Quantum Mechanics Problems and Solutions
Quantum Mechanics Problems and Solutions
Quantum Mechanics
Problems and Solutions
K. Kong Wan
June 23, 2020 10:49 JSP Book - 9in x 6in 00-Solution˙Manual-Prelims
Published by
Jenny Stanford Publishing Pte. Ltd.
Level 34, Centennial Tower
3 Temasek Avenue
Singapore 039190
Email: [email protected]
Web: www.jennystanford.com
All rights reserved. This book, or parts thereof, may not be reproduced in any
form or by any means, electronic or mechanical, including photocopying,
recording or any information storage and retrieval system now known or to
be invented, without written permission from the publisher.
Contents
Preface xi
2 Classical Systems 3
11 Complex Vectors 55
viii Contents
17 Operators in a Hilbert Space H 99
18 Bounded Operators on H 107
19 Symmetric and Selfadjoint Operators in H 115
20 Spectral Theory of Selfadjoint Operators in H 121
21 Spectral Theory of Unitary Operators on H 127
Contents ix
Bibliography 229
June 23, 2020 10:49 JSP Book - 9in x 6in 00-Solution˙Manual-Prelims
Preface
K. Kong Wan
St Andrews
Scotland
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 1
Chapter 2
Classical Systems
Chapter 3
E1 ⊥ E2 = E1 and E 1 − E 2 = ∈.
(a) Write down the probability mass function ℘ and evaluate the
expectation value and the uncertainty.
(b) Write down the corresponding probability distribution function
F(τ ) and sketch a plot of F(τ ) versus τ . What are the values
F(τ ) at τ = 0.9, 1, 2.5, 6 and 6.1?
� �
The uncertainty = variance with an approximate value 2.9.
SQ3(4)(b) The probability distribution function F(τ ) is
piecewise-constant with discontinuous steps occurring at τ =
1, 2, 3, 4, 5, 6. Explicitly F(τ ) is related to ℘(aℓ ) by
⎧
⎪
⎪ 0, τ <1
⎪
⎪
⎪
⎪ 1/6, 1 ≤ τ <2
⎪
⎪
⎪
⎨ 2/6, 2 ≤ τ <3
F(τ ) = 3/6, 3 ≤ τ < 4 .
⎪
⎪
⎪ 4/6, 4 ≤ τ < 5
⎪
⎪
⎪
⎪
⎪ 5/6, 5 ≤ τ < 6
⎪
⎩
1, 6≤τ
The function F(τ ) is continuous from the right for all τ . A plot of
F(τ ) against τ is given below:
F(τ )
0 1 2 3 4 5 6 τ
We have F(0.9) = 0, F (1) = 1/6, F(2.5) = 2/6, F(6) = 1,
F(6.1) = 1.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 4
(2) Eq. (4.19) is part of the defining properties of the measure given
in Eq. (4.17), bearing in mind that g(τ + 0) = g(τ ).
(3) Eq. (4.20) is part of the defining properties of the measure given
in Eq. (4.16).
(4) To prove of Eq. (4.21) we first observe that
Mls, g (τ1 , τ2 ] = Mls, g (τ1 , τ2 ) ∅ {τ2 }
= Mls, g (τ1 , τ2 ) + Mls, g {τ2 }
= Mls, g (τ1 , τ2 ) + g(τ2 ) − g(τ2 − 0).
It follows that
Mls, g (τ1 , τ2 ) = Mls, g (τ1 , τ2 ] − g(τ2 ) − g(τ2 − 0)
= g(τ2 − 0) − g(τ1 ).
1 Papoulis p. 143.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
= b.
The first integral vanishes since the integrand is an odd function of x.
The variance is calculated using Eq. (4.92), i.e.,
→
1 2 2
Var(FG ) = � τ 2 e(τ −b) /2a dτ − E(FG )2
2πa2 −→
→
1 2 2
= � τ 2 e−(τ −b) /2a dτ − b2
2
2πa −→
→
1 2 2 2
= � x + b e−x /2a dx − b2
2
2πa −→
→
1 2 2 2
= � x + 2xb + b2 e−x /2a dx − b2 (�)
2
2πa −→
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
→
1 2
/2a2
= � x 2 e−x dx
2πa2 −→
= a2 ,
due to the integral
→
2 −λx 2 1 π
x e dx = .
−→ 2λ λ
�
The uncertainty is then (FG ) = Var(FG ) = a. Note that there
are three integrals in the expression (�) above. The middle integral
is zero on account of an odd integrand. The last integral is equal to
b2 since the integral of FG (τ ) is equal to 1.
Q4(12) Find the probability distribution function FU (τ ) given by
the density function wU (τ ) in Eq. (4.81).
SQ4(12) The uniform probability distribution function is given by
⎧
τ ⎨0 τ ≤a
FU (τ ) = wU (τ � ) dτ � = (τ − a)/(b − a) x ⇒ (a, b) .
−→ ⎩
1 x⊕b
Chapter 5
Chapter 6
SQ6(1)
(1) To prove Eq. (6.1) we start with the premise v∀1 + u∀ = v∀2 + u.
∀
Adding u∀ −1 on both side this equation we get
v∀1 + u∀ + u∀ −1 = v∀2 + u∀ + u∀ −1
� v∀1 + 0∀ = v∀2 + 0∀ � v∀1 = v∀2 .
Q6(3) Verify that the expression for �u∀ | v∀ ∪ in Eq. (6.13) satisfies
properties SP6.3.1(1), SP6.3.1(2) and SP6.3.1(3) of scalar product.
SQ6(3)
SQ6(4) Let v∀1 and v∀2 be two (non-zero) orthogonal vectors and a1
∀ Then:
and a2 be two real numbers such that a1 v∀1 + a2 v∀2 = 0.
�∀
v 1 | a1 v∀1 + a2 v∀2 ∪ = 0
v 1 | a1 v∀1 + a2 v∀2 ∪ = a1 �∀
� �∀ v 1 | v∀1 ∪ + a2 �∀
v 1 | v∀2 ∪
= a1 �∀
v 1 | v∀1 ∪ = 0
� a1 = 0 � a2 = 0.
The two vectors are therefore linearly independent.
One can further check this by assuming linear dependence, i.e.,
v∀1 = a v∀2 . We will then have
v 1 | v∀1 ∪ = a �∀
�∀ v 1 | v∀2 ∪ = 0,
v 1 | v∀1 ∪ = || v∀1 ||2 �= 0.
which is a contradiction since �∀
Q6(5) Prove Eq. (6.19).
SQ6(5) Take the scalar product of e∀ℓ and v∀ we get
3
3
�e∀ℓ | v∀ ∪ = �e∀ℓ | cℓ� e∀ℓ� ∪ = cℓ� �e∀ℓ | e∀ℓ� ∪
ℓ� =1 ℓ� =1
3
= cℓ� δℓℓ� = cℓ .
ℓ� =1
Note that we have rewritten Eq. (6.19) as 3ℓ� =1 cℓ� e∀ℓ� rather than
3
the original expression ℓ=1 cℓ e∀ℓ which will cause confusion when
taking the scalar product with e∀ℓ . If one wishes to use the original
expression one should take the scalar product with e∀ℓ� which will
show that cℓ� = �e∀ℓ� | v∀ ∪.
Q6(6) Prove the Pythagoras theorem in the forms of Eqs. (6.22)
and (6.27).
SQ6(6) For the Pythagoras theorem in Eq. (6.22) we have
v | v∀ ∪ = v x v x � i∀ | i∀ ∪ + v x v y � i∀ | ∀j ∪ + · · · .
v ||2 = �∀
||∀
The orthonormality of the basis vectors i∀, ∀j , k∀ immediately leads to
the desired result.
For the Pythagoras theorem in Eq. (6.27) we have
3
3
3
�u∀ | v∀ ∪ = � uℓ e∀ℓ | vℓ� e∀ℓ� ∪ = uℓ vℓ� �e∀ℓ | e∀ℓ� ∪.
ℓ=1 ℓ� =1 ℓ, ℓ� =1
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Similarly we have
||∀ v − u∀ + u∀ || ≤ ||∀
v || = ||∀ v − u∀ || + ||u∀ ||
v − u∀ || = ||u∀ − v∀ ||.
� || v∀ || − || u∀ || ≤ ||∀
It follows that
|| u∀ || − || v∀ || ≤ ||u∀ − v∀ ||.
Chapter 7
The
sameresult is obtained
if we interchange M and N , i.e., we get
tr M · N = tr N · M .
Q7(2) Show that the inverse of a square invertible matrix is
unique.
SQ7(2) Let M −1 be an inverse of M and let N be another inverse
of M , i.e., we have
M −1 · M = M · M −1 = I and N · M = M · N = I .
Then
M −1 = M −1 · I = M −1 · M · N = M −1 · M · N = N .
In other words there is no other inverse.
N · M = I and M · C = 0 � C = 0.
Q7(5) Show that the orthogonal matrix R z (θz ) in Eq. (7.149) is not
selfadjoint.
SQ7(5) The matrix R z (θz ) in Eq. (7.149) has real elements. This
means that its transpose R zT (θz ) is equal to its adjoint R†z (θz ). But
R zT (θz ) not equal to R z (θz ). It follows that its adjoint R†z (θz ) is
different from R z (θz ), i.e., R z (θz ) is not selfadjoint.
Q7(6) Prove Eq. (7.160).
SQ7(6) By Eq. (7.81) we get
�U C | U C ∪ = �U † U C | C ∪, �U C 1 | U C 2 ∪ = �U † U C 1 | C 2 ∪.
�U † U C | C ∪ = �C | C ∪, �U † U C 1 | C 2 ∪ = �C 1 | C 2 ∪.
The cross terms are all zero since Pauli matrices anticommute, e.g.,
ax ay σx · σ y + σ y · σx = 02×2 . Next we can verify the fact that the
square of each Pauli matrix is the 2 × 2 identity matrix I 2×2 , e.g.,
0 1 0 1 1 0
σx2 := = = I 2×2 .
1 0 1 0 0 1
It follows that
2
ax σx + ay σ y + az σz = ax2 + ay2 + az2 I 2×2 .
0 1 1 0
σx C α∀ z = · = ,
1 0 0 1
and
0 1 0 1
σx C β∀z = · = .
1 0 1 0
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
2
P·Q = P·Q · Q·P =P·Q·P=P·P·Q
= P · Q.
In other words P · Q is selfadjoint and idempotent. It follows that
P · Q is a projection matrix.
SQ7(9)(c) First, P+Q is clearly selfadjoint. For idempotence we
have
2
P + Q = P2 + Q2 + P · Q + Q · P = P + Q,
since P · Q = Q · P = 0. It follows that P + Q is a projection matrix.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
†
P α∀ y C α∀ y = C α∀ y C α∀ y C α∀ y = C α∀ y ,
†
P α∀ y C β∀y = C α∀ y C α∀ y C β∀y = 0 C β∀y .
†
P β∀y C α∀ y = C β∀y C β∀ C α∀ y = 0 C α∀ y ,
y
†
P β∀y C β∀y = C β∀y C β∀ C β∀y = C β∀y .
y
†
P α∀ z C α∀ z = C α∀ z C α∀ z C α∀ z = C α∀ z ,
†
P α∀ z C β∀z = C α∀ z C α∀ z C β∀z = 0 C β∀z .
1 Isham p. 91.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
†
P β∀z C α∀ z = C β∀z C β∀ C α∀ z = 0 C α∀ z ,
z
†
P β∀z C β∀x = C β∀z C β∀ C β∀z = C β∀z .
x
Chapter 8
∀3
Operations on Vectors in IE
3
∀
36 Operations on Vectors in IE
The product M Â · M Â −1 has elements M Â · M Â −1 km given by
MÂ kℓ
MÂ −1 ℓm
ℓ
= �e∀k | Ae∀ℓ ∪�e∀ℓ | A−1 e∀m ∪
ℓ
= � A† e∀k | e∀ℓ ∪�e∀ℓ | A−1 e∀m ∪ = �A† e∀k | A−1 e∀m ∪
ℓ
3
∀
Operations on Vectors in IE 37
It follows {e∀ℓ } is also a basis for IE∀ 3 and any vector v∀ ⇒ IE∀ 3 can be
written as v∀ = ℓ vℓ e∀ℓ . Then
∀ �
Av∀ = 0 vℓ Ae∀ℓ = ∀
vℓ e∀ �ℓ = 0.
ℓ ℓ
Since e∀ �ℓ
are linearly independent we must have vℓ = 0 ∞ℓ, i.e.,
∀ We can conclude that A is invertible by Theorem 8.2.2(1).
v∀ = 0.
3
∀
38 Operations on Vectors in IE
This means that the transpose of M is equal to M† . We arrive at the
desired result since the transpose of a real square matrix is equal the
adjoint of the matrix. All quantities are real in IE∀ 3 , e.g., �Ae∀l | e∀m ∪ =
�e∀m | Ae∀l ∪.
∀ ) = 0.
Q8(7) Let F be a linear functional on IE∀ 3 . Show that F (0
3
Prove a similar result for linear operators.
SQ8(7) For a linear functional F we have F (au∀ ) = aF (u∀ ). Now,
∀ it follows that
let a = 0 we get F (0u∀ ) = 0F (u∀ ) = 0. Since 0u∀ = 0
∀
F (0 ) = 0.
For a linear operator A we also have A au∀ = a Au∀. Now, let
a = 0 we get A(0u∀ = 0Au∀ = 0. ∀ Since 0u∀ = 0 ∀ it follows that
∀ = 0.
A 0 ∀
3
∀
Operations on Vectors in IE 39
In (2) and (3) above we have used Eq. (6.31) which tells us that
�u∀ | Av∀ ∪ = �u∀ | B ∀ v∀ ⇒ IE∀ 3 � A = B
v∀ ∪ ∞ u, ∀.
3
∀
40 Operations on Vectors in IE
2
Pi∀ e∀ = Pi∀ Pi∀ e∀ = a Pi∀ e∀ = a2 e∀ � a2 = a
� a = 0 or 1.
Eigenvectors corresponding to the eigenvalue 1 is of the form u x i∀
while eigenvectors corresponding to the eigenvalue 0 is of the form
u y ∀j + u z k∀. The eigenvalue 1 is nondegenerate since we do not
count two eigenvectors differing only by a multiplicative constant
as distinct. The eigenvalue 0 is degenerate with degeneracy 2.
Any linear combination of ∀j and k∀ is also an eigenvector of P ∀ i
corresponding to the eigenvalue 0.
