How entangled is a state? For pure states there is one right answer, the entanglement entropy. For mixed states there are many inequivalent answers, each with its own operational meaning, and only a few of them, such as concurrence and negativity, can actually be computed.
What you’ll learn
State the axioms of an entanglement measure and explain why LOCC monotonicity is the central one.
Compute the entanglement entropy and concurrence of pure states, and apply Wootters' formula to two-qubit mixed states.
Define negativity and logarithmic negativity via the partial transpose, and compute them for Werner states.
Compare entanglement of formation, entanglement cost, distillable entanglement, relative entropy of entanglement and squashed entanglement.
What a measure must do
An entanglement measure is a function E on bipartite states such that:
E(ρ)=0 whenever ρ is separable;
LOCC monotonicity: E does not increase under LOCC, even on average over the outcomes of a protocol, ∑kpkE(ρk)≤E(ρ);
E is invariant under local unitaries (implied by 2, since they are reversible LOCC);
normalization: E(Φd)=log2d for a maximally entangled state of local dimension d.
Convexity, additivity and continuity are often added. No single function has every desirable property, which is why there is a zoo.
Pure states: entropy and concurrence
E(ψ)=S(ρA)=−i∑λi2log2λi2∈[0,log2d]
The entanglement entropy in ebits, from the Schmidt coefficients λi: the von Neumann entropy of either marginal. The Rényi versions Sα=1−α1log2∑iλi2α are Schur-concave in the Schmidt probabilities, hence also non-increasing under LOCC on pure states.
DerivationConcurrence of a pure two-qubit state
(Y⊗Y)∣ψ∗⟩=−δ∣00⟩+γ∣01⟩+β∣10⟩−α∣11⟩
Define the spin flip ∣ψ~⟩=(Y⊗Y)∣ψ∗⟩ (complex conjugation in the computational basis) and C(ψ)=⟨ψ∣ψ~⟩. For ∣ψ⟩=α∣00⟩+β∣01⟩+γ∣10⟩+δ∣11⟩:
⟨ψ∣ψ~⟩=−2(αδ−βγ)⟹C(ψ)=2∣αδ−βγ∣=2∣detC∣
Take the overlap with ∣ψ⟩: the spin flip turns the determinant test into a number.
C=2λ1λ2=2detρA,E=h(21+1−C2)
In Schmidt form ∣detC∣=λ1λ2. Since E is a monotonic function of C (h is the binary entropy), the two carry the same information for pure two-qubit states.
Mixed states: the convex roof and Wootters' formula
For mixed states S(ρA) fails as a measure: 2I⊗2I is a product state with S(ρA)=1. The entanglement of formation extends a pure-state measure by the cheapest decomposition, a "convex roof":
EF(ρ)=min{k∑pkE(ψk):k∑pk∣ψk⟩⟨ψk∣=ρ}.
Minimizing over all decompositions is generally intractable, which makes the next theorem remarkable.
Let ρp=p∣Ψ−⟩⟨Ψ−∣+(1−p)I/4. It is diagonal in the Bell basis:
C(ρp)=max(0,21+3p−1)=max(0,23p−1)
Bell states are real eigenvectors of Y⊗Y, so ρ~=ρ for any Bell-diagonal state; the μi are the Bell weights and C=max(0,2wmax−1):
ρpentangled⟺p>31,CHSH violated⟺p>21
So ρp is entangled exactly when p>31, yet it violates the CHSH inequality only for p>21: entanglement and Bell nonlocality are different resources.
Predict before you look
What is the concurrence of the Werner state with p=0.6?
Show the answer
0.4
C=23p−1=21.8−1=0.4. Its entanglement of formation is h(21+0.84)≈h(0.958)≈0.25 ebit.
Negativity: entanglement you can compute
The partial transpose transposes only B's indices: (∣i⟩⟨k∣⊗∣j⟩⟨l∣)TB=∣i⟩⟨k∣⊗∣l⟩⟨j∣. For a separable state, ρTB=∑kpkρkA⊗(ρkB)T is again a valid state, which gives a powerful test:
N(ρ)=2ρTB1−1=ν<0∑∣ν∣,EN(ρ)=log2ρTB1
Negativity (the total weight of the negative eigenvalues ν of ρTB) and logarithmic negativity. Both are entanglement monotones (Vidal & Werner 2002) and cost one eigenvalue decomposition. EN upper-bounds distillable entanglement, is additive, and is not convex (Plenio 2005).
