arXiv subjects

Benjamin X. Shi

Publications and source records attributed to Benjamin X. Shi.

At least 19 recordsLinked to original sources

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 {\AA}. 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

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 {\Delta}-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

How reproducible are first-principles simulations of liquid water?

Liquid water is fundamentally important, and its accurate computer simulation has been the driving force for myriad methodological developments. Ab initio molecular dynamics with forces obtained from density functional theory (DFT) is now a standard tool widely used by researchers. However, we reveal that previous studies of liquid water using the same widely-used density functional (revPBE-D3) exhibit significant discrepancies with one another, varying by over 20% in the diffusion coefficient and 10% in the density, raising fundamental questions about reproducibility. By combining modern long-range machine-learning interatomic potentials that enable robust statistical sampling with carefully converged DFT training data, we resolve these discrepancies, achieving consensus across six diverse community codes. Our predictions differ markedly from previous literature: we show that most previous results overestimate the density and underestimate the diffusion coefficient of revPBE-D3 water due to basis set incompleteness and pseudopotential inconsistencies, coupled with limitations in statistical sampling (in some cases). These benchmark values provide a reliable reference for validating current and future implementations of DFT-based ab initio molecular dynamics. Reaching agreement establishes confidence and credibility and serves as a prerequisite for the systematic assessment of new density functionals and numerical approximations.

physics.chem-ph↗

arXiv2026

Assessing the impact of nodal surface optimization in fixed-node diffusion Monte Carlo on non-covalent interactions

Diffusion quantum Monte Carlo (DMC) and coupled cluster theory [CCSD(T)] are widely-employed benchmark methods for noncovalent interactions (NCIs). However, recent studies have reported notable discrepancies across several hydrogen-bonded and dispersion-dominated systems, raising questions on the accuracy of the approximations underlying each approach. In DMC, the dominant error is expected to stem from the fixed-node approximation, where the nodal surface is typically taken from a single Slater determinant derived from a density functional theory or Hartree-Fock calculation. In this work, we assess the impact of nodal surface optimization on DMC predictions for 12 compounds spanning diverse NCIs, using a recently proposed antisymmetrized geminal power ansatz with natural orbitals. We find improved agreement with CCSD(T) for hydrogen-bonded systems, while having negligible effect for dispersion-dominated systems. These results provide a practical and computationally efficient route to resolving discrepancies in hydrogen-bonded interactions, while offering insight into the remaining differences in dispersion-dominated systems.

physics.chem-ph↗

arXiv2026

From Accurate Quantum Chemistry to Converged Thermodynamics for Ion Pairing in Solution

Quantitative prediction of thermodynamic properties in solution is essential for translating atomistic simulations into reliable chemical insight. As an exemplar system, the behaviour of CaCO$_3$ in water has been widely studied to understand its mineralization in seawater, with potential implications for carbon-capture strategies. However, making accurate computational predictions has been a long-standing challenge, requiring both highly accurate electronic structure methods and extensive statistical sampling. Here, we combine advances in machine learning and electronic structure theory to fully resolve the ion pairing free energy of CaCO$_3$ with explicit solvation. We show that achieving quantitative agreement with experiment requires going beyond the standard density functional theory up to the "gold-standard" coupled cluster theory with single, double, and perturbative triple excitations [CCSD(T)]. We generate a set of systematically improvable models, enabling reliable insights into the initial association mechanism of Ca and CO$_3$ ions prior to nucleation while fully quantifying enthalpic and entropic effects. Our results demonstrate that CCSD(T)-level thermodynamic predictions of complex aqueous systems can now be routinely achieved.

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

Practical and accurate density functionals for transition-metal heterogeneous catalysis

Density functional theory (DFT) underpins modern atomistic simulations of transition-metal surfaces. It can predict key properties linked to catalytic performance, such as adsorption energies and barrier heights, enabling new paradigms in rational catalyst design. These applications require reliable density functionals, however achieving transition-metal chemical accuracy (13 kJ/mol) on these properties remains challenging. We introduce a framework for designing new functionals tailored to catalytic processes on transition-metal surfaces, building on recent non-self-consistent approaches. Within this framework, we develop a hybrid and a double-hybrid functional that achieve unprecedented accuracy, with the latter reaching transition-metal chemical accuracy on average across 39 experimental adsorption reactions. In addition, both functionals demonstrate balanced performance for 17 barrier heights and correct qualitative failures of standard functionals, including CO adsorption on Pt(111) and graphene on Ni(111). They are computationally efficient, readily integrated into existing DFT codes, and supported by open-source workflows to facilitate adoption. More broadly, this framework provides a systematic route towards improved functionals for heterogeneous catalysis and complex materials.

cond-mat.mtrl-sci↗

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↗

arXiv2025

Towards Routine Condensed Phase Simulations with Delta-Learned Coupled Cluster Accuracy: Application to Liquid Water

