RESUMEN
Large-language models (LLMs) such as GPT-4 caught the interest of many scientists. Recent studies suggested that these models could be useful in chemistry and materials science. To explore these possibilities, we organized a hackathon. This article chronicles the projects built as part of this hackathon. Participants employed LLMs for various applications, including predicting properties of molecules and materials, designing novel interfaces for tools, extracting knowledge from unstructured data, and developing new educational applications. The diverse topics and the fact that working prototypes could be generated in less than two days highlight that LLMs will profoundly impact the future of our fields. The rich collection of ideas and projects also indicates that the applications of LLMs are not limited to materials science and chemistry but offer potential benefits to a wide range of scientific disciplines.
RESUMEN
To rank the performance of materials for a given carbon capture process, we rely on pure component isotherms from which we predict the mixture isotherms. For screening a large number of materials, we also increasingly rely on isotherms predicted from molecular simulations. In particular, for such screening studies, it is important that the procedures to generate the data are accurate, reliable, and robust. In this work, we develop an efficient and automated workflow for a meticulous sampling of pure component isotherms. The workflow was tested on a set of metal-organic frameworks (MOFs) and proved to be reliable given different guest molecules. We show that the coupling of our workflow with the Clausius-Clapeyron relation saves CPU time, yet enables us to accurately predict pure component isotherms at the temperatures of interest, starting from a reference isotherm at a given temperature. We also show that one can accurately predict the CO2 and N2 mixture isotherms using ideal adsorbed solution theory (IAST). In particular, we show that IAST is a more reliable numerical tool to predict binary adsorption uptakes for a range of pressures, temperatures, and compositions, as it does not rely on the fitting of experimental data, which typically needs to be done with analytical models such as dual-site Langmuir (DSL). This makes IAST a more suitable and general technique to bridge the gap between adsorption (raw) data and process modeling. To demonstrate this point, we show that the ranking of materials, for a standard three-step temperature swing adsorption (TSA) process, can be significantly different depending on the thermodynamic method used to predict binary adsorption data. We show that, for the design of processes that capture CO2 from low concentration (0.4%) streams, the commonly used methodology to predict mixture isotherms incorrectly assigns up to 33% of the materials as top-performing.
RESUMEN
The heat capacity of a material is a fundamental property of great practical importance. For example, in a carbon capture process, the heat required to regenerate a solid sorbent is directly related to the heat capacity of the material. However, for most materials suitable for carbon capture applications, the heat capacity is not known, and thus the standard procedure is to assume the same value for all materials. In this work, we developed a machine learning approach, trained on density functional theory simulations, to accurately predict the heat capacity of these materials, that is, zeolites, metal-organic frameworks and covalent-organic frameworks. The accuracy of our prediction is confirmed with experimental data. Finally, for a temperature swing adsorption process that captures carbon from the flue gas of a coal-fired power plant, we show that for some materials, the heat requirement is reduced by as much as a factor of two using the correct heat capacity.
