QuIC Seminar 2

The Quantum Teleportation Algorithm

Contents
2.1 Problem Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Multi-Qubit States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Multi-Qubit Gates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 The No-Cloning Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Quantum Teleportation Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.1 Problem Setup

The problem that quantum teleportation looks to solve is as follows: Alice has an unknown qubit |ψ⟩ that
she wants to send to Bob. How can she do this?
It’s interesting to note that this is non-trivial in quantum information science, whereas in classical in-
formation theory the problem is no problem at all. How can we do this classically (i.e., send an unknown
classical bit)? We would simply “look at” the information (i.e., read it by some means) and then send a
message to Bob containing that information.
Why can’t we do this with a qubit? It’s because when we make a measurement, we disturb the state
of the system according to the following measurement postulate of quantum mechanics. We saw a special
case of this in the previous lecture when we measured a qubit in the computational basis. Here, our
measurement operators were M0 := |0⟩⟨0| and M1 := |1⟩⟨1|, which we “sandwiched” between our qubit
to get the probability of obtaining either the zero state (⟨ψ|M0† M0 |ψ⟩ = ⟨ψ|0⟩⟨0|ψ⟩) or the probability
of obtaining the one state (⟨ψ|M1† M1 |ψ⟩ = ⟨ψ|1⟩⟨1|ψ⟩). These measurement operators are Hermitian and
satisfy the completeness relation M0† M0 + M1† M1 = I.

Exercise 11: Verify that the operators M0 := |0⟩⟨0| and M1 := |1⟩⟨1| satisfy the completeness relation
M0† M0 + M1† M1 = I.

Solution 4: Follows by direct computation.

General measurements in quantum mechanics are described by the following, which also provides a rule
for the state of a quantum system after a measurement is made.

Definition 2.1 (Measurement Postulate of Quantum Mechanics). Quantum measurements are de-
scribed by a collection {Mm } of measurement operators that act on the space of the quantum system

9
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 10

being measured (e.g., C2 for qubits) and satisfy the completeness equation
∑
Mm†
Mm = I. (2.1)
m

The index m refers to all measurement outcomes that may occur. If the state of the system is |ψ⟩
immediately before the measurement, then the probability that result m occurs is given by
†
p(m) := ⟨ψ|Mm Mm |ψ⟩. (2.2)

The state of the system immediately after the measurement is given by

M |ψ⟩
√m . (2.3)
p(m)

This postulate tells us how general measurements occur in quantum computing, a special case of which
we saw when measuring qubits in the computational basis in Definition 1.10.
For our particular example of quantum teleportation, (2.3) is the most relevant. This feature says that
the state of Alice’s qubit would change if she measured it. Thus, in the quantum case, she cannot simply
read the information in the qubit and send that to Bob.
Example 1: Completeness Equation is about Preserving Probability.
Why do we require that measurement operators {Mm } satisfy the completeness equation (2.1)? What
does this∑mean? It really expresses the fact that the sum of measurement probabilities must be equal to
one, i.e. m p(m) = 1.
To see this, we can write
∑ ∑
†
p(m) = ⟨ψ|Mm Mm |ψ⟩
m m
( )
∑
†
= ⟨ψ| Mm Mm |ψ⟩.
m
∑
Thus if (2.1) holds, we have m p(m) = ⟨ψ|ψ⟩ = 1, as desired.

Exercise 12: Suppose we measure the qubit |+⟩ = H|0⟩ in the computational basis and obtain the
zero outcome. What is the state immediately after this measurement? Using the irrelevance of global phase
(Theorem 1.3), argue that the qubit is in the |0⟩ state immediately after measurement.
So Alice cannot measure her qubit and send that information to Bob without disturbing the state. Perhaps
she can simply copy the state |ψ⟩ onto another qubit she has, say |ϕ⟩, and send this qubit to Bob? It turns
out that in general this is not possible in quantum mechanics, as described by the no-cloning theorem. To
understand this theorem (which deals with multiple qubits), we first need to understand quantum systems
consisting of more than one qubit.

