arXiv subjects

Eliodoro Chiavazzo

Publications and source records attributed to Eliodoro Chiavazzo.

At least 19 recordsLinked to original sources

arXiv2026

Quantifying Perovskite Solar Cell Degradation via Machine Learning from Spatially Resolved Multimodal Luminescence Time Series

Perovskite solar cells achieve remarkable power conversion efficiencies, yet operational stability remains a major barrier to large-scale deployment. Reliable and rapid assessment of device state of health is therefore essential. Conventional electrical diagnostics, such as illuminated current-voltage (J--V) sweeps, provide accurate performance metrics but are time-consuming and do not resolve spatially localized degradation, motivating non-invasive imaging-based alternatives. A deep-learning framework is introduced to estimate PSC efficiency retention, $R_\mathrm{PCE}=\mathrm{PCE}_t/\mathrm{PCE}_0$, directly from multimodal luminescence imaging acquired during device aging. Each sample combines electroluminescence (EL), open-circuit photoluminescence (PLoc), and short-circuit photoluminescence (PLsc) at an aged state with device-specific reference images at $t=0$, enabling learning of degradation-relevant spatial changes. LumPerNet, a compact convolutional neural network, is benchmarked against a spatially homogenized control in which each luminescence channel is replaced by its spatial average while retaining the same learning framework and leakage-aware protocol. The comparison indicates that global luminescence evolution contains most of the predictive signal, while spatial information provides a secondary contribution to robustness. These results establish spatially resolved luminescence imaging as a practical route for accelerated stability testing and non-invasive degradation monitoring in perovskite photovoltaics.

cond-mat.mtrl-sci↗

arXiv2026

High-Throughput Bayesian Optimization of Cement-Salt Hydrates Composites for Seasonal Thermochemical Energy Storage

Thermochemical energy storage (TCES) based on salt hydrates is a promising route for seasonal heat storage; however, the design of practical sorbent materials remains challenging due to a non-trivial coupling between composition, synthesis feasibility, performance, and cost. Here, focusing on salt-into-matrix cement-based composites, we demonstrate that a high-throughput experimental framework based on Bayesian optimization (BO) can be used to orchestrate the optimization process of composite materials for low-temperature TCES. The explored design space is defined by salt type, salt concentration, water-to-cement ratio, and additive-to-cement ratio, while two competing objectives are pursued in parallel, namely the specific energy and the specific energy cost. The BO-guided campaign identified Pareto-optimal composites based on CaCl$_2$, Zn(NO$_3$)$_2$, and LiCl, highlighting the promise of cement-salt combinations that have been only marginally explored, or not previously reported, in cement-based TCES systems. The best-performing formulation (LiCl-based), achieved an average specific energy of about $\SI{458}{\kilo\joule\per\kilo\gram}$, whereas CaCl$_2$- and Zn(NO$_3$)$_2$-based composites showed lower but still competitive specific energy values combined with more favorable specific energy cost. Overall, the optimized formulations improved the specific energy of previously developed cement-based materials by up to a factor of five, although it remains below that of state-of-the-art composites based on silica gel and expanded vermiculite. Nonetheless, the present materials, notably CaCl$_2$- and Zn(NO$_3$)$_2$-based composites, offer an attractive cost-to-performance balance, highlighting BO as an effective strategy for accelerated TCES materials discovery.

cond-mat.mtrl-sci↗

arXiv2026

Comparing fine-tuning strategies of MACE machine learning force field for modeling Li-ion diffusion in LiF for batteries

Machine learning force fields (MLFFs) are transforming materials science and engineering by enabling the study of complex phenomena, such as those critical to battery operation. In this work, we explore the predictive capabilities of pre-trained and fine-tuned MACE MLFF and compare different fine-tuning strategies for predicting interstitial lithium diffusivity in LiF, a key component in the solid electrolyte interphase in Li-ion batteries. Our results demonstrate that the MACE-MPA-0 foundational model achieves comparable accuracy to well-trained DeePMD, in predicting key diffusion properties based on large scale molecular dynamics simulation, while requiring minimal or no training data. For instance, the MACE-MPA-0 predicts an activation energy $E_a$ of 0.22eV, the fine-tuned model with only 300 data points predicts $E_a =$ 0.20eV, both of which show good agreement with the DeePMD model reference value of $E_a = $ 0.24eV. In this work, we provide a solid test case where fine-tuning approaches, whether using data generated for DeePMD or data produced by the foundational MACE model itself, yield similar robust performance to the DeePMD potential trained with over 40,000 actively learned data, albeit requiring only a fraction of the training data.

