Searcharxiv⌕ Search

arXiv subjects

Venkat Kapil

Publications and source records attributed to Venkat Kapil.

At least 19 recordsLinked to original sources
arXiv2026

GPU infrastructure for ML/MM with electrostatic embedding for enzyme chemistry

We present a fully GPU-native framework for machine learning/molecular mechanics (ML/MM) enhanced sampling simulations with electrostatic embedding, demonstrated on an enzymatic reaction. The framework combines a JAX port of the pretrained MACE-POLAR-1-M model, an electrostatic embedding scheme compatible with the MACE-POLAR architecture, and integration with JAX-MD and PLUMED-JAX for end-to-end compiled enhanced sampling. The implementation is developed using GPT-6.1-Sol within Codex, with reasoning effort set to high. Using replica exchange umbrella sampling, we calculate the free energy profile of nucleophilic substitution in haloalkane dehalogenase in approximately 200 H100 GPU hours, treating the entire protein and reactants with MACE-POLAR-1-M-JAX and most of the solvent with molecular mechanics.

physics.chem-ph↗
arXiv2026

PLUMED-JAX: GPU-native enhanced sampling for atomistic simulations

We present PLUMED-JAX, an AI-assisted adaptation of selected PLUMED collective variables, umbrella sampling, and metadynamics for JAX-based atomistic simulations, developed using gpt-6.1-sol with reasoning effort set to high within the Codex harness. Integration with JAX-MD supports compiled molecular dynamics, batched replicas, walkers sharing a bias, and execution across multiple graphics processing units. We validate the framework against PLUMED and exact free energy benchmarks, and demonstrate applications to butane, silicon crystallization, and lithium transport, with severalfold improvements in computational throughput for both classical forcefields and machine learning interatomic potentials.

physics.chem-ph↗
arXiv2026

Differentiable Electronic Grand Canonical Learning for Electrochemistry

Electronic grand-canonical density functional theory captures how structure and electronic charge respond to electrode potential and, together with statistical mechanics, how these responses give rise to emergent phenomena. However, its computational cost limits the system sizes and sampling times accessible to first- principles simulations. Here, we introduce a differentiable architecture that represents the grand potential and obtains forces and electronic charge as its derivatives, exactly satisfying the mixed-derivative constraints required for thermodynamic consistency, and implement it within the equivariant MACE framework as GC- MACE. We demonstrate more accurate joint learning across multiple electrochemical interfaces and improved transferability to thicker water films relative to previous models, while enabling constant-potential simulations at substantially larger length and time scales than accessible to ab initio molecular dynamics. With broad training data, the framework offers a route to thermodynamically consistent foundation interatomic potentials for the electronic grand canonical ensemble

physics.chem-ph↗
arXiv2026

Low-variance estimators for nuclear quantum mean forces

Quantum nuclear effects on the statistical properties of position-dependent operators can be rigorously captured by molecular dynamics on the learnt potential of mean force of a single path-integral replica. However, the primitive mean-force averages obtained from path integral molecular dynamics with that replica held fixed are noisy. We derive a "scaled gradient" estimator with zero variance for a one-dimensional harmonic oscillator. On our molecular benchmark systems, the Zundel cation at 100 K and 300 K and bulk water at 300 K we observe nearly 5 to 50 times reduction in sampling time. The resulting potentials can be trained on these forces like standard machine learning interatomic potentials and yield stable production simulations from which quantum expectation values of position-dependent observables are obtained as classical ensemble averages.

physics.chem-ph↗
arXiv2026

Isotope Effects at Classical Cost through Mass-Differentiable Machine Learning

Isotope effects govern fractionation and modulate reactivity, with applications from hydrogen energy to environmental science and catalysis, yet predicting them requires resolving small isotope-dependent free-energy differences that remain very challenging for conventional path-integral simulations in complex systems. Here we introduce iso-EPIGS, a path-integral coarse-graining framework built on a mass-differentiable neural network that reconstructs the mass- and temperature-dependent path integral centroid free-energy surface. Classical molecular dynamics on the learned surface yields rigorous isotope-resolved thermodynamics without explicit path-integral sampling. Benchmarks spanning gas, liquid, and crystalline phases, including liquid water and oxalic acid crystal, reproduce reference path-integral isotope free-energy differences, enthalpies, and lattice parameters at near classical computational cost. Crucially, iso-EPIGS trained solely on all-H and all-D isotopologues retains high accuracy for unseen partially deuterated isotopologues, demonstrating robust transferability across nuclear masses. Iso-EPIGS makes accurate isotope-effect simulations feasible for complex systems and lays the groundwork for foundation models of isotope effects across chemical space.