DerivationNegativity of Werner states and pure states
∣Ψ−⟩⟨Ψ−∣TB=21I−∣Φ+⟩⟨Φ+∣
Partial transposition moves the singlet's coherence ∣01⟩⟨10∣ to ∣00⟩⟨11∣:
N(ρp)=max(0,43p−1),EN(ρp)=log221+3p(p>31)
So ρpTB=p(21I−∣Φ+⟩⟨Φ+∣)+41−pI has eigenvalue 41−3p on ∣Φ+⟩ and 41+p (three times) elsewhere:
EN(ψ)=2log2i∑λi
For pure states, ψTB1=(∑iλi)2, so EN is the Rényi-½ entropy of the Schmidt probabilities. For two qubits 2N=C on pure states (and on Werner states), but in general 2N≤C, often strictly, for mixed states.
Check your understanding
A two-qubit state has negativity N(ρ)=0. What can you conclude?
It is separable
It is pure
It may still be bound entangled
Its concurrence may still be positive
Show the answer
A. It is separable
For 2×2 systems PPT is equivalent to separability (Horodecki 1996), so N=0 means separable, and then C=0 too. Bound entangled PPT states exist only in larger dimensions, starting with 2×4 and 3×3.
The zoo, and what is still open
Measure
Idea
Computable?
On pure states
Distillable entanglement ED
best rate of Bell pairs extractable by LOCC
no formula, even for two-qubit Werner states
S(ρA)
Entanglement cost EC
best rate of Bell pairs needed to prepare ρ; equals regularized EF (Hayden, Horodecki & Terhal 2001)
generally unknown
S(ρA)
Entanglement of formation EF
convex roof of S(ρA)
closed form for two qubits (Wootters)
S(ρA)
Relative entropy of entanglement ER
minσseparableS(ρ∥σ)
hard in general: the separable set has no efficient description
S(ρA)
Squashed entanglement Esq
21infI(A;B∣E) over extensions ρABE
hard: the extension is unbounded
S(ρA)
Log-negativity EN
log2ρTB1
yes, one eigendecomposition
2log2∑iλi≥S(ρA)
General inequalities: ED≤Esq≤EC≤EF and ED≤ER≤EF, as well as ED≤EN. The asymptotic measures all coincide on pure states but separate dramatically on mixed ones: bound entangled states have ED=0<EC, an irreversibility with no pure-state analogue.
Key takeaways
Entanglement measures vanish on separable states and never increase under LOCC; on pure states the asymptotically unique measure is S(ρA).
Pure two-qubit concurrence is 2∣αδ−βγ∣; Wootters' formula C=max(0,μ1−μ2−μ3−μ4) gives EF for every two-qubit state.
Negativity counts the negative eigenvalues of ρTB: efficiently computable and an upper bound on distillable entanglement, but blind to PPT entanglement.
Werner states are entangled for p>31 but violate CHSH only for p>21.
Mixed-state measures form a hierarchy (ED≤ER≤EF, ED≤Esq≤EC) and differ on bound entangled states; NPT bound entanglement remains open.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
Question 1
What is the concurrence of cosθ∣00⟩+sinθ∣11⟩?
sin2θ
cos2θ
∣sin2θ∣
2sinθ
Show the answer
C.∣sin2θ∣
C=2∣αδ−βγ∣=2∣cosθsinθ∣=∣sin2θ∣: 0 at θ=0 and 1 at θ=π/4.
Question 2
The Werner state p∣Ψ−⟩⟨Ψ−∣+(1−p)I/4 is entangled if and only if:
p>0
p>21
p>21
p>31
Show the answer
D.p>31
C=max(0,23p−1), and the partial transpose has the eigenvalue 41−3p: both switch on at p=31.
Question 3
Which measure can be computed efficiently for every bipartite state, in any dimension?
Entanglement of formation
Logarithmic negativity
Distillable entanglement
Squashed entanglement
Show the answer
B. Logarithmic negativity
EN=log2ρTB1 needs a single eigenvalue decomposition. The price: it vanishes on PPT entangled states.
Question 4
Why can't S(ρA) serve as an entanglement measure for mixed states?
It can be positive for product states such as 2I⊗2I
It is not invariant under local unitaries
It can be negative
It is not defined for mixed states
Show the answer
A. It can be positive for product states such as 2I⊗2I
Classical uncertainty also makes ρA mixed. For a pure global state, mixedness of ρA signals entanglement; for a mixed global state it need not.
Further reading
W. K. Wootters, "Entanglement of formation of an arbitrary state of two qubits", Phys. Rev. Lett. 80, 2245 (1998).
C. H. Bennett, D. P. DiVincenzo, J. A. Smolin & W. K. Wootters, "Mixed-state entanglement and quantum error correction", Phys. Rev. A 54, 3824 (1996).
G. Vidal & R. F. Werner, "Computable measure of entanglement", Phys. Rev. A 65, 032314 (2002).
M. B. Plenio & S. Virmani, "An introduction to entanglement measures", Quantum Inf. Comput. 7, 1 (2007).
R. Horodecki, P. Horodecki, M. Horodecki & K. Horodecki, "Quantum entanglement", Rev. Mod. Phys. 81, 865 (2009).