cond-mat.mtrl-sci↗

arXiv2026

Pore-level Quantitative Structure-Activity Relationship (QSAR) for Water Permeation Rate in Aquaporins

Aquaporins (AQPs) and aquaglyceroporins (AQGPs) play a crucial role in regulating water transport and solute selectivity across biological membranes. Besides their biological relevance, AQPs have at-tracted growing interest as models for the design of next-generation biomimetic membranes for water filtration. In this work, we present a pore-level Quantitative Structure-Activity Relationship (QSAR) approach that relates structural and physicochemical pore descriptors with experimentally reported water permeation rates across a set of AQ(G)Ps with high-resolution 3D structures. This data-driven methodology, presented here as a proof of concept, introduces a multi-feature framework for determining pore descriptors associated with water transport efficiency in AQ(G)Ps. Applied to two compiled permeation rate datasets, this framework recapitulates determinants previously reported in single-feature studies, while also highlighting additional pore descriptors that emerge as relevant in a multi-variable context. The insights gained through this approach may, in perspective, contribute to advancing the rational design of AQP-based filtration devices and to deepening the molecular understanding of the function of these valuable macromolecules in health and disease.

q-bio.BM↗

arXiv2026

Screening novel cathode materials from the Energy-GNoME database using MACE machine learning force field and DFT

The development of new battery materials, particularly novel cathode chemistries, is essential for enabling next generation energy storage technologies. In this work, we employ a multi-fidelity screening protocol combining the Energy-GNoME confident criteria, foundational MACE machine-learning force fields (MLFF), and physically motivated heuristic filters to identify novel intercalation cathodes for post-lithium batteries, namely: Na-, K-, Mg-, and Ca-ion batteries. Foundational MACE models are used to efficiently asses dynamical stability, thermodynamical stability, average voltage, and theoretical specific energy, enabling a rapid screening of candidates. For the most promising cathodes, voltage predictions are refined using DFT+U calculations. This work delivers three key outcomes: i) establishing and validating a robust high-throughput screening approach for cathode materials with foundational MLFF models; ii) suggestions for cathode candidates for the development of next-generation of batteries; iii) a fair comparison between the MACE predictions and the readily available figures of merit reported in the Energy-GNoME database on the examined materials.

cond-mat.mtrl-sci↗

arXiv2025

Notes on Quantum Computing for Thermal Science

This document explores the potential of quantum computing in Thermal Science. Conceived as a living document, it will be continuously updated with experimental findings and insights for the research community in Thermal Science. By experiments, we refer both to the search for the most effective algorithms and to the performance of real quantum hardware. Those are fields that are evolving rapidly, driving a technological race to define the best architectures. The development of novel algorithms for engineering problems aims at harnessing the unique strengths of quantum computing. Expectations are high, as users seek concrete evidence of quantum supremacy - a true game changer for engineering applications. Among all heat transfer mechanisms (conduction, convection, radiation), we start with conduction as a paradigmatic test case in the field being characterized by a rich mathematical foundation for our investigations.

physics.comp-ph↗

arXiv2025

Charting the landscape of Bardeen-Cooper-Schrieffer superconductors in experimentally known compounds

We perform a high-throughput computational search for novel phonon-mediated superconductors, starting from the Materials Cloud 3-dimensional structure database of experimentally known inorganic stoichiometric compounds. We first compute the Allen-Dynes critical temperature (T$_c$) for 4533 non-magnetic metals using a direct and progressively finer sampling of the electron-phonon couplings. For the candidates with the largest T$_c$, we use automated Wannierizations and electron-phonon interpolations to obtain a high-quality dataset for the most promising 250 dynamically stable structures, for which we calculate spectral functions, superconducting bandgaps, and isotropic Migdal-Eliashberg critical temperatures. For 140 of these, we also provide anisotropic Migdal-Eliashberg superconducting gaps and critical temperatures. The approach is remarkably successful in finding known superconductors, and we find 24 unknown ones with a predicted anisotropic T$_{\rm c}$ above 10~K. Among them, we identify a possible double gap superconductor (p-doped BaB$_2$), a non-magnetic half-Heusler ZrRuSb, and the perovskite TaRu$_3$C, all exhibiting significant T$_{\rm c}$. Finally, we introduce a sensitivity analysis to estimate the robustness of the predictions.