Simulating liquid water to an accuracy that matches its wealth of available experimental data requires both precise electronic structure methods and reliable sampling of nuclear (quantum) motion. This is challenging because applying the electronic structure method of choice - coupled cluster theory with single, double and perturbative triple excitations [CCSD(T)] - to condensed phase systems is currently limited by its computational cost and complexity. Recent tour-de-force efforts have demonstrated that this accuracy can indeed bring simulated liquid water into close agreement with experiment using machine learning potentials (MLPs). However, achieving this remains far from routine, requiring large datasets and significant computational cost. In this work, we introduce a practical approach that combines developments in MLPs with local correlation approximations to enable routine CCSD(T)-level simulations of liquid water. When combined with nuclear quantum effects, we achieve agreement to experiments for structural and transport properties. Crucially, this approach extends beyond constant volume to constant pressure simulations, allowing fundamental properties such as the density to now be predicted by MLP-based CCSD(T) models. Importantly, the approach also handles constant pressure simulations, enabling MLP-based CCSD(T) models to predict isothermal-isobaric bulk properties, such as water's density maximum in close agreement with experiment. Encompassing tests across electronic structure, datasets and MLP architecture, this work provides a practical blueprint towards routinely developing CCSD(T)-based MLPs for the condensed phase.

physics.chem-ph↗

arXiv2025

Accurate and efficient machine learning interatomic potentials for finite temperature modeling of molecular crystals

As with many parts of the natural sciences, machine learning interatomic potentials (MLIPs) are revolutionizing the modeling of molecular crystals. However, challenges remain for the accurate and efficient calculation of sublimation enthalpies - a key thermodynamic quantity measuring the stability of a molecular crystal. Specifically, two key stumbling blocks are: (i) the need for thousands of ab initio quality reference structures to generate training data; and (ii) the sometimes unreliable nature of density functional theory, the main technique for generating such data. Exploiting recent developments in foundational models for chemistry and materials science alongside accurate quantum diffusion Monte Carlo benchmarks, offers a promising path forward. Herein, we demonstrate the generation of MLIPs capable of describing molecular crystals at finite temperature and pressure with sub-chemical accuracy, using as few as $\sim 200$ data structures; an order of magnitude improvement over the current state-of-the-art. We apply this framework to compute the sublimation enthalpies of the X23 dataset, accounting for anharmonicity and nuclear quantum effects, achieving sub-chemical accuracy with respect to experiment. Importantly, we show that our framework can be generalized to crystals of pharmaceutical relevance, including paracetamol and aspirin. Nuclear quantum effects are also accurately captured as shown for the case of squaric acid. By enabling accurate modeling at ambient conditions, this work paves the way for deeper insights into pharmaceutical and biological systems.

physics.comp-ph↗

arXiv2025

Reproducibility of fixed-node diffusion Monte Carlo across diverse community codes: The case of water-methane dimer

Fixed-node diffusion quantum Monte Carlo (FN-DMC) is a widely-trusted many-body method for solving the Schr\"{o}dinger equation, known for its reliable predictions of material and molecular properties. Furthermore, its excellent scalability with system complexity and near-perfect utilization of computational power makes FN-DMC ideally positioned to leverage new advances in computing to address increasingly complex scientific problems. Even though the method is widely used as a computational gold standard, reproducibility across the numerous FN-DMC code implementations has yet to be demonstrated. This difficulty stems from the diverse array of DMC algorithms and trial wave functions, compounded by the method's inherent stochastic nature. This study represents a community-wide effort to assess the reproducibility of the method, affirming that: Yes, FN-DMC is reproducible (when handled with care). Using the water-methane dimer as the canonical test case, we compare results from eleven different FN-DMC codes and show that the approximations to treat the non-locality of pseudopotentials are the primary source of the discrepancies between them. In particular, we demonstrate that, for the same choice of determinantal component in the trial wave function, reliable and reproducible predictions can be achieved by employing the T-move (TM), the determinant locality approximation (DLA), or the determinant T-move (DTM) schemes, while the older locality approximation (LA) leads to considerable variability in results. These findings demonstrate that, with appropriate choices of algorithmic details, fixed-node DMC is reproducible across diverse community codes-highlighting the maturity and robustness of the method as a tool for open and reliable computational science.

physics.comp-ph↗

arXiv2024

An accurate and efficient framework for modelling the surface chemistry of ionic materials

Quantum-mechanical simulations can offer atomic-level insights into chemical processes on surfaces. This understanding is crucial for the rational design of new solid catalysts as well as materials to store energy and mitigate greenhouse gases. However, achieving the accuracy needed for reliable predictions has proven challenging. Density functional theory (DFT), the workhorse quantum-mechanical method, can often lead to inconsistent predictions, necessitating accurate methods from correlated wave-function theory (cWFT). However, the high computational demands and significant user intervention associated with cWFT have traditionally made it impractical to carry out for surfaces. In this work, we address this challenge, presenting an automated framework which leverages multilevel embedding approaches, to apply accurate cWFT methods to the surfaces of ionic materials with computational costs approaching DFT. With this framework, we have reproduced experimental adsorption enthalpies for a diverse set of 19 adsorbate-surface systems. Moreover, we resolve debates on the adsorption configuration of several systems, while offering benchmarks to assess DFT. This framework is open-source, making it possible to more routinely apply cWFT to complex problems involving the surfaces of ionic materials.

