Machine learning-accelerated discovery and design of electrode materials and electrolytes for lithium ion batteries

Energy storage materials, Journal Year: 2024, Volume and Issue: 72, P. 103710 - 103710

Published: Aug. 13, 2024

Language: Английский

---

Uncertainty quantification by direct propagation of shallow ensembles

Machine Learning Science and Technology, Journal Year: 2024, Volume and Issue: 5(3), P. 035006 - 035006

Published: June 17, 2024

Abstract Statistical learning algorithms provide a generally-applicable framework to sidestep time-consuming experiments, or accurate physics-based modeling, but they introduce further source of error on top the intrinsic limitations experimental theoretical setup. Uncertainty estimation is essential quantify this error, and make application data-centric approaches more trustworthy. To ensure that uncertainty quantification used widely, one should aim for are accurate, also easy implement apply. In particular, including an existing architecture be straightforward, add minimal computational overhead. Furthermore, it manipulate combine multiple machine-learning predictions, propagating over modeling steps. We compare several well-established frameworks against these requirements, propose practical approach, which we dub direct propagation shallow ensembles, provides good compromise between ease use accuracy. present benchmarks generic datasets, in-depth study applications field atomistic machine chemistry materials. These examples underscore importance using formulation allows errors without making strong assumptions correlations different predictions model.

Language: Английский

---

i-PI 3.0: A flexible and efficient framework for advanced atomistic simulations

The Journal of Chemical Physics, Journal Year: 2024, Volume and Issue: 161(6)

Published: Aug. 14, 2024

Atomic-scale simulations have progressed tremendously over the past decade, largely thanks to availability of machine-learning interatomic potentials. These potentials combine accuracy electronic structure calculations with ability reach extensive length and time scales. The i-PI package facilitates integrating latest developments in this field advanced modeling techniques a modular software architecture based on inter-process communication through socket interface. choice Python for implementation rapid prototyping but can add computational overhead. In new release, we carefully benchmarked optimized several common simulation scenarios, making such overhead negligible when is used model systems up tens thousands atoms using widely adopted machine learning potentials, as Behler–Parinello, DeePMD, MACE neural networks. We also present features, including an efficient algorithm bosonic fermionic exchange, framework uncertainty quantification be conjunction infrastructure that allows deeper integration electronic-driven simulations, approach simulate coupled photon-nuclear dynamics optical or plasmonic cavities.

Language: Английский

---

Accurate description of ion migration in solid-state ion conductors from machine-learning molecular dynamics

Journal of Materials Chemistry A, Journal Year: 2024, Volume and Issue: 12(19), P. 11344 - 11361

Published: Jan. 1, 2024

Machine-learning molecular dynamics provides predictions of structural and anharmonic vibrational properties solid-state ionic conductors with ab initio accuracy. This opens a path towards rapid design novel battery materials.

Language: Английский

---

Influence of nano-crystallization on Li-ion conductivity in glass Li$$_3$$PS$$_4$$: a molecular dynamics study

Journal of Solid State Electrochemistry, Journal Year: 2024, Volume and Issue: 28(12), P. 4389 - 4399

Published: April 12, 2024

Abstract Understanding the ionic conduction mechanisms in solid electrolyte glasses and glass-ceramics is an important task for improving performance of next-generation all-solid-state batteries. Although many have been proposed, mechanism increased conductivity partially crystallized glass not fully understood. In this study, molecular dynamics was used to analyze strain local ion mobility around crystal nano-particles Li $$_3$$ 3 PS $$_4$$ 4 , which a promising material electrolytes. From analysis results, we find that field generated particles tensile decreases activation energy migration increases conductivity. This study opens possibility by controlling crystallization dispersing field, even though crystalline phase high conducting phase.

