Simple Algorithm
Quantum Computing · Ground State · 16 min

Quantum Information & Communication

No-cloning, teleportation and keys secured by physics.

Copy-paste is the most basic thing you can do with information, and quantum information forbids it. That one restriction, together with entanglement, gives you teleportation, two bits per qubit, and encryption keys whose secrecy is guaranteed by the laws of physics.

What you’ll learn

  • Explain why no device can copy an unknown qubit, and why a CNOT copies ∣0⟩\ket{0} and ∣1⟩\ket{1} but not ∣+⟩\ket{+}.
  • Explain why entanglement alone cannot carry a message, and why that requires the no-cloning theorem.
  • Walk through teleportation (1 Bell pair + 2 bits → 1 qubit) and superdense coding (1 Bell pair + 1 qubit → 2 bits).
  • Describe how BB84 exposes an eavesdropper, and predict the error rate she causes.

Information is physical

In everyday computing, information is abstract: a bit is a bit whether it lives in a transistor, a magnet or ink on paper. Three things are so easy that nobody thinks about them. You can copy a bit, read it without changing it, and send it down a wire.

Qubits break the first two habits. You cannot copy an unknown qubit, and you cannot read it without disturbing it: a measurement returns one bit and collapses the state. In exchange, quantum information comes with a resource that has no classical counterpart: entanglement.

TaskClassical bitQubit
Copy itAlwaysNot if the state is unknown
Read it without disturbing itAlwaysNo: one measurement gives one bit and collapses the state
Send itWire, radio, mailSend the qubit itself, or teleport it using entanglement plus two ordinary bits
Keep it secretSecurity rests on hard math problemsKey distribution whose security rests on physics
This lesson is a tour of the right-hand column.
Think first Suppose you could copy unknown qubits. What would that do to the "you only get one look" rule of measurement?

It would destroy it. Make a million copies of ∣ψ⟩=α∣0⟩+β∣1⟩\ket{\psi}=\alpha\ket{0}+\beta\ket{1}, measure some in one basis and some in others, and you could estimate α\alpha and β\beta as precisely as you like. "No copying" and "only one look" are two faces of the same fact.

Try to copy a qubit

A natural candidate for a quantum copier is the CNOT gate. Put the qubit you want to copy on the control wire and a blank qubit in ∣0⟩\ket{0} on the target. If the control is ∣1⟩\ket{1} the target flips; otherwise nothing happens:

∣0⟩∣0⟩→∣0⟩∣0⟩,∣1⟩∣0⟩→∣1⟩∣1⟩.\ket{0}\ket{0}\to\ket{0}\ket{0},\qquad \ket{1}\ket{0}\to\ket{1}\ket{1}.

That is a perfect copy of either basis state. Build it, check that it copies ∣1⟩\ket{1}, and then feed it a superposition.

Why does it fail? Quantum evolution is linear: whatever the CNOT does to ∣0⟩\ket{0} and to ∣1⟩\ket{1}, it does to each part of a superposition separately:

(α∣0⟩+β∣1⟩)∣0⟩ → α∣00⟩+β∣11⟩.(\alpha\ket{0}+\beta\ket{1})\ket{0}\ \to\ \alpha\ket{00}+\beta\ket{11}.

A true copy would be (α∣0⟩+β∣1⟩)(α∣0⟩+β∣1⟩)(\alpha\ket{0}+\beta\ket{1})(\alpha\ket{0}+\beta\ket{1}), which contains the cross terms αβ∣01⟩+αβ∣10⟩\alpha\beta\ket{01}+\alpha\beta\ket{10}. The two agree only if α\alpha or β\beta is zero. No cleverer circuit escapes this: it is a theorem.

Check your understanding

Which of these can a quantum device actually do?

  1. Copy an unknown qubit perfectly, if the machine is built precisely enough
  2. Prepare a thousand copies of ∣+⟩\ket{+}, if you are told the state is ∣+⟩\ket{+}
  3. Copy an unknown qubit by measuring it, then preparing two qubits in the state you saw
  4. Copy an unknown qubit perfectly by fanning it out with CNOTs onto several blank qubits
Show the answer

B. Prepare a thousand copies of ∣+⟩\ket{+}, if you are told the state is ∣+⟩\ket{+}

A known state can be prepared as often as you like. No-cloning only forbids copying states you do not know.