2.2 Multi-Qubit States

A quantum computer with just one qubit wouldn’t be very powerful, even with superposition. Suppose we
have two qubits in our quantum computer, denoted |ψ⟩ and |ϕ⟩. The total state of the whole computer, |Ψ⟩,
is given as the tensor product of its qubits, |Ψ⟩ = |ψ⟩ ⊗ |ψ⟩. Since each qubit on our two-qubit computer can
be in one of two states (|0⟩ or |1⟩), the possibilities for the state of the whole computer are

10
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 11



 |0⟩ ⊗ |0⟩

|0⟩ ⊗ |1⟩
|Ψ⟩ = . (2.4)

 |1⟩ ⊗ |0⟩


|1⟩ ⊗ |1⟩
But because of superposition, any linear combination of these states is also a valid state on our two qubit
quantum computer. That is to say—the states in (2.4) form a basis for all possible states on a two-qubit
quantum computer:

|Ψ⟩ = a|0⟩ ⊗ |0⟩ + b|0⟩ ⊗ |1⟩ + c|1⟩ ⊗ |0⟩ + d|1⟩ ⊗ |1⟩ (2.5)
such that the complex amplitudes satisfy |a| + |b| + |c| + |d| = 1.
2 2 2 2

This example illustrates the following:

Theorem 2.1 (Bases for Tensor Product Spaces.). Let |i⟩A , i = 0, 1, ..., M − 1 be a basis for the Hilbert
space HA and |j⟩B , j = 0, 1, ..., N − 1 be a basis for the Hilbert space HB . Then, the M N states
|i⟩A ⊗ |j⟩B form a basis for the composite space HAB = HA ⊗ HB . That is, any state |ψ⟩AB ∈ HAB can
be written ∑
|ψ⟩AB = αi,j |i⟩A ⊗ |j⟩B (2.6)
i,j
∑
such that i,j |αi,j |2 = 1. The size of the composite space HAB is thus |HAB | = M N .

This theorem generalizes our previous example of two qubits to arbitrary Hilbert spaces. (In our two
qubit example, HA = HB = C2 and |i⟩A = |j⟩B = {|0⟩, |1⟩}.) A particular note is that the size of the
composite space HAB is the product of the sizes of HA and HB . This leads to exponentially large Hilbert
spaces and the common saying: “Hilbert space is a big place.”
Example 2: The size of a Hilbert Space of n Qubits.
How many basis states are there for the composite Hilbert space consisting of n qubits? Each qubit has
a basis of size two, hence the composite space has 2n basis states.

We can enumerate the basis states of n qubits with bit strings of length n:

|z1 ⟩ ⊗ |z2 ⟩ ⊗ · · · ⊗ |zn ⟩ (2.7)

where each zi ∈ {0, 1}. Because it’s a bit tedious to write out all of the tensor products, shorthand notation
is often used, listed below. First, we can just omit the tensor product symbols to write

|z1 ⟩|z2 ⟩ · · · |zn ⟩ := |z1 ⟩ ⊗ |z2 ⟩ ⊗ · · · ⊗ |zn ⟩. (2.8)

This is shorter, but not as short as it can be: often, we will only write one ket for such a state and
separate the bits zi with commas,

|z1 , z2 , ..., zn ⟩ := |z1 ⟩ ⊗ |z2 ⟩ ⊗ · · · ⊗ |zn ⟩. (2.9)

Or we can even remove the commas to write

|z1 z2 · · · zn ⟩ := |z1 ⟩ ⊗ |z2 ⟩ ⊗ · · · ⊗ |zn ⟩. (2.10)

All of these notations are equivalent.
Example 3: General Two Qubit State in Shorthand Notation
In our new shorthand notation, (2.5) becomes

|ψ⟩ = a|00⟩ + b|01⟩ + c|10⟩ + d|11⟩. (2.11)

11
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 12

How do we explicitly compute a tensor product? Suppose we have two arbitrary qubits
[ ] [ ]
a c
|ψ⟩ = , |ϕ⟩ = . (2.12)
b d
Then, the tensor product |ψ⟩ ⊗ |ϕ⟩ evaluates to

[ ] ac
a|ϕ⟩ ad
|ψ⟩ ⊗ |ϕ⟩ = = 
 bc  . (2.13)
b|ϕ⟩
bd
Since both qubits live in C2 , the composite tensor product state lives in C4 .

