Kekulé on the Nature of Carbon
Kekulé spent the fall of 1855 in Darmstadt pondering his future. Brother Charles had offered to set him up in London as a practical chemist, but Kekulé had regarded that as the most miserable sort of "chemical woodchopping" and "bootpolishing." On a six-week journey, he traveled to Giessen, Marburg, Frankfurt, Wiesbaden, Bonn, Berlin, and Heidelberg to explore the possibility of habilitation, speaking personally with Will, Kopp, Kolbe, Fresenius, Böttger, Baumert, Mitscherlich, Rose, Bunsen, and others. Kekulé decided that Heidelberg offered the best opportunity for an organic theorist since the dominant figure, Bunsen, was an inorganic and physical-chemical empiricist. Ever since Liebig's transfer to Munich and effective retirement from the laboratory, Bunsen's had become the most famous teaching lab, which meant a large number of potential students, hence income. After lengthy negotiations, Charles Kekulé was finally persuaded, against his better judgment, to advance his brother enough money to enable him to attempt "a stupid scholarly career."[31]
In January 1856 Kekulé moved to Heidelberg and two months later had obtained the necessary permissions and passed the examination for
the venia legendi . Bunsen was not in the habit of providing laboratory facilities for Heidelberg Privatdozenten who were not also his assistants. Kekulé was therefore forced to find facilities for himself, and his relationship with Bunsen always remained cool, although each recognized the other's merits. He rented a house on the Hauptstrasse, set up his residence as well as a laboratory and lecture room there, hung out his shingle, and began to take on paying customers for summer semester 1856. His first Praktikant was his old friend Reinhold Hoffmann, his second a promising young man named Adolf Baeyer (who had also studied directly with Bunsen). With Kekulé working alongside them, the small lab was filled to capacity.
Kekulé's rent was reduced the following year by the offer of Emil Erlenmeyer to share the expenses and use of his lecture room, which was considerably more spacious than the laboratory, as it could accommodate ten chairs. Erlenmeyer, a pharmacist who had also earned a chemistry Ph.D. from Liebig in Giessen, had been living in Heidelberg since 1855 doing consulting for artificial fertilizer manufacturers and pursuing, with mixed success, some entrepreneurial activity. Using an advance from his father-in-law for support, he habilitated in 1857 and thereafter entered into an academic career.[32] Kekulé, Baeyer, and Erlenmeyer became good friends and learned much from one another.
According to his later autobiographical sketch, Kekulé wrote down a version of his structure theory shortly after his arrival in Heidelberg (it must have been the spring or summer of 1856), and showed it to two close friends (Baeyer and Erlenmeyer?), both of whom expressed doubts. Deciding that either the time or the theory was not yet ripe, he laid the manuscript away in a drawer. Apparently, he decided it was wiser to concentrate on producing a number of small but respectable empirical studies before coming forward with a broad-ranging theoretical interpretation of organic chemistry.[33]
But Kekulé was incapable of producing empirical studies that did not directly relate to his theoretical obsessions, and the first work to emerge from his new laboratory illustrates this. Sent to Liebig at the end of December 1856, Kekulé's paper attempted to settle the controversial topic of the constitution of the fulminates. Kekulé's analyses and specific constitutional proposal in this paper are less important than the fact that he took the opportunity to defend a marsh gas type that could be used to represent many small organic molecules. He carefully added that he was using the word and concept type not in Gerhardt's positivistic "synoptic" sense but rather in the sense of Dumas' mechanical types. He thus emphasized the structuralist con-
notations of the newer type theory in the tradition of Williamson. The same explicit reference to mechanical, that is, structural, types occurs in Baeyer's first paper from Kekulé's lab.[34]
In the spring and summer of 1857, papers appeared by an ausserordentlicher Professor at Göttingen, Heinrich Limpricht (a former assistant of Wöhler), and two of his assistants, Louis von Uslar (a former assistant of Kolbe) and Otto Mendius.[35] Two years earlier, Limpricht had been the first in Germany to publish a textbook based on Gerhardt's type formulas, but in Kekulé's opinion he had not properly understood the true sense and implications of the newer type theory. These two papers gave Kekulé the opportunity he had been looking for to publish his structural ideas, for Limpricht had opened himself to criticism in several ways, and such a critique provided the rhetorical launching platform for more general considerations.
Kekulé's specific criticisms[36] focused on the concept of copulated compounds, on Gerhardt's "basicity law" to which Limpricht made repeated reference, and on Limpricht's retention of the oxygen equivalent O = 8 within the Williamson-Gerhardt water type. Kekulé was philosophically troubled by Limpricht's use of types in a purely schematic rather than realist sense. This is what he meant when he wrote, seven months later, that "some chemists have fully adopted the external form of the newer type theory, while either misunderstanding or interpreting differently the underlying idea."[37]
Kekulé not only included Limpricht and his students in this group, but also Gerhardt and Odling. All these chemists had often proposed formulas that were impossible to represent consistently by the "theory of polyatomic radicals" as Kekulé conceived it.[38] In 1854 Williamson had made a similar criticism of one of Gerhardt's formulas, and Kekulé was following his friend's lead here; he also thought of Dumas and Wurtz as defenders of this realist version of type theory. The proper criterion for the newer types was whether the formulas could be constructed according to the accepted atomicities of the radicals and elements composing the compound, linked together in atomic-molecular chains. A consistent theory could be constructed following this criterion, Kekulé averred, and that theory would make the hypothesis of copulated compounds unnecessary.
