No abstract

Noble gases are vital industrial chemicals widely used in various applications, such as cryogenics and optical devices. Compared with conventional technologies for the enrichment and separation of noble gases from the atmosphere, emerging methods based on advanced nanoporous materials have advantages in terms of energy and separation efficiency due to their tunable pore structure and chemical affinities. In this work, both sorption and transport properties are calculated via efficient theoretical models to screen large experimental libraries of nanoporous materials to enrich argon, krypton, and xenon from various mixtures. The theoretical predictions are validated by Monte Carlo simulation and molecular dynamics simulation. Promising candidates are identified for both adsorption separation and membrane separation of Ar/Kr, Kr/Xe, and Xe/Ar. The structure−property relation identified in this work provides insights for the design of nanoporous materials in noble gas separation.

Two versions of nonlocal classical density functional theory (cDFT) have been proposed to predict multicomponent gas adsorption in nanoporous materials by using the Lennard-Jones model for gas mixtures and the universal force field for the adsorbents. With the modified fundamental measure theory to describe short-range repulsions or volume-exclusion effects, one version of cDFT adopts the mean-field approximation for van der Waals attraction (here referred to as cDFT-MFA) as commonly used in porous material characterization, and the other version accounts for long-range correlations through a weighted-density approximation (cDFT-WDA). For a number of gas mixtures in MOF-5 (without sub-pores inaccessible to gas molecules), the adsorption isotherms predicted from cDFT-WDA are quantitatively consistent with results from grand canonical Monte Carlo simulation, while cDFT-MFA systematically underestimates the adsorption due to the neglect of correlation effects. Nevertheless, both versions of cDFT outperform the ideal adsorbed solution theory (IAST) at high pressure. Because IAST predicts mixture adsorption using only single-component data, it fails to capture the selective behavior arising from asymmetric interactions among different chemical species. The cDFT calculations are implemented with massively parallel GPU-accelerated algorithms to achieve rapid yet accurate predictions of multicomponent adsorption isotherms with full atomistic details of the adsorbent materials. This work thus provides a theoretical basis for the computational design of adsorption-based separation processes as well as for screening and data-driven inverse design of nanoporous materials.

Gas separation is crucial for industrial production and environmental protection, with metal-organic frameworks(MOFs) offering a promising solution due to their tunable structural properties and chemical compositions. Traditional simulation approaches, such as molecular dynamics, are complex and computationally demanding. Although feature engineering-based machine learning methods perform better, they are susceptible to overfitting because of limited labeled data. Furthermore, these methods are typically designed for single tasks, such as predicting gas adsorption capacity under specific conditions, which restricts the utilization of comprehensive datasets including all adsorption capacities. To address these challenges, we propose Uni-MOF, an innovative framework for large-scale, three-dimensional MOF representation learning, designed for universal multi-gas prediction. Specifically, Uni-MOF serves as a versatile "gas adsorption detector" for MOF materials, employing pure three-dimensional representations learned from over 631,000 collected MOF and COF structures. Our experimental results show that Uni-MOF can automatically extract structural representations and predict adsorption capacities under various operating conditions using a single model. For simulated data, Uni-MOF exhibits remarkably high predictive accuracy across all datasets. Impressively, the values predicted by Uni-MOF correspond with the outcomes of adsorption experiments. Furthermore, Uni-MOF demonstrates considerable potential for broad applicability in predicting a wide array of other properties.

γ-TiAl alloys are the most promising lightweight high-temperature structural materials, but the materials often fail from the surface, which is mainly attributed to the stress state of the material surface. In this paper, the orthogonal experiment method and molecular dynamics modeling are used to choose a set of the best process parameters for supersonic fine particle bombardment. Furthermore, by determining the optimal process parameters, this study examines the influence of residual stress distribution on the mechanical properties of the material under various process conditions. The simulation results reveal that the residual stress distribution is minimally impacted by particle radius, nonetheless, maintaining a moderate level of compressive residual stress within a specific range can substantially augment both the tensile strength and indentation hardness. An increase in the number of particles results in a more uniform distribution of surface residual stresses. Conversely, an increase in the number of impacts causes stress concentration to intensify at the particle's contact point, and thus a deeper distribution of residual stress is observed. This study illustrates how the mechanical properties of polycrystalline γ-TiAl alloy are affected by the process parameters of supersonic fine particle bombardment in terms of atomic size in order to develop and select the optimal supersonic fine particle bombardment parameters.