Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Physical Review A

Quasiprobability methods for nonlinear chemical and optical systems

P. D. Drummond, Crispin W. Gardiner, Daniel F. Walls

Published Aug 1, 1981• 79 citations
Publisher

Abstract

A review of quasiprobability methods for transforming chemical and quantum-optical master equations into Fokker-Planck equations is presented. For cases where conventional representations lead to Fokker-Planck equations with non-positive-definite diffusion coefficients; e.g., sub-Poissonian statistics, a generalization of the representation involving an extension to the complex plane enables analytic results to be obtained for certain nonlinear chemical and optical processes. Alternatively, a different integration measure may be chosen which ensures a positive distribution and Fokker-Planck equation with positive-semidefinite diffusion coefficients. This enables stochastic differential equations to be defined. These methods are applied to two-photon absorption and dispersive bistability in quantum optics where nonclassical photon statistics arise and to two models of nonlinear chemical reactions where sub-Poissonian statistics occur.

DOI
10.1103/physreva.24.914

Related papers

Quasiprobability distributions in dispersive optical bistability

Karl Vogel, H. Risken

For the model of Drummond and Walls [J. Phys. A 13, 725 (1980)] describing dispersive optical bistability, the s-parametrized quasidistributions W(\ensuremath{\alpha},s) introduced by Cahill and Glauber [Phys. Rev. A 117, 1882 (1969)] are investigated. The equation of motion for W(\ensuremath{\alpha},s) generally has derivative terms up to third order. For the Q function (s=-1) and the P function (s=1) the third-order terms vanish. The equation of motion is then a pseudo-Fokker-Planck equation, i.e., a Fokker-Planck equation where the diffusion matrix is not positive definite or semidefinite. It is shown that the equation of motion with the inclusion of third-order-derivative terms can be solved by the matrix continued-fraction method. In particular, results for the Wigner function (s=0) are presented and compared with the results of the Q function (s=-1). Furthermore, the results are compared also with those obtained from the equation of motion for the Wigner function where the third-order-derivative terms have been neglected.

Physical Review A•May 1, 1989•74 citations

Nonequilibrium transitions in quantum optical systems

P. D. Drummond

The topic of this thesis is a theoretical study of nonequilibrium transitions and quantum statistical properties of nonlinear quantum optical systems driven by an external radiation source. In certain limiting cases, a comparison is made between these transitions and the phase transitions found in equilibrium physical systems. In chapters one and two, the mathematical tools are introduced. In operator terms, the time development is described by a Markovian master equation in the interaction picture. This is equivalent to a corresponding time development equation or Fokker-Planck equation in a vector space of c-numbers. In order to deal with the type of Fokker-Planck equation that results, we introduce a quantum classical correspondence resulting in a distribution function over a complex phase-space, which is a generalisation of the Glauber-Sudarshan P-function. In chapter three this is applied to a model of a coherently driven mode with nonlinear dispersion and absorption. We find in the limit of zero temperature, that the spectrum is symmetric relative to the input frequency, and an exact solution is obtained for the distribution function. For a detuned driving field and nonlinear dispersion, optical bistability can occur. In chapter four a model of sub/second harmonic generation is introduced. This has several non-equilibrium transitions, including dispersive optical bistability, and bistable behaviour with coherent phase-locked input to both modes. Exact solutions occur in the limit of zero temperature and adiabatic elimination of one mode. In both chapters three and four, steady-state photon anti-bunching occurs with an absorptive nonlinearity. In chapter five we include interactions between the radiation mode and a fluorescent atomic system. In this case different behaviour occurs depending on the relative decay rates of the individual atoms and of the radiation mode. In the case of a high-Q interferometer, the atomic variables can be adiabatically eliminated. Both dispersive and absorptive bistability can occur. We show by analytic and numerical calculations (in the case of inhomogeneous broadening) that dispersive operation has advantages in requiring a lower atomic density and input field to observe bistability. When the input field has Gaussian (rather than coherent) photon statistics, there is no bistability, but enhanced photon bunching occurs. Finally, there is a different type of behaviour when the field mode decays rapidly and can be adiabatically eliminated. Within the cooperation lifetime the system can be described by a J²-invariant Hamiltonian, giving the special case of only collective damping. The result is a new type of critical point transition in the thermodynamic limit, with the appearance of a family of solutions like Lotka-Volterra cycles for a coherent driving field above threshold.