cond-mat.supr-con↗

arXiv2025

Searching for fixed points of dynamical systems by the Explorative Relaxation Redistribution Method

Being able to effectively locate saddle (and other fixed) points in dynamical systems holds tremendous implications in a number of applications in engineering and science, among which the study of rare events in molecular simulations stands as one of the most prominent field of interest. Although there is a vast literature on methods aiming at addressing this challenge, open issues still remain in several aspects such as computational efficiency and easy implementation. In this work, we suggest that the Relaxation Redistribution Method (RRM) - formerly introduced to solve the invariance equation in the context of stiff dynamical systems ruling detailed chemical kinetics - can be reformulated to locate saddle and other fixed points even for stochastic simulators driven by effective energy gradients. This new formulation of the RRM is referred to as the Explorative Relaxation Redistribution Method (ERRM). Benchmarks based on the popular Mueller-Brown and other potentials are used to test the ERRM.

cond-mat.other↗

arXiv2024

Energy-GNoME: A Living Database of Selected Materials for Energy Applications

Artificial Intelligence (AI) in materials science is driving significant advancements in the discovery of advanced materials for energy applications. The recent GNoME protocol identifies over 380,000 novel stable crystals. From this, we identify over 33,000 materials with potential as energy materials forming the Energy-GNoME database. Leveraging Machine Learning (ML) and Deep Learning (DL) tools, our protocol mitigates cross-domain data bias using feature spaces to identify potential candidates for thermoelectric materials, novel battery cathodes, and novel perovskites. Classifiers with both structural and compositional features identify domains of applicability, where we expect enhanced accuracy of the regressors. Such regressors are trained to predict key materials properties like, thermoelectric figure of merit (zT), band gap (Eg), and cathode voltage ($ΔV_c$). This method significantly narrows the pool of potential candidates, serving as an efficient guide for experimental and computational chemistry investigations and accelerating the discovery of materials suited for electricity generation, energy storage and conversion.

cond-mat.mtrl-sci↗

arXiv2024

Classification-based detection and quantification of cross-domain data bias in materials discovery

It stands to reason that the amount and the quality of data is of key importance for setting up accurate AI-driven models. Among others, a fundamental aspect to consider is the bias introduced during sample selection in database generation. This is particularly relevant when a model is trained on a specialized dataset to predict a property of interest, and then applied to forecast the same property over samples having a completely different genesis. Indeed, the resulting biased model will likely produce unreliable predictions for many of those out-of-the-box samples. Neglecting such an aspect may hinder the AI-based discovery process, even when high quality, sufficiently large and highly reputable data sources are available. In this regard, with superconducting and thermoelectric materials as two prototypical case studies in the field of energy material discovery, we present and validate a new method (based on a classification strategy) capable of detecting, quantifying and circumventing the presence of cross-domain data bias.

cond-mat.other↗

arXiv2024

Learning effective good variables from physical data

We assume that a sufficiently large database is available, where a physical property of interest and a number of associated ruling primitive variables or observables are stored. We introduce and test two machine learning approaches to discover possible groups or combinations of primitive variables: The first approach is based on regression models whereas the second on classification models. The variable group (here referred to as the new effective good variable) can be considered as successfully found, when the physical property of interest is characterized by the following effective invariant behaviour: In the first method, invariance of the group implies invariance of the property up to a given accuracy; in the other method, upon partition of the physical property values into two or more classes, invariance of the group implies invariance of the class. For the sake of illustration, the two methods are successfully applied to two popular empirical correlations describing the convective heat transfer phenomenon and to the Newton's law of universal gravitation.

physics.data-an↗

arXiv2023

Mesoscopic modeling and experimental validation of thermal and mechanical properties of polypropylene nanocomposites reinforced by graphene-based fillers