Exercise 13: Compute the tensor product (|ψ⟩ ⊗ |ϕ⟩) ⊗ |χ⟩ where |χ⟩ := [e f ]T .
What about tensor products of basis states? It is easy to verify using (2.13) that

       
1 0 0 0
 0  1 0 0
|0⟩ ⊗ |0⟩ ≡ |00⟩ =        
0 , |0⟩ ⊗ |1⟩ ≡ |01⟩ = 0 , |1⟩ ⊗ |0⟩ ≡ |10⟩ = 1 , |1⟩ ⊗ |1⟩ ≡ |11⟩ = 0 . (2.14)
0 0 0 1

Note that the binary value of the bit string in the ket denotes the position of the one in the basis vector.
For example “00” in binary is 0 in decimal, and the entry with a one in the tensor product |00⟩ is the 0th
entry. Similarly for “01,” “10,” and “11.” This binary to decimal bijective map leads to yet another notation.
Namely, we can write |00⟩ as |0⟩2 . In the latter ket, the value 0 denotes a decimal value of 0, and the
subscript 2 indicates that there are two bits in the former ket. Similarly, |01⟩ becomes |1⟩2 , |10⟩ becomes
|2⟩2 , and |11⟩ becomes |3⟩2 . In contexts where it is absolutely clear, the subscript on the “decimal notation”
kets may sometimes be dropped.
The tensor product has multiple properties which we’ll cover in future sessions. For example, it is linear,
associative, and distributive. However, it is NOT necessarily commutative.

Exercise 14: For |ψ⟩ and |ϕ⟩ given in (2.12), verify by direct computation that |ψ⟩ ⊗ |ϕ⟩ ̸= |ϕ⟩ ⊗ |ψ⟩.
For now, we are just concerned with how states with multiple qubits work, for which are preliminary
discussion of the tensor product will suffice.
We will mention several important classes of states that appear in quantum computing and quantum
information science.

Definition 2.2 (Bipartite State). A bipartite state is a state on the composite space HAB = HA ⊗ HB
consisting of two “parts” (hence bipartite) HA and HB .

Definition 2.3 (Product State). A product state is a bipartite state on HAB that can be written of
the form |ψ⟩A ⊗ |ϕ⟩B where |ψ⟩A ∈ HA and |ϕ⟩B ∈ HB .

Definition 2.4 (Separable State). A separable state is a convex combination of product states, i.e., a
state that can be written as ∑
ci,j |i⟩A ⊗ |j⟩B (2.15)
i,j
∑
such that i,j pi,j = 1.

12
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 13

Definition 2.5. An entangled state is any state that cannot be written as a separable state.

What’s different about entangled states? Consider the following example.

Suppose Alice and Bob have qubits in Hilbert spaces HA and HB , respectively. If the composite state of
their qubits were
|0⟩A ⊗ |0⟩B , (2.16)
then we would be correct in saying that Alice’s qubit is in the state |0⟩ and Bob’s qubit is in the state |0⟩.
Similarly if the composite state of their qubits were
|1⟩A ⊗ |1⟩B , (2.17)
we would again correctly say that Alice’s qubit is in the |1⟩ state and Bob’s qubit is in the |1⟩ state.
Now, suppose that the composite state of their qubits were
1
√ (|00⟩ + |11⟩) . (2.18)
2
How would you describe Alice’s qubit? Bob’s? The answer is you cannot describe either of their qubits
individually. The state of the entire composite system is specified, but neither component state can be
determined. Alice’s and Bob’s qubits are entangled.
Entanglement plays a key role in the quantum teleportation algorithm and many other algorithms. Before
we finally describe this algorithm, we first need to discuss how quantum gates can act on multiple qubits.
Example 4: Entanglement is Quantum Correlation.
What does “correlation” mean for two classical variables, x and y? It means that they are connected
in some way. For example, if “x increases, then y increases” is an example of two variables that are
(positively) correlated. Negative correlations are of the form “if x increases, y decreases.” In either case,
there is some form of connection between the two variables.
Entanglement is quantum correlation, that is, correlation between two quantum states. Consider the
state given in (2.18), commonly known as a Bell state or EPR pair and denoted
1
|Φ+ ⟩ := √ (|00⟩ + |11⟩) . (2.19)
2
Suppose we measure the first qubit and obtain the zero outcome. What is the state immediately after
measurement? We use Definition 2.1, specifically (2.3) and M0 = |0⟩⟨0| as a measurement on the first
qubit only. Then, the state immediately after the measurement is

