Department of Physics Indian Institute of Technology Madras

PH5842: Adv. QCQI 2025 — Assignment 1 Solutions

Q1: Measurement and Quantum Operations

(i) To find the Kraus operators, we proceed as follows:
Step 1: Action on the basis states: C-NOT gate on environment by system.
USE (|0⟩S |0⟩E ) = |0⟩S |0⟩E , USE (|1⟩S |0⟩E ) = |1⟩S |1⟩E ,
USE (|0⟩S |1⟩E ) = |0⟩S |1⟩E , USE (|1⟩S |1⟩E ) = |1⟩S |0⟩E .

Step 2: We trace out the environment for an initial environment state |0⟩E :
The Kraus operators are given by: Ek = ⟨k|E USE |0⟩E , where k = 0, 1 corresponds
to the environment basis states.
Step 3: Computing the Kraus operators:
For k = 0: E0 = ⟨0|E USE |0⟩E = |0⟩⟨0|S .
For k = 1: E1 = ⟨1|E USE |0⟩E = |1⟩⟨1|S .
Step 4: Verification of completeness: To verify that these Kraus operators form a
valid quantum channel,
we check the completeness relation: E0† E0 + E1† E1 = I. Substituting:

E0† E0 = |0⟩⟨0|, E1† E1 = |1⟩⟨1|,

we can verify that: E0† E0 + E1† E1 = |0⟩⟨0| + |1⟩⟨1| = IS .

(ii) Step 1: Action of USE on |ψ⟩S ⊗ |0⟩E
1
USE (|ψ⟩S ⊗ |0⟩E ) = √ ((σX |ψ⟩S ) ⊗ |0⟩E + (σY |ψ⟩S ) ⊗ (σX |0⟩E )) .
2
Since σX |0⟩E = |1⟩E , we have:
1
USE (|ψ⟩S ⊗ |0⟩E ) = √ ((σX |ψ⟩S ) ⊗ |0⟩E + (σY |ψ⟩S ) ⊗ |1⟩E ) .
2
Step 2: Kraus Operators from Projection onto Environment Basis States {|0⟩, |1⟩}
E0 = ⟨0|E USE |0⟩E ,E1 = ⟨1|E USE |0⟩E .
√ √
From the expression for USE , we find: E0 = σX / 2, E1 = σY / 2.
Step 3: Verification of Completeness Relation: To verify that these Kraus opera-
tors form a valid quantum channel,
we check the completeness relation: E0† E0 + E1† E1 = I. Substituting:

σX † σX 2
σY † σY σ2
       
† σX I † I
E0 E0 = √ √ = = , E1 E1 = √ √ = Y = .
2 2 2 2 2 2 2 2
we can verify that: E0† E0 + E1† E1 = I/2 + I/2 = I.

1
(iii) Kraus operators for the general case: For a system-environment interaction de-
scribed by a unitary USE , followed by a measurement on the environment in the
basis {|ei ⟩}, the Kraus operators {Ek } for the induced quantum channel on the
system are given by: Ek = ⟨ek |USE |e0 ⟩, where |e0 ⟩ is the initial state of the envi-
ronment.
Probability of obtaining outcome k: The probability of obtaining outcome k
after the action of USE and measurement on the environment is: pk = Tr(Ek ρEk† ),
where ρ is the initial state of the system.
Proof: The probability of obtaining outcome k is given by the trace of the joint
system-environment state after USE , projected onto |ek ⟩:
†
pk = Tr((IS ⊗ |ek ⟩⟨ek |)USE (ρ ⊗ |e0 ⟩⟨e0 |)USE ).

Using the cyclic property of trace and the definition of Ek = ⟨ek |USE |e0 ⟩, we get:
†
pk = Tr(⟨ek |USE |e0 ⟩ρ⟨e0 |USE |ek ⟩) = Tr(Ek ρEk† ).

System state corresponding to outcome k: The state of the system ρk cor-

responding to outcome k is: ρk ∝ Ek ρEk† . More precisely:

Ek ρEk†
ρk = .
Tr(Ek ρEk† )

Proof: The unnormalized post-measurement state is:

†
ρ′k = TrE ((IS ⊗ |ek ⟩⟨ek |)USE (ρ ⊗ |e0 ⟩⟨e0 |)USE ).

