FERROCENE: IRONCLAD HISTORY OR RASHOMON* TALE?
Pierre Laszlo and Roald Hoffmann
A critical stance is essential to science. Proving other people wrong is a favorite private and public satisfaction -- there is nothing some people like better. But, excess zeal discounted, doubt serves as a powerful impulse to the advancement of knowledge. We document it here with the discovery of the structure of ferrocene, a story which also plays up the virtue of the spoken word.  We base it on various published (and most fragmentary) accounts, supplemented with some very helpful correspondence from colleagues.
Two groups independently reported at the very end of 1951 preparation of a C10H10Fe from cyclopentadiene. Kealy and Pauson published their contribution in the December 15, 1951 edition of Nature, while Miller, Tebboth, and Tremaine's account of their independent preparation of the identical compound, with the same "traditional" formulation (what would today be called a di-s complex), appeared a little later in the Journal of the Chemical Society. Let's turn to reception of those two papers in the Chemistry Department at Harvard University. As soon as it arrived in the library, during the first couple of weeks in January 1952, R. B. Woodward apparently saw the Kealy and Pauson piece. And he had a hunch that the structure was wrong. He set a graduate student, Myron Rosenblum, preparing not only ferrocene, but also analogs from other transition metals and cyclopentadiene. For this Rosenblum needed ruthenium trichloride. He went to Geoffrey Wilkinson, who had recently taken up an Assistant professorship at Harvard and was occupying a little closet of a room in the Mallinckrodt building, to ask for a sample of ruthenium. Wilkinson "rather brusquely wanted to know" what Rosenblum needed it for. Wilkinson writes :
Clearly, Wilkinson had already also seen the Kealy and Pauson paper and had decided to study this problematic structure of ferrocene.  Wilkinson continues:
"However, the upshot was that Woodward and I had lunch at the Harvard Faculty Club on Monday  and sorted things out. The possibility that the C5H5 ring in the iron compound could possibly undergo Friedel-Crafts or other aromatic reactions simply had not dawned on me, but other than the structure, this seemed to be Bob's main interest, whereas mine was to go on to other transition metals."
One can only bemoan the fact that Woodward left no account of the discovery, and apply all the reservations the science historian has painfully had to learn of the construction of history (yes, even by the very participants) after the fact. Myron Rosenblum offers a somewhat different perspective on the directions taken by the protagonists:
"Woodward certainly had an interest in extending the metallocene series vertically along the periodic table and possibly horizontally as well, since I set up four reactions simultaneously on January 31st. These involved the reactions of NiCl2, CoCl2, CrCl3 and RuCl3 with cyclopentadienyl magnesium bromide. From my research notebook I see that I had actually made two attempts to prepare anhydrous RuCl3 on January 16, five days prior to my first preparation of ferrocene itself.... But in any event, Woodward and Wilkinson had apparently reached a private agreement which apparently left this initiative (extending the ferrocene family) to Geoff. "
Woodward's main contribution to the ferrocene story was the divination of the sandwich structure, and a beautiful piece of insight about its aromaticity. Of the latter Rosenblum writes:
"On a Thursday evening, March 13th (1952), as we were taking our seats before that evening's seminar, we were chatting about some of the chemistry of this new substance, when [RBW] suggested simply, almost matter-of-factly, that I might want to try a Friedel-Crafts reaction on this new substance. ... I put off doing the reaction until Monday, March 17th...the product was a beautiful red crystalline material with an enormous carbonyl stretching absorption in the infrared.  The correct ferrocene structure and its aromaticity were published in the form of a communication in JACS. "
Meanwhile, during the spring of 1952, E.O. Fischer was independently studying the structure of ferrocene in his laboratory at the Technische Hochschule in Munich (with W. Pfab). Basing his conclusions on evidence from X-ray crystallography, and on very reasonable bonding notions, he also concluded that the molecule was best formulated as an iron(II) atom in-between two cyclopentadienide rings as ligands. Fischer and his coworkers immediately synthesized the ferrocenium cation, and started exploring cognate molecules such as the corresponding cobaltocene. The structures proposed by Wilkinson, Woodward, Rosenblum, Fischer and Pfab were truly revolutionary. Contemporary chemists were shocked; here is a recent account by one of the greatest of structural chemists, J. D. Dunitz:
One afternoon, I opened the Library copy of JACS [in Cambridge, England], writes Dunitz, and came across R.B. Woodward's proposal that the molecule consists of two parallel cyclopentadienyl rings with the iron atom sandwiched between them. I was skeptical. Nothing like this had ever been seen before. On my way out of the Library I met Leslie [Orgel] and asked if he had seen this astonishing proposal. He was as skeptical as I was. When we found that the compound was relatively easy to prepare in crystalline form, I decided to determine its crystal structure and so demonstrate the incorrectness [our emphasis] of the proposed molecular structure. Within a few weeks, it became clear to us that Woodward's proposal was correct after all. There was no doubt about it. That was the marvellous thing about crystal structure analysis."
Leslie Orgel went on to devise a MO description of this novel structure, which he published together with Dunitz's structure determination. The title of their paper contained the "sandwich" descriptor   . Which stuck . Thus the ferrocene story unfolded: a conjectured structure was intuitively doubted, refuted and replaced with what we now know as the true structure. But the true structure was deemed so unusual that it itself led to an attempt at refutation. Which failed, and thereby became an outstanding piece of corroborative evidence. Our story is hardly complete! There is something to be learned (will it be done before the traces vanish?) of the complicated interactions that the experimentalists whose names we have mentioned above had with the theoreticians around them .
And the sandwich structure was apparently suggested by William v.E. Doering to Peter Pauson in September 1951, before the synthesis was published ! Also from the outset it was clear that this incredible breakthrough, opening up a whole new field within organometallic chemistry, was worthy of recognition. It led to the award of the Nobel Prize to both Wilkinson and Fischer. Woodward was inexplicably left out; he complained, of course to no avail. Access to the deliberations of the Nobel Committee still decades in the future, Thomas Zydowsky has nevertheless uncovered the fascinating correspondence that ensued, one that will also reveal Woodward's perspective on the initial discovery. But for that story you will have to wait .
We are grateful to many who have shared their memories and writings with us, most importantly Peter Pauson, Jack Dunitz, Myron Rosenblum, and Tom Zydowsky.
 We expand here upon a segment from a recent paper aimed at