M0 |Φ+ ⟩
√ = (|0⟩⟨0| ⊗ I) (|0⟩|0⟩ + |1⟩|1⟩) = |0⟩|0⟩. (2.20)
p(0)

(We use (|0⟩⟨0| ⊗ I) because we only measure the first qubit and do nothing (or act with the identity
operator) on the second qubit.)
Now look at what (2.20) tells us. If we measure the first qubit (Alice’s qubit) to be in the zero state,
then immediately after the measurement the second qubit (Bob’s qubit) will also be in the zero state! The
qubits are somehow correlated in a quantum sense that we call entanglement. If Bob measured his qubit
after, he would always obtain 0 with 100% probability. Alice’s and Bob’s measurement statistics would
be perfectly correlated.
Here’s the interesting thing: Alice’s measurement was local. It only acts on her qubit, she only touches
her qubit, she never looks at Bob’s qubit. However, despite this, the measurement somehow non-locally
affects Bob’s qubit. This is a demonstration of non-locality in quantum mechanics, which means that local
actions can influence objects far away. (Imagine if Alice and Bob created an entangle pair of qubits then
ran off to opposite sides of the world.)

Exercise 15: Suppose that Alice measured her qubit in (2.19) and got the one outcome. Similar to
Example 2.2, show that the state immediately after the measurement is |11⟩.

13
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 14

2.3 Multi-Qubit Gates

Now that we have multi-qubit states, we need to introduce multi-qubit gates. Just as single qubit gates
determine the evolution of single qubits, multi-qubit gates determine the evolution of multiple qubits. Clear
analogues exist in classical information processing. For example, the AND gate inputs two bits and outputs
one bit. The output bit is one if both inputs are one and zero otherwise.
One of the most common two-qubit gates in quantum computing is the Controlled-NOT gate, commonly
abbreviated CNOT. CNOT acts on two qubits, one of which is called control qubit, and the other called the
target qubit. The CNOT gate flips the target qubit if the control qubit is in the one state. If the control
qubit is not in the one state, then CNOT does nothing to the target qubit.
For example, suppose we have a two qubit state |00⟩ that we want to perform CNOT0,1 on. Here, the
subscript means we want the first qubit (indexed 0) as the control and the second qubit (indexed 1) as the
target. Since the first qubit is in the |0⟩ state, nothing happens to the target qubit. The final state is thus
CNOT0,1 |00⟩ = |00⟩. Similarly, CNOT0,1 |01⟩ = |01⟩ because the control qubit is still in the zero state.
The next two cases are CNOT0,1 |10⟩ = |10⟩ and CNOT0,1 |11⟩ = |10⟩. Here, in both cases we flip the
target qubit because the control qubit is in the one state. In summary:

Definition 2.6 (Controlled-NOT Gate). The Controlled-NOT gate CNOT0,1 (controlling on the first
qubit with the second qubit as the target) has the following action on computational basis states:

CNOT0,1 |00⟩ = |00⟩, CNOT0,1 |01⟩ = |01⟩, CNOT0,1 |10⟩ = |11⟩, CNOT0,1 |11⟩ = |10⟩.
(2.21)

Exercise 16: Verify that a matrix representation for CNOT0,1 is

 
1 0 0 0
0 1 0 0 
CNOT0,1 =  0 0 0 1 
 (2.22)
0 0 1 0
in the computational basis {|00⟩, |01⟩, |10⟩, |11⟩}.

