https://arxiv.org/api/fBSdiNH3Ho2Jz91UygLSKPa0mLw 2026-06-13T20:15:47Z 27442 165 15 http://arxiv.org/abs/2605.30889v1 2026-05-29T06:25:47Z

Constructing production-quality machine-learned interatomic potentials (MLIPs) requires balancing accuracy, dynamical stability, and computational throughput under constraints that are not captured by a single training loss. We introduce MLIPilot, an auto-research framework in which tool-calling large language models propose hypotheses, edit MLIP training code, launch HPC jobs, and accept or revert changes using a fixed, physically constrained scorecard. We evaluate MLIPilot on MACE potential optimization using both commercial and open-weight LLM agents, including GPT-5.5, GPT-4.1, Mistral-24B, and Qwen3-32B. The benchmarks span molecular and periodic settings: a QM7-derived dataset for which we generated B3LYP/6-31G(d) energies and forces, and a Cu EMT dataset with periodic copper supercells labeled by ASE's Effective Medium Theory calculator. Across these benchmarks, the strongest agents move initially constraint-violating baselines to accepted models by discovering useful training strategies, including output normalization, loss-function changes, progressive training schedules, and model-capacity adjustments. These results suggest that LLM agents can serve as autonomous operators for scientific machine-learning workflows when their search is constrained by domain-specific validation criteria, shifting part of MLIP development from manual trial-and-error toward auditable, automated experimentation.

2026-05-29T06:25:47Z Etinosa Osaro Santosh Adhikari Stamatia Zavitsanou Kelsey Parker Dario Rocca http://arxiv.org/abs/2605.30872v1 2026-05-29T05:51:44Z

Roaming reactions occur when a molecule enters a near-dissociation region, avoids immediate separation, and later forms products by a pathway not controlled by the conventional tight transition-state bottleneck. Classical studies have shown that roaming is best understood in phase space: inner and outer transition-state structures, together with their invariant manifolds, organize trapping, return, and dissociation. The corresponding quantum question is less settled. Can a single quantum resonance carry a recognizable signature of the classical roaming region? We address this question in Chesnavich's two-degree-of-freedom model for the ion--molecule reaction $\mathrm{CH}_4^+\rightarrow\mathrm{CH}_3^+ + \mathrm{H}$. Resonance states are computed with a complex absorbing potential and analyzed using diagnostics designed to mirror the classical phase-space picture: radial probability weights derived from the tight and outer transition-state structures, radial Husimi projections, angular-momentum channel weights, and coherent-state probes of the classical periodic orbits. One resonance is distinguished from the rest of the computed resonance ensemble. Its wavefunction is concentrated in the projected region between the inner and outer transition-state structures, its radial phase-space distribution is centered at intermediate radius with nearly zero radial momentum, and its angular structure is consistent with a standing rather than a directed rotating component. We interpret this state as a phase-space-localized quantum analogue of classical roaming. The result provides a controlled example in which quantum roaming is identified directly from a resonance wavefunction and its phase-space diagnostics, rather than only from product-state or scattering signatures.

2026-05-29T05:51:44Z 26 pages, 5 figures Stephen Wiggins http://arxiv.org/abs/2605.30831v1 2026-05-29T04:37:00Z

Activity cliffs, structurally similar compounds with large potency differences, are widely treated as intrinsic features of chemical datasets. We argue that apart from target biology, much of our cliff understanding is a consequence of the geometry induced by the chosen molecular representation, not a property of a molecule pair itself. We designed a six-step pipeline to systematically test this hypothesis. The pipeline consists of: assessing pairwise distance geometry, cliff enrichment, activity gradient distribution, persistent homology of the cliff subspace, predictive benchmarking for a chosen pair of an embedding and a metric, and eventually, analysis of the matched molecular pairs and stereoisomers. We applied the pipeline to fifteen configurations of embeddings and metrics to build a benchmark across three distinctive datasets known of activity cliffs challenges. No representation excels on all criteria: Morgan Tanimoto provides the strongest cliff enrichment and cross-scaffold generalization; MolFormer cosine provides the only meaningful stereochemical sensitivity; MACCS and RDKit Dice fingerprints are most sensitive to matched-molecular-pair transformations; ChemBERTa fails uniformly due to embedding collapse. These findings are not a ranking. They reflect the fact that different representations encode different aspects of molecular recognition, and that choosing one implicitly defines what an activity cliff actually is.

