| Peres–Horodecki criterion | |
|---|---|
| Name | Peres–Horodecki criterion |
| Field | Quantum information theory |
| Introduced | 1996 |
| Discoverer | Asher Peres; Michał Horodecki, Paweł Horodecki, Ryszard Horodecki |
| Notable for | separability test via partial transpose |
Peres–Horodecki criterion
The Peres–Horodecki criterion is a practical test for quantum entanglement in bipartite mixed states based on the action of the partial transpose on a density operator. It provides a simple necessary condition for separability and, in low-dimensional systems, a complete characterization: a state is separable iff its partial transpose is positive semidefinite. The criterion plays a central role in quantum information theory and experimental entanglement detection, connecting linear algebraic operations on density matrices to physical properties of composite systems.
The criterion was first proposed by Asher Peres (1996) and later proved to be sufficient in certain cases by the Horodecki family (Michał Horodecki, Paweł Horodecki, Ryszard Horodecki). It addresses the fundamental problem of deciding whether a given density matrix ρ on a composite Hilbert space H_A ⊗ H_B is a separable mixture or an entangled resource for tasks such as quantum teleportation, entanglement distillation, and quantum cryptography. Because separability is an NP-hard problem in general, the Peres–Horodecki criterion supplies an efficiently computable algebraic check that is routinely used in both theoretical analyses and laboratory characterization of entanglement, for example in experiments at institutions like IBM Quantum and Google Quantum AI.
Let ρ be a density operator on the finite-dimensional composite Hilbert space H_A ⊗ H_B with orthonormal bases {|i>_A} and {|j>_B}. The partial transpose with respect to subsystem B, denoted ρ^{T_B}, is defined by transposing only the matrix indices corresponding to B: ρ = ∑_{ijkl} ρ_{ij,kl} |i>_A⟨k| ⊗ |j>_B⟨l| → ρ^{T_B} = ∑_{ijkl} ρ_{ij,kl} |i>_A⟨k| ⊗ |l>_B⟨j|. This operation is linear and basis-dependent but its spectrum and positivity properties are basis-independent. Positivity of ρ^{T_B} (positive partial transpose, or PPT) is the hallmark of the criterion. The partial transpose is closely related to the transpose map and to positive but not completely positive maps, a concept appearing in the Choi–Jamiolkowski isomorphism and studies of map extendibility.
Peres observed that if ρ is separable, i.e. ρ = ∑_k p_k ρ_A^{(k)} ⊗ ρ_B^{(k)} with p_k ≥ 0, then ρ^{T_B} = ∑_k p_k ρ_A^{(k)} ⊗ (ρ_B^{(k)})^T remains positive semidefinite, since transposition preserves positivity. Hence PPT is a necessary condition for separability. The Horodeckis proved sufficiency for 2×2 and 2×3 systems by constructing a decomposition for any PPT state in these dimensions using structural results on positive maps: every positive map on M_2 or M_3 is decomposable into completely positive and completely co-positive parts. The proof leverages the correspondence between positive maps and entanglement witnesses via the Hahn–Banach theorem and the Choi matrix construction, reducing separability to the absence of a positive map that detects entanglement.
In practice, the criterion is applied to bipartite density matrices arising in quantum tomography of systems such as pairs of qubits, qubit–qutrit ensembles, and photonic or ionic implementations. For 2×2 (two-qubit) and 2×3 (qubit–qutrit) systems the PPT test is both necessary and sufficient: any PPT state is separable and any non-PPT state is entangled and distillable. In higher dimensions, PPT violation witnesses free (distillable) entanglement and is used alongside numerical methods—such as semidefinite programming (SDP), convex optimization, and entanglement witnesses—to certify entanglement in quantum computing platforms and quantum optics experiments at facilities like the Max Planck Institute for Quantum Optics or university labs.
The Peres–Horodecki criterion is not sufficient in general: there exist PPT states that are entangled but not distillable, termed bound entanglement. The Horodeckis provided explicit examples of bound entangled states in dimensions ≥ 3×3 using unextendible product bases (UPB) and edge-state constructions. These limitations prompted development of stronger separability criteria and detection tools such as realignment criteria, higher-order positive maps (e.g., reduction map, Breuer–Hall map), and entropic or correlation-based tests. Bound entanglement has implications for tasks like activation of nonlocality and shows subtleties in the resource theory of entanglement.
The PPT test relates to a family of separability criteria including entanglement witnesses, the Peres criterion (original proposal), the realignment criterion (computable cross-norm), and criteria based on positive maps like the Breuer map or the Choi map. Quantitative entanglement measures such as negativity and logarithmic negativity are directly derived from the spectrum of the partial transpose and provide computable upper bounds on distillable entanglement. Connections also exist with the Entanglement of formation and Concurrence in low dimensions where closed formulas are available. The interplay between PPT, positive maps, and resource quantifiers remains an active area in quantum information theory, with ongoing research linking these concepts to quantum channel capacities, LOCC convertibility, and operational tasks in noisy intermediate-scale quantum devices.