arXiv subjects

Shenzhen Xu

Publications and source records attributed to Shenzhen Xu.

At least 19 recordsLinked to original sources

arXiv2026

Energy Partitioning in Dust-catalyzed $\mathrm{H_2}$ and HD Formation Revealed by Molecular Simulations Considering Nuclear Quantum Effects

Molecular hydrogen formation on interstellar dust grains is a key surface process in the interstellar medium, but the redistribution of the recombination energy between the substrate and the nascent molecule remains poorly understood. Here, we use ring-polymer molecular dynamics (RPMD) with a machine-learning force field to investigate energy partitioning during $\mathrm{H_2}$ and $\mathrm{HD}$ formation on graphene at $T=25, 50$ and $100 \mathrm{K}$. We focus on the chemisorbed-H recombination pathway previously identified as the dominant low-temperature channel on bare graphitic surfaces when nuclear quantum effects are included. The desorbing molecule retains the major fraction of the effective surface-mediated released energy, while graphene absorbs a smaller but non-negligible part. This molecular retention fraction is nearly temperature-independent over the investigated range. In contrast, the post-formation molecular kinetic-energy distribution changes more strongly with temperature: rovibrational motion dominates at low temperature, whereas center-of-mass translation becomes increasingly important at $100 \mathrm{K}$. $\mathrm{H_2}$ and $\mathrm{HD}$ exhibit broadly similar total energy retention, with only modest isotope-dependent differences in their internal kinetic-energy partitioning. These results provide an energy-resolved microscopic picture of surface-mediated energy redistribution in $\mathrm{H_2}$/$\mathrm{HD}$ formation, with implications for formation-pumping signatures in high-excitation $\mathrm{H_2}$ lines, vibrationally excited $\mathrm{H_2}$ chemistry, and collisional excitation of coexisting molecules by translationally hot nascent $\mathrm{H_2}$ in cold interstellar gas.

astro-ph.GA↗

arXiv2026

Lee-Yang Theory Guided Force Field Refinement Based on Phase Diagrams

We propose a general framework for automatic force field refinement guided by phase diagrams, grounded in Lee-Yang phase transition theory. The central idea is to directly use the partition function modulus as a phase-diagram-guided optimization target. Evaluating the modulus at points close to the real axis, where the Lee-Yang zeros are mostly associated with the phase transition, is more physically meaningful and avoids the numerical difficulty of explicitly solving for the zeros. This approach requires no system-specific order parameters or response properties for characterizing phase transition points, making it universal across various discontinuous phase transitions and material systems. We validate the method on refining parameters of a Lennard-Jones potential covering both gas-liquid and solid-liquid transitions, and a Cu embedded-atom method potential based on experimental melting curves. The refined force fields reproduce the target phase diagrams with significant improvement across all systems. For Cu, the refinement simultaneously improves predictions of enthalpy and heat capacity, which are observables beyond the optimization target. These results establish Lee-Yang theory as a practical tool for contemporary force field development.

cond-mat.stat-mech↗

arXiv2026

fix pimd/langevin: An Efficient Implementation of Path Integral Molecular Dynamics in LAMMPS

Path integral molecular dynamics (PIMD), which maps a quantum particle onto a fictitious classical system of ring polymers and propagates the "beads" of this extended classical system using molecular dynamics, is widely used to capture nuclear quantum effects (NQEs) in molecular simulations. Accurate PIMD calculations typically require a large number of beads and are therefore computationally demanding. While software packages such as i-PI offer comprehensive PIMD functionality, the high efficiency of simulations driven by machine learning interatomic potentials, such as Deep Potential (DP), calls for more efficient PIMD implementations that fully exploit modern massively parallel supercomputers. Here we present fix pimd/langevin, an efficient PIMD implementation in LAMMPS that supports commonly used features and leverages the Message Passing Interface architecture of LAMMPS to achieve high computational efficiency. We demonstrate the usage and validate the correctness of our code using liquid water as a representative example, and provide a comprehensive overview of the supported features. Then we discuss several important technical aspects of the implementation. Using DP simulations of water as a benchmark, we show that our implementation achieves several-fold acceleration compared to i-PI. Finally, we report strong and weak scaling results that demonstrate the favorable parallel performance of our code.

