Extracting Metal‐Organic Frameworks Data from the Cambridge Structural Database

Most of the synthesized crystal structures accompanying academic publications are deposited in the Cambridge Structural Database (CSD), curated by the Cambridge Crystallographic Data Centre (CCDC) [1]. This database contains data of over 1 million experimentally-obtained crystal structures resulting from X-ray diffraction analyses. Among this myriad of crystalline data are, in particular, those of metal-organic frameworks (MOFs). Over the last two decades, their customizability, combined with the variety of possible cluster-ligand combinations and their relatively straightforward syntheses, has spurred the creation of a fast-growing number of structures covering a range of pore sizes, geometries, internal surface areas, and pore volumes. Our questions include how many MOFs have been synthesized so far? How to specifically extract MOFs from the CSD? What are the structure–property landscapes of these existing MOFs? This chapter answers these questions by presenting the construction of the CSD MOF subset, the world's first automatically-updated dataset of MOFs composed of (as of 2020) nearly 100,000 structures [2]. The CSD tools used in the process are explained and further exploited to demonstrate possible targeted, dynamic searches of sub-classes of MOFs, based on their cluster-ligand linkages, surface functionalities, framework dimensionalities, or chiralities [3]. The extracted structural information is particularly valuable when combined with application-specific descriptors, usually obtained with high-throughput molecular simulations. To ensure the data are suitably clean for this process, we also present ways of removing undesired solvents and adding missing hydrogen atoms [2,4]. The resulting structure–property maps are the primary tools for revealing trends and identifying the best candidates for a given application.