I. I. Kotyukov

On the Influence

of Cyclization

on

Molecular

Rotation

Tomsk.
1929.

I. I. Kotyukov.

On the Influence of Cyclization on Molecular Rotation.

(The theory of the meso-structure of organic compounds IV).

The problem of the dependence of the magnitude of the angle of rotation of the plane of polarization on structure is one of the most enigmatic and complex areas of physical chemistry. Despite the enormous number of works on this question, despite the existence of entire schools occupied solely with it, we still do not have here a single firm regularity. The constant of molecular rotation is the most capricious and whimsical characteristic of a system. The slightest change in structure, temperature, concentration, very sharply, sometimes by a jump, changes the magnitude of the angle of rotation. The constant is influenced in the most unexpected way by the solvent, the slightest impurities, the wavelength of light, and even the time elapsed since the preparation of the system. As for the principal factor—structure—then with the simplest lengthening of a chain (e.g. ethyl—propyl) the constant often changes by a very considerable angle. At present we are completely unable to predict how the constant will change depending on the position of a substituent in the nucleus, depending on the introduction of a double or triple bond; we cannot sum the influences of several asymmetric atoms in a molecule, the influences of several radicals, etc.

The sharpest and crudest change in structure is an increase in the cyclicity of the molecule. It has long been known that cyclization very sharply affects the rotation constant, often changing it severalfold in the vast majority of cases in the direction of increase. However, here too not a few exceptions have been observed, which was especially well studied for camphoric acid and its derivatives.

The theory of the meso-structure of organic compounds*) proposes structures that differ very sharply from the classical ones precisely in their cyclicity. Since in the meso-formula there are bonds obtained as a result of the saturation of the mesovalence of hydrogen by the residue of an atom possessing affinity for it and located in the same molecule, this naturally leads to the formation of a new “mesocycle” in the molecule. All meso-formulas are much more cyclic than the corresponding classical ones. It is therefore natural that, for the physicochemical proof of the new structures, we pay all our attention to properties that are in acute dependence on cyclization. The constant of molecular rotation proves in this respect to be very rewarding material.

The question of the influence of cyclization on rotation as such has actually been raised comparatively rarely, and therefore the material here is extremely scattered. We were able to collect it completely only thanks to our “Index of Optically Active Substances.” In selecting the material we paid attention only to the simplest possible cases of cyclization—with ex-

) Izvestiya S. T. I. T.* 48 (1927); Chem. Centr. 1928.
Chem. Abst. 1928.

falling, where possible, on a simple substance, in most cases \(H_2O\), \(NH_3\), and rarely \(C_6H_5NH_2\). Unfortunately, the two substances were not always tested for rotation in one and the same solvent, partly because of insolubility, partly because cyclization was not the aim of the investigation. In some cases the two substances were taken from different authors, whose works are separated by a considerable interval of time. In all, about 135 comparisons were obtained, and it turned out that both the magnitude and the direction of the rotation upon cyclization change in the most varied manner. For convenience in comparing the magnitude of the effect, we calculated, for all cases, the relative change, which we called the coefficient of cyclization. It is calculated by the formula

\[ K = \frac{M - m}{m}, \]

where \(M\) is the larger constant and \(m\) the smaller one, so that, when the rotation increases, the calculation was made with respect to the non-cyclic substance, and when it decreases, with respect to the cyclic one. The coefficient is considered positive if the rotation increases, and negative if it decreases.

Tables A.I; A.II; B.I; B.II.

As is seen from the tables, cyclization exerts a most unexpected influence, and almost always a very strong one. Decrease and increase are observed in almost equal numbers of cases—59 and 75.

However, the picture changes sharply if we construct mesoformulas for all cases and take into account the degree of both proto- and mesocyclization. With this method of comparison it turns out that cyclization almost always increases the constant. Let us consider the material by individual groups of substances.

