Search

This paper describes a proof-of-concept example for a method that allows the calculation of a similarity index between whole molecular crystal structures; this has been termed ‘structural genetic fingerprinting’. It is based on the use of fingerprint plots derived from Hirshfeld surfaces coupled with cluster analysis and associated multivariate statistics. Using this formalism, it is possible to show quantitatively (using correlation coefficients) that, for example naphthalene is more similar to anthracene than to benzene, and moreover that benzodicoronene is more similar to anthrabenzonaphthopentacene than naphthalene is to anthracene. Whereas the correlation coefficients themselves obtained say nothing about the ways in which the patterns of intermolecular interactions are similar or different for two different structures, the fingerprint plots do contain such information. In principle this method for quantifying structural similarities of whole molecular crystal structures should be both robust and generally applicable. This method is potentially applicable to datasets consisting of many hundreds or even thousands of structures.

Cambridge Structural Database (CSD) searches can produce thousands of ’hits’ especially if a simple fragment is used, and as a result, processing and interpreting the results becomes a considerable task, and one in which it is easy to make mistakes. Cluster analysis using dendrograms, metric multidimensional scaling and suitable visualization tools can reduce the workload to a few hours withminimal user intervention, and thusminimal user bias. The real beauty of the method is in its interactivity and scalability: it allows you to go from an overview of the entire dataset and easily spot any outliers or errors, to the ability to ‘drill down’ within a cluster, looking atmore andmore detailed differences between different fragments in a statistically rigorous way. The formalism and the associated program are also invaluable in crystal engineering and structure prediction environments since it can readily identify the basic motifs in a structure and how they pack in the unit cell. The computer program we have developed to do this is dSNAP [1]. It is available free of charge from the web site http://www.chem.gla.ac.uk/snap.

Intermolecular hydrogen bonding plays an important role in forming anisotropic interactions in condensed systems, and subtle competition between H-bond acceptors/donors can lead to dramatically different solid-state structures. There is significant national and international research effort directed at the study of hydrogen bonding in the solid state. Much of this work has focussed on the static structures adopted by molecular hydrogen-bonded systems, but the importance of hydrogen atom migration through hydrogen bonds between molecules has also been identified. It is becoming increasingly apparent that the positions of the protons involved in hydrogen bonds are highly susceptible not only to chemical environment, but also to the effects of temperature and pressure. A collaborative project has been set up with the aim of preparing molecular adducts in which a proton migrates between the donor and acceptor species as temperature and/or pressure is varied. In this poster we will describe our experimental techniques [1] and present results for some of our variable temperature studies of the 1:1 salt formed between 2-iodoaniline and picric acid [2].