physics.chem-ph↗
arXiv2026

Dyna-Mat: End-to-end benchmarking of foundation machine learning interatomic potentials in finite-temperature ensembles

Foundation machine learning interatomic potentials (MLIPs) are increasingly being used as drop-in replacements for first-principles calculations, enabling simulations of materials at length and time scales that were previously inaccessible. However, due to lack of ground truth data, their accuracy on structural and dynamical observables in finite thermodynamic ensembles is yet to be established. Here, we introduce Dyna-Mat-v1.0, a benchmark dataset of condensed-phase first-principles molecular dynamics trajectories designed to test foundation MLIPs at realistic finite-temperature conditions. Using this dataset, we evaluate 15 foundation MLIPs across four model tiers by comparing both single-point energy and force errors on first-principles configurations and observables generated from MLIP-driven trajectories. We find that "on average" models with lower single-point force errors also yield lower errors for structural and dynamical observables. However, there are individual systems for which low force errors lead to qualitative failures in the predicted structure. Pressure remains poorly described across most models, pointing to limitations in the density functional theory stress labels available in current large-scale training datasets. Finally, we construct an accuracy-cost Pareto frontier to identify the best trade-offs for molecular dynamics with foundation MLIPs, finding that the latest generation of cross-trained models is close to Pareto-optimal according to the accuracy metrics considered here. Overall, Dyna-Mat-v1.0 shows that end-to-end finite-temperature validation is essential for quantifying the predictive behaviour of foundation MLIPs, and provides a simple, scalable route for assessing them beyond static and harmonic benchmarks relevant to materials design.

cond-mat.mtrl-sci↗
arXiv2026

Controlling the phase behaviour of ultraconfined water via bilayer graphene stacking

Water confined within nanoscale capillaries exhibits phase behaviour and transport properties that differ substantially from bulk, and these effects are commonly interpreted as consequences of geometric confinement and reduced dimensionality. Here we show that confinement topology alone is insufficient to predict the behaviour of nanoconfined water. Using machine learning interatomic potentials with first-principles accuracy, we compute the density-temperature phase diagram of water confined within bilayer graphene nanocapillaries and compare AA and AB stacking arrangements, which differ only by a lateral shift of 1.4 Å. Despite this minimal structural change, AA stacking can stabilise different ice polymorphs, can increase the melting temperature by more than 100 K, can enhance proton transfer, and alters the onset of superionic behaviour relative to AB stacking. We trace these effects to stacking-induced changes in the hydrogen-bond network associated with modifications to the lateral free energy landscape and neighbouring O-O separations. Our results demonstrate that even subtle atomistic variations in the confining walls can qualitatively reshape the physical and chemical behaviour of nanoconfined water, with implications for the interpretation and control of fluids under angstrom-scale confinement.

physics.chem-ph↗
arXiv2026

Six Open Questions in Machine-Learned Interatomic Potential Foundation Models

Machine-learned interatomic potentials (MLIPs) have had a profound impact on molecular modelling in recent years, promising to resolve the long-standing tension between the scale and accuracy of simulations. There has been a proliferation of new models and designs, and recently the paradigm of ``foundational'' MLIPs has become prevalent. Broadly speaking, foundation models are trained on large diverse datasets and promise to work well for new systems with minimal updates required. However, in such a new and fast moving field, there are many unanswered questions. In this article, we set out to articulate and explore what we see as the most important among these questions. We start by developing a working definition for foundational MLIPs and use this definition to frame the subsequent open questions. Despite the rapid progress in the field of MLIP models, we believe that these are fundamental questions which will continue to define cutting edge research in MLIPs in the years to come.

cond-mat.mtrl-sci↗
arXiv2026

Distilling first-principles accuracy into compact machine learning potentials for condensed-phase chemistry

