A5032671940- Machine Learning in Materials Science
- Advanced Battery Materials and Technologies
- Spectroscopy and Quantum Chemical Studies
- Theoretical and Computational Physics
- Material Dynamics and Properties
- nanoparticles nucleation surface interactions
- Advancements in Battery Materials
- Advanced battery technologies research
- Electrochemical Analysis and Applications
- Perovskite Materials and Applications
- Electrocatalysts for Energy Conversion
- Supercapacitor Materials and Fabrication
- Phase Equilibria and Thermodynamics
- Conducting polymers and applications
- Freezing and Crystallization Processes
- Advanced Physical and Chemical Molecular Interactions
- Protein Structure and Dynamics
- Liquid Crystal Research Advancements
- Anomaly Detection Techniques and Applications
- Glass properties and applications
- Crystallography and molecular interactions
- Quantum Dots Synthesis And Properties
- Quantum many-body systems
- Scientific Computing and Data Management
- Advanced Thermodynamics and Statistical Mechanics
University of Cambridge
2016-2024
University of California, Berkeley
2021-2023
Lawrence Berkeley National Laboratory
2020-2023
Kavli Energy NanoScience Institute
2021
Machine-learned force fields have transformed the atomistic modelling of materials by enabling simulations ab initio quality on unprecedented time and length scales. However, they are currently limited by: (i) significant computational human effort that must go into development validation potentials for each particular system interest; (ii) a general lack transferability from one chemical to next. Here, using state-of-the-art MACE architecture we introduce single general-purpose ML model,...
Ion adsorption at solid–water interfaces is crucial for many electrochemical processes involving aqueous electrolytes including energy storage, separations, and electrocatalysis. However, the impact of hydronium (H3O+) hydroxide (OH–) ions on ion surface charge distributions remains poorly understood. Many fundamental studies supercapacitors focus non-aqueous to avoid addressing role functional groups electrolyte pH in altering uptake. Achieving microscopic level characterization interfacial...
By adopting a perspective informed by contemporary liquid-state theory, we consider how to train an artificial neural network potential describe inhomogeneous, disordered systems. We find that potentials based on local representations of atomic environments are capable describing some properties liquid-vapor interfaces but typically fail for depend unbalanced long-ranged interactions build up in the presence broken translation symmetry. These same cancel translationally invariant bulk,...
We apply a scientific machine learning (ML) framework to aid the prediction and understanding of nanomaterial formation processes via joint spectral-kinetic model. this study nucleation growth two-dimensional (2D) perovskite nanosheets. Colloidal nanomaterials have size-dependent optical properties can be observed in situ, all which make them good model for complex nucleation, growth, phase transformation 2D perovskites. Our results demonstrate that form through two at nanoscale: either...
Ion adsorption at solid–water interfaces is the key process that underlies for many electrochemical processes including energy storage, separations, and electrocatalytic applications involving aqueous electrolytes. However, impact of hydronium (H3O+) hydroxide (OH–) ions on ion surface charge distributions are poorly understood, fundamental studies double layer capacitors focus non-aqueous electrolytes to avoid addressing role functional groups pH in altering uptake. This particularly true...
Relaxation times and transport processes of many glass-forming supercooled liquids exhibit a super-Arrhenius temperature dependence. We examine this phenomenon by computer simulation the Lewis-Wahnström model for ortho-terphenyl. propose microscopic definition single-molecule cage-breaking transition show that, when correlation behaviour is taken into account, these rearrangements are sufficient to reproduce correct translational diffusion constants over an intermediate range in regime. that...
Nanoelectrochemical devices have become a promising candidate technology across various applications, including sensing and energy storage, provide new platforms for studying fundamental properties of electrode/electrolyte interfaces. In this work, we employ constant-potential molecular dynamics simulations to investigate the impedance gold-aqueous electrolyte nanocapacitors, exploiting recently introduced fluctuation-dissipation relation. particular, relate frequency-dependent these...
The emergence of observable properties from the organisation underlying potential energy landscape is analysed, spanning a full range complexity self-organising to glassy and jammed systems. examples include atomic molecular clusters, β-barrel protein, GNNQQNY peptide dimer, models condensed matter that exhibit structural glass formation jamming. We have considered measures based on several different properties, namely, Shannon entropy, an equilibrium thermodynamic measure uses sample local...
Finding the optimal alignment between two structures is important for identifying minimum root-mean-square distance (RMSD) them and as a starting point calculating pathways. Most current algorithms aligning are stochastic, scale exponentially with size of structure, performance can be unreliable. We present complementary methods corresponding to isolated clusters atoms condensed matter described by periodic cubic supercell. The first method (Go-PERMDIST), branch bound algorithm, locates...
