j) return 0;
return sqrt((double)(j*(j+1)-m*(m-1)));
}
void expect(double *a,int j,double st){//evaluate <Sx> and <Sx2 >
double s1=0,s2=0;
for(int n=-j;n<=j;n++){
int nm2=n-2,nm1=n-1,np1=n+1,np2=n+2;
if(nm2>=-j) s2+=alpham(j,n)*alpham(j,nm1)*a[nm2]*a[n];
if(np2<=j) s2+=alphap(j,n)*alphap(j,np1)*a[np2]*a[n];
s2+=(alphap(j,nm1)*alpham(j,n)+alpham(j,np1)*alphap(j,n))*pow(a[n],2);
if(nm1>=-j) s1+=alpham(j,n)*a[nm1]*a[n];
if(np1<=j) s1+=alphap(j,n)*a[np1]*a[n];
}
s1*=.5; s2*=.25;
printf("%f %f %f %f\n",s1,j*st,s2,.5*j*(1-st*st)+pow(j*st,2));
}
int main(void){
double pi=acos(-1),theta[NT]={80, 120, 30};
double *D = new double[2*J+1];
double *E = new double[2*J+1];
double **Z = new double*[2*J+1];//allocate storage for square matrix
for(int i=0; i<2*J+1; i++) Z[i] = new double[2*J+1];
for(int it=0; it<3; it++){
theta[it]=theta[it]*pi/180;
double ct=cos(theta[it]), st=sin(theta[it]);
for(int m=-J; m<=J; m++){
D[J+m]=m*ct;//diagonal elements of matrix
if(m>-J) E[J+m]=st*.5*alpham(J,m);//sub-diagonal elements
}
tridiag(D,E,2*J+1,Z);//finds evalues & ekets of tridiagonal matrix
48 Problems
int mm;
for(int i=0; i<2*J+1; i++){
if(fabs(D[i]-J)<.05) mm=i; // identify eket m=J
}
expect(Z[mm]+J,J,st);
}
}
7.14∗ We have that ∂ ∂
L+ ≡ Lx + iLy = eiφ + i cot θ . (7.5)
∂θ ∂φ
From the Hermitian nature of Lz = −i∂/∂φ we infer that derivative operators are anti-Hermitian.
So using the rule (AB)† = B † A† on equation (7.5), we infer that
∂ ∂
L− ≡ L†+ = − +i cot θ e−iφ .
∂θ ∂φ
This
R argument
R and the result it leads to is wrong. Obtain the correct result by integrating by parts
dθ sin θ dφ (f ∗ L+ g), where f and g are arbitrary functions of θ and φ. What is the fallacy in
the given argument?
Soln: Z Z Z Z
∗ ∗ iφ ∂g ∂g
dθ sin θ dφ (f L+ g) = dθ sin θ dφ f e + i cot θ
∂θ ∂φ
Z Z Z Z
iφ ∗ ∂g ∂g
= dφ e dθ sin θf + i dθ cos θ dφ f ∗ eiφ
∂θ ∂φ
Z Z
∂(sin θf ∗ )
= dφ eiφ [sin θ f ∗ g] − dθ g
∂θ
Z Z
∗ iφ ∂(f ∗ eiφ )
+ i dθ cos θ [f e g] − dφ g
∂φ
The square brackets vanish so long f, g are periodic in φ. Differentiating out the products we get
Z Z Z Z Z
∗ iφ ∂f ∗ ∗
dθ sin θ dφ (f L+ g) = − dφ e dθ sin θg + dθ cos θgf
∂θ
Z Z ∗ Z
iφ ∂f iφ ∗
− i dθ cos θ dφ e g + i dφ e gf
∂φ
The two integrals containing f ∗ g cancel as required leaving us with
Z Z Z Z ∗ Z Z
∗ iφ ∂f ∂f ∗
dθ sin θ dφ (f L+ g) = − dθ sin θ dφ ge + i cot θ = dθ sin θ dφ g(L− f )∗
∂θ ∂φ
where ∂ ∂
L− = −e−iφ − i cot θ .
∂θ ∂φ
The fallacy is the proposition that ∂/∂θ is anti-Hermitian: the inclusion of the factor sin θ in the
integral prevents this being so.
P
7.15∗ By writing h̄2 L2 = (x × p) · (x × p) = ijklm ǫijk xj pk ǫilm xl pm show that
h̄2 L2 1
p2 = 2
+ 2 (r · p)2 − ih̄r · p . (7.6)
r r
By showing that p · r̂ − r̂ · p = −2ih̄/r, obtain r · p = rpr + ih̄. Hence obtain
h̄2 L2
p2 = p2r + . (7.7)
r2
Give a physical interpretation of one over 2m times this equation.
Problems 49
Soln: From the formula for the product of two epsilon symbols we have
X
h̄2 L2 = (δjl δkm − δjm δkl )xj pk xl pm
jklm
X
= xj pk xj pk − xj pk xk pj .
jk
The first term is X X X
xj pk xj pk = xj (xj pk + [pk , xj ])pk = xj (xj pk − ih̄δjk )pk
jk jk jk
= r2 p2 − ih̄r · p.
The second term is X X
xj pk xk pj = xj (xk pk − ih̄)pj
jk jk
X X
= xj (pj xk pk + ih̄δjk pk ) − 3ih̄ xj pj
jk j
which establishes that Al |E, li ∝ |E, l + 1i is a state with the same energy but more angular
momentum. To see whether there is an upper limit on the angular momentum, we evaluate
2
|Al |E, li| = hE, l|A†l Al |E, li = E > 0
so there is no limit to the angular momentum. Physically, when there is no confining potential, with
a given energy the particle can move at a given speed along a path that is as far from the origin as
it pleases. Hence its angular momentum is not constrained by its energy.
P
7.20∗ Show that [Ji , Lj ] = i k ǫijk Lk and [Ji , L2 ] = 0 by eliminating Li using its definition
L = h̄−1 x × p, and then using the commutators of Ji with x and p.
Soln:
h̄[Ji , Lj ] = ǫjkl [Ji , xk pl ] = ǫjkl ([Ji , xk ]pl + xk [Ji , pl ])
= ǫjkl (iǫikm xm pl + iǫiln xk pn ) = i(ǫklj ǫkmi xm pl + ǫljk ǫlni xk pn )
= i(δlm δji − δli δjm )xm pl + i(δjn δki − δji δkn )xk pn
= i(x · pδij − xj pi + xi pj − x · pδij ) = i(xi pj − xj pi )
But
ih̄ǫijk Lk = iǫijk ǫklm xl pm = iǫkij ǫklm xl pm = i(δil δjm − δim δjl )xl pm = i(xi pj − xj pi )
7.21∗ In this problem you show that many matrix elements of the position operator x vanish when
states of well-defined l, m are used as basis states.
P These results will lead to selection rules for electric
dipole radiation. First show that [L2 , xi ] = i jk ǫjik (Lj xk + xk Lj ). Then show that L · x = 0 and
using this result derive
X
[L2 , [L2 , xi ]] = i ǫjik Lj [L2 , xk ] + [L2 , xk ]Lj = 2(L2 xi + xi L2 ). (7.10)
jk
By squeezing this equation between angular-momentum eigenstates hl, m| and |l′ , m′ i show that
0 = (β − β ′ )2 − 2(β + β ′ ) hl, m|xi |l′ , m′ i,
where β ≡ l(l + 1) and β ′ = l′ (l′ + 1). By equating the factor in front of hl, m|xi |l′ , m′ i to zero,
and treating the resulting equation as a quadratic equation for β given β ′ , show that hl, m|xi |l′ , m′ i
must vanish unless l + l′ = 0 or l = l′ ± 1. Explain why the matrix element must also vanish when
l = l′ = 0.
Soln: X X X
[L2j , xi ] = (Lj [Lj , xi ] + [Lj , xi ]Lj ) = i ǫjik (Lj xk + xk Lj )
j j jk
X X X
h̄L · x = ǫijk xj pk xi = ǫijk (xj xi pk + xj [pk , xi ]) = ǫijk (xj xi pk − ih̄xj δki )
ijk ijk ijk
P
Both terms on the right side of this expression involve ik ǫijk Sik where Sik = Ski so they vanish
by Problem 7.3. Hence x · L = 0 as in classical physics.
Now
X X
[L2 , [L2 , xi ]] = i ǫjik [L2 , (Lj xk + xk Lj )] = i ǫjik (Lj [L2 , xk ] + [L2 , xk ]Lj )
jk jk
X
=− ǫjik ǫlkm (Lj {Ll xm + xm Ll } + {Ll xm + xm Ll }Lj )
jklm
X
=− (δjm δil − δjl δim )(Lj {Ll xm + xm Ll } + {Ll xm + xm Ll }Lj )
jlm
X
=− (Lj {Li xj + xj Li } + {Li xj + xj Li }Lj − Lj {Lj xi + xi Lj } − {Lj xi + xi Lj }Lj )
j
X X
=− (Lj Li xj + xj Li Lj ) − L2 xi − (Lj xi Lj + Lj xi Lj ) − xi L2
j j
52 Problems
where to obtain the last line we have identified occurrences of L · x and x · L. Now
X X X
Lj Li xj = (Lj xj Li + Lj [Li , xj ]) = i ǫijk Lj xk
j j jk
P P
Similarly, j xj Li Lj = i jk ǫjik xk Lj . Moreover
X X X
Lj xi Lj = ([Lj , xi ]Lj + xi Lj Lj ) = i ǫjik xk Lj + xi L2
j j jk
X X
= (Lj [xi , Lj ] + Lj Lj xi ) = i ǫijk Lj xk + L2 xi
j jk
7.23∗ Let
1 (l + 1)h̄
Al ≡ √ ipr − + mωr . (7.11)
2mh̄ω r
be the ladder operator of the three-dimensional harmonic oscillator and |E, li be the stationary state
of the oscillator that has energy E and angular-momentum
√ quantum number l. Show that if we
write Al |E, li = α− |E − h̄ω, l + 1i, then α− = L − l, where L is the angular-momentum quantum
number of a circular orbit of energy E. Show similarly that if A†l−1 |E, li = α+ |E + h̄ω, l − 1i, then
√
α+ = L − l + 2.
Soln: Taking the mod-square of each side of Al |E, li = α− |E − h̄ω, l + 1i we find
2 † Hl 3 E
|α− | = hE, l|Al Al |E, li = hE, L| − (l + 2 ) |E, li = − (l + 23 ).
h̄ω h̄ω
3
In the case l = L, |α− |2 = 0, so L = (E/h̄ω) − 2 and therefore |α− |2 = L − l as required. We can
choose the phase of α− at our convenience.
Similarly
α2+ = hE, l|Al−1 A†l−1 |E, li = hE, l|(A†l−1 Al−1 + [Al−1 , A†l−1 ])|E, li
Hl−1 Hl − Hl−1 E
= hE, l| − (l + 12 ) + + 1 |E, li = −l+ 1
2 =L−l+2
h̄ω h̄ω h̄ω
7.24∗ Show that the probability distribution in radius of a particle that orbits in the three-
dimensional harmonic oscillator
p potential on a circular orbit with angular-momentum quantum
number l peaks at r/ℓ = 2(l + 1), where
r
h̄
ℓ≡ . (7.12)
2mω
Derive the corresponding classical result.
Soln: The radial wavefunctions of circular orbits are annihilated by Al , so Al |E, li = 0. In the
position representation this is
∂ 1 l+1 r
+ − + 2 u(r) = 0
∂r r r 2ℓ
Using the integrating factor,
Z
l r
exp dr − + 2 = r−l exp r2 /4ℓ2 , (7.13)
r 2ℓ
2 2 2
/2ℓ2
to solve the equation, we have u ∝ rl e−r /4ℓ . The radial distribution is P (r) ∝ r2 |u|2 = r2(l+1) e−r .
Differentiating to find the maximum, we have
√
2(l + 1)r2l+1 − r2(l+1) r/ℓ2 = 0 ⇒ r = 2(l + 1)1/2 a
For the classical result we have
mv 2 lh̄
mrv = lh̄ and = mω 2 r ⇒ r = v/ω =
r mrω
so r = (lh̄/mω)1/2 = (2l)1/2 ℓ in agreement with the QM result when l ≫ 1.
7.25∗ A particle moves in the three-dimensional harmonic oscillator potential with the second
largest angular-momentum quantum number possible at its energy. Show that the radial wavefunc-
tion is r
l 2l + 1 −x2 /4 h̄
u1 ∝ x x − e where x ≡ r/ℓ with ℓ ≡ . (7.14)
x 2mω
How many radial nodes does this wavefunction have?
54 Problems
Soln: From Problem 7.24 we have that the wavefunction of the circular orbit with angular mo-
2 2
mentum l is hr|E, li ∝ rl e−r /4ℓ . So the required radial wavefunction is
hr|E + h̄ω, l − 1i ∝ hr|A†l−1 |E, li
∂ l+1 r 2 2 rl+1 rl+1 2 2
∝ − − + 2 rl e−r /4ℓ = −lrl−1 + 2 − (l + 1)rl−1 + 2 e−r /4ℓ
∂r r 2ℓ 2ℓ 2ℓ
2 2 r 2l + 1 2 2l + 1
= rl e−r /4ℓ − ∝ xl e−x /4 x −
ℓ2 r x
√
This wavefunction clearly has one node at x = 2l + 1.
7.26 A box containing two spin-one gyros A and B is found to have angular-momentum quantum
numbers j = 2, m = 1. Determine the probabilities that when Jz is measured for gyro A, the values
m = ±1 and 0 will be obtained.
What is the value of the Clebsch–Gordan coefficient C(2, 1; 1, 1, 1, 0)?
Soln: We have to add two spin-one systems. We know that |2, 2i = |1i|1i and we create |2, 1i from
this by applying J− = J1− + J2− :
p p
J− |2, 2i = 2(2 + 1) − 2(2 − 1)|2, 1i = (J1− |1i)|1i+|1i(J2− |1i) = 1(1 + 1) − 1(1 − 1) (|0i|1i + |1i|0i)
i.e.,
1
|2, 1i = √ (|0i|1i + |1i|0i)
2
1
It follows that PA (1) = PA (0) = 2 , while PA (−1) = 0. From the above equation and the definition
of the Clebsch–Gordan coefficient
√
C(2, 1; 1, 1, 1, 0) = 1/ 2.
7.27 The angular momentum of a hydrogen atom in its ground state is entirely due to the spins of
the electron and proton. The atom is in the state |1, 0i in which it has one unit of angular momentum
but none of it is parallel to the z-axis. Express this state as a linear combination of products of the
spin states |±, ei and |±, pi of the proton and electron. Show that the states |x±, ei in which the
electron has well-defined spin along the x-axis are
1
|x±, ei = √ (|+, ei ± |−, ei) . (7.15)
2
By writing
|1, 0i = |x+, eihx+, e|1, 0i + |x−, eihx−, e|1, 0i, (7.16)
express |1, 0i as a linear combination of the products |x±, ei|x±, pi. Explain the physical significance
of your result.
Soln: We obtain the state |1, 0i by applying the ladder operator s− = s−,e + s−,p to |1, 1i =
|+, ei|+, pi. We easily find
1
|1, 0i = √ (|−, ei|+, pi + |+, ei|−, pi)
2
The states of a spin-half particle that have well-defined spin on the x-axis are the eigenkets of sx .
Writing |x+i = a+ |+i + a− |−i we have that (a+ , a− ) is the eigenvector of the Pauli matrix σx with
eigenvalue +1:
0 1 a+ a+
=
1 0 a− a−
√ √
It follows that a+ = a− = 1/ 2. Similarly, we find that |x−i = (|+i − |−i)/ 2. We can now
straightforwardly calculate
hx+, e|1, 0i = 12 (h+, e| + h−, e|) (|−, ei|+, pi + |+, ei|−, pi)
= 12 (|−, pi + |+, pi).
Problems 55
(1) (2)
S+ S− = (Sx(1) + iSy(1) )(Sx(2) − iSy(2) ) = Sx(1) Sx(2) + Sy(1) Sy(2) + i(Sy(1) Sx(2) − Sx(1) Sy(2) )
Similarly,
(1) (2)
S− S+ = Sx(1) Sx(2) + Sy(1) Sy(2) − i(Sy(1) Sx(2) − Sx(1) Sy(2) )
Adding these expressions we obtain the desired relation.
