arXiv subjects

Lukáš Grajciar

Publications and source records attributed to Lukáš Grajciar.

8 recordsLinked to original sources

arXiv2026

Synthetic pre-training of graph-network models for predicting solid-state NMR parameters

Nuclear magnetic resonance (NMR) is a powerful probe of atomic structure, but accurate quantum-mechanical predictions of tensorial NMR parameters are computationally demanding. This creates a bottleneck both for direct quantum-mechanical studies and for collecting high-quality training data for machine-learning (ML) models. Here, we introduce a synthetic pre-training and fine-tuning protocol for graph-based ML models of solid-state NMR parameters. We first pre-train models on synthetic tensorial data, as obtained using an existing ML model, and subsequently fine-tune those models on new ground-truth data. We observe a pronounced improvement in data efficiency when pre-training and fine-tuning span the same compositional and configurational space, and we carry out initial experiments regarding chemical transferability. Our work outlines a route toward future data-efficient training workflows for tensorial ML models for solid-state NMR, combining inexpensive synthetic supervision with targeted first-principles refinement.

cond-mat.mtrl-sci↗

arXiv2026

An Accurate Tensorial Model for Prediction of Full Zeolite NMR Spectra

Solid state nuclear magnetic resonance (ss-NMR) is one of the most sensitive and popular techniques for structure elucidation in geometrically complex crystalline materials, such as zeolites. Synergistic support from computational modelling is vital to interpret experimental spectra, and relate ss-NMR to atomistic models. Nevertheless, computational predictions are hindered by the high expense of calculating magnetic shielding (MS) and electric field gradient (EFG) tensors from first principles. In this work, we leverage a novel tensorial machine learning approach to train a general model for predicting complete NMR tensors. We demonstrate the utility of the approach for a diverse dataset of zeolitic materials and NMR-active nuclei ($^{27}$Al, $^{29}$Si, $^{17}$O, $^{23}$Na and $^{1}$H), predicting all NMR observables to a high degree of precision. These observables are then translated into predictions of the full $^{27}$Al and $^{29}$Si ss-nMR spectra for the exemplary zeolite RTH. Thus, this work opens a pathway to accurate, high-throughput NMR simulation for large-scale and realistic models of chemically complex zeolites.

physics.chem-ph↗

arXiv2025

A Simple and Scalable Kernel Density Approach for Reliable Uncertainty Quantification in Atomistic Machine Learning

Machine learning models are increasingly used to predict material properties and accelerate atomistic simulations, but the reliability of their predictions depends on the representativeness of the training data. We present a scalable, GPU-accelerated uncertainty quantification framework based on $k$-nearest-neighbor kernel density estimation (KDE) in a PCA-reduced descriptor space. This method efficiently detects sparsely sampled regions in large, high-dimensional datasets and provides a transferable, model-agnostic uncertainty metric without requiring retraining costly model ensembles. The framework is validated across diverse case studies varying in: i) chemistry, ii) prediction models (including foundational neural network), iii) descriptors used for KDE estimation, and iv) properties whose uncertainty is sought. In all cases, the KDE-based score reliably flags extrapolative configurations, correlates well with conventional ensemble-based uncertainties, and highlights regions of reduced prediction trustworthiness. The approach offers a practical route for improving the interpretability, robustness, and deployment readiness of ML models in materials science.

physics.chem-ph↗

arXiv2023

A reactive neural network framework for water-loaded acidic zeolites

Under operating conditions, the dynamics of water and ions confined within protonic aluminosilicate zeolite micropores are responsible for many of their properties, including hydrothermal stability, acidity and catalytic activity. However, due to high computational cost, operando studies of acidic zeolites are currently rare and limited to specific cases and simplified models. In this work, we have developed a general reactive neural network potential (NNP) attempting to cover the entire class of acidic zeolites, including the full range of experimentally relevant water concentrations and Si/Al ratios. This NNP combines dramatic sampling acceleration, retaining the reference metaGGA DFT level, with the capacity for discovery of new chemistry, such as collective defect formation mechanisms at the zeolite surface. Furthermore, we exemplify how the NNP can be used as a basis for further extensions/improvements which include data-efficient adoption of higher-level (hybrid) references via $Δ$-learning and the acceleration of rare event sampling via automatic construction of collective variables. These developments represent a significant step towards accurate simulations of realistic catalysts under operando conditions.

cond-mat.mtrl-sci↗

arXiv2023

Differentiable Simulations for Enhanced Sampling of Rare Events

Simulating rare events, such as the transformation of a reactant into a product in a chemical reaction typically requires enhanced sampling techniques that rely on heuristically chosen collective variables (CVs). We propose using differentiable simulations (DiffSim) for the discovery and enhanced sampling of chemical transformations without a need to resort to preselected CVs, using only a distance metric. Reaction path discovery and estimation of the biasing potential that enhances the sampling are merged into a single end-to-end problem that is solved by path-integral optimization. This is achieved by introducing multiple improvements over standard DiffSim such as partial backpropagation and graph mini-batching making DiffSim training stable and efficient. The potential of DiffSim is demonstrated in the successful discovery of transition paths for the Muller-Brown model potential as well as a benchmark chemical system - alanine dipeptide.

