ELECTROMERISM, A CASE OF CHEMICAL ISOMERISM RESULTING FROM A DIFFERENCE IN DISTRIBUTION OF VALENCE ELECTRONS 1 RECENT advances in our knowledge of the structure of matter have made it possible for an organic chemist to address a group of non-organic chemists and of physicists upon this subject without apologizing. During a period which is not far behind us in the past, not only the validity, but, possibly, even the utility of employing structure conceptions requiring atoms and their arrangements was brought into question; so that the organic chemist, who has maintained an abiding faith in atoms and a confidence in his ability to decipher something of their arrangements in molecules, became aware of an indulgent smile whenever he broached this subject except in the company of his own confrères.

With this inheritance, it is natural to expect that the organic chemist would welcome any discoveries which make our conception of atoms and of the mechanism by which atoms combine to form molecules more concrete; and that he would be among the first to seek to apply these concepts to special problems in his own field.

With a feeling of keen satisfaction, therefore, we learn through the work of Bragg that, in a diamond crystal, each carbon atom is surrounded by four other carbon atoms placed equidistant from it. These atoms are grouped around the central carbon atom as the four corners of a 1 An address prepared for the symposium on the "Structure of Matter,'' held at the meeting of the American Association for the Advancement of Science in New York City, December, 1916.

regular tetrahedron are arranged around its center. Thus, the tetravalent character of the carbon atom manifests itself clearly. Furthermore, when a model of a diamond crystal is examined, it is discovered that the atoms appear to arrange themselves in rings of six. These relationships suggest at once well known fundamental theories of the organic chemist.

Through the writings of J. J. Thomson,2 Stark, Abegg and many others, the conviction has been reached, that the forces which hold the atoms together, commonly called chemical affinity, are chiefly, if not wholly, electrical in character. The impetus to this interpretation has come from the discovery that electricity itself possesses atomic structure, and that our material atoms appear to be composed of units of positive and negative electrical atoms nicely balanced in the neutral atom. As Carl Barus says:5

an

Not only has energy possessed herself of inertia, but with ever stronger insistence she is usurping the atomic structure once believed to be among the very insignia of matter. Contemporaneously matter, itself, the massive, the indestructible, endowed by Lavoisier with a sort of physical immortality, recedes ever more into the background among the shades of velocity and acceleration.

Electrochemical theories have not been lacking in the development of chemistry. For many years the electrochemical theory of Berzelius was a guide in the interpretation of chemical phenomena. There is, perhaps, no greater tragedy recorded in the annals of science during the past one hundred years than that which overtook Berzelius at the close of his active career as

2 J. J. Thomson, Phil. Mag., March, 1904, 27, 757 (1914), etc.

3 J. Stark, "Die Elektrizitut im Chemischen Atom," Leipzig, 1915.

4 Abegg, Z. Anorg. Chem., 39, 330 (1904); 50, 309 (1906).

SCIENCE, N. S., Vol. XL., 727, 1914.

leader of chemical thought. We of to-day know best why this theory failed, and why we are now busily engaged in formulating a new electrochemical theory, as well as a new electrophysical theory. In fact, J. Stark in his recent work, "Die Elektrizität im chemischen Atom," gives a eulogy of Berzelius, and points out the many striking qualitative resemblances which the theory of Berzelius bears to his own.

The special purpose of my remarks today precludes any detailed discussion of the various theories concerning the structure of the atom. This phase of the subject has already been presented in the morning meeting of this symposium. It may be said that all theories agree upon a positive core or nucleus associated with negative electrons, the atoms of negative electricity. Thomson presents hypotheses concerning the possible arrangements within the atom, while Stark limits his treatment chiefly to the surface layer. The surface layer, he says, contains an excess of positive electricity. In the neutral atom one or more electrons, called valence electrons, are held close to the surface of the atom by this positive charge. Compounds are formed, when the lines of force from one or more of these valence electrons reach out and end on the positive areas of other atoms. In the case of strongly polar compounds, an electron is almost wholly drawn over to the atom which it then holds combined.