SQ8(9)(d) The given matrix M P̂ i∀ is clearly selfadjoint and idempo
tent. It follows that M P̂ i∀ is a projection matrix by Definition 7.7.5(1).
By the same argument presented for the preceding question we can
see that the matrix has two eigenvalues 1 and 0. The eigenvector
corresponding to the eigenvalue 1 is
⎛ ⎞
1
⎝0⎠,
0
and the eigenvectors corresponding to the eigenvalue 0 are linear
combinations of
⎛ ⎞ ⎛ ⎞
0 0
⎝ 1 ⎠ and ⎝ 0 ⎠ .
0 1
The eigenvalue 1 is nondegenerate while the eigenvalue 0 is
degenerate with degeneracy 2.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 9
Then
� r 2 = 1 � r = ±1.
Since
we get
v | Aw
�∀ v � Aw
∀ ∪ = 0 ∞∀ ∞w
∀ =0 ∀ � A =
0.
where
This agrees with the expression for �u∀ | Rz (θz )v∀ ∪ obtained earlier.
Q9(12) Prove Eq. (9.70).
SQ9(12) Let
Since e∀ �21 and e∀ �22 are orthonormal the coefficients must satisfy
We can also express e∀21 and e∀21 in terms of e∀ �21 and e∀ �21 , i.e.,
β2 e∀ �21 − α2 e∀ �22 β1 e∀ �21 − α1 e∀ �22
e∀21 = , e∀22 = .
β2 α1 − α2 β1 β1 α2 − α1 β2
Since e∀21 and e∀22 are
Since β22 + β12 = α22 + α12 = 1 we get α12 + β12 = 1 which in turn
implies α22 + β22 = 1.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
These results imply that A† A and A A† are selfadjoint. Moreover we
have
v | A A† v∀ ∪ = �A† v∀ | A† v∀ ∪ = || A† v∀ ||2 ⊕ 0,
�∀
v | A† A v∀ ∪ = �Av∀ | Av∀ ∪ = || Av∀ ||2 ⊕ 0.
�∀
These two equations hold for all v∀, including the eigenvectors of A† A
and A A† . It follows that the product operators do not have negative
eigenvalues, i.e., both product operators are positive.
Q9(15) Show that the expression of the inverse in Eq. (9.86)
satisfy the condition A A−1 = II .
−1
SQ9(15) Taking the expressions for A in Eq. (9.79) and A in Eq.
(9.86) we get
−1
A A = am� P Â (am� ) −1 Â
am P (am )
m� m
= −1
am� am P Â (am� )P Â (am )
m� , m
−1 Â
= am� am P (am� )δm� m
m� , m
= P Â (am ) = II .
m
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
We have used the result P Â (am� ) P Â (am ) = P Â (am ) δm� m and the
spectral decomposition of the identity in Eq. (9.80).
Q9(16) The square root A of a positive operator A is defined by
2
Eq. (9.87). Show that
A = A.
SQ9(16) From Eq. (9.87) we get
2 � �
A = am P Â (am ) am� P Â (am� )
m m�
� �
= am am� P Â (am ) δmm�
m, m�
= am P Â (am ) = A.
m
We have used the result P Â (am ) P Â (am� ) = P Â (am ) δmm� and the
spectral decomposition of the A in Eq. (9.79).
Q9(17) Show that the quadratic form Q(A, u∀ ) generated by a
positive selfadjoint operator in any u∀ is non-negative.
SQ9(17) Since the quadratic form Q(A, u∀ ) is given by �u∀ | A u∀ ∪ we
get, for a positive selfadjoint operator A,
Q(A, u∀ ) = �u∀ | am P Â (am ) u∀ ∪, am ⊕ 0
m
= am �u∀ | P Â (am ) u∀ ∪
m
⊕ 0,
since the eigenvalues am of a positive selfadjoint operator A is non
negative by definition 9.4.4(2) and
2
�u∀ | P Â (am ) u∀ ∪ = �u∀ | P Â (am ) u∀ ∪
= �P Â (am )u∀ | P Â (am ) u∀ ∪
= || P Â (am )u∀ ||2 ⊕ 0.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 10
SQ10(1)
(1) For the first state vector u∀1 = e∀1 the probability of getting the
value a1 on a measurement of A is 1 with expectation value
E(A, u∀1 ) = a1 and zero uncertainty.
(2) For the second state vector u∀2 = e∀2 the probability of getting
the value a1 on a measurement of A is 0 with expectation value
E(A, u∀2 ) = a2 and zero uncertainty.
� �
(3) For the third state vector u∀3 = e∀1 + 2 e∀3 / 3 the probability
of getting the value a1 on a measurement of A is 1/3 with
52 Probability, Selfadjoint Operators, Unit Vectors and the Need for Complexness
expectation value
1
E(A, u∀3 ) = a1 + 2a3 .
3
The variant can be calculated in accordance with Eq. (3.33), i.e.,
we have
2 1 2 2
Var(A, u∀3 ) = a1 − E(A, u∀3 ) + a3 − E(A, u∀3 )
3 3
4 2 2 2
= a1 − a3 + − a1 + a3
27 27
2 2
= a1 − a3 .
9
The uncertainty is then given by
�
2
( A, u∀3 ) = Var(A, u∀3 ) = | a1 − a3 |.
3
The variant can also be calculated in accordance with Eq. (3.36),
i.e., we have
1 2 2
Var( A, u∀3 ) = a12 + a32 − E(A, u∀
3 3
1 2 1 2
= a1 + 2a32 − a1 + 2a3
3 9
2 2
= a1 − a3 .
9
SQ10(2) Writing
3
3
A = aℓ P e∀ℓ and A� = a�ℓ� Pe∀ℓ� .
ℓ=1 ℓ� =1
we get
3
3
3
A A� = aℓ a�ℓ� Paℓ Paℓ� = aℓ aℓ� � Paℓ δℓℓ� = aℓ aℓ� Paℓ .
ℓ, ℓ� =1 ℓ, ℓ� =1 ℓ=1
3
3
3
A� A = aℓ� � aℓ Paℓ� Paℓ = aℓ� � aℓ Paℓ δℓℓ� = aℓ� aℓ Paℓ .
ℓ� , ℓ=1 ℓ, ℓ� =1 ℓ=1
It follows that A A − A A =
� �
0, i.e., A and A commute. �
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Probability, Selfadjoint Operators, Unit Vectors and the Need for Complexness 53
Chapter 11
Complex Vectors
Q11(2) Let
1 1
ǫ∀1 = � i∀c + i ∀j c , ǫ∀2 = � i∀c − i ∀j c , ǫ∀3 = k∀ c .
2 2
(a) Write down the matrix representations of the above vectors
in the orthonormal basis {i∀c , ∀j c , k∀ c }.
56 Complex Vectors
Complex Vectors 57
SQ11(2)(c) The proof is the same as that in SQ11(1). Since ǫ∀1 , ǫ∀2
and ǫ∀3 form an orthonormal basis we can express any vector ζ∀ as
3
a linear combination of ǫ∀1 , ǫ∀2 and ǫ∀3 , i.e., we have ζ∀ = ∀ℓ .
ℓ=1 ζℓ ǫ
Taking the scalar product of ǫ∀ℓ and ζ∀ we get
3
3
�∀ǫℓ | ζ∀ ∪c = �∀ǫℓ | ζℓ� ǫ∀ℓ� ∪c = ζℓ� �∀ǫℓ | ǫ∀ℓ� ∪c = ζℓ .
ℓ� =1 ℓ� =1
Chapter 12
|| ζ∀ + η∀ ||2 = �ζ∀ + η∀ | ζ∀ + η∀ ∪
= �ζ∀ | ζ∀ ∪ + �η∀ | η∀ ∪ + �ζ∀ | η∀ ∪ + �η∀ | ζ∀ ∪
= ||ζ∀ ||2 + ||η∀ ||2 + 2 × real part of �η∀ | ζ∀ ∪
≤ ||ζ∀ ||2 + ||η∀ ||2 + 2 × |�η∀ | ζ∀ ∪|.
1 Roman p. 419. Prugovečki pp. 19–20.
2 Work out ||ζ∀ + η∀ ||2 = �ζ∀ + η∀ | ζ∀ + η∀ ∪ and ||ζ∀ − η∀ ||2 = �ζ∀ − η∀ | ζ∀ − η∀ ∪. Note that
�ζ∀ | η∀ ∪ + �η∀ | ζ∀ ∪ = 2× the real part of �ζ∀ | η∀ ∪ which is less than or equal to 2|�ζ∀ | η∀ ∪|.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 13
SQ13(4)
(1) For uniqueness let us suppose there are two such decomposi
tions
ζ∀ = ζ∀ S∀ + ζ∀ S∀� and ζ∀ = ζ∀ �S∀ + ζ∀ �S∀� .
Then we have
∀ = ζ∀ ∀ − ζ∀ �∀ + ζ∀ ∀� − ζ∀ �∀� .
0 S S S S
The two vectors (ζ∀ S∀ − ζ∀ �S∀) ⇒ S∀ and (ζ∀ S∀� − ζ∀ �S∀� ) ⇒ S∀� are
orthogonal and hence linearly independent, i.e., their sum
cannot be zero. It follows that the two vectors must be zero
separately, i.e., we must have
ζ∀ S∀ = ζ∀ �S∀ and ζ∀ S∀� = ζ∀ �S∀� .
This proves the uniqueness of the decomposition.
(2) We need to show the existence of such a decomposition. Since
S∀ is a subspace there exists an orthonormal basis { ε∀ j , j =
1, 2, · · · , M ≤ N } for S∀. Let
M
ζ∀ S∀ = ε j | ζ∀ ∪ ε∀ j
�∀ and ζ∀ S∀� = ζ∀ − ζ∀ S∀.
j =1
Clearly ζ∀ S∀ lies in the subspace S∀. Also ζ∀ S∀� lies in S∀� since
M
ε j | ζ∀ S∀� ∪ = �∀ǫ j | ζ∀ ∪ − �∀
�∀ εj | ε j � | ζ∀ ∪ ε∀ j � ∪
�∀
j � =1
M
ε j | ζ∀ ∪ −
= �∀ ε j � | ζ∀ ∪ �∀
�∀ ε j | ε∀ j � ∪
j � =1
M
ε j | ζ∀ ∪ −
= �∀ ε j � | ζ∀ ∪ δ j j � = 0.
�∀
j � =1
SQ13(5)1
N
P S∀2 P S∀1 ζ∀ = P S∀2 P S∀1 ζ∀ = P S∀1 ζ∀ ∞ζ∀ ⇒ VV∀ .
This means P S∀2 PS∀1 = P S∀1 . Secondly, if S∀1 ≥ S∀2 , projecting onto
S∀2 and then projecting onto S∀1 is the same as projecting onto
S∀1 alone. We can prove this by taking the adjoint of the equality
PS∀2 PS∀1 = P S∀1 , i.e., we have
† † † †
PS∀1 = P S∀ = P S∀2 PS∀1 = PS∀ P S∀ = P S∀1 P S∀2 .
1 1 2
(b) If ||η∀ || = ||P S∀1 η∀ ||, then ||P S∀� η∀ || = 0 which implies P S∀� η∀ =
1 1
∀
0.
1 Prugovečki p. 202. Roman Vol. 2 p. 538, p. 569.
2 See P.13.1(1) on the definition of norm of an operator on VV∀ N .
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
This in turn implies η∀ ⇒ S∀2 , i.e., any η∀ ⇒ S∀1 must also be in S∀2 . In
other words we have S∀1 ≥ S∀2 .
We have therefore showed that
Eq. (13.20) � Eq. (13.25) � Eq. (13.26) � Eq. (13.20).
Secondly we have
�η∀1 | ζ∀2 ∪ ζ∀1 η∀2 ξ∀ = �η∀1 | ζ∀2 ∪ �η∀2 | ξ∀ ∪ ζ∀1 .
(1) If P S∀1 and P S∀2 are orthogonal then ζ∀2 has no projection onto ζ∀1
since �ζ∀1 | ζ∀2 ∪ = 0.
∀
P S∀1 P S∀2 ζ∀ = P S∀1 P S∀2 ζ∀ = P S∀1 ζ∀2 = 0.
We also have
N
N
N
�η∀ | P ε∀ℓ η∀ ∪ = �η∀ | �∀
ε ℓ | η∀ ∪ ε∀ℓ ∪ = ε ℓ | η∀ ∪|2 .
| �∀
ℓ=1 ℓ=1 ℓ=1
It follows that
N N
�η∀ | P ε∀ℓ η∀ ∪ = �η∀ | P ε∀ℓ η∀ ∪ = �η∀ | η∀ ∪.
ℓ=1 ℓ=1
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Then
�A ζ∀ | η∀ ∪ = �η∀ | A ζ∀ ∪�
1
= �η∀ + ζ∀ | A(η∀ + ζ∀ ) ∪ + i �η∀ + i ζ∀ | A(η∀ + i ζ∀ ) ∪
2
−(i + 1) �η∀ | A η∀ ∪ + �ζ∀ | A ζ∀ ∪ .
Then
3
3
3
A� ζ∀ = aℓ P ε∀ℓ ζ∀ = aℓ �∀
ε ℓ | v∀ ∪∀
εℓ = aℓ ζℓ ε∀ℓ .
ℓ=1 ℓ=1 ℓ=1
We see that Aζ∀ = A� ζ∀ for every vector ζ∀. It follows that A� = A, i.e.,
the operator A has a spectral decomposition.
The spectral decomposition of A in Theorem 13.3.2(2) is just
a restatement of that of Theorem 13.3.2(1) by grouping all the
projectors of degenerate eigenvalues.
Q13(11) Show that the eigenprojectors of a selfadjoint operator
commutes with each other and that a selfadjoint operator commutes
with its eigenprojectors.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
For the inverse we have, again using Eqs. (13.48) and (13.49),
ei am P Â (am ) e−i am� P Â (am� )
m m�
= e i am −i am�
e P (am� )P Â (am )
Â
m, m�
= ei am e−i am� P Â (am� )δm� m
m, m�
= P Â (am ) = II
m
−1
� ei am P Â (am ) = e−i am� P Â (am� )
m m�
i  −1
† −1
� e = e−i  � ei  = ei  = e−i  .
and
[A, B
] = am bm� P Â (am )P B̂ (bm� ) − am bm� P B̂ (bm� )P Â (am )
m, m� m, m�
= Â
am bm� [P (am ), P (bm� )]. B̂
m, m�
It follows that
[P Â (am ), P B̂ (bm� )] = ] =
0 ∞ m, m� � [A, B 0.
