| partial transpose | |
|---|---|
| Name | Partial transpose |
| Type | Mathematical operation |
| Field | Quantum information theory |
| Introduced | 1980s–1990s |
| Related | Peres–Horodecki criterion, entanglement measure, negativity (quantum entanglement) |
partial transpose
The partial transpose is a linear algebra operation on composite Hilbert space operators that transposes one subsystem while leaving the other(s) unchanged. It is a central tool in the study of quantum entanglement because it provides a simple, often computable criterion for separability and entanglement detection in bipartite states. Its mathematical simplicity and operational implications have made it influential in quantum information theory and experimental tests of nonclassical correlations.
The partial transpose acts on density operators ρ defined on a tensor product Hilbert space H_A ⊗ H_B. Given fixed orthonormal bases { |i⟩_A } for H_A and { |μ⟩_B } for H_B, the matrix elements of ρ are ρ_{iμ,jν} = ⟨i,μ|ρ|j,ν⟩. The partial transpose with respect to subsystem B, denoted ρ^{T_B}, is defined by swapping indices on B: (ρ^{T_B})_{iμ,jν} = ρ_{iν,jμ}. This operation is basis-dependent in its matrix representation but basis-independent in the sense of similarity under local unitary transformations unitaries on the transposed subsystem. The map is linear and positive on pure-product operators but not completely positive; it corresponds to the tensor product of the identity map on A with the ordinary transpose map on B, id_A ⊗ T_B. The transpose map T is an example of a positive but not completely positive linear map, a concept developed in the theory of operator algebras and used in separability theory by Asher Peres and the Horodeckis.
The partial transpose underpins the Peres–Horodecki criterion (also called the positive partial transpose or PPT criterion). Peres observed that for any separable state ρ_sep = ∑_k p_k ρ_A^k ⊗ ρ_B^k, the partial transpose ρ_sep^{T_B} remains positive semidefinite. Horodecki et al. later proved that PPT is necessary and sufficient for separability in 2×2 and 2×3 systems but only necessary in higher dimensions; states with non-positive partial transpose (NPT) are necessarily entangled. The partial transpose thus provides an operational test linked to structural results in operator theory and has motivated studies of bound entanglement, where states are PPT but still non-separable—a phenomenon studied by the Horodeckis and observed to have implications for quantum cryptography and entanglement distillation protocols.
For two-qubit systems (H_A ≅ H_B ≅ C^2), partial transpose detects all entanglement: a two-qubit density matrix ρ is entangled iff ρ^{T_B} has a negative eigenvalue. Canonical examples include the Bell states and Werner states: the singlet state |Ψ^-⟩ leads to an NPT partial transpose, while certain mixed Werner states transition from NPT to PPT at a critical mixing parameter. In higher dimensions (e.g., C^3 ⊗ C^3), there exist PPT entangled states discovered in constructive examples by the Horodecki family and via unextendible product bases by Bennett et al.. Partial transpose may produce non-physical operators with negative spectrum, and the number and magnitude of negative eigenvalues carry quantitative information relevant to entanglement measures.
In quantum information practice, partial transpose is used for state certification, entanglement verification, and theoretical analyses of quantum channels. In quantum state tomography, reconstructed density matrices are tested via partial transpose to certify entanglement in experiments performed at institutions such as MIT, Caltech, IQOQI Vienna and industrial labs like IBM Quantum and Google Quantum AI. The PPT test is computationally light and often serves as a first-pass diagnostic before applying more sophisticated criteria like semidefinite-programming-based separability tests. In quantum communication and quantum key distribution research, detecting NPT states can validate resources for entanglement-based protocols, while PPT bound-entangled states have been proposed for tasks like private key extraction under restricted operations.
Negativity and logarithmic negativity are entanglement monotones directly derived from the spectrum of the partial transpose. The negativity N(ρ) = (||ρ^{T_B}||_1 − 1)/2 equals the absolute sum of negative eigenvalues and quantifies the degree to which ρ^{T_B} fails to be positive. The logarithmic negativity E_N(ρ) = log_2 ||ρ^{T_B}||_1 provides an upper bound on distillable entanglement and is computable via singular value decomposition or eigenvalue solvers, making it popular in numerical studies of many-body systems, conformal field theories, and condensed-matter models studied at places like Perimeter Institute and Max Planck Institute for Quantum Optics.
Computing partial transpose and derived measures relies on linear algebra libraries and eigenvalue solvers. For moderate system sizes, dense matrix diagonalization using LAPACK or ARPACK implementations suffices; for larger composite systems, tensor-network methods (e.g., matrix product states) and randomized numerical linear algebra speed up evaluations. Semidefinite programming (SDP) appears in relaxations of separability tests beyond PPT; packages such as CVX, SDPA, and SCS are widely used. Research groups at IBM Research and academic groups led by researchers like John Preskill and Fernando Brandão have developed algorithms exploiting symmetries to reduce computational cost and probe entanglement in many-body models.
Physically, the partial transpose lacks a direct local quantum operation interpretation because transpose is anti-unitary; nevertheless, its eigenstructure captures nonclassical correlations with experimental signatures. Experiments in photonic systems, trapped ions, and superconducting qubits have used state tomography followed by partial transpose to report entanglement generation. Understanding PPT versus NPT distinctions also has justice and equity implications in resource allocation: accessible, robust entanglement detection methods like PPT lower the barrier for resource-limited labs and underrepresented institutions to certify quantum resources, informing fairer dissemination of quantum technologies and collaborative research across institutions.
Category:Quantum information theory Category:Quantum entanglement