--- title: Quantum Fokker–Planck Master Equation url: https://www.emergentmind.com/topics/quantum-fokker-planck-master-equation type: topic --- # Quantum Fokker–Planck Master Equation In the literature surveyed here, the quantum Fokker–Planck master equation denotes a family of irreversible quantum evolution equations that combine drift, diffusion, and dissipation in operator space, phase space, or enlarged system–detector spaces. Depending on the framework, the dynamical object may be a density operator, a Wigner function, a reduced diagonal density functional, or an element of a noncommutative \(L^p\) space; correspondingly, the generator may appear as a GKLS dissipator, a Wigner–Poisson/Fokker–Planck PDE, or an operator-valued Fokker–Planck equation for classical filter variables coupled to quantum backaction [2305.05805, 1811.12460, 2106.05718, 2110.09159]. ## 1. Scope and competing definitions A central feature of the subject is that there is no single universally fixed “quantum Fokker–Planck master equation.” Rather, the term covers several mathematically distinct but structurally related constructions. Some are obtained by canonical quantization of classical Fokker–Planck generators; some arise from Wigner transforms of Lindblad or quantum Brownian-motion equations; some are exact or approximate reduced equations for continuously monitored systems; and some are formulated directly on noncommutative \(L^p\) spaces over von Neumann algebras [2305.05805, 2308.11580, 2106.05718]. | Formulation | State variable | Representative structure | |---|---|---| | Operator-space FP/GKLS | \(\rho\) | current–divergence form, transformable to Lindblad form | | Wigner/phase-space FP | \(W(x,p)\) or \(w(x,\xi)\) | drift, friction, diffusion, pseudo-differential potential | | Noncommutative FP | \(f_t \in L^2(M)\) | \(H_q=C+V\) built from derivations and positive maps | | Measurement-feedback QFPME | \(\hat\rho(D)\) or \(\hat\rho(G)\) | quantum term plus FP drift–diffusion in detector/filter variables | In the operator-space approach, the quantum Fokker–Planck structure is a continuity equation in operator space. In phase-space approaches, the equation acts on a Wigner function and resembles a Kramers or Ornstein–Uhlenbeck equation supplemented by quantum potential terms. In measurement-and-feedback theory, the Fokker–Planck variable is neither position nor momentum but a classical detector output or filtered signal, and the equation propagates a joint system–signal state [2305.05805, 1811.12460, 2110.09159]. A further distinction concerns what is meant by “quantum.” In some works the evolution remains fully quantum at the level of \(\rho\) or \(W\). In others, a classical Fokker–Planck equation is derived from an underlying Schrödinger dynamics after coarse-graining away coherences in a preferred basis. That latter use retains a quantum origin but not a fully quantum reduced state description [2505.22693]. ## 2. Operator formulations, GKLS structure, and noncommutative semigroups One operator-theoretic construction begins from a canonical quantization of the classical Fokker–Planck equation and writes the dissipative dynamics in a current–divergence form, \[ i\hbar\,\dot{\rho} = [H,\rho] -\sum_{j}\Big([a_j,J_j^\dagger]+[a_j^\dagger,J_j]\Big), \qquad J_j=i\,\gamma_j\Big(g_j\,\rho+[a_j^\dagger,\rho]\Big), \] with \[ g_j=\frac{e^{-\beta_j H}a_j e^{+\beta_j H}-a_j}{i\hbar\,\beta_j}. \] This