Using the definition of Ek = ⟨ek |USE |e0 ⟩, we have: ρ′k = Ek ρEk† .

Normalizing this state, we get:

ρ′k Ek ρEk†
ρk = = .
Tr(ρ′k ) Tr(Ek ρEk† )

Q2: Bit-flip channel

(i) The action of the bit-flip channel is described by the quantum operation:
E(ρ) = (1 − p)ρ + pσX ρσX ,
where p is the probability of a bit-flip, and σX is the Pauli-X operator.
between a pure state |ψ⟩ and a mixed state ρ is given by:
The fidelity p
F (|ψ⟩, ρ) = ⟨ψ|ρ|ψ⟩.
Calculate the fidelity between an initial pure state |ψ⟩ and the final state after the
action of the bit-flip channel.

2
 Step 1: Final State After Bit-Flip Channel: The initial pure state is represented
as: ρin = |ψ⟩⟨ψ|. After applying the bit-flip channel, the final state becomes:

ρf = E(ρin ) = (1 − p)|ψ⟩⟨ψ| + pσX |ψ⟩⟨ψ|σX .

Step 2: Fidelity Calculation: p The fidelity between the initial pure state and the
final state is: F (|ψ⟩, ρf ) = ⟨ψ|ρf |ψ⟩. Substituting for ρf , we get:
p
F (|ψ⟩, ρf ) = ⟨ψ|[(1 − p)|ψ⟩⟨ψ| + pσX |ψ⟩⟨ψ|σX ]|ψ⟩.

Expanding this expression:

p
F (|ψ⟩, ρf ) = (1 − p)⟨ψ||ψ⟩⟨ψ||ψ⟩ + p⟨ψ|(σX |ψ⟩)(⟨ψ|σX )|ψ⟩.

Since |ψ⟩ is normalized (⟨ψ|ψ⟩ = 1), this simplifies to:

p
F (|ψ⟩, ρf ) = (1 − p) + p|⟨ψ|σX |ψ⟩|2 .

Let us denote x = |⟨ψ|σX |ψ⟩|2 , where 0 ≤ x ≤ 1. p

The fidelity can then be written as: F (|ψ⟩, ρf ) = (1 − p) + px.
Step 3: Minimum Fidelity: To find the minimum fidelity, we minimize F (|ψ⟩, ρf )
with respect to x. The minimum value of F (|ψ⟩, ρf ) occurs when x = 0, since
p > 0. Thus, the minimum fidelity is:
√
Fmin = F (|ψ⟩, (1 − p)|ψ⟩ + pσX | ⊥ψ ⟩) = Fmin 1 − P

This minimum fidelity occurs for states that are eigenstates of σz (i.e., |0⟩, and |1⟩),
which are most affected by the bit-flip operation.

(ii) Step 1: Eigenstates of σx are given by {|+⟩, |−⟩}.

The corresponding projectors are: P+ = |+⟩⟨+|, P− = |−⟩⟨−|
Step 2: Express σx in terms of P+ and P− =⇒ σx = P+ − P−
Step 3: Substitute into the bit-flip channel equation:

E(ρ) = (1 − p)ρ + p(P+ − P− )ρ(P+ − P− )

Step 4: Expand the terms: E(ρ) = (1−p)ρ+p(P+ ρP+ +P− ρP− −P+ ρP− −P− ρP+ )
Step 5: Using the completeness of projectors: (P+ + P− = I), we simplify:
• The terms P+ ρP− + P− ρP+ = −P+ ρP+ − P− ρP− + ρ.
• The remaining terms simplify to: E(ρ) = (1 − p)ρ + p(2P+ ρP+ + 2P− ρP− − ρ)
Thus: E(ρ) = (1 − 2p)ρ + 2p(P+ ρP+ + P− ρP− )
This representation provides an interesting interpretation of the bit-flip channel:
with probability (1 − 2p), the state remains unchanged, and with probability 2p,
the state is projected onto either the |+⟩ or |−⟩ basis.

