| reduced density matrix | |
|---|---|
| Name | Reduced density matrix |
| Field | Quantum physics |
| Introduced | Early 20th century |
| Notable use | Quantum information theory, condensed matter physics |
reduced density matrix
A reduced density matrix is the operator describing the state of a subsystem obtained by tracing out degrees of freedom of a larger composite quantum system. It captures local statistics, correlations, and mixedness due to entanglement or classical ignorance, and is central to understanding von Neumann's formulation of quantum statistical mechanics and modern quantum information tasks. Reduced density matrices underpin practical analyses in quantum decoherence, quantum measurement theory, and many-body condensed matter studies.
The reduced density matrix ρ_A for subsystem A of a bipartite system AB is the unique positive semidefinite operator on A that reproduces expectation values of observables acting only on A. Physically, ρ_A encodes all accessible information when degrees of freedom of subsystem B are ignored or inaccessible, as in open systems studied by the Lindblad formalism or when an observer has limited resolution. Early conceptual foundations trace to John von Neumann and later to pragmatic formulations in statistical mechanics and the study of quantum correlations by researchers such as Erwin Schrödinger and Albert Einstein during debates on completeness of quantum description.
Given a composite Hilbert space H = H_A ⊗ H_B and a global density operator ρ_{AB}, the reduced density matrix ρ_A is defined by the partial trace over H_B: ρ_A = Tr_B(ρ_{AB}). It is a trace-class operator with Tr(ρ_A) = 1, Hermiticity ρ_A = ρ_A†, and eigen-decomposition ρ_A = ∑_i p_i |ψ_i⟩⟨ψ_i|. The eigenvalues {p_i} are the occupation probabilities and determine the von Neumann entropy S(ρ_A) = −Tr(ρ_A log ρ_A), a measure used in works by Clauset, Shalizi, Newman-style information analyses and foundational papers by Nielsen and Chuang. The formalism connects to Gibbs states and thermal ensembles via marginalization from global equilibrium states in statistical physics.
The partial trace is a linear map Tr_B : B(H_A ⊗ H_B) → B(H_A) characterized by Tr_B(|a⟩⟨a'| ⊗ |b⟩⟨b'|) = ⟨b'|b⟩ |a⟩⟨a'|. For a pure global state |Ψ⟩ ∈ H_A ⊗ H_B with Schmidt decomposition |Ψ⟩ = ∑_i √λ_i |u_i⟩_A |v_i⟩_B (as used in analyses by Erhard Schmidt and popularized in textbooks by Peres, Asher and Nielsen, Michael A.), one obtains ρ_A = ∑_i λ_i |u_i⟩⟨u_i|. This derivation is the basis for quantitative measures of entanglement such as the entanglement entropy found in studies by groups at IQIM and research from MIT and Caltech.
A subsystem is in a pure state if and only if ρ_A is a rank-1 projector, otherwise it is mixed. Mixedness can arise either from classical probabilistic mixtures (as in convex combinations studied in Gleason's theorem) or from entanglement with an environment. The purity P = Tr(ρ_A^2) and participation ratio are used as diagnostics in work by William Wootters and others. For bipartite pure global states, spectra of reduced density matrices on A and B coincide, a fact employed in the study of Schmidt rank and in computational approaches by groups at Los Alamos National Laboratory and IBM Quantum.
Reduced density matrices provide the operational bridge between entanglement theory and observable consequences: nonzero von Neumann entropy of ρ_A signals entanglement in a global pure state, a cornerstone in protocols like quantum teleportation and superdense coding developed by researchers including Bennett, Charles H. and Brassard, Gilles. In decoherence theory (papers by Wojciech Zurek), tracing over environment degrees of freedom yields a reduced density matrix whose off-diagonal coherence terms decay, explaining emergence of classicality and preferred pointer bases. In measurement theory, the post-measurement state for a subsystem is described by updated reduced density matrices via selective or non-selective measurement maps tied to the Kraus representation and completely positive trace-preserving maps studied in operator theory.
Reduced density matrices are indispensable in quantum information tasks: entanglement quantification, channel capacities, and error correction (as in work by Peter Shor and Daniel Gottesman). In many-body physics, reduced density matrices underlie the Density Matrix Renormalization Group (DMRG) method introduced by Steven R. White and the study of entanglement spectra in topological phases investigated by Haldane, F. D. M. and collaborators. Thermalization, eigenstate thermalization hypothesis (ETH) research at institutions like Perimeter Institute uses reduced density matrices to assess local equilibration. Quantum chemistry uses reduced density matrices (one- and two-particle RDMs) in electronic structure methods derived from the work of H. A. Mayer and A. J. Coleman.
Practically, computing reduced density matrices relies on exact diagonalization, tensor network methods (matrix product states used in DMRG and projected entangled pair states), quantum Monte Carlo estimators, and density functional approximations in electronic structure codes from Gaussian and Quantum ESPRESSO. Example computations include obtaining the one-body reduced density matrix for fermionic systems to study off-diagonal long-range order (relevant to BCS theory of superconductivity) and computing entanglement entropy in spin chains such as the Heisenberg model and Transverse-field Ising model. Ongoing advances in quantum computing platforms from Google Quantum AI and Rigetti increasingly allow experimental tomography of reduced density matrices for verification of entanglement and benchmarking of devices.
Category:Quantum mechanics Category:Quantum information theory