106
Part Three · A · Foundations of quantum information

Entanglement theory: measures, distillation, monotones

Entanglement behaves like money. Two distant labs that can only do local operations and talk by phone can spend it, to teleport a qubit, or move it around, or concentrate many weakly entangled pairs into a few strong ones. What they can never do is create it from nothing. That single rule, entanglement cannot increase under local operations and classical communication, turns entanglement from a curiosity into a measurable resource, with an exchange rate.

On the plate: a singlet mixed with noise, and how much entanglement survives.

LOCC

Local operations and classical communication: each party may do any quantum operation on its own system and send ordinary messages, as many rounds as it likes. Anything achievable this way creates only separable states (Chapter 50). So a good entanglement measure must never increase under LOCC, on average: it must be an entanglement monotone. For pure states, the entropy of either half (Chapter 49) is the unique sensible measure, in ebits. Established

Detecting and measuring it

For two qubits, three tools: the PPT test (a state is entangled exactly when its partial transpose has a negative eigenvalue; true for 2 × 2 and 2 × 3 systems, not larger); the negativity, the size of that negative part; and the concurrence, from William Wootters’ 1998 formula, which gives the entanglement of formation exactly. Mix a singlet with white noise, ρ=p∣Ψ−⟩⟨Ψ−∣+(1−p)1/4\rho = p|\Psi^-\rangle\langle\Psi^-| + (1 - p)\mathbb 1/4, and set p=p = 0.60. Concurrence is max⁡(0,(3p−1)/2)\max(0, (3p - 1)/2): zero below p=1/3p = 1/3, as Chapter 50 found.

Distillation

Given many copies of a weakly entangled state, can LOCC make fewer, better ones? Yes. Charles Bennett and colleagues’ 1996 recurrence protocol takes two pairs of fidelity FF with the singlet, applies CNOTs and measures one pair; when the results agree, the kept pair has higher fidelity. Start at 0.75 and iterate: 0.79, 0.83, 0.86… toward 1, at the cost of discarding at least half the pairs each round. It works whenever F>1/2F > 1/2. Quantum repeaters (Chapter 153) depend on it.

Strangely, some entangled states cannot be distilled at all. The Horodeckis found such bound entangled states in 1998, in higher dimensions: entanglement that exists but cannot be extracted into pure pairs. Established

Which conversions are possible

For pure states, Michael Nielsen proved in 1999 an exact criterion: ∣ψ⟩|\psi\rangle can be converted to ∣ϕ⟩|\phi\rangle by LOCC with certainty if and only if the Schmidt coefficients of ∣ψ⟩|\psi\rangle are majorized by those of ∣ϕ⟩|\phi\rangle: the more spread out the coefficients, the more entanglement, and you can only go from more spread to less. Entanglement theory, like thermodynamics, is about which transformations are allowed.

What you now have

LOCC and entanglement monotones, the PPT criterion, negativity and concurrence, the Werner state and its threshold, distillation by the recurrence protocol, bound entanglement, and Nielsen’s majorization theorem.

· · ·

Three questions to carry with you.

Why can’t LOCC create entanglement?tap
Local operations and classical messages can only produce mixtures of product states.
When does the PPT test fully decide entanglement?tap
For 2 × 2 and 2 × 3 systems.
What is bound entanglement?tap
Entanglement that no LOCC protocol can distill into pure pairs.

And one to work out. What is the concurrence of the Werner state with p=0.6p = 0.6?

C =

Sources: R. Horodecki et al., Rev. Mod. Phys. 81, 865 (2009). M. B. Plenio and S. Virmani, Quantum Inf. Comput. 7, 1 (2007). W. K. Wootters, Phys. Rev. Lett. 80, 2245 (1998). C. H. Bennett et al., Phys. Rev. Lett. 76, 722 (1996). M. Horodecki, P. Horodecki and R. Horodecki, Phys. Rev. Lett. 80, 5239 (1998). M. A. Nielsen, Phys. Rev. Lett. 83, 436 (1999). Checked by code: Werner concurrence (3p−1)/2 and negativity (3p−1)/4; recurrence from F = 0.75: 0.789, 0.827, 0.864, 0.896, 0.923.

Next · Chapter 107
Nonlocality and contextuality as resources

Contextuality may be the "magic" that makes quantum computers powerful.

Turn the page