Semiclassical limit of quantum dynamics with rough Born-Oppenheimer potentials: regularity of molecular potential surfaces for general molecules and deterministic initial data

The semiclassical limit problem asks whether solutions of the Schrödinger equation $i \varepsilon \partial_t \psi^\varepsilon = -(\varepsilon^2/2) \Delta \psi^\varepsilon + U \psi^\varepsilon$ converge, as the rescaled Planck constant $\varepsilon \to 0$, to classical Hamiltonian dynamics in the same potential $U$. Ambrosio, Figalli, Friesecke, Giannoulis and Paul (Communications on Pure and Applied Mathematics, 2011) established this convergence for molecular Coulomb potentials $U = U_s + U_b$ with $U_b$ Lipschitz and $\nabla U_b$ of bounded variation, transport being understood in the DiPerna-Lions-Ambrosio sense of regular Lagrangian flows, but only for randomly averaged initial data ('almost all' initial data) and under a regularity hypothesis on the Born-Oppenheimer electronic ground-state surface $U_b$ that they could verify only for diatomic molecules. The mixed-state (density matrix) extension was settled by Figalli, Ligabò and Paul (2012) under a deterministic operator bound that excludes pure states, and the nonlinear self-consistent extension has been advanced, in different formulations, by the quantitative Hartree-to-Vlasov program of Lafleche, Saffirio, Chong, and Golse-Paul. The surviving open core is precise: (i) prove that the ground-state Born-Oppenheimer surface of a general molecule with three or more nuclei satisfies the BV-gradient regularity hypothesis, which requires new geometry of electronic eigenvalue crossing sets; (ii) establish semiclassical convergence for deterministic, fast-concentrating pure-state initial data in the BV framework, where classical trajectories through crossings are non-unique and known selection principles fail; (iii) as an auxiliary item, verify regularity of the entropic measures on spaces of probability measures (von Renesse and Sturm in dimension one, Sturm in general dimension, published 2026).

Background

Semiclassical analysis studies the limit $\varepsilon \to 0$ of the Schrödinger equation $i \varepsilon \partial_t \psi^\varepsilon = -(\varepsilon^2/2) \Delta \psi^\varepsilon + U \psi^\varepsilon$, where $\varepsilon$ is the rescaled Planck constant. The object connecting quantum states to classical phase space is the Wigner transform, a phase-space quasi-probability distribution associated to $\psi^\varepsilon$ (its nonnegative mollification, the Husimi transform, is the rigorous tool of the theory), whose weak limits — Wigner measures — were studied by Lions and Paul (1993): for smooth $U$ and well-prepared data, Wigner transforms converge to solutions of the classical Liouville equation $\partial_t \mu + p \cdot \nabla_x \mu - \nabla U(x) \cdot \nabla_p \mu = 0$ transported along Hamiltonian trajectories. For rough $U$ the underlying ordinary differential equation has no classical flow: DiPerna and Lions (1989) and Ambrosio (2004) showed that vector fields of bounded variation (BV: the distributional derivative is a finite measure) still define an essentially unique regular Lagrangian flow, a generalized flow unique up to Lebesgue-negligible sets of initial data whose push-forward of Lebesgue measure stays controlled by Lebesgue measure, and Ambrosio and Figalli (2010) extended well-posedness of the continuity equation to measure initial data. The physical motivation is molecular dynamics: by the Born-Oppenheimer approximation, the nuclei (positions $x \in \mathbb{R}^{3M}$ for a molecule with $M$ nuclei) move in the effective potential $U = U_s + U_b$, where $U_s$ is the explicit repulsive Coulomb term between nuclei and $U_b(x)$ is the lowest eigenvalue of the electronic Hamiltonian $H_{\mathrm{el}}(x)$, i.e. the electronic ground-state energy as a function of the nuclear positions (the ground-state potential energy surface); $U_b$ is smooth where the electronic ground level is nondegenerate — by the analytic perturbation theory of Kato and Hunziker, a nondegenerate eigenvalue depends analytically on the nuclear positions — but loses analyticity at eigenvalue crossings. Ambrosio, Friesecke and Giannoulis (2010) reduced the problem to this two-scale structure and proved the dimer case, in which symmetry leaves a one-parameter analytic family whose crossings form a discrete set. Ambrosio, Figalli, Friesecke, Giannoulis and Paul (Communications on Pure and Applied Mathematics, 2011) then proved the main theorem: for $U_b$ globally bounded and Lipschitz with $\nabla U_b \in BV_{\mathrm{loc}}$, the Husimi transforms of randomly averaged initial data — families $\psi^\varepsilon_{0,w}$ indexed by a reference probability measure $\mathbb{P}$ on the space of initial data, satisfying a no-concentration-in-mean condition equivalent to the uniform operator bound $(1/\varepsilon^n) \int_W \rho^{\psi^\varepsilon_{0,w}} \, d\mathbb{P}(w) \le C \, \mathrm{Id}$, where $\rho^\psi$ denotes the orthogonal projection onto $\psi$ and $W$ indexes the initial data — converge to the Liouville solution transported by the regular Lagrangian flow, and measure initial data are admitted via a regularity condition $\int \mu \, d\nu(\mu) \le C \mathcal{L}^d$ on the reference measure $\nu$ on the space of probability measures (the average of the initial data measures $\mu$ with respect to $\nu$ is dominated by a constant multiple of Lebesgue measure $\mathcal{L}^d$). The hypothesis $\nabla U_b \in BV_{\mathrm{loc}}$ was verified only for diatomic molecules, and the authors explicitly left open: the verification for general molecules, the extension to mixed states (density matrices evolving by the Heisenberg-von Neumann equation $i \varepsilon \partial_t \rho^\varepsilon = [H_\varepsilon, \rho^\varepsilon]$), the nonlinear self-consistent case $U = U_0 * |\psi|^2$, deterministic (non-averaged) initial data, and the regularity of the entropic measures of von Renesse and Sturm (a canonical probability measure on the space of probability measures, constructed rigorously in dimension one in 2009 and in general dimension by Sturm, published in 2026), which are the natural candidates for the reference measure.