3
Q3: Unital maps:

(i) To show that these channels are unital, we need to prove that E(I) = I for each
channel. Bit-flip channel: The bit-flip channel is described by:

E(ρ) = (1 − p)ρ + pσX ρσX

For the identity operator:

E(I) = (1 − p)I + pσX IσX = (1 − p)I + pI = I

Phase-flip channel: The phase-flip channel is described by:

E(ρ) = (1 − p)ρ + pσZ ρσZ

For the identity operator:

E(I) = (1 − p)I + pσZ IσZ = (1 − p)I + pI = I

Depolarizing channel: The depolarizing channel is described by:

p
E(ρ) = (1 − p)ρ + (σX ρσX + σY ρσY + σZ ρσZ )
3
For the identity operator:
p
E(I) = (1 − p)I + (σX IσX + σY IσY + σZ IσZ )
3
p
= (1 − p)I + (I + I + I) = (1 − p)I + pI = I
3
Therefore, all three channels (bit-flip, phase-flip, and depolarizing) are unital as
they leave the identity operator invariant.

(ii) We want to show that if E is a unital channel, then E(ρ) ≺ ρ. This means we need
to prove that the eigenvalue spectrum of E(ρ), denoted as ΛE(ρ) , is majorized by
the eigenvalue spectrum of ρ, denoted as Λρ . In other words, we need to find a
doubly stochastic matrix D such that ΛE(ρ) = DΛρ .
Step 1: Eigendecomposition: Start by considering the eigendecomposition of
the initial state ρ: X
ρ= λi |i⟩⟨i|
i

where λi are the eigenvalues of ρ (elements of Λρ ) and |i⟩ are the corresponding
eigenvectors.
Step 2: Action of the Unital Channel: Apply the unital channel E to ρ:
!
X X
E(ρ) = E λi |i⟩⟨i| = λi E(|i⟩⟨i|)
i i

4
Step 3: Expressing in terms of Kraus Operators: Substitute the operator-
sum representation of the quantum channel:

Ek |i⟩⟨i|Ek†
X
E(|i⟩⟨i|) =
k
!
Ek |i⟩⟨i|Ek† λi |i⟩⟨i| Ek† = Ek ρEk†
X X X X X
So, E(ρ) = λi = Ek
i k k i k

Step 4: Relating Eigenvalues: Let’s denote the eigenvalues P of E(ρ) as µj

(elements of ΛE(ρ) ), and its eigenvectors as |j ′ ⟩. Thus, E(ρ) = j µj |j ′ ⟩⟨j ′ |. We
want to relate the µj to the λi .
Step 4: Key Insight: Expressing Eigenvectors: A crucial step involves
expressing the eigenvectors |j ′ ⟩ of E(ρ) in terms of the eigenvectors |i⟩ of ρ. This
is where the unital property comes into play. Since E is unital, k Ek Ek† = I.
P

Step 5: Constructing the Doubly Stochastic Matrix: Define a matrix D

with elements: X
Dji = |⟨j ′ |Ek |i⟩|2
k
We need to show that this matrix D is doubly stochastic.
Step 6: Proof of Doubly Stochasticity:
• Non-negativity: Dji ≥ 0 since it’s the square of an absolute value.
† ′
• Column Sum: ′ 2 ′
P P P P P
j Dji = j k |⟨j |Ek |i⟩| = k j ⟨i|Ek |j ⟩⟨j |Ek |i⟩ =
P † P † P †
k ⟨i|Ek Ek |i⟩ = ⟨i|( k Ek Ek )|i⟩ = ⟨i|I|i⟩ = 1 (because k Ek Ek = I for a
CPTP map).
• Row Sum: i Dji = i k |⟨j ′ |Ek |i⟩|2 = k i ⟨j ′ |Ek |i⟩⟨i|Ek† |j ′ ⟩ = k ⟨j ′ |Ek Ek† |j ′ ⟩ =
P P P P P P

⟨j ′ |( k Ek Ek† )|j ′ ⟩ = ⟨j ′ |I|j ′ ⟩ = 1 (because k Ek Ek† = I for a unital map).

P P

Step 7: Relating Eigenvalue P Spectra: Now, show that ΛE(ρ) = DΛρ . This
means showing that µj = i Dji λi .
Consider: µj = ⟨j ′ |E(ρ)|j ′ ⟩ = ⟨j ′ | k Ek ρEk† |j ′ ⟩ = k ⟨j ′ |Ek ( i λi |i⟩⟨i|)Ek† |j ′ ⟩ =
P P P
P ′ † ′ P P ′ 2
P
i,k λi ⟨j |Ek |i⟩⟨i|Ek |j ⟩ = i λi k |⟨j |Ek |i⟩| = i λi Dji

Thus, µj = i Dji λi , which in matrix form is ΛE(ρ) = DΛρ .

