arXiv subjects

Chiara Panosetti

Publications and source records attributed to Chiara Panosetti.

8 recordsLinked to original sources

arXiv2026

Learning to Converge: Warm-Starting DFTB Self-Consistent Charges with Machine Learning

Semiempirical electronic structure methods such as Density-Functional Tight-Binding (DFTB) offer a computationally efficient approach to molecular and materials simulations, bridging the gap between first-principles accuracy and classical force field speed while retaining full access to electronic properties. However, DFTB calculations based on self-consistent charge (SCC) schemes can still suffer from slow convergence, particularly for complex molecular and materials systems, making the iterative procedure a significant bottleneck in large-scale simulations and high-throughput workflows. We present a machine learning approach that accelerates DFTB simulations by predicting optimal initial atomic charges. Using element-specific models based on the Smooth Overlap of Atomic Positions descriptor and kernel ridge regression, we train charge models on reference calculations and demonstrate that ML-predicted initial charges consistently and significantly improve SCC convergence across diverse chemical systems including organic molecules, biomolecules, water clusters, transition metal oxides and solid electrolytes.

cond-mat.mtrl-sci↗

arXiv2026

Adaptive Slater Koster Parameters: Crossing Oxidation States with Density Functional Tight Binding

We propose to adapt the confined pseudo-atomic orbitals underpinning the precalculated Slater-Koster (SK) interaction tables in Density Functional Tight Binding (DFTB) to local atomic environments. We demonstrate significant improvement in electronic structure and energetics in the application to a partially oxidized Ni surface and Li insertion into graphite, where we assign optimal SK parameters to metal atoms in different oxidation states. Further analysis reveals the smoothness of the SK integrals across the varying oxidation states. Exploiting this, we introduce a site-resolved machine-learning scheme for fully adaptive DFTB. Using atomic descriptors and simple regression architectures already established in the context of machine-learning interatomic potentials, our scheme achieves 95% band-structure accuracy across all Ni-O binary compositions in the Materials Project.

cond-mat.mtrl-sci↗

arXiv2026

Determination of Density Functional Tight Binding Models for Cerium-containing Materials

We have developed Density Functional Tight Binding models for cerium and cerium oxide with generalized gradient approximation and hybrid functionals that accurately predict the electronic band structure of different cerium polymorphs as well as the insulating oxide. We show that determination of a many-body repulsion energy with the Chebyshev Interaction Model for Simulation correctly predicts cerium allotropic energetic ordering with a minimal training created from small-scale molecular dynamics calculations of a single phase, only. Our approach determines DFTB models for both electronic and material property predictions that retain a high degree of transferability.

cond-mat.mtrl-sci↗

arXiv2022

The intrinsic electrostatic dielectric behaviour of graphite anodes in Li-ion batteries -- across the entire functional range of charge

Lithium-graphite intercalation compounds (Li-GICs) are the most common anode material for modern Li-ion batteries. However, the dielectric response of this material in the electrostatic limit (and its variation with the state of charge (SOC)) has not been investigated to a satisfactory degree, especially not for the higher SOC range. Nevertheless, said dielectric behaviour is a highly desired property, particularly as an input parameter for charged kinetic Monte Carlo simulations, one of the most promising modeling techniques for energy materials. In this work, we employ our recent DFTB parametrization for Li-GICs based on a machine-learned repulsive potential in order to overcome the computational hurdles of sampling the long-ranged Coulomb interactions within this material, as experienced by the charge carriers within. This approach is rather novel due to computational cost, but best suited for investigating our specific property of interest. For the first time, we discover a mostly linear dependency of the relative permittivity on the SOC, from ca. 7 at SOC 0% to ca. 25 at SOC 100%. In doing so, we also present a straightforward approach that can be used in future research for other intercalation compounds -- once sufficiently fast and long-ranged computational methods, e.g. linear-scaling DFT, a good DFTB parametrization, or atomic potentials with inbuilt electrostatics, become available. However, while the presented qualitative behaviour is robust and our results compare favourably with the few experimental studies available, we stress that quantitative results are strongly dependent on our estimation of the partial charge transfer from the intercalated Li-ion to the carbon host structure, and need to be verified by further experiments and calculations. Yet, our research shows that in principle, two measurements -- one at low and one at high SOC -- should suffice for that purpose.

cond-mat.mtrl-sci↗

arXiv2021

Revisiting the storage capacity limit of graphite battery anodes: spontaneous lithium overintercalation at ambient pressure