His first sketch of this theory—which we know today as structure theory—is an intentionally incomplete version, dated 15 August 1857 and published in the November issue of Liebig's Annalen .[39] Kekulé here attempted to demolish the surviving remnants of copula theory and at the same time provide the first systematic exposition of the "theory of polyatomic radicals," whose origin and proper interpretation he ascribed mostly to Williamson. Molecules are "contiguous jux-
tapositions" (Aneinanderlagerungen ) of atoms, he stated, constructed according to atomic valences, which he called "atomicities" or "affinity units."[40] He thus traced the atomicities of radicals to the atomicities of the constituent atoms, which were monobasic as in hydrogen and halogen, dibasic as in oxygen and sulfur, or tribasic as in nitrogen and phosphorus. A footnote[41] asserts that carbon is "tetrabasic or tetratomic," as in CH4 or CCl4 . He defined a radical as simply that portion of a molecule that persists unchanged in any given reaction; thus different radicals may be assumed in the same molecule, depending on the reaction being considered. They are a theoretical expedient only, nonexistent in and of themselves.
Limpricht published a reply to this paper, not so much denying the truth of Kekulé's assertions but suggesting that Kekulé had distorted his views in opposing them.[42] Limpricht's reply in turn provided the inducement for Kekulé to write a more complete version of his theory, dated 16 March 1858, which was published two months later and grandly entitled "On the Constitution and Metamorphoses of Chemical Compounds and on the Chemical Nature of Carbon."[43] After a polemical introduction in which Kekulé elaborated and clarified his earlier criticisms, he boldly declared
I consider it necessary, and given the present state of chemical knowledge in many cases possible, to go back to the elements themselves of which compounds are composed, in order to account for the properties of chemical substances. I no longer consider the principal task of the times to be the determination of atomic groupings, which due to certain properties can be considered as radicals, in this way assigning compounds to a few types which thus have scarcely any more significance than pattern-formulas. I believe on the contrary that we must extend our considerations to the constitutions of the radicals themselves, that we must determine the relations of the radicals among each other, and that we must derive the nature of the radicals as well as that of their compounds from the nature of the elements. My earlier considerations on the nature of the elements and the basicity of the atoms serve as the point of departure for these views.[44]
Now, as he had failed to do earlier, he explored the implications of considering carbon as a tetratomic element. Just as examples were known in which atoms of a single element could link to each other—especially in the elemental gases hydrogen, oxygen, nitrogen, and chlorine—carbon could reasonably be supposed to do the same. For compounds containing two carbon atoms, which possess between them eight "affinity units," two units, one on each atom, must be used to link them together. Hence, six affinities remain for attachment to other
atoms, and C2 is indeed a "hexatomic radical" as in, for example, the ethyl series. Every subsequent carbon atom added to the "skeleton" (this neologism first occurs in this article) requires two valences to be incorporated into the chain and so adds a total of two free valences to bond hydrogen or other atoms.[45]
Of course, other bonding patterns exist as well. Hydrogen atoms may be only indirectly bonded to the carbon group, that is, they may be held only through the intermediacy of a polyatomic atom such as oxygen or nitrogen, which is itself bonded to the carbon chain—such as in alcohols. Alternatively, more than one carbon group may be indirectly held together through the intermediacy of such a polyatomic atom—such as in ethers. In either case, those atoms that are indirectly or incompletely bonded to carbon, that is, those atoms in which not all of their affinity units (valences) are bonded to one and the same carbon group, Kekulé began to call "typical" atoms (that is, atoms of the type). Typical atoms in this sense thus included the replaceable hydrogen atoms of alcohols, acids, and amines, the oxygen atoms of alcohols, ethers, acids, and esters (except the carbonyl oxygens), and the nitrogen atoms of amines. Kekulé later defined a radical as a group consisting of the basic hydrocarbon skeleton plus all the atoms directly and completely bonded to it, i.e., excluding only the typical atoms.[46]
This definition of radicals is an extension of Liebig's and Laurent's idea of distinguishing atoms that are "inside or outside of the radical," the distinction having been provided by the type notation. For many years Kekulé used this more theoretical definition for taxonomic purposes, while simultaneously applying his more empirical and flexible definition stated in the previous paragraph. But the former definition is only slightly less general than the latter, for atoms "of the radical" are only occasionally attacked, which is to say that most chemical reactions exclusively involve atoms "of the type," which compose what modern chemists refer to as functional groups.