P

Step : Conclusion: Since we have found a doubly stochastic matrix D such that
ΛE(ρ) = DΛρ , we have shown that E(ρ) ≺ ρ. Therefore, the final state after the
action of a unital channel is majorized by the initial state.
In
P summary: The key to this proof is recognizing that the unital property
†
( i Ei Ei = I) allows you to construct a doubly stochastic matrix that directly
relates the eigenvalue spectra of the initial and final states. This demonstrates that
the action of a unital channel increases the mixedness (or leaves it unchanged) of
the quantum state.

5
Q4: Single-qubit channels
Phase-flip channel: The phase-flip channel EP is described by the Kraus operators:E0 =
√ √
pσZ , E1 = 1 − pI
To find the effect on the Bloch vector, we need to calculate ρ′ = E(ρ) = i Ei ρEi† .
P
First, let’s express ρ in terms of the Bloch vector:
1
ρ = (I + rx σx + ry σy + rz σz )
2
Now, let’s apply the channel:

ρ′ = pσz ρσz + (1 − p)ρ

   
1 1
= pσz (I + rx σx + ry σy + rz σz ) σz + (1 − p) (I + rx σx + ry σy + rz σz )
2 2

Simplifying (noting that σz σx σz = −σx and σz σy σz = −σy ):

1
ρ′ = (I + (1 − 2p)rx σx + (1 − 2p)ry σy + rz σz )
2
Therefore, the new Bloch vector is: r⃗′ = ((1 − 2p)rx , (1 − 2p)ry , rz )
Depolarizing channel: The depolarizing channel ED is described by the Kraus opera-
tors:
r √ √ √
3p p p p
E0 = 1 − I, E1 = σx , E2 = σy , E3 = σz
4 2 2 2
Following a similar process. Apply the channel:

 
′ 3p p p p
ρ = 1− ρ + σx ρσx + σy ρσy + σz ρσz (1)
4 4 4 4
p
= (1 − p)ρ + (σx ρσx + σy ρσy + σz ρσz ) (2)
4
Simplify (noting that σi σj σi = −σj for i ̸= j):

1 p
ρ′ = (1 − p) (I + rx σx + ry σy + rz σz ) + (3I − rx σx − ry σy − rz σz )
2 4
Combining terms:
1
ρ′ = (I + (1 − p)rx σx + (1 − p)ry σy + (1 − p)rz σz )
2

Therefore, the new Bloch vector is: r⃗′ = ((1 − p)rx , (1 − p)ry , (1 − p)rz )
Pictorial representation: Pictorial representation of effects on the Bloch sphere due
to these noist channels:

6
 • The phase-flip channel shrinks the x and y components of the Bloch vector by a
factor of (1 − 2p), while leaving the z component unchanged. This results in a
flattening of the Bloch sphere into an ellipsoid along the z-axis.

• The depolarizing channel shrinks the entire Bloch vector uniformly by a factor of
(1 − p). This results in a uniform shrinking of the Bloch sphere towards the center.

These effects can be visualized as:

• Phase-flip: The Bloch sphere becomes a prolate spheroid (elongated along z-axis).

• Depolarizing: The Bloch sphere becomes a smaller sphere centered at the origin.

The severity of the deformation increases with p, with p = 0 leaving the sphere unchanged
and p = 1 (for depolarizing) shrinking it to a point at the center.
Note: Check out the appendix on the effect of the Amplitude damping channel on the
Bloch sphere.