Accurate machine learning interatomic potentials (MLIPs) have made first-principles-quality potential energy surfaces increasingly accessible for condensed-phase chemistry, but their inference cost can still limit the sampling needed to compute experimentally relevant observables. In this work, we combine transfer learning and knowledge distillation to construct compact "student" models that retain the accuracy of much larger "teacher" models obtained by applying transfer learning to foundation models. The resulting students reduce production simulation cost by roughly an order of magnitude, making high-accuracy sampling practical for challenging condensed-phase problems. We demonstrate this across three problems of increasing sampling complexity: finite-temperature NPT simulations of ice Ih, classical and path-integral simulations of liquid water over 240-370 K, and path-integral umbrella-sampling simulations of water dissociation at the anatase TiO2(101)/water interface. In all cases, the distilled students reproduce the target observables of their teachers more reliably than models of the same size trained directly on the limited reference data. The liquid-water student, distilled from a Δ-learned CCSD(T)-quality teacher, reproduces thermodynamic, structural, transport, and nuclear quantum properties over the full temperature range studied. At the TiO2/water interface, distillation makes PIMD umbrella sampling practical and shows that nuclear quantum effects lower the dissociation barrier by roughly 2 kcal/mol and shift the molecular-dissociated free energy difference into quantitative agreement with recent solid-state 17O NMR measurements. Our work demonstrates how knowledge distillation can make accurate MLIPs practical for the sampling methods needed to connect condensed-phase reaction thermodynamics with experiment, notably for interfacial chemistry and catalysis.

physics.chem-ph↗
arXiv2026

General Learning of the Electric Response of Inorganic Materials

We introduce \texttt{MACE-Field}, a field-aware, $O(3)$-equivariant interatomic potential that learns a single electric enthalpy functional $\mathcal F(\{\mathbf R\},\mathbf E)$ and obtains $\mathbf P$, $Z^*$, and $\boldsymbolα$ by exact differentiation. A uniform field couples to latent equivariant features inside the \texttt{MACE} backbone, while the scalar energy readout preserves Maxwell reciprocity, the acoustic sum rule, and crystal tensor symmetries by construction. Because this coupling is a plug-in on top of standard \texttt{MACE}, existing energy/force foundation models can be upgraded to become field-aware. Benchmarked against semilocal DFT/DFPT reference data, a directly trained cross-chemistry ferroelectric model reproduces the same-branch Berry-phase and spontaneous polarisations across diverse inorganic crystals. Starting from the multihead foundation model \texttt{mace-mp-mh-0} and its OMAT-PBE head, joint fine-tuning on dielectric, ferroelectric, and replay data yields \texttt{MACE-Field-MH-0} foundation models, which predict $Z^*$, $\boldsymbolα$, derived dielectric constants, and cross-chemistry polarisation trends with fidelity that captures branch-resolved polarisation and spontaneous-polarisation, while retaining strong force-field accuracy. Further, single-material \texttt{MACE-Field} models and \texttt{MACE-Field-MH-0} reproduce \ce{BaTiO3} hysteresis loops and $α$-quartz infrared, Raman, and dielectric spectra from finite-field molecular dynamics, comparable to DFPT. These results show that a simple, physics-informed field coupling can endow atomistic foundation models with transferable dielectric and ferroelectric response, while targeted single-material training remains advantageous for the most quantitative spectroscopic predictions.

cond-mat.mtrl-sci↗
arXiv2026

Rigorous Quantum Thermodynamics from Entropic Path Integral Coarse-Graining

Nuclear quantum effects (NQEs) remain a major challenge for molecular simulations, as rigorous treatment requires imaginary-time path-integral methods with heavy computational overhead. Neglecting NQEs leads to systematic errors in thermodynamic properties and failures in predicting isotope effects, quantum tunnelling, and anharmonic zero-point motion. Here, we introduce entropic path-integral coarse-graining (EPIGS), which enables rigorous quantum thermodynamics at the cost of classical simulations by training size- and temperature-transferable effective potentials utilising absolute centroid free energy and entropy. Central to EPIGS is an instanton-based free-energy perturbation scheme that enables efficient and accurate evaluation of the centroid free energy and entropy for large systems, making construction of the EPIGS training dataset practical. Benchmarks against full path-integral simulations on representative hydrogen-bonded systems, including liquid water, show that EPIGS reproduces quantum free energies and enthalpies within 0.2 meV/atom at near-classical computational cost. EPIGS provides a highly accurate, scalable and low-cost framework for quantum thermodynamic simulations of complex systems across temperatures.

physics.chem-ph↗
arXiv2026

Efficient first-principles modeling of complex molecular crystals at sub-chemical accuracy

