By Howard Colquhoun, University of Reading

In 1978, while a lecturer at the University of Sheffield, Fraser was awarded an SRC Cooperative Research Grant to work "on secondment" at the ICI Corporate Research Laboratory in Cheshire, where he was to be based in the Catalysis Group led by Warren Hewertson. This award was intended to allow him to work more or less full time on research at ICI, and he and his family moved to Chester to be nearer the Lab. However, I do recall that Fraser was actually still teaching at Sheffield, on a day-visit sort of basis, and that somehow he also continued to maintain a research group there. (Fraser - then as now - always seemed to be doing about two-and-a-half normal people's jobs quite satisfactorily, as was once also said of Sir John Cockcroft).

At the Corporate Lab he worked in a windowless, broom-cupboard-sort-of-an-office at the top of the building, from which he supervised a group of PhD students, including David Holland and Tom Crawshaw who were working on carbohydrate and crown ether derivatives as potential ligands and auxiliaries for chiral catalysis. This group was based in lab C04, where I also happened to be working - on dinitrogen-transition metal chemistry - having joined the Corporate Lab in 1977 (after an MA in Cambridge, a PhD in London with Bernard Aylett and a postdoc at Warwick with Malcolm Wallbridge).

One day in early 1980, Fraser and I were discussing crown ether chemistry in the context of Tom Crawshaw's project, and the idea gradually emerged that ammonia, when co-ordinated to a transition metal centre, might form three strong(ish) hydrogen bonds to 18-crown-6 in exactly the same way that an alkylammonium ion does. Neither of us can recall this key conversation in detail - Fraser seems to think I suggested the idea, whereas I am quite convinced that he produced it - but in any event it probably arose from our thinking that (i) most catalysts are transition-metal based, (ii) ammonia binds as a ligand to all sorts of transition metals, and (iii) a chiral crown ether derivative, if it did indeed hydrogen bond to co-ordinated ammonia, might act as a sort of supramolecular chiral ligand and thus be able to influence the enantiomeric selectivity of a chiral-catalytic reaction.

Fraser was also aware of instances where attempts to complex 18-crown-6 directly to transition metal cations had instead produced crystals in which the crown ether was found to be hydrogen-bonded, in the solid state, to metal-coordinated water. So, as the postulated crown-ammonia-metal interaction seemed a quite reasonable and potentially very exciting idea, I said I would go away and have a look in the literature. This produced lots of evidence, going back many years, for coordinated ammonia being able to form hydrogen bonds in a general sort of way, e.g., to solvents, and to counterions in cationic complexes, but nothing at all about its possible interactions with crown ethers.

An especially pretty technique for detecting hydrogen bonding to coordinated ammonia had been described by Joseph Chatt (a particular scientific hero to us, and often an unknowing godfather to our work - see below) who had found that the infrared stretching frequency of the carbonyl ligand in the cationic complex [Ru(NH₃)₅(CO)]²⁺ was highly dependent on the nature of the counterion. This frequency was much lower for chloride or bromide than it was for hexafluorophosphate or tetrafluoroborate, and Chatt had shown that this reflected the strength of hydrogen bonding between the anion and coordinated ammonia in the complex cation. Hydrogen bonding was much stronger for the small halide ions than for the larger counterions, and this led to increased electron density on the metal and so in turn to a reduction in the CO stretching frequency.

I knew about Chatt's work and, as a test of the predicted crown-ammonia-metal interaction, I suggested we should use a complex with just one ammonia ligand, [CpFe(NH₃)(CO)₂]⁺[BPh₄]^-, in which the CO ligands would act as infrared-probes. As a control we used the corresponding complex in which pyridine replaced the NH₃. Addition of dibenzo-18-crown-6 to a solution of the ammonia complex indeed produced a marked shift of the two carbonyl bands to lower frequency, but the same experiment with the pyridine complex left the carbonyl bands entirely unmoved, consistent with strong H-bonding between the crown ether and the ammonia ligand. (I don't remember needing to ask anyone's permission to start this work: the Corporate Lab was an amazingly flexible research environment in those days).

However, I was not able to isolate any specific adduct from the iron-crown system, and so we next decided to look at an even simpler complex [trans-(Me₃P)PtCl₂(NH₃)] which had no charge and so no counterion to worry about. It was also more symmetrical than the iron complex and so potentially more likely form crystals as an adduct. More persuasively still, Chatt and Venanzi had shown a quarter of a century earlier, using infrared methods, that strong intermolecular hydrogen bonds existed in the solid state between the ammonia and chloride ligands in this type of platinum complex.

I was fortunate in being able to synthesise [trans-(Me₃P)PtCl₂(NH₃)] very easily, by making use of ICI's unique store of research materials in the "Chatt Archive". This collection of transition metal compounds dated from the early 1950s, when Chatt had led a pioneering ICI research group which effectively laid the foundations of modern coordination chemistry. One of the platinum-phosphine compounds we used even carried a pencilled label, "L. Venanzi"! The archive, which at that time still survived in the basement of the Corporate Lab, had been moved from Chatt's former laboratory at The Frythe in Hertfordshire, a country house establishment which, before ICI established a research centre there after the war, was home to the Technical Section of SOE. It was apparently known as "Station IX" and specialised in the design of secret weapons for the French Resistance.

Our new platinum-ammonia complex quickly gave crystalline adducts with both 18-crown-6 and dibenzo-18-crown-6, and David Williams at Imperial College, London - subsequently, if not already by then, one of the world's great crystallographers - was able to solve their structures and so prove conclusively the existence of multiple hydrogen bonds between the molecular components.