We choose to orient the z-axis along a. Then H becomes
K 1 (1) (2) (1) (2) (1) (2) (1) (2)
H= (S S + S S ) + S S − 3S S . (7.18)
a 2 + − − + z z z z
The eigenkets of S 2 and Sz are the three spin-one kets |1, 1i, |1, 0i and |1, −1i and the single spin-zero
ket |0, 0i. We multiply each of these kets in turn by H:
K 1 (1) (2) (1) (2)
H|1, 1i = H|+i|+i = 2 (S+ S− + S− S+ ) − 2Sz(1) Sz(2) |+i|+i
a
K
= − |1, 1i
2a
(i)
which uses the fact that S+ |+i = 0. Similarly H|1, −1i = H|−i|−i = −(K/2a)|1, −1i.
1 K 1 (1) (2) (1) (2) (1) (2)
H|1, 0i = H √ (|+i|−i + |−i|+i) = √ (S + S − + S − S + ) − 2S z S z (|+i|−i + |−i|+i)
2 2a 2
K 1 K
=√ 2 + 1 (|+i|−i + |−i|+i) = a |1, 0i
2a
where we have used S+ |−i = |+i, etc. Finally
1 K 1 (1) (2) (1) (2) (1) (2)
H|0, 0i = H √ (|+i|−i − |−i|+i) = √ (S + S − + S − S + ) − 2S z S z (|+i|−i − |−i|+i)
2 2a 2
K
=√ − 21 + 21 (|+i|−i − |−i|+i) = 0
2a
The given initial condition
1
|ψi = |+i|−i = √ (|1, 0i + |0, 0i),
2
which is a superposition of two stationary states of energies that differ by K/a. By analogy with the
symmetrical-well problem, we argue that after time πh̄/∆E = πh̄a/K the particle spins will have
reversed.
56 Problems
Problems
8.1 Some things about hydrogen’s gross structure that it’s important to know (ignore spin through-
out):
(a) What quantum numbers characterise stationary states of hydrogen?
(b) What combinations of values of these numbers are permitted?
(c) Give the formula for the energy of a stationary state in terms of the Rydberg R. What is the
value of R in eV?
(d) How many stationary states are there in the first excited level and in the second excited level?
(e) What is the wavefunction of the ground state?
(f) Write down an expression for the mass of the reduced particle.
(g) The wavefunction hx|ni of any state with principal quantum number n contains an exponential
in r = |x|. Write down the scale-length of this exponential in terms of the Bohr radius a0 .
(h) We can apply hydrogenic formulae to any two charged particles that are electrostatically bound.
How does the ground-state energy then scale with (i) the mass of the reduced particle, and (ii)
the charge Ze on the nucleus? (iii) How does the radial scale of the system scale with Z?
Soln:
(a) n principal quantum number, l angular-momentum number, m magnetic quantum number
(b) −l ≤ m ≤ l, l < n, n > 0
(c) En = −R/n2 ; R = 13.6 eV
(d) with n = 2 have l = 0 (1 state) and l = 1 (3 states) so 4 states in all. For n = 3 have three
equivalent states plus 5 states with l = 2, so 9 states in all.
(e) Ae−r/a0
(f) µ = me mp /(me + mp ) ≃ me
(g) na0
(h) (i) E ∝ µR (ii) E ∝ Z 2 R (iii) a ∝ Z −1 .
8.2 Show, by induction or otherwise, that there are n2 stationary states of hydrogen with energy
E = −R/n2 .
Soln: The difference N (n + 1) − N (n) between the number of states in the n + 1 and n levels is
2L + 1, where L = n is the extra angular momentum quantum number. So N (n + 1) − N (n) =
2n + 1 = (n + 1)2 − n2 . Thus if N (n) = n2 for some given n, it holds for n + 1. But it certainly
holds for n = 1 and the proof is complete.
8.3 In the Bohr atom, electrons move on classical circular orbits that have angular momenta lh̄,
where l = 1, 2, . . . Show that the radius of the first Bohr orbit is a0 and that the model predicts the
correct energy spectrum. In fact the ground state of hydrogen has zero angular momentum. Why
did Bohr get correct answers from an incorrect hypothesis?
Soln:
mv 2 e2
= and lh̄ = mvr
r 4πǫ0 r2
Eliminating v
l2 h̄2 e2 4πǫ0 l2 h̄2
= ⇒ r = = l 2 a0
mr3 4πǫ0 r2 me2
as required. By the virial theorem the energy will be half the potential energy, so
e2 e2 R
El = − =− l−2 = − 2
8πǫ0 r 8πǫ0 a0 l
as required. Bohr was rescued by zero-point motion. It happens that for a Kepler potential the
zero-point energy is equal to the Bohr energy.
8.4 Show that the speed of a classical electron in the lowest Bohr orbit (Problem 8.3) is v = αc,
where α = e2 /4πǫ0 h̄c is the fine-structure constant. What is the corresponding speed for a hydrogen-
like Fe ion (atomic number Z = 26)? Given these results, what fractional errors must we expect in
the energies of states that we derive from non-relativistic quantum mechanics.
Soln: In hydrogen
h̄ e2
v= = = αc
ma0 4πǫ0 h̄
In this formula one power of e is for the charge on the proton and one for the electron. So v ∝ Z.
So in Fe v = 26/137c ≃ 0.2c.
Problems 57
The relativistic formula for kinetic energy differs from the Newtonian formula by terms that
are O(v 2 /c2 ) smaller than the Newtonian terms. So we expect our non-relativistic results to be in
error by ∼ 10−4 for hydrogen and by 4% for iron.
8.5 Show that Bohr’s hypothesis (that a particle’s angular momentum must be an integer multiple
of h̄), when applied to the three-dimensional harmonic oscillator, predicts energy levels E = lh̄ω
with l = 1, 2, . . . Is there an experiment that would falsify this prediction?
Soln: From classical mechanics
mv 2
= mω 2 r ⇒ v 2 = ω 2 r2
r
Eliminating r from the Bohr condition mvr = lh̄, we find v 2 = lh̄ω/m. By the virial theorem
E = 2KE = mv 2 = lh̄ω. This prediction could be tested if we could measure the change in ground-
state energy when ω is varied. For example, we could compare the binding energies of the 12 CO and
14
CO, which should have identical springs and therefore different values of ω.
8.6 Show that the electric field experienced by an electron in the ground state of hydrogen is of
order 5 × 1011 V m−1 . Why is it impossible to generate comparable macroscopic fields using charged
electrodes. Lasers are available that can generate beam fluxes as big as 1022 W m−2 . Show that the
electric field in such a beam is of comparable magnitude.
Soln:
e 1.6 × 10−19
E= 2 = = 5.1 × 1011 V m−1
4πǫ0 a0 4π × 8.9 × 10−12 (5.3 × 10−11 )2
When an electron moves a0 down this field, it gains eEa0 = 5.3 × 5.1 eV of energy, which exceeds
the binding energies of the outer electrons of atoms. So such a field would strip electrons from
any material exposed to it (such as the generating electrodes) and the subsequent current would
discharge the electrodes.
The energy flux of an electromagnetic wave is given by the Poynting vector
1 E2
I= |E × B| =
µ0 cµ0
where we’ve used E/c = B in an electromagnetic wave. Thus
p
E ∼ Iµ0 c ∼ (1022 × 4π × 10−7 × 3 × 108 )1/2 = 1.9 × 1012 V m−1
8.7 Positronium consists of an electron and a positron (both spin-half and of equal mass) in orbit
around one another. What are its energy levels? By what factor is a positronium atom bigger than
a hydrogen atom?
Soln: We have 2 2
1 e
Eground = 2 µ
4πǫ0 h̄
so Eground ∝ µ. For positronium µ = 12 me rather than ∼ me in H. So the energy scale of positronium
is half that for H. Since a0 ∝ 1/µ, positronium is twice as large as hydrogen.
8.8 The emission spectrum of the He+ ion contains the Pickering series of spectral lines that is
analogous to the Lyman, Balmer and Paschen series in the spectrum of hydrogen.
Balmer i = 1, 2, . . . 0.456806 0.616682 0.690685 0.730884
Pickering i = 2, 4, . . . 0.456987 0.616933 0.690967 0.731183
The table gives the frequencies (in 1015 Hz) of the first four lines of the Balmer series and the first
four even-numbered lines of the Pickering series. The frequencies of these lines in the Pickering series
are almost coincident with the frequencies of lines of the Balmer series. Explain this finding. Provide
a quantitative explanation of the small offset between these nearly coincident lines in terms of the
reduced mass of the electron in the two systems. (In 1896 E.C. Pickering identified the odd-numbered
lines in his series in the spectrum of the star ζ Puppis. Helium had yet to be discovered and he
believed that the lines were being produced by hydrogen. Naturally he confused the even-numbered
lines of his series with ordinary Balmer lines.)
58 Problems
Indeed, the analogous integral representation of hn, l, m|n′ , l′ , m′ i must vanish for n 6= n′ for then
the states correspond to different e-values, and for l = l′ and m = m′ the only the integral over r
can vanish.
To prove orthogonality with respect to l we recall that |n, l, mi is an eigenket of the radial
Hamiltonian
h̄2 p2r l(l + 1) 2Z
Hl = + − .
2µ h̄2 r2 ra0
Hence
l
1 d 2 dun l(l + 1) l 2Z l 2µEn l Z2 l
− 2 r + u n − u n = u n = u
r dr dr r2 ra0 h̄2 n2 a20 n
We bring this equation into Sturm-Liouville form by multiplying through by −r2 :
l
d 2 dun 2Zr Z 2 r2
r + + 2 2 uln = l(l + 1)uln
dr dr a0 n a0
′
The required orthogonality relation follows because uln and uln are eigenfunctions for distinct eigen-
values of the Sturm-Liouville operator on the left of the last equation.
8.12 Show that for hydrogen the matrix element h2, 0, 0|z|2, 1, 0i = −3a0 . On account of the non-
zero value of this matrix element, when an electric field is applied to a hydrogen atom in its first
excited state, the atom’s energy is linear in the field strength (§10.1.2).
Soln: From Tables Table 7.2 and Table 8.1,
Z Z Z r
1 2 1 r −r/2a0 6 2 r
h200|z|210i = dr r dθ sin θ dφ r cos θ √ e cos θ √ 1 − e−r/2a0
(2a0 )3 3 a0 8π 4π 2a0
Z Z
a0
= 4 −x
dx x e (1 − 2 x) dθ sin θ cos2 θ
1
8
a0
= (4! − 21 5!) 23 = −3a0
8
q
8.13∗ From equation (8.50) show that l′ + 21 = (l + 12 )2 − β and that the increment ∆ in l′ when
l is increased by one satisfies ∆2 + ∆(2l′ + 1) = 2(l + 1). By considering the amount by which the
solution of this equation changes when l′ changes from l as a result of β increasing from zero to a
small number, show that
2β
∆=1+ 2 + O(β 2 ). (8.6)
4l − 1
Explain the physical significance of this result.
Soln: The given eqn is a quadratic in l′ :
p q
′2 ′ ′ −1 ± 1 + 4l(l + 1) − 4β
l +l −l(l +1)+β = 0 ⇒ l = ⇒ l′ + 12 = (l + 21 )2 − β, (8.7)
2
where we’ve chosen the root that makes l′ > 0.
Squaring up this equation, we have
(l′ + 21 )2 = (l + 12 )2 − β ⇒ (l′ + ∆ + 21 )2 = (l + 23 )2 − β
Taking the first eqn from the second yields
∆2 + 2(l′ + 12 )∆ = (l + 23 )2 − (l + 12 )2 = 2(l + 1)
This is a quadratic equation for ∆, which is solved by ∆ = 1 when l′ = l. We are interested in
the small change δ∆ in this solution when l′ changes by a small amount δl′ . Differentiating the
equation, we have
2∆δl′
2∆δ∆ + 2∆δl′ + (2l′ + 1)δ∆ = 0 ⇒ δ∆ = −
2∆ + 2l′ + 1
Problems 61
dΦ eh̄2 X
=− h[Si , Sj ]Lj + Sj [Si , Lj ]i
dr 2rm2e c2 j
dΦ eh̄2 X
=− iǫijk hSk Lj i
dr 2rm2e c2
jk
P
Cancelling ih̄ from each side and using jk ǫijk Sk Lj = −(S × L)i on the right we obtain
d hSi dΦ eh̄
= hS × Li
dt dr 2rm2e c2
which agrees with (8.74) for Q = −e and B = 0.
8.15∗ (a) A particle of mass m moves in a spherical potential V (r). Show that according to
classical mechanics
d dV der
(p × Lc ) = mr2 , (8.8)
dt dr dt
where Lc = r × p is the classical angular-momentum vector and er is the unit vector in the radial
direction. Hence show that when V (r) = −K/r, with K a constant, the Runge–Lenz vector
Mc ≡ p × Lc − mKer is a constant of motion. Deduce that Mc lies in the orbital plane, and that
for an elliptical orbit it points from the centre of attraction to the pericentre of the orbit, while it
vanishes for a circular orbit.
(b) Show that in quantum mechanics (p × L)† − p × L = −2ip. Hence explain why in quantum
mechanics we take the Runge–Lenz vector operator to be
M ≡ 21 h̄N − mKer where N ≡ p × L − L × p. (8.9)
P
Explain why we can write down the commutation relation [Li , Mj ] = i k ǫijk Mk .
(c) Explain why [p2 , N ] = 0 and why [1/r, p × L] = [1/r, p] × L. Hence show that
1 1
[1/r, N] = i 3 (r2 p − x x · p) − pr2 − p · x x 3 . (8.10)
r r
62 Problems
= (p × L)i − 2ipi .
We want the Runge–Lenz vector to be a Hermitian operator, so we apply the principle that 12 (AB +
BA) is Hermitian even when [A, B] 6= 0 and write
X
Mi = 12 h̄ ǫijk (pj Lk + Lk pj ) − mKer = 21 h̄(p × L − L × p) − mKer
jk
Problems 63
M is a (pseudo) vector operator, so its components have the standard commutation relations with
the components of L.
(c) p2 is a scalar so it commutes with L, and of course it commutes with p, so it must commute
with both p × L and L × p. As a scalar 1/r commutes with L, so
ih̄ ih̄
[1/r, p × L] = [1/r, p] × L = − 3 [r2 , p] × L = − 3 x × L.
2r r
Similarly,
1
[1/r, L × p] = −ih̄L × x 3 .
r
Now
1 X 1 X 1X
(x × L)i = ǫijk ǫklm xj xl pm = ǫijk ǫlmk xj xl pm = (δil δjm − δim δjl )xj xl pm
h̄ h̄ h̄
jklm jklm jlm
1
= (xi x · p − r2 pi )
h̄
and
1 X 1 X 1X
(L × x)i = ǫijk ǫjlm xl pm xk = ǫjki ǫjlm xl pm xk = (δkl δim − δkm δil )xl pm xk
h̄ h̄ h̄
jklm jklm klm
1X 1X
= (xk pi xk − xi pk xk ) = (pi xk xk + ih̄δik xk − pk xi xk − ih̄δki xk )
h̄ h̄
k k
1
= pi r2 − p · x xi
h̄
Hence
n 1 1
[1/r, N] = [p × L, 1/r] − [L × p, 1/r] = i − 3 (xi x · p − r2 pi ) + pi r2 − p · x xi 3 } (8.16)
r r
(d)
X
[p2 , (er )n ] = [p2 , xn /r] = (pj [pj , xn /r] + [pj , xn /r]pj )
j
X
= (pj [pj , xn ]/r + pj xn [pj , 1/r] + [pj , xn ]/rpj + xn [pj , 1/r]pj )
j
X δjn xj δjn xj
= ih̄ −pj + pj xn 3 − pj + xn 3 pj
j
r r r r
X
1 1 xj xj
= ih̄ − pn + pn + pj 3 xn + xn 3 pj
r r j
r r
(e)
p2 K 1
[H, M] = − , 2 h̄{p × L − L × p} − mKer
2m r
The results we have in hand imply that when we expand this commutator, there are only two
non-zero terms, so
1
2 1 1
[H, M] = − 2 K p , er − 2 h̄K , p × L − L × p
r
X
1 1 1 xj xj 1 2
2
1
= 2 ih̄K p + p − pj 3 x + x 3 pj + 3 x x · p − r p − pr − p · x x 3
r r j
r r r r
=0
This result shows: (i) that the eigenvalues of the Mi are good quantum numbers – if the particle
starts in an eigenstate of Mi , it will remain in that state; (ii) the unitary transformations Ui (θ) ≡
64 Problems
exp(−iθMi ) are dynamical symmetries of a hydrogen atom. In particular, these operators turn
stationary states into other stationary states of the same energy.
