| reduced density matrix | |
|---|---|
| Name | Reduced density matrix |
| Field | Quantum physics |
| Related | Density matrix; Partial trace; Entanglement |
reduced density matrix
A reduced density matrix is the density operator describing a subsystem of a larger quantum system obtained by tracing out degrees of freedom of the remainder. It encapsulates all observable statistics for that subsystem when the global state is known only partially, and is essential for describing open quantum system dynamics, entanglement measures, and quantum statistical ensembles.
The reduced density matrix (often called reduced density operator) ρ_A represents the state of subsystem A extracted from a global density operator ρ_{AB} on the Hilbert space H_A ⊗ H_B. It is positive semidefinite, self-adjoint, and has unit trace when the global state is normalized. Important properties include convexity under mixtures, monotonicity under completely positive trace-preserving maps such as quantum channels, and invariance of eigenvalues under global unitary transforms that act nontrivially only on traced-out subsystems. The spectral decomposition of ρ_A yields its eigenstates and eigenvalues (the subsystem's occupation probabilities), which are used to compute von Neumann entropy and other information-theoretic quantities.
Formally, given a composite system with state ρ_{AB} on H_A ⊗ H_B, the reduced density matrix for A is ρ_A = Tr_B(ρ_{AB}), where Tr_B denotes the partial trace over H_B. In coordinates, if {|i_A⟩} and {|μ_B⟩} are orthonormal bases, the matrix elements are (ρ_A)_{ij} = Σ_μ ⟨i_A, μ_B| ρ_{AB} |j_A, μ_B⟩. For pure global states |Ψ⟩, ρ_{AB}=|Ψ⟩⟨Ψ|, and the nonzero spectra of ρ_A and ρ_B coincide (Schmidt decomposition), relating to the Schmidt rank and Schmidt decomposition theorem. The von Neumann entropy S(ρ_A) = −Tr(ρ_A log ρ_A) quantifies uncertainty and is central to thermodynamics and quantum information theory; it reduces to Shannon entropy for classical diagonal states.
The partial trace operation is linear and defined by its action on product operators: Tr_B(O_A ⊗ O_B) = O_A Tr(O_B). For a pure state |Ψ⟩ ∈ H_A ⊗ H_B, one computes ρ_A by expanding |Ψ⟩ in a product basis and tracing out B. The Schmidt decomposition provides an explicit form |Ψ⟩ = Σ_k √λ_k |u_k⟩_A ⊗ |v_k⟩_B, giving ρ_A = Σ_k λ_k |u_k⟩⟨u_k|. This equivalence underpins many results, including the fact that bipartite pure-state entanglement is fully characterized by the eigenvalue spectrum {λ_k} of the reduced states. Foundational work in this area draws on methods used by John von Neumann and later formalized in quantum information theory by researchers such as Niels Bohr (conceptual), Hugh Everett (relative state), and modern textbooks like Nielsen and Chuang.
Reduced density matrices often represent mixed states, that is, statistical ensembles of pure states arising from ignorance or entanglement. Distinct ensembles can yield the same ρ_A (ensemble equivalence), illustrating that ρ_A is the complete description for expectation values of subsystem observables. Purity P(ρ)=Tr(ρ^2) and linear entropy S_L = 1 − P measure mixedness; purity equals 1 for pure states and decreases as the subsystem becomes mixed. Other measures include the von Neumann entropy and Rényi entropies. In many-body contexts, the eigenvalue distribution of reduced density matrices, often analyzed via random matrix theory by groups like those at Princeton University and Los Alamos National Laboratory, informs thermalization and eigenstate thermalization hypothesis studies.
Reduced density matrices are the primary tool to quantify bipartite entanglement for pure states via entropy of entanglement S(ρ_A). For mixed states, entanglement measures such as entanglement of formation, concurrence, and negativity depend on reduced states or their extensions. The presence of classical and quantum correlations is diagnosed by comparing ρ_{AB} with ρ_A ⊗ ρ_B and computing mutual information I(A:B)=S(ρ_A)+S(ρ_B)−S(ρ_{AB}). Reduced density matrices also appear in studies of decoherence pioneered by Wojciech Zurek and in operational tasks like quantum teleportation experiments by groups at IBM and Google Quantum AI where subsystem states and their fidelities are measured.
In quantum statistical mechanics, reduced density matrices describe subsystems in contact with reservoirs, leading to canonical and grand-canonical reduced states when the environment is large and thermal. The formalism underlies derivations of master equations such as the Lindblad equation and Redfield equations for open quantum system dynamics, often derived using projection operator techniques associated with researchers like Rudolf Zwanzig and Hendrik Anthony Kramers-style approaches. Reduced states are crucial to quantum thermodynamics (work extraction, equilibration), transport phenomena in condensed matter (lattice models studied at CERN and university research groups), and quantum impurity problems solved by numerical renormalization group methods.
Numerical computation of reduced density matrices is routine in exact diagonalization, density matrix renormalization group (DMRG), tensor network methods, and quantum Monte Carlo. For finite spin chains (e.g., Heisenberg model, Ising model), one computes ρ_A for contiguous blocks to analyze entanglement scaling and area laws. In quantum chemistry, reduced one- and two-particle density matrices encode electronic correlations and are used in configuration interaction and coupled cluster calculations at institutions like Lawrence Berkeley National Laboratory. On quantum hardware, tomography protocols reconstruct reduced density matrices for subsystems; compressed sensing and shadow tomography reduce measurement cost, developed by researchers including Patrick Hayden and groups at MIT and Caltech.
Category:Quantum mechanics Category:Quantum information theory