physics.chem-ph↗

arXiv2025

Adsorption of volatiles on dust grains in protoplanetary disks

The adsorption of volatile molecules onto dust grain surfaces fundamentally influences dust-related processes, including condensation of gas-phase molecules, dust coagulation, and planet formation in protoplanetary disks. Using advanced ab-initio density functional theory with r$^2$SCAN+rVV10 van der Waals functionals, we calculate adsorption energies of H$_2$, H$_2$O, and CO on carbonaceous (graphene, amorphous carbon) and silicate (MgSiO$_3$) surfaces. Results reveal fundamentally different adsorption mechanisms: weak physisorption on carbonaceous surfaces ($|\Delta\epsilon_{\rm ad}|\sim 0.1-0.2~{\rm eV}$) versus strong chemisorption on silicates ($|\Delta\epsilon_{\rm ad}|\sim 0.5-1.5~{\rm eV}$) via coordination bonds. Kinetic Monte Carlo simulations incorporating these energies demonstrate divergent surface evolution: carbonaceous grains exhibit distinct condensation radius compared to silicates, while the cocrystal of H$_2$O and CO significantly increases the desorption temperature of CO. The actual radii of gas-phase molecule depletion could thus be a comprehensive result of temperatures, chemical compositions, and even evolution tracks. Meanwhile, silicates maintain chemisorbed molecular coatings throughout most disk regions. Such dichotomy in surface coverage could also provide a natural mechanism for carbon depletion in inner planetary systems.

astro-ph.EP↗

arXiv2025

Automatic Refinement of Force Fields Based on Phase Diagrams

Exact characterization of phase transitions requires sufficient configurational sampling, necessitating efficient and accurate potential energy surfaces. Molecular force fields with computational efficiency and physical interpretability are desirable but challenging to refine for complex interactions. To address this, we propose a force field refinement strategy with phase diagrams as top-down optimization targets based on automatic differentiation. Using gas-liquid co-existence as a paradigm, we employ an enhanced sampling technique and design a differentiable loss function to evaluate force fields' depiction of phase diagrams. The refined force fields produce gas-liquid phase diagrams matching well with targets for two modeling systems, which confirms our approach as an effective automated force field development framework for phase transition studies.

cond-mat.stat-mech↗

arXiv2025

Interstellar Dust-Catalyzed Molecular Hydrogen Formation Enabled by Nuclear Quantum Effects

Molecular hydrogen (H$_2$) is one of the key chemical species that controls and shapes a wide spectrum of astrophysical processes from galaxy evolution to planet formation. Although catalyzation on dust grain surfaces is the dominant formation channel of H$_2$ in the interstellar medium, its efficiency across $20-200~\rm K$ has remained not fully understood. Here, using multiscale simulations combining ab-initio-level machine learning force fields, constrained path-integral Monte Carlo, and kinetic Monte Carlo, we perform a systematic, quantum-mechanical study of the full H$_2$ formation sequence, including hydrogen adsorption, diffusion, association and desorption. We explicitly consider the decoupling of gas and dust temperatures, making our results applicable to photon-dominated regions (PDRs) and dense cold clouds. Our results show that on the bare, crystalline surfaces studied here (graphitic and silicate grains), physisorbed hydrogen is negligible, and nuclear quantum effects (NQEs) in chemisorbed hydrogen atoms are essential for efficient formation at low temperatures, overcoming the classical Boltzmann suppression. This work presents a quantitative NQEs-inclusive study on silicate surfaces (exemplified by enstatite) and graphitic grains, revealing surface-specific adsorption behavior. These findings provide a first-principles quantum foundation for interstellar H$_2$ formation, complementing empirical multipliers, and enable new observational constraints on dust composition and molecular cloud evolution. The framework also extends to other astrochemical reactions on dust grains under full NQEs.