Entanglement cannot send messages

As you saw in the Entanglement topic, Alice and Bob can share a Bell pair 12(∣00⟩+∣11⟩)\tfrac{1}{\sqrt2}(\ket{00}+\ket{11}). If Alice measures her qubit and gets 0, Bob's qubit is instantly ∣0⟩\ket{0}; if she gets 1, it is ∣1⟩\ket{1}, however far apart they are.

Can Alice use this to send Bob a message faster than light? No. She cannot choose her outcome, so Bob just sees a random 0 or 1, exactly as he would if Alice had done nothing at all. The correlation shows up only when the two compare their lists, and that needs an ordinary channel, limited by the speed of light. This is the no-signalling principle.

Think first Suppose Bob could clone. Alice measures her half of the Bell pair either in ZZ (leaving Bob's qubit in ∣0⟩\ket{0} or ∣1⟩\ket{1}) or in XX (leaving it in ∣+⟩\ket{+} or ∣−⟩\ket{-}). How could Bob read her choice instantly?

He makes 1000 copies of his qubit and measures all of them in ZZ. If Alice chose ZZ, his qubit is ∣0⟩\ket{0} or ∣1⟩\ket{1}, so all 1000 results agree. If she chose XX, it is ∣±⟩\ket{\pm}, and the results are a random 50/50 mix. One glance would tell him her choice: a message sent faster than light. Relativity forbids that, so perfect cloning must be impossible, and it is.

Teleportation

Alice holds a qubit in an unknown state ∣ψ⟩\ket{\psi} and wants Bob to have it. She cannot measure it (one look reveals one bit, not α\alpha and β\beta), and she cannot copy it. But if they share a Bell pair, she can teleport it:

  1. Alice makes a joint measurement of ∣ψ⟩\ket{\psi} and her half of the pair (a CNOT, then an H, then two ordinary measurements).
  2. She gets two random bits and sends them to Bob over any classical channel: phone, email, carrier pigeon.
  3. Depending on the bits, Bob applies XX, ZZ, both or neither to his half. Now his qubit is in the state ∣ψ⟩\ket{\psi}.
Check your understanding

Alice has just measured, but her two bits have not reached Bob yet. What can Bob learn about ∣ψ⟩\ket{\psi} from his qubit right now?

  1. Nothing: on its own, his qubit is in a completely random state
  2. Everything: teleportation is instantaneous
  3. The probabilities ∣α∣2\abs{\alpha}^2 and ∣β∣2\abs{\beta}^2, but not the relative phase
  4. He can find ∣ψ⟩\ket{\psi} by trying all four corrections and seeing which one works
Show the answer

A. Nothing: on its own, his qubit is in a completely random state

Bob's qubit is ∣ψ⟩\ket{\psi}, X∣ψ⟩X\ket{\psi}, Z∣ψ⟩Z\ket{\psi} or XZ∣ψ⟩XZ\ket{\psi}, each with probability ¼. Averaged together these give a perfectly random qubit, whatever ∣ψ⟩\ket{\psi} is. The two bits are what turn this noise into ∣ψ⟩\ket{\psi}.

Superdense coding: two bits in one qubit

Teleportation spends one Bell pair and two classical bits to move one qubit. Run the trade the other way and you get superdense coding: one Bell pair plus one transmitted qubit carry two classical bits.

Alice and Bob share a Bell pair in advance. To send two bits, Alice applies one of four operations to her half only: II, XX, ZZ or ZXZX. Each turns the pair into a different one of the four Bell states. She mails her single qubit to Bob, who now holds both halves. He undoes the entangling circuit (CNOT, then H), measures, and out come both bits.

