The best imperfect copier and the 5/6 fidelity limit.
No-cloning says perfect copies are impossible. It does not say how close you can get. For qubits the answer is a crisp number, 5/6, and three independent routes reach it: an explicit machine, symmetry and representation theory, and the impossibility of signalling.
What you’ll learn
Formulate cloning as an optimization over channels, with single-copy and global fidelities, and justify restricting to covariant cloners.
Derive the Bužek–Hillery cloner's output, its shrinking factor η=2/3 and its fidelity 5/6.
Prove that 5/6 is optimal for symmetric universal 1→2 qubit cloning from positivity and no-signalling.
Apply the general N→M formula in dimension d, and connect its M→∞ limit to optimal state estimation.
Derive the optimal asymmetric trade-off and relate cloning to eavesdropping in QKD.
Cloning as an optimization
A 1→2 cloner is a channel (a CPTP map) T:L(Cd)→L(Cd⊗Cd). For a pure input ψ=∣ψ⟩⟨ψ∣ write ρAB=T(ψ), with marginals ρA=TrBρAB and ρB=TrAρAB. Two figures of merit:
FA(ψ)=⟨ψ∣ρA∣ψ⟩,Fglob(ψ)=⟨ψψ∣ρAB∣ψψ⟩.
A cloner is universal if these do not depend on ψ, and symmetric if FA=FB.
For qubits, covariance pins the marginals down completely. A qubit channel that commutes with every rotation can only rescale the Bloch vector, so each copy is a depolarized version of the input:
ρA=ηAψ+(1−ηA)2I,FA=21+ηA.
The Bloch vector of each copy is the input's, multiplied by the shrinking factorηA.
The Bužek–Hillery machine
Bužek and Hillery (1996) wrote down a unitary acting on the input a, a blank b in ∣0⟩, and one machine qubit x:
Here ∣Ψ+⟩=(∣01⟩+∣10⟩)/2, and ∣↑⟩,∣↓⟩ are orthonormal machine states. The two right-hand sides are orthonormal, so the map extends to a unitary on all eight basis states.
DerivationReading off the copies
ρab(0)=32∣00⟩⟨00∣+31∣Ψ+⟩⟨Ψ+∣
Trace out the machine. Its two states are orthogonal, so the branches add incoherently.
ρa(0)=32∣0⟩⟨0∣+31⋅2I=65∣0⟩⟨0∣+61∣1⟩⟨1∣
Use Trb∣Ψ+⟩⟨Ψ+∣=I/2. By symmetry ρb(0) is the same.
ρa(ψ)=32ψ+31⋅2Ifor every pure ψ
The machine is covariant, so the same shrinking factor η=32 holds for every input. (Check one case in the reveal below.)
F=21+η=65,Fglob=⟨ψψ∣ρab∣ψψ⟩=32
Each copy has fidelity 5/6, and a joint test of "both outputs equal ψ" passes with probability 2/3. The copies are correlated: Fglob=FAFB.
Think first Check universality on an input that is not a basis state: compute ⟨+∣ρa∣+⟩ for the input ∣+⟩.
By linearity the output is 21[(32∣00⟩+31∣Ψ+⟩)∣↑⟩+(31∣Ψ+⟩+32∣11⟩)∣↓⟩]. Project qubit a onto ∣+⟩ in each machine branch. The ∣↑⟩ branch leaves 21(63∣0⟩b+61∣1⟩b), of squared norm 125, and the ∣↓⟩ branch contributes another 125. Total: ⟨+∣ρa∣+⟩=65.
The general two-qubit output for an input with Bloch vector m and equal depolarized marginals. Without loss of generality the cloner is covariant (twirling) and swap-symmetric (randomly swapping the outputs preserves both fidelities when they are equal).
tjk(m)=tδjk+t′mjmk
Covariance makes tjk a rotation-covariant tensor built from m alone; swap symmetry makes it symmetric, which excludes ϵjklml.
No-signalling: let Bob's input be half of a singlet. By measuring along any axis n she likes, Alice steers it to ±n with probability ½ each. Bob's averaged output must not depend on her choice, so t′=0. (For a map already known to be linear this is automatic; the point is that no-signalling alone forces it.)
specρab(z^)=41{1+2η+t,1−2η+t,1+t,1−3t}
With t′=0 and m=z^ the operator is diagonal in {∣00⟩,∣11⟩,∣Ψ+⟩,∣Ψ−⟩}. Use ∑jσj⊗σj=2F−I, which is +1 on the triplet and −3 on the singlet.
1−3t≥0 and 1−2η+t≥0⟹η≤21+t≤32,F=21+η≤65
Positivity of ρab. Equality forces t=31 and η=32; the spectrum becomes {32,0,31,0}, exactly the Bužek–Hillery output 32∣00⟩⟨00∣+31∣Ψ+⟩⟨Ψ+∣. ■
Check your understanding
A symmetric universal 1→2 qubit cloner has correlation parameter t=0.2. What is the largest shrinking factor that positivity allows?
η=0.6
η=2/3
η=0.4
η=0.8
Show the answer
A.η=0.6
The binding eigenvalue is 1−2η+t≥0, so η≤(1+t)/2=0.6; the other constraint, t≤31, is satisfied. Only t=31 allows the optimum η=32.
Representation theory gives the same answer and much more. Werner (1998) showed that, for global fidelity, the optimal universal N→M cloner is unique:
where SM projects onto the symmetric subspace of (Cd)⊗M and d[K] is its dimension. Its global fidelity is d[N]/d[M]; for d=2, N=1, M=2 that is 2/3, as found above. Keyl and Werner (1999) proved that the same machine also maximizes the single-copy fidelity.