Lewis classifies compounds into polar and non-polar, but in a footnote remarks: It must not be assumed that any one compound corresponds wholly, and at all times, to any one type.

He distinguishes between valence number and polar number. Valence number he defines as the number of positions, or regions, or points (bond termini) on the

6 G. N. Lewis, J. Am. Chem. Soc., 38, 762 (1916).

atom at which attachment to corresponding points on other atoms may occur. Polar number is the number of negative electrons which an atom has lost (in an algebraic sense).

The evidence of, perhaps, indeed, the cause of the mobility of polar compounds is the freedom of one especially important atom, the atom of electricity, or the electron, to move from one position to another.

From a study of the reactions of chemical compounds, and in particular of organic compounds, it seems doubtful whether the classification 'into polar and non-polar based upon physical values, such as the dielectric constants of compounds in the gaseous state, is of any more significance than the terms electrolyte and nonelectrolyte were to the older supporters of the theory of Arrhenius. In time, it came to be known that it was no easy matter to draw the dividing line between these two classes, and that one class seemed to merge imperceptibly into the other. So, with polar and non-polar compounds, it seems theoretically probable that there is no perfectly non-polar compound, unless it be a molecule composed of two like univalent atoms, such as hydrogen; and that other

7 Stark ("Die Elektrizität im Chemischen Atom," p. 29) says: "Between the properties 'dielectric' and 'conducting' there is a connection. In a dielectric medium, since there are positive and negative 'Quanten' bound to one another, it follows that the medium may become conducting when, through proper application of energy from without, the 'Quanten' pairs become partially dissociated, or ionized; that is, into freely moving positive and negative 'Quanten.' Conversely, the ions of a conducting medium by mutual union to form 'Quanten' pairs may make the medium dielectric; and in general a material medium is at the same time dielectric and conducting, so that by assigning a dielectric constant and a specific conductivity, the medium is characterized for a finite electric field and a finite electrical current.' ""

8 Bohr concludes that the hydrogen molecule consists of two hydrogen nuclei (at a distance apart of 0.60 × 10-8 cm.), and two electrons which

compounds are polar in varying degrees, depending upon the mutual attractions between valence electrons and the positive surfaces or cores of the atoms combined, and upon the distances to which these electrons, in forming such compounds, are deflected from their normal positions relative to the positive areas of the uncombined atoms themselves.

Even before the electron theory had been proposed, an application of the theory of ion formation and charges upon ions led to the recognition of polar characteristics in compounds not known to be ionogens.

In a study of chloroamines, RNHCl and RNCI, Seliwanow observed that, during hydrolyses, the chlorine in these compounds was replaced by hydrogen; and that they interacted with hydrogen iodide with the liberation of two equivalents of iodine for each equivalent of combined chlorine,

R2NCI +2 HI=R2NH + HCl + I2.

Usually, during hydrolysis, combined chlorine in organic compounds is exchanged for hydroxyl and has no tendency to liberate iodine from hydriodic acid. Seliwanow ascribed this peculiar behavior of the chlorine atom in chloroamines to the fact that, even in combination, it existed as "hypochlorous chlorine." He pointed out that the chlorine atoms in nitrogen trichloride, NC, also showed the same peculiar behavior.

In 1901, Noyes and Lyon,10 in performing Hofmann's well-known lecture experiment for demonstrating the composition of ammonia, observed that, under certain favorable conditions, the amount of nitrogen liberated as free nitrogen was about one sixth, instead of one third, the volume revolve in an orbit in a plane perpendicular to the line joining the nuclei.

9 Seliwanow, Ber., 25, 3612 (1892).

10 W. A. Noyes, A. C. Lyon, J. Am. Chem. Soc., 23, 460 (1901).

and, to account for the reaction between ammonia and chlorine, assumed that ammonia may ionize in two ways,

+
N++ 3H2NH,2N = + 3H+
+

and, furthermore, that the chlorine molecule may ionize to give both positive and negative chlorine ions.