•Jan 1, 1979•0 citations

Quasiprobability distributions for open quantum systems within the Lindblad theory

Аурелиан Исар, W. Scheid, Aurel Sǎndulescu

The Lindblad master equation for the damped quantum harmonic oscillator is transformed into Fokker–Planck equations for quasiprobability distributions. A comparative study is made for the Glauber P representation, the antinormal-ordering Q representation and the Wigner W representation. It will be proven that the variances for the damped harmonic oscillator found with these representations are the same. By solving the Fokker–Planck equations in the steady state, it will be shown that the quasiprobability distributions are two-dimensional Gaussians with widths determined by the diffusion coefficients.

Journal of Mathematical Physics•Aug 1, 1991•23 citations

Quantum theory of optical bistability. III. Atomic fluorescence in a low- Q cavity

H. J. Carmichael, Daniel F. Walls, P. D. Drummond +1 more

We specialize the quantum-statistical theory of optical bistability developed in "Optical Bistability. II" to a low-$Q$ cavity. From a linearized Fokker-Planck equation derived in the positive $P$ representation, the steady-state correlation matrix and spectrum of fluctuations are calculated. Expressions which normally assume a positive-definite diffusion matrix are shown to be valid for nonpositive-definite diffusion. For absorptive bistability we find agreement with results derived previously by quite different methods. More generally we extend previously published results by including atomic and cavity detunings. We compare the transmitted spectrum for dispersive bistability with that for absorptive bistability, and photon antibunching in dispersive bistability is discussed.

Physical Review A•Jun 1, 1983•35 citations

Generalised P-representations in quantum optics

P. D. Drummond, Crispin W. Gardiner

A class of normal ordering representations of quantum operators is introduced, that generalises the Glauber-Sudarshan P-representation by using nondiagonal coherent state projection operators. These are shown to have practical application to the solution of quantum mechanical master equations. Different representations have different domains of integration, on a complex extension of the usual canonical phase-space. The 'complex P-representation' is the case in which analytic P-functions are defined and normalised on contours in the complex plane. In this case, exact steady-state solutions can often be obtained, even when this is not possible using the Glauber-Sudarshan P-representation. The 'positive P-representation' is the case in which the domain is the whole complex phase-space. In this case the P-function may always be chosen positive, and any Fokker-Planck equation arising can be chosen to have a positive-semidefinite diffusion array. Thus the 'positive P-representation' is a genuine probability distribution. The new representations are especially useful in cases of nonclassical statistics.

Journal of Physics A Mathematical and General•Jul 1, 1980•716 citations

Quasiprobability density diffusion equations for the quantum Brownian motion in a potential

Mulligan, Bernard P. J.

Wigner's [E. P. Wigner, Phys. Rev., 1932, 40, 749] representation of the density operator as a c-number quasiprobablity distribution in phase space allowing quantum mechanical averages involving the density matrix to be calculated as phase space averages just as classical averages originally used by him to obtain quantum mechanical corrections to classical thermodynamic equilibrium i.e. to the Maxwell-Boltzmann distribution so applying to closed quantum systems is extended to open quantum systems comprising a canonical ensemble of Brownian particles in a potential. This is accomplished via an idea of Gross and Lebowitz [E. P. Gross and J. L. Lebowitz, Phys. Rev. 1956, 104, 1528]. They suggested that using Wigner's representation the connection between classical and quantum collision kernels, (i.e. in classical mechanics the Stosszahlanzatz describing the bath-particle interaction in the open system in the Boltzmann equation for the single particle distribution function) is much more transparent than in the density operator formalism. Moreover the quantum kernel should closely correspond to the classical one. Hence the idea developed in this Thesis that in the quantum Brownian motion the collision term in a quantum master equation in Wigner's representation should be described by a Kramers-Moyal like expansion truncated at the second term (leading of course in the classical limit to the Fokker-Planck equation) as in the classical Brownian motion. Imposition of the Wigner equilibrium distribution as the stationary solution of this equation (which is akin to the Fokker-Planck equation) in the manner used by Einstein [A. Einstein, in R. H. Fürth, Ed., Investigations on the Theory of the Brownian Movement, Methuen, London, 1926; reprinted Dover, New York, 1954] used to calculate diffusion coefficients in the Fokker-Planck equation by imposing the Maxwell-Boltzmann distribution as the stationary distribution then leads to the most important results of this Thesis. Namely the diffusion coefficients in the master equation become functions of the quantum parameter and the derivatives of the potential. Moreover all the solution techniques (matrix continued fractions etc) developed for the Fokker-Planck equation carry over to the quantum case as is illustrated by calculating the reaction rate and dynamical structure factor of a particle in a periodic cosine potential where the results are in agreement with those predicted by quantum reaction rate theory.