Mono- and dibasic polyoxy acids (A.I Nos. 3, 20, 22, 29, 32, 35, 36, 37, 41; A.II Nos. 6, 26, 27; B.I Nos. 4, 15, 42) almost all increase the rotation upon cyclization. No. 15.B.I was taken from different authors (1924 and 1911). In No. 42 the salt is taken into account for the acid. Thus the only exception is B.I.4—rhodeohexonic acid. If the carboxyl in these acids has oxygen with a residue, then it has no hydrogen, for, owing to the lightness of the OH group, all hydrogens sit firmly. Thus the mesoformula of these acids does not differ from the protoformula, and here cyclization always increases the rotation.

The next large group consists of derivatives of camphoric acid.

(A.I.4, 21; A.II.1, 2, 11, 14; B.I.2, 6, 9, 12, 13, 16, 18, 19, 20, 21, 26, 27, 29, 30, 31, 32, 33, 34, 36, 37, 38, 39, 41, 43, 45, 48; B.II.1, 3, 6, 9, 10). In the enormous majority of cases the rotation here falls. However, the mesoformula of camphoric acid is more cyclic than that of its anhydride, so that, according to mesotheory, here there is a decrease in cyclicity:

[structural formulas]

Camphoric acid (three cycles).

\[ \begin{array}{c} \text{CH}_3 \qquad\qquad \text{CO}\\ \diagdown \qquad\quad \diagup \qquad\quad \diagdown\\ \text{C} \qquad\qquad\qquad \text{O}\\ \diagup \qquad\qquad\qquad\diagup\\ \text{CH}_2 \qquad\qquad \text{C}\!\cdot\!(\text{CH}_3)_2\\ \diagdown \qquad\quad \diagup\\ \text{CH}_2\!-\!\text{CH}\!-\!\text{CO} \end{array} \]

Its anhydride (two rings).

As is clear from the formulae, camphoric acid has three rings, whereas the anhydride has only two. The meso-formula is strongly supported by the circumstance that dehydrocamphoric acid

\[ \begin{array}{cc} \begin{array}{c} \text{CH}{=}\text{C}\!\cdot\!\text{COOH}\\ \phantom{\text{CH}{=}}\vert\\ \phantom{\text{CH}{=}}\text{C}\!\cdot\!(\text{CH}_3)_2\\ \phantom{\text{CH}{=}}\vert\\ \text{CH}_2-\text{C}\!\cdot\!\text{COOH} \end{array} & \begin{array}{c} \text{H}\qquad\qquad\qquad\text{H}\\ \diagup\ \diagdown \qquad\qquad \diagup\ \diagdown\\ \text{O}\qquad \text{CH}\!-\!\text{C}\qquad\qquad \text{O}\\ \Vert\qquad \diagdown\qquad \vert\qquad\quad \diagup\!\!\Vert\\ \text{C}\!-\!\text{C}\qquad\text{C}\!-\!\text{C}\\ \vert\qquad \vert\qquad \diagdown\qquad \vert\\ \text{OH}\quad \text{CH}_3\qquad \text{C}\qquad \text{OH}\\ \qquad\qquad\qquad (\text{CH}_3)_2 \end{array} \end{array} \]

does not give an anhydride, which, in comparison with the easy anhydrization of camphoric acid, cannot but appear very strange. The impossibility of obtaining an anhydride here is explained by the fact that, according to the meso-theory, the H at the double bond possesses very strong meso-valency. It binds the carbonyl oxygen so firmly that the possibility of removing water disappears entirely because of the steric remoteness of the hydroxyl groups. According to this scheme the meso-formulae of the other derivatives will also be constructed, so that the substances of category A will be exceptions to the general rule of an increase in the angle with increasing cyclicity. As for the substances A.I.4; A.II.2 and 14, i.e. O-chloro- and bromophenyl-substituted compounds, they are, in general, unusually capricious in their rotation: they change both magnitude and sign depending on the solvent, so that they cannot be regarded as typical. This is an accidental exception which, upon special closer investigation, may also fit into the system.

A very interesting deviation is found in α-naphthyl-α-campheramic acid (A.I.21), especially if it is compared with the corresponding β-derivative (B.I.39), which on cyclization behaves normally from the point of view of the meso-theory.

We are trying to discover the cause of these exceptions experimentally.

