174 Quantum circuits

4.2 Single qubit operations

The development of our quantum computational toolkit begins with operations on the
simplest quantum system of all – a single qubit. Single qubit gates were introduced in
Section 1.3.1. Let us quickly summarize what we learned there; you may find it useful
to refer to the notes on notation on page xxiii as we go along.
A single qubit is a vector |ψ = a|0 + b|1 parameterized by two complex numbers
satisfying |a|2 + |b|2 = 1. Operations on a qubit must preserve this norm, and thus are
described by 2×2 unitary matrices. Of these, some of the most important are the Pauli
matrices; it is useful to list them again here:
     
0 1 0 −i 1 0
X≡ ; Y ≡ ; Z≡ . (4.1)
1 0 i 0 0 −1
Three other quantum gates will play a large part in what follows, the Hadamard gate
(denoted H), phase gate (denoted S), and π/8 gate (denoted T ):
     
1 1 1 1 0 1 0
H=√ ; S= ; T = . (4.2)
2 1 −1 0 i 0 exp(iπ/4)
√
A couple of useful algebraic facts to keep in mind are that H = (X + Z)/ 2 and S = T 2 .
You might wonder why the T gate is called the π/8 gate when it is π/4 that appears in
the definition. The reason is that the gate has historically often been referred to as the
π/8 gate, simply because up to an unimportant global phase T is equal to a gate which
has exp(±iπ/8) appearing on its diagonals.
 
exp(−iπ/8) 0
T = exp(iπ/8) . (4.3)
0 exp(iπ/8)
Nevertheless, the nomenclature is in some respects rather unfortunate, and we often refer
to this gate as the T gate.
Recall also that a single qubit in the state a|0 + b|1 can be visualized as a point (θ, ϕ)
on the unit sphere, where a = cos(θ/2), b = eiϕ sin(θ/2), and a can be taken to be real
because the overall phase of the state is unobservable. This is called the Bloch sphere
representation, and the vector (cos ϕ sin θ, sin ϕ sin θ, cos θ) is called the Bloch vector.
We shall return to this picture often as an aid to intuition.

Exercise 4.1: In Exercise 2.11, which you should do now if you haven’t already done
it, you computed the eigenvectors of the Pauli matrices. Find the points on the
Bloch sphere which correspond to the normalized eigenvectors of the different
Pauli matrices.
The Pauli matrices give rise to three useful classes of unitary matrices when they are
exponentiated, the rotation operators about the x̂, ŷ, and ẑ axes, defined by the equations:
 
−iθX/2 θ θ cos θ2 −i sin θ2
Rx (θ) ≡ e = cos I − i sin X = (4.4)
2 2 −i sin θ2 cos θ2
 
−iθY /2 θ θ cos θ2 − sin θ2
Ry (θ) ≡ e = cos I − i sin Y = (4.5)
2 2 sin θ2 cos θ2
 −iθ/2 
−iθZ/2 θ θ e 0
Rz (θ) ≡ e = cos I − i sin Z = . (4.6)
2 2 0 eiθ/2
 Single qubit operations 175

Exercise 4.2: Let x be a real number and A a matrix such that A2 = I. Show that
exp(iAx) = cos(x)I + i sin(x)A. (4.7)
Use this result to verify Equations (4.4) through (4.6).
Exercise 4.3: Show that, up to a global phase, the π/8 gate satisfies T = Rz (π/4).
Exercise 4.4: Express the Hadamard gate H as a product of Rx and Rz rotations and
eiϕ for some ϕ.
If n̂ = (nx , ny , nz ) is a real unit vector in three dimensions then we generalize the
previous definitions by defining a rotation by θ about the n̂ axis by the equation
   
θ θ
Rn̂ (θ) ≡ exp(−iθ n̂ · σ /2) = cos I − i sin (nx X + ny Y + nz Z) , (4.8)
2 2
where σ denotes the three component vector (X, Y, Z) of Pauli matrices.

Exercise 4.5: Prove that (n̂ · σ )2 = I, and use this to verify Equation (4.8).
Exercise 4.6: (Bloch sphere interpretation of rotations) One reason why the
Rn̂ (θ) operators are referred to as rotation operators is the following fact, which
you are to prove. Suppose a single qubit has a state represented by the Bloch
vector λ. Then the effect of the rotation Rn̂ (θ) on the state is to rotate it by an
angle θ about the n̂ axis of the Bloch sphere. This fact explains the rather
mysterious looking factor of two in the definition of the rotation matrices.
Exercise 4.7: Show that XY X = −Y and use this to prove that
XRy (θ)X = Ry (−θ).
Exercise 4.8: An arbitrary single qubit unitary operator can be written in the form
U = exp(iα)Rn̂ (θ) (4.9)
for some real numbers α and θ, and a real three-dimensional unit vector n̂.
1. Prove this fact.
2. Find values for α, θ, and n̂ giving the Hadamard gate H.
3. Find values for α, θ, and n̂ giving the phase gate
 
1 0
S= . (4.10)
0 i
An arbitrary unitary operator on a single qubit can be written in many ways as a
combination of rotations, together with global phase shifts on the qubit. The following
theorem provides a means of expressing an arbitrary single qubit rotation that will be
particularly useful in later applications to controlled operations.