2026-05-29T04:37:00Z Pawel Dabrowski-Tumanski Bartosz Topolski Dariusz Plewczynski Tomasz Jetka http://arxiv.org/abs/2509.07890v2 2026-05-28T21:29:24Z

Near a detailed-balance equilibrium, the perturbed mass-action dynamics of a chemical reaction network (CRN) map exactly onto an electrical-flow problem on the bipartite species-reaction graph: chemical potentials become electrical potentials, Onsager coefficients become conductances, and the instantaneous Gibbs free-energy consumption equals the dissipated electrical energy. We exploit this map to design quantum walk algorithms that decide species reachability, sample reachable species, approximate any individual steady-state reaction flux, and estimate the total Gibbs dissipation. The first three follow from standard electrical-flow quantum walks; the last is non-trivial because the chemical flow is not the minimum-energy electrical flow on the same graph. We resolve this via a new use of alternative neighbourhoods in multidimensional quantum walks, which forces the walker onto the mass-action flow whenever the network is $σ-M$ rigid. In an adjacency-matrix QRAM access model the algorithms achieve up to a quadratic speedup over classical methods -- for example $Ω(n^{3/2})$ vs $Ω(n^2)$ for reachability -- and dissipation-aware bounds tighten this further when the perturbation is concentrated.

2025-09-09T16:14:16Z Seenivasan Hariharan Sebastian Zur Sachin Kinge Lucas Visscher Kareljan Schoutens Stacey Jeffery http://arxiv.org/abs/2601.15662v2 2026-05-28T18:45:25Z

Chemically fueled supramolecular systems provide a versatile platform for generating nonequilibrium structures and dynamical instabilities, including chemical oscillations and traveling waves reminiscent of biological organization. However, a minimal mechanistic framework capable of capturing the emergence of such spatiotemporal order is still lacking. Here, we develop a minimal reaction-transport framework for fuel-driven supramolecular polymerization that couples activation-deactivation chemistry with cooperative assembly, fragmentation, and polymer length-dependent diffusion. The model captures autonomous oscillations arising through a Hopf bifurcation and demonstrates how temporal instabilities evolve into spatial self-organization upon inclusion of transport. We show that the nonlinear interplay between reaction kinetics and state-dependent mobility gives rise to traveling polymerization fronts, oscillatory wave dynamics, and complex spatiotemporal patterns. The propagating fronts exhibit near-ballistic dynamics, revealing a fundamentally nonequilibrium transport mechanism emerging from reactive feedback and dynamically evolving diffusivity. These findings establish a minimal physical framework connecting dissipative self-assembly, nonlinear transport, and active matter, while providing design principles for programmable supramolecular materials capable of autonomous spatiotemporal organization.

2026-01-22T05:22:11Z Akta Singh Nayana Mukherjee Jagannath Mondal Pushpita Ghosh http://arxiv.org/abs/2603.28373v4 2026-05-28T18:20:51Z

This perspective provides a succinct history of Fermi's golden rule (FGR), overview of its derivation, assumptions, and representative forms. Major applications of FGR, mostly in the field of chemical physics, are reviewed. These illustrate the broad applicability and success of FGR. Ambiguities and open issues encountered in practical applications of FGR are clarified. Recent advances in generalizations of FGR and computational methods for practical applications are addressed.

2026-03-30T12:40:21Z 26 pages, 6 figures Seogjoo J. Jang Goun Kim Young Min Rhee http://arxiv.org/abs/2606.11227v1 2026-05-28T16:50:15Z

Collective emission in light-harvesting assemblies is governed by the local transition dipole and finite geometry of emitting units, a fact that point-dipole approximation obscures. To go beyond this picture, we develop a non-Hermitian Hamiltonian using the quantum electrodynamic dyadic Green's tensor for a purple bacteria. We construct it for the isolated 24-bacteriochlorophyll conical frustum and its P42$_1$2 crystallographic assembly. The P42$_1$2 unit-cell symmetry is found to invert the bright-dark ordering of the single ring, placing subradiant states at the low-energy end and revealing the entire crystal to be the energy-harvesting entity. Tilt-driven switching is activated only in crystal geometries where the finite dipole-carrier (LH2) lies perpendicular to the growth plane. Vacancy and orientational disorder work only in cooperation to renormalize the switching threshold from higher polar angles to lower values.

2026-05-28T16:50:15Z Javed Akhtar Himangshu Prabal Goswami http://arxiv.org/abs/2605.30165v1 2026-05-28T16:18:49Z

The kinetic isotope effect (KIE) is the conventional probe for quantum tunneling, yet its composite nature conflates tunneling with zero-point energy and classical kinetics. Here, we introduce the tunneling phase diagram, a machine-learning framework that decouples true tunneling strength by decoding the nonlinear relationship between KIE and the tunneling factor (\k{appa}). With exceptional fidelity (R^2 > 0.98, RMSE = 0.21), this framework reveals an anomalous high KIE-low \k{appa} spanning 300-600 K, thereby defining a paradigm for the quantitative assessment of quantum tunneling.