Q5: Measurement and Quantum Operations

(i) Trace-Preserving Condition: For a quantum channel to be trace-preserving, the

Kraus operators must satisfy: X †
Ei Ei = I
i

Calculate this sum for the given Kraus operators:

   √ 
1 0 0 p
E0 = √ , E1 =
0 1−p 0 0
   
1 0 0 0
E0† E0
= †
, E1 E1 =
0 1−p 0 p
     
† † 1 0 0 0 1 0
E0 E0 + E1 E1 = + = =I
0 1−p 0 p 0 1
Therefore, the channel is trace-preserving.
Unital Condition: For a quantum channel to be unital, it must satisfy:

Ei Ei† = I
X

Let’s calculate this sum:

   
1 0 p 0
E0 E0† = , E1 E1† =
0 1−p 0 0
     
1 0 p 0 1+p 0
E0 E0† + E1 E1† = + =
0 1−p 0 0 0 1−p

7
 This sum is not equal to the identity matrix unless p = 0. Therefore, the amplitude
damping channel is not unital for non-zero p.
Conclusion: The amplitude damping channel described by the given Kraus oper-
ators is trace-preserving but not unital.

(ii) Spontaneous emission as amplitude damping: To analyze the Jaynes-Cummings

Hamiltonian for a two-level atom interacting with a quantized mode of the elec-
tromagnetic field in a cavity, we’ll proceed step by step.
Rewriting the Hamiltonian: The total Hamiltonian is given as
ℏω0
Htot = σZ + ℏωa† a + g(a† σ− + aσ+ )
2

Now, introduce the number operator:

σZ
ℏωN = ℏω(a† a + )
2
Rearranging this:
ℏω
ℏωa† a = ℏωN − σZ
2
Substituting this back into our total Hamiltonian:
ℏω0 ℏω
Htot = σZ + (ℏωN − σZ ) + g(a† σ− + aσ+ )
2 2
Simplifying:
ℏ(ω0 − ω)
Htot = ℏωN + σZ + g(a† σ− + aσ+ )
2
(ω0 −ω)
Defining the detuning δ = 2 , we get:

Htot = ℏωN + δσZ + g(a† σ− + aσ+ )

This is the desired form of the Hamiltonian.

To proof: Time-evolution operator for the atom-field system is given by:
 
δ
U (t) = e−iHt = e−iδt |00⟩⟨00| + cos Ωt + i sin Ωt |01⟩⟨01|
Ω
 
δ g
+ cos Ωt − i sin Ωt |10⟩⟨10| − i sin Ωt(|01⟩⟨10| + |10⟩⟨01|)
Ω Ω

where Ω ≡ g 2 + δ 2 and δ = (w02−w) .

p

Step 1: Hamiltonian Matrix Representation

The Hamiltonian is given by:
 
Htot = ℏwN + δσZ + g a† σ− + aσ+ ,

8
 σZ (w0 − w)
where, N = a† a + and, δ = .
2 2
We consider the basis states {|00⟩, |01⟩, |10⟩, |11⟩}, corresponding to
{|0⟩field |0⟩atom , |0⟩field |1⟩atom , |1⟩field |0⟩atom , |1⟩field |1⟩atom }.
The matrix representation of the Hamiltonian in this basis is:
 ℏω 
2 +δ 0 0 g
 0 − ℏω
2 −δ 0 0 
H=  0 3ℏω

0 2 +δ 0 
2 −δ
ℏω
g 0 0

Step 2: Time-Evolution Operator

The time-evolution operator is U (t) = e−iHt/ℏ . Since the Hamiltonian is block-
diagonal, we can exponentiate it.
ω
• The |01⟩ sector is: ei( 2 +δ)t |01⟩⟨01|
3ω
• The |10⟩ sector is: e−i( 2
+δ)t
|10⟩⟨10|
• The |00⟩ and |11⟩ sector is a 2x2 matrix. The relevant block is:
 ℏω 
2 +δ g
H2x2 =
2 −δ
ℏω
g
 
t

iδ sin( ℏt Ω) ig sin( ℏt Ω)
itω cos Ω − −
e−iH2x2 t/ℏ = e− 2ℏ  ℏ
ig sin( ℏt Ω)
Ω Ω

iδ sin( ℏt Ω) 
t
− Ω cos ℏ Ω + Ω
p
where Ω = g2 + δ2.
Step 3: Reconstructing the Full Operator
Combining all the parts, we get refer fig.1:
 
−i( ω +δ)t − itω δ
U (t) = e 2 |01⟩⟨01| + e 2ℏ cos(Ωt) − i sin(Ωt) |00⟩⟨00|+
Ω
 
itω δ itω g
e− 2ℏ cos(Ωt) + i sin(Ωt) |11⟩⟨11| − ie− 2ℏ sin(Ωt) (|00⟩⟨11| + |11⟩⟨00|) +
Ω Ω
3ω
e−i( 2
+δ)t
|10⟩⟨10|