Very remarkable is the identical rotation of α-hydrindone β-benzoyl-O-carbonic acid (A.II.3) and of the corresponding spiran. From the point of view of the meso-theory the cyclicity here is the same in both:

\[ \begin{array}{c} \text{CH}_2 \qquad\qquad \text{CH}_2\\ \diagup \qquad \diagdown \qquad \diagup \qquad \diagdown\\ \text{C}_6\text{H}_4 \qquad \text{C} \qquad \text{C}_6\text{H}_4\\ \diagdown \qquad \diagup\ \diagdown \qquad \diagup\\ \text{C}\!-\!\text{O}\cdots\text{H} \qquad\quad \\ \vert \qquad\qquad\qquad \diagdown\\ \text{OH}\qquad\qquad\quad \text{CO} \end{array} \]

Acid

\[ \begin{array}{ccccccccc} && \mathrm{CH_2} &&& \mathrm{CH_2} \\ & / && \backslash && / && \backslash \\ \mathrm{C_6H_4} &&& \mathrm{C} &&&& \mathrm{C_6H_4} \\ & \backslash && / && \backslash && / \\ & \mathrm{C} {-} \mathrm{O} &&&& \mathrm{CO} \\ & | \\ & \mathrm{O} \end{array} \]

Spiran.

Derivatives of tartaric acid are divided into two categories: those substituted in the hydroxyl increase rotation (A.I.1, 23, 30, 42; A.II.8, 10, 20), while those substituted in the carboxyl lower it (B.I.1, 3, 5, 7, 23—25). The latter include exclusively aromatic amides, and cyclization here proceeds with the loss either of aniline or of its substituent. Since the diamide of tartaric acid possesses, in comparison with the acid itself, a very high activity, its meso-formula will be

\[ \begin{array}{c} \mathrm{C} \\ \mathrm{O} \\ \diagdown \quad \diagup \\ \mathrm{OH.CH} \qquad \mathrm{NH_2} \\ \mathrm{OH.CH} \qquad \mathrm{NH_2} \\ \diagup \quad \diagdown \\ \mathrm{C} \\ \mathrm{O} \end{array} \]

Then there will be no difference in cyclicity between its substituted derivative and the corresponding imide, and the fall in activity is explained only by the loss of the heavy aromatic radical.

In some cases a high value of rotation is undoubtedly motivated by association, for example (B.I.8) in propylene glycol and in the substance (B.II.7). The remaining examples can be brought under one or another of the schemes considered above, and especially under camphoric acid. Individual cases sometimes do not yield to explanation and plainly contradict the meso-theory. Thus, for example, mono- and diphenylantaric acids, upon cyclization to anhydrides, decrease their activity, whereas cyclicity increases. It is possible that we do not perceive what is at issue here.

On the basis of the foregoing, the difference in the angle of rotation in optically active stereoisomers becomes comprehensible—a fact entirely inexplicable for the classical theory.

Furthermore, it is clear that increasing the unsaturation of a molecule leads to an increase in rotation only by virtue of the appearance of mesovalency in the H atom adjacent to the double bond and of the subsequent cyclization. This is why a triple bond, near which there is no hydrogen, does not increase rotation.

Thus, an analysis of all known cases of cyclization from the standpoint of the theory of meso-structure leads us to the inevitable conclusion that rotation is increased in the enormous majority of cases. Exceptions to this rule are isolated, and it is possible that, upon special investigation, they may be incorporated into the system. Every time that we

we have an unexpected jump in the magnitude of rotation, it follows that mesocyclization should be sought, which in the present respect is entirely equivalent to cyclization through the principal valences. It is possible that the magnitude of this influence depends on the strength of the bonds in the cycle and, for meso-bonds, will be weaker than for the principal-valence bonds.

If cyclization, irrespective of whether it occurred through mesovalence or through a principal valence, invariably increases rotation, then in such a case we are entitled to expect, among substances containing an asymmetric carbon atom not situated in a cycle, a certain minimum of rotation. In other words, if one constructs meso-formulas for all known optically active substances and arranges them according to the number of cycles, whether mesocycles or protocycles, they should also be arranged approximately according to the magnitude of molecular rotation. Non-cyclic substances will have the smallest rotation, and among strongly rotating ones there will be none of them. This proposition has in fact been borne out. Here, however, we must make a number of reservations. At present about 12,000 optically active substances are known. The “construction” of meso-formulas is based not on any firm experimental foundations, but on logical considerations of a theoretical character—considerations whose starting points will probably change when the appropriate experimental material becomes available. Therefore we considered it rational to make, so to speak, a “rough sketch” of these formulas for the time being, without giving them any final form. How precisely this is done systematically we indicate at the end of the article. Moreover, the figures available in the literature are far from always reliable, owing to the slight certainty as to the optical purity of the system.