(2) To prove the converse let us assume that [ A, B ]= 0. From the
 Â
result A P (am ) = am P (am ) in Eq. (13.118) proved in SQ13(12)
we deduce that
A B P Â (am ) ζ∀ = B A P Â (am ) ζ∀ = am B P Â (am ) ζ∀ .
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
� P Â (am ) ζ∀ = B
P Â (am )B P Â (am ) ζ∀ ∞ζ∀
� P Â (am )B
P Â (am ) = B
P Â (am ).
� P Â (am )B
P Â (am ) = P Â (am )B
.
P Â (am ) = P Â (am )B
We get B , i.e.,
, P Â (am )] =
[B 0 and similarly [A, P B̂ (bm )] =
0.
What we have shown is that if a selfadjoint operator A
then
commutes with another selfadjoint operator B
(1) A commutes with every eigenprojector P B̂ (bm ) of B
, and
(2) B commutes with every eigenprojector P (am ) of A.
Â
SQ13(16)
†
By definition the adjoint matrix M U� has elements
†
εℓ | U ε∀n ∪� = �U ε∀n | ε∀ℓ ∪.
MU� nℓ = MU�� ℓn = �∀
†
The product matrix of M U� and M U� has elements
†
MU�mn MU� = ε m | U ε∀n ∪�U ε∀n | ε∀ℓ ∪
�∀
nℓ
n n
= εn | U † ε∀ℓ ∪
�U † ε∀m | ε∀n ∪�∀
n
SQ13(18) For the operator in Eq. (13.70) we can write down its
adjoint, using Eq. (13.22), i.e.,
N
�
U † = ε∀ℓ ε∀ ℓ .
ℓ=1
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Then
N
N
�
U † U = ε∀ℓ ε∀ �ℓ ε∀ k ε∀k
ℓ=1 k=1
N
N
= ε∀ℓ ε∀ �ℓ ε∀ �k ε∀k = ε �ℓ | ε∀ �k ∪ ε∀ℓ ε∀k
�∀
ℓ, k=1 ℓ, k=1
N
= ε∀k ε∀k since �∀
ε �ℓ | ε∀ �k ∪ = δkℓ
k=1
= II .
Hence U is invertible (see SQ13(17)) with its adjoint equal to its
inverse, i.e., U † = U −1 . Hence U is unitary. We have used Eq. (13.23).
Q13(19) Prove Eq. (13.74) on the preservation of commutation
relations.
SQ13(19) To prove the preservation of commutation relations in
Eq. (13.75) we let A� = U AU † , B � = U B
U † and C� = U CU † . Then
� ] = A� B
[ A� , B � − B � A� = U AU † U B U † − U B U † U AU †
= U A BU † − U B AU † = U A B−B A U † = U CU †
= C� .
Q13(20) Prove Eq. (13.77) on the preservation of the quadratic
form.
SQ13(20) To prove the preservation of quadratic form in Eq.
(13.75) we let A� = U AU † and ζ∀� = U ζ∀. Then
�ζ∀� | A� ζ∀� ∪ = �U ζ∀ | U AU † U ζ∀ ∪ = �U ζ∀ | U Aζ∀ ∪ = �ζ∀ | Aζ∀ ∪.
Next we have
− i at
dξ∀ (t) d U (t)ξ∀ (0) d e ¯ ξ∀ (0)
i� = i� = i�
dt dt dt
− i at − i at
=a e ¯ ∀
ξ (0) = e ¯ ∀
A ξ (0)
− i at − i at
= e ¯ A ξ∀ (0) = A e ¯ ξ∀ (0)
= A ξ∀ (t).
This shows that ξ∀ (t) is a solution of Eq. (13.100).
The question assumes that A has a nondegenerate eigenvalue a.
A similar proof can be carried out for an eigenvector corresponding
to a degenerate eigenvalue. We will then use the expression of U in
Eq. (13.49).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 14
Then we have
1 �
αz | Sy α∀ z ∪ = �∀
MŜ y 11 = �∀ α y + β∀y | α∀ y − β∀y ∪ = 0.
2 2
i � �
αz | Sy β∀z ∪ = − �∀
MŜ y 12 = �∀ α y + β∀y | α∀ y + β∀y ∪ = −i .
2 2 2
i � �
MŜ y 21 = �β∀z | Sy α∀ z ∪ = �∀ α y − β∀y | α∀ y − β∀y ∪ = i .
2 2 2
1 �
MŜ y 22 = �β∀z | Sy β∀z ∪ = − �∀ α y − β∀y | α∀ y + β∀y ∪ = 0.
2 2
1 1
M P̂ β∀ 11
αz | P β∀y α∀ z ∪ = � �∀
= �∀ αz | β∀ y ∪ = .
y
2 2
i i
M P̂ β∀ 12
αz | P β∀y β∀z ∪ = � �∀
= �∀ αz | β∀y ∪ = .
y
2 2
1 i
M P̂ β∀ 21
= �β∀z | P β∀y α∀ z ∪ = � �β∀z | β∀y ∪ = − .
y
2 2
i 1
M P̂ β∀ 22
= �β∀z | P β∀y β∀z ∪ = � �β∀z | β∀y ∪ = .
y
2 2
This is the projection matrix in Eq. (7.181).
Q14(3) Find the probability distribution function for the x
component spin values in the state given by the vector α∀ z . What is
the corresponding expectation value?
SQ14(3) Equations (14.56) and (14.58) apply here. The state
vector α∀ z is related to the eigenvectors of Sx by Eq. (14.25), i.e., we
have
1 1
α∀ x = � (α∀ z + β∀z ), β∀x = � (α∀ z − β∀z ).
2 2
1 1 1 c+ c+ + c− 1
M P̂ α∀ x C η∀ = =
2 1 1 c− 2 1
1
= � (c+ + c− )C α∀ x .
2
The result shows that the matrix M P̂ α∀ x projects the column vector C η∀
onto the unit column vector C α∀ x .
Next we need to evaluate M P̂ β∀ C η∀, M P̂ α∀ yC η∀ and M P̂ β∀ C η∀.
x y
To evaluate M P̂ β∀ C η∀ we have:
x
To evaluate M P̂ α∀ yC η∀ we have:
(1) Using Eq. (7.180) for the expression of M P̂ α∀ y we get
† †
M P̂ α∀ y C η∀ = C α∀ y C α∀ y C η∀ = C α∀ y C α∀ y C η∀
= �C α∀ y | C η∀∪C α∀ y ,
where3
1 1 1
C α∀ y = � and �C α∀ y | C η∀∪ = � (c+ − i c− ).
2 i 2
2 See Eq. (7.116) for the expression of C
β∀x .
3 See Eq. (7.117) for the expression of C
α∀ y .
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
† †
M P̂ β∀ C η∀ = C β∀y C β∀ C η∀ = C β∀y C β∀ C η∀ = �C β∀y | C η∀∪C β∀y .
y y y
where
1 1 1
C β∀y = � and �C β∀y | C η∀∪ = � (c+ + i c− ).
2 −i 2
dF Â (η∀, τ )
ω Â (η∀, τ ) = = ℘ Â (η∀, a1 ) δ(τ − a1 )
dτ
+ ℘ Â (η∀, a2 ) δ(τ − a2 ) + · · · + ℘ Â (η∀, aN ) δ(τ − aN ).
4 See Eq. (7.117) for the expression of C which agrees with Eq. (14.40).
β∀ y
5 See Eq. (4.74).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 15
= F (τ2 ) − F (τ1 ).
Q15(2) Using Eqs. (15.15), (15.19) and (15.20) prove Eqs. (15.21),
(15.22) and (15.23).
SQ15(2)
(1) To prove Eq. (15.21) we have
M [τ1 , τ2 ] = M {τ1 } ∅ (τ1 , τ2 ] = M ({τ1 }) + M (τ1 , τ2 ]
= F (τ1 ) − F (τ1 − 0) + F (τ2 ) − F (τ1 )
= F (τ2 ) − F (τ1 − 0).
(2) To prove Eq. (15.22) we have
[τ1 , τ2 ] = M
M [τ1 , τ2 ) ∅ {τ2 }
=M [τ1 , τ2 ) + M ({τ2 }).
[τ1 , τ2 ) = M
M [τ1 , τ2 ] − M ({τ2 }) = F (τ2 ) − F (τ1 − 0)
− F (τ2 ) − F (τ2 − 0)
= F (τ2 − 0) − F (τ1 − 0).
(3) To prove Eq. (15.23) we have
[τ1 , τ2 ) = M
M {τ1 } ∅ (τ1 , τ2 )
=M ({τ1 }) + M (τ1 , τ2 ) .
M (τ1 , τ2 ) = M [τ1 , τ2 ) − M ({τ1 })
= F (τ2 − 0) − F (τ1 − 0) − F (τ1 ) − F (τ1 − 0)
= F (τ2 − 0) − F (τ1 ).
Chapter 16
SQ16(6)1 Since
�
|φ(x ) + ψ(x )|2 = φ(x ) + ψ(x ) φ(x ) + ψ(x )
= |φ(x )|2 + |ψ(x )|2 + φ � (x )ψ(x ) + φ(x )ψ(x )� ,
�
|φ(x ) − ψ(x )|2 = φ(x ) − ψ(x ) φ(x ) − ψ(x )
= |φ(x )|2 + |ψ(x )|2 − φ � (x )ψ(x ) − φ(x )ψ(x )� ,
we get, by adding the above two equations,
|φ(x ) + ψ(x )|2 = 2 |φ(x )|2 + |ψ(x )|2 − |φ(x ) − ψ(x )|2
≤ 2 |φ(x )|2 + |ψ(x )|2 .
Since |φ(x )|2 and |ψ(x )|2 are separately integrable over IR their sum
is integrable over IR . It follows that |φ(x ) + ψ(x )|2 is integrable over
IR and that φ(x ) + ψ(x ) is a member of L2 (IR ).
Similarly |φ(x ) ± i ψ(x )| are square-integrable and hence are
members of L2 (IR ).
1 Fano pp. 240–241.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Since all the terms on the right of the equality sign are integrable
over IR we conclude that φ(x )� ψ(x ) is integrable over IR .
Q16(7) Show that the sequence of vectors f∀n defined by functions
fn (x ) in Eq. (16.60) is a Cauchy sequence in the space C∀().3
SQ16(7)4 The function fn (x ) has the value 0 for x ≤ 1 − 1/n so
that in the limit as n � → the function has the value zero for all x <
1. It follows that in the limit as n, m � → the function fn (x ) − fm (x )
also has the value zero for all x < 1. Given that fn (x ) and fm (x ) have
the value 1 for all x ⊕ 1 and hence their difference
fn (x ) − fm (x ) has
the value 0 for all x ⊕ 1 we can deduce that fn (x ) − fm (x ) � 0 for
all x as n, m � → and that
2
| fn (x ) − fm (x )|2 dx � 0 as n, m � →.
0
2 Fano p. 241.
3 Fano p. 251.
4 Fano pp. 250–251.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
two quantities: (1) �ϕ∀ | called a bra and (2) |φ∀ ∪ called a ket. Their
product forms a bracket �ϕ∀ | φ∀ ∪ which is the scalar product. Explain
how we can interpret bras and kets in terms of vectors and linear
functionals.5
SQ16(10) A ket is just a vector by definition, i.e., |φ∪ := φ∀. Since
a bra �ϕ∀ | operates on a ket φ∀ produces the scalar product �ϕ∀ |
φ∀ ∪ which is equal to the linear functional F ϕ∀ generated by ϕ∀, i.e.,
F ϕ∀ (φ∀ ) = �ϕ∀ | φ∀ ∪ we can see that the bra �ϕ∀ | is identifiable with
the linear functional F ϕ∀ .
Q16(11) Explain the concept of separability in a Hilbert space.
SQ16(11) An infinite-dimensional vector space does not nec
essarily possess a countable orthonormal basis. The concept of
separability of a space is defined by the existence of a countable
orthonormal basis, i.e., a Hilbert space is separable if there is a
countable orthonormal basis for the space.
5 Jauch p. 32.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 17
∀
Operators in a Hilbert Space H
N
N
N
||Au∀ || = || cℓ Aǫ∀ℓ || ≤ |cℓ | ||Aε∀ℓ || ≤ |cℓ | M
ℓ=1 ℓ=1 ℓ=1
N
=M |cℓ | ≤ MN,
ℓ=1
1 Halmos p. 177.
∀
100 Operators in a Hilbert Space H
� ||φ∀S∀ || ≤ ||φ∀ || = 1.
Since P S∀ φ∀ = φ∀S∀ we deduce that
||P S∀ φ∀ || ≤ ||φ∀ || = 1 ∞ φ∀.
The upper bounded of ||PS∀ φ∀ || is 1 for any φ∀ ⇒ S∀. It follows that the
norm of PS∀ is 1.
Q17(3) Are projectors invertible? Are they reducible?
SQ17(3) Projectors are not invertible since they violate Theorem
17.5(1). The projector P S∀ onto the subspace S∀ would have zero
projection of any vector in the orthogonal complement S∀� of S∀, i.e.,
φ∀ ⇒ S∀� � PS∀ φ∀ = 0. ∀
∀
The subspace S is clearly invariant under PS∀ , i.e., φ∀ ⇒ S∀ �
P S∀ φ∀ ⇒ S∀. Since PS∀ is selfadjoint and bounded Theorem 17.9(1)
∀
applies, i.e., P ∀ is reducible by S.
S
∀
Operators in a Hilbert Space H 101
(3) For Y1, 0 in Eq. (16.65) we have −i � ∂Y1, 0 /∂ϕ = 0, since Y1, 0 is
independent of ϕ.
(4) For Y1, 1 in Eq. (16.66) we have −i � ∂Y1, 1 /∂ϕ = �Y1, 1 . The value
� comes from differentiating exp(i ϕ) in Y1, 1 .
We can conclude that the spherical harmonics in Eqs. (16.64) to
(16.67) are eigenfunctions of Lz corresponding to eigenvalues 0, −�,
0 and � respectively.2
Q17(5) Prove properties P17.10(1), P17.10(2) and P17.10(3) for
a pair of operators a and a† in Definitions 17.10(1) and 17.10(2).
SQ17(5) For property P17.10(1) we have
�ϕ∀1 | aϕ∀0 ∪ = 0 and �aϕ∀1 | ϕ∀0 ∪ = �ϕ∀0 | ϕ∀0 ∪ = 1.