astro-ph.GA↗

arXiv2025

Consistent Modeling of Non-equilibrium Dust Sublimation and the Interactions with Dust Evolution in the Inner Regions of Protoplanetary Disks

The inner regions of protoplanetary disks are host to the sublimation of dust grains, a process traditionally modeled using equilibrium thermodynamics. We demonstrate through ab-initio density functional theory (DFT) and kinetic Monte Carlo (KMC) simulations that silicate dust sublimation is inherently a non-equilibrium kinetic process. The binding energies and vibrational frequencies governing desorption, calculated for MgSiO3 and other compositions, reveal that sublimation timescales far exceed local dynamical times, allowing grains to persist in a superheated state. This kinetic inhibition results in a broad, dynamic sublimation front whose location and morphology are strongly regulated by radial advection and dust coagulation. Our coupled simulations, integrating sublimation with advection and grain evolution, show that the front varies radially by a factor of four with accretion rate and exhibits a vertically stratified, bowl-shaped structure. These findings imply that the inner disk dust distribution, thermal structure, and subsequent planet formation are profoundly influenced by the kinematics and kinetics of dust grains, necessitating a departure from equilibrium prescriptions in disk models and interpretations of inner rim observations.

astro-ph.EP↗

arXiv2025

Electrochemistry-Enhanced Dynamic Paths Sampling Unveiling Nuclear Quantum Effects in Electrocatalysis

Proton-coupled electron transfers (PCET) are elementary steps in electrocatalysis. However, accurate calculations of PCET rates remain challenging, especially considering nuclear quantum effects (NQEs) under a constant potential condition. Statistical sampling of reaction paths is an ideal approach for rate calculations, however, is always limited by the rare-event issue. Here we develop an electrochemistry-driven quantum dynamics approach enabling realistic enhanced paths sampling under constant potentials without a priori defined reaction coordinates. We apply the method in modeling the Volmer step of the hydrogen evolution reaction, and demonstrate that the NQEs exhibit more than one order of magnitude impact on the computed rate constant, indicating an essential role of NQEs in electrochemistry.

physics.chem-ph↗

arXiv2025

Observing Nucleation and Crystallization of Rocksalt LiF from Molten State through Molecular Dynamics Simulations with Refined Machine-Learned Force Field

Lithium fluoride (LiF) is a critical component for stabilizing lithium metal anode and high-voltage cathodes towards the next-generation high-energy-density lithium batteries. Recent modeling study reported the formation of wurtzite LiF below about 550 K (J. Am. Chem. Soc. 2023, 145, 1327-1333), in contrast to experimental observation of rocksalt LiF under ambient conditions. To address this discrepancy, we employ molecular dynamics (MD) simulations with a refined machine-learned force field (MLFF), and demonstrate the nucleation and crystallization of rocksalt LiF from the molten phase at temperatures below about 800 K. The rocksalt phase remains stable in LiF nanoparticles. Complementary density functional theory (DFT) calculations show that dispersion interactions are essential for correctly predicting the thermodynamic stability of rocksalt LiF over the wurtzite phase on top of the commonly used PBE functional. Furthermore, we show that inclusion of virial stresses--alongside energies and forces--in the training of MLFFs is crucial for capturing phase nucleation and crystallization of rocksalt LiF under the isothermal-isobaric ensemble. These findings underscore the critical role of dispersion interactions in atomistic simulations of battery materials, where such effects are often non-negligible, and highlight the necessity of incorporating virial stresses during the training of MLFF to enable accurate modeling of solid-state systems.

cond-mat.mtrl-sci↗

arXiv2025

ABACUS: An Electronic Structure Analysis Package for the AI Era

ABACUS (Atomic-orbital Based Ab-initio Computation at USTC) is an open-source software for first-principles electronic structure calculations and molecular dynamics simulations. It mainly features density functional theory (DFT) and molecular dynamics functions and is compatible with both plane-wave basis sets and numerical atomic orbital basis sets. ABACUS serves as a platform that facilitates the integration of various electronic structure methods, such as Kohn-Sham DFT, stochastic DFT, orbital-free DFT, and real-time time-dependent DFT, etc. In addition, with the aid of high-performance computing, ABACUS is designed to perform efficiently and provide massive amounts of first-principles data for generating general-purpose machine learning potentials, such as DPA models. Furthermore, ABACUS serves as an electronic structure platform that interfaces with several AI-assisted algorithms and packages, such as DeePKS-kit, DeePMD, DP-GEN, DeepH, DeePTB, HamGNN, etc.