formulation makes the quantum analogues of probability current and detailed balance explicit, and the same paper shows that the resulting dissipator can be transformed into Lindblad form with a positive Kossakowski matrix, thereby ensuring trace preservation and positivity of the density operator [2305.05805]. In that same framework, thermal contact is encoded by choosing \[ C_{\alpha}=A_{\alpha}, \qquad B_{\alpha}=e^{-\beta_{\alpha}H}A_{\alpha}e^{+\beta_{\alpha}H}, \] so that detailed balance is expressed as \[ B_{\alpha}\rho_{\mathrm{eq}}=\rho_{\mathrm{eq}}C_{\alpha}, \qquad \rho_{\mathrm{eq}}\propto e^{-\beta H}. \] The Gibbs state is then a fixed point, and the Fokker–Planck structure yields explicit expressions for entropy production and entropy flux. The same work emphasizes that the entropy production rate need not vanish for a closed system except in equilibrium [2305.05805]. A different, more abstract version is developed on \(\sigma\)-finite von Neumann algebras equipped with Haagerup \(L^p\) spaces. There the quantum Fokker–Planck equation is \[ \frac{d f_t}{dt} = -\sum_{1\le i,j\le n} d_{ij}\,\delta_i^{\,*}\delta_j(f_t)+V(f_t) =:H_q(f_t), \] where the “quantized Laplacian” \[ C=-\Big[\sum_{1\le i,j\le n} d_{ij}\,\delta_i^{\,*}\delta_j\Big] \] is built from weak\(^*\)-closed derivations, while the potential term is \(V=T-\mathrm{id}\) for a unital Schwarz map \(T\). Under Detailed Balance II, \(H_q\) generates conservative \(L^2(M)\)-Markov \(C_0\)-semigroups, induced by integrable Markov semigroups on \(M\); if \(C\) commutes with \(\Delta^{1/2}\), the semigroups satisfy DBII as well [2106.05718]. That noncommutative theory also gives an asymptotic analysis by spectral methods. If \(-H_q\) has a spectral gap \(y>0\), then \[ \|e^{tH_q}f_0-(1-S_{H_q})f_0\| \le e^{-yt}\|e_{(y,\infty)}f_0\|, \] and even without a gap one has strong convergence to \((1-S_{H_q})f_0\) as \(t\to\infty\). Convergence in \(L^1\) is supplemented by a noncommutative Csiszár–Kullback inequality, \[ \|f-g\|_1 \le 2\,\mathrm{tr}\!\left( g\, \big| (g^{-1/2}fg^{-1/2})\log(g^{-1/2}fg^{-1/2}) \big| \right), \] which controls approach to equilibrium in terms of relative entropy [2106.05718]. By contrast, the high-temperature Markovian Caldeira–Leggett equation, \[ \frac{d\rho}{dt} = -\frac{i}{\hbar}[H_S,\rho] -\frac{i\gamma}{\hbar}[x,\{p,\rho\}] -\frac{2m\gamma k_B T}{\hbar^2}[x,[x,\rho]], \] is explicitly identified as not being of Lindblad form; the low-temperature and positivity issues that accompany this structure motivate several later reformulations [1008.4402]. ## 3. Wigner-space and phase-space equations In kinetic and open-systems theory, the quantum Fokker–Planck master equation often appears in Wigner representation. For a reduced density matrix \(\rho(t,x,y)\), the Wigner function is \[ w(t,x,\xi)=\frac{1}{(2\pi)^3}\int_{\mathbb{R}^3} \rho\Big(t,x+\tfrac{\eta}{2},x-\tfrac{\eta}{2}\Big)e^{-i\eta\cdot\xi}\,d\eta, \] and the nonlinearity enters through the Wigner pseudo-differential operator \(\Theta[V]\). In the non-Markovian Unruh–Zurek and Hu–Paz–Zhang models with self-consistent Poisson coupling, the equations take the form \[ T[w]-\Omega_0^2(x\cdot\nabla_\xi)w-2\gamma\,\mathrm{div}_\xi(\xi w)+\Theta[V]w = c(t)\Delta_\xi w-d(t)\nabla_x\!\cdot\nabla_\xi w \] and \[ T[w]-\big(\Omega^2-2a(t)\big)(x\cdot\nabla_\xi)w-2b(t)\,\mathrm{div}_\xi(\xi w)+\Theta[V]w = c(t)\Delta_\xi w-d(t)\nabla_x\!