In other words we have �ϕ∀1 | aϕ∀0 ∪ �= �aϕ∀1 | ϕ∀0 ∪. This implies
that a is not selfadjoint since Eq. (17.102) which is necessary for
selfadjointness is violated. Similarly we have
�ϕ∀1 | a† ϕ∀0 ∪ = �ϕ∀1 | ϕ∀1 ∪ = 1 and
�
�a† ϕ∀1 | ϕ∀0 ∪ = 2�ϕ∀2 | ϕ∀0 ∪ = 0.
Since �ϕ∀1 | a† ϕ∀0 ∪ �= �a† ϕ∀1 | ϕ∀0 ∪ the operator a† is not selfadjoint.
For property P17.10(2) we have
ϕ∀0 = a† a ϕ∀0 = 0 ϕ∀0 .
N
� �
ϕ∀n = a† a ϕ∀n = a† n ϕ∀n−1 = n (n − 1) + 1 ϕ∀(n−1)+1
N
= nϕ∀n ∞n = 1, 2, 3, · · · .
corresponding
These results show that ϕ∀n are the eigenvectors of N
to eigenvalues n = 0, 1, 2, · · · .
For property P17.10(3) we have, by the definition of the
operators, for n ⊕ 1,
�
aa† ϕ∀n = a n + 1 ϕn+1 = (n + 1)ϕ∀n ,
�
a† a ϕ∀n = a n ϕ∀n−1 = nϕ∀n
� aa† − a† a ϕ∀n = ϕ∀n
� [ a, a† ] = II
� [ a† , a] = −II .
2 See
SQ36(1) in Exercises and Problems for Chapter 36 for calculations in Cartesian
coordinates.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
102 Operators in a Hilbert Space H
∀
Operators in a Hilbert Space H 103
∀ z is
corresponding to the eigenvalue z. Show also that the vector
normalised.
Finally show that
∀z | N
� ∀ z ∪ = | z |2 ,
= a† a.
where N
SQ17(7) First we have
→
1 2 zn
∀ z = exp
a − |z| � aϕ∀n
2 n=0
n!
→
1 2 zn �
= exp − |z| � n ϕ∀n−1
2 n=1
n!
→
1 2 zn−1
= exp − |z| z� ϕ∀n−1
2 n=1
(n − 1)!
! →
"
1 zn−1
= z exp − | z |2 � ϕ∀n−1
2 n=1
(n − 1)!
! →
"
1 2 zn
= z exp − |z| � ϕ∀n = z ∀ z,
2 n=0
n!
∀ z is an eigenvector of the annihilation operator a
showing that
corresponding to the eigenvalue z.
For normalisation we have
→
(zn )� zm
� ∀ z ∪ = exp(−| z |2 )
∀z | � � �ϕ∀n | ϕ∀m ∪
n, m=0
n! m!
→
(zn )� zm
= exp(−| z |2 ) � � δnm
n, m=0
n! m!
→
(| z |2 )n
= exp(−| z |2 )
n=0
n!
= exp(−| z | ) exp(|z|2 ) = 1.
2
∀z = z
Using the result a ∀ z we get
∀z | N
� ∀ z ∪ = �
∀ z | a a
† ∀ z∪
= �a ∀ z ∪ = �z
∀ z | a ∀ z | z
∀ z ∪ = z� z �
∀z |
∀z∪
= | z |2 .
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
104 Operators in a Hilbert Space H
Then we have
�
�∀
ϕn |
a∀ z ∪ = � ∀ z∪ =
a† ϕ∀ n | n + 1 �∀ ∀ z∪
ϕn+1 |
�
� z �∀ ∀z ∪ =
ϕn | n + 1 �∀ ∀ z ∪,
ϕn+1 |
∀ z∪ = � z ∀ z ∪.
ϕn+1 |
� �∀ �∀
ϕn |
n+1
This is valid for any n. Explicitly we have:
∀
Operators in a Hilbert Space H 105
∀
106 Operators in a Hilbert Space H
Chapter 18
∀
Bounded Operators on H
∀
108 Bounded Operators on H
Q18(3) Show that plane waves f p (x) in Eq. (18.12) is not square
integrable over IR.
�
SQ18(3) The plane wave�f p (x) has an absolutely value of 1/ 2π �
for all x, i.e., | f p (x)| = 1/ 2π � and | f p (x)|2 = 1/2π �. Hence the
function is not integrable over IR since
→ →
2 1
| f p (x)| dx = dx = →.
−→ 2π � −→
SQ18(4)
(1) Let PS∀ be a projector onto the subspace S∀ of a Hilbert space H ∀.
∀
Let ϕ∀m be an orthonormal basis for S and let ψ∀ n be an orthonormal
basis for S∀� . Then the set of vectors {ϕ∀m , ψ∀ n } form an orthonormal
basis for H∀ . In accordance of Eq. (18.14) the trace of P ∀ is given by
S
tr(P S∀ ) = �ϕ∀m | P S∀ ϕ∀m ∪ + �ψ∀ n | PS∀ ψ∀ n ∪
m n
= �ϕ∀m | P S∀ ϕ∀m ∪,
m
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
Bounded Operators on H 109
∀
110 Bounded Operators on H
is a density operator.
Hence D
Q18(6) Show that the density operator D z in Eq. (18.29) can be
decomposed in terms of the eigenprojectors P α∀ y and Pβ∀y of Sy
shown in Eq. (14.38) as
z = 1 P α∀ y + 1 P ∀ .
D
2 2 βy
Find
Q18(7) the matrix representations (density matrices) in
basis α∀ z , β∀z of the density operators D
z and D
x in Eqs. (18.29)
and (18.30).
SQ18(7) Using the formula in Eq. (13.107) the matrix elements of
the density matrix D z for D z are easily worked out to be
1 1
D z 11 = �∀
αz | D z α∀ z ∪ = �∀
αz | P α∀ z + P β∀z α∀ z ∪ = ,
2 2
1
D z 12 = �∀
αz | D z β∀z ∪ = �∀αz | P α∀ z + P β∀z β∀z ∪ = 0,
2
z α∀ z ∪ = �β∀z | 1 P α∀ z + P ∀ α∀ z ∪ = 0,
D z 21 = �β∀z | D βz
2
1 1
D z 22 = �β∀z | Dz β∀z ∪ = �β∀z |
P α∀ z + P β∀z β∀z ∪ = .
2 2
The desired density matrix is
1/2 0 1 10
Dz = = .
0 1/2 2 01
The density matrix D x for D x can be similarly worked out. Since
x = D
D z by Eq. (18.32) the same density matrix is obtained, i.e., we
have
1 10
Dx = .
2 01
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
Bounded Operators on H 111
We also have
� � � �
tr BD = D
�ϕ∀ℓ | B ϕ∀ℓ ∪ = U † U D
�ϕ∀ℓ | U B U † ϕ∀ℓ ∪
ℓ ℓ
= † D
�U ϕ∀ℓ | B U ϕ∀ℓ ∪ =
† D
�ϕ∀ �ℓ | B ϕ∀ �ℓ ∪,
ℓ ℓ
where ϕ∀ �ℓ †
= U ϕ∀ℓ . These new vectors, being unitary transforms of
the basis vectors of an orthonormal basis, form a new orthonormal
basis. Since the trace is independent of the choice of basis, we get
� �
D
�ϕ∀ �ℓ | B ϕ∀ �ℓ ∪ = tr B
D � tr BD = tr B D .
ℓ
∀
112 Bounded Operators on H
SQ18(12)
d d d d
[ (IR ] = −i � x
x (IR), p − x = −i � x − 1+x = i �.
dx dx dx dx
d d d d
[
x⊗ (IR),
⊗
(IR)]
p ⊗
= i� p− p = i� 1+ p −p = i �.
⊗ dp dp dp dp
∀
Bounded Operators on H 113
Chapter 19
∀
116 Symmetric and Selfadjoint Operators in H
∀
Symmetric and Selfadjoint Operators in H 117
where E → →
D, ℓ () are the corresponding eigenvalues of H D () in
Eq. (19.44). Then assume that the expression in Eq. (�) can be
calculated by
→
�ϕ∀λ=0, n=0 () | cℓ E → →
∀ D,
D, ℓ ()ϕ ℓ ()∪
ℓ=1
→
= cℓ E → →
∀λ=0, n=0 () | ϕ∀ D,
D, ℓ ()�ϕ ℓ ()∪
ℓ=1
→
= | c ℓ |2 E →
D, ℓ (). (���)
ℓ=1
meaningful?
3 Equation
�(19.34) shows that the function which defines ϕ∀λ=0, n=0 () has a constant
value 1/ L for all x ⇒ , i.e., the function violates the Dirichlet boundary condition
in Eq. (17.31).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
118 Symmetric and Selfadjoint Operators in H
(1) The sum in Eq. (���) over odd integers values of ℓ, i.e., ℓ =
1, 3, 5 · · · , diverges, i.e.,
→ →
� 2
2 2 π 2 �2 ℓ2
2 →
| cℓ | E D, ℓ () =
ℓ=odd ℓ=odd
ℓπ 2mL2
→
4�2
= = →.
ℓ=odd
mL2
H D
∀
→ ()ϕ∀λ=0, n=0 () = 0(). (����)
∀
Symmetric and Selfadjoint Operators in H 119
not give a finite value for the norm of the formal vector expression,
i.e., carrying out formal calculation we get
→ →
� cℓ E →
D, ℓ ∀
()ϕ →
ℓ () | cm E → ∀m→ ()∪
D, m ()ϕ
ℓ=1 m=1
→
→
2
= cℓ� cm E → →
D, ℓ ()E m ()δℓm = |cℓ |2 E →
D, ℓ ()
ℓ, m=1 ℓ
→ →
� 2 2
2
2 2 π 2 �2 ℓ2
= |cℓ |2 E →
D, ℓ () = = →.
ℓ=odd ℓ=odd
ℓπ 2mL2
So the expression in Eq. (���) would not produce a meaningful result
due to the fact that Eq. (��) fails to define a vector. Any attempt to get
round this would fail to produce any meaning result.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 20
∀
122 Spectral Theory of Selfadjoint Operators in H
(2) The identity operator II has one eigenvalue, i.e., 1. This means
that F IÎ (τ ) =
0 for τ < 1 and at τ = 1 the spectral function
would jump to a value F IÎ (1). The function remains constant for
all τ ⊕ 1 as there are no eigenvalues above 1. It follows that
F 0̂ (1) is equal to its value at → which is II . In other words we
have !
0 τ <1
IÎ
F (τ ) = .
II τ ⊕ 1
(3) The projector operator P has two eigenvalue, i.e., 0 and 1. This
means that F P̂ (τ ) = 0 for τ < 0 and at τ = 0 the spectral
function would jump to a value F P̂ (0). The function F P̂ (τ )
remains constant for all τ ⊕ 1 as there are no eigenvalues above
1, i.e., F P̂ (1) is equal to its value at → which is II . So we have
F P̂ (τ ) = 0 for τ < 0 and F P̂ (τ ) = II for τ ⊕ 1.
The function F P̂ (τ ) for τ ⇒ [0, 1) can be determined as follows:
(a) Being a selfadjoint operator the projector P has a spectral
decomposition in the form of Eq. (20.20). Now let the
eigenvalues of P be denoted by a1 and a2 , i.e., a1 = 0 and
a2 = 1, and let the eigenprojectors associated with these
two eigenvalues be denoted by P P̂ (a1 ) and P P̂ (a2 ). These
eigenprojectors are related to the spectral function F P̂ (τ ),
i.e., we have, by Eq. (15.10),
∀
Spectral Theory of Selfadjoint Operators in H 123
SQ20(2) For the singleton set {0} and {1} we get, by Eq. (15.20),4
P̂ ({0}) =
M F P̂ (0) − F P̂ (0 − 0) = II − P ,
P̂ ({1}) =
M F P̂ (1) − F P̂ (1 − 0) = P .
The same results can be obtained directly from Eq. (20.19).
Finally, since the spectral function is constant over any interval
not containing 0 or 1, we get M P̂ () =
0 for such intervals.
∀
124 Spectral Theory of Selfadjoint Operators in H
SQ20(3)
(1) Let ϕ∀ be an eigenvector associated with an eigenvalue a of a
selfadjoint operator A. We have
�ϕ∀ | Aϕ∀ ∪ = �Aϕ∀ | ϕ∀ ∪ � �ϕ∀ | aϕ∀ ∪ = �aϕ∀ | ϕ∀ ∪
� a�ϕ∀ | ϕ∀ ∪ = a� �ϕ∀ | ϕ∀ ∪
� a = a� � a ⇒ IR .
(2) Let ϕ∀1 and ϕ∀2 be two eigenvectors associated with eigenvalue a1
and a2 respectively. We have
�Aϕ∀1 | ϕ∀2 ∪ = �ϕ∀1 | Aϕ∀2 ∪
� �a1 ϕ∀1 | ϕ∀2 ∪ = �ϕ∀ | a2 ϕ∀2 ∪
� ( a1 − a2 )�ϕ∀1 | ϕ∀2 ∪ = 0
� �ϕ∀1 | ϕ∀2 ∪ = 0 if a1 =
� a2 .
∀
Spectral Theory of Selfadjoint Operators in H 125
Q20(4) In L∀2 (IR) show that for every φ(x) ⇒ C c→ (IR) we have
d f (x)
[ f ( (IR) ]φ∀ := i �
x ), p φ(x).
dx
SQ20(4)
d d
[ f ( (IR)]φ∀ := −i � f (x)
x ), p − f (x) φ(x)
dx dx
d d f (x) d
= −i � f (x) − + f (x) φ(x)
dx dx dx
d f (x)
= i� φ(x).
dx
Q20(5) Show that selfadjoint operators having a discrete spec
trum are reducible by their eigensubspaces.
SQ20(5) A selfadjoint operator with a discrete spectrum com
mutes with its eigenprojectors. Let P Â (am ) be the projector onto
the eigensubspace S∀ Â (am ) associated with the eigenvalue am of a
selfadjoint operator A. Then P Â (am ) commutes with A. Theorem
17.9(1) then tells us that A is reducible by S∀ Â (am ).
Q20(6) Show that in L∀2 (IR) the position operator x (IR) is
(IR) is also reducible.5
reducible and the momentum operator p
SQ20(6) Generally a selfadjoint operator commutes with its
spectral projectors. For the position operator x in L∀2 (IR) its spectral
projector associated with an interval is given by Eq. (20.28), i.e.,
χ . The subspace S∀x̂ () associated with the projector is defined
by all the functions in L2 (IR) vanishing outside the interval . Since
χ commute with x Theorem 17.9(1) tells us x is reducible by the
5 This
is in contrast to the fact discussed in E20.7(1) that together the position and
momentum operators form an irreducible set.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
∀
126 Spectral Theory of Selfadjoint Operators in H
subspace S∀x̂ (). The same analysis in the momentum space shows
that the momentum operator is similarly reducible.
x in L∀2 (IR ) constitutes a
Q20(7) Show that the position operator
complete set of selfadjoint operators.
