GILBERT NEWTON LEWIS MEDAL ACCEPTANCE
By Linus Pauling
California Section of the American Chemical Society
Berkeley, November 27, 1951
It is with great pleasure and appreciation that I receive the first Gilbert Newton Lewis Medal of the California Section of the American Chemical Society. I have for decades been a great admirer of Gilbert Newton Lewis. I remember when I first learned about his existence -- I am afraid that it was a rather shadowy existence for me, and that it was some time before I considered him a real person, rather than a mythical figure, one of those distant beings who wrote scientific papers that could be published in the Journal of the American Chemical Society and similar journals.
When I was eighteen years old, in 1919, and was unable to return to Oregon Agricultural College to continue my studies, at the end of my sophomore year, I was offered full-time appointment as assistant in quantitative analysis. I had a very heavy teaching load -- long hours in the quant lab, and also some classroom periods with several sections of quant students. In addition, I prepared unknowns for analysis by students and carried out miscellaneous jobs in the field of quantitative analysis. There was some time, however, for me to follow my own inclinations, and the circumstance that my office was also the chemistry library, rarely entered by anyone, gave rise to a habit of reading the chemical literature. My attention was first attracted by the papers by Irving Langmuir on the electronic structure of molecules, and his reference to Lewis's 1916 paper sent me back three years in the literature of this paper. The illuminating concept of the shared electron pair chemical bond, as formulated by Lewis, made an immediate appeal to my imagination; and I believe that I decided then that the field of molecular structure and the nature of the chemical bond would be one to which I would devote attention in my later life.
The greatness of Lewis's fundamental contribution, the idea of the shared electron pair, can be realized in this modern world only if we try to put ourselves in Lewis's place. In his 1916 paper, "the Atom and the Molecule," we read, with, I think, some shock, that Lewis was not certain that the atomic number of helium is 2 -- he mentioned that, although Moseley's work supported the value of 2, the value 4, suggested by Rydberg, who had inserted two unknown elements between hydrogen and helium, might nevertheless be the correct one.
It is interesting that Lewis in his 1916 paper emphasized the cubical atom, rather than the tetrahedral atom. Toward the end of his paper he discusses the tetrahedral atom, pointing out that the formation of triple bonds and other chemical facts support the tetrahedral atom rather than the cubical atom. He mentioned that in reaching the conclusion that the tetrahedral atom is a superior concept he had had the benefit of discussions with members of the Department in the University of California, including such men as Gerald Branch.
In his Silliman Lecture, Lewis also discussed the question of absolute truth. This is an interesting question in science -- for example, was it truthful to say, before 1905, that bodies move in ways corresponding to Newton's laws of motion? This matter in relation to Lewis's tetrahedral atom was raised by a question asked me by one of my freshman students last week. he asked which was the s orbital and which three were the p orbitals in the four tetrahedral bonds formed by a tetrahedral carbon atom, as in methane. I replied, of course, that chemists now described the four orbitals of the L shell of a tetrahedral carbon in methane as four equivalent tetrahedral orbitals, hybrids of the s and p orbitals -- and then I pointed out that if it had ben the good fortune of Gilbert Newton Lewis to have determined the detailed electronic structure of atoms from chemical properties we would doubtless consider the four tetrahedral orbitals as fundamental, and would say that physicists (spectroscopists) sometimes found it convenient to hybridize these tetrahedral orbitals into one s orbital and three p orbitals. The student then asked whether the s orbital and p orbitals were the real ones, and the tetrahedral orbitals the hybrid ones, or the other way around. I explained the situation again, and pointed out that the one set had just as great significance as the other set of orbitals. Here we have an example, I think, of what Lewis had in mind in discussing absolute truth, and why he said that he sympathized with the preacher who began his sermon with the words "Paradoxical as it may seem to thee, O Lord, . . ."
After thirty-five years, the fundamental ideas about the electronic structure of molecules that were expressed by Lewis still retain their validity and significance. Since 1916 the physicists have elucidated the electronic structure of atoms and, through the theory of quantum mechanics, have provided a sound theoretical basis for the shared electron pair bond. I feel that the details of the electronic structure of molecules are all contained within the theory of quantum mechanics. In attempting to foresee the future of the Lewis theory -- the future of structural chemistry -- I have no hope that a radical new theory of molecular structure will be developed. When, after months of labor, a reasonably good solution of the Schrödinger equation is obtained for some molecule, it is always found that it agrees reasonably well with the structure of the molecule as induced from experiment; the calculated values of properties of substances, as given by their molecular structure and the theory of quantum mechanics, agree with the observed properties of substances, as given by their molecular structure and the theory of quantum mechanics, agree with the observed properties as closely as the degree of approximation in the solution of the Schrödinger equation justifies. Perhaps some improved approximate methods of the wave equation will be found. I feel sure that Gilbert Newton Lewis would not be greatly interested in these laborious calculations, these approximate methods -- except, probably, the Monte Carlo method, which might well have appealed to him.