Nevertheless, a fairly definite result is obtained. For simplicity of accounting we selected only substances containing one or two identical asymmetric carbon atoms. It is convenient to call the latter by the apharetic term ascarat. Thus we are dealing with mono- and diascarate substances. Of the former, 2400 have been prepared, and of the latter, 500. From both groups we selected only non-cyclic and monocyclic ones.

The more complex substances were not considered. We divided the mono-ascarates into three categories:

1. ascarat not a member of a cycle (“amesic” substances), 510 items.
2. ascarat a member of one mesocycle (monomesocyclic), 176 items, and
3. ascarat a member of one protocycle (monoprotocyclic), 80 items.

The diascarates are divided into the same categories—59, 18, and 13 respectively.

According to the magnitude of molecular rotation, the named categories are distributed as follows:

Table.

The data of the table have been reduced to the appended graph, with the angles plotted on the abscissa and, on the ordinate, the number of substances rotating by the corresponding angle.

As is evident from the table and the graph, all amesic substances possess a very small angle of rotation—namely \(0\)—\(35^\circ\), with the overwhelming majority of them (\(65\%\)) at \(5\)—\(25^\circ\). The relations are especially sharp in the case of dicarboxylic substances. As for the meso- and protocyclic substances, they are distributed rather uniformly over the graph from \(0\)—\(130^\circ\).

We hope that the calculation proposed here is the first real approach to the construction of a theory embracing the entire mass of prepared optically active compounds. There are very few sharp exceptions here; in fact they are isolated, and special investigation will explain in detail the cause of their deviation.

With this article we conclude the exposition of the dogmatic part of the theory of meso-structure. We regard it as an enormous, all-embracing working hypothesis. The experimental future will show how far all that has been set forth above is valid, and how far it may be extended to fields not touched upon here.

Especially interesting in this respect appear to be the cis-trans isomerism of naphthenes and the magnitude of the constant of electrolytic dissociation.

As the principal conclusion from all the preceding material we put forward the following proposition: the force of the mesovalency of hydrogen is a function of the intensity of exchange by means of neighboring carbon atoms. This dependence appears because carbon gives to carbon one bond more than a quarter of its entire affinity. From this point of view, all radicals of organic chemistry are divided into two groups: mesophoric, possessing a residual means, and amesic—residues not possessing it.

Residues occur at H, O, N, Cl. Residues at H are borne by groupings with tertiary H, with H directed toward the nucleus, with secondary H adjacent to a negative group, an oxime group, and possibly an aldehyde group. Residues at O are borne by the carbonyl of carboxyl, the isonitroso group, the nitro group, hydroxyl adjacent to a negative group, and carbonyl adjacent to such a group. Nitrogen and the haloids are evidently always mesophoric. It is possible that in the carboxyl group the mesophore is hydroxyl, and not carbonyl.

Amesic groups have no residues. These are negatively directed H; the same H at secondary and tertiary C with a positive nearest group (\(\mathrm{OH}\), \(\mathrm{NH_2}\), etc.); hydroxyl and carbonyl at C with a positive group; and, finally, fatty saturated radicals.

The fundamental assumption is the occurrence of a bond between the residue—chiefly H—and the atom having affinity for it. If this takes place in one molecule, it causes the formation of a new ring, in the overwhelming majority of cases five- or six-membered. If this takes place between different molecules, it leads to the formation of a molecular compound.

A formula with the designation of such bonds, called a “mesoformula,” is constructed as follows. First it is necessary to establish, in final form, the protoformula, i.e. the distribution of the principal valencies. Then the distribution formula is established, i.e. the protovalencies are differentiated. They are subdivided into three approximate groups: strong, ordinary, and weak. Their strength is established on the basis of the usual theories of exchange by means: the theory of negative groups, the theory of unsaturation, the theory of directedness, etc.