(f) (i)
X X X
[Mi , L2 ] = [Mi , L2j ] = ([Mi , Lj ]Lj + Lj [Mi , Lj ]) = i ǫijk (Mk Lj + Lj Mk ) 6= 0.
j j jk
so we do not expect to know the total angular momentum when the atom is in an eigenstate of any
of the Mi . P P
(ii) [Li , M 2 ] = 2
j [Li , Mj ] = i jk ǫijk (Mk Mj + Mj Mk ) = 0, so there is a complete set of
mutual eigenstates of L2 , Lz and M 2 .
(g)
X X X X
[(p × L)i , pm ] = ǫijk [pj Lk , pm ] = ǫijk pj [Lk , pm ] = i ǫijk ǫkmn pj pn = i ǫkij ǫkmn pj pn
jk jk jkn jkn
X
2
=i (δim δjn − δin δjm )pj pn = i(p δim − pi pm )
nj
= pi (p × L)s − (L × p)s pi
(8.18)
The standardised sum of cross products in equation (8.17) is now
i − 2pi (p × L)s + 2(L × p)s pi + pi (p × L)s + pi (p × L)s − (L × p)s pi − (L × p)s pi
Problems 65
and is manifestly zero. The last term in (8.17) has to vanish because it alone is symmetric in is,
and it’s not hard to show that it does:
X
p·N+N·p= ǫijk pi (pj Lk − Lj pk ) + (pj Lk − Lj pk )pi
ijk
The first and last terms trivially vanish because they are symmetric in ij and ik,respectively. The
remaining terms can be written
X X
− ǫijk pi Lj pk + ǫijk pj Lk pi
ijk jki
and they cancel.
Since er = x/r and in (8.10) we already have [1/r, N] we prepare for calculating [Ni , er ] by
calculating
X X X X
[(p × L)i , xj ] = ǫist [ps Lt , xj ] = ǫist (ps [Lt , xj ] + [ps , xj ]Lt ) = i ǫist ps ǫtjn xn − h̄δsj Lt
st st st n
X X X
=i (δij δsn − δin δsj )ps xn − h̄ ǫijt Lt = i p · x δij − pj xi − h̄ ǫijt Lt
sn t t
Similarly
X
[(L × p)i , xj ] = −i x · p δij − xi pj − h̄ ǫijt Lt
t
so
X
[Ni , xj ] = i (p · x + x · p)δij − (pj xi + xi pj ) − 2h̄ ǫijt Lt
t
Now we can compute
[Ni , (er )j ] = [Ni , xj /r] = [Ni , xj ]/r + xj [Nj , 1/r]
δij 1 2h̄ X
= i (p · x + x · p) − (pj xi + xi pj ) − ǫijt Lt
r r r t
+ xj [(p × L)i , 1/r] − xj [(L × p)i , 1/r]
δij 1 2h̄ X
= i (p · x + x · p) − (pj xi + xi pj ) − ǫijt Lt
r r r t
xj 1
+ 3 (xi x · p − r2 pi ) − xj (pi r2 − p · x xi ) 3
r r
when we calculate [Ni , (er )j ] − [Nj , (er )i ] all terms above that are symmetric in ij and will vanish
and we find
1 4h̄ X
[Ni , (er )j ] − [Nj , (er )i ] = i −(pj xi + xi pj − pi xj − xj pi ) − ǫijt Lt
r r t
1 1 1
− (xj pi − xi pj ) − (xj pi − xi pj ) + (xj p · x xi − xi p · x xj ) 3
r r r
X
4h̄ 1 1 1
=i − ǫijt Lt − (pj xi − pi xj ) − (xj pi − xi pj ) + (xj p · x xi − xi p · xxj ) 3
r t r r r
(8.19)
Now
X X X X
xi pk xk xj = xi (xj pk − ih̄δjk )xk = xi xj pk xk − ih̄xi xj = xj (pk xi + ih̄δki )xk − ih̄xi xj
k k k k
X
= xj pk xk xi
k
66 Problems
so the terms with dot products in (8.19) cancel. Finally [1/r, pj ] = −ih̄xj /r3 so
1 1
(xj pi − xi pj ) = xj (pi /r − ih̄xi /r3 ) − xi (pj /r − ih̄xj /r3 ) = (xj pi − xi pj )
r r
so the terms with factors 1/r in (8.19) cancel and we are left with
4ih̄ X
[Ni , (er )j ] − [Nj , (er )i ] = − ǫijt Lt (8.20)
r t
From the definition of M we have
[Mi , Mj ] = [ 21 h̄Ni − mK(er )i , 12 h̄Ni − mK(er )i ] = 41 h̄2 [Ni , Nj ] − 21 mKh̄([Ni , (er )j ] + [(er )i , Nj ])
= 41 h̄2 [Ni , Nj ] − 12 mKh̄ [Ni , (er )j ] − [Nj , (er )i ] .
since the components of e commute with each other. We obtain the required result on substituting
from equations (8.12) and (8.13).
A physical consequence of (8.14) is that we will not normally be able to know the values of
more than one component of M – but we can in the exceptional case of completely radial orbits
(L2 |ψi = 0).
Problems
9.1 This problem is all classical electromagnetism, but it gives physical insight into quantum
physics. It is hard to do without a command of Cartesian tensor notation (Appendix B). A point
charge Q is placed at the origin in the magnetic field generated by a spatially confined current
distribution. Given that
Q r
E= (9.1)
4πǫ0 r3
and B = ∇ × A with ∇ · A = 0, show that the field’s momentum
Z
P ≡ ǫ0 d3 x E × B = QA(0). (9.2)
Write down the relation between the particle’s position and momentum and interpret this relation
physically in light of the result you have just obtained.
Hint: write E = −(Q/4πǫ0 )∇r−1 and B = ∇ × A, expand the vector triple product and
integrate each of the resulting terms by parts so as to exploit in one ∇R· A = 0 and in
H the other
∇2 r−1 = −4πδ 3 (r). The tensor form of Gauss’s theorem states that d3 x ∇i T = d2 Si T no
matter how many indices the tensor T may carry.
Soln: In tensor notation with ∂i ≡ ∂/∂xi
Z
Q
P=− d3 x ∇r−1 × (∇ × A)
4π
becomes Z
Q X
Pi = − d3 x ǫijk ∂j r−1 ǫklm ∂l Am
4π
jklm
Z X
Q
=− d3 x ǫkij ǫklm ∂j r−1 ∂l Am
4π
jklm
Z X
Q
=− d3 x (δil δjm − δim δjl )∂j r−1 ∂l Am
4π
jlm
Z X
Q
=− d3 x (∂j r−1 ∂i Aj − ∂j r−1 ∂j Ai )
4π j
In the first integral we use the divergence theorem to move ∂j from r−1 to ∂i Aj and in the second
integral we move the ∂j from Aj to ∂j r−1 . We argue that the surface integrals over some bounding
sphere of radius R that arise in this process vanish as R → ∞ because their integrands have explicit
Problems 67
R−2 scaling, so they will tend to zero as R → ∞ so long as A → 0 no matter how slowly. After this
has been done we have Z
Z
Q 3 −1 3 −1
Pi = − − d x (r ∂j ∂i Aj + d x ∂j ∂j r Ai
4π
The first integral vanishes because ∂j Aj = ∇ · A = 0 and in the second integral we note that
∂j ∂j r−1 = ∇2 r−1 = −4πδ 3 (x), so evaluation of the integral is trivial and yields the required expres-
sion.
The physical interpretation is that when we accelerate a charge, we have to inject momentum
into the emag field that moves with the charge. So the conserved momentum p associated with the
particle’s motion includes both the momentum in the particle and that in the field, which we have
just shown to be QA(x), where x is the particle’s location. Thus p = mẋ + QA. The particle’s
kinetic energy is H = 21 mẋ2 = (p − QA)2 /2m.
9.2 From equation (9.32) show that the normalised wavefunction of a particle of mass m that is
in the nth Landau level of a uniform magnetic field B is
2 2
rn e−r e /4rB −inφ
hx|ni = √ n+1
, (9.3)
2(n+1)/2 n! π rB
p
where rB = h̄/QB. Hence show that the expectation of the particle’s gyration radius is
√ (n + 21 )!
hrin ≡ hn|r|ni = 2 rB . (9.4)
n!
Show further that
δ ln hrin 1
≃ (9.5)
δn √ 2n
and thus show that in the limit of large n, hri ∝ E, where E is the energy of the level. Show that
this result is in accordance with the correspondence principle. R R
2 2
Soln: We obtain the required expression R ∞ forz hx|ni by evaluating dr r r2n e−r /2rB (n+r2 /2rB 2
) dφ
with the help of the definition z! = 0 dx x e−x . Then
R 2 2 Z √
dr r2 r2n e−r /2rB 2π 23/2 rB ∞ dz n+1/2 −z 2rB
hrin = 2n+2 = z e = (n + 12 )!
2n+1 n! π rB n! 0 2 n!
√
where z ≡ r/ 2rB .
3
δ lnhrin (n + 23 )!n! n + 32 1 + 2n 1
= lnhrin+1 − lnhrin = ln 1 = ln = ln 1 ≃
δn (n + 2 )!(n + 1)! n+1 1+ n 2n
Summing this expression over several increments we have that the overall change ∆ ln(hrin ) =
∆ ln(n1/2 ), so for n ≫ 1, hrin ∼ n1/2 ∼ E 1/2 . Classically
2
mv 2 QB
= QvB ⇒ E = 21 mv 2 = 21 m r2
r m
so r ∝ E 1/2 as in QM.
9.3 Show that in the gauge in which the magnetic vector potential is A = 21 B× x the wavefunction
of the nth Landau level of gyration about the point a is
2 2
hx|n, ai = eiQ(B×a)·x/2h̄ {(x − ax ) − i(y − ay )}n e−|x−a| /4rB . (9.6)
′ 1
Soln: For the gauge in which the vector potential is A = 2 B × (x − a) the substitution x → x − a
in equation (9.32) yields that the wavefunction of the nth Landau level of oscillation about a is
2 2
hx|n′ , ai ∝ {(x − ax ) − i(y − ay )}n e−|x−a| /4rB
.
68 Problems
The scalar function Λ that effects the gauge change A′ → A that we need is Λ = 12 x · B × a because
A = A′ + ∇Λ. Thus the required wavefunction is the one just given multiplied by
eiQΛ/h̄ = eiQ(B×a)·x/2h̄
as required.
9.4 A particle of charge Q is confined to move in the xy plane, with electrostatic potential φ = 0
and vector potential A satisfying
∇ × A = (0, 0, B). (9.7)
Consider the operators ρx , ρy , Rx and Ry , defined by
1
ρ= êz × (p − QA) and R = r − ρ, (9.8)
QB
where r and p are the usual position and momentum operators, and êz is the unit vector along B.
Show that the only non-zero commutators formed from the x and y components of these are
2 2
[ρx , ρy ] = irB and [Rx , Ry ] = −irB , (9.9)
2
where rB = h̄/QB.
The operators a, a† , b and b† are defined via
1 1
a= √ (ρx + iρy ) and b= √ (Ry + iRx ). (9.10)
2 rB 2 rB
Evaluate [a, a†] and [b, b† ]. Show that for suitably defined ω, the Hamiltonian can be written
H = h̄ω a† a + 12 . (9.11)
Given that there exists a unique state |ψi satisfying
a|ψi = b|ψi = 0, (9.12)
what conclusions can be drawn about the allowed energies of the Hamiltonian and their degeneracies?
What is the physical interpretation of these results?
Soln: We have
1 1
ρx = − (py − QAy ) ; ρy = (px − QAx )
QB QB
so
1 1
[ρx , ρy ] = − 2
[(py − QAy ), (px − QAx )] = ([py , Ax ] + [Ay , px ])
(QB) QB 2
−ih̄ ∂Ax ∂Ay ih̄ 2
= 2
− = = irB
QB ∂y ∂x QB
Also
1 1 ih̄ 2
[ρx , x] = − [py , x] = 0 ; [ρx , y] = − [py , y] = = irB
QB QB QB
1 2
[ρy , x] = [px , x] = −irB ; [ρy , y] = 0
QB
So
2
[ρx , Rx ] = [ρx , x − ρx ] = 0 ; [ρx , Ry ] = [ρx , y − ρy ] = irB (1 − 1) = 0
and similarly [ρy , Ri ] = 0. Finally
2
[Rx , Ry ] = [x − ρx , y − ρy ] = −[ρx , y] − [x, ρy ] + [ρx , ρy ] = −irB
Now
1 1
[a, a† ] = 2 [ρx + iρy , ρx − iρy ] = − 2 2i[ρx , ρy ] = 1
2rB 2rB
1 1
[b, b† ] = 2 [Ry + iRx , Ry − iRx ] = 2 2i[Rx , Ry ] = 1
2rB 2rB
Problems 69
1 1
a† a = 2 (ρx − iρy )(ρx + iρy ) = 2 (ρ2x + ρ2y + i[ρx , ρy ])
2rB 2rB
1 |p − QA|2 2 1 |p − QA|2
= 2 − rB = − 12
2rB (QB)2 2h̄ QB
But
|p − QA|2 QB
H= = (a† a + 21 )h̄ = (a† a + 21 )h̄ω,
2m m
where ω ≡ QB/m.
Given a|ψi = 0, we have H|ψi = (a† a + 12 )h̄ω|ψi = 12 h̄ω|ψi so |ψi is a stationary state of energy
1 †
2 h̄ω. Given any stationary state |Ei of energy |Ei we have that a |Ei is a stationary state of energy
E + h̄ω:
Ha† |Ei = h̄ωa† (aa† + 21 )|Ei = h̄ωa† (a† a + [a, a† ] + 21 )|Ei = a† (H + h̄ω)|Ei = (E + h̄ω)a† |Ei,
so the energies are h̄ω × ( 21 , 32 , 52 , . . .).
Since [ρi , Rj ] = 0 we have [a, b] = [a, b† ] = 0 etc, and it follows that [b† , H] = 0. Hence b† |Ei is
a stationary state of energy E.
To show that it is distinct from |Ei, we define the number operator N ≡ b† b. N |ψi = 0 because
b|ψi = 0. Given that N |ni = n|ni, we have that
N (b† |ni) = b† (bb† |ni) = b† (b† b + [b, b† ])|ni = b† (N + 1)|ni = (n + 1)b† |ni
Thus by repeatedly applying b† to |ψi we can create distinct states of energy 12 h̄ω, and by applying
it to a† |ψi we can make distinct states of energy 32 h̄ω, etc. Hence all energy levels are infinitely
degenerate.
Physically the energy levels are degenerate because there are an infinite number of possible
locations of the gyrocentre of a particle that is gyrating with a given energy.
9.5 Using cylindrical polar coordinates (R, φ, z), show that the probability current density associ-
ated with the wavefunction (9.3) of the nth Landau level is
2 2
h̄R2n−1 e−R /2rB R2
J(R) = − n+1 2n+2 n + 2 êφ , (9.13)
2 πn!mrB 2rB
p
where rB ≡ h̄/QB. Plot J as a function of R and interpret your plot physically.
Soln: In cylindrical polars,
∂ 1 ∂ ∂
∇= , , ,
∂R R ∂φ ∂z
so we obtain the required expression for J of the Landau level when we take |ψ| and θ from (9.3)
and use A = 12 BRêφ .
The figure shows the current density as a function of radius for the first six Landau levels. The
ground state is in a class by itself. For the excited states the characteristic radius R increases with
70 Problems
√
n roughly as n as predicted by classical physics, while the peak magnitude of the flow is almost
independent of n because J ∝ ω|ψ 2 | and ω ∼ constant and |ψ | 2 ≃ 1/R as expected from classical
mechanics.
9.6 In classical electromagnetism the magnetic moment of a planar loop of wire that has area A,
normal n̂ and carries a current I is defined to be
µ = IAn̂. (9.14)
Use this formula and equation (9.13) to show that the magnetic moment of a charge Q that is in
a Landau level of a magnetic field B has magnitude µ = E/B, where E is the energy of the level.
Rederive this formula from classical mechanics.