physics.chem-ph↗

arXiv2024

Systematic discrepancies between reference methods for non-covalent interactions within the S66 dataset

The accurate treatment of non-covalent interactions is necessary to model a wide range of applications, from molecular crystals to surface catalysts to aqueous solutions and many more. Quantum diffusion Monte Carlo (DMC) and coupled cluster theory with single, double and perturbative triple excitations [CCSD(T)] are considered two widely-trusted methods for treating non-covalent interactions. However, while they have been well-validated for small molecules, recent work has indicated that these two methods can disagree by more than 7.5 kcal/mol for larger systems. The origin of this discrepancy remains unknown. Moreover, the lack of systematic comparisons, particularly for medium-sized complexes, has made it difficult to identify which systems may be prone to such disagreements and the potential scale of these differences. In this work, we leverage the latest developments in DMC to compute interaction energies for the entire S66 dataset, containing 66 medium-sized complexes with a balanced representation of dispersion and electrostatic interactions. Comparison to previous CCSD(T) references reveals systematic trends, with DMC predicting stronger binding than CCSD(T) for electrostatic-dominated systems, while the binding becomes weaker for dispersion-dominated systems. We show that the relative strength of this discrepancy is correlated to the ratio of electrostatic and dispersion interactions, as obtained from energy decomposition analysis methods. Finally, we have pinpointed model systems: the hydrogen-bonded acetic acid dimer (ID 20) and dispersion-dominated uracil-cyclopentane dimer (ID 42), where these discrepancies are particularly prominent. These systems offer cost-effective benchmarks to guide future developments in DMC, CCSD(T) as well as the wider electronic structure theory community.

physics.chem-ph↗

arXiv2024

Basis set incompleteness errors in fixed-node diffusion Monte Carlo calculations on non-covalent interactions

Basis set incompleteness error (BSIE) is a common source of error in quantum chemistry (QC) calculations, but it has not been comprehensively studied in fixed-node Diffusion Monte Carlo (FN-DMC) calculations. FN-DMC, being a projection method, is often considered minimally affected by basis set biases. Here, we show that this assumption is not always valid. While the relative error introduced by a small basis set in the total FN-DMC energy is minor, it can become significant in binding energy ($E_{\rm b}$) evaluations of weakly interacting systems. We systematically investigated BSIEs in FN-DMC-based binding energy ($E_{\rm b}$) evaluations using the A24 dataset, a well-known benchmark set of 24 non-covalently bound dimers. Contrary to common expectations, we found that BSIEs in FN-DMC evaluations of $E_{\rm b}$ are indeed significant when small localized basis sets, such as cc-pVDZ, are employed. We observed that BSIEs are larger in dimers with hydrogen-bonding interactions and smaller in dispersion-dominated interactions. We also found that augmenting the basis sets with diffuse orbitals, using counterpoise (CP) correction, or both, effectively mitigates BSIEs.

physics.chem-ph↗

arXiv2024

Nuclear quantum effects induce superionic proton transport in nanoconfined water

Recent work has suggested that nanoconfined water may exhibit superionic proton transport at lower temperatures and pressures than bulk water. Using first-principles-level simulations, we study the role of nuclear quantum effects in inducing this superionicity in nanoconfined water. We show that nuclear quantum effects increase the ionic conductivity of nanoconfined hexatic water, leading to superionic behaviour at lower temperatures and pressures than previously thought possible. Our work suggests that superionic water may be accessible in graphene nanocapillary experiments.

cond-mat.mtrl-sci↗

arXiv2024

Data-efficient fine-tuning of foundational models for first-principles quality sublimation enthalpies

Calculating sublimation enthalpies of molecular crystal polymorphs is relevant to a wide range of technological applications. However, predicting these quantities at first-principles accuracy -- even with the aid of machine learning potentials -- is a challenge that requires sub-kJ/mol accuracy in the potential energy surface and finite-temperature sampling. We present an accurate and data-efficient protocol based on fine-tuning of the foundational MACE-MP-0 model and showcase its capabilities on sublimation enthalpies and physical properties of ice polymorphs. Our approach requires only a few tens of training structures to achieve sub-kJ/mol accuracy in the sublimation enthalpies and sub 1 % error in densities for polymorphs at finite temperature and pressure. Exploiting this data efficiency, we explore simulations of hexagonal ice at the random phase approximation level of theory at experimental temperatures and pressures, calculating its physical properties, like pair correlation function and density, with good agreement with experiments. Our approach provides a way forward for predicting the stability of molecular crystals at finite thermodynamic conditions with the accuracy of correlated electronic structure theory.

cond-mat.mtrl-sci↗