The market quest for fast-charging, safe, long-lasting and performant batteries drives the exploration of new energy storage materials, but also promotes fundamental investigations of materials already widely used. Presently, revamped interest in anode materials is observed -- primarily graphite electrodes for lithium-ion batteries. Here, we focus on the upper limit of lithium intercalation in the morphologically quasi-ideal highly oriented pyrolytic graphite (HOPG), with a LiC$_6$ stoichiometry corresponding to 100\% state of charge (SOC). We prepared a sample by immersion in liquid lithium at ambient pressure and investigated it by static $^7$Li nuclear magnetic resonance (NMR). We resolved unexpected signatures of superdense intercalation compounds, LiC$_{6-x}$. These have been ruled out for decades, since the highest geometrically accessible composition, LiC$_2$, can only be prepared under high pressure. We thus challenge the widespread notion that any additional intercalation beyond LiC$_6$ is not possible under ambient conditions. We monitored the sample upon calendaric aging and employed ab initio calculations to rationalise the NMR results. The computed relative stabilities of different superdense configurations reveal that non-negligible overintercalation does proceed spontaneously beyond the currently accepted capacity limit.

cond-mat.mtrl-sci↗

arXiv2020

DFTB modelling of lithium intercalated graphite with machine-learned repulsive potential

Lithium ion batteries have been a central part of consumer electronics for decades. More recently, they have also become critical components in the quickly arising technological fields of electric mobility and intermittent renewable energy storage. However, many fundamental principles and mechanisms are not yet understood to a sufficient extent to fully realize the potential of the incorporated materials. The vast majority of concurrent lithium ion batteries make use of graphite anodes. Their working principle is based on intercalation---the embedding and ordering of (lithium-) ions in the two-dimensional spaces between the graphene sheets. This important process---it yields the upper bound to a battery's charging speed and plays a decisive role for its longevity---is characterized by multiple phase transitions, ordered and disordered domains, as well as non-equilibrium phenomena, and therefore quite complex. In this work, we provide a simulation framework for the purpose of better understanding lithium intercalated graphite and its behaviour during use in a battery. In order to address the large systems sizes and long time scales required to investigate said effects, we identify the highly efficient, but semi-empirical Density Funtional Tight Binding (DFTB) as a suitable approach and combine particle swarm optimization (PSO) with the machine learning (ML) procedure Gaussian Process Regression (GPR) to obtain the necessary parameters. Using the resulting parametrization, we are able to reproduce experimental reference structures at a level of accuracy which is in no way inferior to much more costly ab initio methods. We finally present structural properties and diffusion barriers for some exemplary system states.

physics.comp-ph↗

arXiv2019

A Practical Guide to Surface Kinetic Monte Carlo Simulations

This review article is intended as a practical guide for newcomers to the field of kinetic Monte Carlo (KMC) simulations, and specifically to lattice KMC simulations as prevalently used for surface and interface applications. We will provide worked out examples using the kmos code, where we highlight the central approximations made in implementing a KMC model as well as possible pitfalls. This includes the mapping of the problem onto a lattice and the derivation of rate constant expressions for various elementary processes. Example KMC models will be presented within the application areas surface diffusion, crystal growth and heterogeneous catalysis, covering both transient and steady-state kinetics as well as the preparation of various initial states of the system. We highlight the sensitivity of KMC models to the elementary processes included, as well as to possible errors in the rate constants. For catalysis models in particular, a recurrent challenge is the occurrence of processes at very different timescales, e.g. fast diffusion processes and slow chemical reactions. We demonstrate how to overcome this timescale disparity problem using recently developed acceleration algorithms. Finally, we will discuss how to account for lateral interactions between the species adsorbed to the lattice, which can play an important role in all application areas covered here.

physics.comp-ph↗

arXiv2016

Global Structure Search for Molecules on Surfaces: Efficient Sampling with Curvilinear Coordinates

Efficient structure search is a major challenge in computational materials science. We present a modification of the basin hopping global geometry optimization approach that uses a curvilinear coordinate system to describe global trial moves. This approach has recently been shown to be efficient in structure determination of clusters [Nano Letters 15, 8044-8048 (2015)] and is here extended for its application to covalent, complex molecules and large adsorbates on surfaces. The employed, automatically constructed delocalized internal coordinates are similar to molecular vibrations, which enhances the generation of chemically meaningful trial structures. By introducing flexible constraints and local translation and rotation of independent geometrical subunits we enable the use of this method for molecules adsorbed on surfaces and interfaces. For two test systems, trans-$β$-ionylideneacetic acid adsorbed on a Au(111) surface and methane adsorbed on a Ag(111) surface, we obtain superior performance of the method compared to standard optimization moves based on Cartesian coordinates.

cond-mat.mtrl-sci↗