cond-mat.mtrl-sci↗

arXiv2024

Observing Li Nucleation at Li Metal-Solid Electrolyte Interface in All-Solid-State Batteries

Benefiting from the significantly improved energy density and safety, all-solid-state lithium batteries (ASSLBs) are considered one of the most promising next-generation energy technologies. Their practical applications, however, are strongly impeded by the Li dendrite formation. Despite this recognized challenge, a comprehensive understanding of Li dendrite nucleation and formation mechanism remains elusive. In particular, the initial locations of Li dendrite formation are still ambiguous: do Li clusters form directly at the Li anode surface, or inside the bulk solid electrolyte (SE), or within the solid-electrolyte interphase (SEI)? Here, based on the deep-potential molecular dynamics simulations combined with enhanced sampling techniques, we investigate the atomic-level mechanism of Li cluster nucleation and formation at the Li anode/SE interface. We observe that an isolated Li cluster initially forms inside the SEI between the Li6PS5Cl SE and the Li metal anode, located ~1 nm away from the Li anode/SEI boundary. The local electronic structure of the spontaneously formed SEI is found to be a key factor enabling the Li cluster formation within SEI, in which a significantly decreased bandgap could facilitate electronic conduction through the SEI and reduce Li+ ions to metallic Li atoms therein. Our work therefore provides atomic-level insights into Li-dendrite nucleation at anode/SE interfaces in ASSLBs, and could guide future design for developing Li-dendrite-inhibiting strategies.

cond-mat.mtrl-sci↗

arXiv2024

A Machine-Learning Accelerated Grand Canonical Sampling Framework for Nuclear Quantum Effects in Constant Potential Electrochemistry

Proton-coupled electron transfer (PCET) is the key step for energy conversion in electrocatalysis. Atomic-scale simulation acts as an indispensable tool to provide a microscopic understanding of PCET. However, consideration of the quantum nature of transferring protons under an exact grand canonical (GC) constant potential condition is a great challenge for theoretical electrocatalysis. Here, we develop a unified computational framework to explicitly treat nuclear quantum effects (NQEs) by a sufficient GC sampling, further assisted by a machine learning force field adapted for electrochemical conditions. Our work demonstrates a non-negligible impact of NQEs on PCET simulations for hydrogen evolution reaction at room temperature, and provides a physical picture that wave-like quantum characteristic of the transferring protons facilitates the particles to tunnel through classical barriers in PCET paths, leading to a remarkable activation energy reduction compared to classical simulations. Moreover, the physical insight of NQEs may reshape our fundamental understanding of other types of PCET reactions in broader scenarios of energy conversion processes.

cond-mat.mtrl-sci↗

arXiv2024

Observation of dendrite formation at Li metal-electrolyte interface: A machine-learning enhanced constant potential framework

Uncontrollable dendrites growth during electrochemical cycles leads to low Coulombic efficiency and critical safety issues in Li metal batteries. Hence, a comprehensive understanding of the dendrite formation mechanism is essential for further enhancing the performance of Li metal batteries. Machine learning accelerated molecular dynamics (MD) simulations can provide atomic-scale resolution for various key processes at an ab-initio level accuracy. However, traditional MD simulation tools hardly capture Li electrochemical depositions, due to lack of an electrochemical constant potential (ConstP) condition. In this work, we propose a ConstP approach that combines a machine learning force field with the charge equilibration method to reveal the dynamic process of dendrites nucleation at Li metal anode surfaces. Our simulations show that inhomogeneous Li depositions, following Li aggregations in amorphous inorganic components of solid electrolyte interphases, can initiate dendrites nucleation. Our study provides microscopic insights for Li dendrites formations in Li metal anodes. More importantly, we present an efficient and accurate simulation method for modeling realistic ConstP conditions, which holds considerable potential for broader applications in modeling complex electrochemical interfaces.

