PHAS0069

Advanced Quantum Theory

Problem Sheet 2 – Model Answers

1. Solve the following problems related to temporal evolution:

(a) Consider two state vectors |ψ (1) (t)i and |ψ (2) (t)i connected by a unitary,
time-dependent transformation T (t) so that |ψ (2) (t)i = T (t)|ψ (1) (t)i. The
temporal evolution of both states is governed by the time-dependent Schrödinger
equation, with Hamiltonians H (1) (t) and H (2) (t), respectively. Show that

dT (t) †
H (2) (t) = T (t)H (1) (t)T † (t) + i~ T (t).
dt
[3]
Model Answer: To prove that we need to consider that the time-dependent
Schrödinger equation remains invariant under the transformation T (t), i.e.,
that
d
i~ |ψ (1) i = H (1) |ψ (1) i
dt
and that
d
i~ |ψ (2) i = H (2) |ψ (2) i
dt
hold. Let us then consider
d (2) d 
T (t)|ψ (1) (t)i .

i~ |ψ i = i~
dt dt
This gives
d|ψ (1) (t)i
 
dT (t) (1)
i~ |ψ (t)i + T (t) .
dt dt
(1 mark)

Since d|ψ (1) (t)i/dt = H (1) |ψ (1) (t)i/(i~), we find

 
dT (t) (1)
i~ + T (t)H |ψ (1) (t)i = H (2) |ψ (2) (t)i = H (2) T (t)|ψ (1) (t)i.
dt

(1 mark)

This implies that

dT (t)
i~ + T (t)H (1) = H (2) T (t).
dt
 Acting from the right with T † (t) and using the fact that this transformation
is unitary gives
dT (t) †
H (2) = i~ T (t) + T (t)H (1) T † (t).
dt
(1 mark)
(b) Using |ψ (2) (t)i = T (t)|ψ (1) (t)i and the fact that this relation should be
valid for all times, show that the time evolution operators associated to
each state vector transform as U (2) (t, t0 ) = T (t)U (1) (t, t0 )T † (t0 ). [2]
Model answer: The transformation being valid for all times means that for
an initial time t0 such that |ψ (n) (t)i = U (n) (t, t0 )|ψ (n) (t0 )i, with n = 1, 2,
|ψ (2) (t0 )i = T (t0 )|ψ (1) (t0 )i must hold. This implies that

|ψ (2) (t)i = U (2) (t, t0 )T (t0 )|ψ (1) (t0 )i.

(1 mark)
Acting with T † (t) from the left gives

|ψ (1) (t)i = U (1) (t, t0 )|ψ (1) (t)i = T † (t)U (2) (t, t0 )T (t0 )|ψ (1) (t0 )i.

This means that

U (1) (t, t0 ) = T † (t)U (2) (t, t0 )T (t0 ),

(1 mark)
as requested.

2. Consider an arbitrary state vector |ψi and the basis {|ri} and {|pi} associ-
ated with the position and momentum representations, respectively. Show that
hp|R̂|ψi = i~∇p hp| ψi, where R̂ is the position operator and ∇p the gradient
in momentum space.
Hint: closure relations in {|ri} and properties of Fourier transforms are very
useful. [2]
Model answer: Here we need to start using a closure relation in {|ri} on the
left-hand side of the above expression. This gives

e−ip·r/~
Z Z
hp|R̂|ψi = d rhp| ri hr|R̂|ψi = d3 r
3
rhr| ψi
(2π)3/2
(1 mark).
Note, however, that
X ∂ Z −ip·r/~ −ip·r/~
 Z
3 e 3 e
i~∇p hp| ψi = i~ êi dr hr| ψi = d r rhr| ψi ,
∂pi (2π)3/2 (2π)3/2
which proves the statement (1 mark).
3. Two quantum systems a and b are prepared in a joint state |Ψi = |ψa i|ψb i. Let
the set {|φk i} be a basis of states for system a, {|νj i} a basis for system b and A
an operator on the Hilbert space of a with the following spectral decomposition:
N
X
A = λ1 |φ1 ihφ1 | + λ2 |φn ihφn |.
n=2

(a) Write down an expression for the projector acting on the total Hilbert space
and which is associated to the measurement outcome λ2 . Derive an expres-
sion for the probability of obtaining λ2 in a measurement of A on the state
|Ψi.

Model answer:Let’s denote Pλ2 the projector associated to λ2 , then acting

on the total Hilbert space this projector reads:
N
X X N X
X
Pλ2 ⊗ 1b = |φn ihφn | |νj ihνj | = |φn i|νj ihνj |hφn |
n=2 j n=2 j

Then the probability of obtaining λ2 = hΨ|Pλ2 ⊗ 1b |Ψi becomes:

PN P
hΨ|Pλ2 ⊗ 1b |Ψi = n=2 j hψa |hψb ||φn i|νj ihνj |hφn ||ψa i|ψb i
PN X
2
= n=2 |hψa |φn i| hψb |νj ihνj |ψb i
j
| {z }
=1
PN
= n=2 |hψa |φn i|2

(b) Derive an expression for the state |Ψ0 i of the global system after the mea-
surement. Has the state of the system b changed?