Step 4: Addressing the Discrepancy: The target equation in the equation,

the discrepancy stems from a difference in the zero-point energy reference. We
can adjust the zero-point energy of the Hamiltonian, effectively shifting the energy
scale. By performing the transformation H → H + ℏω 2 I, we can eliminate the
explicit ℏω terms. Note that it also assumes that |11⟩ state is not reachable, or
not relevant in the dynamics. By making this adjustment to the Hamiltonian

9
 Figure 1: alternate using mathematica function MatrixExp

and dropping the |11⟩ term, we arrive at the desired equation. The modified
Hamiltonian matrix is:  
−δ 0 0
H′ =  0 δ g 
0 g −δ
Then refer fig.2,
 
−iHt −iδt δ
U (t) = e =e |00⟩⟨00| + cos Ωt + i sin Ωt |01⟩⟨01|
Ω
 
δ g
+ cos Ωt − i sin Ωt |10⟩⟨10| − i sin Ωt(|01⟩⟨10| + |10⟩⟨01|)
Ω Ω

To show that tracing out the field mode at zero-detuning (δ = 0) results in an

amplitude damping channel, we need to:
• Set δ = 0 in the time-evolution operator.
• Apply this operator to an initial state |ψ⟩ = α|0⟩ + β|1⟩ of the atom.
• Trace out the field mode.

10
 Figure 2: Using mathematica function MatrixExp

At δ = 0, the time-evolution operator simplifies to: Applying this to |ψ⟩ ⊗ |0⟩:

U (t)(α|00⟩ + β|01⟩) = α|00⟩ + β(cos(gt)|01⟩ − i sin(gt)|10⟩)

Tracing out the field mode gives the reduced density matrix for the atom:

|α| + |β|2 sin2 (gt) αβ ∗ cos(gt)

 2 
ρatom =
α∗ β cos(gt) |β|2 cos2 (gt)

This is precisely the form of an amplitude damping channel with damping param-
eter p = sin2 (gt). The Kraus operators for this channel are:
   
1 0 0 sin(gt)
E0 = , E1 =
0 cos(gt) 0 0

These match the standard form of Kraus operators for the amplitude damping
channel with p = sin2 (gt).
Thus, we have shown that tracing out the field mode at zero-detuning indeed
results in an amplitude damping channel for the atom.

(iii) Circuit Analysis: To analyze the given circuit model and derive the Kraus
operators as follows.
The circuit consists of two qubits: a system qubit (S) and an environment qubit
(E). The environment qubit starts in the |0⟩ state and is measured in the compu-
tational basis at the end. The joint system-environment unitary is composed of
two controlled operations:
• Controlled-Ry (θ) gate: Applies Ry (θ) = exp(−iθY /2) to E if S is in |1⟩
• CNOT gate: Flips E if S is in |1⟩

11
 Deriving the Kraus Operators: Let’s calculate the effect of this circuit on an arbi-
trary input state |ψ⟩ = α|0⟩ + β|1⟩ of the system qubit.
• Initial state: |ψ⟩|0⟩ = (α|0⟩ + β|1⟩)|0⟩
• After Controlled-Ry (θ): α|0⟩|0⟩ + β|1⟩(cos(θ/2)|0⟩ + sin(θ/2)|1⟩)
• After CNOT: α|0⟩|0⟩ + β|1⟩(cos(θ/2)|1⟩ + sin(θ/2)|0⟩)
Now, we measure the environment qubit. The two possible outcomes correspond
to our Kraus operators:
• E0 (measuring |0⟩ in E): |0⟩⟨0|⊗⟨0|+|1⟩⟨1|⊗⟨0|Ry (θ) = |0⟩⟨0|+sin(θ/2)|1⟩⟨1|
• E1 (measuring |1⟩ in E): |1⟩⟨1| ⊗ ⟨1|Ry (θ)CNOT = cos(θ/2)|1⟩⟨1|
We can write these in matrix form:
   
1 0 0 cos(θ/2)
E0 = , E1 =
0 sin(θ/2) 0 0

Comparison with Amplitude Damping Channel: Comparing these Kraus operators

to those of the amplitude damping channel given in part (i):
   √ 
1 0 0 p
E0 = √ , E1 =
0 1−p 0 0
We can see that our derived Kraus operators match the amplitude damping channel
√ √ √
if we set: sin(θ/2) = 1 − p, cos(θ/2) = p =⇒ θ = 2 cos−1 ( p).
Conclusion: The given circuit indeed implements the amplitude damping channel.
The parameter p in the amplitude damping channel is related to the rotation angle
θ in the circuit by p = cos2 (θ/2). This circuit provides a physical model for how
amplitude damping can arise from a system-environment interaction, where the
environment is represented by a single qubit.