Soln: Eq. (9.13) gives the probability current density J. We multiply it by Q to get the electrical
current density QJ. The current dI = QJ(r)dr through the annulus of radius r contributes magnetic
moment dµ = πr2 dI. Summing over annuli we obtain
Z ∞ Z ∞
Qh̄ 2n+1 −r 2 /2rB
2 r2
µ = Qπ dr r2 J(r) = − n+1 2n+2 dr r e n + 2
0 2 n!mrB 0 2rB
Z ∞
h̄Q h̄Q h̄Q
=− dz z n e−z (n + z) = − [n n! + (n + 1)!] = (n + 12 )
2n!m 0 2n!m m
1
But E = (n + 2 )h̄QB/m so µ = E/B.
Classically, I = Q/τ = Qω/2π = Q2 B/2πm and r2 = 2mE/Q2 B 2 , so µ = πr2 I = E/B.
Problems
10.1 A harmonic oscillator with mass m and angular frequency ω is perturbed by δH = ǫx2 . (a)
What is the exact change in the ground-state energy? Expand this change in powers of ǫ up to order
ǫ2 . (b) Show that the change given by first-order perturbation theory agrees with
√ the exact result to
O(ǫ) (c) Show that the first-order change in the ground state is |bi = −(ǫℓ2 / 2h̄ω)|E2 i. (d) Show
that second-order perturbation theory yields an energy change Ec = −ǫ2 h̄/4m2 ω 3 in agreement with
the exact result.
Soln: (a) The perturbed Hamiltonian is {p2 + (mω ′ x)2 }/2m, where ω ′2 = ω 2 + 2ǫ/m. Hence the
ground-state energy changes by
r !
2ǫ ǫ ǫ 2
1 ′ 1 1 1
δE = 2 h̄(ω − ω) = 2 h̄ω 1+ − 1 = 2 h̄ω − 2 mω 2 + · · · .
mω 2 mω 2
(b) Perturbation theory yields
2
h0|ǫx2 |0i = ǫℓ2 h0| A + A† |0i = ǫℓ2 h0| A2 + AA† + A† A + A†2 |0i = ǫℓ2 h0|AA† |0i
h̄
= ǫℓ2 = ǫ
2mω
which agrees with the binomial expansion of theP exact to order ǫ.
(c) The change in the ground-state is |bi = k6=0 bk |Ek i, where
† 2 †2 √
2 hEk |(A + A ) |E0 i 2 hEk |A |E0 i ǫℓ2 2
bk = ǫℓ = −ǫℓ =− δk2 .
−kh̄ω kh̄ω 2h̄ω
where the second equality exploits the facts that k 6= 0 and A|E0 i = 0.
(d) We have
(ǫℓ2 )2 (ǫℓ2 )2 ǫ2 h̄
Ec = hE0 |δH|bi = ǫℓ2 hE0 |(A + A† )2 |bi = − √ hE0 |(A + A† )2 |E2 i = − =−
2h̄ω h̄ω 4m2 ω 3
10.2 The harmonic oscillator of Problem 10.1 is perturbed by δH = ǫx. Show that the perturbed
Hamiltonian can be written
1 2 2 2 2 ǫ2
H= p +m ω X − 2 ,
2m ω
Problems 71
where X = x + ǫ/mω 2 and hence deduce the exact change in the ground-state energy. Interpret
these results physically.
What value does first-order perturbation theory give? From perturbation theory determine the
coefficient b1 of the unperturbed first-excited state in the perturbed ground state. Discuss your
result in relation to the exact ground state of the perturbed oscillator.
Soln: The potential energy is
1 2 2 2ǫx 1 2 ǫ 2 ǫ 2 1 2 2 2 ǫ2
V (x) = 2 mω x + = 2 mω x+ − = m ω X − 2
mω 2 mω 2 mω 2 2m ω
as required. The eigenergies of {p2 + (mωX)2 }/2m are (n + 12 )h̄ω, so the energies of the perturbed
Hamiltonian are simply
ǫ2
En = (n + 21 )h̄ω −
2mω 2
Physically, the oscillating body has just shifted its point of equilibrium through to x = −ǫ/mω 2 and
oscillates at the same frequency about this new point of equilibrium.
First-order perturbation theory gives δ1 En = ǫhn|x|ni = 0.
The first-order change to the ground state is
X h0|∆|ni X h0|(A + A† )|ni ǫℓ ǫℓ
|bi = |ni = ǫℓ |ni = − h0|A|1i|1i = − |1i.
n
E0 − En n
E0 − En h̄ω h̄ω
The perturbed wavefunction is
2 2
ǫℓ ǫℓ x e−x /4ℓ
ψ(x) ≃ hx|0i − hx|1i = 1 −
h̄ω h̄ω ℓ (2πℓ2 )1/4
This peaks at x < 0 just like the exact stationary state, a displaced Gaussian.
10.3 The harmonic oscillator of Problem 10.1 is perturbed by δH = ǫx4 . Show that the first-order
change in the energy of the nth excited state is
2
2 h̄
δE = 3(2n + 2n + 1)ǫ . (10.1)
2mω
Hint: express x in terms of A + A† .
Soln: Using (3.19)
2
h̄
δEn = hn|ǫx2 |ni = ǫ hn|(A + A† )4 |ni
2mω
When expanding (A + A† )4 we need retain only terms with equal powers of A and A† and we
need every possible ordering of these factors just once. We find these by writing down A2 A†2 and
migrating each A† to the left one step at a time
2
h̄
δEn = ǫ hn|(A2 A†2 + AA† AA† + A† A2 A† + A† AA† A + AA†2 A + A†2 A2 )|ni
2mω
2
h̄
=ǫ (n + 2)(n + 1) + (n + 1)2 + (n + 1)n + n2 + (n + 1)n + n(n − 1)
2mω
2
h̄
=ǫ (n + 1)(n + 2 + n + 1 + n) + n2 + n(n + 1 + n − 1)
2mω
2
h̄
=ǫ 3(n + 1)2 + 3n2
2mω
10.4 The infinite square-well potential V (x) = 0 for |x| < a and ∞ for |x| > a is perturbed by
the potential δV = ǫx/a. Show that to first order in ǫ the energy levels of a particle of mass m are
unchanged. Show that even to this order the ground-state wavefunction is changed to
1 16ǫ X n
ψ1 (x) = √ cos(πx/2a) + 2 √ (−1)n/2 2 sin(nπx/2a),
a π E1 a n=2,4, (n − 1)3
72 Problems
where E1 is the ground-state energy. Explain physically why this wavefunction does not have well-
defined parity but predicts that the particle is more likely to be found on one side of the origin than
the other. State with reasons but without further calculation whether the second-order change in
the ground-state energy will be positive or negative.
Soln: The stationary states of the unperturbed system have well-defined parity, and the pertur-
bation has odd parity. Consequently δEn = hn|δV |ni = 0 for all n.
Expanding
P the perturbed ground-state |1′ i in terms of the unperturbed stationary state as
′
|1 i = n bn |ni, we have that
hn|δV |0i
bn = −
En − E1
The numerators are non-zero only for states of odd parity, that √ is even values of n. The ground-
state wavefunction is u1 (x) = A cos(πx/2a), where A = 1/ a, and the odd-parity excited-state
wavefunctions are un = A sin(nπx/2a), so En = h̄2 n2 π 2 /8a2 m = n2 E1
Z a Z π/2
ǫ 4ǫ
bn = − dx sin(nπx/2a)x cos(πx/2a) = − dθ θ sin(nθ)x cos(θ)
E1 a2 (n2 − 1) −a E1 π 2 (n2 − 1) −π/2
Z π/2
4ǫ
=− dθ θ {sin[(n + 1)θ] + sin[(n − 1)θ]}
E1 π 2 (n2 − 1) 0
4ǫ sin[(n + 1)π/2] sin[(n − 1)π/2]
=− +
E1 π 2 (n2 − 1) (n + 1)2 (n − 1)2
n = 2r is an even number, and sin[(2r + 1)π/2] = (−1)r , so the two terms in the big bracket have
opposite signs and can be combined:
(−1)n/2 4ǫ 1 1
bn = − −
E1 π 2 (n2 − 1) (n + 1)2 (n − 1)2
r
(−1) 4ǫ −4n
=−
E1 π (n − 1) (n − 1)2
2 2 2
If A1 = A2 this simplifies to
q
λ = A1 + 12 B1 ǫ + ± 21 B12 + 4B22 ǫ
in agreement with perturbation theory. If A1 6= A2 we expand the radical to first order in ǫ
1 1 B1 2
λ = 2 (A1 + A2 + B1 ǫ) ± 2 (A1 − A2 ) 1 + ǫ + O(ǫ )
A1 − A2
A1 + B1 ǫ if +
=
A2 if −
again in agreement with perturbation theory
10.9∗ For the P states of hydrogen, obtain the shift in energy caused by a weak magnetic field
(a) by evaluating the Landé g factor, and (b) by use equation (10.28) and the Clebsch–Gordan
coefficients calculated in §7.6.2.
Soln: (a) From l = 1 and s = 12 we can construct j = 23 and 21 so we have to evaluate two values
of gL . When j = 32 , j(j + 1) = 15/4, and when j = 12 , j(j + 1) = 3/4, so
4
3 1 l(l + 1) − s(s + 1) for j = 23
gL = 2 − 2 = 23
j(j + 1) 3 for j = 21
So
2 for j = 32 , m = 32
EB /(µB B) = mgL = 32 for j = 23 , m = 12
1
3 for j = 21 , m = 12
with the values for negative m being minus the values for positive m.
(b) We have | 32 , 32 i = |+i|11i so h 23 , 23 |Sz | 32 23 i = 12 and EB /(µB B) = m + hψ|Sz |ψi = 32 + 21 = 2
in agreement with the Landé factor. Similarly
q q
| 32 , 12 i = 13 |−i|11i + 23 |+i|10i ⇒ h 23 , 12 |Sz | 32 , 12 i = 13 (− 12 ) + 32 12 = 61
1
so EB /(µB B) = 2 + 61 = 32 Finally
q q
| 12 , 21 i = 23 |−i|11i − 13 |+i|10i ⇒ h 12 , 12 |Sz | 21 , 12 i = 23 (− 12 ) + 11
23 = − 16
1 1 1
so EB /(µB B) = 2 − 6 = 3
10.10 The 2 × 2 Hermitian matrix H has positive eigenvalues λ1 > λ2 . The vectors (X, Y ) and
(x, y) are related by
X x
H· = .
Y y
Show that the points (λ1 X, λ2 Y ) and (x, y) are related as shown in Figure 10.4. How does this
result generalise to 3 × 3 matrices? Explain the relation of Rayleigh’s theorem to this result.
76 Problems
Soln: With the coordinate axes parallel to the eigenvectors of H, the matrix is diagonal with λ1
and λ2 its non-vanishing elements. So H(X, Y ) = (x, y) reads
λ1 X = x ; λ2 Y = y
If we place (X, Y ) on the unit circle, so X = cos θ, Y = sin θ, (x, y) will be on the plotted ellipse
because
x2 y2
2 + 2 = X2 + Y 2 = 1
λ1 λ2
Moreover x = λ1 X1 as shown. If H were a 3 × 3 matrix, (X, Y, Z) would be a point on the unit
sphere, and (x, y, z) would lie on an ellipsoid.
The significance for Rayleigh’s theorem is that (X, Y )† H(X, Y ) = (X, Y )† (x, y) is the dot
product of the lines to (X, Y ) and to (x, y). This dot product is clearly stationary in the vicinity of
the principal axes of the ellipse because there is perfect symmetry between the way the two vectors
come into alignment as the axis is approached, and depart from alignment after the axis has been
passed.
10.11 We find an upper limit on the ground-state energy of the harmonic oscillator from the trial
wavefunction ψ(x) = (a2 + x2 )−α . Using the substitution x = a tan θ, or otherwise, show that when
α=1 Z ∞ Z ∞ Z ∞
dx |ψ|2 = 41 πa−3 dx x2 |ψ|2 = 14 πa−1 dx |pψ|2 = 18 πh̄2 a−5 . (10.3)
0 0 0
Hence show that hψ|H|ψi/hψ|ψi is minimised by setting a =√2l/4 ℓ, where ℓ is the characteristic
length of the oscillator. Show that our upper limit on E0 is h̄ω/ 2. Plot the final trial wavefunction
and the actual ground-state wavefunction and infer how α should be changed to obtain a better trial
wavefunction.
Soln: Z ∞ Z Z π/2
dx
dx |ψ|2 = 2 + x2 )2
= a −3
dθ cos2 θ = 41 πa−3
0 (a 0
Z ∞ Z 2 Z π/2
x
2 2
dx x |ψ| = dx 2 =a −1
dθ sin2 θ = 14 πa−1
0 (a + x2 )2 0
Z Z
2 2 x2 2 −5
|p|ψi| = 4h̄ dx 2 = 4h̄ a dθ, cos4 θ sin2 θ
(a + x2 )4
Z π/2
π/2
= 21 h̄2 a−5 dθ (1 + cos 2θ) sin2 2θ = 21 h̄2 a−5 ( 41 π + 61 [sin3 2θ]0 ) = 81 h̄2 a−5
0
So
1 −5 2
8 πa h̄ /2m + 21 mω 2 14 πa−1 h̄2 −2 1
hHi = 1 −3
= a + 2 mω 2 a2
4 πa
4m
4 2 2 2
which has its minimum when a = h̄ /2m ω . Substing this value of a into hHi produces the stated
upper limit.
Problems 77
The trial wavefunction is too narrow in its core and its wings are too prominent. It could be
improved by increasing α from unity.
10.12∗ Show that with 2 2 −2
√ the trial wavefunction ψ(x) = (a + x ) the variational principle yields
an upper limit E0 < ( 7/5)h̄ω ≃ 0.529 h̄ω on the ground-state energy of the harmonic oscillator.
Soln: We set x = a tan θ and have
Z ∞ Z π/2 Z π/2
2 −7 6 −7
dx |ψ| = a dθ cos θ = a dθ { 21 (1 + cos 2θ)}3
0 0 0
Z π/2
= 81 a−7 dθ (1 + 3 cos 2θ + 3 cos2 2θ + cos3 2θ) = 81 a−7 12 π(1 + 23 ) = 5π/32
0
where we have used the facts (i) that an odd power of a cosine averages to zero over (0, π) and (ii)
that cos2 θ has average value 12 over this interval.
Similarly
Z ∞ Z π/2 Z π/2
dx x2 |ψ|2 = a−5 dθ cos4 θ sin2 θ = a−5 dθ 21 (1 + cos 2θ) 14 sin2 2θ
0 0 0
Z π/2
= 18 a−5 dθ (sin2 2θ + cos 2θ sin2 2θ) = 81 a−5 ( 14 π + 61 [sin3 2θ]) = π/32
0
and
−2
hx|p|ψi = −ih̄ 2x
(a2 + x2 )3
so
Z ∞ Z π/2 Z π/2
2 −9 2 2 −9
2
dx |pψ| = 16h̄ a 8
dθ cos θ sin θ = 16h̄ a dθ 18 (1 + cos 2θ)3 14 sin2 2θ
0 0 0
Z π/2 Z π/2
= 12 h̄2 a−9 dθ (sin2 2θ + 3 cos2 2θ sin2 2θ) + dθ cos 2θ(3 + 1 − sin2 2θ)
0 0
= 21 h̄2 a−9 14 π(1 + 43 ) + 23 sin3 2θ − 10
1
sin5 2θ 7 2 −9
= 32 h̄ a π
Hence
7 2 −9
32 h̄ a π/2m + 12 mω 2 32
1 −5
a π h̄2 7 −2 1 2 2
hHi = 5 −7 = a + 10 mω a
32 a π 2m 5
∂ hHi h̄2
0= = − 57 a−3 + 15 mω 2 a
∂a m
2 √
4 h̄ 1/4 √ 7
a =7 ⇒ a=7 2ℓ hHi = h̄ω
mω 5
10.13 Show that for the unnormalised spherically symmetric wavefunction ψ(r) the expectation
value of the gross-structure Hamiltonian of hydrogen is
Z Z !,Z
2
h̄2 2 dψ e2 2
hHi = dr r − dr r|ψ| dr r2 |ψ|2 . (10.4)
2me dr 4πǫ0
For the trial wavefunction ψb = e−br show that
h̄2 b2 e2 b
hHi = − ,
2me 4πǫ0
and hence recover the definitions of the Bohr radius and the Rydberg constant.
78 Problems
Soln: In H we express the kinetic energy operator in terms of pr and L2 and discard the L2 term
on account of the spherical symmetry of ψ.