Molecules can form myriad crystalline polymorphs, each with distinct properties affecting their performance across diverse applications, from pharmaceuticals to functional materials and more. Predicting the thermodynamically most stable polymorph from first principles remains a formidable challenge. It requires methods that scale to large, technologically-relevant molecules while achieving very high accuracy (below 1 kJ/mol) on relative lattice energies. Such accuracy, often termed sub-chemical accuracy, is generally beyond the reach of the workhorse density functional theory (DFT). In this work, we introduce a framework, combining advances in correlated wavefunction theory (cWFT) and the many-body expansion, to deliver accurate, cost-effective predictions of complex molecular crystals. For 23 organic molecules and 13 ice polymorphs, we predict crystal lattice energies to within experimental uncertainties at costs comparable to hybrid DFT, while being several orders of magnitude more efficient than previous cWFT approaches. We extend this approach to a set of large, drug-like molecules including axitinib and ROY, previously inaccessible to cWFT and where DFT is insufficient, achieving sub-chemical accuracy on the relative energies between challenging polymorphs. With the reference data generated throughout this work, we have been able to further parametrize a DFT functional with unprecedented accuracy aligning with our predictions. This cWFT framework as well as DFT functional are made openly available, providing new ranking tools to facilitate efficient high-throughput screening of molecular crystal polymorphs.

cond-mat.mtrl-sci↗
arXiv2026

MACE-POLAR-1: A Polarisable Electrostatic Foundation Model for Molecular Chemistry

Accurate modelling of electrostatic interactions and charge transfer is fundamental to computational chemistry, yet most machine learning interatomic potentials (MLIPs) rely on local atomic descriptors that cannot capture long-range electrostatic effects. We present a new electrostatic foundation model for molecular chemistry that extends the MACE architecture with explicit treatment of long-range interactions and electrostatic induction. Our approach combines local many-body geometric features with a non-self-consistent field formalism that updates learnable charge and spin densities through polarisable iterations to model induction, followed by global charge equilibration via learnable Fukui functions to control total charge and total spin. This design enables an accurate and physical description of systems with varying charge and spin states while maintaining computational efficiency. Trained on the OMol25 dataset of 100 million hybrid DFT calculations, our models achieve chemical accuracy across diverse benchmarks, with accuracy competitive with hybrid DFT on thermochemistry, reaction barriers, conformational energies, and transition metal complexes. Notably, we demonstrate that the inclusion of long-range electrostatics leads to a large improvement in the description of non-covalent interactions and supramolecular complexes over non-electrostatic models, including sub-kcal/mol prediction of molecular crystal formation energy in the X23-DMC dataset and a fourfold improvement over short-ranged models on protein-ligand interactions. The model's ability to handle variable charge and spin states, respond to external fields, provide interpretable spin-resolved charge densities, and maintain accuracy from small molecules to protein-ligand complexes positions it as a versatile tool for computational molecular chemistry and drug discovery.

physics.chem-ph↗
arXiv2026

Effective classical potential for quantum statistical averages

We present an effective potential that allows quantum thermal expectation values of a position-dependent observable to be estimated as a classical ensemble average of the corresponding function. We follow the approach of Feynman and Hibbs, but perform the mean-field treatment of quantum fluctuations about the path starting point rather than the path centroid. Furthermore, rather than performing a full variational optimization of the potential, we explore approximate functional forms that yield a numerical robustness. The resulting closed-form potential is exact in the classical and harmonic limits; benchmarks against exact position distributions for one-dimensional quartic, Morse, and double-well potentials, show good agreement for potentials with harmonic support.

quant-ph↗
arXiv2026

Harmonic-to-anharmonic thermodynamic integration made simple using REG TI

Standard harmonic-to-anharmonic thermodynamic integration (TI) is known to develop a near singularity in the integrand for solids exhibiting diffusive degrees of freedom, such as rotating functional groups or migrating defects. This pathology results in numerical challenges for estimating absolute free energies within a single thermodynamic cycle. In this work, we introduce a simple regularization that removes this singularity and yields a well-behaved integrand that can be accurately evaluated on a uniform grid. The approach -- termed Regularized End-point Gradient (REG) TI -- is demonstrated on a model system and on predicting the relative stability of paracetamol polymorphs for which quasi-free methyl rotations lead to a near singularity in standard TI. We expect REG TI to simplify anharmonic free energy calculations for solids and to potentially enable their automation.