The development of nanocomposites relies on structure-property relations, which necessitate multiscale modeling approaches. This study presents a modelling framework that exploits mesoscopic models to predict the thermal and mechanical properties of nanocomposites starting from their molecular structure. In detail, mesoscopic models of polypropylene (PP) and graphene based nanofillers (Graphene (Gr), Graphene Oxide (GO), and reduced Graphene Oxide (rGO)) are considered. The newly developed mesoscopic model for the PP/Gr nanocomposite provides mechanistic information on the thermal and mechanical properties at the filler-matrix interface, which can be then exploited to enhance the prediction accuracy of traditional continuum simulations by calibrating the thermal and mechanical properties of the filler-matrix interface. Once validated through a dedicated experimental campaign, this multiscale model demonstrates that with the modest addition of nanofillers (up to 2 wt.%), the Young's modulus and thermal conductivity show up to 35% and 25% enhancement, respectively, while the Poisson's ratio slightly decreases. Among the different combinations tested, PP/Gr nanocomposite shows the best mechanical properties, whereas PP/rGO demonstrates the best thermal conductivity. This validated mesoscopic model can contribute to the development of smart materials with enhanced mechanical and thermal properties based on polypropylene, especially for mechanical, energy storage, and sensing applications.

cond-mat.mtrl-sci↗

arXiv2023

Enhancing ReaxFF for Molecular Dynamics Simulations of Lithium-Ion Batteries: An interactive reparameterization protocol

Lithium-ion batteries (LIBs) are crucial for the green economy, powering portable electronics, electric vehicles, and renewable energy systems. The solid-electrolyte interphase (SEI) is vital for LIB operation, performance, and safety. SEI forms due to thermal instability at the anode-electrolyte interface, with electrolyte reduction products stabilizing it as an electrochemical buffer. This article aims to enhance the parametrization of the ReaxFF force field for accurate molecular dynamics (MD) simulations of SEI in LIBs. Focus is on Lithium Fluoride (LiF), an inorganic salt with favorable properties in the passivation layer. The protocol heavily relies on Python libraries for atomistic simulations, enabling robust automation of reparameterization steps. The proposed configurations and dataset enable the new ReaxFF to accurately represent the solid nature of LiF and improve mass transport property prediction in MD simulations. Optimized ReaxFF surpasses previous force fields by adjusting lithium diffusivity, resulting in a significant improvement in room temperature prediction by two orders of magnitude. However, our comprehensive investigation reveals ReaxFF's strong sensitivity to the training set, challenging its ability to interpolate the potential energy surface. Consequently, the current ReaxFF formulation is suitable for modeling specific phenomena by utilizing the proposed interactive reparameterization protocol and constructing a dataset. This work is an important step towards refining ReaxFF for precise reactive MD simulations, shedding light on challenges and limitations in force field parametrization. The demonstrated limitations underscore the potential for developing more advanced force fields through our interactive reparameterization protocol, enabling accurate and comprehensive MD simulations in the future.

cond-mat.mtrl-sci↗

arXiv2023

The impact of physicochemical features of carbon electrodes on the capacitive performance of supercapacitors: A machine learning approach

Hybrid electric vehicles and portable electronic systems use supercapacitors for energy storage owing to their fast charging discharging rates, long life cycle, and low maintenance. Specific capacitance is regarded as one of the most important performance-related characteristics of a supercapacitor's electrode. In the current study, Machine Learning (ML) algorithms were used to determine the impact of various physicochemical properties of carbon-based materials on the capacitive performance of electric double-layer capacitors. Published experimental datasets from 147 references (4899 data entries) were extracted and then used to train and test the ML models, to determine the relative importance of electrode material features on specific capacitance. These features include current density, pore volume, pore size, presence of defects, potential window, specific surface area, oxygen, and nitrogen content of the carbon-based electrode material. Additionally, categorical variables as the testing method, electrolyte, and carbon structure of the electrodes are considered as well. Among five applied regression models, an extreme gradient boosting model was found to best correlate those features with the capacitive performance, highlighting that the specific surface area, the presence of nitrogen doping, and the potential window are the most significant descriptors for the specific capacitance. These findings are summarized in a modular and open-source application for estimating the capacitance of supercapacitors given, as only inputs, the features of their carbon-based electrodes, the electrolyte and testing method. In perspective, this work introduces a new wide dataset of carbon electrodes for supercapacitors extracted from the experimental literature, also giving an instance of how electrochemical technology can benefit from ML models.

cond-mat.mtrl-sci↗

arXiv2023

Leveraging Composition-Based Material Descriptors for Machine Learning Optimization