In the same number of the Journal of the American Chemical Society, Stieglitz11 commented upon the work of Noyes and Lyon, and put forth arguments to show that this reaction,

H2O+Cl2 HCl + HOCI,

a reversible reaction, was, at the same time, an ionic reaction. In other words, hypochlorous acid may ionize in two ways, amphoterically,

HOCH + OCIHOCHO+ Cl+. The chlorine molecule, therefore, must yield negative chlorine ions, Cl-, and, also, positive chlorine ions, Cl+.

These deductions, expressed originally by Noyes and Lyon, as well as by Stieglitz, In terms of ion formation, have since been translated into the language of the electron theory of valence. Thus, the chlorine mole

11 Stieglitz, J. Am. Chem. Soc., 23, 797 (1901).

the symbol, Cl − + Cl.

The striking difference in behavior of derivatives of positive chlorine and of negative chlorine may be illustrated by comparing the two compounds, nitrogen trichlovirtue of the family relationship of nitroride and phosphorus trichloride, which, by gen and phosphorus in the periodic system, and the similarity in the formulas of the two chlorides, would be expected to resemble one another in chemical behavior about as closely as any two compounds could. At the same time, the illustration will serve to explain the significance of the statement made in an earlier part of this paper, viz., that the polar characteristics of compounds may be revealed by a study of their chemical interactions.

If the electronic formulas,

are assigned to these two substances, we obtain formulas which, unlike those in general use, show why it is that the two compounds are most dissimilar in chemical deportment; why nitrogen trichloride, when hydrolyzed, gives ammonia and hypochlorous acid, while phosphorus trichloride yields phosphorus acid and hydrogen chloride; why the chlorine atom in nitrogen trichloride chlorine in phosphorus trichloride does possesses oxidizing properties, while the not. The oxidizing value of a positive chlorine atom corresponds to a gain of two negative electrons, if a negative chlorine ion is the final stage in the change.

Br,C+H + HO +Br,

I¿C−+ I + ɤÃ=

I¿С − + H + HO−+I. Iodine monochloride reacts as follows: I+C+H+- OH = HO~+I+H+ − CI. In fact, there is no difficulty in finding among organic compounds countless cases in which the polarity manifests itself clearly during chemical changes. Thus, in the case of alkyl cyanides, RCN, it may be asked what indication there is in the formula itself to lead chemists to predict, unerringly, that the products of hydrolysis of

such a compound are always ammonia and a carboxylic acid. Pure speculation would suggest that at least four different sets of products are possible:

(a) RC(OH), and NH,;
(b) RCH(OH), and NH2OH;
(c) RCH2OH and NH(OH)2;
(d) RCH, and N(OH)3.

But the substances expressed under (a) are the only ones ever realized. That this decision is not inherent in the formula is emphasized all too forcibly by the fact that these four sets of products are the very ones which beginning students offer to explain the hydrolysis of an alkyl cyanide. In terms of the electron conception valence, the explanation lies in the fact that the nitriles are polar compounds of the formula:

products formed are metallic mercury, an alcohol, and a hydrocarbon. At about 200°, mercury diethyl decomposes to give mercury and butane. This dissociation implies that the mercury atom in these dialkyls either possesses, or readily assumes, the condition of reduction which it has in the metallic state, viz., with an equal number of positive and negative "charges.' This suggests, also, that the two ethyl groups may be one negative and the other positive:

=

HgCH Hg + CH ̧ + − С2H ̧.

When mercury diethyl is heated with acetic acid, further evidence in support of this inference is furnished; a quantitative yield of metallic mercury is found, and in addition, ethane and ethyl alcohol (or acetic

ethyl ester). These changes may be ex

pressed in terms of the electron conception of valence as follows:

+ - C2H H+-OH + +-C2H, H+-ОH + он + он

Zn

=

он

If, therefore, the atoms in compounds may function positively or negatively, in general a univalent atom, A, may be represented by two electronic symbols, A+ and