Problem Statement

The open problem is to complete the semiclassical-limit theorem of Ambrosio, Figalli, Friesecke, Giannoulis and Paul for physically general molecular systems, in three linked components. (1) Regularity of Born-Oppenheimer surfaces for general molecules: prove that for a molecule with an arbitrary number $M \ge 2$ of nuclei (with positive charges) and finitely many electrons interacting through Coulomb forces, the electronic ground-state potential $U_b(x)$ (the lowest eigenvalue of the electronic Hamiltonian $H_{\mathrm{el}}(x)$ for each nuclear configuration $x \in \mathbb{R}^{3M}$) is globally bounded and Lipschitz with $\nabla U_b \in BV_{\mathrm{loc}}$, the regularity hypothesis the semiclassical-limit theorem requires; the expectation that $U_b$ is smooth on an open set of full Lebesgue measure of $\mathbb{R}^{3M} \setminus S$, where $S$ is the nuclear collision set (configurations where two or more nuclei coincide), is context for this component and does not by itself resolve it. This is currently established only for $M = 2$ (dimers), where symmetry reduces the problem to a one-parameter analytic family whose eigenvalue crossings form a discrete set; the core difficulty is the geometry of the crossing set where the electronic ground level becomes degenerate, which for $M \ge 3$ has codimension at least two and admits no analytic eigenvalue selection. (2) Deterministic initial data beyond random averaging: in the same potential class (repulsive Coulomb $U_s$ plus Lipschitz $U_b$ with $\nabla U_b \in BV_{\mathrm{loc}}$), establish semiclassical convergence, including the identification of the selected solution when the flow is non-unique, for the deterministic class of fast-concentrating pure states, namely coherent states (minimal-uncertainty wave packets) centered at a phase-space point, whose associated classical trajectories meet the crossing set where the flow is essentially non-unique. The mixed-state case is settled under the operator bound $\rho^\varepsilon \le C \varepsilon^n \, \mathrm{Id}$, which precisely excludes pure states; Paul's conjecture that normalized data have Husimi transforms unbounded in $L^\infty$ uniformly in $\varepsilon$ frames the obstruction. (3) Auxiliary: verify that the entropic measures on the space $\mathscr{P}(\mathbb{R}^d)$ of probability measures (von Renesse and Sturm in dimension one; Sturm's construction in general dimension) are regular in the sense of the source theorem, i.e. satisfy $\int \mu \, d\nu(\mu) \le C \mathcal{L}^d$ for some constant $C$. Components (1) and (2) are the load-bearing open core; (3) is an auxiliary expectation stated by the source authors. A complete resolution of component (1) or of component (2) resolves the task on its own; resolving (3) alone does not.