With the development of giant calculating machines, theoretical chemistry may go into a new period -- a period in which theoretical values of the properties of substances, information about the structure of molecules and crystals, will be obtained by reasonably accurate solution of the Schrödinger wave equation. I am sure that this procedure will not, however, constitute an important part of the field of structural chemistry even in the distance future. I believe that we need to take up again the characteristic chemical method of attack on the problems of the nature of the world -- the method of induction from chemical facts, the formulation of generalizations based on experience, and independent of a priori arguments. In 1916 Lewis said that, as chemists, we should not be distracted in our task by a priori arguments such as that electrostatic repulsion of eight electrons in an atom would force them to the corners of a cube, but should rather make inductions independently from the facts of chemistry -- namely, in this case, that there are pairs of electrons at tetrahedron corners. For twenty years, since the first applications of quantum mechanics to chemistry, it has been thought hardly proper to attack chemical problems in an imaginative and independent way. Physicists, and even some more chemists, have not taken kindly to the theory of metals that I formulated a few years ago, in which copper is assigned a metallic valency of 5.44, and other elements are assigned similar unexpected valencies. I foresee that a period of significant development in structural chemistry can take place in the future through the renewed effort to apply this characteristic chemical method of advance, the formulation of chemical principles by induction from chemical facts.
At the present time the theory of the electronic structure of molecules is in large part still qualitative. For example, we may discuss the structure of a molecule or complex ion such as the sulfate ion in terms first of the Lewis Structure, as given in 1916 -- the structure in which the sulfur atom shares one electron pair with each of the four oxygen atoms that surround it tetrahedrally. then, arguing largely from the observed inter-atomic distances and from the principle of approximate electrical neutrality of all atoms, we may assign to each of the sulfur-oxygen bonds some double-bond character (or, rather, pi-bond character, since each of the two different pi electron pairs of an oxygen atom may enter into pi-bond formation with the sulfur atom), with use of orbitals of the sulfur atom exceeding the octet (that is, other than the sp3 orbitals).
I suggest it may be possible to formulate a quantitative theory of electronic structure -- only an approximate theory, to be sure -- by discussing the energy of a molecule or crystal in an empirical, approximate way. Quadratic functions are so convenient mathematically that it will without doubt be desirable to make a first attempt at an approximate quantitative theory of molecular structure by constructing an energy function in which quadratic terms represent the important structural features, and quadratic cross terms represent their interactions with one another. For example, in the sulfate ion we may begin with a Lewis structure, in which the four single bonds are assigned a certain amount of partial ionic character. If we represent the partial ionic character as a parameter, x, say, a quadratic term of the form 1/2k(x-x0)2 would be written in the energy function for the molecule, in which x0 is a normal amount of partial ionic character, as given by the difference in electronegativity of sulfur and oxygen. the coefficient k would then express the tendency of the bond to approximate the normal amount of partial ionic character -- the value of k, like that of x0, would be formulated by consideration of empirical information. Then, in addition, terms of the form 1/2k'nΠ2 would be introduced to represent the resistance of the sulfur atom to forming pi bonds with the use of its unstable orbitals, beyond the sp3 subshell. A term 1/2k"Σ2 would then be introduced for each atom, in which Σ is the total electrical charge on the atom. This term would represent the tendency for the atom to have the minimum charge -- that is, k" is the constant for that atom, of the electroneutrality principle. There is empirical evidence that the individual charges conferred on the atom by its various bonds need to be considered separately, and that terms of the sort 1/2k'''xixj need to be introduced, in which xi and xj are the charges placed on the sulfur atom by two oxygen atoms i and j. The introduction of terms of this sort corresponds to a picture of the sulfur atom such that the electrons involved in the bonds to different oxygen atoms are localized in space, rather than being distributed with spherical symmetry about the nucleus of the sulfur atom. This picture is, of course, a reasonable one.
The actual electronic structure of the complex ion, in terms of this chemical method of description, would then be found by minimizing the energy with respect to the various parameters contained in the energy expression. Properties of sulfates could then be predicted, as determined by the distribution of energy, the amount of pi-bond character, and so on. From the moments of inertia and vibrational frequencies thermodynamic properties could be calculated; an in general a system of this sort might be applied so widely as to permit it to be described as constituting an essentially complete approximate theory of the structure and properties of substances.
Perhaps I am pessimistic in feeling that quantum mechanics as known at the present time constitutes theory of molecular structure, and that we cannot hope for the discovery of a greatly simplified theory of structural chemistry. I wish that I could believe that I am pessimistic, and that a find, accurate, simple theory that would not require giant calculating machines for its application to the molecules that the chemist is interested in, such as penicillin or hemoglobin, could be discovered. Whatever the future of structural chemistry may turn out to be, I know that the discovery of the shared electron pair bond by Gilbert Newton Lewis in 1916 will remain forever in the minds of chemists as marking the beginning of the period of modern development of theoretical structural chemistry.
November
27, 1951