ProtocolYou spendYou get
Teleportation1 Bell pair + 2 classical bits1 qubit delivered
Superdense coding1 Bell pair + 1 qubit sent2 classical bits delivered

Keys secured by physics

Quantum key distribution (QKD) lets Alice and Bob grow a shared random key for encryption, with its secrecy guaranteed by quantum mechanics. In BB84 (Bennett & Brassard, 1984):

  1. Alice sends single photons, each carrying a random bit in a randomly chosen basis: ZZ (∣0⟩\ket{0}, ∣1⟩\ket{1}) or XX (∣+⟩\ket{+}, ∣−⟩\ket{-}).
  2. Bob measures each photon in a basis he picks at random.
  3. In public they compare bases, never bits, and keep only the rounds where the bases matched, about half. That is the sifted key.
  4. They sacrifice a random sample of it to estimate the error rate. Too many errors means someone was listening, and they abort.

An eavesdropper, Eve, is stuck like a would-be copier: she cannot clone the photons and she does not know the bases. Whenever she measures in the wrong basis, she resends a photon in the wrong state, and in a kept round Bob then gets a wrong bit half the time.

Predict before you look

Eve intercepts every photon, measures it in a random basis and resends what she saw. In the sifted key, what percentage of Bob's bits disagree with Alice's?

Show the answer

25 %

Eve picks the wrong basis half the time. When she does, the photon she resends is in the wrong basis for Bob, so his result is a coin flip, wrong half the time. The error rate is 12×12=25%\tfrac12\times\tfrac12=25\%. Without Eve (and with perfect equipment) it is 0%0\%.

The rules of the game, and what comes next

The whole topic fits in four rules:

  • Unknown quantum states cannot be copied.
  • They cannot be read without disturbance, which is what makes eavesdropping detectable.
  • Entanglement creates correlations, but by itself it cannot carry a message.
  • Resources trade at fixed rates: 1 Bell pair + 2 bits → 1 qubit (teleportation), and 1 Bell pair + 1 qubit → 2 bits (superdense coding).
Excited State lessonWhat you will do there
No-Cloning & No-SignallingProve no-cloning two ways (from linearity and from inner products), and prove that no local operation can signal.
Teleportation & Superdense CodingDerive both protocols line by line, build them as circuits, and see what noisy entanglement does to them.
Quantum Key Distribution: BB84 & E91Analyze eavesdropping, and distill keys from entanglement and Bell tests.
Quantum Information TheoryVon Neumann entropy, quantum channels, the Holevo bound and channel capacities.
At the Rydberg State level these branch further: optimal approximate cloning, quantum money, gate teleportation, quantum repeaters, QKD security proofs and device-independent QKD.
Think first Teleportation needs a Bell pair shared between the two ends. Fiber absorbs photons exponentially with distance, and no-cloning forbids amplifying them. So how could you ever share entanglement across a continent?

Teleport the entanglement itself. Create Bell pairs over short links, then measure at intermediate stations to stitch them into one long pair (entanglement swapping), purifying away errors as you go. That is a quantum repeater, one of the Rydberg-level deep dives in this topic.

Key takeaways

  • No-cloning: no device copies unknown quantum states. A CNOT "copier" turns a superposition into entanglement instead.
  • No-signalling: entanglement gives correlations, not messages. A perfect cloner would break this, which is one reason it cannot exist.
  • Teleportation: 1 Bell pair + 2 classical bits move 1 qubit. The original is destroyed, and nothing travels faster than the bits.
  • Superdense coding: 1 Bell pair + 1 transmitted qubit deliver 2 classical bits.
  • BB84: eavesdropping disturbs the photons. A full intercept-resend attack causes about 25% errors in the sifted key.

Check yourself

In the interactive lesson these are scored with your confidence, and what you miss comes back for review.

Question 1

A CNOT receives ∣+⟩\ket{+} on the control and a blank ∣0⟩\ket{0} on the target. What is the output state?