2026-05-28T16:18:49Z 17 pages, 7 figures Xinrui Yang Zhigang Wang http://arxiv.org/abs/2605.00222v2 2026-05-28T15:54:09Z

Chemical reaction datasets such as USPTO suffer from substantial incompleteness, frequently missing byproducts, co-reactants, and stoichiometric coefficients. This limits their applicability and reliability in downstream applications. Here, we introduce CompleteRXN, a large-scale supervised benchmark for reaction completion under realistic missing-data conditions. We construct a dataset of aligned incomplete and atom-balanced reactions by mapping USPTO records to curated mechanistic reactions. We evaluate representative baselines, including a novel encoder-decoder reaction completion model with constrained decoding, the Constrained Reaction Balancer (CRB), and a recent algorithmic method, SynRBL. On our CompleteRXN benchmark, the CRB achieves high performance across splits of increasing difficulty, reaching 99.20% equivalence accuracy on the random split and 91.12% on the extreme out-of-distribution split. SynRBL produces many balanced and chemically plausible completions, but with lower accuracy on the benchmark test splits. Across all methods, performance degrades with increasing incompleteness. We observe a substantial drop when evaluating on reactions outside the benchmark (full uncurated USPTO), highlighting the gap between benchmark performance and practical robustness and motivating future work.

2026-04-30T20:53:29Z Gabriel Vogel Minouk Noordsij Evgeny Pidko Jana M. Weber http://arxiv.org/abs/2508.11311v2 2026-05-28T15:49:43Z

The squeezed photons, as a quantum-correlated light with reduced noise, have emerged as a great resource for sensing the structures of matter. Here we study the transient absorption (TA) scheme using the squeezed photons whose spectral correlation of amplitudes can be tailored. A microscopic theory is developed, revealing a highly time-energy-resolved nature of the signal that is not attainable by conventional TA scheme. Such a capability is elaborated by applying to monolayer transition metal dichalcogenide materials (TMDs), achieving a real-time monitoring of valley excitons and their dynamics. Moreover, we show the intermediate squeezing regime-not the strong squeezing-which the time-resolved spectroscopy is in favor of. Our work offers a new paradigm for studying nonequilibrium dynamics of matter, in light of the photocatalysis and optoelectronics.

2025-08-15T08:29:25Z 7 pages, 3 figures Jiahao Joel Fan Feihong Liu Dangyuan Lei Zhedong Zhang http://arxiv.org/abs/2605.18265v2 2026-05-28T15:06:22Z

Resonances from electron attachment to $\text{NO}_2$ have been identified in the total electron scattering cross sections (TCS) measured with a magnetically confined electron transmission apparatus. The corresponding absolute electron attachment cross sections have been derived from the TCS values through a self-consistent scattering channel deconvolution process. The positions of some of these resonances are consistent with previous elastic scattering calculations and dissociative electron attachment experiments. However, both the observed positions and the magnitude of the present resonance cross sections are in contradiction with the available recommended TCS values thus suggesting that electron scattering cross section databases need to be updated. Keywords: electron scattering resonances, negative ion formation, electron attachment cross sections

2026-05-18T12:00:21Z 18 pages, 2 figures and 2 tables. Published in Chemical Physics Letters Chemical Physics Letters 896 (2026) 142878 Ana I. Lozano Francisco Blanco Juan C. Oller Paulo Limão-Vieira Gustavo García 10.1016/j.cplett.2026.142878 http://arxiv.org/abs/2602.14975v3 2026-05-28T14:58:18Z

Following our previous work (J. Phys. Chem. Lett., 2026, 17, 5, 1288-1295), we propose the DMTS-NC approach, a distilled multi-time-step (DMTS) strategy using non-conservative (NC) forces to further accelerate atomistic molecular dynamics simulations using foundation neural network models such as FeNNix-Bio1. There, a dual-level reversible reference system propagator algorithm (RESPA) formalism couples a target accurate conservative potential to a simplified distilled representation optimized for the production of non-conservative forces. Despite being non-conservative, the distilled architecture is designed to enforce key physical priors, such as equivariance under rotation and cancellation of atomic force components. These choices facilitate the distillation process and therefore improve drastically the robustness of simulation, significantly limiting abnormal discrepancies between the two models, thus achieving excellent agreement with the forces data. Overall, the DMTS-NC scheme is found to be more stable and efficient than its conservative counterpart with additional speedups reaching 15-30% over DMTS. Requiring no fine-tuning steps, it is easier to implement and can be pushed to the limit of the systems physical resonances to maintain accuracy while providing maximum efficiency. We obtain additional speedup by combining hydrogen mass repartitioning (HMR), High Hydrogen Friction (HHF) to further extended the largest timestep up to 10fs of our schemes while conserving stability and accuracy. As for DMTS, DMTS-NC is applicable to any neural network potential and can be applied to approaches that are computationally heavier than FeNNix-Bio1. We show a proof of principle applying the approach to the distillation of MACE-OFF23 with consequent speedups ranging from 3.66 to 5.64 compared to single timestep.