Exercise 17: Prove, without using matrix multiplication, that CNOT0,1 squares to the identity.

Exercise 18: Write down the action of CNOT1,0 (control and target qubit reversed) on the computa-
tional basis states {|00⟩, |01⟩, |10⟩, |11⟩}. Write out the matrix in this basis as well.
In future seminars, we’ll see more two-qubit and multi-qubit gates. In general, m qubit gates are just
unitary operators on m qubits. For the quantum teleportation algorithm, the CNOT gate is all we need to
discuss.

2.4 The No-Cloning Theorem

We now have the ingredients—namely, multi-qubit states and multi-qubit gates—to prove the no-cloning
theorem. Recall the problem of quantum teleportation: Alice has some unknown qubit that she wants to
send to Bob. We already ruled out measuring the qubit to see what it was, for we saw that this would
inherently modify the qubit. Why don’t we just make a copy of it? Well, unfortunately this is what the
no-cloning theorem forbids. Casually, it states that we cannot clone arbitrary quantum states by unitary
evolution. A formal statement of the theorem is given below.

Theorem 2.2 (No-Cloning Theorem). Suppose U is a unitary operator that serves as a “quantum copy
machine.” That is, U is a two-qubit operator that copies (or clones) the first qubit into the second qubit

U |ψ⟩|s⟩ = |ψ⟩|ψ⟩ (2.23)

14
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 15

where |s⟩ is some arbitrary initial qubit. Then, U can only clone qubits parallel or orthogonal to |ψ⟩.

Proof. The proof of this theorem is actually rather simple. Since the theorem is about the ability of U to
clone other states, let’s right down (2.23) for some other arbitrary state |ϕ⟩,

U |ϕ⟩|s⟩ = |ϕ⟩|ϕ⟩. (2.24)

Now we’re going to “compare” (2.23) and (2.24) in some sense by taking an inner product between them.
First, if we consider the left-hand sides of these equations, then we have

⟨ϕ| ⊗ ⟨s|U † U |ψ⟩ ⊗ |s⟩ = ⟨ϕ|ψ⟩⟨s|s⟩ = ⟨ϕ|ψ⟩ (2.25)

where we have used that U is unitary and that inner products “distribute” across tensor products. If we
consider the right-hand sides of (2.23) and (2.24) when taking inner products, then we have

(⟨ϕ| ⊗ ⟨ϕ|) (|ψ⟩ ⊗ |ψ⟩) = ⟨ϕ|ψ⟩⟨ϕ|ψ⟩ = ⟨ϕ|ψ⟩2 . (2.26)

But of course the expressions (2.25) and (2.26) must be equal, i.e.

⟨ϕ|ψ⟩ =< ϕ|ψ⟩2 . (2.27)

This equation is of the form x2 = x (where x ≡ ⟨ϕ|ψ⟩) which has only two solutions: x = 0 or x = 1. If
x = 1, then |ϕ⟩ and |ψ⟩ are parallel. If x = 0, then |ϕ⟩ and |ψ⟩ are orthogonal. Thus, U can only possibly
clone qubits parallel or orthogonal to |psi⟩, which completes the proof.
So quantum information forbids Alice explicitly reading her qubit and sending that information to Bob,
and also forbids Alice cloning her qubit and sending the clone to Bob. (This would not work in general, as
we have seen, and we have no information about the qubit she wants to send.) There is still a way for Alice
to send her unknown qubit to Bob via the interesting result of quantum teleportation.

2.5 Quantum Teleportation Algorithm

Quantum teleportation involves an intermediary qubit between Alice and Bob. It relies on quantum en-
tanglement, specifically creating an EPR pair that is “shared” between Alice and Bob. The circuit for the
teleportation algorithm is shown below.

|ψ⟩ • H ✌✌✌ •

|0⟩ H • ✌✌✌ •

|0⟩ X Z

Figure 2.1: Circuit diagram for the quantum teleportation algorithm.