\cdot\nabla_\xi w, \] with time-dependent integral coefficients encoding memory of the bath and a nonlinear Hartree potential satisfying \(-\Delta_xV=n\), \(n=\int w\,d\xi\). The analysis constructs explicit Gaussian propagators, proves local existence by Banach fixed point, and obtains global mild solutions by kinetic energy estimates and Lieb–Thirring control of the density [1811.12460]. For an Ohmic quantum Langevin dynamics, a simpler reduced formulation arises by integrating out momentum and studying the momentum-independent Wigner function \(W(q,t)\). The resulting exact reduced equation is \[ \frac{\partial W(q,t)}{\partial t} = -\Omega(t)\,\frac{\partial}{\partial q}\big[q\,W(q,t)\big] +\frac{1}{2}\,\mathcal{D}(t)\,\frac{\partial^2 W(q,t)}{\partial q^2}, \] with \(\Omega(t)=\dot g(t)/g(t)\) and \(\mathcal D(t)=\dot\Sigma(t)-2\Sigma(t)\Omega(t)\). The adjoint density-matrix equation becomes \[ \frac{\partial \hat{\rho}}{\partial t} = -\Omega(t)\left( \hat{\rho} +\frac{i}{2\hbar}\{\hat q,[\hat p,\hat\rho]\} \right) -\frac{1}{2\hbar^2}\mathcal D(t)[\hat p,[\hat p,\hat\rho]], \] which is considerably simpler than the full phase-space Schramm–Jung–Grabert equation while preserving the exact \(q\)-statistics [1712.02780]. A related non-Markovian Caldeira–Leggett construction starts instead from a classical non-Markovian Fokker–Planck equation and quantizes it by a “dynamical quantization” procedure. In position representation the result is \[ i\hbar \frac{\partial \rho}{\partial t} = [V(x_1)-V(x_2)]\rho -\frac{\hbar^2}{2m}\Big(\partial_{x_1}^2-\partial_{x_2}^2\Big)\rho -i\hbar\gamma(x_1-x_2)\Big(\partial_{x_1}-\partial_{x_2}\Big)\rho -\frac{D_\hbar(\infty)I(t)}{\hbar^2}(x_1-x_2)^2\rho, \] with \(D_\hbar(t)=2\gamma m E_\hbar(\infty)I(t)\), \(E_\hbar(\infty)=\frac{\hbar\omega}{2}\coth\!\big(\frac{\hbar\omega}{2k_BT}\big)\), and a concrete non-Markovian choice \(I(t)=1-e^{-t/t_c}\). In Wigner representation for a free particle, this becomes a quantum Rayleigh equation with time-dependent momentum diffusion [1008.4402]. Another phase-space version follows from a Markovian quantum Langevin equation for a trapped particle in a quantum-thermal Ohmic bath. In the underdamped harmonic regime one obtains the semiclassical Kramers-type equation \[ \partial_t\mathcal P = \left[ -v\partial_q +\frac{\gamma}{m}\partial_v v +\frac{V'(q)}{m}\partial_v +\frac{\gamma k_B T_{\rm eff}}{m^2}\partial_v^2 \right]\mathcal P, \] with \[ T_{\rm eff} = \frac{\hbar\Omega}{2k_B\,\coth\!\left(\frac{\hbar\Omega}{2k_BT}\right)}. \] For generic nonlinear \(V(q)\), the quantum Fokker–Planck equation acquires the higher-derivative correction \[ \partial_t\mathcal P = \left[ -v\partial_q +\frac{\gamma}{m}\partial_v v +\frac{\gamma k_B T_{\rm eff}}{m^2}\partial_v^2 +\partial_v(\partial_v\tilde F) \right]\mathcal P, \] where \(\tilde F\) is given by an infinite series in \((\hbar\partial_v/m)^{2n}V^{(2n+1)}\). For harmonic confinement, the series truncates exactly; for nonlinear potentials, truncations must be treated cautiously because of Pawula’s theorem [2308.07423]. ## 4. Memory, regularity, and Markovian limits Non-Markovianity enters these equations in different ways. In the UZ and HPZ Wigner–Poisson models, it appears through integral coefficients \(a(t)\), \(b(t)\), \(c(t)\), and \(d(t)\) built from bath spectral densities, thermal factors \(\coth(\beta\theta/2)\), and