SQ20(7)6 The proof presented below is based on an application of
Theorem 20.6(1), i.e., we want to show that any bounded operator
on L∀2 (IR ) commuting with x is a function of x.
Let φ(x ) be a function in L2 (IR ) and let ψ(x ) be a real positive
function in L2 (IR ), i.e., ψ(x ) > 0 for all x . Let φ∀ and ψ∀ be the vectors
in L∀2 (IR ) defined by φ(x ) and ψ(x ). We can introduce a function
G(x ) = φ(x )/ψ(x ) since ψ(x ) > 0 for all x . This function G(x )
defines a multiplication operator G in L∀2 (IR ) by
φ(x )
Gϕ∀ := ϕ(x ),
ψ(x )
φ(x ) (x )
φ∀ = B
B Gψ∀ = G B ∀ :=
ψ∀ = G (x ) = φ(x ).
ψ(x ) ψ(x )
is a multiplication operator corresponding to the
It follows that B
function f (x ) = (x )/ψ(x ). By Definition 20.6(3) the position
operator constitutes a complete set of selfadjoint operators in
L∀2 (IR ).
6 Jordan p. 58.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 21
¨
Q21(2) Equation (10.27), which is a time dependent Schrodinger
equation, can be written in the form of Eq. (21.18) which follows
from Stone’s theorem as a vector equation in the Hilbert space
L∀2 (IR), i.e.,
dφ∀(t) φ∀(t),
i� =H
dt
∀
128 Spectral Theory of Unitary Operators on H
− i tH �
where U (t) = e ¯ is a one-parameter family of unitary operators
in Eq. (21.13).1 Show that2
d A(t)
i� = [A(t), H
].
dt
SQ21(3)
d A(t) dU †(t) dU (t)
i� = i� A U (t) + U †(t) A
dt dt dt
† †
= −H U (t) AU (t) + U (t) A H U (t)
= −H A(t) + A(t)H = [A(t), H ].
Chapter 22
[ ] = i �.
x, p
[ ] ≥ i � II ,
x, p
or
[ ] φ∀ = i � φ∀
x, p (�)
for an appropriate set of vectors φ∀ in L∀2 (IR ). What are the conditions
φ∀ must satisfy in order for the equality to hold?
SQ22(1) What we want is the domain D ∀ [ ] of the com-
x, p
mutator [ ]. The domains D
x, p ∀ (
x ) and D ∀(p ) of the position and
momentum operators are given by Eqs. (17.13) and (17.49). The
commutator is − p
xp x . So we must specify the domains of
xp
and p x first in terms that of D ∀ (
x ) and D ∀(p ), and then specify the
domain of the commutator. We have, by Eq. (17.58),
D∀ (
xp ∀(p
) := φ∀ ⇒ L∀(IR) : φ∀ ⇒ D ), p ∀ (
φ∀ ⇒ D x) .
∀(p
D ∀ (
x ) := φ∀ ⇒ L∀(IR) : φ∀ ⇒ D x φ∀ ⇒ D
x ), ∀(p) .
Finally we have, by Eq. (17.56), the domain
D∀ [ ] = D
x, p ∀ (
xp) ⊥ D∀(p
x ).
∀ ∀
Vectors φ in Eq. (�) must belong to D [ ] .
x, p
Q22(2) Consider a particle confined in an infinite potential well
of width [0, L]. All its wave functions φ(x) must vanish outside the
well, i.e., φ(x) = 0 ∞x ⇒ / [0, L]. Hence the state space of the trapped
particle is taken to be L∀2 (), = [0, L] rather than L∀2 (IR).
This result violates the inequality in Eq. (22.28), i.e., the inequality
cannot be applied to eigenvectors of p λ=0 (). The reason for
this violation lies in the domain of the commutator discussed in
SQ22(1), i.e., the commutator [ λ=0 () ] is not defined on the
x ( ), p
eigenvector ϕ∀λ=0, n () in Eq. (19.34). To see this let the commutator
acts on a vector φ∀ ⇒ L∀2 (), i.e., we have
[ λ=0 () ]φ∀ =
x ( ), p λ=0 () φ∀ − p
x () p x () φ∀.
λ=0 ()
For above operation to be meaningful we require that:
(1) The vector φ∀ must in the domain of p λ=0 (), i.e., apart from
differentiability the corresponding function φ(x) ⇒ L2 () must
satisfy the periodic boundary condition φ(0) = φ(L).
(2) The function xφ(x) corresponding to the vector x ()φ∀ must
also satisfy the periodic boundary condition.
Clearly the function xϕλ=0, n ()(x) corresponding to the vector
x ( ) ϕ∀λ=0, n (), i.e.,
% &
x 2nπ
� exp i x ,
L L
does not satisfy the periodic boundary condition. This means
x ( ) ϕ∀λ=0, n () is not in the domain of p
that the vector λ=0 ().
Consequently pλ=0 () cannot act on x () ϕ∀λ=0, n (). It follows that
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
the commutator cannot act on ϕ∀λ=0, n (), which in turn implies that
the inequality in Eq. (22.28) cannot be applied to ϕ∀λ=0, n ().
A similar problem arises with the uncertainty relation in the
Hilbert space L∀2 (Ca ) for a particle in circular motion. This is
discussed in §28.3.2.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 23
x⊗ (IR)
⊗
x (IR ) U F−1 , ⊗p
= U F (IR)
⊗
(IR ) U F−1 ,
= U F p
where U F is the Fourier transform operator defined by Eq. (18.54).
Since U F is unitary all the measured values of physical quantities,
Chapter 24
Let η∀ (−) , η∀ (0) and η∀ (+) be unit vectors in S∀(−) , S∀(0) and S∀(+) ,
respectively.1 A vector in H ∀ � is of the form
136 Direct Sums and Tensor Products of Hilbert Spaces and Operators
L− η∀ (−)� = η∀ (0)� ,
L− η∀ (0)� = η∀ (−)� , ∀ �.
L− η∀ (+)� = 0
L+ η∀ (+)� = η∀ (0)� ,
L+ η∀ (0)� = η∀ (+)� , ∀�.
L+ η∀ (−)� = 0
L=
L− +
L+ .
∀ (−) � 0
η∀ (+)� = 0 ∀ (0) � η∀ (+) .
two-dimensional space.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Direct Sums and Tensor Products of Hilbert Spaces and Operators 137
�η∀ (−)� |
L− η∀ (−)� ∪ = 0, �
L− η∀ (−)� | η∀ (−)� ∪ = 0.
�η∀ (−)� |
L− η∀ (0)� ∪ = 1, �
L− η∀ (−)� | η∀ (0)� ∪ = 1.
�η∀ (−)� |
L− η∀ (+)� ∪ = 0, �
L− η∀ (−)� | η∀ (+)� ∪ = 0.
�η∀ (0)� |
L− η∀ (−)� ∪ = 1, �
L− η∀ (0)� | η∀ (−)� ∪ = 1.
�η∀ (0)� |
L− η∀ (0)� ∪ = 0, �
L− η∀ (0)� | η∀ (0)� ∪ = 0.
�η∀ (0)� |
L− η∀ (+)� ∪ = 0, �
L− η∀ (0)� | η∀ (+)� ∪ = 0.
�η∀ (+)� |
L− η∀ (−)� ∪ = 0, �
L− η∀ (+)� | η∀ (−)� ∪ = 0.
L− η∀ (0)� ∪ = 0, �
�η∀ (+)� | L− η∀ (+)� | η∀ (0)� ∪ = 0.
�η∀ (+)� |
L− η∀ (+)� ∪ = 0, �
L− η∀ (+)� | η∀ (+)� ∪ = 0.
Similarly we have
�η∀ (−)� |
L+ η∀ (−)� ∪ = 0, �
L+ η∀ (−)� | η∀ (−)� ∪ = 0.
�η∀ (−)� |
L+ η∀ (0)� ∪ = 0, �
L+ η∀ (−)� | η∀ (0)� ∪ = 0.
�η∀ (−)� |
L+ η∀ (+)� ∪ = 0, �
L+ η∀ (−)� | η∀ (+)� ∪ = 0.
�η∀ (0)� |
L+ η∀ (−)� ∪ = 0, �
L+ η∀ (0)� | η∀ (−)� ∪ = 0.
�η∀ (0)� |
L+ η∀ (0)� ∪ = 0, �
L+ η∀ (0)� | η∀ (0)� ∪ = 0.
�η∀ (0)� |
L+ η∀ (+)� ∪ = 1, �
L+ η∀ (0)� | η∀ (+)� ∪ = 1.
�η∀ (+)� |
L+ η∀ (−)� ∪ = 0, �
L+ η∀ (+)� | η∀ (−)� ∪ = 0.
�η∀ (+)� |
L+ η∀ (0)� ∪ = 1, �
L+ η∀ (+)� | η∀ (0)� ∪ = 1.
�η∀ (+)� |
L+ η∀ (+)� ∪ = 0, �
L+ η∀ (+)� | η∀ (+)� ∪ = 0.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
138 Direct Sums and Tensor Products of Hilbert Spaces and Operators
Direct Sums and Tensor Products of Hilbert Spaces and Operators 139
∀ ∗ can be written as
∀ ∗ in H
SQ24(3)(a) An arbitrary vector
∀∗ =
cmn ϕ∀m ∗ ϕ∀n .
m, n
Then
� cmn ϕ∀m ∗ ϕ∀n | cm� n� ϕ∀m� ∗ ϕ∀n� ∪
m, n m� , n�
= ∀ ∗ ||2 < →.
|cmn |2 = ||
m, n
Since
Up cmn ϕ∀m ∗ ϕ∀n = cmn ϕ∀n ∗ ϕ∀m ,
m, n m, n
we have
||Up cmn ϕ∀m ∗ ϕ∀n ||2
m, n
=� cmn ϕ∀n ∗ ϕ∀m | cm� n� ϕ∀n� ∗ ϕ∀m� ∪
m, n m� , n�
= |cmn | = || ||2 .
2 ∀∗
m, n
Hence the operator Up preserves the norm and it is defined on every
∀ ∗ . Hence it is bounded operator of the norm ||Up || = 1.
vector in H
SQ24(3)(b) Given an arbitrary vector ∀ � = m, n cmn ϕ∀m ∗ ϕ∀n ⇒
H∀ ∗ we can establish the following results:
140 Direct Sums and Tensor Products of Hilbert Spaces and Operators
(2) From SQ24(a) we know that Up preserves the norm of vectors
∀ ∗ . It is invertible by Theorem 17.5(1), since
in H
Up ∀∗
∀∗ = 0 � ∀ ∗ || = ||
||Up ∀ ∗ || = 0
∀∗.
∀∗ = 0
�
� ∀ ∗ ∪ = �
∀∗ | ∀ ∗ | Up Up
∀ ∗ ∪ = �Up†
∀ ∗ | Up
∀∗∪
∀∗ |
= �Up Up† ∀∗∪
by Eq. (18.2).
� Up U p† = II
∀∗ |
� ∀ ∗ | Up
∀ ∗ ∪ = �Up ∀ ∗ ∪ = �U p† Up
∀∗ |
∀∗∪
.
� U p† Up = II
(4) Since the operator has been shown to be bounded, preserve the
norm of all vectors, invertible with its inverse equal its adjoint it
is therefore unitary by Theorem 18.3(1).
∀ � ∗ = m� , n� cm
(5) For selfadjointness let �
∀m� ∗ ϕ∀n� . Then:
� n� ϕ
� ∀ �∗∪ = �
∀ ∗ | Up cmn ϕ∀m ∗ ϕ∀n | �
cm ∀n� ∗ ϕ∀m� ∪
� n� ϕ
m, n m� , n�
� �
= cmn cm � n� δmn� δnm� .
m, n, m� , n�
�Up ∀ �∗∪ = �
∀∗ | cmn ϕ∀n ∗ ϕ∀m | �
cm ∀m� ∗ ϕ∀n� ∪
� n� ϕ
m, n m� , n�
� �
= cmn cm � n� δnm� δmn� .
m, n, m� , n�
∀ � ∗ ∪ = �Up
∀ ∗ | Up
It follows that � ∀ � ∗ ∪ which implies the
∀∗ |
selfadjointness of Up .
From Up = U p−1 we get Up = Up† = U p−1 . All this confirms the
previous conclusion that Up is unitary by Theorem 18.3(1).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Direct Sums and Tensor Products of Hilbert Spaces and Operators 141
∀∗ =
SQ24(3)(c) Given an arbitrary vector m, n cmn ϕ∀m ∗ ϕ∀n in
H∀ we have
∗
∗
Up A ∗ B p
∀
U † = p A ∗ B
U c mn ∀
ϕ n ∗ ∀
ϕ m
m, n
= Up cmn Aϕ∀n ∗ B
ϕ∀m
m, n
= cmn B ϕ∀m ∗ Aϕ∀n
m, n
∗ A
= B cmn ϕ∀m ∗ ϕ∀n .
m, n
1 2
2
P (s) II + 2IIUp + U p2 = P (s) .
=
4
(a) 2 1 2
P = II − 2IIUp + U p2 = P (a) .
4
Hence P and P are projector. They are orthogonal since
(s) (a)
1 2
P(s) P (a) = II − U p + Up − Up2 = 0.
4
The followings are examples of vectors unchanged by these
projectors:
(1) Vectors of the form
∀ ∗s =
cmn ϕ∀m ∗ ϕ∀n + ϕ∀n ∗ ϕ∀m
m, n
142 Direct Sums and Tensor Products of Hilbert Spaces and Operators
These vectors are shown in Eq. (33.19) and they are said to be
symmetrical in Definition 33.3.1(1).
(2) Vectors of the form
∀ ∗a =
cmn ϕ∀m ∗ ϕ∀n − ϕ∀n ∗ ϕ∀m
m, n
Chapter 25
Pure States
Q25(2) Explain why only unit vectors are used for state descrip
tion in Postulate 25.1(PS).1
SQ25(2) Theorem 22.1(1) tells us that a unit vector in a Hilbert
space together with the spectral function of a selfadjoint operator in
the Hilbert space can generate a probability distribution function. It
follows that if we describe a state by a unit vector and an observable
by a selfadjoint operator we can generate a probability distribution
function which can be taken to describe the probability distribution
of the values of the observable. Such a description of states and
observables is demonstrated in the model theory for electron spin
in §14.1.1. A formal statement that observables correspond to the
selfadjoint operators is given by Postulate 26.1(OV).