In this study, we evaluate several classifiers and focus on selecting a minimal set of appropriate material features. Our objective is to propose and discuss general strategies for reducing the number of descriptors required for material classification. The first strategy involves testing whether the critical temperature of the target material property is invariant with respect to binary groups of composition-based features. We also propose a multi-objective optimization procedure to reduce the set of composition-based material descriptors. The latter procedure is found to be particularly useful when applied to Bayesian classifiers. We test the proposed strategies focusing on low-temperature superconductors material data extracted from a public database.

cond-mat.other↗

arXiv2022

Double Diffusion Maps and their Latent Harmonics for Scientific Computations in Latent Space

We introduce a data-driven approach to building reduced dynamical models through manifold learning; the reduced latent space is discovered using Diffusion Maps (a manifold learning technique) on time series data. A second round of Diffusion Maps on those latent coordinates allows the approximation of the reduced dynamical models. This second round enables mapping the latent space coordinates back to the full ambient space (what is called lifting); it also enables the approximation of full state functions of interest in terms of the reduced coordinates. In our work, we develop and test three different reduced numerical simulation methodologies, either through pre-tabulation in the latent space and integration on the fly or by going back and forth between the ambient space and the latent space. The data-driven latent space simulation results, based on the three different approaches, are validated through (a) the latent space observation of the full simulation through the Nyström Extension formula, or through (b) lifting the reduced trajectory back to the full ambient space, via Latent Harmonics. Latent space modeling often involves additional regularization to favor certain properties of the space over others, and the mapping back to the ambient space is then constructed mostly independently from these properties; here, we use the same data-driven approach to construct the latent space and then map back to the ambient space.

stat.ML↗

arXiv2021

Minimal set of crystallographic descriptors for sorption properties in hypothetical Metal Organic Frameworks: Role in sequential learning optimization

Several studies have been recently reported in the literature on sorption properties of MOFs with a number of organic sorbates, such as ethanol and methanol. Surprisingly, still few studies have been reported on water sorbate despite its large availability, low cost and environmental sustainability, and the screening of a large number of hypothetical MOFs-water working pairs for engineering applications is still challenging. Based on a recently reported database of over 5000 hypothetical MOFs, a first contribution of this study is the identification of the minimal set of crystallographic descriptors underpinning the most important sorption properties of MOFs for \ch{CO2} and, importantly, for \ch{H2O}. Furthermore, a comprehensive comparison of several Sequential Learning (SL) algorithms for MOFs properties optimization is carried out and the role played by the above minimal set of crystallographic descriptors clarified. In sorption-based energy transformations, thermodynamic limits of important figures of merit (e.g. maximum specific energy) depend both on operating conditions and equilibrium sorption properties in a wide range of sorbate coverage values. The access to the latter properties is often incomplete, with essential quantities such as equilibrium adsorption isotherms spanning over the full sorbate coverage range and values of the isosteric heat being only partially available. As a result, this may prevent the computation of objective functions during the optimization procedure. We propose a fast procedure for optimizing specific energy in a closed sorption energy storage system with the only access to the water Henry coefficient at a fixed temperature value and to the specific surface area.

cond-mat.other↗

arXiv2017

Passive high-yield seawater desalination at below one sun by modular and low-cost distillation

Although seawater is abundant, desalination is energy-intensive and expensive. Using the sun as an energy source is attractive for desalinating seawater; however, the performance of state-of-the-art passive devices is unsatisfactory when operated at less than one sun ($<$ $1$ $kW$ $m^{-2}$). Here, we present a completely passive, modular, and low-cost solar thermal distiller for seawater desalination. Each distillation stage is made of two opposed hydrophilic layers separated by a hydrophobic microporous membrane, and it does not require further mechanical ancillaries. Under realistic laboratory and outdoor conditions, we obtained a distillate flow rate of almost $3$ $L$ $m^{-2}$ $h^{-1}$ from seawater at less than one sun - twice the yield of recent passive device reported in the literature. In perspective, theoretical modelling suggests that the distiller has the potential to further doubling the peak flow rate observed in the current experiments. This layout can satisfy freshwater needs in isolated and impoverished communities, as well as realize self-sufficient floating installations or provide freshwater in emergency conditions.

cond-mat.other↗