Theorem 4.1: (Z-Y decomposition for a single qubit) Suppose U is a unitary

operation on a single qubit. Then there exist real numbers α, β, γ and δ such that
U = eiα Rz (β)Ry (γ)Rz (δ). (4.11)
176 Quantum circuits

Proof
Since U is unitary, the rows and columns of U are orthonormal, from which it follows
that there exist real numbers α, β, γ,and δ such that
 i(α−β/2−δ/2) 
e cos γ2 −ei(α−β/2+δ/2) sin γ2
U= . (4.12)
ei(α+β/2−δ/2) sin γ2 ei(α+β/2+δ/2) cos γ2
Equation (4.11) now follows immediately from the definition of the rotation matrices and
matrix multiplication.

Exercise 4.9: Explain why any single qubit unitary operator may be written in the
form (4.12).

Exercise 4.10: (X-Y decomposition of rotations) Give a decomposition

analogous to Theorem 4.1 but using Rx instead of Rz .
Exercise 4.11: Suppose m̂ and n̂ are non-parallel real unit vectors in three
dimensions. Use Theorem 4.1 to show that an arbitrary single qubit unitary U
may be written
U = eiα Rn̂ (β)Rm̂ (γ)Rn̂ (δ), (4.13)
for appropriate choices of α, β, γ and δ.
The utility of Theorem 4.1 lies in the following mysterious looking corollary, which
is the key to the construction of controlled multi-qubit unitary operations, as explained
in the next section.

Corollary 4.2: Suppose U is a unitary gate on a single qubit. Then there exist unitary
operators A, B, C on a single qubit such that ABC = I and U = eiα AXBXC,
where α is some overall phase factor.

Proof
In the notation of Theorem 4.1, set A ≡ Rz (β)Ry (γ/2), B ≡ Ry (−γ/2)Rz (−(δ + β)/2)
and C ≡ Rz ((δ − β)/2). Note that
γ   γ    
δ+β δ−β
ABC = Rz (β)Ry Ry − Rz − Rz =I. (4.14)
2 2 2 2
Since X 2 = I, and using Exercise 4.7, we see that
 γ   γ   
δ+β δ+β
XBX = XRy − XXRz − X = Ry Rz . (4.15)
2 2 2 2
Thus
γ  γ     
δ+β δ−β
AXBXC = Rz (β)Ry Ry Rz Rz (4.16)
2 2 2 2
= Rz (β)Ry (γ)Rz (δ) . (4.17)
Thus U = eiα AXBXC and ABC = I, as required.

Exercise 4.12: Give A, B, C, and α for the Hadamard gate.

 Controlled operations 177

Exercise 4.13: (Circuit identities) It is useful to be able to simplify circuits by

inspection, using well-known identities. Prove the following three identities:
HXH = Z; HY H = −Y ; HZH = X. (4.18)
Exercise 4.14: Use the previous exercise to show that HT H = Rx (π/4), up to a
global phase.
Exercise 4.15: (Composition of single qubit operations) The Bloch
representation gives a nice way to visualize the effect of composing two rotations.
(1) Prove that if a rotation through an angle β1 about the axis n̂1 is followed by a
rotation through an angle β2 about an axis n̂2 , then the overall rotation is
through an angle β12 about an axis n̂12 given by
c12 = c1 c2 − s1 s2 n̂1 · n̂2 (4.19)
s12 n̂12 = s1 c2 n̂1 + c1 s2 n̂2 − s1 s2 n̂2 × n̂1 , (4.20)
where ci = cos(βi /2), si = sin(βi /2), c12 = cos(β12 /2), and s12 = sin(β12 /2).
(2) Show that if β1 = β2 and n̂1 = ẑ these equations simplify to
c12 = c2 − s2 ẑ · n̂2 (4.21)
s12 n̂12 = sc(ẑ + n̂2 ) − s n̂2 × ẑ ,
2
(4.22)
where c = c1 and s = s1 .
Symbols for the common single qubit gates are shown in Figure 4.2. Recall the basic
properties of quantum circuits: time proceeds from left to right; wires represent qubits,
and a ‘/’ may be used to indicate a bundle of qubits.

 
1 1 1
Hadamard √
2 1 −1 
0 1
Pauli-X
1 0
 
0 −i
Pauli-Y
i 0
 
1 0
Pauli-Z
0 −1
 
1 0
Phase
0 i
 
1 0
π/8
0 eiπ/4

Figure 4.2. Names, symbols, and unitary matrices for the common single qubit gates.

4.3 Controlled operations

‘If A is true, then do B’. This type of controlled operation is one of the most useful in
computing, both classical and quantum. In this section we explain how complex controlled
operations may be implemented using quantum circuits built from elementary operations.