A vector which is not normalised cannot be used in Theorem
22.1(1) to produce a probability distribution function. It needs to
be normalised first.
Chapter 26
(1) They are all compatible, i.e., they are all simultaneously
measurable.
(2) Classical observables are related to the state in that they are
described by numerical functions defined on the classical state
space, i.e., they are numerical functions of the state. It follows
that a given state would determine the values of all observables.
In other words a classical system in a given state would possess
a value of every observables. These values can be revealed by
measurement.
(3) As the the state evolves in time observables would evolve in
time accordingly. The values of observables at a later time are
determined by their initial values.
(1) They are not all compatible, i.e., they are not all simultaneously
measurable. Only a limited number of quantum observables are
compatible.
Quantum Mechanics: Problems and Solutions
K. Kong Wan
c 2021 Jenny Stanford Publishing Pte. Ltd.
Copyright
ISBN 978-981-4800-72-3 (Paperback), 978-0-429-29647-5 (eBook)
www.jennystanford.com
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
1 See Definition 29.1.2(1) and Q29(6) in Exercises and Problems for Chapter 29.
2 See §20.5 for discussion on functions of selfadjoint operators. Definition 13.3.3(1),
i.e., Eqs. (13.37) and (13.38), applies to discrete observables.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 27
Canonical Quantisation
3
∂ A ∂(BC ) ∂ A ∂(BC )
{A, BC } = −
j =1
∂ x j ∂ pcj ∂ pcj ∂ x j
3
∂A ∂B ∂A ∂B
= C− C
j =1
∂ x j ∂ pcj ∂ pcj ∂ x j
3
∂ A ∂C ∂ A ∂C
+ B −B
j =1
∂ x j ∂ pcj ∂ pcj ∂ x j
= {A, B}C + B{A, C }.
The equations for {A + B, C } and {A B, C } are similarly proved.
Q27(3) Verify the Poisson bracket relations in Eq. (27.61) be
tween the components of the canonical angular momentum L∀ c .
SQ27(3) Using properties in Eqs. (27.54) to (27.59) and the
Poisson brackets of the canonical variables in Eq. (27.60) we get
{Lcx , Lcy } = {ypcz − zpcy , zpcx − x pcz }
= {ypcz , zpcx − x pcz } − {zpcy , zpcx − x pcz }
= {ypcz , zpcx } + {zpcy , x pcz } = −ypcx + x pcy
= Lcz ,
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
SQ27(6) First let us have a quick look at how Eq. (27.70) can
i and
come about. Using the selfadjointness of Q P j and treating Eqs.
(17.100) and (17.101) as equalities we get
† †
i,
[Q P j ]† = Q Pj −
i i =
PjQ PjQi − Q i i,
P j = −[ Q P j ].
i,
This shows that [Q P j ] is not selfadjoint. Let us suppose
P j ] = �II .
j,
[Q
[Q P j ] = �II � [ Q
j, P j ]† = �II † = �II = [ Q
j, i,
Pj ]
� −[ Q j , P j , ] = [ Q j , P j ],
2 See §16.1.2.4 for the definition of Schwartz functions. See also E.17.3.2.5(2),
Eq. (17.62) and P27.2(4).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
We have
[ (n+1) ] = [
x, p n p
x, p ] = [ n ] p
x, p n [
+ p ]
x, p
(n−1)
n
= i� np + (p
p ) i�
n,
= i � (n + 1) p
x (n+1) ] = [ p
,
[p x n]
, x +xn[ p ,
x]
(n−1)
= −i � n x x +
x n (−i �)
x n.
= −i � (n + 1)
Step 3: We can now conclude that the desired equations are true
for all n.
Q27(9) Verify the commutation relations in Eqs. (27.106) and
(27.107).
SQ27(9) Consider an observable A which is a polynomial in
x and
, i.e.,
p
N
M
A = xnp
cmn m.
n=1 m=1
[ ] = i� ∂ H .
x, H ] = −i � ∂ H .
, H
and [ p
∂p ∂x
[ ho ] = i � ∂ H ho = i � p
x, H ,
∂p m
[p ho ] = −i � ∂ H ho = −i �mω2
, H x.
x
∂
[
Lx , z − z p
Ly ] = [ y p x −
y , z p xpz ]
= [ y p x −
z , z p xpz ] − [ z p x −
y , z p xpz ]
= [ y p x ] + [ z p
z , z p y ,
xpz ]
= −i �y p x + i � xp
y = i � Lz .
Lz ,
[ Lx ] = [
xpy − y p z − z p
x , y p y ]
= [
xpy , y p y ] − [ y p
z − z p z − z p
x , y p y ]
= [
xp z ] + [ y p
y , y p x , z p
y ]
x pz + i �z px = i � Ly .
= −i �
Ly ,
[ x −
Lz ] = [ z p xpz ,
xpy − y p
x ]
= [ z p
x ,
xp x ] − [
y − y p xpz ,
xpy − y p
x ]
= [ z p
x ,
xpy ] + [ xpz , y p
x ]
= −i � z p z = i �
y + i � y p Lx .
4 See SQ27(3).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
[
Lz ,
L2 ] = [
Lz ,
L2x ] + [
Lz ,
L2y ] + [
Lz ,
L2z ]
= [
Lz ,
L2x ] + [
Lz ,
L2y ]
=Lx [Lz ,
Lx ] + [Lz ,
Lx ]
Lx
Ly [ Lz , Ly ] + [ Lz , Ly ] Ly
= i� Lx Ly
Ly + Lx − i � Ly Lx
Lx + Ly
=0.
� → � 2π � π � �
∂ 1
∀ |�
� ∀ ∪ = −i �
pr (r, θ, ϕ)� + (r, θ, ϕ)r 2 sin θ drdθ dϕ.
0 0 0 ∂r r
6 We assume that φ(r) and ψ(r) are differentiable with respect to r, i.e., they are
absolutely continuous in r. See Wan pp. 174–175 for more details.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
The vector
θ (Ca )ϕ∀n := θ ϕn (θ) does not satisfy the periodic boundary
condition in Eq. (17.37).10 This means that θ (Ca )ϕ∀n (Ca ) is not in the
domain of L(Ca ). Hence the commutator is not defined on ϕ∀n (Ca ).
Chapter 28
αx | P α∀z α∀ x ∪ = 1/2.
℘ Ŝ z (αxs , �/2) = �∀
1 See §14.1.1 and §36.3 for the theory for electron spin.
Pˆ ()φ∀ ∪
M P̂ (φ∀, ) = �φ∀ | M
⎧
⎪ �φ∀ | II − P φ∀ ∪
⎨ if = {0},
= �φ∀ | P φ∀ ∪ if = {1},
⎪
⎩ �φ∀ | 0φ
∀∪ if does not contain 0 or 1.
⎧
⎨ 1 − �φ∀ | Pφ∀ ∪ if
⎪ = {0},
= �φ∀ | P φ∀ ∪ if = {1},
⎪
⎩0 if does not contain 0 or 1.
where
P Ĥ (Ca ) (E 0 ) = |ϕ∀0 Ca ∪�ϕ∀0 Ca |,
P Ĥ (Ca ) (E n ) = |ϕ∀n Ca ∪�ϕ∀n Ca | + |ϕ∀−n Ca ∪�ϕ∀−n Ca |, n ⊕ 1.
This is as expected. The kinetic energy does not take any negative
values. The probability of the kinetic energy having a value in the
range (−→, τ ] for τ > 0 is the same as the probability of the kinetic
energy having a value in the range [0, τ ]. This is the same as the
� �
momentum having a value in the range [− τ , τ ].
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 29
Time Evolution
φ∀(t) = U (H
, t)φ∀(0)
−iE t
= e ¯ ℓ Pϕ∀ℓ φ∀(0)
ℓ
−iE t
= e ¯ ℓ cℓ ϕ∀ℓ , cℓ = �ϕ∀ℓ | φ∀(0)∪.
ℓ
¨
Q29(2) Verify Eq. (29.12) in the Schrodinger picture and Eq.
(29.33) in the Heisenberg picture.
SQ29(2)
¨
In the Schrodinger picture the expectation value
E A, φ s (t) at time t is given by �φ∀(t) | Aφ∀(t) ∪, where A is time-
d A(t) 1
�φ∀ | φ∀ ∪ = �φ∀ | [ A(t), H
] φ∀ ∪.
dt i�
For brevity we have omitted the subscripts Sch for Schrodinger¨
picture quantities and Hei for Heisenberg quantities. We have also
assumed that the initial and the evolved vectors are in the domain of
relevant operators, e.g., φ∀(t) is in the domain of A.
Q29(3) The Hamiltonian of a system at time t = 0 is given in terms
of a pair of annihilation and creation operators a and a† by
= 1
H a† a + �ω.
2
(a) Using the method of induction and the commutation relation of
a and a† , prove
H − �ω)n ,
n a = a (H n = 0, 1, 2, 3, · · · . (�)
¨
(c) The annihilation operator in the Schrodinger picture, denoted
by a Sch (t), is related to aHei (t) by
� �
aSch t) = e H t/i � aHei (t) e− H t/i � .
By expanding the exponential in term of a series, i.e.,1
→ n
� 1 t n,
e H t/i � = H
n=0
n! i�
(t) = U (t)† H
(0)U (t) = �ω † 1
H aHei (t)aHei (t) + .
2
It follows that
% &
† 1
[ aHei (t), H (t)] = aHei (t), �ω aHei (t)aHei (t) +
2
†
= �ω [aHei (t), aHei (t)
aHei (t)]
= �ω aHei (t).
The Heisenberg equation for aHei (t) is
d (t)] = �ω aHei (t)
aHei (t) = [aHei (t), H
i�
dt
d
� aHei (t) = −i ω aHei (t).
dt
The solution is clearly
aHei (t) = aHei (0)e−i ωt ,
3 The Hamiltonian is time-independent, � (t), i.e.,we have
even though it is written as H
� (t) = H
H � (0) = H
� , as explicitly shown in Q29(3)(c).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
¨
SQ29(3)(c) The annihilation operator in the Schrodinger picture,
denoted by aSch (t), is related to aHei (t) by
� �
aSch (t) = e H t/i � aHei (t) e− H t/i � .
Our problem is to work out the right-hand side and show explictly
that aSch (t) is time-independent. To work out the right-hand side we
first consider4
�
e H t/i � aHei (t)
→
n
1 H t
= aHei (t)
n=0
n! i �
→
n
1 t
H
−i ωt
=e aHei (0)
n=0
n! i�
→ n
1 t n a since aHei (0) = a
= e−i ωt H
n=0
n! i�
→
1 t n − �ω)n
= e−i ωt a(H by Eq. (�) in Q29(3)(a)
n=0
n!
i�
→
n
1 (H − �ω) t
−i ωt
=e a
n=0
n! i�
� �
= e−i ωt a e(H −�ω)t/i � = a e H t/i � .
It follows that
� �
e H t/i � aHei (t) e− H t/i � = a = aSch (t),
which is time-independent as expected.
Q29(4) Let A be the selfadjoint operator representing an observ
¨
able A in the Schrodinger picture and let F Â (τ ) be its spectral
function. Let the corresponding operator at time t in the Heisenberg
picture be denoted by AHei (t) and let F Â Hei (t) (τ ) be its spectral
4 Making used of Eq. (29.19) which relate a
�Hei (t) to a�Hei (0).
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
function. How are these two spectral functions related and how are
the probability distribution functions generated by these spectral
functions in a given state related?
SQ29(4) At time t = 0 let the state vector be φ∀ and the operator
for the observable in question be A. At time t > 0 let the evolved
state vector be denoted by φ∀(t) in the Schrodinger
¨ picture. In the
Heisenberg picture the evolved vector φ∀Hei (t) at time t remains φ∀,
i.e., we have φ∀Hei (t) = φ∀. So, we can simply rewrite φ∀Hei (t) as φ∀
without the variable t. The two vectors φ∀(t) and φ∀Hei are related by
Eq. (29.36), i.e., we have by φ∀(t) = U (t)φ∀Hei . The evolved operator
remains A in the Schrodinger ¨ picture, while in the Heisenberg
picture the evolved operator becomes AHei (t) = U (t)† AU (t) as
shown in Eq. (29.39). For brevity of notation we have omitted the
subscripts Sch for quantities in the the Schrödinger picture.
In accordance of Theorem 15.3(2) the spectral functions F Â (τ )
and F Â Hei (t) (τ ) of A and AHei (t) are related by the unitary
transformation generated by the unitary operator U (t) in the same
way the operators of the two pictures are related by Eq. (29.43), i.e.,
¨
The probability distribution function in the Schrodinger picture in a
given state at time t is given by
H = 1 p 2 + V (
x ).
2m
x ) to be a polynomial function of
Assuming V ( x show that in the
Heisenberg picture, we have5
d
m �φ∀ | x φ∀ ∪ = �φ∀ | p
φ∀ ∪,
dt
d dV (
x)
�φ∀ | p
φ∀ ∪ = −�φ∀ | φ∀ ∪.
dt dx
Establish the following Ehrenfest’s theorem:
d2 ∀ dV (
x)
m x φ∀ ∪ = �φ∀ | F φ∀ ∪, where F = −
�φ | .
dt2 dx
Discuss the physical significance of this result,
SQ29(5) From Eq. (29.32) we get6
d (t)
∂H 1
x (t) =
= p (t),
dt (t)
∂p m
d ∂H (t) ∂ V (x (t))
(t) = −
p =− .
dt x (t)
∂ x
∂
It follows that
d2 1 d d2 ∂ V (
x (t))
2
x (t) =
p (t) � m 2
x (t) = −
.
dt m dt dt x
∂
Taking the expectation values on both sides we get
d2 ∀ d
m x (t)φ∀ ∪ = �φ∀ | F φ∀ ∪,
�φ |
dt2 dt
where
dV (
x (t))
F = − ,
dx (t)
which is the Ehrenfest’s theorem. This is the equivalence of Newton’s
second law in classical mechanics. The significance here is that it
5 The subscript Hei for Heisenberg picture quantities is omitted for brevity. We also
assume that φ∀ is in an appropriate domain of all the operators involved.