cond-mat.mtrl-sci↗

arXiv2023

A Spin-dependent Machine Learning Framework for Transition Metal Oxide Battery Cathode Materials

Owing to the trade-off between the accuracy and efficiency, machine-learning-potentials (MLPs) have been widely applied in the battery materials science, enabling atomic-level dynamics description for various critical processes. However, the challenge arises when dealing with complex transition metal (TM) oxide cathode materials, as multiple possibilities of d-orbital electrons localization often lead to convergence to different spin states (or equivalently local minimums with respect to the spin configurations) after ab initio self-consistent-field calculations, which causes a significant obstacle for training MLPs of cathode materials. In this work, we introduce a solution by incorporating an additional feature - atomic spins - into the descriptor, based on the pristine deep potential (DP) model, to address the above issue by distinguishing different spin states of TM ions. We demonstrate that our proposed scheme provides accurate descriptions for the potential energies of a variety of representative cathode materials, including the traditional Li$_x$TMO$_2$ (TM=Ni, Co, Mn, $x$=0.5 and 1.0), Li-Ni anti-sites in Li$_x$NiO$_2$ ($x$=0.5 and 1.0), cobalt-free high-nickel Li$_x$Ni$_{1.5}$Mn$_{0.5}$O$_4$ ($x$=1.5 and 0.5), and even a ternary cathode material Li$_x$Ni$_{1/3}$Co$_{1/3}$Mn$_{1/3}$O$_2$ ($x$=1.0 and 0.67). We highlight that our approach allows the utilization of all ab initio results as a training dataset, regardless of the system being in a spin ground state or not. Overall, our proposed approach paves the way for efficiently training MLPs for complex TM oxide cathode materials.

cond-mat.mtrl-sci↗

arXiv2018

Factors controlling oxygen interstitial diffusion in the Ruddlesden-Popper oxide La2-xSrxNiO4+delta

The development of Ruddlesden-Popper (RP) oxides as oxygen exchange and transport materials will benefit from mechanistic understanding of how oxygen transports through these materials. Using Density Functional Theory, we found there are two distinct oxygen interstitial diffusion mechanisms involving two different oxygen interstitial species that can be active in La2-xSrxNiO4+delta, and, we believe, in hyperstoichiometric RP oxides in general. The first mechanism is the previously proposed interstitialcy-mediated mechanism, which consists of diffusing oxide interstitials. The second mechanism is newly discovered in this work, and consists of both oxide and peroxide interstitial species. This mechanism exhibits a similar or possibly lower migration barrier than the interstitialcy mechanism for high oxidation states. Which mechanism contributes to the oxygen interstitial diffusion is the result of the change in relative stability between the oxide interstitial and peroxide interstitial. The stability of the oxide and peroxide, and the competition between the two oxygen diffusion mechanisms, is sensitive to the overall oxidation state of the system. Therefore, the oxygen diffusion mechanism is a function of the material composition, oxygen off-stoichiometry, operating temperature and oxygen partial pressure. We also examined the effect of epitaxial strain on both oxygen diffusion mechanisms, and found that tensile and compressive epitaxial strain of up to 2% had less than 100 meV/(% strain) effects on oxygen interstitial formation, migration, and activation energies, and that total achievable activation energy reductions are likely less than 100 meV for -2% to +2% epitaxial strain. The presented understanding of factors governing interstitial oxygen diffusion has significant implications for the engineering of RP oxides in numerous alternative energy technologies.

cond-mat.mtrl-sci↗

arXiv2017

Ab-initio modeling of electrolyte molecule Ethylene Carbonate decomposition reaction on Li(Ni,Mn,Co)O2 cathode surface