Appendix:
Effect of Amplitude damping channel on Bloch sphere: If the Bloch vector of the
initial qubit density operator is ⃗r = (rx , ry , rz ), obtain the Bloch vector of the final
density operator in terms of θ, rx , ry and rz .
!
1+rz rx −iry
I + ⃗r · ⃗σ 2 2
ρin = = rx +ir y 1−rz
2 2 2

Ei ρin Ei† = E0 ρin E0T + E1 ρin E1T

X
ρout = E(ρin ) =
i

12
 ! !
  1+rz rx −iry    1+rz rx −iry 
1 0 2 2 1 0 0 sin θ/2 2 2 0 0
ρout = rx +iry 1−rz + rx +iry 1−rz
0 cos θ/2 2 2
0 cos θ/2 0 0 2 2
sin θ/2 0
! 
  1+rz rx −ir y
  rx −iry 
1 0 2 2 cos θ/2 0 sin θ/2 2 sin θ/2 0
= rx +iry 1−rz + 1−rz
0 cos θ/2 2 2 cos θ/2
0 0 0 2 sin θ/2
! 
1+rz rx −iry 1−rz

2 2 cos θ/2 4 (1 − cos θ) 0
= rx +iry +
2 cos θ/2 1−r 4 (cos θ + 1)
z 0 0
!
1 rx −iry
= 4 (3 + rz + (rz − 1) cos θ) 2 cos θ/2
rx +iry 1
2 cos θ/2 4 (1 − rz )(1 + cos θ/2)

rz′ = 2( 14 (3 + rz + (rz − 1) cos θ)) − 1 = 12 (1 − cos θ) + 12 rz (1 + cos θ) = sin2 θ/2 + rz cos2 θ/2
⃗r′ = (rx cos θ/2, ry θ/2, sin2 θ/2 + rz cos2 θ/2)

Alternative to Q5 part (iii): In figure 3, the first wire represents a ‘system’ (S) qubit
and the second wire represents the ’environment’ (E). The environment qubit starts out
in |0⟩ and is measured in the {|0⟩, |1⟩} in the end. The joint system-environment (ES)
unitary is described by a sequence of controlled-unitaries as specified in the figure.
UES
ρin • ρout

|0⟩ RY (θ) •

Figure 3: Circuit for a Quantum Channel

     
−i θ2 Y θ θ cos θ/2 − sin θ/2
RY (θ) = e = Iˆ cos − iY sin =
2 2 sin θ/2 cos θ/2
C − RY (θ) = IE ⊗ |0⟩⟨0| + RY (θ) ⊗ |1⟩⟨1|
C − N OT (reverse) = |0⟩⟨0| ⊗ IS + |1⟩⟨1| ⊗ XS

∴ USE = C − N OT (reverse) × C − RY (θ)

= (|0⟩⟨0| ⊗ IS + |1⟩⟨1| ⊗ XS )(IE ⊗ |0⟩⟨0| + RY (θ) ⊗ |1⟩⟨1|)
= |0⟩⟨0| ⊗ |0⟩⟨0| + |0⟩⟨0|RY (θ) ⊗ |1⟩⟨1| + |1⟩⟨1| ⊗ |0⟩⟨0| + |1⟩⟨1|RY (θ) ⊗ |1⟩⟨1|
 
0 sin θ/2
∴ E0 ≡E ⟨0|U |0⟩E = |0⟩⟨0| + (⟨0|RY |0⟩)|1⟩⟨1| =
0 0
 
1 0
E1 ≡E ⟨1|U |0⟩E = (⟨1|RY |0⟩)|0⟩⟨1| =
0 cos θ/2

13