No-Cloning Theorem
The no-cloning theorem is a law of quantum physics that says you cannot build a device that makes an identical copy of an arbitrary unknown quantum state. If someone hands you a qubit whose state you don’t already know, there is no operation, measurement, or trick that produces a second qubit in the same state while leaving the first untouched. Copying is something classical information takes for granted, and it turns out to be impossible for unknown quantum information. That single limit has two consequences a security professional actually cares about: it’s the physics that lets QKD catch an eavesdropper, because an interceptor can’t quietly duplicate the photons and pass the originals along, and it’s the reason quantum error correction can’t protect data the easy classical way, by copying a bit and taking a majority vote.
The short version:
- The no-cloning theorem forbids copying an unknown quantum state. You can copy a state you already know how to prepare; you can’t copy one handed to you blind.
- It follows from a basic property of quantum mechanics (linearity), so it’s a hard rule of the physics, not a limit of today’s equipment.
- It’s the reason QKD works: an eavesdropper can’t clone the photons carrying the key to read a copy while forwarding the originals undisturbed, so tampering leaves a trace.
- It’s a reason quantum error correction is hard: the classical “store three copies and vote” trick is off the table, so error correction had to be invented from a completely different starting point.
- It does not trap quantum information forever. You can move an unknown state from one place to another by quantum teleportation, but the move destroys the original, which keeps no-cloning intact.
An everyday way to picture it
Imagine a document printed in ink that scrambles itself the instant any scanner’s light touches it. If you already know what the document says, you can print as many fresh copies as you like from the original text. But if a stranger hands you one and asks you to reproduce it without being told its contents, you’re stuck: the only way to read it is to shine light on it, and the moment you do, the pattern rearranges and the information you were trying to capture is gone. You can’t photocopy what you’re only allowed to glance at once, and glancing wrecks it. An unknown quantum state is that self-erasing document. Reading it to copy it changes it, so a faithful duplicate of an unknown state can never be made.
What is the no-cloning theorem?
The no-cloning theorem is the proof, first published in 1982, that no physical process can duplicate an arbitrary unknown quantum state. Wootters and Zurek stated it in the title of their paper as plainly as it can be put: “A single quantum cannot be cloned.”
Source: William K. Wootters and Wojciech H. Zurek, “A single quantum cannot be cloned,” Nature 299, 802-803, 1982, nature.com.
The result was reached independently the same year by Dennis Dieks, working from the question of whether quantum entanglement could be used to signal faster than light. The answer, no, rests directly on the impossibility of copying, which is one reason the theorem sits so deep in the structure of the physics.
Source: Dennis Dieks, “Communication by EPR devices,” Physics Letters A 92, 271-272, 1982, sciencedirect.com.
The precise claim matters, because the theorem is narrower than “you can never copy quantum information,” and the boundary is where all the useful consequences live:
- An unknown state can’t be cloned. If you’re handed a qubit and told nothing about how it was prepared, there is no operation that yields two qubits both in that state.
- A known state can be copied freely. If you know the recipe that made the state, you can run that recipe again as many times as you want and produce identical copies. The prohibition covers duplicating a state whose description you lack, while preparing states you already know how to make stays perfectly allowed.
- Measuring doesn’t rescue you. The obvious workaround, measure the unknown qubit to learn its state and then re-create it, fails, because measurement collapses the state to a single value and destroys the blend you were trying to read. You learn one bit, not the full state, and the original is gone.
So the theorem is really a statement about information not already in your possession, and quantum mechanics offers no way to extract a perfect copy of it.
Why can’t you copy a qubit?
You can’t copy an unknown qubit because quantum mechanics is linear, and a machine that copied every state would have to break that linearity. This is the heart of the standard proof, and it needs no equations to follow the logic.
Source: Michael A. Nielsen and Isaac L. Chuang, “Quantum Computation and Quantum Information,” 10th Anniversary Edition, Cambridge University Press, 2010, Box 12.1, “The no-cloning theorem.”
The reasoning runs in three moves:
- Suppose a universal copier existed. Assume some device takes any input qubit and a blank second qubit, and reliably leaves both in the input’s state. Suppose it works for the pure state 0 and also for the pure state 1, copying each perfectly.