Fraser and I decided to try and extend these results by working with larger crown ethers, up to dibenzo-30-crown-12, and by using different co-ligands as 1H NMR probes for adduct formation. We had worked out by this time that, when the crown ether interacted with co-ordinated ammonia, the aromatic rings of the crown produced ring-current shifts in resonances associated with ligands adjacent to the ammonia. This ring-current shift in fact proved an exquisitely sensitive tool for detecting such interactions, and it has played an absolutely key role in much subsequent research. We had also realised that the metal-ammonia-crown binding phenomenon was really a special case of "second-sphere coordination", a concept going back to the earliest days of coordination chemistry.

Fraser's group in Sheffield (specifically I think John Wolstenholme) was co-opted to synthesise the larger crown ethers, and I designed a series of transition metal complexes containing both multiple ammonia ligands (to increase the number of potential hydrogen bonds) and also suitable probe ligands for 1H NMR work. These ligands included 2,2'-bipyridyl which I chose because I suspected - probably wrongly - that it could act as a pi-acid ligand and so perhaps have the capacity to increase the hydrogen-bonding potential of any ammonia co-ligands. We despatched these compounds to Sheffield (I think Fraser must have taken them in his car) for NMR experiments with the larger dibenzo-crown ethers.

One day shortly after this, coming back from lunch, I met Fraser who excitedly told me that one of the larger crown ethers, dibenzo-30-crown-10, when added to a solution of the cationic bipyridyl platinum complex [Pt(bipy)(NH₃)₂]²⁺, was producing absolutely enormous shifts (more than a ppm) in the bipyridyl proton NMR resonances, implying the existence of an extremely strong set of interactions between the two components.

Obviously we needed to try and crystallise something to see what was going on, and after Fraser brought some dibenzo-30-crown-10 back from Sheffield I eventually managed to isolate some goodish crystals of a 1:1 adduct with [Pt(bipy)(NH₃)₂][PF₆]₂. The crystals were sent off to David Williams in London, and a couple of weeks later he produced the structure which, sure enough, showed an awful lot of hydrogen bonds. To our astonishment however, it also showed a perfect triple pi-stack between the electron-deficient bipyridyl and both the electron-rich aromatic rings of the crown ether. We only needed one look at the structure to see it was a charge-transfer complex! Regrettably, either of us had previously spotted that the adduct was a rather deeper shade of yellow than its parent platinum complex (to be fair the difference was not that great), but we soon showed that this additional depth of colour corresponded to a clearly identifiable charge-transfer band in the visible spectrum.

A year or so later, after we had patented and then published these results, it was time for Fraser to return to Sheffield. He gave a final seminar in the Corporate Lab conference room, setting out all the results achieved during his time there. At the end of the talk my one-time ICI boss, Eric Goodings (an outstandingly far-sighted chemist who had actually retired by then, but had come in specially for Fraser's talk) pointed out that ICI's bipyridyl-based herbicides paraquat and diquat were well known to form charge-transfer complexes, and why didn't we see if dibenzo-30-crown-10 would bind to these herbicides in the same "sandwiching" sort of way we had found for coordinated 2,2'-bipyridyl?

Since diquat was based on the same 2,2'-bipyridyl unit as our platinum complex, we thought Eric's idea well worth a try, and we quickly obtained a sample of diquat from ICI's Plant Protection Division at Jealott's Hill. I converted it to the hexafluorophosphate salt (more soluble in organic solvents than the commercial bromide salt), dissolved this up in dichloromethane and tried adding an equivalent of dibenzo-30-crown-10. The solution immediately changed from colourless to deep red. This observation strongly suggested formation of a charge-transfer complex, but I failed to isolate anything interesting from this solution (the two components always tending, in my impatient hands, to crystallise separately). This was I think the last experimental work I did with Fraser, as by this time I was heavily involved in ICI's new research programme on high-temperature aromatic polymers and had little free time left for what was then still a very academic topic. However, some months later John Maud - working with Fraser in Sheffield - did manage to isolate the dark red diquat-crown-ether complex which I had seen briefly in solution. David Williams eventually solved its structure (a particularly difficult crystallographic problem) and showed that it contained a triple pi-stack entirely analogous to that of the platinum bipyridyl adduct: the resulting papers on diquat complexation carried, inter alia, the name of Eric Goodings!

Fraser, David Williams and I pursued research on second-sphere coordination for a few more years, mainly through a CASE PhD studentship (the student was Simon Doughty, who did a very nice job), and in 1986 we wrote a comprehensive review of our work in this field for Angewandte Chemie.

I spent some 20 years in high-performance polymer research, both industrial and academic, before returning almost accidentally a few years ago to work on the interactions of electronically complementary pi-systems. This is now in the context of using tweezer-type aromatic molecules to "read" polymer sequence-information (ironically at the University of Reading). Once again, the key result leading into this work - the complementary structure of a tweezer-macrocyle complex - emerged from the outstanding structural work of David Williams, by then Professor of Crystallography at Imperial College.

Meanwhile Fraser, still working closely with David, had managed to raise ARC funds to design and synthesise crown-ether-based receptors for paraquat, and from 1982 it was this extremely difficult and challenging work (in which I was not involved) which led Fraser, first in Sheffield, then in Birmingham and now at UCLA, into the world of the mechanical bond and on to the now famous catenanes, rotaxanes, molecular knots, rings and functional molecular machinery which have come to define the Stoddart Era in Chemistry.

June 05, 2007