Sep 152007

I worked with Keith for 40 years. When I came to Oxford in 1967 as a fresh post doc, Keith was already the Lecturer in Chemical Crystallography, running a successful research group alongside those of Tiny Powell and Dorothy Hodgkin. Although he was not my supervisor, I quickly learned that it was he who held all the practical knowledge that kept the laboratory running. He had already accumulated a wide circle of colleagues throughout crystallography, so that when he could not answer a question himself, he was always on first-name terms with someone who could. Within a few years of my arrival, Tiny retired and Keith became Head of the Chemical Crystallography Laboratory. Under Keith’s stewardship the laboratory blossomed, building on the foundations laid down by Tiny and Dorothy. The laboratory attracted academic visitors from all over the world, with most of these people remaining life-long friends of Keith and Lesley.

Keith was born in Barnsley in 1934 and was a pupil at Barnsley Holgate Grammar School before coming up to Wadham College at Oxford. It was during his Part II (research project) year with Bob Williams that his interest in transition metal chemistry and in crystallography began to develop (An Investigation into Some Transition Metal Complexes with Hydrazine, Wadham, 1957). The Part II thesis also shows an interest in twinning and polymorphism, interests which were to develop during his D.Phil studies. He presented his D.Phil thesis in 1959, after only two years of very productive research. The thesis shows very clearly both Keith’s curiosity about the solid state, and his determination to work at problems until they were solved.

The thesis is also is a window onto the almost unbelievable world in which Science moved ahead without computers, but which was also being revolutionised as he worked. His first crystallographic computations were carried out using punched (Hollerith) cards on mechanical tabulators, but by the end of his project he was writing programs for the Ferranti Mercury computers. In spite of the fact that the programs were written in Autocode (a semi-machine language) Keith realised the importance of the user interface, and produced a program with a surprisingly modern keyword-based input. This early brush with computers convinced him that he did not really like writing programs, but, paradoxically, this later enabled him to father a number of important software projects. Because he was not interested in encoding the answer to a problem, his imagination was liberated to push crystallographic and chemical computing to new frontiers.

In the early 1970’s he raised a grant with John Rollett to employ Bob Carruthers to create a program for the refinement of twinned data – a program that eventually grew into CRYSTALS. It was their misfortune that the program emerged at just the time that photographic methods (which provide copious information about twinning) were being replaced by serial diffractometers (which made the handling of twins rather difficult) so that the significance of their creation was not appreciated. Keith was also instrumental in the formation of the Molecular Graphics Group, leading to the launch of the Journal of Molecular Graphics. He was consultant to James Crabbe and John Appleyard as they created the Desktop Molecular Modeller, and worked with Bob Diamond, Tom Koetzle and Jane Richardson on an enormously ambitious project with IBM and Oxford Electronic Publishing- the CD-ROM based Molecular Structures in Biology. This was a project at the very cutting edge of technology. ChemX, from Chemical Design, and Cobra, from Oxford Molecular, were developments of academic projects he started.

Keith was much more than a software innovator. It was as a chemical crystallographer that he worked in laboratories in Sweden (with Sixten Abrahamsson), in Rome with Sandro Vaciago, in Yugoslavia with Boris Kamenar, and many times in North America with Jenny Glusker and Abe Clearfield. In addition to his fascination with the chemistry of copper and molybdenum, Keith was drawn to studying charge transfer complexes, cyclopentadienyl complexes, pharmaceutical materials and liquid crystals. He had a long collaboration with Andy Vinter and Anthony Roe at Smith, Kline & French, during which he was instrumental in the milestone solution of the structure of cimetidine from powder data. In the years leading to his retirement he reprised his old interest in the organic solid state, and was a strong advocate for the complimentary use of X-ray crystallography and solid state NMR.

Although he rarely appeared on stage at BCA meetings, Keith was an accomplished speaker, with the knack of appearing to talk with the audience rather than at them.  In the laboratory, however, he was much more than just a talker – he was a doer. It was not unusual to find him with his sleeves rolled up, half way inside an X-ray generator or re-building a vacuum line. He was a hands-on scientist, and loved practical work of every kind. This led him into doing research in all the difficult corners of crystallography, so that his experience was both deep and wide.

His other real love was teaching, and it is perhaps through this that he will be most fondly remembered. He had the knowledge and insight to debate problems with the best students, and the patience and understanding to help the weaker ones. Even after his retirement he continued to come into the laboratory, and clearly took great satisfaction from sharing his expertise with young research workers and visitors.

Keith was what one might call an old-fashioned scientist, coming from a culture that has now slipped away. For his generation it was not necessary, or even seemly, to promote oneself at every opportunity. He was not interested in joining the International Lecture Circus. Because of this, he was able to devote much of his time and energy to the well-being of his colleagues, staff and students, and to providing the environment and encouragement needed for first class research. The measure of his success is both the scope of his published work, and the large number of crystallographers in both the UK and the rest of the world whose training started or developed in the Chemical Crystallography Laboratory, Oxford.

In spite of his BBC accent, Keith remained a Yorkshire man all his life – bluff on the outside but gentle within. He could not abide evasiveness or deceit. He told things as they were. I was once summoned to see the Head of Inorganic Chemistry. The day before the interview Keith took me aside and suggested that I might look neater than usual for the occasion. Almost certainly good advice, but its effect was not lasting. He had the same attitude to science. If he saw a flaw in a project or results he would say so, usually coming directly to the point. This directness did upset some people, but most of the crystallographic community valued his judgments.

His astute assessment of situations was not restricted to science. Keith was always active in the affairs of Oriel College, being a fellow from 1966 until 2001, and then an Emeritus Fellow until his early death. He took on the roles of Vice Provost, Acting Provost, Senior Tutor and finally Tutor for Graduates. He was always willing to listen to other people’s problems, and if appropriate, offer advice or help. Underneath the brooding eyebrows there lived a shy and kindly man. He will be remembered as a remarkable scientist and a loyal friend.

David Watkin

 Posted by on September 15, 2007  News
Apr 092007

Acta. Cryst. (2007), 63, 303-308. [ doi:10.1107/S0108768106055212 ]

A new polymorph of 2,4-dihydroxybenzoic acid is reported. The structure was characterized by multiple-temperature X-ray diffraction and solid-state DFT computations. The material shows a geometric pattern of hydrogen bonding consistent with cooperativity between the intermolecular carboxylic acid dimer and intramolecular hydrogen bonds. The presence of proton disorder within this hydrogen-bond system, which would support such a cooperative model, was not fully ruled out by the initial X-ray studies. However, solid-state calculations on the three possible end-point tautomers indicate that the dominant crystallographically observed configuration is substantially lower in energy than the other tautomers (by at least 9 kJ mol-1), indicating that no disorder should be expected. It is therefore concluded that no disorder is observed either in the intra- or intermolecular hydrogen bonds of the title compound and that the cooperativity between the hydrogen bonds is not present within the temperature range studied.

Electronic reprints

Publisher’s copy