arXiv subjects

Raffaele Cheula

Publications and source records attributed to Raffaele Cheula.

5 recordsLinked to original sources

arXiv2026

Fine-tuning universal machine learning potentials for transition state search in surface catalysis

Determining transition states (TSs) of surface reactions is central to understanding and designing heterogeneous catalysts but remains computationally prohibitive with density functional theory (DFT). While machine learning potentials (MLPs) offer significant speedups, task-specific models have limited transferability across catalytic systems, and universal MLPs (uMLPs) lack the accuracy needed for reactive configurations. Here, we present a workflow based on active learning to iteratively fine-tune uMLPs for DFT-quality TS search. Using 250 TSs from the CO2 hydrogenation reaction network on metal and single-atom alloy surfaces, we first benchmark TS search algorithms, identifying the Sella algorithm as most robust, and propose a modification (Bond-Aware Sella) that substantially improves its success rate. We then explore sequential and batch active-learning strategies for fine-tuning and show that DFT-quality TS structures can be found using only 8 DFT single-point calculations on average per structure. This demonstrates the viability of fine-tuned uMLPs for high-throughput catalyst screening.

cond-mat.mtrl-sci↗

arXiv2025

How accurate are foundational machine learning interatomic potentials for heterogeneous catalysis?

Foundational machine learning interatomic potentials (MLIPs) are being developed at a rapid pace, promising closer and closer approximation to ab initio accuracy. This unlocks the possibility to simulate much larger length and time scales. However, benchmarks for these MLIPs are usually limited to ordered, crystalline and bulk materials. Hence, reported performance does not necessarily accurately reflect MLIP performance in real applications such as heterogeneous catalysis. Here, we systematically analyze zero-shot performance of 80 different MLIPs, evaluating tasks typical for heterogeneous catalysis across a range of different data sets, including adsorption and reaction on surfaces of alloyed metals, oxides, and metal-oxide interfacial systems. We demonstrate that current-generation foundational MLIPs can already perform at high accuracy for applications such as predicting vacancy formation energies of perovskite oxides or zero-point energies of supported nanoclusters. However, limitations also exist. We find that many MLIPs catastrophically fail when applied to magnetic materials, and structure relaxation in the MLIP generally increases the energy prediction error compared to single-point evaluation of a previously optimized structure. Comparing low-cost task-specific models to foundational MLIPs, we highlight some core differences between these model approaches and show that -- if considering only accuracy -- these models can compete with the current generation of best-performing MLIPs. Furthermore, we show that no single MLIP universally performs best, requiring users to investigate MLIP suitability for their desired application.

cond-mat.mtrl-sci↗

arXiv2025

Interpretable machine learned predictions of adsorption energies at the metal--oxide interface

The conversion of $\mathrm{CO_2}$ to value-added compounds is an important part of the effort to store and reuse atmospheric $\mathrm{CO_2}$ emissions. Here we focus on $\mathrm{CO_2}$ hydrogenation over so-called inverse catalysts: transition metal oxide clusters supported on metal surfaces. The conventional approach for computational screening of such candidate catalyst materials involves a reliance on density functional theory (DFT) to obtain accurate adsorption energies at a significant computational cost. Here we present a machine learning (ML)-accelerated workflow for obtaining adsorption energies at the metal--oxide interface. We enumerate possible binding sites at the clusters and use DFT to sample a subset of these with diverse local adsorbate environments. The data set is used to explore interpretable and black-box ML models with the aim to reveal the electronic and structural factors controlling adsorption at metal--oxide interfaces. Furthermore, the explored ML models can be used for low-cost prediction of adsorption energies on structures outside of the original training data set. The workflow presented here, along with the insights into trends in adsorption energies at metal--oxide interfaces, will be useful for identifying active sites, predicting parameters required for microkinetic modeling of reactions on complex catalyst materials, and accelerating data-driven catalyst design.

cond-mat.mtrl-sci↗

arXiv2025

Transition States Energies from Machine Learning: An Application to Reverse Water-Gas Shift on Single-Atom Alloys

Obtaining accurate transition state (TS) energies is a bottleneck in computational screening of complex materials and reaction networks due to the high cost of TS search methods and first-principles methods such as density functional theory (DFT). Here we propose a machine learning (ML) model for predicting TS energies based on Gaussian process regression with the Wasserstein Weisfeiler-Lehman graph kernel (WWL-GPR). Applying the model to predict adsorption and TS energies for the reverse water-gas shift (RWGS) reaction on single-atom alloy (SAA) catalysts, we show that it can significantly improve the accuracy compared to traditional approaches based on scaling relations or ML models without a graph representation. Further benefitting from the low cost of model training, we train an ensemble of WWL-GPR models to obtain uncertainties through subsampling of the training data and show how these uncertainties propagate to turnover frequency (TOF) predictions through the construction of an ensemble of microkinetic models. Comparing the errors in model-based vs DFT-based TOF predictions, we show that the WWL-GPR model reduces errors by almost an order of magnitude compared to scaling relations. This demonstrates the critical impact of accurate energy predictions on catalytic activity estimation. Finally, we apply our model to screen new materials, identifying promising catalysts for RWGS. This work highlights the power of combining advanced ML techniques with DFT and microkinetic modeling for screening catalysts for complex reactions like RWGS, providing a robust framework for future catalyst design.

cond-mat.mtrl-sci↗

arXiv2024

Accessing Numerical Energy Hessians with Graph Neural Network Potentials and Their Application in Heterogeneous Catalysis

Access to the potential energy Hessian enables determination of the Gibbs free energy, and certain approaches to transition state search and optimization. Here, we demonstrate that off-the-shelf pretrained Open Catalyst Project (OCP) machine learned potentials (MLPs) determine the Hessian with great success (58 cm$^{-1}$ mean absolute error (MAE)) for intermediates adsorbed to heterogeneous catalyst surfaces. This enables the use of OCP models for the aforementioned applications. The top performing model, with a simple offset correction, gives good estimations of the vibrational entropy contribution to the Gibbs free energy with an MAE of 0.042 eV at 300 K. The ability to leverage models to capture the translational entropy was also explored. It was determined that 94% of randomly sampled systems had a translational entropy greater than 0.1 eV at 300 K. This underscores the need to go beyond the harmonic approximation to consider the entropy introduced by adsorbate translation, which increases with temperature. Lastly, we used MLP determined Hessian information for transition state search and found we were able to reduce the number of unconverged systems by 65% to 93% overall convergence, improving on the baseline established by CatTSunami.

cond-mat.mtrl-sci↗