physics.chem-ph↗
arXiv2025

On the role of nuclear quantum effects on the stability of peptides

Nuclear quantum effects (NQEs) arising from the light mass of hydrogen can influence the structure and stability of hydrogen-bonded biomolecules, yet their role in determining peptide and protein folding remains unclear. Experiments show that substituting H$_2$O with D$_2$O often stabilizes folded states, but the microscopic mechanism associated with this phenomena remains unresolved. Through ab initio-level path-integral molecular dynamics simulations enabled by machine-learning interatomic potentials, we address the fundamental question of the role of NQEs in peptides by investigating both their overall impact and isotope substitution effects. Overall, NQEs systematically destabilize compact three-dimensional structures across peptide systems, independent of secondary structure type or side-chain interactions. Contrary to the conventional picture that places central importance on hydrogen bonds, we find that the dominant destabilization instead arises from the quantum C-H vibrations. In addition, we reveal microscopic insights into the stabilization of folded peptides upon H$_2$O to D$_2$O substitution, showing that the H/D isotope substitution of active peptide hydrogens, previously considered unimportant, produces free-energy changes within the range of experimentally observed shifts. These findings provide a new interpretation of isotope effects in biological systems, indicating that seemingly small H$\to$D substitutions within peptides can be as important as, or even outweigh, solvent contributions.

physics.chem-ph↗
arXiv2025

Vibrational Spectra of Materials and Molecules from Partially-Adiabatic Elevated-Temperature Centroid Molecular Dynamics

Centroid molecular dynamics (CMD) incorporates nuclear quantum statistics into the calculation of vibrational spectra. However, when performed in Cartesian coordinates, CMD shows unphysical artifacts in certain vibrational bands, known as the curvature problem. Recent work showed that CMD spectra can be freed from the curvature problem by evolving the ring-polymer centroid on a potential of mean force (PMF) calculated at an elevated temperature ($T_e$-CMD). Here we present a partially-adiabatic implementation of $T_e$-CMD (PA-$T_e$-CMD), which eliminates the need for precomputed PMFs and instead yields the centroid force 'on the fly'. We introduce a two-temperature path-integral Langevin thermostat to achieve a temperature separation between centroid and internal modes of the ring polymer. Because it is paramount that the elevated temperature be chosen as low as possible for a given physical temperature in this formulation, we present a general scheme for its determination. We benchmark PA-$T_e$-CMD against exact vibrational spectra for the isolated water monomer and discuss its performance for challenging anharmonic systems: the carbonic acid fluoride molecule (CAF) and the methylammonium lead iodide perovskite (MAPI). We conclude that PA-$T_e$-CMD mitigates the curvature problem and the steep increase in computational cost with decreasing temperature of conventional path-integral methods. We observe energy leakage from the hot internal modes to high-frequency centroid modes in some cases, which, nevertheless, only compromises the spectral lineshapes at lower temperatures. While an adiabatic setup based on a coarse-grained centroid PMF is still preferable when a good pre-trained PMF can be easily obtained, PA-$T_e$-CMD presents a low-barrier single-shot setup for any system.

physics.chem-ph↗
arXiv2025

The Good, the Bad, and the Ugly of Atomistic Learning for "Clusters-to-Bulk" Generalization

Training machine learning interatomic potentials (MLIPs) on total energies of molecular clusters using differential or transfer learning is becoming a popular route to extend the accuracy of correlated wave-function theory to condensed phases. A key challenge, however, lies in validation, as reference observables in finite-temperature ensembles are not available at the reference level. Here, we construct synthetic reference data from pretrained MLIPs and evaluate the generalizability of cluster-trained models on ice-Ih, considering scenarios where both energies and forces and where only energies are available for training. We study the accuracy and data-efficiency of differential, single-fidelity transfer, and multi-fidelity transfer learning against ground-truth thermodynamic observables. We find that transferring accuracy from clusters to bulk requires regularization, which is best achieved through multi-fidelity transfer learning when training on both energies and forces. By contrast, training only on energies introduces artefacts: stable trajectories and low energy errors conceal large force errors, leading to inaccurate microscopic observables. More broadly, we show that accurate reproduction of microscopic structure correlates strongly with low force errors but only weakly with energy errors, whereas global properties such as energies and densities correlate with low energy errors. This highlights the need to incorporate forces during training or to apply careful validation before production. Our results highlight the promise and pitfalls of cluster-trained MLIPs for condensed phases and provide guidelines for developing - and critically, validating - robust and data-efficient MLIPs.

physics.chem-ph↗