Language: Английский

---

Transport coefficients from equilibrium molecular dynamics

The Journal of Chemical Physics, Journal Year: 2025, Volume and Issue: 162(6)

Published: Feb. 13, 2025

The determination of transport coefficients through the time-honored Green–Kubo theory linear response and equilibrium molecular dynamics requires significantly longer simulation times than those properties while being further hindered by lack well-established data-analysis techniques to evaluate statistical accuracy results. Leveraging recent advances in spectral analysis current time series associated with trajectories, we introduce a new method estimate full (diagonal as well off-diagonal) Onsager matrix from single model. This approach, based on knowledge distribution Onsager-matrix samples frequency domain, unifies evaluation diagonal (conductivities viscosities) off-diagonal (e.g., thermoelectric) within comprehensive framework, improving reliability coefficient estimation for materials ranging molten salts solid-state electrolytes. We validate this against existing approaches using benchmark data cesium fluoride liquid water conclude our presentation computation various Li3PS4 electrolyte.

Language: Английский

---

Critical outlook on separator layers for solid-state lithium batteries: Solid electrolyte materials, anode interface engineering, & scalable separator production

Journal of Power Sources, Journal Year: 2025, Volume and Issue: 643, P. 237014 - 237014

Published: April 18, 2025

Language: Английский

---

Tracking Li atoms in real-time with ultra-fast NMR simulations

Published: April 16, 2024

We present for the first time a multiscale machine learning approach to jointly simulate atomic structure and dynamics with corresponding solid state Nuclear Magnetic Resonance (ssNMR) observables. study use-case of spin-alignment echo (SAE) NMR exploring Li-ion diffusion within electrolyte material Li3PS4 (LPS) by calculating quadrupolar frequencies 7Li. SAE probes long-range down microsecond-timescale hopping processes. Therefore only few force field schemes are able capture length scales required accurate comparison experimental results. By using new class interatomic potentials, known as ultra-fast potentials (UFPs), we efficiently access timescales beyond microsecond regime. In tandem, have developed model predicting full 7Li electric gradient (EFG) tensors in LPS. combining long timescale trajectories from UFP our EFG tensors, extract autocorrelation function (ACF) during Li diffusion. decay constants ACF both crystalline β-LPS amorphous LPS, find that predicted rates on same order magnitude those dynamics. This demonstrates potential finally make predictions experimentally relevant temperatures, opens avenue crystallography: dynamical simulations accessing polycrystalline glass ceramic materials.

Language: Английский

---

Thermal conductivity of Li3PS4 solid electrolytes with ab initio accuracy

Physical Review Materials, Journal Year: 2024, Volume and Issue: 8(6)

Published: June 12, 2024

The vast amount of computational studies on electrical conduction in solid-state electrolytes is not mirrored by comparable efforts addressing thermal conduction, which has been scarcely investigated despite its relevance to management and (over)heating batteries. reason for this lies the complexity calculations: one hand, diffusion ionic charge carriers makes lattice methods formally unsuitable, due lack equilibrium atomic positions needed normal-mode expansion. On other prohibitive cost large-scale molecular dynamics (MD) simulations heat transport large systems at ab initio levels hindered use MD-based methods. In paper, we leverage recently developed machine-learning potentials targeting different functionals (PBEsol, ${\mathrm{r}}^{2}\text{SCAN}$, PBE0) a state-of-the-art formulation Green-Kubo theory multicomponent compute conductivity promising electrolyte, ${\mathrm{Li}}_{3}{\mathrm{PS}}_{4}$, all polymorphs ($\ensuremath{\alpha}, \ensuremath{\beta}$, $\ensuremath{\gamma}$). By comparing MD estimates with low-temperature, nondiffusive $\ensuremath{\gamma}\ensuremath{-}{\mathrm{Li}}_{3}{\mathrm{PS}}_{4}$, highlight strong anharmonicities negligible nuclear quantum effects, hence further justifying even phases. Finally, ion-conducting $\ensuremath{\alpha}$ $\ensuremath{\beta}$ phases, where approach mandatory, our indicate weak temperature dependence conductivity, glass-like behavior effective local disorder characterizing these Li-diffusing

Language: Английский

---

Inorganic solid electrolytes for all-solid-state sodium/lithium-ion batteries: recent development and applications

Journal of Materials Chemistry A, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 14, 2024

This review provides a comprehensive overview of recent advancements in preparation techniques and electrolyte engineering. It also discusses the integration both single- multi-phase electrolytes ASSBs future research potentials.

Language: Английский

---