2
pr e2
hHi = hψ| − |ψi
2m 4πǫ0 r
Now Z ∗
dψ ψ∗ dψ ψ
hψ|p2r |ψi = |pr |ψi|2 = h̄2 dr r2 + +
dr r dr r
Z 2 Z ∗
2 2 dψ 2 2 ψ dψ ψ dψ ∗ |ψ|2
= h̄ dr r + h̄ dr r + + 2
dr r dr r dr r
The second integral is equal to
Z
d(r|ψ|2 )
dr = [r|ψ|2 ]∞0 = 0.
dr
[Alternatively
2
∂ 1 ∂ 1 ∂ 2 ∂ 2 1 ∂ 2 ∂ψ
p2r ψ = −h̄2 + + ψ = −h̄2 + ψ = −h̄ r
∂r r ∂r r ∂r2 r ∂r r2 ∂r ∂r
Hence Z
1 ∂ ∂ψ
hψ|p2r |ψi = −4πh̄2 dr r2 ψ ∗ 2 r2
r ∂r ∂r
Integrating by parts
∞ Z
∂ψ ∂ψ ∗ 2 ∂ψ
hψ|p2r |ψi = −4πh̄2 ψ ∗ r2 − dr r
∂r 0 ∂r ∂r
The square bracket vanishes, so the required result follows (the integral on the bottom is added to
permit ψ to be improperly normalised).]
Putting ψ = e−br we easily find
2 2Z Z ,Z
h̄ b 2 −2br e2 −2br
hHi = dr r e − dr re dr r2 e−2br
2m 4πǫ0
2 Z Z , Z
h̄ 2 −x e2 −x 1
= dx x e − dx xe dx x2 e−x
16mb 16πǫ0 b2 8b3
which yields the required results when the integrals are replaced by 2! and 1! and the expression
simplified. hHi is stationary w.r.t. b when b = me2 /4πǫ0 h̄2 ; the Bohr radius is b−1 as required. The
Rydberg is the value of hHi for this value of b.
2 2
10.14∗ Using the result proved in Problem 10.13, show that the trial wavefunction ψb = e−b r /2
yields −8/(3π)R as an estimate of hydrogen’s ground-state energy, where R is the Rydberg constant.
2 2 2 2
Soln: With ψ = e−b r /2 , dψ/dr = −b2 re−b r /2 , so
2 4Z Z ,Z
h̄ b 4 −b2 r 2 e2 −b2 r 2 2 2
hHi = dr r e − dr re dr r2 e−b r
2m 4πǫ0
2 Z Z , Z
h̄ 4 −x2 e2 −x2 1 2
= dx x e − dx xe dx x2 e−x
2mb 4πǫ0 b2 b3
Now " #∞
Z 2
−x2 e−x
dx xe = = 21
−2
0
Z " # ∞ Z √
−x2
2 xe −x2 π
dx x2 e−x = 1
+ 2 dx e =
−2 4
0
Z " #∞ Z √
3 −x2
2 x e 2 3 π
dx x4 e−x = + 23 dx x2 e−x =
−2 8
0
Problems 79
so √ ,√
h̄2 3 π e2 1 π 3h̄2 b2 e2 b
hHi = − = −
2mb 8 4πǫ0 b2 2 4b3 4m 2π 3/2 ǫ0
At the stationary point of hHi b = me2 /(3π 3/2 ǫ0 h̄2 ). Plugging this into hHi we find
2 2
3h̄2 m2 e4 e2 me2 8 m e 8
hHi = 2 4 − 3/2 2 =− = R
4m 9π ǫ0 h̄
3 2π ǫ0 3π ǫ0 h̄
3/2 3π 2 4πǫ0 3π
10.15 Show that the stationary point of hψ|H|ψi associated with an excited state of H is a saddle
point. Hint: consider the state |ψi = cos θ|ki + sin θ|li, where θ is a parameter.
Soln: Let’s investigate the nature of the extremum in hHi at |ψi = |ki. We vary |ψi to |ψi =
cos θ|ki + sin θ|li and easily find that
hψ|H|ψi = cos2 θEk + sin2 θEl = Ek + (El − Ek ) sin2 θ
If El > Ek , θ = 0 is a minimum of hHi, while it’s a maximum if El < Ek . Since the general variation
of ψ is a linear combination of the |li for all l 6= k, it follows that ψ = |ki is a saddle point of hHi.
10.16 At early times (t ∼ −∞) a harmonic oscillator of mass m and natural angular frequency ω
2 2
is in its ground state. A perturbation δH = Exe−t /τ is then applied, where E and τ are constants.
a. What is the probability according to first-order theory that by late times the oscillator transi-
tions to its second excited state, |2i?
b. Show that to first order in δH the probability that the oscillator transitions to the first excited
state, |1i, is
πE 2 τ 2 −ω2 τ 2 /2
P = e , (10.5)
2mh̄ω
c. Plot P as a function of τ and comment on its behaviour as ωτ → 0 and ωτ → ∞.
Soln: The equation of motion of the amplitude ak (t) to be in the state |ki has the equation of
motion
i
ȧk = − eikωt hk|δH|0i
h̄
On parity grounds p h2|x|0i = 0, so the amplitude to be found in the second excited state is always 0.
By expressing x = h̄/2mω(A + A† ) we easily show that
r
h̄ −t2 /τ 2
h1|δH|0i = ǫ e
2mω
Putting this into the equation of motion of a1 and integrating w.r.t. t gives
r Z t
i h̄ ′ ′2 2
at (t) = − ǫ dt′ eiωt −t /τ
h̄ 2mω −∞
√ 2 2
We evaluate the integral as per Box 2.3 and find that for t = ∞ it comes to τ πe−ω τ /4 . The
desired result now follows as |a1 (∞)|2 .
The transition probability is small if ωτ ≪ 1 because then the perturbation lasts for less than
a period and does no work since the oscillator barely moves in this time. It is small when ωτ ≫ 1
because then the perturbation is applied slowly and the adiabatic principle applies: the oscillator
80 Problems
stays in the same stationary state throughout. Classically in this limit the effects of the perturbation
at different phases of oscillation cancel to high accuracy.
10.17 A particle of mass m executes simple harmonic motion at angular frequency ω. Initially it
is in its ground state but from t = 0 its motion is disturbed by a steady force F . Show that at time
t > 0 and to first order in F the state is
|ψ, ti = e−iE0 t/h̄ |0i + a1 e−iE1 t/h̄ |1i
where Z t
i ′
a1 = √ dt′ F (t′ )eiωt .
2mh̄ω 0
Calculate hxi (t) and show that your expression coincides with the classical solution
Z t
x(t) = dt′ F (t′ )G(t − t′ ),
0
where the Green’s function is G(t − t′ ) = sin[ω(t − t′ )]/mω. Show that a suitable displacement of
the point to which the oscillator’s
P spring is anchored could give rise to the perturbation.
Soln: Writing |ψ, ti = k ak e−iEk t/h̄ |ki, the equation of motion are
X X X
ih̄ ȧk e−iEk t/h̄ |ki = kak e−iEk t/h̄ V |ki = − ak F e−iEk t/h̄ x|ki.
k k
Braing through by hm| we find
iF X iF −iE0 t/h̄
ȧm e−iEm t/h̄ = ak e−iEk t/h̄ hm|x|ki ≃ e hm|x|0i,
h̄ h̄
k
where the second equality
p for the first time restricts the calculation to the first order. Since x =
ℓ(A + A† ) with ℓ = h̄/2mω it follows that the rhs vanishes for m 6= 1 and we have
Z t Z
i ′ ′ iωt′ iℓ t ′ ′
a1 ≃ h1|x|0i dt F (t )e = dt F (t′ )eiωt .
h̄ 0 h̄ 0
The expectation of x is
hxi (t) = ℓ h0|eiE0 t/h̄ + a∗1 eiE1 t/h̄ h1| (A + A† ) |0ie−iE0 t/h̄ + a1 e−iE1 t/h̄ |1i = ℓ(a1 e−iωt + a∗1 eiωt )
Z Z t
2ℓ2 t ′ ′ 1
= 2ℓℜ(a1 e−iωt ) = dt F ℜ ieiω(t −t) = dt′ F sin[ω(t − t′ )]
h̄ 0 mω 0
The classical Green’s function solution is
Z t
1
hxi (t) = dt′ , sin[ω(t − t′ )]F (t′ )
mω 0
in perfect agreement with the quantum result.
At t > 0 the force on the mass is
d d d 1
F − mω 2 x = − (−F x + 21 mω 2 x2 ) = − 21 m (ωx − F/mω)2 = − ( mω 2 X 2 )
dx dx dX 2
where X = x − F/mω 2 , so the perturbation could be the result of suddenly shifting the anchor point
from the origin to x = F/mω 2 .
10.18∗ A particle of mass m is initially trapped by the well with potential V (x) = −Vδ δ(x),
where Vδ > 0. From t = 0 it is disturbed by the time-dependent potential v(x, t) = −F xe−iωt . Its
subsequent wavefunction can be written
Z
|ψi = a(t)e−iE0 t/h̄ |0i + dk {bk (t)|k, ei + ck (t)|k, oi} e−iEk t/h̄ , (10.6)
Problems 81
where E0 is the energy of the bound state |0i and Ek ≡ h̄2 k 2 /2m and |k, ei and |k, oi are, respectively
the even- and odd-parity states of energy Ek (see Problem 5.17). Obtain the equations of motion
Z
ih̄ ȧ|0ie−iE0 t/h̄ + dk ḃk |k, ei + ċk |k, oi e−iEk t/h̄
Z (10.7)
= v a|0ie−iE0 t/h̄ + dk (bk |k, ei + ck |k, oi) e−iEk t/h̄ .
Given that the free states are normalised such that hk ′ , o|k, oi = δ(k − k ′ ), show that to first order
in v, bk = 0 for all t, and that
iF sin(Ωk t/2) Ek − E0
ck (t) = hk, o|x|0i eiΩk t/2 , where Ωk ≡ − ω. (10.8)
h̄ Ωk /2 h̄
Hence show that at late times the probability that the particle has become free is
2πmF 2 t |hk, o|x|0i|2
Pfr (t) = . (10.9)
h̄3 k Ωk =0
Given that from Problem 5.17 we have
√ mVδ 1
hx|0i = Ke−K|x| where K = 2 and hx|k, oi = √ sin(kx), (10.10)
h̄ π
show that r
K 4kK
hk, o|x|0i = . (10.11)
π (k 2 + K 2 )2
Hence show that the probability of becoming free is
p
8h̄F 2 t Ef /|E0 |
Pfr (t) = 2 , (10.12)
mE0 (1 + Ef /|E0 |)4
where Ef > 0 is the final energy. Check that this expression for Pfr is dimensionless and give a
physical explanation of the general form of the energy-dependence of Pfr (t)
Soln: When we substitute the given expansion of |ψi in stationary states of the unperturbed Hamil-
tonian H0 into the tise, the terms generated by differentiating the exponentials in time cancel on
H0 |ψi. The given expression contains the surviving terms, namely the derivatives of the amplitudes
a, bk and ck on the left and on the right v|ψi. In the first order approximation we put a = 1 and
bk = ck = 0 on the right. Then we bra through with hk ′ , e| and hk ′ , o| and exploit the orthonormality
of the stationary states to obtain equations for ḃk (t) and ċk (t). The equation for ḃk is proportional
to the matrix element hk, e|v|0i, which vanishes by parity because v is an odd-parity operator. Then
we replace v by −xF e−iωt and have
Z t Z t
iF ′ iF eiΩk t − 1
ck (t) = dt′ ċk = hk, o|x|0i dt′ ei[(Ek −E0 )/h̄−ω]t = hk, o|x|0i
0 h̄ 0 h̄ iΩk
iF sin(Ωk t/2)
= hk, o|x|0i eiΩk t/2 .
h̄ Ωk /2
The probability that the particle is free is
Z Z
F2 sin2 (Ωk t/2)
Pfr (t) = dk |ck |2 = 2 dk |hk, o|x|0i|2 .
h̄ (Ωk /2)2
As t → ∞ we have sin2 xt/x2 → πtδ(x), so at large t
Z
F2 F 2 |hk, o|x|0i|2 πt
Pfr (t) = 2 dk |hk, o|x|0i|2 πtδ(Ωk /2) = 2
h̄ h̄ d(Ωk /2)/dk Ωk =0
82 Problems
The required result follows when we substitute into the above k 2 /K 2 = Ef /|E0 | and h̄4 K 2 =
(2mE0 )2 .
Regarding dimensions, [F ] = E/L and [h̄] = ET , so
(E/L)2 ET T ET 2 M L2 T −2 T 2
[Pfr ] = 2
= 2
= .
ME ML M L2
Pfr (t) is small for small E because at such energies the free state, which always has a node at
the location of the well, has a long wavelength, so it is practically zero throughout the region of scale
2/K within which the bound particle is trapped. Consequently for small E the coupling between the
bound and free state is small. At high E the wavelength of the free state is much smaller than 2/K
and the positive and negative contributions from neighbouring half cycles of the free state nearly
cancel, so again the coupling between the bound and free states is small. The coupling is most
effective when the wavelength of the free state is just a bit smaller than the size of the bound state.
10.19∗ A particle travelling with momentum p = h̄k > 0 from −∞ encounters the steep-sided
potential well V (x) = −V0 < 0 for |x| < a. Use the Fermi golden rule to show that the probability
that a particle will be reflected by the well is
V2
Preflect ≃ 0 2 sin2 (2ka),
4E
where E = p2 /2m. Show that in the limit E ≫ V0 this result is consistent with the exact reflection
probability derived in Problem 5.10. Hint: adopt periodic boundary conditions so the wavefunctions
of the in and out states can be normalised.
Soln: We consider a length L of the x axis where L ≫ a and k = 2nπ/L, where n ≫ 1 is an
integer. Then correctly normalised wavefunctions of the in and out states are
1 1
ψin (x) = √ eikx ; ψout (x) = √ e−ikx
L L
The required matrix element is
Z Z a
1 L/2 ikx ikx sin(2ka)
dx e V (x)e = −V0 dx e2ikx = −V0
L −L/2 −a Lk
so the rate of transitions from the in to the out state is
2π 2π sin2 (2ka)
Ṗ = g(E)hout|V |ini = g(E)V02
h̄ h̄ L2 k 2
2 2 2
Now we need the density of states g(E). E = p /2m = h̄ k /2m is just kinetic energy. Eliminating
k in favour of n, we have
L √
n= 2mE
2πh̄
Problems 83
Now the cross section dσ is defined by Ṗ = dσ × incoming flux = (v/L3 )dσ = (h̄k/mL3 )dσ. Putting
everything together, we find
Z 2 3
h̄k 1 3 −ik′ ·x ik·x 2π L dk 2
dσ = 6 d xe V (x)e k2 d Ω
mL3 L h̄ 2π dE
Z 2
mk dk/dE 3 −ik′ ·x ik·x
⇒ dσ = d x e V (x)e .
(2π)2 h̄2
Eliminating k with h̄2 k dk = mdE we obtain the desired expression.
The Born approximation is valid providing the unperturbed wavefunction is a reasonable ap-
proximation to the true wavefunction throughout the scattering potential. That is, we must be able
to neglect “shadowing” by the scattering potential.
10.21 Given that a0 = h̄/(αme c) show that the product a0 k of the Bohr radius and the wavenum-
ber of a photon of energy E satisfies
E
a0 k = . (10.16)
αme c2
5
Hence show that the wavenumber kα of an Hα photon satisfies a0 kα = 72 α and determine λα /a0 .
7
What is the connection between this result and our estimate that ∼ 10 oscillations are required to
complete a radiative decay. Does it imply anything about the way the widths of spectral lines from
allowed atomic transitions vary with frequency?
Soln: From a0 = h̄/(αme c) we have a0 k = h̄k/(αme c) = h̄ω/(αme c2 ) = E/(αme c2 ) as required.
An Hα photon is emitted as n = 3 → n = 2, so has energy
1 1 5
Eα = R − = 36 R
4 9
But by (8.68) R = 21 α2 me c2 , and the required result follows. We have λα = 1.2 × 104 a0 ≃ 656 nm.
An antenna that is significantly smaller than the wavelength of the radiation it is trying to emit
is very small. This calculation shows that H atoms are inefficient atennae, so it’s not surprising
that it takes many periods of the oscillating charge distribution to radiate a quantum of energy.