Given such an ambitious program, the epistemological question of the determinability of atomic arrangements within molecules through the study of chemical reactions was a point that Kekulé needed to address directly. Kekulé was a student of the structural agnostic Gerhardt and of the cautious realist Williamson (who had studied directly under Auguste Comte, the father of positivism). Consequently, he was sensitive to the need for epistemic caution. He emphasized that the determination of the actual physical arrangements of atoms was as yet impossible. The chemist needs to exercise extreme care in drawing conclusions about "constitutions" from the study of chemical reactions since reactions necessarily produce intermolecular and some-
times also intramolecular rearrangements. "Rational formulas" are nothing but "reaction formulas," he proclaimed, and can be nothing more in the current state of the science.[47]
But Kekulé believed, pace Gerhardt, that the chemist is by no means powerless to gain information on constitutions, especially if one proceeds while simultaneously renouncing the attempt to determine actual spatial atomic arrangements. At the end of his 1858 paper he proposed three axioms (above and beyond simple application of the valence rules) that he believed were justified to be used in determination of constitutions. First, homologous series are to be constructed schematically by using the "simplest arrangement" of carbon atoms, in which each carbon atom in a chain is connected to each of its neighbors by a single affinity unit (Kekulé essentially ignored olefinic and aromatic compounds in this article). Second, in reactions that proceed without loss of carbon, it is to be assumed that the constitution of the carbon skeleton of the reactant is preserved in the product. Finally, when the carbon skeleton is broken up, aspects of its original constitution may be deduced from the constitutions of the products.[48]
The last words of Kekulé's article were a formalist obeisance to positivist ideals, a safe cautionary note for a young man attempting against bad odds to make a career—moreover, a statement that appears to have been closely modeled on the final paragraph of Wurtz' July 1855 paper. Kekulé wrote
Finally, I think I should emphasize that I myself place only a subordinate value on considerations of this sort. But since, in the total absence of exact-scientific principles in chemistry, one must for the moment be content with conceptions based on probabilities and convenience, it seemed appropriate to communicate these views, since, it appears to me, they give a simple and rather general expression especially for the most recent discoveries, and therefore their application may perhaps facilitate the discovery of new facts.[49]
Of course, Kekulé did succeed in his career, his first call coming just four months after his second structure theory paper appeared. When Jean Servais Stas traveled to Germany to seek a successor to D. J. B. Mareska at the University of Ghent, his attention was first drawn to Limpricht. However, both Liebig and Bunsen pointed out to him that Kekulé was far more cosmopolitan, would have no trouble in lecturing in French, and had already earned a sterling teaching reputation. (When Privatdozent Ludwig Carius took over the organic chemistry course after Kekulé's departure from Heidelberg, he was shocked to see Kekulé's eighty auditors gradually dwindle to six!)[50] Impecunious,
energetic, and ambitious, Kekulé was overjoyed at the call, even though the exigencies of becoming a Belgian scholar were daunting, and the Ghent laboratory proved to be stripped bare upon his arrival in the middle of November 1858.
A letter from Kekulé to Erlenmeyer gives a good impression of his activities during his first semester in Ghent, chief of which was the writing of a textbook of organic chemistry:
I have again been working regularly until one or two o'clock, often, as I did in Heidelberg, until 11:30 or 12 on my textbook and only then beginning to piece together a lecture. . . . In the morning at 9 I go to the laboratory, lecture three times a week from 10 to 11:30 . . . at 12 a frugal bite to eat in the lab, at 4:30 dinner—the high point of the day, and my only pleasure in life other than cigars (genuine Havannas, I can afford them now!)[51] —then a strong cup of coffee, and around 6 or 6:30 back to my digs, first for digestion and then for work. My lifestyle is simple. No diversions except Sundays, when I eat around one and then take my required constitutional walk. No longer do I go to theater; concerts and balls are neglected; no beer; no social visits. If with such a lifestyle I fail to achieve the dignity of "scholar" it is not my fault; of "man" there is little enough left over.[52]
In his old age, Kekulé recalled that Liebig told him that he who does not ruin his health with overwork would never make it in academic chemistry. Kekulé averred that for years he followed this advice to the letter.
Kekulé did of course achieve "scholarship," one sign of which was the appearance of his textbook. Like Kolbe's, it was published in fascicles, the first (240 pages) appearing in June 1859; after 1867, when it was about two-thirds completed, Kekulé ceased writing. Also like Kolbe's, the first fascicle contained an interesting and perceptive, if one-sided, history of recent chemical theory. It also contained a longer revised version of Kekulé's theory of the atomicity of the elements (structure theory) and unveiled in published form Kekulé's idiosyncratic sausage formulas.
This work enjoyed tremendous success, far outstripping Kolbe's textbook. Not only did the fascicles appear faster, but it was both modern and imaginative in its outlook, and chemists took notice. It proved an effective means of publicizing and propagating the views of the newer French-English school to new audiences, especially to German students. After 1865, it also helped quickly to establish Kekulé's benzene theory. When the first fascicle appeared, Rudolf Fittig said he read it with such excitement that he could hardly put it down. His testimony is significant, not least because he was a student of Wöhler and Limpricht
at Göttingen and so had no direct or indirect personal connection with Kekulé. He was referring to this period when he wrote, toward the end of his life, that "in a certain sense all of us were Kekulé's students."[53] This was not true for Kolbe, however, who later commented that he barely glanced through it since, as he said, he was sure he could learn nothing from it.[54]