- Feed it a superposition. Now hand that same device a qubit in a superposition, a blend of 0 and 1 at once. Because quantum operations are linear, the device has to act on the blend as the sum of how it acts on 0 and on 1 separately. That produces an entangled two-qubit state, one where the two qubits are correlated with each other.
- Compare to what a real copy would be. A genuine copy of the superposition would be two independent qubits, each carrying the full blend on its own. The entangled result from step 2 is a different state entirely. The two don’t match, so the assumed universal copier can’t exist.
The contradiction is unavoidable as long as quantum evolution stays linear, and linearity is one of the most thoroughly confirmed facts in all of physics. That’s why no-cloning is treated as a foundational constraint rather than an engineering hurdle: better hardware doesn’t get around it, because there’s nothing wrong with the hardware. The rule is baked into the theory itself.
A useful way to hold the idea: copying and reading are really the same forbidden move. To copy an unknown state you’d have to fully characterize it, and to fully characterize it you’d have to measure it, and measuring collapses it. The no-cloning theorem and the collapse of the wave function are two faces of one restriction, that you can’t learn everything about a quantum state without disturbing it.
What can you actually do with an unknown quantum state?
The theorem forbids duplication, but plenty of operations on unknown states are allowed, and telling the two apart is what keeps the concept from being mis-stated. Here’s the boundary as a table, contrasting a classical bit with an unknown qubit.
| Operation | Classical bit | Unknown qubit | Why |
|---|---|---|---|
| Read it without disturbing it | Allowed, freely | Forbidden | Reading a qubit is measurement, which collapses the state |
| Make an exact copy of it | Allowed, freely | Forbidden | The no-cloning theorem (Wootters and Zurek, 1982) |
| Copy a state you already know how to prepare | Allowed | Allowed | Re-run the known preparation recipe as many times as you like |
| Move it to another location | Allowed by copying | Allowed by teleportation, but the original is destroyed | Teleportation transfers the state without cloning, consistent with the theorem |
| Apply reversible operations to it | Allowed | Allowed | Quantum gates transform a state without needing to copy or read it |
Sources: William K. Wootters and Wojciech H. Zurek, “A single quantum cannot be cloned,” Nature 299, 802-803, 1982, nature.com. Michael A. Nielsen and Isaac L. Chuang, “Quantum Computation and Quantum Information,” 10th Anniversary Edition, Cambridge University Press, 2010, teleportation in Section 1.3.7.
The teleportation row is the one people trip over. Quantum teleportation genuinely moves an unknown state from one qubit to another, but it consumes the source: at the end, the original qubit is left blank and the state survives only at the destination, so there’s still only one copy in existence at any moment. No-cloning is preserved exactly because teleportation destroys as it transfers. Quantum information can travel; it just can’t be duplicated.
Why does the no-cloning theorem matter for cryptography?
It matters because it’s the physical guarantee behind QKD’s eavesdropper detection, and it’s a core reason quantum error correction is difficult. The same rule that helps one quantum security technology stands in the way of another, and both are worth spelling out.
On the defensive side, QKD encodes secret key bits into single photons sent over a dedicated optical link. The reason a passive wiretap can’t work is no-cloning. Tapping an ordinary fiber undetectably means splitting off a copy of the signal and letting the original pass through untouched, and against single photons carrying unknown quantum states, that copy can’t be made. An interceptor is forced to actually measure the photons, and measuring in the wrong basis disturbs them, injecting errors the two legitimate parties see when they compare a sample of their results. So the eavesdropper faces a wall built from two rules at once: they can’t clone the photons to measure a duplicate, and they can’t measure the originals without leaving fingerprints.
Source: NSA Cybersecurity, “Quantum Key Distribution (QKD) and Quantum Cryptography (QC)“.
On the engineering side, no-cloning is a big part of why building a fault-tolerant quantum computer is so hard, which loops back to the cryptographic threat. Classical computers fight noise by copying: store a bit three times, and if one copy flips, the other two outvote it. That entire strategy is forbidden for quantum information, because you can’t make the copies to vote on.
Source: Michael A. Nielsen and Isaac L. Chuang, “Quantum Computation and Quantum Information,” 10th Anniversary Edition, Cambridge University Press, 2010, Chapter 10, quantum error correction.