Moreover, this argument predicts that atoms should be more efficient radiators of X-rays, and less
efficient radiators of microwaves. So X-ray lines should have larger natural widths than microwave
lines.
10.22 Equation (10.75) implies that x± act as ladder operators for Jz . Why did we not use these
operators in §7.1?
Soln: x+ does create a new eigenstate of Jz :
J+ (x+ |jmi = (x+ Jz + [Jz , x+ ])|jmi = (x+ Jz + x+ )|jmi = (m + 1)x+ |jmi
It isn’t a substitute for J+ because it doesn’t commute with J 2 and consequently, x+ |jmi is not an
eigenstate of J 2 .
10.23 Given that a system’s Hamiltonian is of the form
p2
H= + V (x) (10.17)
2me
show that [x, [H, x]] = h̄2 /me . By taking the expectation value of this expression in the state |ki,
show that
X h̄2
|hn|x|ki|2 (En − Ek ) = , (10.18)
2me
n6=k
where the sum runs over all the other stationary states.
The oscillator strength of a radiative transition |ki → |ni is defined to be
2me
fkn ≡ 2 (En − Ek )|hn|x|ki|2 . (10.19)
h̄
Problems 85
P
Show that n fkn = 1. What is the significance of oscillator strengths for the allowed radiative
transition rates of atoms?
Soln: For the given form of H the only thing in H that does not commute with x is p2 /2m, so
1 1 ih̄ h̄2
[x, [H, x]] = [x, [p2 , x]] = [x, 2p[p, x]] = − [x, 2p] =
2m 2me 2me me
as required. Taking the expectation value of this expression
h̄2
= hk|[x, [H, x]]|ki = hk|[x, Hx − xH]|ki = hk|(xHx − Hx2 − x2 H + xHx)|ki
me
P
Now we insert I = n |nihn| as suitable points:
h̄2 X
= (2hk|x|nihn|Hx|ki − hk|Hx|nihn|x|ki − hk|x|nihn|xH|ki)
me n
X
= 2(En − Ek )|hn|x|ki|2 .
n
We can clearly drop the term n = k from the sum and the result follows.
By (10.72) the rate of radiative decay from |ki to |ni is proportional to |hn|x|ki|2 . So fkn is
an indication of the fraction of atoms in the state |ki which will decay to the state |ni. It is only
a rough indication on account of both the factor En − Ek that’s included in the definition of fkn
(in order to deliver the f -sum rule) and the presence in (10.72) of of the radiation density dρ/dν,
which is nothing to do with the atom and generally depends strongly on ν (and therefore on n).
Nonetheless, a small oscillator strength implies that a transition will be rare.
Problems
11.1 Show that when the state of a pair of photons is expanded as
X
|ψi = bnn′ |ni|n′ i, (11.1)
nn′
where {|ni} is a complete set of single-photon states, the expansion coefficients satisfy bnn′ = bn′ n .
Soln: Photons are bosons so the wavefunction has to be unchanged by swapping the labels of
the photons. We closely follow the derivation of equation (11.12): we multiply (11.1) through by
hx, x′ , m, m′ | to obtain the two-photon wavefunction. Then we swap x, m with x′ , m′ and subtract
the resulting amplitude from our original amplitude. By exchange symmetry the difference vanishes.
Then in the second sum we swap the dummy indices n and n′ . Finally we extract from the sums
the common factor hx, m||nihx′ , m′ ||n′ i and argue that the remaining coefficient bnn′ − bn′ n of each
basis ket must independently vanish.
11.2 Explain the physical content of writing the wavefunction of a pair of electrons in the form
ψ++ (x, x′ )
′
ψ (x, x )
hx, x′ |ψi = −+ . (11.2)
ψ+− (x, x′ )
′
ψ−− (x, x )
Which of these functions vanishes when the pair is a spin singlet? What relation holds between
the non-zero functions? Suppose |ψi for a spin singlet can be expanded in terms of products of the
single-particle states |u, ±i and |v, ±i in which the individual electrons are in the states associated
with spatial amplitudes u(x) and v(x) with Sz returning ± 12 . Show that
|ψi = 12 (|u, −i|v, +i − |v, +i|u, −i − |u, +i|v, −i + |v, −i|u, +i)
and explain why this expansion is consistent with exchange symmetry.
Given the four single-particle states |u, ±i and |v, ±i, how many linearly independent entangled
states of a pair of particles can be constructed if the particles are not identical? How many linearly
86 Problems
independent states are possible if the particles are identical fermions? Why are only four of these
states accounted for by the states in the first excited level of helium?
Soln: A pair of electrons requires a four-component wavefunction since for every pair of points
(x, x′ ) we need to specify the amplitude for obtaining each of the four possible spin outcomes:
sz = ± 21 at x and at x′ . For a spin singlet ψ++ = ψ−− = 0 and ψ−+ = −ψ+− . The only
independent function, say ψ−+ (x, x′ ) = 12 (u(x)v(x′ ) − u(x′ )v(x)). We substitute this into the given
form of the wavefunction
0
1 u(x)v(x′ ) − u(x′ )v(x)
hx, x′ |ψi =
2 −u(x)v(x′ ) + u(x′ )v(x)
0
On shifting to Dirac notation we obtain the given ket because hx, x′ |ui|vi = u(x)v(x′ ), etc.
Swapping the particle labels in the first term yields the second term, which indeed has the
opposite sign. Ditto the third and fourth terms.
From four single-particle states {|ni} (n = 0, . . . , 3) we have 16 basis vectors |ni|n′ i for the
entangled states, so the space of P entangled states of non-identical particles is 16 dimensional. The
general linear combination |ψi = nn′ ann′ |ni|n′ i has 16 expansion coefficients ann′ , but for identical
Fermions the Pauli conditions allow all the non-vanishing ones to be determined from the 6 coeffi-
cients with n < n′ . So the space of states compatible with the Pauli conditions is six-dimensional.
In the case of helium, the spin triplets times (|ui|vi − |vi|ui) form three of these states and the
spin singlet times (|ui|vi + |vi|ui) forms a fourth. The missing two states are the spin singlet times
|ui|ui or |vi|vi, which between them form the ground and second excited states.
11.3 P In Chapter 6 we saw that when a state of a composite system has a non-trivial expansion
|ψi = ij cij |A; ii|B; ji in terms of products of states |A; ii and |B; ji of the individual systems
it does not automatically follow that the systems are entangled. By recalling the property that
the matrix cij will have if the systems are not entangled, show that any two electrons are always
entangled.
Soln: A and B are not entangled if every column of the matrix cij is a multiple of the first column.
When A and B are both electrons, we show that this property implies that all cij = 0, which is
incompatible with the requirement for normalisation.
Exchange symmetry requires cij = −cji so there are zeros down the diagonal. In particular the
top element of the first column must vanish, and in the absence of entanglement, there must be a
zero at the top of every column. The second element of the second column vanishes, so either this
column is zero times the first column, or the second element of the first column is zero, which will
be reflected in zeros in the second place of every column. If, on the other hand, the whole second
column vanishes, then by the antisymmetry of c, the entire second row must vanish. Repeating this
argument for each row in turn, we show that the absence of correlation forces every element of c to
zero.
11.4∗ In terms of the position vectors xα , x1 and x2 of the α particle and two electrons, the centre
of mass and relative coordinates of a helium atom are
mα xα + me (x1 + x2 )
X≡ , r1 ≡ x1 − X, r2 ≡ x2 − X, (11.3)
mt
where mt ≡ mα + 2me . Write the atom’s potential energy operator in terms of the ri .
Show that
∂ ∂ ∂ ∂
= + +
∂X ∂xα ∂x1 ∂x2
(11.4)
∂ ∂ me ∂ ∂ ∂ me ∂
= − = −
∂r1 ∂x1 mα ∂xα ∂r2 ∂x2 mα ∂xα
and hence that the kinetic energy operator of the helium atom can be written
2
h̄2 ∂ 2 h̄2 ∂ 2 ∂2 h̄2 ∂ ∂
K =− − + 2 − − , (11.5)
2mt ∂X2 2µ ∂r21 ∂r2 2mt ∂x1 ∂x2
where µ ≡ me (1 + 2me /mα ). What is the physical interpretation of the third term on the right?
Explain why it is reasonable to neglect this term.
Problems 87
11.6 The H− ion consists of two electrons bound to a proton. Estimate its ground-state energy by
adapting the calculation of helium’s ground-state energy that uses the variational principle. Show
that using single-particle wavefunctions u(x) ∝ e−r/a the expectation of the Hamiltonian is
a0
hHia = R(2x2 − 11 4 x) where x ≡ a . (11.6)
Hence find that the binding energy of H− is ∼ 0.945R. Will H− be a stable ion?
Soln: The energy −R of an electron with wavefunction ∝ e−r/a0 that is bound to a proton com-
prises R of kinetic energy and −2R of potential energy. When we change the scale length in ψ to
a, the kinetic energy scales as (a0 /a)2 and the potential energy scales as a0 /a. So the energy of two
electrons moving around a proton is
2R{(a0 /a)2 − 2(a0 /a)}
plus the energy of the electron-electron interaction. We obtain the latter energy by setting aZ = a
in Box 11.1: the electron-electron energy is
5 a0 e2 5 a0
8 a 4πǫ a = 4 a R.
0 0
Adding this energy to our expression for the electron-proton energy we obtain the required total.
hHia is stationary when
d hHia 11 2 112
0= = R(4x− 11 4 ) ⇒ x = 16 11
⇒ hHia = R(2( 16 ) − 11 11 11 11 11
4 16 ) = R 16 ( 8 − 4 ) = −R 128
dx
as required. If we accept this result, H− will be unstable because there is enough energy to eject one
of the electrons so the other one can form an H atom. On the other hand, the variational principle
delivers an upper limit on E, so the binding energy could in principle exceed R.
11.7∗ Assume that a LiH molecule comprises a Li+ ion electrostatically bound to an H− ion, and
that in the molecule’s ground state the kinetic energies of the ions can be neglected. Let the centres
of the two ions be separated by a distance b and calculate the resulting electrostatic binding energy
under the assumption that they attract like point charges. Given that the ionisation energy of Li
is 0.40R and using the result of Problem 11.6, show that the molecule has less energy than that of
well separated hydrogen and lithium atoms for b < 4.4a0 . Does this calculation suggest that LiH is
a stable molecule? Is it safe to neglect the kinetic energies of the ions within the molecule?
Soln: When the LI and H are well separated, the energy required to strip an electron from the Li
and park it on the H− is E = (0.4 + 1 − 0.955)R = 0.445R. Now we recover some of this energy
by letting the Li+ and H− fall towards each other. When they have reached distance b the energy
released is
e2 a0
= 2R
4πǫ0 b b
This energy equals our original outlay when b = (2/0.445)a0 = 4.49a0, which establishes the required
proposition.
In LiH the Li-H separation will be <∼ 2a0 , because only at a radius of this order will the electron
clouds of the two ions generate sufficient repulsion to balance the electrostatic attraction we have
been calculating. At this separation the energy will be decidedly less than that of isolated Li and
H, so yes the molecule will be stable.
In its ground state the molecule will have zero angular momentum, so there is no rotational
kinetic energy to worry about. However the length of the Li-H bond will oscillate around its equi-
librium value, roughly as a harmonic oscillator, so there will be zero-point energy. However, this
energy will suffice only to extend the bond length by a fraction of its equilibrium value, so it does
not endanger the stability of the molecule.
11.8∗ Two spin-one gyros are in a box. Express the states |j, mi in which the box has definite
angular momentum as linear combinations of the states |1, mi|1, m′ i in which the individual gyros
have definite angular momentum. Hence show that
1
|0, 0i = √ (|1, −1i|1, 1i − |1, 0i|1, 0i + |1, 1i|1, −1i). (11.7)
3
Problems 89
By considering the symmetries of your expressions, explain why the ground state of carbon has l = 1
rather than l = 2 or 0. What is the total spin angular √ momentum of a√C atom? √
Soln: We have that J− |2, 2i = 2|2, 1i, J− |2, 1i = 6|2, 0i, J− |1, 1i = 2|1, 0i, J− |1, 0i = 2|1, −1i.
We start from
|2, 2i = |1, 1i|1, 1i
and apply J− to both sides, obtaining
√ 1
2|2, 1i = 2(|1, 0i|1, 1i + |1, 1i|1, 0i) ⇒ |2, 1i = √ (|1, 0i|1, 1i + |1, 1i|1, 0i)
2
Applying J− again we find
1
|2, 0i = √ (|1, −1i|1, 1i + 2|1, 0i|1, 0i + |1, 1i|1, −1i)
6
Next we identify |1, 1i as the linear combination of |1, 1i|1, 0i and |1, 0i|1, 1i that’s orthogonal to
|2, 1i. It clearly is
1
|1, 1i = √ (|1, 0i|1, 1i − |1, 1i|1, 0i)
2
We obtain |1, 0i by applying J− to this
1
|1, 0i = √ (|1, −1i|1, 1i − |1, 1i|1, −1i)
2
and applying J− again we have
1
|1, −1i = √ (|1, −1i|1, 0i − |1, 0i|1, −1i)
2
Finally we have that |0, 0i is the linear combination of |1, −1i|1, 1i, |1, 1i|1, −1i and |1, 0i|1, 0i that’s
orthogonal to both |2, 0i and |1, 0i. By inspection it’s the given expression.
The kets for j = 2 and j = 0 are symmetric under interchange of the m values of the gyros,
while that for j = 1 is antisymmetric under interchange. Carbon has two valence electrons both in
an l = 1 state, so each electron maps to a gyro and the box to the atom. When the atom is in the
|1, 1i state, for example, from the above the part of the wavefunction that described the locations
of the two valence electrons is
1
hx1 , x2 |1, 1i = √ (hx1 |1, 0ihx2 |1, 1i − hx1 |1, 1ihx2 |1, 0i)
2
This function is antisymmetric in its arguments so vanishes when x1 = x2 . Hence in this state of
the atom, the electrons do a good job of keeping out of each other’s way and we can expect the
electron-electron repulsion to make this state (and the other two l = 1 states) lower-lying than the
l = 2 or l = 0 states, which lead to wavefunctions that are symmetric functions of x1 and x2 .
Since the wavefunction has to be antisymmetric overall, for the l = 1 state it must be symmetric
in the spins of the electrons, so the total spin has to be 1.
11.9∗ Suppose we have three spin-one gyros in a box. Express the state |0, 0i of the box in which
it has no angular momentum as a linear combination of the states |1, mi|1, m′ i|1, m′′ i in which the
individual gyros have well-defined angular momenta. Hint: start with just two gyros in the box,
giving states |j, mi of the box, and argue that only for a single value of j will it be possible to get
|0, 0i by adding the third gyro; use results from Problem 11.8.
Explain the relevance of your result to the fact that the ground state of nitrogen has l = 0.
Deduce the value of the total electron spin of an N atom.
Soln: Since when we add gyros with spins j1 and j2 the resulting j satisfies |j1 − j2 | ≤ j ≤ j1 + j2 ,
we will be able to construct the state |0, 0i on adding the third gyro to the box, only if the box has
j = 1 before adding the last gyro. From Problem 11.8 we have that
1
|0, 0i = √ (|1, −1i|1, 1i − |1, 0i|1, 0i + |1, 1i|1, −1i),
3
90 Problems
where we can consider the first ket in each product is for the combination of 2 gyros and the second
ket is for the third gyro. We use Problem 11.8 again to express the kets of the 2-gyro box as linear
combinations of the kets of individual gyros:
1 1 1
|0, 0i = √ √ |1, −1i|1, 0i − |1, 0i|1, −1i |1, 1i − √ |1, −1i|1, 1i − |1, 1i|1, −1i |1, 0i
3 2 2
1
+ √ |1, 0i|1, 1i − |1, 1i|1, 0i |1, −1i ,
2
1
= √ |1, −1i|1, 0i|1, 1i − |1, 0i|1, −1i|1, 1i − |1, −1i|1, 1i|1, 0i
6
+ |1, 1i|1, −1i|1, 0i + |1, 0i|1, 1i|1, 0i − |1, 1i|1, 0i|1, 0i
This state is totally antisymmetric under exchange of the m values of the gyros.
When we interpret the gyros as electrons and move to the position representation we find that
the wavefunction of the valence electrons is a totally antisymmetric function of their coordinates,
x1 , x2 , x3 . Hence the electrons do an excellent job of keeping out of each other’s way, and this will
be the ground state. To be totally antisymmetric overall, the state must be symmetric in the spin
labels of the electrons, so the spin states will be |+i|+i|+i and the states obtained from this by
application of J− . Thus the total spin will be s = 32 .