time integrals over the full past of the system. This modifies both drift and dissipation and produces propagators with singular short-time scaling. In particular, the paper derives \[ D_{UZ}(t)\sim t^6,\quad a_{UZ}(t)\sim t^{-4},\quad b_{UZ}(t)\sim t^{-3},\quad c_{UZ}(t)\sim t^{-2}, \] and \[ D_{HPZ}(t)\sim t^6,\quad a_{HPZ}(t)\sim t^{-5},\quad b_{HPZ}(t)\sim t^{-4},\quad c_{HPZ}(t)\sim t^{-2}, \] with an associated significant lack of Sobolev regularity relative to Markovian counterparts. The global theory therefore relies on mild formulations, time-dependent Gaussian bounds, continuity equations, and kinetic-energy control rather than on uniform parabolic smoothing [1811.12460]. A gravitational variant yields a non-Markovian quantum master equation and associated Wigner Fokker–Planck equation for two quantum masses interacting with a graviton bath at all temperatures. The graviton spectral density is \[ J(\omega)=\frac{32\pi^4\alpha^2}{15}\,\omega\,\Theta(\Lambda-\omega), \] so it is Ohmic at the field level, but the system coupling is quadratic in the separation coordinates, which generates quartic derivative structures and energy-basis decoherence terms. In the high-temperature short-memory regime the master equation reduces to the ABH structure \(-\Gamma_{ABH}[H_0,[H_0,\rho]]\), whereas in the low-temperature non-Markovian regime the off-diagonal momentum-space elements decay as \[ \rho(p,p';t)\simeq \rho(0)\exp\!\left[-a_0\ln(t/\tau_0)-\frac{6a_0}{\beta^2}t^2+\cdots\right]. \] The zero-temperature part is logarithmic in time, and the temperature-dependent part is quadratic in time, which differs sharply from the exponential Markovian behavior [2504.11991]. A different limit appears in derivations from reversible Schrödinger dynamics. Expanding the state in the eigenbasis of a time-independent \(H_0\), one may write the populations \(P_k(t)=|a_k(t)|^2\) as \[ P_k(t)=\sum_l T_{kl}P_l(t)+\sigma_k(t), \qquad T_{kl}=|U_{kl}|^2, \] and then neglect the coherence correction \(\sigma_k(t)\) under coarse-graining. This yields a classical master equation \[ \partial_t P_k(t)=\sum_{l\neq k}\big[W_{kl}P_l(t)-W_{lk}P_k(t)\big], \] and, in the continuum limit, a Fokker–Planck equation \[ \partial_t P(x,t) = -\partial_x\big[A(x,t)P(x,t)\big] +\partial_x^2\big[D(x,t)P(x,t)\big]. \] A plausible implication is that some occurrences of “quantum Fokker–Planck” in the literature refer not to a fully quantum reduced state equation, but to a classical diffusion equation derived from an underlying quantum dynamics after basis-dependent decoherence and coarse-graining [2505.22693]. ## 5. Continuous measurement, feedback, and signal filtering In continuous-measurement theory, the quantum Fokker–Planck master equation describes the joint dynamics of a quantum system and a classical detector output. For continuous monitoring of a Hermitian observable \(\hat A\) with measurement strength \(\lambda\), and for a detector with finite bandwidth \(\gamma\) modeled by an Ornstein–Uhlenbeck filter, the operator-valued density \(\hat\rho_t(D)\) obeys \[ \partial_t\hat\rho_t(D) = \mathcal L(D)\hat\rho_t(D) +\lambda\mathcal D[\hat A]\hat\rho_t(D) -\gamma\partial_D\,\mathcal A(D)\hat\rho_t(D) +\frac{\gamma^2}{8\lambda}\partial_D^2\hat\rho_t(D), \] where \[ \mathcal A(D)\hat\rho=\frac12\{\hat A-D,\hat\rho\}. \] Here \(\mathcal L(D)\) may depend nonlinearly on the filtered outcome \(D\), so