2026-02-16T17:59:44Z Journal of Chemical Theory and Computation, 2026 Nicolaï Gouraud Côme Cattin Thomas Plé Olivier Adjoua Louis Lagardère Jean-Philip Piquemal 10.1021/acs.jctc.6c00653 http://arxiv.org/abs/2511.22324v3 2026-05-28T14:24:01Z

Quantum computing has the potential to transform simulations of quantum many-body problems at the heart of electronic structure theory. Efficient quantum algorithms to compute the eigenstates of fermionic Hamiltonians, such as quantum phase estimation, rely critically on high-accuracy initial state preparation. While several state preparation algorithms have been proposed for fermionic ground states, the preparation of excited states remains a major challenge, limiting the applicability of quantum algorithms to photochemistry and photophysics. In this contribution, we describe a physically motivated adiabatic state preparation technique for low-lying bright excited states using the explicit coupling between electrons and photons. Our approach systematically converges to the first bright excited state and can target different symmetry sectors by changing the photon polarization. We demonstrate the preparation of high-fidelity excited states for the Hubbard model and methylene molecule across a range of correlation regimes, and perform a successful hardware implementation for a model Hamiltonian.

2025-11-27T10:54:27Z 15 pages, 8 figures Hugh G. A. Burton Maria-Andreea Filip http://arxiv.org/abs/2512.24816v2 2026-05-28T13:58:57Z

We present a generalizable scale-bridging computational framework that enables predictive modeling of insertion-type electrode materials from atomistic to device scales. Applied to sodium manganese hexacyanoferrate, a promising cathode material for grid-scale sodium-ion batteries, our methodology employs an active-learning strategy to train a Moment Tensor Potential through iterative hybrid grand-canonical Monte Carlo--molecular dynamics sampling, robustly capturing configuration spaces at all sodiation levels. The resulting machine learning interatomic potential accurately reproduces experimental properties including volume expansion, operating voltage, and sodium concentration-dependent structural transformations, while revealing a four-order-of-magnitude difference in sodium diffusivity between the rhombohedral (sodium-rich) and tetragonal (sodium-poor) phases at 300 K. We directly compute all critical parameters -- temperature- and concentration-dependent diffusivities, interfacial and strain energies, and complete free-energy landscapes -- to feed them into pseudo-2D phase-field simulations that predict phase-boundary propagation and rate-dependent performances across electrode length scales. This multiscale workflow establishes a blueprint for rational computational design of next-generation insertion-type materials, such as battery electrode materials, demonstrating how atomistic insights can be systematically translated into continuum-scale predictions.

2025-12-31T12:04:38Z 26 pages, 14 figures Yuan-Chi Yang Eric Woillez Quentin Jacquet Ambroise van Roekeghem 10.1103/ydcg-9cy4 http://arxiv.org/abs/2605.29924v1 2026-05-28T13:36:33Z

Ideal density-functional approximations (DFAs) should account for dynamic, static, and nondynamic correlation. While common DFAs struggle with the latter two, the Ziegler-Rauk-Baerends-Daul multiplet sum method (MSM) provides a pragmatic way to include static correlation. In this article, we use diagrammatic MSM density-functional theory (diag MSM DFT) using the two-orbital two-electron model (TOTEM) to extend MSM DFT to include nondynamic correlation without relying on symmetry arguments. Building on previous formulations [A. Ponra, C. Bakasa, A.J. Etindele,and M.E. Casida, J. Chem. Phys. 159, 244306 (2023); M.E. Casida, A. Ponra, C. Bakasa, and A.J.Etindele, J. Chem. Phys. 162, 144317 (2025)] that lacked relaxation effects, this article incorporates relaxation via nonorthogonal configuration interaction (NOCI). We demonstrate that this modified diag MSM DFT produces an accurate ground-state potential energy curve (PEC) for lithium hydride (LiH), even at the ionic-to-open-shell-singlet avoided crossing characterized by significant charge transfer. This encouraging result suggests that the model can be extended to (at least) other singly and multiply-bonded diatomic molecules, while providing insight into a novel way to include strong correlation in DFT.

2026-05-28T13:36:33Z 31 pages, 15 Figures Mark E Casida Abraham Ponra Gadzikano Munyuki Bharathi Natarajan