Entropy-production fluctuation theorem for a generalized Langevin particle in crossed electric and magnetic fields

Technical Summary: Entropy-production fluctuation theorem for a generalized Langevin particle in crossed electric and magnetic fields

Problem Statement
This work investigates the statistical fluctuations of total entropy production for a charged Brownian particle confined in a harmonic trap and driven out of equilibrium by crossed electric and magnetic fields. The system is characterized by non-Markovian dynamics, where the particle interacts with a structured bath that induces delayed friction (memory effects) and colored thermal noise, constrained by the fluctuation–dissipation theorem. The magnetic field is constant and perpendicular to the plane of motion, coupling the transverse coordinates, while the electric field is time-dependent, providing the external driving.

While detailed fluctuation theorems (DFT) for total entropy production have been established for Markovian systems (standard Langevin equations) in similar configurations, and fluctuation relations for non-Markovian work have been studied, an explicit demonstration of the DFT for total entropy production in this specific non-Markovian (Generalized Langevin Equation, or GLE) setting with crossed fields has not been previously reported.

Methodology
The authors employ an exact analytical approach based on the linearity of the GLE and the Gaussian nature of the noise. The methodology proceeds as follows:

1. Exact Solution of GLE: The authors solve the linear GLE using Laplace transformation. They demonstrate that for equilibrium initial conditions (canonical distribution in position, Maxwellian in velocity), the phase-space probability density (PSPD) remains Gaussian for all finite times.
2. Factorization: Due to the decoupling of the $z$-axis dynamics from the $(x, y)$ plane, the PSPD factorizes. The planar marginal density $P(\tilde{r}, t)$ is derived explicitly in terms of the fluctuation variables around the mean trajectory.
3. Stochastic Thermodynamics Framework: The study defines the total entropy production $\Delta s_{tot}$ as the sum of the medium entropy change ($\Delta s_m = Q/T$) and the system entropy change ($\Delta s = -\ln[P(\tilde{r}, \tau)/P(\tilde{r}_0)]$).
4. Two Solvable Protocols: The DFT is proven analytically for two distinct driving protocols:
   • Direct Forcing: The particle is driven by a prescribed time-dependent electric force $\tilde{F}_e(t)$.
   • Dragged Trap: The center of the harmonic trap is dragged according to a time-dependent protocol $\tilde{\lambda}(t)$, which is mathematically equivalent to the direct force protocol via $\tilde{F}_e(t) = k\tilde{\lambda}(t)$.
5. Statistical Verification: The authors prove that for both protocols, the total entropy production is a Gaussian random variable. They derive the mean and variance of this distribution, showing that the variance is exactly twice the mean ($\sigma^2 = 2\langle \Delta s_{tot} \rangle$). This specific mean-variance relationship is the necessary and sufficient condition for the DFT to hold.

Key Contributions and Results

• Proof of DFT in Non-Markovian Regime: The primary result is the analytical proof that the total entropy production obeys the detailed fluctuation theorem, $P(\Delta s_{tot})/P(-\Delta s_{tot}) = e^{\Delta s_{tot}}$, for a charged particle in crossed fields subject to memory effects. This extends previous Markovian results to the GLE framework.
• Gaussian Structure Preservation: The work establishes that the linear GLE dynamics preserves the Gaussian form of the phase-space distribution. This structural property ensures that the system entropy, defined via the instantaneous probability density, retains a canonical form relative to the fluctuation variables, facilitating the exact calculation of entropy production statistics.
• Equivalence of Protocols: The paper demonstrates that both the direct-force and dragged-trap protocols share the same structural mechanism: the linearity of the GLE preserves Gaussianity, and specific covariance identities enforce the variance-mean relation required by the fluctuation theorem.
• Numerical Validation: Using an exponentially decaying memory kernel, the authors perform Monte Carlo simulations to visualize the physical response.
  • Trajectory Response: Simulations show how memory induces a lag in the particle's response to the trap movement, while the magnetic field generates a transverse displacement (Lorentz force effect). The Monte Carlo covariance of the trajectory ensemble matches the theoretical boundaries derived from the embedded Markovian system.
  • Entropy Statistics: Numerical histograms of entropy production for forward and reverse protocols confirm the predicted Gaussian distribution. The symmetry function $\ln[P(\Delta s_{tot})/P(-\Delta s_{tot})]$ is linear with a slope close to unity, validating the DFT in the presence of memory and magnetic coupling.

Significance and Claims
The paper claims to provide the first explicit demonstration of the detailed fluctuation theorem for total entropy production in a non-Markovian setting involving crossed electric and magnetic fields. The significance lies in bridging the gap between Markovian fluctuation theorems and complex, memory-dependent environments (such as viscoelastic media or structured baths) where white-noise approximations fail.

The authors note that while the theoretical framework is general, the specific application to magnetically driven non-Markovian systems remains largely untested experimentally. They suggest that future work could involve confronting these predictions with experiments in viscoelastic environments where magnetic fields play a crucial role, though they do not propose specific experimental setups themselves. The work serves as a rigorous theoretical foundation for understanding irreversibility in structured, non-Markovian stochastic systems.