N → M, any dimension, and the estimation limit
For N identical inputs and M≥N outputs in dimension d, the optimal universal single-copy fidelity is (Gisin–Massar 1997 for qubits; Werner 1998 and Keyl–Werner 1999 in general):
FN→M(d)=MN+(1−MN)N+dN+1
For qubits this becomes F=M(N+2)M(N+1)+N, i.e. shrinking factor ηN→M=M(N+2)N(M+2).
N→M
d
Optimal F
1→2
2
5/6≈0.833
1→3
2
7/9≈0.778
2→3
2
11/12≈0.917
1→∞
2
2/3 (state estimation)
1→2
3
3/4
1→2
d
2(d+1)d+3
Predict before you look
What is the optimal universal single-copy fidelity for 1→3 qubit cloning?
Show the answer
0.778
F=M(N+2)M(N+1)+N=3⋅33⋅2+1=97≈0.778, between the 1→2 value 5/6 and the estimation limit 2/3.
Asymmetric cloners and the frontier
Drop the symmetry requirement: how good can copy B be if copy A must have shrinking factor ηA? An elegant construction superposes the two trivial cloners.
DerivationThe optimal asymmetric trade-off
∣Ψ⟩RABC=a∣Φ+⟩RA∣Φ+⟩BC+b∣Φ+⟩RB∣Φ+⟩AC,a,b≥0
Work with the Choi state: a reference R is maximally entangled with the input. In the first branch copy Ais the input and B is half of a fresh pair (junk); in the second the roles swap. C is the machine.
⟨ΦRA+ΦBC+∣ΦRB+ΦAC+⟩=21⟹a2+b2+ab=1
Normalization. One checks TrABC∣Ψ⟩⟨Ψ∣=I/2, so ∣Ψ⟩ is the Choi state of a valid isometry; it is covariant by construction, since each Φ+ is invariant under Uˉ⊗U.
FeA=(⟨Φ+∣RA⊗IBC)∣Ψ⟩2=(a+2b)2
The entanglement fidelity of the channel to A, using (⟨Φ+∣RA⊗I)∣Φ+⟩RB∣Φ+⟩AC=21∣Φ+⟩BC.
ηA=34FeA−1=a(a+b),ηB=b(a+b)
For a depolarizing channel Fe=(1+3η)/4; then use a2+b2+ab=1.
ηA2+ηB2+ηAηB=ηA+ηB
Eliminating a and b gives an ellipse through (1,0), (0,1) and the symmetric optimum (32,32). Equivalently FA=1−2b2 and FB=1−2a2. This family is optimal (Niu–Griffiths 1998; Cerf 2000).
Key takeaways
Optimal cloners can be taken covariant; qubit copies are then depolarized, ρ=ηψ+(1−η)I/2, with F=(1+η)/2.
Bužek–Hillery: η=2/3, F=5/6, Fglob=2/3, equivalently T(ρ)=32S(ρ⊗I)S. Positivity plus no-signalling proves it optimal.
General law: FN→M=MN+(1−MN)N+dN+1, approaching the estimation limit N+dN+1 as M→∞.
Asymmetric cloners trace the ellipse ηA2+ηB2+ηAηB=ηA+ηB: the same trade-off an eavesdropper faces.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
Question 1
The Bužek–Hillery output for input ∣0⟩ is ρab=32∣00⟩⟨00∣+31∣Ψ+⟩⟨Ψ+∣. What is its global fidelity ⟨00∣ρab∣00⟩?
5/6
25/36
2/3
1/3
Show the answer
C.2/3
Only the first term overlaps ∣00⟩, so Fglob=32=d[1]/d[2]. The copies are correlated, so Fglob=FAFB.
Question 2
For qutrits (d=3), what does the optimal universal 1→M single-copy fidelity tend to as M→∞?
2/3
3/4
1/3
1/2
Show the answer
D.1/2
The limit is the estimation fidelity N+dN+1=42=21 for N=1, d=3.
Question 3
In the no-signalling proof of the 5/6 bound, what forces the coefficient t′ of mjmkσj⊗σk to vanish?
Positivity of ρab
Bob's averaged output cannot depend on the axis along which Alice steers his input
Swap symmetry of the two outputs
Normalization, Trρab=1
Show the answer
B. Bob's averaged output cannot depend on the axis along which Alice steers his input
Averaging over Alice's two outcomes leaves the term t′∑jknjnkσj⊗σk, which depends on her chosen axis n. No-signalling demands independence from n, so t′=0.
Question 4
On the optimal asymmetric boundary ηA2+ηB2+ηAηB=ηA+ηB, copy A must have ηA=0.9. What is the best possible ηB, approximately?
ηB≈0.35
ηB≈0.10
ηB≈0.67
ηB≈0.50
Show the answer
A.ηB≈0.35
Solve ηB2−0.1ηB−0.09=0: ηB=20.1+0.37≈0.354, so FB≈0.68. Making one copy nearly perfect leaves the other only slightly better than measure-and-prepare (η=31).
Further reading
V. Bužek & M. Hillery, "Quantum copying: Beyond the no-cloning theorem", Physical Review A54, 1844 (1996).
N. Gisin & S. Massar, "Optimal quantum cloning machines", Physical Review Letters79, 2153 (1997).
R. F. Werner, "Optimal cloning of pure states", Physical Review A58, 1827 (1998).
N. Gisin, "Quantum cloning without signaling", Physics Letters A242, 1 (1998).
V. Scarani, S. Iblisdir, N. Gisin & A. Acín, "Quantum cloning", Reviews of Modern Physics77, 1225 (2005).