11.10∗ Consider a system made of three spin-half particles with individual spin states |±i. Write
down a linear combination of states such as |+i|+i|−i (with two spins up and one down) that is
symmetric under any exchange of spin eigenvalues ±. Write down three other totally symmetric
states and say what total spin your states correspond to.
Show that it is not possible to construct a linear combination of products of |±i which is totally
antisymmetric.
What consequences do these results have for the structure of atoms such as nitrogen that have
three valence electrons?
Soln: There are just three of these product states to consider because there are just three places
to put the single minus sign. The sum of these states is obviously totally symmetric:
1
|ψi = √ |+i|+i|−i + |+i|−i|+i + |−i|+i|+i
3
Three other totally symmetric state are clearly |+i|+i|+i and what you get from this ket and the
one given by everywhere interchanging + and −. These four kets are the kets | 32 , mi.
A totally antisymmetric state would have to be constructed from the same three basis kets used
above, so we write it as
|ψ ′ i = a|+i|+i|−i + b|+i|−i|+i + c|−i|+i|+i
On swapping the spins of the first and the third particles, the first and third kets would interchange,
and this would have to generate a change of sign. So a = −c and b = 0. Similarly, by swapping the
spins on the first and second particles, we can show that a = 0. Hence |ψi = 0, and we have shown
that no nonzero ket has the required symmetry.
States that satisfy the exchange principle can be constructed by multiplying a spatial wave-
function that is totally antisymmetric in its arguments by a totally symmetric spin function. Such
states have maximum total spin. In contrast to the situation with helium, conforming states can-
not be analogously constructed by multiplying a symmetric wavefunction by an antisymmetric spin
function.
Problems 91
Problems
12.1 We have derived approximate expressions for the change in the energies of stationary states
when an electric or magnetic field is applied. Discuss whether the derivation of these results implicitly
assumed the validity of the adiabatic principle.
Soln: We did not need to assume the validity of the adiabatic approximation because really we are
investigating the eigenvalues and eigenkets of a one-parameter family of time-independent Hamilto-
nians. However, for vividness we imagined varying this parameter over some time interval, and unless
that interval is long the system will not be at all times in an eigenstate of the current Hamiltonian.
12.2 In §12.2 we assumed that the potential energy of air molecules is infinitely large inside a
bicycle pump’s walls. This cannot be strictly true. Give a reasoned order-of-magnitude estimate for
the potential in the walls, and consider how valid it is to approximate this by infinity.
Soln: The atoms of the cylinder’s walls are essentially touching each other. So a gas atom can
penetrate the walls only by appreciably shrinking itself and/or squeezing down the sizes of the wall’s
atoms. The energy required to do this will be on the order of the kinetic energy of the valence
electrons of those atoms, which is a few eV. The only handy source of energy is the thermal kinetic
energy of the gas atoms, 23 kB T ≃ 0.039(T /300 K) eV. So at room temperature the gas atoms haven’t
enough energy to get in there. At a few thousand degrees penetration would be a significant effect,
had the walls not melted/vaporised.
12.3 Explain why E/ω is an adiabatic invariant of a simple harmonic oscillator, where ω is the
oscillator’s angular frequency. Einstein proved this result in classical physics when he was developing
the “old quantum theory”, which involved quantising adiabatic invariants such as E/ω and angular
momentum. Derive the result for a classical oscillator by adapting the derivation of the wkbj
approximation to the oscillator’s equation of motion ẍ = −ω 2 x.
Soln: E/ω = (n+ 21 )h̄ depends only on which stationary state the system is in. So it is an adiabatic
invariant. Rt
To derive this result classically, we define φ = 0 dt ω and write x(t) = X(t)eiφ . Differentiating
and substituting the result in ẍ = −ω 2 x we obtain
Ẍ + 2iω Ẋ + iX ω̇ = 0.
The wkbj approximation is to drop Ẍ. Integrating the remainder we have 2[ln X] = −[ln ω] so
X 2 ω = constant. The energy is the potential energy when x is equal to the amplitude X, so
E ∝ ω 2 X 2 ∝ ω, so E/ω is adiabatically invariant.
12.4 Consider a particle that is trapped in a one-dimensional potential well V (x). If the particle is
in a sufficiently highly excited state of this well, its typical de Broglie wavelength may be sufficiently
smaller than the characteristic lengthscale of the well for the wkbj approximation to be valid.
Explain why it is plausible that in this case
Z
1 x2 ′ p
dx 2m{E − V (x′ )} = nπ, (12.1)
h̄ x1
where
H the E − V (xi ) = 0 and n is an integer. Relate this condition to the quantisation rule
dx px = nh used in the “old quantum theory”. R
Soln: According to the wkbj approximation, the left side is dx k so it is the change in phase of ψ
between turning points. A stationary state is a resonant mode of the cavity formed by the potential
well, and a mode is defined by the condition that the wave interferes constructively with itself when
it returns to its starting point after crossing the cavity out and back. So the wkbj approximation
suggests that for a stationary state twice the left side should be an integer multiple of 2π. The
square root in the integral is precisely the classical momentum px , so our condition is exactly that
of the old quantum theory.
12.5 Show that the “old quantum theory” (Problem 12.4) predicts that the energy levels of the
harmonic oscillator are nh̄ω rather than (n + 12 )h̄ω. Comment on the dependence on n of the
fractional error in En .
Soln: Our quantisation condition is
Z X q Z X p
1 1 2 x2 ) = mω
2 nh = dx 2m(E − 2 mω dx X 2 − x2 ,
−X −X
92 Problems
p
where X ≡ 2E/mω 2 . Substituting x = X cos θ and integrating from 0 to π we find that the
integral evaluates to 12 πX 2 = πE/mω 2 and the required result follows.
The condition for the wkbj approximation to be valid is that the de Broglie wavelength λ is
short compared to the distance over which the potential changes significantly. The wavelength of the
ground state is ∼ 4X, so the approximation does not apply. We can consider that the wavefunction
is negligible a short distance beyond the classical turning point X, and the wavefunction of the
nth excited state has n nodes, so 2X ≃ (n + 1)λ/2. Hence the quality of the wkbj approximation
increases rapidly with n.
12.6 Suppose the charge carried by a proton gradually decayed from its current value, e, being at
a general time f e. Write down an expression for the binding energy of a hydrogen atom in terms of
f . As α → 0 the binding energy vanishes. Explain physically where the energy required to free the
electron has come from.
When the spring constant of an oscillator is adiabatically weakened by a factor f 4 , the oscillator’s
energy reduces by a factor f 2 . Where has the energy gone?
In Problems 3.14 and 3.15 we considered an oscillator in its ground state when the spring
constant was suddenly weakened by a factor f = 1/16. We found that the energy decreased from
1
2 h̄ω to 0.2656h̄ω not to h̄ω/512. Explain physically the difference between the sudden and adiabatic
cases.
Soln: For hydrogen E = −R/n2 and R ∝ e4 (8.29). However, two of these powers of e relate
to the charge on the electron, so the binding energy decays as E = −f 2 R/n2 . If we imagine the
the proton’s charge has been evaporating, work has to be done to move positive charge out of the
electron’s sphere of negative charge. If the proton’s charge has simply been cancelled, a debt (the
energy released when the electron approached the proton) has simply been cancelled.
Imagine slowly tightening the spring of the oscillator. You would have to do work pulling the
oscillating mass in, and this work would manifest itself as an increase in the oscillator’s energy.
Conversely, when the spring is weakened, work is done on the mechanism that weakens the spring.
In the limit that the spring constant is made negligible, you can extract all the oscillator’s energy
because the mass moves slower and slower and finishes with negligible kinetic energy, and therefore
(by the virial theorem) negligible potential energy.
In the sudden case the mass retains whatever part of its energy is in kinetic form as the spring
is cut. So averaged over all phases of oscillation, it retains about half of its energy even when f is
reduced to zero.
12.7 Photons are trapped inside a cavity that has perfectly reflecting walls which slowly recede,
increasing the cavity’s volume V. Give a physical motivation for the assumption that each photon’s
frequency ν ∝ V −1/3 . Using this assumption, show that the energy density of photons u ∝ V −4/3
and hence determine the scaling with V of the pressure exerted by the photons on the container’s
walls.
Black-body radiation comprises an infinite set of thermally excited harmonic oscillators – each
normal mode of a large cavity corresponds to a new oscillator. Initially the cavity is filled with
black-body radiation of temperature T0 . Show that as the cavity expands, the radiation continues
to be black-body radiation although its temperature falls as V −1/3 . Hint: use equation (6.14).
Soln: Each photon is an excitation of a normal mode of the cavity. The number of nodes of this
mode is invariant as the cavity expands, so the wavelength increases in proportion to the lengths
of the cavity’s sides, i.e., λ ∝ V 1/3 . Consequently, the energy hν of each photon scales as V −1/3
and since the number of photons is invariant, their density scales as V −1 . Hence the energy density
scales as V −4/3 . By conservation of E, the energy lost by the photons must equal the work P dV
done on the mechanism that retains the walls. Thus
dV −1/3 = P dV ⇒ P ∝ V −4/3
so the photon gas has adiabatic index γ = 43 .
During an adiabatic change the probabilities pi of individual quantum states are invariant, so the
sum of these – the partition function Z – is invariant also. The partition function (6.14) of a single
mode is a function of the product βω = 2πβν, so a rescaling ν ∝ V −1/3 must be associated with a
rescaling β ∝ V 1/3 of the inverse temperature, so we have T ∝ V −1/3 . Since the partition functions
of all normal modes retain the black-body form, the radiation as a whole remains black-body during
the expansion.
12.8 Show that when a charged particle gyrates with energy E in a uniform magnetic field of flux
density B, the magnetic moment µ = E/B is invariant when B is changed slowly. Hint: recall
Problems 93
Problem 9.6. By applying the principle that energy must be conserved when the magnetic field is
slowly ramped up, deduce whether a plasma of free electrons forms a para- or diamagnetic medium.
Soln: In Problem 9.6 we showed that µ = E/B = n + 21 h̄Q/m, so it depends only on which Landau
level the particle is in. Since this level is adiabatically invariant, µ is too.
Since E/B is adiabatically invariant, the particle’s kinetic energy E has to increase
R with B. So
the coil that generates B has to supply more energy than just theR field energy, d3 x B 2 /2µ0 . By
Faraday’s law the work done to drive current I through the coil is dt IdΦ/dt, where Φ ∝ B is the
flux linked. So a larger current is required to generate a given B when electrons are gyrating in the
B field than otherwise. Hence these electrons must be generating a B field that partially cancels
that generated by the driving current I. That is, the electrons constitute a diamagnetic medium.
Problems
e ± defined by equation (13.23) obey
13.1 Show that the operators Ω
HΩ e± = Ωe ± (HK ± iǫ) ∓ iǫ. (13.1)
e
Soln: The operators Ω± are defined as
Z ∓∞
e ± ≡ limǫ→0+ 1 + i
Ω dτ U † (τ )V e±ǫτ /h̄ UK (τ ) .
h̄ 0
As in the text, it is convenient to approach this problem by multiplying Ω e ± on the left by U (t) and
differentiate wrt t before setting t = 0. We obtain
Z ∓∞
e ± − 1) = limǫ→0+ i
U (t)(Ω dτ U † (τ − t)V e±ǫτ /h̄ UK (τ )
h̄ 0
Z
i ∓∞
= limǫ→0+ dτ U † (τ − t)V e±ǫ(τ −t)/h̄ UK (τ − t) × e±ǫt/h̄ UK (t) .
h̄ 0
Introducing τ ′ = τ − t and being careful about the lower limit of the integral, this expression may
be written as
Z ∓∞ Z 0
′
lim+ + dτ ′ U † (τ ′ )V e±ǫτ /h̄ UK (τ ′ ) e∓ǫt/h̄ UK (t)
ǫ→0 0 −t
Z 0
′
e ± − 1) + lim
= (Ω dτ ′ U † (τ ′ )V e±ǫτ /h̄
UK (τ ′ ) e∓ǫt/h̄ UK (t) .
ǫ→0+ −t
Differentiating wrt t and setting t = 0 gives
d e ± − 1) = H(Ω
e ± − 1)
ih̄ U (t)(Ω
dt t=0
for the left hand side, while we get
d e ± − 1)(HK ∓ ǫ) − V
ih̄ rhs = (Ω
dt t=0
from the right hand side. Equating these two and rearranging gives the desired result.
13.2 Obtain the first- and second-order contributions to the S-matrix from the Feynman rules
given in §13.3.
Soln: At first order, the Feynman diagram has only one vertex and no propagators. The rules
immediately give
hp′ |S (1) |pi = −2πiδ(Ep − Ep′ )hp′ |V |pi ,
which is the Born approximation. At second order the Feynman diagram has two vertices separated
by a propagator, so we obtain
Z
hp′ |V |kihk|V |pi
hp′ |S (2) |pi = −2πiδ(Ep − Ep′ ) d3 k
Ep − Ek + iǫ
94 Problems
as found in the text by the more laborious procedure of expanding the S-matrix.
13.3 Derive the Lippmann–Schwinger equation
1
|±i = |Ei + V |±i, (13.2)
E − H K ± iǫ
where |±i are in and out states of energy E and |Ei is a free-particle state of the same energy. In
the case that the potential V = V0 |χihχ| for some state |χi and constant V0 , solve the Lippmann–
Schwinger equation to find hχ|±i.
Soln: The Lippmann–Schwinger equation is usually derived by rearranging the tise E|ψi = (H K +
V )|ψi to give (E − H K )|ψi = V |ψi and then arguing that the correct way to treat the non-invertible
operator (E − H K ) is to use an iǫ prescription such that
1
|ψi = |Ei + V |ψi ,
E − H K ± iǫ
where |Ei is any state in the kernel of (E − H K ). Here we will give an alternative derivation using
the operators Ω e ± . From the results of 13.1, we know that
H|±i = H Ω e ± |Ei = Ω
e ± (H K ± iǫ)|Ei ∓ iǫ|Ei
where in the first equality we used the relation |±i = Ω e ± |Ei between the in / out states and the
free states. Since we are acting on energy eigenstates, we have H|±i = (E ± iǫ)|±i ∓ iǫ|Ei, or
(E − H K ± iǫ)|±i = ±iǫ|Ei + V |±i .
The operator (E −H K ±iǫ) is invertible since the Hermitian operator HK always has real eigenvalues.
Thus we have
1 1
|±i = [±iǫ|Ei + V |±i] = |Ei + V |±i ,
E − H K ± iǫ E − H K ± iǫ
where the last line follows since H K |Ei = E|Ei by definition. This is the Lippmann–Schwinger
equation.
In the case V = V0 |χihχ|, contracting the Lippmann–Schwinger equation with hχ| and inserting
a complete set of eigenstates of H K gives
X V0 |hχhE ′ ||2
hχ|±i = hχ|Ei + hχ|±i
E − E ′ ± iǫ
E
′
so rearranging we find
hχ|Ei
hχ|±i = P V0 |hχhE ′ ||2
.
1− E′ E−E ′ ±iǫ
This solves for hχ|±i in terms of the coefficients of the (known) state |χi in the energy basis.
13.4 A certain central potential V (r) falls as r−n at large distances. Show that the Born approx-
imation to the total cross-section is finite if n > 2. Is this a problem with
R the Born approximation?
Soln: The Born cross-section is finite if and only if the integral I ≡ d3 x e−iq·x/h̄ V (x) is finite.
For a central potential, writing q · x = qr cos θ this integral is
Z Z Z
sin(qr) 4π
I = 4π r2 dr V (r) ≤ dr |r sin(qr)V (r)| = dr |rV (r)| .
qr q
The integral on the right converges at large r provided n > 2. This restriction is not so much a
problem for the Born approximation as a difficulty in applying scattering theory as a whole: our
whole framework assumed that the interaction term was negligible at large distances. Important
potentials such as the Coulomb potential fail this condition, and to compute the scattering we first
screen the potential as by a factor e−αr before taking the α → 0 limit at the end of the calculation.
In practice the Coulomb potential is always shielded. Its long-range nature is seen as a divergence
in the resulting momentum space amplitude as p → 0.
Problems 95
13.5 Compute the differential cross-section in the Born approximation for the potential V (r) =
V0 exp(−r2 /2r02 ). For what energies is the Born approximation justified?