6 The subscript H ei for Heisenberg picture quantities is omitted for brevity.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 30
Q30(1) An electron spin is in state αxs .1 What is the state and the
state vector immediately after a measurement of the z-component
spin resulting in the value �/2?
SQ30(1) Immediately after a measurement of the z-component
spin resulting in the value �/2 state is αzs by Postulate 30.1.1(PPDO).
The corresponding state vector is α∀ z .
Q30(2) Find the subspace associated with the projector F x̂ (x2 ) −
F x̂ (x1 ) on L∀2 (IR ).
SQ30(2) The subspace consists of all the vectors ϕ∀ in L∀2 (IR )
satisfying the equation
F x̂ (x2 ) − F x̂ (x1 ) ϕ∀ = ϕ.
∀
Chapter 31
1 This means that the transition as a time evolution does not satisfy Postulate
29.1.2(TESP) which applies to quantum evolution in the Schrödinger picture.
2 See Q18(9).
(1) The transition shown in Eq. (31.1) means that an initial state
vector is transformed into many different final state vectors. It
follows that the transition cannot be a unitary transformation
from the initial state vector.
(2) A one-dimensional projector cannot evolve unitarily to a density
operator which is not a projector.
= �ϕ∀1 | B P ∀ φ∀ ∪ = �φ∀ | B
P ∀ ϕ∀1 ∪ = �φ∀ | B φ∀ ∪.
φ φ
Chapter 32
Superselection Rules
Being reducible by S∀(n) the operator B re would commute with P (n)
by Theorem 17.9(1) so that the terms in the second sum above can
re P (n) P (m) . Since the projectors are orthogonal these
be written as B
terms vanish. We get
re =
B P (n) B
re P(n) = (n) ,
B
n n
� η∀ (0)� = b0 η∀ (0)� .
H
It follows from Eq. (29.4) that in the Schrödinger picture the evolved
state vector for the initial state vector η∀0� is given by
− ib t
η∀ (0)� (t) = e ¯ 0 η∀ (0)� .
Here both the initial and the final states are pure. The state vector
η∀ (0)� (t) remains in the subspace S∀(0) for all times.
where H ∀ and H
∀ (1) is spanned by a unit vector η∀ (1) ⇒ H ∀ (2) is spanned
∀ which is orthogonal to η∀ (1) . Let
by a unit vector η∀ (2) ⇒ H L be an
∀
operator on H defined by
(c) Discuss how the evolution may cause a transition from a pure
state to a mixture.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
SQ32(3)(b) First the state vector η∀(t) satisfies the initial condi
tion, i.e., η∀(t = 0) = η(0).
∀ ¨
It also satisfies the Schrodinger equation,
i.e.,
dη∀(t)
i� = i λ cos(λt/�) η∀ (1) + λ sin(λt/�) η∀ (2) .
dt
H η∀(t) = λ sin(λt/�) η∀ (2) + i λ cos b(λt/�) η∀ (1) .
dη∀(t) η∀(t).
� i� =H
dt
Chapter 33
Many-Particle Systems
Chapter 34
Conceptual Issues
Q34(1) Verify that ∀ (q, m) (t) in Eq. (34.22) satisfies the initial
condition in Eq. (34.16) and the Schrödinger equation (34.21).
SQ34(1) By setting t = 0 in Eq. (34.22) we can clearly see that
∀ (q, m) (t) satisfies the initial condition in Eq. (34.16). To prove
∀ (q, m) (t) satisfies the Schrodinger
¨ equation we first have
d (q, m)
i� ∀ (t)
dt
= iρc− ϕ∀− ∗ − sin(ρt/�) η∀ (0)� − i cos(ρt/�) η∀ (−)�
+ iρc2 ϕ∀+ ∗ − sin(ρt/�) η (0)� − i cos(ρt/�) η∀ (+)�
= ρc− ϕ∀− ∗ − i sin(ρt/�) η∀ (0)� + cos(ρt/�) η∀ (−)�
+ ρc2 ϕ∀+ ∗ − i sin(ρt/�) η (0)� + cos(ρt/�) η∀ (+)� .
The results above show that ∀ (q, m) (t) satisfies the Schrodinger
¨
equation, i.e., it satisfies Eq. (34.21).
Quantum Mechanics: Problems and Solutions
K. Kong Wan
c 2021 Jenny Stanford Publishing Pte. Ltd.
Copyright
ISBN 978-981-4800-72-3 (Paperback), 978-0-429-29647-5 (eBook)
www.jennystanford.com
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Chapter 35
1
ϕ∀H 2 (x) = � a† ϕ∀H 1
2
�
λ � d
:= x− ϕH 1 (x)
2 mω dx
�
λ 1 d
= xϕH 1 (x) − ϕ (x)
2 λ dx H 1
�
λ� 1 d
= 2λ x 2 ϕH 0 (x) − ϕH 0 (x) + x ϕH 0 (x)
2 λ dx
1 2 1
= � λ x ϕH 0 (x) − ϕH 0 (x) + x 2 ϕH 0 (x)
2 λ
1
= � (2λ x 2 − 1) ϕH 0 (x) = ϕH 2 .
2
Q35(5) Verify Eq. (35.14) and (35.15).
SQ35(5) To verify Eq. (35.14) we have
λ i i
a† a = x−
p x+
p
2 mω mω
λ 1 i i
= x2 +
2
2 +
p
xp−
p x
2 (mω) mω mω
λ 1 i
= x2 +
2
2 +
p [ ]
x, p
2 (mω) mω
mω 1 �
= x2 +
2
2 −
p II .
2� (mω) mω
To verify Eq. (35.15) we have
1 2 1 1
�ωa† a = + mω2
p x 2 − �ω
2m 2 2
� H ho = �ω N + 1 .
2
Q35(6) Verify that φ(x, t) in Eq. (35.39) is normalised and that
¨
it also satisfies the Schrodinger equation for time evolution of the
harmonic oscillator.
SQ35(6) Using Eqs. (35.43) for |φ(x, t) |2 and following formula of
integration
→
−a(x−b)2 π
e dx =
−→ a
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
2
p 1 1 2
φ(x, t) = �ω + pc (t) + i ω pc (t) x − xc (t)
2m 2 2m
1 1
− mω2 x 2 + mω2 x xc (t) − mω2 xc2 (t) φ(x, t).
2 2
(3) Comparing the above results we get
∂φ(x, t) ho φ(x, t).
i� =H
∂t
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
¨
In other words the wave function φ(x, t) satisfies the Schrodinger
equation.
Q35(7) Verify Eqs. (35.44), (35.45).
SQ35(7) Using Eqs. (35.23) and (35.27) for φ(x, t) and |φ(x, t) |2
we get
→
�φ∀(t) |
x φ∀(t)∪ = x |φ(x, t) |2
−→
→
1 2
λ 2
= x e−λ x−xc (t)
dx.
−→ π
We can rewrite the integral as the sum of two integrals, i.e.,
−λ x−xc (t)2
1
→
λ 2
x − xc (t) e dx
−→ π
→
1 2
λ 2 1
+ xc (t) e− 2 λ x−xc (t)
dx.
−→ π
The first integral can be written as
→
1
λ 2
2
ye−λ y dy, y = x − xc (t).
−→ π
This integral vanishes on account of the integrand being an odd
function of y. Since the function φ(x, t) is normalised the second
integral is equal to xc (t), i.e., we get
�φ∀(t) |
x φ∀(t)∪ = xc (t).
To calculate �φ∀(t) | p
φ∀(t)∪ we first observe that
φ(x, t) = pc (t) + i mω x − xc (t) φ(x, t).
p
2
sin ωt sin ωt 2 2
x+
(cos ωt)
p = (cos ωt)2
x2 +
p
mω mω
(sin ωt)(cos ωt)
+ (
xp+ p
x ),
mω
(sin ωt)2 + (cos ωt)2 = 1.
where
�
γ =− (1 − i )g.
2mω
(b) In the Heisenberg picture the annihilation and creation opera
†
tors become aHei (t) and aHei (t) and the Hamiltonian becomes
H Hei (t) = �ω a† (t)aHei (t) + 1 + γ aHei (t) + γ � a† (t).
Hei
2 Hei
SQ35(12)(a) ¨
In the Schrodinger picture we have
λ i
aSch := x Sch +
Sch ,
p
2 mω
λ i
a†Sch := x Sch −
Sch .
p
2 mω
†
We can calculate the product aSch aSch in terms of Sch as in
x Sch and p
SQ35(5), i.e.,
mω i i
a†Sch aSch = x Sch −
Sch
p x Sch +
p
2� mω mω Sch
mω 2 1 2 i
=
x Sch + Sch
p + (x p −p Sch
x Sch )
2� (mω) 2 mω Sch Sch
mω 2 1 2 1
= x Sch +
Sch
p − .
2� 2m�ω 2
It follows that
1 2 1 1
�ω a†Sch aSch = + mω2
p 2
x Sch − �ω.
2m Sch 2 2
We can express Sch in terms of the annihilation and creation
x Sch and p
operators, i.e.,
�
x Sch =
aSch + a†Sch ,
2mω
�mω
Sch = −i
p aSch − a†Sch .
2
It follows that
1 � †
1 �mω
x Sch + =
p aSch + aSch + −i aSch − a†Sch
mω Sch 2mω mω 2
� �
= 1 − i aSch + 1 + i a†Sch
2mω 2mω
1
� −g x Sch +
p
mω Sch
� �
= −g 1 − i aSch − g 1 + i a†Sch
2mω 2mω
= γ aSch + γ � a†Sch
Sch = �ω a† aSch + 1 + γ aSch + γ � a† .
�H Sch
2 Sch
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
d
i� a (t) = �ωaHei (t) + γ � , or
dt Hei
d i
a (t) = −i ωaHei (t) − γ � .
dt Hei �
It follows that:
d d
aHei (t)ei ωt = a (t) ei ωt + i ωaHei (t)ei ωt
dt dt Hei
i �
= −i ωaHei (t) − γ ei ωt + i ωaHei (t)ei ωt
�
i
= − γ � ei ωt .
�
Integrating the above equation, i.e.,
d i
a (t)ei ωt = − γ � ei ωt ,
dt Hei �
from 0 to t we get
1 � i ωt 1 �
aHei (t)ei ωt − aHei (0) = − γ e + γ
�ω �ω
1 �
� aHei (t) = aHei (0)e−i ωt − γ 1 − e−i ωt .
�ω
Chapter 36
Angular Momenta
Q36(1) The spherical harmonics Y1, 1 (θ, ϕ), Y1, 0 (θ, ϕ), Y1, −1 (θ, ϕ)
are given by Eqs. (16.65) to (16.67). In Cartesian coordinates, these
functions are given by1
3 x − iy
Y1, −1 (x, y, z) = ,
8π r
3 z
Y1, 0 (x, y, z) = ,
4π r
3 x + iy
Y1, 1 (x, y, z) = − .
8π r
(a) Using the expression in Cartesian coordinates for the quan
tised orbital angular momentum operator L cz in Eq. (27.86),
verify by explicit calculations that Y1, 1 (x, y, z), Y1, 0 (x, y, z) and
Y1, −1 (x, y, z) are the eigenfunctions of L cz corresponding to
eigenvalues �, 0 and −�.
(b) Consider the following coordinate transformations:
x � z� , y � x �, z � y� .
1 Zettili
pp. 293–934. For convenience we have used the same symbols, e.g., Y1, 0 for
the functions in Cartesian coordinates.
Lz Y∀1, −1 := x p
y − y p
x Y1, −1
3 1
= −i � (−i xr 2 − x 2 y + i x y 2 )
8π r 3
−(yr 2 − yx 2 + i x y 2 )
3 1
= −i � 3
− i xr 2 − yr 2
8π r
3 1
= −� {x − i y} = −� Y1, −1 .
8π r
Next we have
3 ∂ z 3 z ∂ 3 zx
x Y∀1, 0 := −i �
p = −i � − 2 r = i� .
4π ∂ x r 4π r ∂x 4π r 3
3 ∂ z 3 z ∂
y Y∀1, 0 := −i �
p = −i � − r
4π ∂ y r 4π r2 ∂y
3 zy
= i� .
4π r 3
Lz Y∀1, 0
:= x py − y p
x Y1, 0
3 zy 3 zx
= x i� 3
− y i� = 0.
4π r 4π r 3
Finally we have
3 ∂ x + iy 3 1 x + iy ∂
x Y∀1, 1 := i �
p = −i � − r
8π ∂x r 8π r r2 ∂x
3 1 x + iy x
= i� −
8π r r2 r
3 1 2
= i� 3
r − x2 − i xy ,
8π r
3 ∂ x + iy 3 i x + iy ∂
y Y∀1, 1 := i �
p = i� − r
8π ∂y r 8π r r2 ∂y
3 i x + iy y
= i� −
8π r r2 r
3 1 2
= i� 3
ir − x y − i y 2 ,
8π r
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Lz Y∀1, 1 := x p
y − y p
x Y1, 1
3 1 2
= i� i xr − x 2 y − i x y 2
8π r 3
= − yr 2 − yx 2 − i x y 2
3 1 2
= i� 3
i xr − yr 2
8π r
3 1
= −� x + i y = � Y1, 1 .
8π r
SQ36(1)(b)
(1) Consider a transformation from the original coordinate system
(x, y, z) to a new coordinate system (x � , y � , z� ):
x � z� , y � x � , z � y � .
We can see that the new z� -axis coincides with the original x-axis,
and that
r = x 2 + y 2 + z2 = x �2 + y �2 + z�2 = r � .
Physically we expect the angular momentum about the x-axis to
be the same as the angular momentum about the z� -axis. We can
confirm this mathematically by showing that
∂ ∂
Lx = y p z − z py = −i � y − z
∂z ∂y
is equal to
∂ ∂
Lz� = x � p x � = −i � x � � − y � � .
y � − y � p
∂y ∂x
This is obvious since x � = y, y � = z. It follows that the eigenfunctions
and eigenvalues of Lx are the same in terms of the dashed
coordinates as that of Lz� .
(2) We know the eigenfunctions Yℓ�, mℓ (x � , y � , z� ) and eigenvalues
of
Lz� in the transformed coordinates (x � , y � , z� ) as they are of the
same form as those of the eigenfunctions and eigenvalues of Lz in
the original coordinates, i.e., we have
� � � � 3 x � − i y� � � � � 3 z�
Y1, −1 (x , y , z ) = �
, Y1, 0 (x , y , z ) = ,
8π r 4π r �
3 x � + i y�
Y1,� 1 (x � , y � , z� ) = − .
