| Schmidt decomposition | |
|---|---|
| Name | Schmidt decomposition |
| Field | Quantum mechanics |
| Introduced | 1906 |
| Introduced by | Erhard Schmidt |
| Related | Singular value decomposition, Quantum entanglement |
Schmidt decomposition
The Schmidt decomposition is a canonical representation of pure states in a bipartite quantum system that expresses any vector in a tensor product Hilbert space as a sum of orthogonal product vectors with nonnegative coefficients. It is foundational in Quantum mechanics and Quantum information science because it reveals correlations between subsystems, simplifies analyses of entanglement, and connects to linear algebraic tools such as the Singular value decomposition.
For a composite system with finite-dimensional Hilbert spaces H_A and H_B, any pure state |ψ⟩ ∈ H_A ⊗ H_B admits a Schmidt decomposition |ψ⟩ = ∑_i λ_i |u_i⟩_A ⊗ |v_i⟩_B, where {λ_i} are nonnegative real numbers (the Schmidt coefficients) and {|u_i⟩_A}, {|v_i⟩_B} are orthonormal bases of the respective subsystems. The number of nonzero Schmidt coefficients is the Schmidt rank, an integer-valued invariant under local unitary operations. The decomposition is unique up to degeneracies in the Schmidt coefficients and phase choices for the local vectors. The Schmidt coefficients squared equal the nonzero eigenvalues of the reduced density matrices ρ_A = Tr_B(|ψ⟩⟨ψ|) and ρ_B = Tr_A(|ψ⟩⟨ψ|).
The Schmidt decomposition follows from linear algebra via the Singular value decomposition (SVD) of the coefficient matrix representing |ψ⟩ in product bases. Given bases {|i⟩_A}, {|j⟩_B} and coefficients ψ_{ij}, the SVD yields unitary matrices U, V and nonnegative singular values σ_k such that ψ = U Σ V†, producing Schmidt coefficients λ_k = σ_k. Important properties include: equality of nonzero eigenvalue spectra of ρ_A and ρ_B (purification property), invariance of Schmidt rank under local invertible operations, and additivity relationships under tensor products. For infinite-dimensional systems one requires care with convergence and continuous spectra; the original work by Erhard Schmidt predates quantum applications, while later formal treatments appear in functional analysis texts and in works by John von Neumann and others. The Schmidt decomposition is closely related to the Takagi factorization for symmetric tensors and to Schmidt-mode analysis in continuous-variable systems such as quantum optics.
Schmidt decomposition is used widely across theoretical and experimental domains. In quantum information theory it diagnoses entanglement of pure states, informs protocols like quantum teleportation and entanglement swapping, and underpins resource accounting in quantum communication tasks. In many-body physics and condensed matter, Schmidt spectra of matrix product states and tensor network ansätze quantify correlations and drive algorithms such as the density matrix renormalization group (DMRG) developed by Steven R. White. Experimental implementations in quantum optics, trapped ions (e.g., groups at National Institute of Standards and Technology and IonQ research), superconducting circuits (e.g., IBM Quantum), and ultracold atoms exploit Schmidt-mode decompositions to design state tomography and entanglement verification protocols. The decomposition also aids analyses of decoherence, open quantum systems, and quantum thermodynamics by making subsystem entropies explicit.
The Schmidt coefficients determine entanglement measures for pure states: the von Neumann entropy S(ρ_A) = -∑ λ_i^2 log λ_i^2 (often called entanglement entropy) and the Rényi entropies are functions of {λ_i}. The Schmidt rank provides a coarse discrete indicator of entanglement classes under stochastic local operations and classical communication (SLOCC). In quantum communication, Schmidt-number-based bounds constrain channel capacities and entanglement distillation protocols; these links connect to work on entanglement witnesses, concurrence (for low dimensions), and operational tasks studied at institutions such as Perimeter Institute and QuTech. The decomposition also appears in proofs of majorization relations and in formulations of the Schmidt decomposition theorem used in quantum resource theories.
Numerical construction of Schmidt decompositions reduces to SVD of matrices whose dimensions scale with subsystem Hilbert spaces. Efficient algorithms exploit sparsity, symmetry, and tensor decompositions: iterative SVD, randomized SVD, and tensor network techniques (MPS, PEPS). Software libraries and platforms implement these methods in contexts like DMRG (e.g., ITensor), quantum simulation packages (e.g., QuTiP), and machine-learning-aided state compression. For large-scale many-body systems, truncation of small Schmidt coefficients yields controlled approximations; error bounds relate discarded Schmidt spectrum weight to fidelity loss. Computational challenges include handling continuous-variable systems, degeneracies, and numerical instability in near-product states.
Generalizations include multipartite Schmidt-like decompositions (Schmidt decompositions fail to generalize straightforwardly beyond bipartite pure states), tensor rank and canonical polyadic decompositions for higher-order tensors, and continuous Schmidt modes for infinite-dimensional systems such as frequency-entangled photons. The lack of a unique canonical form for multipartite systems motivates alternative frameworks: Schmidt measure, Schmidt number hierarchies, and entanglement monotones. Limitations arise for mixed states, where the convex roof extension is needed to relate mixed-state entanglement to pure-state decompositions, and for operational interpretations in noisy, open systems. There are social and ethical dimensions: equitable access to quantum technologies leveraging entanglement—developed in labs at CERN, MIT, and publicly funded institutions—requires attention to allocation of resources, inclusion in education, and the social impact of quantum-enabled surveillance or computational advantages.
Category:Quantum mechanics Category:Quantum information theory Category:Linear algebra