Model answer:
(Pλ ⊗ 1b )|Ψi
|Ψ0 i = p 2
hΨ|Pλ2 ⊗ 1b |Ψi
PN
hφn |ψa i
= qP n=2 |φn i|ψb i = |ψa0 i|ψb i.
N 2
n=2 |hψa |φn i|

The state of system B hasn’t changed.

 (c) Discuss whether the predictions in (a) and (b) would change depending of
the choice of basis for the system b. Will the results of measurements on
the systems a and b be correlated? Justify your answer.

Model answer: The predictons in (a) and (b) would not change depending
on the basis set for b. Also, if we find the probability of a particular eigen-
value of an operator acting on B the result will be analogus i.e. it will be
independent on the basis set of a. This therefore means that measurement
outcomes are uncorrelated.

[4]
Marks 2 marks for part (a). 1 mark for each part (b) and (c).

4. In this question, we shall consider experiments which generates pure states prob-
abilistically. Such a probability distribution over pure states can be represented
by a (single) mixed state in the density matrix formalism. Write down a matrix
representation for the density operator in each of the following three cases. You
may express them in bra and ket notation or use a matrix representation, but in
either case simplify your expressions as much as possible. Note that |φ1 i, |φ2 i
and |φ3 i are an orthonormal set of states. Write down a matrix representation for
the density operator considering that the following states are each prepared with
probability 1/3,
√
• (|φ1 i + |φ2 i)/ 2,
√
• (|φ1 i + exp(i2π/3)|φ2 i)/ 2,
√
• (|φ1 i + exp(−i2π/3)|φ2 i)/ 2.

[2]

Model Answer: The density matrix is:

11 11
ρ= (|φ1 i + |φ2 i)(hφ1 | + hφ2 |) + (|φ1 i + exp[i2π/3]|φ2 i)(hφ1 | + exp[−i2π/3]hφ2 |)
32 32
11
+ (|φ1 i + exp[−i2π/3]|φ2 i)(hφ1 | + exp[i2π/3]hφ2 |)
  3 2
1
= (|φ1 ihφ1 | + |φ2 ihφ2 |)
2
 
1 1 0
=
2 0 1
 (2 marks)

5. The combined state of a pair of two-level atoms, A and B, is given by the density
matrix:
1 1
ρ = |gA , gB ihgA , gB | + |gA , eB ihgA , eB |.
2 2
(a) Calculate the reduced density matrix operator for the each of the two-level
systems.

Model answer: Using 1B = |gB ihgB | + |eB iheB |

ρA = TrB (ρ1B ) = hgb |ρ|gb i + hgb |ρ|gb i
= 21 |gA ihgA | + 21 |gA ihgA | = |gA ihgA |.

Similarly, ρB = TrA (ρ1A ) = hgA |ρ|gA i+heA |ρ|eA i but in this case heA |ρ|eA i =
0.
Then
1 1
ρB = hgA |ρ|gA i = |gB ihgB | + |eB iheB |.
2 2
B is an statistical mixture of the ground and excited state.
(b) Calculate the purity of system A and indicate whether the state of the two
atoms is a product or an entangled state. Justify your answer.

Model answer: The purity for a general state ρ is given by Purity= Tr(ρ2 ).
Since ρA and ρB are diagonal in the basis chosen i.e.
   
1 0 1/2 0
ρA = , ρB = .
0 0 0 1/2
the Purity is very simple to compute: Tr(ρ2A ) = 1 and thererefore it cannot
be in an entangled state with B. You can also notice that the join state of A
and B, ρ, can be written as a mixed product state i.e. ρ = ρA ⊗ ρB :
1
ρ = |gA ihgA | ⊗ (|gB ihgB | + |eB iheB |).
2
If we compute the purity of B we have Tr(ρ2B ) = 1/2. Based alone of the
purity of B we cannot conclude anything about the entanglement between
A and B.

[3]
Marks: Total marks is 3. 2 for part (a) and 1 for part (b).
6. A physicist runs two experiments A and B to prepare quantum systems in a va-
riety of initial states. In experiment A she uses a probabilistic machine that can
prepare a single quantum system in one of n possible pure states {|Ψ1 i, |Ψ2 i, · · · , |Ψn i}
with corresponding probabilities {p1 , p2 , · · · , pn }. In experiment B, instead, she
generates m non-interacting quantum systems, each of them prepared in its cor-
responding lower energy state {|Φ1 i, |Φ2 i, · · · , |Φm i}. Let ρA and ρB be the
density matrix operators for the quantum states prepared in experiments A and
B, respectively.