8π r�
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Since Lz� =
Lx these are also eigenfunctions of Lx . Written in terms
of the original coordinates these functions become
� 3 y − iz � 3 x
Y1, −1 = , Y1, 0 = ,
8π r 4π r
3 y + iz
Y1,� 1 = − .
8π r
We can conclude that the required eigenfunctions Lx , denoted more
conveniently by X 1, 1 , X 1, 0 , X 1, −1 are
3 y − iz
X 1, −1 = ,
8π r
3 x
X 1, 0 = ,
4π r
3 y + iz
X 1, 1 = − .
8π r
Q36(2) Suppose L 2c and L cz are measured giving the eigenvalues
2� and −�, respectively. A measurement of
2
L cx is then made.
What are the possible results of the measurement of L cx ? Find the
probability of each of these possible results.
SQ36(2) The first measurement serves as a state preparation
process, assuming the projection postulate. From the measured
values of L2 and
Lz we deduce that ℓ = 1, mℓ = −1. The projection
postulate tells us that the state after the first measurement is
described by the eigenfunction
3 x − iy
Y1, −1 = .
8π r
It also follows that the possible values of Lx must be −�, 0, �, since
the orbital angular momentum quantum number ℓ is 1. According
to Postulate 28.1(PDDO) the probability of obtaining any of these
results are:
℘−1 = �Y∀1, −1 | P X∀ 1, −1 Y∀1, −1 ∪
= �Y∀1, −1 | |X∀ 1, −1 ∪�X∀ 1, −1 | Y∀1, −1 ∪
℘0 = �Y∀1, −1 | P X∀ 1, 0 Y∀1, −1 ∪
= �Y∀1, −1 | |X∀ 1, 0 ∪�X∀ 1, 0 | Y∀1, −1 ∪
= �Y∀1, −1 | X∀ 1, 0 ∪ � X∀ 1, 0 | Y∀1, −1 ∪
2
= �X∀ 1, 0 | Y∀1, −1 ∪ .
℘1 = �Y∀1, −1 | P X∀ 1, 1 Y∀1, −1 ∪
= �Y∀1, −1 | |X∀ 1, 1 ∪�X∀ 1, 1 | Y∀1, −1 ∪
= �Y∀1, −1 | X∀ 1, 1 ∪ � X∀ 1, 1 | Y∀1, −1 ∪
2
= �X∀ 1, 1 | Y∀1, −1 ∪ .
and
1 1 1
℘−1 = |c−1 |2 =, ℘0 = |c0 |2 = , ℘1 = |c1 |2 = .
4 2 4
The total probability is 1 as it must be.
3 Zettili
pp. 296–297. Such a system is known as a rigid rotator which can be used to
model a diatomic molecule. For the state space of a quantum rigid rotator, see the
comment on a footnote in §27.8.
4 A general and explicit expression of Y
ℓ, mℓ (θ, ϕ) is given by Eq. (36.42).
5 See Eq. (36.8) and the discussion there.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
[A, B
+ C ] = [A, B
] + [A, C ],
[ A, B
C ] = [A, B
] C + B
[A, C ],
J x,
we can calculate the commutation relation [ J y ] as follows:
�2 † † † †
[
J x,
Jy ] = [ (a1 a2 + a2 a1 ), (
a1 a2 − a2 a1 ) ]
4i
�2 † † †
= a1 a2 − a2 a1 ) ]
[ (a1 a2 , (
4i
† † †
+ [ a2 a1 , (
a1 a2 − a2 a1 ) ]
�2 † † † †
= − [ a1 a2 , a2 a1 ] + [ a2 a1 , a1 a2 ]
4i
�2 † † † †
= − a1 [ a2 , a2 a1 ] − [a1 , a2 a1 ] a2
4i
† † † †
+ a2 [ a1 , a1 a2 ] + [ a2 , a1 a2 ] a1
�2 † †
† †
= − a1 a1 + a2 a2 + a2 a2 − a1 a1
4i
�2 † †
= a2 a2 − a1 a1 = i � J z.
2i
�2
† † † † † †
J x2 = (a1 a2 )2 + (a2 a1 )2 + a1 a2 a2 a1 + a2 a1 a1 a2
4
�2
† † 1 (N
2 + 1) + N
2 (N
1 + 1) ,
= (a1 a2 )2 + (a2 a1 )2 + N
4
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
�2 † 2 † 1 (N
2 + 1) − N 2 (N
1 + 1) ,
J y2 = − (a1 a2 ) + (a2 a1 )2 − N
4
2
� 2 + N 2 − N1 N2 − N 2 N
1 ,
J z2 = N 1 2
4
2
�
�
J2= 2 + N
N 1
2 + N
2
1 N2 + N 2 N
1 + 2N 1 + 2N
2
4
�2
2 2 + 2 N
1 + N2
= N1 + N
4
N
N
= + 1 �2 .
2 2
To verify Eq. (36.60) we observe that
� † † † †
J+ = Jx + iJy = (a1 a2 + a2 a1 ) + (a1 a2 − a2 a1 )
2
†
= � a1 a2 .
� † † † †
J− = Jx − iJy = (a1 a2 + a2 a1 ) − (a1 a2 − a2 a1 )
2
†
= � a2 a1 .
We get
†
J + | j, m∪ = � a1 a2 ϕ∀n1 , n2 = � (n1 + 1) n2 ϕ∀n1 +1, n2 −1 . (�)
Writing ϕ∀n1 −1, n2 +1 in the notation | j �� , m�� ∪ and using Eq. (36.62)
we get
1 �� 1
j �� =(n + n��2 ) = (n1 − 1) + (n2 + 1) = j,
2 1 2
1 1
m�� = (n��1 − n��2 ) = (n1 − 1) − (n2 + 1) = m − 1,
2 2
n1 (n2 + 1) = ( j + m)( j − m + 1),
⎛ ⎞2 ⎛ ⎞
1 0 0 1 0 0
MŜ2(1)z = �2 ⎝ 0 0 0 ⎠ = �2 ⎝ 0 0 0 ⎠
0 0 −1 0 0 1
⎛ ⎞
2 0 0
�2 ⎝
= 0 0 0 ⎠.
2
0 0 2
It follows that
⎛ ⎞
1 0 0
MŜ 2(1) = MŜ2(1)x + MŜ2(1)y + MŜ2(1)z = 2�2 ⎝ 0 1 0 ⎠ .
0 0 1
αx | β∀z ∪ = |�β∀z | α∀ x ∪ |2 .
= �β∀z | α∀ x ∪ �∀
1 1
β∀z = c+ � (α∀ z + β∀z ) + c− � (α∀ z − β∀z )
2 2
1 1
= � (c+ + c− )α∀ z + � (c+ − c− )β∀z
2 2
1
� (c+ + c− ) = 0, � (c+ − c− ) = 1
2
� �
� c+ = 1/ 2, c− = −1/ 2.
The required probabilities are |c+ |2 = 1/2 and |c− |2 = 1/2 which
agree with previous results.
1 0
C α∀ z = , C β∀z = ;
0 1
1 1 1 1
C α∀ x = � , C β∀x = � ;
2 1 2 −1
1 1 1 1
C α∀ y = � , C β∀y = � .
2 i 2 −i
Next wehave the matrix representations of Sz , Sx and Sy in the basis
α∀ z , β∀z given by Eqs. (36.143) to (36.145), i.e.,
1 1 0
MŜ z = � ,
2 0 −1
1 0 1
MŜ x = � ,
2 1 0
1 0 −i
MŜ y = � .
2 i 0
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
� 0 1 1 � −1
MŜ x C β∀x = � = �
2 2 1 0 −1 2 2 1
�
= − C β∀x ,
2
� 0 −i 1 � 1
MŜ y C α∀ y = � = �
2 2 i 0 i 2 2 i
�
= C α∀ y ,
2
� 0 −i 1 � −1
MŜ y C β∀y = � = �
2 2 i 0 −i 2 2 i
�
= − C β∀y .
2
Q36(13)7 A unit vector n∀ in the 3-dimensional IE∀ 3 aligned at an
angle θ to the z-axis on the x z plane is given by n∀ = sin θ i∀ + cos θ k∀.
The spin operator Sn∀ in the direction of the unit vector n∀ is given by
Sn∀ = n∀ · S, where
n∀ · S = nx Sx + n y Sy + nz Sz = sin θ Sx + cos θ Sz .
7 See Zettili pp. 298–316 for more examples. IE∀ 3 corresponds to the physical space we
live in.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
1 cos θ sin θ
MŜ n∀ = � ,
2 sin θ − cos θ
1 cos θ sin θ
= � .
2 sin θ − cos θ
8 Thenotation C η∀n∀ in Eq. (�) shows that the spin aligned in the direction n∀ should be
denoted by η∀n∀+ .
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
1 cos(θ/2) �
= � = C η∀ ,
2 sin(θ/2) 2 n∀+
and
1 cos θ sin θ − sin(θ/2)
MŜ n∀ C η∀n−
∀
= �
2 sin θ − cos θ cos(θ/2)
1 − cos θ sin(θ/2) + sin θ cos(θ/2)
= �
2 − sin θ sin(θ/2) − cos θ cos(θ/2)
1 sin(θ/2) �
= � = − C η∀n∀− .
2 − cos(θ/2) 2
These results shows that C η∀n∀+ and C η∀n∀− are eigenvectors of MŜ n∀
corresponding to eigenvalues �/2 and −�/2 respectively.
SQ36(13)(b) The required state vector must be the eigenvector
of the spin operator Sn∀ corresponding to the eigenvalue �/2. In
matrix representation this means that the matrix representation of
the state vector must be the eigenvector of MŜ n∀ corresponding to
the eigenvalue �/2. This eigenvector has been shown in SQ36(13)(a)
to be C η∀n∀+ in Eq. (�). This explains the notation, i.e., η∀n∀ + is the state
vector for the spin aligned along the n∀ direction and C η∀n∀+ is its matrix
representation in basis α∀ z , β∀z .
SQ36(13)(c) We can express the state vector η∀n∀ + in terms of a
linear combination of the spin-up state vector α∀ z and spin-down
state vector β∀z . In matrix terms this means that
cos(θ/2) 1 0
= cos(θ/2) + sin(θ/2) .
sin(θ/2) 0 1
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
The required probabilities for spin up and spin down are given
by the squares of the coefficients in the above expression, i.e., the
probabilities are given respectively by
2 2
cos(θ/2) and sin(θ/2) .
We can check that total probability is 1, i.e.,
2 2
cos(θ/2) + sin(θ/2) = 1.
(a) Show that the Pauli matrix σ y possesses the following proper
ties9 :
σ y0 = σ y2 = σ y4 = · · · = σ y2k = I 2×2 ,
σ y = σ y3 = σ y5 = · · · = σ y2k+1 ,
→
→
1 1
ecσ y = c 2k I 2×2 + c2k+1 σ y , (�)
k=0
(2k)! k=0
(2k + 1)!
→
� 1 � �n
ecσ y = cσ y .
n!
n=0
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
(c) Let
1
U (θ) = e− 2 i θσ y .
→
1 n
ecσ y = cσ y
n=0
n!
→
→
1 2k 1 (2k+1)
= cσ y + cσ y
k=0
(2k)! k=0
(2k + 1)!
→
→
1 1
= c 2k I 2×2 + c(2k+1) σ y .
k=0
(2k)! k=0
(2k + 1)!
cos(θ/2) sin(θ/2)
.
− sin(θ/2) cos(θ/2)
Chapter 37
∀ = − 1 y B i∀ + 1 x B ∀j .
A
2 2
Using the expression for the curl in Eq. (27.33) we get, due to A z = 0,
∀ = ∂ Ay ∂ Ax ∂ Ay ∂ Ax
�×A − i∀ + ∀j + − k∀
∂z ∂z ∂x ∂y
= B k∀.
Q37(2) Using Eqs. (37.8) to (37.10), show that the magnetic field
in Eq. (37.7) is derivable from the vector potential in Eq. (37.14) and
that the magnetic field in Eq. (37.16) is derivable from the vector
potential in Eq. (37.17).
SQ37(2) In cylindrical coordinates the vector potential A ∀ in Eq.
(37.14) has a non-zero component only in the θ direction, i.e.,
1
Ar = 0, A z = 0 and A θ (r) = Br.
2
Then Eqs. (37.8), (37.9) and (37.10) tell us that:
(1) Since r and A θ are independent of z and A z = 0 we get Br = 0.
1 − 1 i ωt 1
η∀(t) = � e 2 α∀ z + e 2 i ωt β∀z ,
2
where ω = eB/m. What are the spin orientations initially at t = 0
and later at t = π/2ω?
SQ37(3) ¨
The Pauli-Schrodinger equation is
d (s) η∀(t) = eB Sz η∀(t).
i� η∀(t) = H
dt m
The given vector η∀(t) satisfies the initial condition, i.e., we have
1
∀ = 0) = � (α∀ z + β∀z ).
η∀(0) = η(t
2
The given vector η(t)∀ also satisfies the Pauli-Schrodinger ¨ equation,
i.e., we have
dη∀(t) 1 d − 1 i ωt 1
i� = i�� e 2 α∀ z + e 2 i ωt β∀z
dt 2 dt
$
1 1 − 12 i ωt 1 1
i ωt ∀
= i�� − iω e α∀ z + i ω e 2 βz
2 2 2
1 1 1 1
= �ω � e− 2 i ωt α∀ z − e 2 i ωt β∀z .
2 2
η∀(t) = eB �1 e− 21 i ωt Sz α∀ z + e 21 i ωt Sz β∀z
H
m 2
$
eB 1 1 � 1 −�
= � e− 2 i ωt α∀ z + e 2 i ωt β∀z
m 2 2 2
� eB 1 1 1
= � e− 2 i ωt α∀ z − e 2 i ωt β∀z
2 m 2
dη∀(t) eB
= i� since = ω.
dt m
t = π/2ω we have
1 − 1 iω π 1 π
η∀(π/2ω) = � e 2 2ω α∀ z + e 2 i ω 2ω β∀z
2
1 −i π π
= � e 4 α∀ z + ei 4 β∀z
2
1 −i π π π
= � e 4 α∀ z + ei 4 ei 4 β∀z
2
1 −i π π
= � e 4 α∀ z + ei 2 β∀z
2
1 −i π
= � e 4 α∀ z + i β∀z .
2
According Eq. (36.138) this represents a spin orientation along
positive y-direction. The complex phase factor does not affect this
conclusion. The spin evolves from x-direction to y-direction.
June 23, 2020 10:50 JSP Book - 9in x 6in SolutionsManual(C2019)
Bibliography