We assume the the first qubit (the top qubit in the diagram), which belongs to Alice, starts out in the
|ψ⟩ state, since this is what the quantum teleportation algorithm assumes. The bottom two qubits belong
to Bob. The first thing Bob does is create an EPR pair (2.19) with his two qubits, doing a Hadamard on
one then performing a controlled not gate.
To see that this circuit indeed creates an EPR pair, note that we start out with both qubits in the |0⟩⊗|0⟩
state. We then perform a Hadamard gate on the first qubit and do nothing on the second qubit. Thus, the
overall multi-qubit gate that we perform is H ⊗ I. The action of this gate on the |0⟩ ⊗ |0⟩ state is
1 1
(H ⊗ I)|0⟩ ⊗ |0⟩ = H|0⟩ ⊗ I|0⟩ = √ (|0⟩ + |1⟩) ⊗ |0⟩ = √ (|00⟩ + |10⟩) (2.28)
2 2

15
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 16

This is almost an EPR pair. To make it exactly an EPR pair, we can perform a CNOT, controlling on the
first qubit. Doing so yields
1 1 1
CNOT0,1 √ (|00⟩ + |10⟩) = √ (CNOT0,1 |00⟩ + CNOT0,1 |10⟩) = √ (|00⟩ + |11⟩) (2.29)
2 2 2
In the first step, we used linearity, and in the second step we used the definition of the CNOT gate (Defini-
tion 2.21).
We now have Bob’s qubits in an EPR pair. The next step is to have Alice perform a CNOT gate between
her qubit and one of Bob’s qubits. (This is the “shared” qubit part that we mentioned before.) Note that
the state of the entire system (Alice + Bob’s qubits) is
1
√ |ψ⟩ ⊗ (|00⟩ + |11⟩) . (2.30)
2
How can we perform a CNOT gate with arbitrary state |ψ⟩? The key is to express it in terms of the
computational basis states
|ψ⟩ = α|0⟩ + β|1⟩ (2.31)
for some unknown coefficients α, β ∈ C. If we distribute this term into the entire state (2.30), we get four
terms describing the whole state
1
√ (α|000⟩ + α|011⟩ + β|100⟩ + β|111⟩) (2.32)
2
This expression enables us to easily perform a CNOT gate between the first and second qubits. Doing so,
we obtain the state
1
√ (α|000⟩ + α|011⟩ + β|110⟩ + β|101⟩) (2.33)
2
We then perform a Hadamard gate on the first qubit to get
1
|Ψ⟩ := (α|000⟩ + α|100⟩ + α|011⟩ + α|111⟩ + β|010⟩ − β|110⟩ − β|101⟩ + β|001⟩) . (2.34)
2
These are all the gate operations that we need to perform before we measure. To get a better idea of the
possible outcomes we can get when we measure, we can group terms containing the same state for the first
two terms. There are four possibilities for the first two qubits, namely |00⟩, |01⟩, |10⟩, and |11⟩. Grouping
these terms together, we can rewrite the state as
1
|Ψ⟩ = (|00⟩(α|0⟩ + β|1⟩ + |01⟩(α|1⟩ + β|0⟩ + |10⟩(α|0⟩ − β|1⟩ + |11⟩(α|1⟩ − β|0⟩) . (2.35)
2

Exercise 19: Expand each term in (2.35) to get the same state in (2.34), thus verifying they are equal.

Now Alice measures her two qubits (the first two qubits). Suppose she gets the |00⟩ outcome. Then,
Bob’s qubit is in the state α|0⟩ + β|1⟩, which is |ψ⟩ exactly. In this case, we are done, and Alice’s unknown
qubit |ψ⟩ has been teleported to Bob.
This is not the only possible measurement outcome Alice could obtain, however. Suppose instead she
measures |01⟩. By looking at (2.35), we see that Bob’s qubit in this case would be α|1⟩ + β|0⟩. This is almost
|ψ⟩, but the amplitudes are flipped. How can we get |ψ⟩ exactly from this state? We can perform a NOT
gate X(α|1⟩ + β|0⟩) = α|0⟩ + β|1⟩ = |ψ⟩. If Bob performs a NOT gate on his qubit in this case, then he
again obtains |ψ⟩ exactly.
The other two measurement cases can easily be worked out as well.

