arXiv:2610.16864
Exact Gaussian--Hermite closure of the mapping quantum--classical Liouville equation
Sahand Nikzat Ali Nassimi, Rony Modak Parai
Abstract
In the Meyer--Miller--Stock--Thoss mapping representation, the Liouville operator of the quantum--classical Liouville equation contains an excess coupling term in addition to the Poisson bracket. Neglecting this term yields the Poisson bracket mapping equation (PBME). We expand the mapping dependence of the density in Hermite functions about the Gaussian factor common to all physical mapping densities. In this basis, the excess coupling term acts purely as a raising operator and exactly cancels the raising contribution generated by the back--reaction force. Consequently, the exact dynamics does not generate higher Hermite orders and closes on a finite, sparse system of equations for the coefficient fields. Within the physical sector, these equations reduce to the subsystem quantum--classical Liouville equation. In contrast, the PBME retains the raising contribution of the force, modifying the back reaction within the physical sector while continuously generating higher Hermite orders. For Tully's dual avoided crossing model, direct propagation of the coefficient equations confirms this finite closure and agrees with independent solutions of the subsystem equation. Comparison with converged full quantum dynamics further shows that, across all regimes examined, the PBME is less accurate than the quantum--classical description for populations, coherences, and momentum distributions.
AI Summary
The paper asks how the excess-coupling term in the Meyer–Miller–Stock–Thoss mapping quantum–classical Liouville equation (QCLE) controls closure, and what is lost when it is omitted in the Poisson bracket mapping equation (PBME). Expanding the mapping density in Gaussian-weighted Hermite polynomials shows that the excess term raises Hermite order by two and exactly cancels the raising contribution of the subsystem back-reaction force. The remaining operator preserves or lowers order, making every space with maximum order $K$ invariant. This yields finite, sparse equations for coefficient fields on bath phase space. Physical densities occupy an invariant subspace at orders zero and two, where these equations reproduce the subsystem QCLE. Unlike the previously established physical-sector invariance, the closure applies to arbitrary finite Hermite support. PBME retains the raising contribution, both modifying back reaction within the physical sector and generating fourth and higher orders. Symbolic checks and direct coefficient propagation test this structure. Benchmarks use Tully’s two-state dual avoided crossing with one bath coordinate and six incident momenta, $P_0=10,20,30,40,70,100$, alongside independent QCLE solvers and converged quantum dynamics. PBME is less accurate for populations, coherences and momentum distributions throughout these tests, though not for every phase-space density metric. A fourth-order leakage rate does not quantitatively predict PBME error. The closure generalizes to arbitrary subsystem and bath sizes, but grid propagation remains restricted to low-dimensional benchmarks and costs more than solving the subsystem QCLE directly. Numerical error comparisons concern only this model; the supplied text leaves numerical error magnitudes unresolved as LaTeX macros.
A large language model, Anthropic's Claude, model Opus 5.5, was used to help with writing the manuscript and checking the simulation codes.
Error found: The supplied manuscript omits inspectable numerical results through unresolved macros and empty figures, preventing quantitative verification of its simulation-based closure checks and PBME-versus-QCLE accuracy comparisons.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.