Quantum error correction had to be reinvented from scratch to route around the prohibition. Instead of copying one qubit’s state into others, it spreads a single logical state across a block of entangled physical qubits and finds errors through indirect “syndrome” measurements that reveal where a fault occurred without ever reading the protected data. That workaround is expensive: each reliable logical qubit costs on the order of a thousand physical ones, which is exactly why a machine big enough to run Shor’s algorithm against RSA sits in the millions-of-qubits range that no hardware is close to. So no-cloning, by making error correction costly, is one of the quiet reasons a code-breaking quantum computer doesn’t exist yet.
Common misconceptions
- “No-cloning means quantum information can never be copied at all.” The prohibition is only on unknown states. A state you know how to prepare can be reproduced as many times as you want by re-running its preparation.
- “You could just measure the qubit and rebuild it from the reading.” Measurement collapses the state to a single value, so you recover one bit of a much richer state and destroy the original in the process. That’s not a copy.
- “No-cloning means you can’t move quantum data from place to place.” You can, by teleportation. The catch is that teleportation destroys the source as it transfers the state, so only one copy ever exists, which is exactly what keeps the theorem true.
- “It’s a limitation of current technology that better hardware will fix.” No-cloning follows from the linearity of quantum mechanics, one of the most tested facts in physics. It’s a rule of the theory, so no future device evades it.
- “No-cloning makes QKD unbreakable in practice.” It makes the idealized protocol secure. Real QKD hardware has been attacked through side-channels in the detectors and optics, so security depends on the engineering of the device, not the theorem alone.
- “The theorem is just about photons or just about qubits.” It applies to any unknown quantum state of any system. Photons and qubits are where it shows up in cryptography, but the proof is completely general.
Questions people ask
Do I need physics to understand the no-cloning theorem? No. One sentence covers what you need: you can’t make an identical copy of a quantum state you don’t already know, because reading it to copy it changes it. That’s the whole idea, and it’s enough to see why an eavesdropper on a quantum channel gets caught and why quantum error correction can’t just copy and vote.
Why can’t you copy a qubit if you can copy any file on your computer? A file is classical information, stored as robust voltages you can read and duplicate at will without disturbing them. A qubit in an unknown superposition can’t be read without collapsing it, and the no-cloning theorem proves no operation duplicates it without reading it. Classical copying works precisely because classical bits, unlike quantum states, survive being looked at.
How does no-cloning let QKD detect an eavesdropper? An eavesdropper can’t split off a perfect copy of the single photons carrying the key, so to learn anything they have to measure the originals, and measuring disturbs them and injects detectable errors (NSA). The two legitimate parties compare a sample of their results, see the elevated error rate, and discard the key. Without no-cloning, a passive undetectable tap would be possible and QKD’s whole guarantee would collapse.
Does the no-cloning theorem stop quantum computers from working? No, it shapes how they have to be built. Quantum algorithms operate on states with reversible gates and never need to copy them, so the computation itself is fine. Where no-cloning bites is error correction, which can’t use the classical copy-and-vote method and had to be redesigned around entanglement and syndrome measurement instead.
Can quantum states be moved without copying them? Yes, through quantum teleportation, which transfers an unknown state from one qubit to another. The process destroys the source qubit’s state as it recreates it at the destination, so at no point do two copies exist (Nielsen and Chuang, 2010). That’s why teleportation and no-cloning coexist without contradiction.
Who proved the no-cloning theorem, and when? Wootters and Zurek published it in Nature in 1982 under the title “A single quantum cannot be cloned” (nature.com), and Dieks reached the same result independently that year while analyzing whether entanglement could send faster-than-light signals (Physics Letters A 92, 271). The proof rests on the linearity of quantum mechanics.
Does no-cloning have anything to do with post-quantum cryptography? Not directly. PQC is math-based software whose security rests on hard computational problems, not on quantum physics, so it doesn’t lean on no-cloning the way QKD does. The theorem’s relevance to the migration is indirect: by making error correction expensive, it’s one of the reasons the quantum computer that would force the migration is still years away.
Everything here is the map, given freely. When your team needs the quantum threat translated into a cryptographic inventory and a dated plan for your own estate, that’s the work I do. Request an alignment briefing.
Last verified 2026-07-09 · Maintained by Addie LaMarr, LaMarr Labs.