B. V. Tronov and L. P. Kulev.

Determination of the electromotive force of reactions with metals as a method for studying organic compounds.

The task of studying any organic compound found in nature or obtained artificially is divided into the following parts:

1) Isolation of the given compound from the mixture, its purification, and determination of its purity.
2) Qualitative and quantitative analysis of the substance and determination of its molecular weight.
3) Determination of the structure of the molecule, derivation of the structural formula, and, if necessary, determination of the spatial configuration.
4) Synthesis, i.e., preparation from simple substances or from simpler compounds, which finally resolves the question of the structure of the molecule of the given substance.

The fundamental and most often the most difficult question is the determination of structure. This is carried out:

1) By determining the physical constants of the substance: melting point, boiling point, specific gravity, refractive index, surface tension, rotation of the plane of polarization of light, etc. For some of these constants, only qualitative regularities have so far been established, expressing their dependence on structure. For other constants the connection with the structure of the molecule can be expressed by quantitative formulas. Such, for example, are the complex constants: molecular refraction:

\[ R = \frac{n^2 - 1}{n^2 + 2} \cdot \frac{M}{d} \]

(\(n\)—refractive index, \(M\)—mol. wt., \(d\)—specific gravity)

and parachor

\[ P = \frac{M}{d}\gamma^{\frac{1}{4}} \]

(\(\gamma\)—surface tension).

These constants make it possible to distinguish quite definitely, for example, a double bond, a triple bond, give an idea of the degree of branching of the chain, etc.

2) Some physical methods, such as X-rays, electron waves, make it possible almost directly to determine the spatial configuration of molecules.

However, in their practical significance in chemistry, all purely physical methods yield (at least for the time being) to

3) chemical methods of investigation.

Chemical methods of determination may be subdivided into:

a) Reactions of decomposition of a complex molecule into simpler ones. Such are: hydrolysis of compounds having an ester or amide linkage; the action of strong oxidizing agents; strong heating of the substance under study alone or with hydrogen. By studying the decomposition products, we learn from what parts the skeleton of the molecule of the given compound consists.

b) Qualitative and quantitative determination of functional groups, i.e., the most active atomic groups in the molecule: carboxyl, alcoholic hydroxyl, aldehyde and ketone carbonyl, nitro group, amino group, double and triple bonds, etc. Such determinations supplement the material obtained by applying the preceding group of reactions.

c) Determination of the activity of functional groups. The activity of one or another group of atoms depends so much on its position in the molecule and, in general, on the composition and structure of the entire molecule that sometimes this group gives almost none of the reactions characteristic of it. Thus there are amines that do not add acids at all (triphenylamine), alcohols (for example, tertiary amyl alcohol) that do not react or almost do not react with acetic anhydride¹, ketones that do not react with bisulfite, and compounds with double bonds that do not add bromine.

Such dependence on the structure of the molecule makes determination of the chemical activity of functional groups an extremely important means for investigating organic compounds. In the present state of organic chemistry, when methods for cleavage of molecules, as well as for the qualitative and quantitative determination of the most important functional groups, have already been fairly well developed, the principal task for further development of the science in this field should be considered precisely the development of methods for determining activity.

The principal methods for determining the activity of functional groups are:

1) Measurement of the rate of reactions in which the given group participates. The rate of reactions can be determined directly, for example, from the quantity of products formed, from the decrease in concentration of the reacting components, or from the change in some physical property. Sometimes, instead of an absolute determination of the rate of one process, the ratio of the rates of two or several opposing or generally different reactions is determined. Such are: the study of the equilibrium system:

$$ \text{alcohol} + \text{acid} \rightleftarrows \text{ester} + \text{water}. $$

(simultaneously proceeding opposite reactions);

the study of the products of cleavage of a simple ether by hydroiodic acid:

$$ \mathrm{CH_3 - O - C_2H_5 + HJ} \begin{matrix} \nearrow \mathrm{CH_3J + HOC_2H_5}\\ \searrow \mathrm{CH_3OH + JC_2H_5} \end{matrix} $$

For a complete characterization of the course of the process, the results of such determinations must be supplemented by direct measurement of the rate of at least one of the reactions.

2) Determination of the degree of dissociation.