(a) Let Pj and Pk be the projectors associated to the states |Ψj i and |Ψk i pro-
duced by A. Discuss whether the product Pj Pk vanishes. Justify your
answer. [1]

Model answer: Pj Pk = |Ψj ihΨj |Ψk ihΨk |. Since the states |Ψj i and |Ψk i
are in general not orthogonal, then hΨj |Ψk i =
6 0 in general. Hence Pj Pk
does not vanish in general.

(b) Derive expressions for ρA and ρB and justify your answers in detail. [2]
Model answer: In experiment A the physicist is generating an statistical
mixture of quantum states, hence
n
X
ρA = pj |Ψj ihΨj |.
j=1

In experiment B the physicist is generating non interacting systems, each

of them in a specific pure quantum state, hence

ρB = |Φ1 ihΦ1 | ⊗ |Φ2 ihΦ2 | ⊗ · · · ⊗ |Φm ihΦm |.

Marks: 1 for each correct density matrix operator and 1 for justification.

(c) Let OA be a Hermitian operator describing an observable of the system

produced by A. Derive an expression for the expected value of OA . [1]
Model answer:
n
X
hOA i = Tr[ρA OA ] = pj hΨj |OA |Ψj i.
j=1
 k
(d) Let OB be a Hermitian operator describing an observable of the kth system
k
produced by B. Derive an expression for the expected value of OB . [1]
Model answer:
k
hOB i = hΦk |OB |Φk i.
Jusitfication: the reduced state of the kth system is ρk = |Φk ihΦk | i.e.
sytem k is in a pure state.

7. In this question we will explore why we can neglect term S2 (x) in the expansion
of S(x) in the derivation of the WKB wavefunction, when the potential V (x) is
slowly varying with respect to the local de Broglie wavelength.
In the lectures we showed that the TISE may be written, using the Ansatz
ψ(x) = exp[iS(x)] as:

i~ 00 1 0 2
− S (x) + S (x) + V (x) − E = 0.
2m 2m
where 0 denotes the derivative with respect to x, and 00 the second derivative with
respect to x.
We then replaced S(x) with the series expansion S(x) = j ~j Sj (x) and de-
P

rived zeroth and first order equations from which we derived expressions:
i
S00 (x) = p(x) S1 (x) = log p(x)
2
where for simplicity I have neglected the ± signs.

(a) By considering terms in the above equation proportional to ~2 derive the

second order equation:

iS100 (x) − S10 (x)2

S20 (x) =
2S00 (x)

Model answer: We consider only the terms in our rewritten TISE which
are to the second order in ~, e.g. proportional to ~2 . This gives us:
i~ 1 2 0 2
− ~S100 (x) + ~ (S1 (x) + 2S00 (x)S20 (x)) = 0.
2m 2m
We cancel the ~2 and 2m factors and rearrange to achieve the above ex-
pression.
(b) Using the definitions for S00 (x) and S1 (x) above show that:

1 3p0 (x)2 2p00 (x)

 
0
S2 (x) = −
8 p(x)3 p(x)2

We need to compute S10 (x) and S100 (x).

i p0 (x)
S10 (x) =
2 p(x)
via the chain rule. Using the product rule or quotient rule, we achieve:
i p00 (x) p0 (x)2
 
00
S1 (x) = −
2 p(x) p(x)2
and substituting this into the expression derived in 1(a) recovers the correct
expression.

(c) And verify that the following expression is a solution to this differential
equation:
1 p0 (x) 1 x p0 (y)2
Z
S2 (x) = − − dy,
4 p(x)2 8 p(y)3
where y is a dummy integration variable.

To verify, we differentiate this expression once:

1 p00 (x) p0 (x)2 1 p0 (x)2
 
0
S2 (x) = − − 2 − .
4 p(x)2 p(x)3 8 p(x)3
using the product or quotient rules and the fundamental theorem of calcu-
lus. Simplifying we verify that the expression is indeed a solution to the
above differential equation.

(d) To be able to neglect the term containing S2 (x) in the series expansion for
S(x) we need that |~S2 (x)|  1. We will focus solely on the first term
in S2 (x) (a similar analysis can be performed on the second term, but is
more complicated due to the integral). If the local de Broglie wavelength
is defined λ(x) = h/p(x) show that
 0 
 p (x) 
 p(x)2   1
~ 
 implies  
1 0 
 λ (x)  1.
 2π 

We differentiate λ(x):

hp0 (x)
λ0 (x) = −
p(x)2
and substitute it into the first inequality. Upon cancelling the h factor we
achieve the second inequality.

[4]
Marks: total marks are 4. 1 mark for each numeral and partial marks for partial
answers.