the formalism covers both linear and nonlinear feedback. In the fast-measurement limit \(\gamma\gg\Gamma\), detector elimination yields a reduced Markovian master equation \[ \partial_t\hat\rho_t = \big[\mathcal L_0+\lambda\mathcal D[\hat A]+\gamma^{-1}\mathcal L_{\mathrm{corr}}\big]\hat\rho_t, \] which generalizes the Wiseman–Milburn equation and supports analytic treatments of threshold engines and driven qubits [2110.09159]. That construction has been extended to arbitrary linear filtering. If the filter state \(G'(t)\in\mathbb R^n\) satisfies \[ dG'(t)=M\,G'(t)\,dt+b\,z(t)\,dt, \] with \(z(t)=\langle A\rangle(t)+\xi(t)\), then the joint operator-valued density \(\hat\rho(G,t)\) obeys \[ \partial_t \hat\rho(G) = -\frac{i}{\hbar}[H(G),\hat\rho(G)] +\lambda \mathcal D[A]\hat\rho(G) -\nabla_G\!\cdot\!\left[\frac12\{bA+MG,\hat\rho(G)\}\right] +\frac12\nabla_G\!\cdot\!\left(\frac{bb^T}{4\lambda}\nabla_G\hat\rho(G)\right). \] This encompasses single-pole low-pass filters, cascaded low-pass filters, band-pass filters, and finite-order ODE kernels. In harmonic-oscillator cooling with one filtered layer, the ensemble-averaged steady-state energy is \[ \langle H_1\rangle_\infty = \frac{\hbar\omega}{2}\left(\frac{\lambda}{\gamma}+\frac{\gamma}{4\lambda}\right), \] with optimum \(\gamma=2\lambda\), for which \(\langle H_1\rangle_\infty=\hbar\omega/2\). Two- and three-layer cascades and band-pass filters admit explicit analytical formulas and reveal bandwidth-dependent stability boundaries [2509.16841]. These feedback QFPMEs differ conceptually from phase-space quantum Brownian-motion equations. The Fokker–Planck variable is now the filtered signal rather than a particle coordinate or momentum, the diffusion tensor is \(bb^T/(4\lambda)\), and the Lindblad term \(\lambda\mathcal D[A]\) represents measurement backaction. The structure is therefore a hybrid quantum–classical transport equation rather than a Wigner kinetic equation [2110.09159, 2509.16841]. ## 6. Numerical reformulations and domain-specific extensions For numerical work on Markovian open systems, the Wigner–Fokker–Planck equation can be transformed by Fourier transform in momentum to symmetrized position coordinates \(u(x,\eta,t)=\rho(x+\eta/2,x-\eta/2,t)\). Starting from \[ w_t+k\cdot\nabla_x w+\Theta[V]\{w\}=Q_{FP}\{w\}, \qquad Q_{FP}\{w\}=\Delta_k w+\nabla_k\cdot(kw)+\Delta_x w, \] the transform removes the nonlocal pseudo-differential potential term and yields the local convection–diffusion equation \[ \widehat{w}_t+i\nabla_\eta\!\cdot\nabla_x\widehat{w}+i\widehat{w}\,\delta V = -\eta^2\widehat{w}-\eta\cdot\nabla_\eta\widehat{w}+\Delta_x\widehat{w}. \] In one dimension, decomposing \(\widehat w=R+iI\) gives a coupled convection–diffusion system suitable for NIPG-DG discretization. The paper reports that for piecewise linear DG functions the method exhibits \(O(h)\) convergence in \(L^2\) for the harmonic steady state and \(O(h^2)\) projection error, and it identifies a cost crossover in favor of the transformed equation for meshes with \(m>19\) intervals in momentum or \(\eta\) [2308.11580]. Quantum Fokker–Planck methods also appear in cosmology. For a massless scalar field with quartic interaction in de Sitter space, coarse-graining at \[ k_c(t)=\epsilon a(t)H \] and integrating out short modes yields a functional equation for the reduced diagonal density \(P_L[\Phi_L;t]\), \[ \partial_t P_L = \int_{k