physics.chem-ph↗

arXiv2022

Accurate large-scale simulations of siliceous zeolites by neural network potentials

The computational discovery and design of zeolites is a crucial part of the chemical industry. Finding highly accurate while computationally feasible protocol for identification of hypothetical zeolites that could be targeted experimentally is a great challenge. To tackle the challenge, we trained neural network potentials (NNP) with the SchNet architecture on a structurally diverse database of density functional theory (DFT) data. This database was iteratively extended by active learning to cover not only low-energy equilibrium configurations but also high-energy transition states. We demonstrate that the resulting reactive NNPs retain the accuracy of the DFT reference for thermodynamic stabilities, vibrational properties, and reactive and non-reactive phase transformations. The novel NNPs outperforms specialized, analytical force fields for silica, such as ReaxFF, by order(s) of magnitude in accuracy, while speeding up the calculations in comparison to DFT by at least three orders of magnitude. As a showcase, we screened an existing zeolite database containing 330 thousand structures and revealed more than 20 thousand additional hypothetical frameworks in the thermodynamically accessible range of zeolite synthesis. Hence, our NNPs are expected to be essential for future high-throughput studies on the structure and reactivity of hypothetical and existing zeolites.

cond-mat.mtrl-sci↗

arXiv2022

Understanding chemical reactions via variational autoencoder and atomic representations

On the time scales accessible to atomistic numerical modelling, chemical reactions are considered rare events. Atomistic simulations are typically biased along a low-dimensional representation of a chemical reaction in an atomic structure space, i.e., along the collective variable, to accelerate sampling of these improbable events. However, suitable collective variables are often complicated to guess due to the complexity of the transitions. Therefore, we present an automatic method of generating robust collective variables from atomic representation vectors, using either fixed Behler-Parrinello functions or representations extracted from pre-trained machine learning potentials. Variational autoencoder with these representations as inputs is trained while its latent space with arbitrary dimension gives us the set of collective variables. The resulting collective variables inherit all necessary invariances from the atomic representations and can be trained entirely unsupervised. The method's effectiveness is demonstrated using three different chemical reactions, one being the complex hydrolysis of a heterogeneous aluminosilicate catalyst. Lastly, we consider the method in the context of unseen atomic structure prediction, efficiently creating structures for different values of collective variables in a generative model fashion.

physics.chem-ph↗

arXiv2019

Control of spintronic and electronic properties of bimetallic and vacancy-ordered vanadium carbide MXenes via surface functionalization

MXenes are 2D transition metal carbides with high potential for overcoming limitations of conventional two-dimensional electronics. In this context, various MXenes have shown magnetic properties suitable for applications in spintronics, yet the number of MXenes reported so far is far smaller than their parental MAX phases. Therefore, we have studied the structural, electronic and magnetic properties of bimetallic and vacancy-ordered MXenes derived from a new $(V_{2/3}Zr_{1/3})_2AlC $ MAX phase to assess whether MXene exfoliation would return stable magnetic materials. In particular, we have investigated the properties of pristine and surface-functionalized $(V_{2/3}Zr_{1/3})_2CX_2$ bimetallic and $(V_{2/3} \square_{1/3})_2CX2$ vacancy-ordered MXenes with X = O, F and OH. Our density functional theory (DFT) calculations showed that modifying the MXene stoichiometry and/or MXene surface functionalization changes MXene properties. After testing all possible combinations of metallic motifs and functionalization, we identified $V_{2/3}Zr_{1/3})_2CX_2$, $(V_{2/3}\square_{1/3})_2CF_2$ and $(V_{2/3}\square_{1/3})_2C(OH)_2$ as stable structures. Among them, $(V_{2/3}Zr_{1/3})_2CO_2$ MXene is predicted to be an FM intrinsic half-semiconductor with a remarkably high Curie temperature ($T_C$)) of 270 K. The $(V_{2/3}Zr_{1/3})_2C(OH)_2$ MXene exhibits a rather low work function (WF) (1.37 eV) and is thus a promising candidate for ultra-low work function electron emitters. Conversely, the $(V_{2/3}\square_{1/3})_2CF_2$ MXene has a rather high WF and hence can be used as a hole injector for Schottky-barrier-free contact applications. Overall, our proof-of-concept study shows that theoretical predictions of MXene exfoliation and properties support further experimental research towards developing spintronics devices.

cond-mat.mtrl-sci↗