•Jan 1, 2007•1 citation
A review of quasiprobability methods for transforming chemical and quantum-optical master equations into Fokker-Planck equations is presented. For cases where conventional representations lead to Fokker-Planck equations with non-positive-definite diffusion coefficients; e.g., sub-Poissonian statistics, a generalization of the representation involving an extension to the complex plane enables analytic results to be obtained for certain nonlinear chemical and optical processes. Alternatively, a different integration measure may be chosen which ensures a positive distribution and Fokker-Planck equation with positive-semidefinite diffusion coefficients. This enables stochastic differential equations to be defined. These methods are applied to two-photon absorption and dispersive bistability in quantum optics where nonclassical photon statistics arise and to two models of nonlinear chemical reactions where sub-Poissonian statistics occur.
For the model of Drummond and Walls [J. Phys. A 13, 725 (1980)] describing dispersive optical bistability, the s-parametrized quasidistributions W(\ensuremath{\alpha},s) introduced by Cahill and Glauber [Phys. Rev. A 117, 1882 (1969)] are investigated. The equation of motion for W(\ensuremath{\alpha},s) generally has derivative terms up to third order. For the Q function (s=-1) and the P function (s=1) the third-order terms vanish. The equation of motion is then a pseudo-Fokker-Planck equation, i.e., a Fokker-Planck equation where the diffusion matrix is not positive definite or semidefinite. It is shown that the equation of motion with the inclusion of third-order-derivative terms can be solved by the matrix continued-fraction method. In particular, results for the Wigner function (s=0) are presented and compared with the results of the Q function (s=-1). Furthermore, the results are compared also with those obtained from the equation of motion for the Wigner function where the third-order-derivative terms have been neglected.
The topic of this thesis is a theoretical study of nonequilibrium transitions and quantum statistical properties of nonlinear quantum optical systems driven by an external radiation source. In certain limiting cases, a comparison is made between these transitions and the phase transitions found in equilibrium physical systems. In chapters one and two, the mathematical tools are introduced. In operator terms, the time development is described by a Markovian master equation in the interaction picture. This is equivalent to a corresponding time development equation or Fokker-Planck equation in a vector space of c-numbers. In order to deal with the type of Fokker-Planck equation that results, we introduce a quantum classical correspondence resulting in a distribution function over a complex phase-space, which is a generalisation of the Glauber-Sudarshan P-function. In chapter three this is applied to a model of a coherently driven mode with nonlinear dispersion and absorption. We find in the limit of zero temperature, that the spectrum is symmetric relative to the input frequency, and an exact solution is obtained for the distribution function. For a detuned driving field and nonlinear dispersion, optical bistability can occur. In chapter four a model of sub/second harmonic generation is introduced. This has several non-equilibrium transitions, including dispersive optical bistability, and bistable behaviour with coherent phase-locked input to both modes. Exact solutions occur in the limit of zero temperature and adiabatic elimination of one mode. In both chapters three and four, steady-state photon anti-bunching occurs with an absorptive nonlinearity. In chapter five we include interactions between the radiation mode and a fluorescent atomic system. In this case different behaviour occurs depending on the relative decay rates of the individual atoms and of the radiation mode. In the case of a high-Q interferometer, the atomic variables can be adiabatically eliminated. Both dispersive and absorptive bistability can occur. We show by analytic and numerical calculations (in the case of inhomogeneous broadening) that dispersive operation has advantages in requiring a lower atomic density and input field to observe bistability. When the input field has Gaussian (rather than coherent) photon statistics, there is no bistability, but enhanced photon bunching occurs. Finally, there is a different type of behaviour when the field mode decays rapidly and can be adiabatically eliminated. Within the cooperation lifetime the system can be described by a J²-invariant Hamiltonian, giving the special case of only collective damping. The result is a new type of critical point transition in the thermodynamic limit, with the appearance of a family of solutions like Lotka-Volterra cycles for a coherent driving field above threshold.