Exercise 20: If Alice measures |10⟩, prove that Bob’s qubit is in the state α|0⟩ − β|1⟩. Further, prove
that he can obtain |ψ⟩ by performing a Pauli-Z gate on his qubit. That is, Z(α|0⟩−β|1⟩) = |ψ⟩ := α|0⟩+β|1⟩.

16
QUIC SEMINAR 2. THE QUANTUM TELEPORTATION ALGORITHM 17

Exercise 21: If Alice measures |11⟩, prove that Bob’s qubit is in the state α|1⟩ − β|0⟩. Further, prove
that he can obtain |ψ⟩ by performing a NOT gate and then performing Pauli-Z gate on his qubit. That is,
ZX(α|1⟩ − β|0⟩) = |ψ⟩ := α|0⟩ + β|1⟩. Also prove that he can first do Z and then do X to still obtain |ψ⟩.
(In general Z and X anticommute, i.e. ZX = −XZ, which you may want to prove by direct computation.)
Thus, in all four measurement cases, Bob is able to obtain exactly Alice’s unknown qubit |ψ⟩ without
either knowing what the actual state is. For clarity, we enumerate all measurement possibilities in the table
below.

Alice measures: Bob performs:

|00⟩ I
|01⟩ X
|10⟩ Z
|11⟩ ZX

Table 2.1: All possible measurement outcomes of the quantum teleportation algorithm.

Note that the controlled-gates after the measurements on Alice’s qubits in Figure 2.1 perform exactly these
operations. After the measurement, we have classical information, which we represent in circuit diagrams
with two lines (instead of one line for qubits). The controlled-X and controlled-Z are thus conditional on
these measurement outcomes. If the first qubit comes out to be a zero, then we do not perform an X gate
on Bob’s qubit. If the first qubit does come out to be zero, then we do perform an X gate. Similarly for the
controlled-Z in the standard sense of controlled operations.
Note that there must be some form of (classical) communication between Alice and Bob after Alice makes
measurements and before Bob performs operations. If Alice measures |00⟩, she has to tell Bob “I measured
|00⟩,” at which point Bob knows to do nothing. Similarly for the other three possibilities. Without this
communication, Bob would have no idea what to do.
In this sense, it is important to emphasize that no information, classical or quantum, is really being
“teleported” here, despite the name of the algorithm. Alice’s classical “communication” is of course limited
by the speed of light. Although “teleport” may imply instantaneous information transfer, that is not what is
happening here. It is interesting nonetheless that neither Alice nor Bob know anything about the actual state
of the qubit |ψ⟩ being teleported. We’ve shown that |ψ⟩ could be any arbitrary qubit and the teleportation
algorithm works just fine.

Example 5: Quantum Teleportation with a Particular Qubit

In the teleportation circuit shown in Figure 2.1, we assumed |ψ⟩ was already input to the circuit. In this
setting, we think of Alice in a lab setting performing some experiment that produces an unknown qubit.
Practically with quantum circuits, we always have control over what states we have in our algorithms.
Let’s suppose we wanted to construct an explicit circuit to teleport a given qubit, just as an illustrative
example. Let’s arbitrarily pick Alice’s qubit to be |1⟩.
Conventionally, all qubits start in the ground state |0⟩ at the beginning of a quantum algorithm. So
we need some way of preparing |1⟩ from |0⟩. This can be done with a NOT gate X.
The entire quantum teleportation circuit for this particular example thus takes the following form
shown in Figure 2.2. An example sending arbitrary states can be found at this link on the visual quantum
circuit simulator Quirk by Craig Gidney. [citation needed]

|0⟩ X • H ✌✌✌ •

|0⟩ H • ✌✌✌ •

|0⟩ X Z

Figure 2.2: Quantum teleportation algorithm for the particular qubit |1⟩.

17