References

1. Luigi Ambrosio, Alessio Figalli, Gero Friesecke, Johannes Giannoulis, Thierry Paul. Semiclassical limit of quantum dynamics with rough potentials and well-posedness of transport equations with measure initial data. Communications on Pure and Applied Mathematics 64 (2011) 1199-1242. DOI 10.1002/cpa.20371; arXiv:1006.5388. https://doi.org/10.1002/cpa.20371
2. Luigi Ambrosio, Alessio Figalli, Gero Friesecke, Johannes Giannoulis, Thierry Paul. Corrigendum: Semiclassical limit of quantum dynamics with rough potentials and well-posedness of transport equations with measure initial data. Communications on Pure and Applied Mathematics 66 (2013) 646-647. DOI 10.1002/cpa.21440. https://doi.org/10.1002/cpa.21440
3. Luigi Ambrosio, Gero Friesecke, Johannes Giannoulis. Passage from quantum to classical molecular dynamics in the presence of Coulomb interactions. Communications in Partial Differential Equations 35 (2010) 1490-1515. DOI 10.1080/03605301003657835; arXiv:0907.1205. https://doi.org/10.1080/03605301003657835
4. Alessio Figalli, Marilena Ligabò, Thierry Paul. Semiclassical limit for mixed states with singular and rough potentials. Indiana University Mathematics Journal 61 (2012) 193-222. DOI 10.1512/iumj.2012.61.4670; arXiv:1012.2483. https://doi.org/10.1512/iumj.2012.61.4670
5. Thierry Paul. Recent results in semiclassical approximation with rough potentials. Survey for Microlocal Methods in Mathematical Physics and Global Analysis, Tübingen, June 2011. HAL hal-00623921. https://hal.science/hal-00623921
6. Agissilaos Athanassoulis, Thierry Paul. On the selection of the classical limit for potentials with BV derivatives. Journal of Dynamics and Differential Equations 25 (2013) 33-47. DOI 10.1007/s10884-012-9284-z. https://doi.org/10.1007/s10884-012-9284-z
7. François Golse, Thierry Paul. Semiclassical evolution with low regularity. Journal de Mathématiques Pures et Appliquées 151 (2021) 257-311. DOI 10.1016/j.matpur.2021.02.008. https://doi.org/10.1016/j.matpur.2021.02.008
8. François Golse, Thierry Paul. Time dependent quantum perturbations uniform in the semiclassical regime. Indiana University Mathematics Journal 72 (2023) 1273-1296. DOI 10.1512/iumj.2023.72.9363. https://doi.org/10.1512/iumj.2023.72.9363
9. Laurent Lafleche. Propagation of moments and semiclassical limit from Hartree to Vlasov equation. Journal of Statistical Physics 177 (2019) 20-60. DOI 10.1007/s10955-019-02356-7; arXiv:1809.04544. https://doi.org/10.1007/s10955-019-02356-7
10. Laurent Lafleche. Global semiclassical limit from Hartree to Vlasov equation for concentrated initial data. Annales de l'Institut Henri Poincaré C, Analyse non linéaire 38 (2021) 1739-1762. DOI 10.1016/j.anihpc.2021.01.004; arXiv:1902.08520. https://doi.org/10.1016/j.anihpc.2021.01.004
11. Laurent Lafleche, Chiara Saffirio. Strong semiclassical limits from Hartree and Hartree–Fock to Vlasov–Poisson equations. Analysis & PDE 16 (2023) 891-926. DOI 10.2140/apde.2023.16.891; arXiv:2003.02926. https://doi.org/10.2140/apde.2023.16.891
12. Jacky J. Chong, Laurent Lafleche, Chiara Saffirio. From many-body quantum dynamics to the Hartree–Fock and Vlasov equations with singular potentials. Journal of the European Mathematical Society 26 (2024) 4923-5007. DOI 10.4171/jems/1478; arXiv:2103.10946. https://doi.org/10.4171/jems/1478
13. Clotilde Fermanian-Kammerer, Patrick Gérard, Caroline Lasser. Wigner measure propagation and conical singularity for general initial data. Archive for Rational Mechanics and Analysis 209 (2013) 209-236. DOI 10.1007/s00205-013-0622-z. https://doi.org/10.1007/s00205-013-0622-z
14. Victor Chabu. Semiclassical analysis of the Schrödinger equation with conical singularities. Asymptotic Analysis 103 (2017) 165-220. DOI 10.3233/ASY-171423. https://doi.org/10.3233/ASY-171423
15. Victor Chabu. Semiclassical analysis of the Schrödinger equation with singular potentials. PhD thesis, Université Paris-Est, 2016. HAL tel-01531040. https://theses.hal.science/tel-01531040
16. Max-K. von Renesse, Karl-Theodor Sturm. Entropic measure and Wasserstein diffusion. The Annals of Probability 37 (2009) 1114-1191. DOI 10.1214/08-AOP430. https://doi.org/10.1214/08-AOP430
17. Karl-Theodor Sturm. Wasserstein diffusion on multidimensional spaces. The Annals of Probability 54 (2026), no. 2. DOI 10.1214/25-AOP1774; arXiv:2401.12721. https://doi.org/10.1214/25-AOP1774
18. Pierre-Louis Lions, Thierry Paul. Sur les mesures de Wigner. Revista Matemática Iberoamericana 9 (1993) 553-618. DOI 10.4171/RMI/143. https://doi.org/10.4171/RMI/143
19. R. J. DiPerna, P. L. Lions. Ordinary differential equations, transport theory and Sobolev spaces. Inventiones Mathematicae 98 (1989) 511-547. DOI 10.1007/BF01393835. https://doi.org/10.1007/BF01393835
20. Luigi Ambrosio. Transport equation and Cauchy problem for BV vector fields. Inventiones Mathematicae 158 (2004) 227-260. DOI 10.1007/s00222-004-0367-2. https://doi.org/10.1007/s00222-004-0367-2
21. Luigi Ambrosio, Alessio Figalli. Almost everywhere well-posedness of continuity equations with measure initial data. Comptes Rendus Mathématique 348 (2010) 249-252. DOI 10.1016/j.crma.2010.01.018. https://doi.org/10.1016/j.crma.2010.01.018
22. Dongfen Bian, Emmanuel Grenier, Laurent Laflèche, Quoc-Hung Nguyen. Quantum Landau damping with Coulomb repulsion in the whole space. Preprint arXiv:2609.02040 (2026). https://arxiv.org/abs/2609.02040