The constant of electrolytic dissociation is most often determined from the electrical conductivity of a solution of the given substance in a suitable solvent (usually in water). Both electrolytic dissociation and the decomposition of a molecule into neutral atoms or groups can be determined from the decrease in molecular weight, from the color of the substance itself or in solution, and by other methods. Sometimes activity is determined by making measurements not on the substance itself, but on some readily obtainable derivative of it. Thus, a good means for assessing the activity of nitrogen in amines is determination of the dissociation constant of the bases formed when water is added to them:

$$ \mathrm{CH_3NH_2 + H_2O \rightleftarrows CH_3NH_3OH \rightleftarrows [CH_3NH_3]' + (OH)'} $$

3) Determination of the heat of formation of a compound from the elements (i.e., strictly, from the simple substances corresponding to these elements). The heat of formation is calculated from the heat of combustion of the substance in oxygen. This method serves as a good means of recognizing certain active groups, for example carbon atoms joined to one another by double or triple bonds; it sometimes gives an idea of the degree of activity of a double or triple bond (for example, benzene has a smaller store of energy than might be expected for three double bonds, and this fully corresponds to its comparatively low chemical activity); but the results obtained indicate only the activity of the entire molecule in general. The heat of formation of a compound from the elements is not directly connected with how active one or another part of the molecule should be, nor, moreover, with respect to one or another reagent. Naturally, we can sometimes make very little use of the heat of formation for elucidating the structure of a substance.

In this respect better is:

4) Determination of the heat evolved in a definite reaction, for example, the heat of addition of bromine to a double bond, the heat of reaction with metals or with organometallic compounds of substances having active hydrogen. The figures obtained characterize the properties of only one functional group, and these properties, of course, must depend in a definite way both on the nature of this group and on its position and on the composition and structure of the remaining parts of the molecule.

5) Determination of the electromotive force arising in a reaction, if the latter is carried out in such a way that the chemical energy is converted into electrical energy.

The electromotive force, as is known, is a measure of the free energy of the process, i.e., of that part of the energy evolved which can be converted into work. The indicated method in principle thus has the same significance as the preceding one, where the total amount of energy liberated is determined. The numbers obtained are in general far from identical, and the application of both methods to one and the same reaction should, of course, yield especially valuable results for elucidating structure.

The determination of the activity of functional groups from the electromotive force arising in reactions has hitherto been applied very little in organic chemistry. The works available in the literature on this question concerned only oxidation–reduction reactions; for example, the quinhydrone electrode, based on such a reaction, has found rather wide application in electrochemistry.

Meanwhile in organic chemistry, as also in mineral chemistry, one can find very many processes which, under a corresponding arrangement of the experiment, can give an electric current and in which an electromotive force must undoubtedly arise.

Wishing to choose a simple and convenient method for determining activity, one encompassing as many diverse functional groups as possible, we settled on the determination of the electromotive force of reactions of organic compounds with metals. Metals, especially the most active of them, such as the alkali metals, can act on the following groups of organic compounds:

1) Compounds with mobile hydrogen.

Among these compounds, in some, as in carboxylic acids, sulfonic acids, and phenols with strongly electronegative substituents, the mobility of hydrogen is so great that it can be determined by electrical conductivity or, in general, by methods that directly determine the degree of electrolytic dissociation. In alcohols, amines, acetylene and cyc-

of cyclopentadiene hydrocarbons, electrolytic dissociation occurs to a degree far too slight for direct measurement. With metals, however, especially with such metals as sodium, these substances react very noticeably. This reaction is analogous to the dissolution of metals in acid, on which all galvanic cells are based; i.e., it should be entirely amenable to electrometric measurement.

2) Halide compounds.

In the reaction of organic halide compounds with metals, an electron is lost by the metal atom, which passes into the state of a positive ion. The halide, which was part of the neutral molecule of the organic halide, takes up this electron and becomes a negative ion:

\[ \mathrm{CH_3I + Na \rightarrow CH_3 + Na^{\bullet} + I'} \]

If this reaction is written as follows:

\[ \mathrm{CH_3I + 2Na \rightarrow CH_3Na + NaI'} \]

\[ \mathrm{CH_3Na + ICH_3 \rightarrow CH_3{-}CH_3 + NaI'} \]

then the electron process here too will be one and the same. The transfer of an electron from one atom to another must, of course, likewise lead to the appearance of an electromotive force.

3) Unsaturated compounds.

