Hydrogen bond competition between chemical groups: new methodology and the Cambridge Structural Database

Abstract

Abstract The probabilities of formation of intermolecular hydrogen bond interactions between chemical groups have been studied using new methodology to extract information from the Cambridge Structural Database (CSD). Data for 41052 crystal structures containing at least one strong hydrogen bond donor have been analysed using the RPluto program, assigning 108 chemical group codes to atoms, and creating tables of hydrogen contacts for subsequent analysis using the Access relational database software. This has enabled the study of competition effects where there are specified limited numbers of chemical groups in a structure, which is often difficult with the standard CSD search program, ConQuest. There are sufficiently high numbers of certain combinations of groups to make significant observations of the preference of a given donor for choices of acceptor atoms. For example, COOH…COOH contacts are frequently disrupted by groups such as keto carbonyl, whereas CONH…CONH is very robust and is seldom disrupted. There are a surprising number of structures that do not present any intermolecular hydrogen bond interactions, often due to intramolecular hydrogen bonding taking preference. There is a tendency to use as many groups as possible to build intermolecular hydrogen bond networks. An estimate is made of the relative strengths of self-association interactions which are, in decreasing order, CONH, COOH, alcoholic OH and phenolic OH. The frequency of preferred contacts in some cases allows one to predict the most probable contacts for a given molecule with a specific combination and ratio of chemical groups.