S.M.J. Rogge

Soft polymorphic materials, such as metal–organic frameworks (MOFs) and covalent organic frameworks (COFs), often display distinct anisotropy. Yet, their phase transition behavior has been predominantly characterized under isotropic stimuli, such as temperature or pressure variations, up to now. In this work, we employed the Cauchystat to investigate how MIL-53(Al) and COF-5, two prototypical soft porous crystals, respond to anisotropic stresses instead. For MIL-53(Al), we showed that normal stresses induce a phase transition already at stresses below the critical hydrostatic pressure, depending on the directionality of the applied stress. For COF-5, we determined the critical shear stress needed to induce a layer instability, leading to delamination. In both cases, we highlighted the importance of selecting adequate values of the Cauchystat control parameters to obtain accurate predictions. Based on these insights, we formulated best practices to simulate phase transitions in soft porous crystals under nonhydrostatic loadings, which is required for, e.g., nanosensors and -dampers.

After the development of the famous “Transformer” network architecture and the meteoric rise of artificial intelligence (AI)-powered chatbots, large language models (LLMs) have become an indispensable part of our daily activities. In this rapidly evolving era, “all we need is attention” as Google’s famous transformer paper’s title [Vaswani et al., Adv. Neural Inf. Process. Syst. 2017, 30] implies: We need to focus on and give “attention” to what we have at hand, then consider what we can do further. What can LLMs offer for immediate short-term adaptation? Currently, the most common applications in metal–organic framework (MOF) research include automating literature reviews and data extraction to accelerate the material discovery process. In this perspective, we discuss the latest developments in machine-learning and deep-learning research on MOF materials and reflect on how their utilization has evolved within the LLM domain from this standpoint. We finally explore future benefits to accelerate and automate materials development research.

Postcombustion carbon capture provides a high-potential pathway to reduce anthropogenic CO₂ emissions in the short term. In this respect, nanoporous materials, such as covalent organic frameworks (COFs), offer a promising platform as adsorbent beds. However, due to the modular nature of COFs, an almost unlimited number of structures can possibly be synthesized. To efficiently identify promising materials and reveal performance trends within the COF material space, we present a computational high-throughput screening of 268,687 COFs for their ability to efficiently and selectively separate CO₂ from the flue gas of power plants using a pressure swing adsorption process. Furthermore, we demonstrate that our screening can be significantly accelerated using machine learning to identify a set of promising materials. These are subsequently characterized with high accuracy, taking into account the effects of competitive adsorption and coadsorption. Our screening reveals that imide, (keto)enamine, and (acyl)hydrazone COFs are particularly interesting for carbon capture. Additionally, the best-performing materials are 3D COFs possessing strong CO₂ adsorption sites between aromatic rings at opposite sides of pores with a diameter of 1.0 nm. In 2D COFs, a significant influence of the framework chemistry, such as imide linkages or fluoro groups, is observed. Our design rules can guide experimental researchers to construct high-performing COFs for CO₂ capture.