Soln: In the Born approximation we have f (p → p′ ) = −4π 2 h̄mhp′ |V |pi. Converting to a position
basis Z
1
hp′ |V |pi = d3 x V (x)eiq·x
(2πh̄)3
and with out potential this is
Y3 Z Y3 2 Z
1 x2i e−(qi r0 ) /2 1 2 2
V0 dxi exp − 2 + iqi xi = V0 dxi exp − 2 (xi − iqi r0 )
i=1
2π 2r0 i=1
2πh̄ 2r0
by completing the square in the exponential. Using the fact that q = 2k sin(θ/2) this gives
mV0 r03
f (p → p′ ) = −4π 2 3 2
exp −2(kr0 )2 sin2 (θ/2)
(2π) 2 h̄
and the differential cross-section is
dσdΩ = |f (p → p′ )|2 = 2πm2 |V0 |2 r06 h̄4 exp −4(kr0 )2 sin2 (θ/2)
for the Gaussian potential.
13.6 When an electron scatters off an atom, the atom may be excited (or even ionised). Consider
an electron scattering off a hydrogen atom. The Hamiltonian may be written as H = H0 + H1 where
p̂2 e2 p̂2
H0 = 1 − + 2 (13.3)
2M 4πǫ0 r1 2m
is the Hamiltonian of the hydrogen atom (of mass M ) whose electron is described by coordinate r1 ,
together with the kinetic Hamiltonian of the scattering electron (of mass m), while
e2 1 1
H1 = − (13.4)
4πǫ0 |r1 − r2 | r2
is the interaction of the scattering electron with the atom.
By using H0 in the evolution operators, show that in the Born approximation the amplitude
for a collision to scatter the electron from momentum p2 to p′2 whilst exciting the atom from the
state |n, l, mi to the state |n′ , l′ , m′ i is
f (p2 ; n, l, m → p′2 ; n′ , l′ , m′ )
Z
4π 2 h̄m (13.5)
=− d3 r1 d3 r2 e−iq2 ·r2 hn′ , l′ , m′ |r1 ihr1 |n, l, miH1 (r1 , r2 ),
(2πh̄)3
where q2 is the momentum transferred to the scattering electron. (Neglect the possibility that the
two electrons exchange places, and the recoil of the hydrogen atom.)
Compute the differential cross-section for the |1, 0, 0i → |2, 0, 0i transition and show that when
the energy of the scattering particle is very high it varies as cosec12 (θ/2). Hint: perform the d3 r2
integral by including a factor e−αr2 and then let α → 0.
Soln: In our derivation of the S-matrix, all that we assumed about the unperturbed Hamiltonian
was that we knew its complete set of eigenstates. In the present case, this is true of the hydrogen
atom Hamiltonian H0 and so we can use this in the evolution operators. Let |n, l, m; p2 i be the state
in which the hydrogen atom is found in state |n, l, mi and the scattering electron has momentum
p2 . Then the Born approximation to the desired amplitude is
−4πh̄2 mhn′ , l′ , m′ ; p2 ′ |H1 |n, l, m; p2 i
Z
= −4πh̄2 m d3 r1 d3 r2 hn′ , l′ , m′ ; p2 ′ |r1 ; r2 i H1 (r1 , r2 )hr1 , r2 |n, l, m; p2 i
96 Problems
where we have used the fact that the perturbing Hamiltonian H1 is diagonal in the position basis.
But hr1 , r2 |n, l, m; p2 i = hr1 |n, l, mi hr2 |p2 i and the second factor is just a plane wave. We therefore
get Z
4π 2 h̄m
f =− d3 r1 d3 r2 e−iq2 ·r2 hn′ , l′ , m′ |r1 ihr1 |n, l, miH1 (r1 , r2 ),
(2πh̄)3
as required.
Let us write ψ1 (r1 ) = hr1 |1, 0, 0i and ψ2 (r1 ) = hr1 |2, 0, 0i. Then the desired scattering amplitude
4π 2 h̄m
is − (2πh̄)3 I, where I is the integral
Z
I = d3 r1 d3 r2 e−iq2 ·r2 H1 (r1 , r2 )ψ2 (r1 ) ψ1 (r1 )
where q = (p2 − p2 ′ )/h̄ is the momentum transfer for the scattering electron. Let us consider the
r2 integral first. We have
Z Z
d3 r2 e−iq2 ·r2 H1 (r1 , r2 ) = lim d3 r2 e−iq2 ·r2 −αr2 H1 (r1 , r2 )
α→0+
and with this convergence factor we can take each term in H1 separately. For the first term we have
Z Z
3 −iq2 ·r2 −αr2 1 4π ∞ 4π
lim d r2 e = lim dr2 sin(qr2 )e−αr2 = 2 .
α→0+ r2 α→0+ q 0 q
The same integral can be used to compute the d3 r2 integral involving the second term in H1 upon
using the substitution r′ 2 = r2 − r1 . Thus we find
Z Z
e2 3 iq·r1
e2
I= d r1 e − 1 ψ2 (r3 )ψ1 (r1 ) = d3 r1 eiq·r1 ψ2 (r3 )ψ1 (r1 )
ǫ0 q 2 ǫ0 q 2
where in the second equality we have used the orthogonality of the hydrogenic wavefunctions. Using
the wavefunctions
2 2 r
ψ1 (r) = √ e−r/a ψ2 (r) = p e−r/2a 1 −
4πa3 4π(2a)3 2a
where a is the Bohr radius a = 4πǫ0 /h̄2 µ, we find
Z Z
1 r
dr eiq·r ψ2 (r)ψ1 (r) = 3 dr eiq·r e−3r/2a 1 −
a π 2a
Z ∞
4 r
= 3 dr re−3r/2a 1 − sin(qr)
a q 0 2a
so that Z ∞
4π 2 h̄m 4e2 −3r/2a r
f =− dr re 1 − sin(qr) .
(2πh̄)3 (aq)3 ǫ0 0 2a
The remaining integral can be performed by writing the sin(qr) as exponentials and then integrating
by parts to remove the polynomial factor in r. The result is
−3
′
√ 2 2 2 2 9
f (ψ1 (r1 ), p2 → ψ2 (r1 ), p2 ) = −8 2ma e h̄ (qa) +
4
and the differential cross-section is the mod-square of this.
To compute the high energy behaviour, note that
2
h̄q = |p2 − p2 ′ | = p22 + (p′2 ) − 2p2 p′2 cos θ
where θ is the scattering angle. This is an inelastic collision – the hydrogen atom has been excited –
(p22 − (p′2 )2 )/2m will differ from zero by three quarters of a Rydberg. However, in the limit that the
Problems 97
scattering electron has very high energy we may neglect this difference and treat h̄q ≃ 2p sin2 (θ/2)
as in elastic collisions. Then we find
dσ ae 4
→ 2 7 m2 (qa)−6 ∝ csc12 (θ/2)
dΩ h̄
as was to be shown.
13.7 Use the optical theorem to show that the first Born approximation is not a valid description
of forward scattering.
Soln: The optical theorem states that the imaginary part of the forward scattering amplitude
f (p → p) governs the total cross-section
p
Imf (p → p) = σ(p) .
4πh̄
However, in the first Born approximation, f (p → p) = −4π 2 h̄mhp|V |pi which is real since V is
Hermitian. Using the first Born approximation in the optical theorem would lead us to conclude
that the total cross section must always vanish for any potential!
13.8 A particle scatters off a hard sphere, described by the potential
n
V (r) = ∞ for |r| ≤ a (13.6)
0 otherwise.
By considering the form of the radial wavefunction u(r) √ in the region r > a, show that the phase
shifts are given by tan δl = jl (ka)/nl (ka), where k = 2mE/h̄ and jl (kr) and nl (kr) are the two
independent solutions (spherical Bessel functions) of the second-order radial equation
1 d 2 d l(l + 1) 2mE
r u(r) = − 2 u(r), (13.7)
r2 dr dr r2 h̄
with jl (kr) regular at the origin.
In the limit kr → 0, show that these functions behave as
jl (kr) ∼ (kr)l nl (kr) ∼ 1/(kr)l+1 . (13.8)
l l+1
In fact, the limiting values are jl (kr) → (kr) /(2l + 1)!! and nl (kr) → −(2l − 1)!!/(kr) where
(2l +1)!! ≡ (2l +1)(2l −1)(2l −3) · · · 3·1. Use this to show that in the low-energy limit, the scattering
is spherically symmetric and the total cross-section is four times the classical value.
Soln: In the region r > a outside the hard sphere, the wavefunction must satisfy the free Schrodinger
equation. For spherically symmetric potentials we have seen that the S-matrix has no dependence
on the azimuthal angle, and so we can expand the wavefunction in terms of Legendre polynomials
as ∞
X
ψ(r) = (2l + 1)il Pl (cos θ)ul (r)
l=0
where ul (r) is a solution of the radial equation. We can thus write it as a linear combination
ul (r) = al jl (kr) + bl nl (kr)
of the two spherical Bessel functions. Applying the boundary condition ul (a) = 0 so that the
wavefunction vanishes at the surface of the hard sphere we find bl = −al jl (ka)/nl (ka). We must
now relate this solution to the scattering amplitude. Using the asymptotic behaviour
1 1
lim jl (kr) = cos(kr − (l + 1)π/2) lim nl (kr) = sin(kr − (l + 1)π/2)
r→∞ kr r→∞ kr
we see that as r → ∞ our radial wavefunction behaves as
al Al
lim ul (r) = (nl (ka) cos(kr−(l+1)π/2)−jl (ka) sin(kr−(l+1)π/2) = sin(kr−lπ/2+∆l )
r→∞ nl (ka)kr kr
98 Problems
where tan ∆l ≡ jl (ka)/nl (ka) and Al = al /nl (ka). To see that this ∆l is nothing but the phase shift
δl defined in the scattering amplitude, recall from (eq 13.45) that f (θ) is just the amplitude of the
outgoing spherical wave in the asymptotic region:
eikr
lim ψ(r) = eikr cos θ + f (θ) .
r→∞ r
(We are using the fact that, for spherical potentials, the scattering amplitude is independent of the
azimuthal angle to write f (p → p′ ) as f (θ).) From the expansion
∞
X ∞
X cos(kr − (l + 1)π/2)
eikr cos θ = (2l + 1)il jl (kr)Pl (cos θ) → (2l + 1)il
kr
l=0 l=0
together with the asymptotic behaviour of the spherical Bessel function, we see that f (θ) satisfies
X∞ ∞
X
eikr al
(2l + 1)il jl (krPl (cos θ) + f (θ) = (2l + 1)il Pl (cos θ) sin(kr − lπ/2 + ∆l ) , (13.9)
r nl (ka)kr
l=0 l=0
where we have equated our two Pexpressions for the asymptotic form of the wavefunction. The
partial wave expansion kf (θ) = l (2l + 1)Pl (cos θ)eiδl sin δl of the scattering amplitude allows us to
combine the terms on the left into a single sum over l, and for (13.9) to hold it must then be that
the coefficients of e±ikr are separately equal. This gives the two equations
X il X
(2l + 1)Pl (x) ei(l+1)π/2 = (2l + 1)Pl (x)il 2i Al eilπ/2−∆l
2
l l
X l
X
i
(2l + 1) eiδl sin δl + e−i(l+1)π/2 Pl (x) = (2l + 1)Pl (x)il 2i Al e−ilπ/2+i∆l .
2
l l
The first of these enables us to solve for Al , finding Al = ei∆l . Using this in the second shows that
indeed ei∆l = eiδl so we have shown that
jl (ka)
tan δl =
nl (ka)
for hard sphere scattering, as required.
In the opposite limit kr → 0, we can neglect the final term in the radial equation
d du
r2 = l(l + 1) − (kr)2 u .
dr dr
Writing u(r) = Ul (r)/r we have that U (r) satisfies
d2 U U
2
= l(l + 1) 2
dr (r)
in this limit, which is solved by U (r) = A(kr)l+1 + B(kr)−l or
u(r) = A(kr)l + B(kr)−(l+1) ,
which are therefore the limits of the spherical Bessel functions jl (kr) and nl (kr), respectively. (The
power of k is determined either by dimensional analysis or by relabelling r in the equation for U .)
The total scattering cross-section is
4π 2 X
σtot = 2 (2l + 1) sin2 δl
k
l
2 2 2
Since sin δl = tan δl /(1 + tan δl ) = jl (ka) /(jl (ka)2 + nl (ka)2 ), our previous result shows that
2
(ka)4l+2
lim sin2 δl (k) = .
k→0 1 + (ka)4l+2
Problems 99
Thus in this limit only the l = 0 term survives. The total cross-section becomes
4π 2
σtot = (ka)2 = 4π 2 a2
k2
This is four times greater than the classical value πa2 , the area of a disc presented by the sphere in
the plane orthogonal to the direction of the incoming particle.
13.9 Show that in the Born approximation the phase shifts δl (E) for scattering off a spherical
potential V (r) are given by
Z
2mk ∞ 2
δl (E) ≃ − 2 dr r2 V (r) (jl (kr)) . (13.10)
h̄ 0
When is the Born approximation valid? Hint: with q = 2k sin(θ/2) we have
∞
sin(qr) X
= (2l + 1)jl (kr)2 Pl (cos θ).
qr
l=0
Soln: In the Born approximation the scattering amplitude f (p → p′ ) = −4π 2 h̄mhp′ |V |pi and for
a spherically symmetric potential this is
Z
f (θ) = −4π 2 h̄m d3 r′ d3 r hp′ |r′ ihr′ |V |rihr|pi
Z
1 2m sin(qr)
=− 2 dr r2 V (r)
4π h̄ qr
where q = |p′ − p|/h̄ = 2k sin(θ/2) is the momentum transfer. Using the given expression for
sin(qr)/(qr) in terms of spherical Bessel functions and equating this to the partial wave expansion
we find
X Z
4π 2 h̄pm X
(2l + 1)Pl (cos θ)eiδl sin δl = − (2l + 1) dr r2 V (r)jl (kr)2 Pl (cos θ) .
(2πh̄)3
l l
For this equation to hold for all θ we must be able to equate coefficients of Pl (cos θ). Furthermore,
in the Born approximation we expect the phase shift to be small. Hence we find
Z ∞
iδl mp
δl (E) ∼ e sin δl = − dr r2 V (r)jl (kr)2
2πh̄2 0
as was to be shown.
13.10 Two α particles collide. Show that when the α particles initially have equal and opposite
momenta, the differential cross-section is
dσ
= |f (θ) + f (θ − π)|2 . (13.11)
dΩ
Using the formula for f (θ) in terms of partial waves, show that the differential cross-section at
θ = π/2 is twice what would be expected had the α particles been distinguishable.
A moving electron crashes into an electron that is initially at rest. Assuming both electrons are
in the same spin state, show that the differential cross-section falls to zero at θ = π/4.
Soln: Since the α particles are indistinguishable, we must add the amplitude for scattering through
an angle θ to the amplitude for scattering through an angle π − θ, since these two possibilities lead
to indistinguishable final states. This immediately gives the desired formula for the differential
cross-section.
The lth Legendre polynomial obeys Pl (− cos θ) = (−1)l Pl (cos θ), and so is an odd or even
function of cos θ depending on whether l is odd or even. In particular, all odd Legendre polynomials
vanish when θ = π/2, so at θ = π/2 only the even partial waves survive. Hence f (π/2) = f (−π/2)
100 Problems
and so the differential cross-section is 4|f (π/2)|2 . By contrast, for distinguishable particles we would
add the probabilities rather than the amplitudes:
dσ
= |f (θ)|2 + |f (π − θ)|2 .
dΩ dist
Thus we would obtain 2|f (π/2)|2 at θ = π/2, so the differential cross-section in the indistinguishable
case is twice as big here.
In the case of electrons, Fermi statistics mean that the total wavefunction is antisymmetric
under the exchange of all quantum numbers of the two particles. Since they are in the same spin
state, the scattering amplitude must be antisymmetric under exchange of the final momenta of the
particles. Thus the differential cross-section is now
dσ
= |f (θ) − f (π − θ)|2
dΩ
where θ is the scattering angle in the centre of mass frame. In contrast to the bosonic α particles,
this expression now vanishes when θ = π/2 because the two scattering amplitudes cancel each other
out. A scattering angle of π/2 in the centre of mass frame corresponds to a scattering angle of π/4
in the frame in which one of the electrons was initially at rest.