  1. ∣+⟩∣+⟩\ket{+}\ket{+}: two copies
  2. ∣+⟩∣0⟩\ket{+}\ket{0}: unchanged
  3. 12(∣00⟩+∣11⟩)\tfrac{1}{\sqrt2}(\ket{00}+\ket{11}): an entangled Bell state
  4. A random choice between ∣00⟩\ket{00} and ∣11⟩\ket{11}
Show the answer

C. 12(∣00⟩+∣11⟩)\tfrac{1}{\sqrt2}(\ket{00}+\ket{11}): an entangled Bell state

By linearity, 12(∣0⟩+∣1⟩)∣0⟩→12(∣00⟩+∣11⟩)\tfrac{1}{\sqrt2}(\ket{0}+\ket{1})\ket{0}\to\tfrac{1}{\sqrt2}(\ket{00}+\ket{11}). That is one entangled state, not two copies of ∣+⟩\ket{+}, which would also contain ∣01⟩\ket{01} and ∣10⟩\ket{10}.

Question 2

During teleportation, Alice's two classical bits get lost in transit. What does Bob have?

  1. The state ∣ψ⟩\ket{\psi}: the quantum part already arrived
  2. One of four scrambled versions of ∣ψ⟩\ket{\psi}, and no way to tell which: effectively a random qubit
  3. Nothing at all: the Bell pair was destroyed
  4. A copy of ∣ψ⟩\ket{\psi}, while Alice keeps the original
Show the answer

B. One of four scrambled versions of ∣ψ⟩\ket{\psi}, and no way to tell which: effectively a random qubit

Bob's qubit is ∣ψ⟩\ket{\psi}, X∣ψ⟩X\ket{\psi}, Z∣ψ⟩Z\ket{\psi} or XZ∣ψ⟩XZ\ket{\psi} with equal probability. The bits tell him which correction to apply; without them, his qubit (averaged over the four cases) is completely random.

Question 3

In superdense coding, how many qubits does Alice send during the protocol, and how many classical bits does Bob learn?

  1. 1 qubit sent, 2 bits learned
  2. 1 qubit sent, 1 bit learned
  3. 2 qubits sent, 2 bits learned
  4. 1 qubit sent, any number of bits if Alice picks her gate precisely enough
Show the answer

A. 1 qubit sent, 2 bits learned

Alice sends only her half of the pre-shared Bell pair: one qubit. Bob's Bell measurement reveals which of four states she made: two bits. The other half was distributed earlier, which is why Holevo's one-bit-per-qubit limit is not violated.

Question 4

In BB84 with no eavesdropper and perfect equipment, about what fraction of Alice's photons end up in the sifted key?

  1. All of them
  2. About a quarter
  3. About three quarters
  4. About half
Show the answer

D. About half

Bob's randomly chosen basis matches Alice's with probability ½, and only those rounds are kept. (A further sample is then sacrificed to estimate the error rate.)

Question 5

Why can't Eve simply copy each BB84 photon, pass the original on to Bob, and measure her copy after the bases are announced?

  1. She could, but copying would slow the photons down and be noticed
  2. Copying is possible, but her copy would always come out in the wrong basis
  3. No device can copy unknown states drawn from a set containing non-orthogonal pairs such as ∣0⟩\ket{0} and ∣+⟩\ket{+}
  4. She can, which is why BB84 needs an extra layer of classical encryption
Show the answer

C. No device can copy unknown states drawn from a set containing non-orthogonal pairs such as ∣0⟩\ket{0} and ∣+⟩\ket{+}

A copier that worked on ∣0⟩\ket{0}, ∣1⟩\ket{1}, ∣+⟩\ket{+} and ∣−⟩\ket{-} would have to copy non-orthogonal states, which the no-cloning theorem forbids. Any attempt to extract information disturbs the photons, and the disturbance shows up as errors in the sifted key.

Go deeper

Further reading

  • M. A. Nielsen & I. L. Chuang, Quantum Computation and Quantum Information, §1.3.5–1.3.7, §2.3 and §12.6 (Cambridge, 10th anniv. ed., 2010).
  • C. H. Bennett & G. Brassard, "Quantum cryptography: Public key distribution and coin tossing", Proceedings of the IEEE International Conference on Computers, Systems and Signal Processing (1984).