Organic molecules can be stable in distinct crystalline forms, known as polymorphs, which have significant consequences for industrial applications. Here, we predict the polymorphs of benzene computationally an accurate anisotropic model parametrized to reproduce electronic structure calculations. We adapt basin-hopping global optimization procedure case unit cells, simultaneously optimizing molecular coordinates and cell parameters locate multiple low-energy structures from a variety...
We study the dynamical behaviour of a computer model for viscous silica, archetypal strong glass former, and compare its diffusion mechanism with earlier studies fragile binary Lennard-Jones liquid. Three different methods analysis are employed. First, temperature time scale dependence constant is analysed. Negative correlation particle displacements influences transport properties in silica as well liquids. suggest that difference between Arrhenius super-Arrhenius diffusive results from...
Using molecular dynamics simulations and methods of importance sampling, we study the thermodynamics sodium chloride in aqueous premelting layer formed spontaneously at interface between ice its vapor. We uncover a hierarchy time scales that characterize relaxation this system, spanning picoseconds ionic motion to tens or hundreds nanoseconds associated with fluctuations liquid–crystal their presence. find ions distort both local interfaces, incurring restoring forces result preferentially...
Coarse-grained models developed for polycyclic aromatic hydrocarbons based on the Paramonov–Yaliraki potential have been employed to investigate finite temperature thermodynamics, out-of-equilibrium dynamics, energy landscapes, and rearrangement pathways of coronene octamer.
We use energy landscape methods to investigate the response of a supercooled liquid random pinning. classify structural similarity different minima using measure overlap. This analysis reveals correspondence between distinct particle packings (which are characterised via overlap) and funnels on disconnectivity graphs). As number pinned particles is increased, we find crossover from glassy behavior at low pinning structure-seeking high pinning, in which all thermally accessible structurally...
We present a theoretical framework to explain how interactions between redox mediators and electrolyte components impact electron transfer kinetics, thermodynamics, catalytic efficiency. Specifically focusing on ionic association, we use DBBQ as case study demonstrate these effects. Our analytical equations reveal observed potential rate constants evolve with Li+ concentration, evidencing different activity mechanisms. Experimental validation shows that bounds 3 ions in its reduced state 1...
The kinetics and thermodynamics of the electrochemical reactions redox mediators for lithium–air batteries depend on ionic association strength with Li + ions specific pathways, potentially affecting energetic efficiency devices.
With the emergence of Foundational Machine Learning Interatomic Potential (FMLIP) models trained on extensive datasets, transferring data between different ML architectures has become increasingly important. In this work, we examine extent to which training optimised for one machine-learning forcefield algorithm may be re-used train models, aiming accelerate FMLIP fine-tuning and reduce need costly iterative training. As a test case, an organic liquid mixture that is commonly used as solvent...
Nanoelectrochemical devices have become a promising candidate technology across various applications, including sensing and energy storage, provide new platforms for studying fundamental properties of electrode/electrolyte interfaces. In this work, we employ constant-potential molecular dynamics simulations to investigate the impedance gold-aqueous electrolyte nanocapacitors, exploiting recently-introduced fluctuation-dissipation relation. particular, relate frequency-dependent these...
We apply a scientific machine learning framework to aid the prediction and understanding of nanomaterial formation processes via joint spectral-kinetic model. this study nucleation growth 2D perovskite nanosheets. Colloidal nanomaterials have size-dependent optical properties, can be observed in situ, all which make them good model for complex nucleation, phase transformation perovskites. Our results demonstrate that form through two at nanoscale: either layer-by-layer chemical exfoliation...
Lithium-air batteries are appealing candidates for high-energy electric vehicle applications because they have a theoretical capacity 3 to 5 times higher than li-ion batteries. They composed of lithium anode and porous conductive matrix (usually carbon-based), where oxygen is reduced during discharge oxidized charge. One the major issues high overpotentials needed recharge battery due isolating nature solid product, peroxide. The use redox mediators (RMs) charge reactions has gained...
Li-O 2 batteries are very promising candidates for electromobility because of their high theoretical energy density. Moreover, they composed abundant materials, as opposed to Li-ion batteries, which would make them a greener and cheaper alternative. The principal reaction in is the formation lithium peroxide during discharge via oxygen reduction (ORR), that then oxidised charge by evolution (OER). One major challenges related insulating nature product, significantly limits capacity hinders...
Using molecular dynamics simulations and methods of importance sampling, we study the thermodynamics sodium chloride in aqueous premelting layer formed spontaneously at interface between ice its vapor. We uncover a hierarchy timescales that characterize relaxation this system, spanning picoseconds ionic motion to 10s-100s nanoseconds associated with fluctuations liquid-crystal their presence. find ions distort both local interfaces, incurring restoring forces result preferentially residing...