After the establishment of the carbon bonds, the affinity between C and H is distributed, making use of the “basic proposition” cited above. Finally, meso-bonds are closed between the residues obtained; among several possibilities, those most probable according to the affinity of the residues are chosen. Here the number of possible “mesoisomers” is established. The mesoformula obtained in this way, i.e. by pure reasoning, is confirmed experimentally, if, of course, the corresponding material is not available in the literature. The experimental proof is carried out as follows.

If we need to prove the presence of a meso-bond between any two radicals \(R_1\) and \(R_2\) in a molecule, then we must prepare an optically active substance of the formula

\[ \begin{array}{ccccc} & R_1 & & R_2 & \\ & \diagdown & & \diagup & \\ & & C & & \\ & \diagup & & \diagdown & \\ CH_3 & & & & C_2H_5 \end{array} \]

If the angle of rotation is less than \(25^\circ\), then the substance is “amesic,” and there is no meso-bond. The presence of an H residue in some radical is proved by the occurrence of “mesoisomerism” and by the number of mesoisomers. It is necessary to prepare the oxime

\[ \begin{array}{ccc} C_6H_5.C & - & R \\ \| & & \\ \mathrm{HO}\;N & & \end{array} \]

If there are two isomers here, then \(R\) will contain an H-residue. The same applies to the residue at oxygen. For final confirmation, substances of the same structures are now prepared, but the residue suspected to be H is replaced by “amesic” \(CH_3\). Then the angle of rotation must fall, and the mesoisomerism must also disappear. From a comparison of all the material, the final mesoformula is then constructed.

In many cases, optically active substances and the corresponding oximes can be found in the literature, and the mesoformula can be constructed, so to speak, by reasoning alone. However, this cannot of course be carried out systematically; and therefore there is here an enormous field for experimentation.

The proposed theory creates an especially keen interest in the optical activity and isomerism of oximes.

The present work arose and was, in principle, carried out within the walls of the laboratory of colloid chemistry at the Voronezh Agricultural Institute, under the direction of A. V. Dumanskii. It could come to light only thanks to the utterly exceptional cultural environment created in the laboratory by the exalted personality of Anton Vladimirovich. An extraordinary scientific “tolerance,” extreme delicacy and caution in dealing with another “scientific soul,” the immediate granting of the fullest “freedom of research” in the presence of “one’s own” ideas, all-round support, assistance, and unfailing sympathy—this is what accompanied me throughout my stay in the laboratory and made it possible to develop the mesotheory. I offer to the deeply respected and dear Anton Vladimirovich my greatest gratitude for his rare participation and attention shown to my work.

The theory of mesostructure in manuscript was reviewed by Academician N. D. Zelinsky. There is no doubt that, had it not been for Nikolai Dmitrievich’s authoritative support, the work would not have seen the light of day. Nikolai Dmitrievich pointed out to me a large number of errors in the work, many of which, incidentally, owing to my lesser competence and in part to haste, I still was unable to correct. Yet the positive things that I heard from so great an authority gave me the necessary confidence that there is a grain of truth in my work. I take this opportunity to express here to the deeply respected and most highly esteemed Nikolai Dmitrievich my profound gratitude.

Tomsk, October 15, 1928.

A. Coefficient of elongation. I. The sign does not change.

A. Coefficient positive. II. Sign changes

V. The coefficient is negative. The sign does not change.

V. The coefficient is negative. II. The sign changes.

For the article by Prof. Ts. Ts. Kotyukov.

Legend

—○— monosaccharide [[unclear: handwritten descriptor]]
--○-- “ mesocyclic
--●-- “ protio-cyclic
--×-- disaccharide [[unclear: handwritten descriptor]]

Axes

Vertical axis: \(A\)

Horizontal axis: molecular rotation, \(0^\circ, 5^\circ, 10^\circ, 15^\circ, 20^\circ, 25^\circ, 50^\circ, 75^\circ, 100^\circ, 125^\circ, 150^\circ\).