The Lindblad master equation for the damped quantum harmonic oscillator is transformed into Fokker–Planck equations for quasiprobability distributions. A comparative study is made for the Glauber P representation, the antinormal-ordering Q representation and the Wigner W representation. It will be proven that the variances for the damped harmonic oscillator found with these representations are the same. By solving the Fokker–Planck equations in the steady state, it will be shown that the quasiprobability distributions are two-dimensional Gaussians with widths determined by the diffusion coefficients.
We specialize the quantum-statistical theory of optical bistability developed in "Optical Bistability. II" to a low-QQQ cavity. From a linearized Fokker-Planck equation derived in the positive PPP representation, the steady-state correlation matrix and spectrum of fluctuations are calculated. Expressions which normally assume a positive-definite diffusion matrix are shown to be valid for nonpositive-definite diffusion. For absorptive bistability we find agreement with results derived previously by quite different methods. More generally we extend previously published results by including atomic and cavity detunings. We compare the transmitted spectrum for dispersive bistability with that for absorptive bistability, and photon antibunching in dispersive bistability is discussed.
A class of normal ordering representations of quantum operators is introduced, that generalises the Glauber-Sudarshan P-representation by using nondiagonal coherent state projection operators. These are shown to have practical application to the solution of quantum mechanical master equations. Different representations have different domains of integration, on a complex extension of the usual canonical phase-space. The 'complex P-representation' is the case in which analytic P-functions are defined and normalised on contours in the complex plane. In this case, exact steady-state solutions can often be obtained, even when this is not possible using the Glauber-Sudarshan P-representation. The 'positive P-representation' is the case in which the domain is the whole complex phase-space. In this case the P-function may always be chosen positive, and any Fokker-Planck equation arising can be chosen to have a positive-semidefinite diffusion array. Thus the 'positive P-representation' is a genuine probability distribution. The new representations are especially useful in cases of nonclassical statistics.
Wigner's [E. P. Wigner, Phys. Rev., 1932, 40, 749] representation of the density operator as a c-number quasiprobablity distribution in phase space allowing quantum mechanical averages involving the density matrix to be calculated as phase space averages just as classical averages originally used by him to obtain quantum mechanical corrections to classical thermodynamic equilibrium i.e. to the Maxwell-Boltzmann distribution so applying to closed quantum systems is extended to open quantum systems comprising a canonical ensemble of Brownian particles in a potential. This is accomplished via an idea of Gross and Lebowitz [E. P. Gross and J. L. Lebowitz, Phys. Rev. 1956, 104, 1528]. They suggested that using Wigner's representation the connection between classical and quantum collision kernels, (i.e. in classical mechanics the Stosszahlanzatz describing the bath-particle interaction in the open system in the Boltzmann equation for the single particle distribution function) is much more transparent than in the density operator formalism. Moreover the quantum kernel should closely correspond to the classical one. Hence the idea developed in this Thesis that in the quantum Brownian motion the collision term in a quantum master equation in Wigner's representation should be described by a Kramers-Moyal like expansion truncated at the second term (leading of course in the classical limit to the Fokker-Planck equation) as in the classical Brownian motion. Imposition of the Wigner equilibrium distribution as the stationary solution of this equation (which is akin to the Fokker-Planck equation) in the manner used by Einstein [A. Einstein, in R. H. Fürth, Ed., Investigations on the Theory of the Brownian Movement, Methuen, London, 1926; reprinted Dover, New York, 1954] used to calculate diffusion coefficients in the Fokker-Planck equation by imposing the Maxwell-Boltzmann distribution as the stationary distribution then leads to the most important results of this Thesis. Namely the diffusion coefficients in the master equation become functions of the quantum parameter and the derivatives of the potential. Moreover all the solution techniques (matrix continued fractions etc) developed for the Fokker-Planck equation carry over to the quantum case as is illustrated by calculating the reaction rate and dynamical structure factor of a particle in a periodic cosine potential where the results are in agreement with those predicted by quantum reaction rate theory.