Electrolyte decomposition reactions on Li-ion battery electrodes contribute to the formation of solid electrolyte interphase (SEI) layers. These SEI layers are one of the known causes for the loss in battery voltage and capacity over repeated charge/discharge cycles. In this work, density functional theory (DFT)-based ab-initio calculations are applied to study the initial steps of the decomposition of the organic electrolyte component ethylene carbonate (EC) on the (10-14) surface of a layered Li(Nix,Mny,Co1-x-y)O2 (NMC) cathode crystal, which is commonly used in commercial Li-ion batteries. The effects on the EC reaction pathway due to dissolved Li+ ions in the electrolyte solution and different NMC cathode surface terminations containing absorbed hydroxyl -OH or fluorine -F species are explicitly considered. We predict a very fast chemical reaction consisting of an EC ring-opening process on the bare cathode surface, the rate of which is independent of the battery operation voltage. This EC ring-opening reaction is unavoidable once the cathode material contacts with the electrolyte because this process is purely chemical rather than electrochemical in nature. The -OH and -F adsorbed species display a passivation effect on the surface against the reaction with EC, but the extent is limited except for the case of -OH bonded to a surface transition metal atom. Our work implies that the possible rate-limiting steps of the electrolyte molecule decomposition are the reactions on the decomposed organic products on the cathode surface rather than on the bare cathode surface.

cond-mat.mtrl-sci↗

arXiv2017

Iron Partitioning between Ferropericlase and Bridgmanite in the Earth's Lower Mantle

Earth's lower mantle is generally believed to be seismically and chemically homogeneous because most of the key seismic parameters can be explained using a simplified mineralogical model at expected press-temperature conditions. However, recent high-resolution tomographic images have revealed seismic and chemical stratification in the middle-to-lower parts of the lower mantle. Thus far, the mechanism for the compositional stratification and seismic inhomogeneity, especially their relationship with the speciation of iron in the lower mantle, remains poorly understood. We have built a complete and integrated thermodynamic model of iron and aluminum chemistry for lower mantle conditions, and from this model has emerged a stratified picture of the valence, spin and composition profile in the lower mantle. Within this picture the lower mantle has an upper region with Fe3+ enriched bridgmanite with high-spin ferropericlase and metallic Fe, and a lower region with low-spin, iron-enriched ferropericlase coexisting with iron-depleted bridgmanite and almost no metallic Fe. The transition between the regions occurs at a depth of around 1600km and is driven by the spin transition in ferropericlase, which significantly changes the iron partitioning and speciation to one that favors Fe2+ in ferropericlase and suppresses Fe3+ and metallic iron formation These changes lead to lowered bulk sound velocity by 3-4% around the mid-lower mantle and enhanced density by ~1% toward the lowermost mantle. The predicted chemically and seismically stratified lower mantle differs dramatically from the traditional homogeneous model.

cond-mat.mtrl-sci↗

arXiv2017

Nanoscale Voltage Enhancement at Cathode Interfaces in Li-ion Batteries

Interfaces are ubiquitous in Li-ion battery electrodes, occurring across compositional gradients, regions of multiphase intergrowths, and between electrodes and solid electrolyte interphases or protective coatings. However, the impact of these interfaces on Li energetics remains largely unknown. In this work, we calculated Li intercalation-site energetics across cathode interfaces and demonstrated the physics governing these energetics on both sides of the interface. We studied the olivine/olivine-structured LixFePO4/LixMPO4 (x=0 and 1, M=Co, Ti, Mn) and layered/layered-structured LiNiO2/TiO2 interfaces to explore different material structures and transition metal elements. We found that across an interface from a high- to low-voltage material the Li voltage remains constant in the high-voltage material and decays approximately linearly in the low-voltage region, approaching the Li voltage of the low-voltage material. This effect ranges from 0.5-9nm depending on the interfacial dipole screening. This effect provides a mechanism for a high-voltage material at an interface to significantly enhance the Li intercalation voltage in a low-voltage material over nanometer scale. We showed that this voltage enhancement is governed by a combination of electron transfer (from low- to high-voltage regions), strain and interfacial dipole screening. We explored the implications of this voltage enhancement for a novel heterostructured-cathode design and redox pseudocapacitors.

cond-mat.mtrl-sci↗