Advanced Computational Modeling of CO2 Solubility in Ionic Liquids Using an XAI Methodology

Identification of Metal-Organic Frameworks for CO2 Capture from Humid Flue Gas: Integrating Molecular Simulation, Machine Learning, and Experimental Synthesis.

Increasing global CO2 emissions are driving efforts to develop advanced carbon capture materials. Metal-organic frameworks (MOFs) show great promise as CO2 adsorbents, yet maintaining performance under humid flue gas conditions remains a major challenge. Herein, we present an integrated computational high-throughput screening workflow that begins with a library of over 110,000 experimental MOFs or MOF-like structures, explicitly includes the effects of water in adsorption simulations, and incorporates machine learning-based stability analysis to identify promising candidates from among existing MOFs. Guided by this workflow, we synthesized a top-performing MOF that demonstrates exceptional tolerance to humid conditions, maintaining high CO2 uptake at elevated relative humidity. We further revealed key structure-property relationships and structural motifs that offer valuable design principles for next-generation MOFs for CO2 capture in humid conditions.

Black-Box and Interpretable Artificial Intelligence Models for Hydrogen Uptake Across Various Metal–Organic Frameworks

Hydrogen (H2) is expected to play a critical role in modern industry, particularly in ammonia synthesis, petroleum refining, and low-carbon transportation. The safe storage of H2 remains a major challenge due to its low volumetric density under ambient conditions. Metal–Organic Frameworks (MOFs), highly porous crystalline materials, have emerged as promising H2 storage candidates owing to their high surface areas and tuneable pore structures. Molecular simulations such as grand canonical Monte Carlo or density functional theory are costly and limited in exploring large material spaces, motivating efficient predictive tools to accelerate discovery. Here, Machine Learning (ML) techniques are compared to an explainable artificial intelligence (XAI) approach using symbolic regression (SR), trained on 10,123 experimentally measured H2 adsorption datapoints from real-world MOFs. The best performing model achieved a goodness of fit of 0.9986 with lower computational demand, but reduced interpretability, addressed using XAI analysis and clustering. SR achieves a lower goodness of fit of 0.914 but produces a physically meaningful equation highlighting structural features driving high gravimetric efficiencies. These results demonstrate strong ML capability for predicting how MOF properties and environmental conditions affect H2 uptake. This offers engineers and researchers a practical means of screening potential MOFs for H2 storage applications, with the XAI analyses providing additional confidence in the predictions. They allow researchers to understand the physical reasoning behind each output, assess the reliability of individual predictions, and make fully informed decisions, enabling predictive models to be acted upon with confidence in real-world contexts.

Methodological Guide for Understanding CO2 Capture Quantification and Characterization: A Short Review

The atmospheric concentration of carbon dioxide has surpassed 420 ppm, intensifying global warming and the need for effective mitigation strategies. Carbon capture is a central technology in this effort, traditionally implemented with aqueous amine systems but now expanding to a diverse array of solid and liquid sorbents, including zeolites, metal-organic frameworks, covalent organic frameworks, porous polymers, ionic liquids, deep eutectic solvents, and others. The rapid diversification of capture materials has expanded the analytical toolbox for quantifying CO2 uptake and identifying binding mechanisms but has also created challenges in selecting the most appropriate methodology. Gravimetric and volumetric methods are widely used for porous solids, calorimetry provides access to adsorption energies, and spectroscopic approaches such as nuclear magnetic resonance and infrared spectroscopy resolve molecular speciation in reactive liquids and functionalized sorbents. In this context, the aim of this mini-review is to identify and compare methods, techniques, and instruments capable of performing CO2 capture while simultaneously or subsequently quantifying the amount captured, thus linking the capture process with its analytical evaluation. This work systematizes these methodologies, outlining technical principles, material applicability, limitations, and representative examples, with the goal of guiding method selection and promoting consistent, comparable evaluation across emerging CO2 capture materials.

Target-driven discovery of metal hydrides using a high-throughput computational framework: from operational targets to sustainability assessment

Metal hydrides based on multi-principal element alloys (MPEAs) offer tunable hydrogen storage properties, but their vast compositional space poses significant design challenges. A high-throughput computational strategy is proposed to identify alloys under predefined operating conditions, demonstrated here for hydrogen gas separation. CALPHAD equilibrium calculations were performed in the Ti-V-Cr-Nb system using a Python TM script with PanPython (Pandat TM SDK). Single-phase body-centered cubic (BCC) alloys were identified and evaluated using a computational framework to predict pressure-composition-temperature (PCT) diagrams. Candidates were screened based on operational requirements and ranked using multi-criteria decision analysis (MCDA), considering economic, environmental, and societal indicators. The highest-ranked alloy, Ti 20 V 12.5 Cr 20 Nb 47.5 , was synthesized and characterized. Hydrogen storage measurements using a volumetric apparatus confirmed reversible hydrogen uptake of 1.2 H/M (1.68 wt.%) under the targeted operating conditions. This integrated framework enables the identification of hydrides operating under predefined conditions while supporting sustainability-driven decision-making, paving the way for accelerated discovery of hydrogen materials and the advancement of sustainable energy technologies.

Molecular Simulations and Data-Driven Predictive Modeling of Mercury Adsorption in Special Solvents

Mercury contamination presents critical challenges to the oil and gas industry due to its toxicity and severe environmental and health repercussions. Traditional mercury removal methods often suffer from inefficiency, limited selectivity, and suboptimal predictive capabilities for adsorption performance. This study addresses these challenges by integrating molecular simulations and data-driven predictive modeling to evaluate and optimize mercury adsorption in specialized solvents, with a focus on ionic liquids. Leveraging the exceptional thermal stability and selectivity of ionic liquids, Turbomole and COSMO-RS were employed to calculate sigma surfaces and profiles, providing insights into molecular properties relevant to mercury adsorption. Key predictors, including viscosity, density, activity coefficient, and Henry's constant, were derived from COSMO-RS outputs and incorporated into regression models using MATLAB's Curve Fitting Toolbox. Multiple Linear Regression (MLR) and Multiple Non-Linear Regression (MNLR) were applied to establish relationships between these predictors and the mercury adsorption capacity and efficiency. Sigma surface and profile analyses revealed that molecules with higher electronegativity exhibited superior mercury adsorption. MNLR also effectively captured nonlinear interactions, yielding more accurate adsorption capacity and efficiency predictions than MLR with an adjusted R2 of 1 and 0.7105. Among the predictors, viscosity and activity coefficient emerged as the most reliable for adsorption capacity and efficiency predictions, registering an RMSE of 1.129 and 1.590, respectively. This study highlights the potential of combining computational chemistry with advanced data analytics to overcome limitations of current methods and enhance mercury removal strategies. Despite dataset limitations, the findings demonstrate the promise of integrating molecular simulations and data analytics for designing efficient mercury mitigation strategies, laying a foundation for future improvements in predictive accuracy and industrial applications.

Ionic Liquids for Hydrocarbon/Thiophene Separation: Extensive Analysis of Structure–Property Correlations