It is known that aldehydes and ketones are capable of adding metals to the oxygen of the carbonyl group. In this process, sometimes the carbon atoms of two molecules of the carbonyl compound couple with one another by the liberated units of affinity, with formation of the alcoholate of the corresponding pinacol:

\[ \begin{array}{c} \mathrm{CH_3}\\ |\\ \mathrm{C{=}O + Mg + O{=}C} \\ | \quad \quad \diagup \quad \diagdown\\ \mathrm{CH_3 \quad CH_3 \quad CH_3} \end{array} \;\longrightarrow\; \begin{array}{c} \mathrm{CH_3 \quad CH_3}\\ \diagdown \quad \diagdown\\ \mathrm{C{-}C}\\ / \quad \backslash\\ \mathrm{CH_3 \quad O{-}Mg{-}O \quad CH_3}\\ \quad \quad \backslash \quad |\\ \quad \quad \mathrm{C}\\ \quad \quad |\\ \quad \quad \mathrm{CH_3} \end{array} \]

Sometimes such coupling does not occur, and a metal-ketyl is formed—a compound with trivalent carbon:

\[ \begin{array}{c} \mathrm{C_6H_5}\\ |\\ \mathrm{C{=}O + Na}\\ |\\ \mathrm{C_6H_5} \end{array} \;\longrightarrow\; \begin{array}{c} \mathrm{C_6H_5}\\ |\\ \mathrm{C{-}O{-}Na}\\ \diagdown\\ \mathrm{C_6H_5} \end{array} \]

In both cases there is reason to think that the metal has passed into the state of a positive ion (even if it has not finally separated off), i.e., has lost an electron (or electrons), and hence here too an electromotive force may arise. In the case of double and triple bonds between carbon and nitrogen or between two carbon atoms, such addition products cannot be isolated (apart from some cases of the addition of sodium to the double bonds of unsaturated hydrocarbons of the benzene series), for example:

\[ \mathrm{(C_6H_5)_2C{=}CH.C_6H_5 + 2Na \rightarrow (C_6H_5)_2CNa{-}CHNa\,C_6H_5.} \]

However, many substances that do not directly add metal can nevertheless, under the influence of sodium, react by passing into polymers. Thus, for example, nitriles polymerize:

\[ 3\mathrm{C}_6\mathrm{H}_5-\mathrm{C}\equiv\mathrm{N}(+\mathrm{Na})\ \to\ \text{[[unclear: structural formula of triphenyl triazine]]} \]

Metallic sodium is the best polymerizing agent in the polymerization of butadiene into rubber:

\[ n\mathrm{CH}_2=\mathrm{CH}-\mathrm{CH}=\mathrm{CH}_2(+\mathrm{Na})\quad (-\mathrm{CH}_2-\mathrm{CH}=\mathrm{CH}-\mathrm{CH}_2-)n. \]

It is quite probable that here there first occurs an unstable addition of the Na atom to the more electronegative of the atoms joined by a double or triple bond:

\[ \begin{aligned} \mathrm{C}_6\mathrm{H}_5-\mathrm{C}\equiv\mathrm{N}\ .\ .\ .\ .\ .\mathrm{Na}\\ \mathrm{CH}_2=\mathrm{CH}-\mathrm{CH}=\mathrm{CH}_2\ .\ .\ .\ .\mathrm{Na} \end{aligned} \]

These intermediate products can be constructed in the same way as metallic ketyls, and in their formation one could likewise expect the appearance of an electromotive force.

Arrangement of the experiments.

As the most active of the readily available metals, we chose sodium.

The apparatus itself, in which the reaction was carried out, consisted of a thick-walled glass test tube with two electrodes. One electrode was a cylindrical platinum plate lying against the walls of the test tube. The other electrode had to have a fresh surface of metallic sodium. For this purpose, from a piece of sodium 9 mm thick a cylinder was cut out by means of a short glass tube. The tube was inserted into the test tube with the substance poured into it (5–6 cubic cm); then, already in the liquid, the metal was pressed out by means of a brass rod.

The platinum plate and the brass rod were connected to the poles of a sensitive voltmeter. Since we had to deal with the most varied magnitudes of electromotive force, from ten-thousandths of a volt to whole volts, it proved most convenient to use two instruments: a sensitive millivoltmeter with a scale up to 16 mv, and a voltmeter up to 3 v. Both were connected to a common commutator, from which wires ran to the sodium and the platinum.

Some organic compounds are very active toward sodium, and the reaction with them is accompanied by a large evolution of heat. Since maintaining a constant temperature would have required very complicated arrangements, we diluted such substances with a neutral solvent—benzene. The temperature during the reaction was not measured. For the most part it was close to 15–20°.

Determination of the electromotive force with the aid of a voltmeter is, of course, less accurate than that given by potentiometers and electrometers of various systems. Strictly speaking, the voltmeter readings express not the difference of potentials, but the strength of the current passing through it. These readings give a sufficiently accurate value for the electromotive force only at a comparatively small co-

of the resistance of the system under study, appreciably less than the resistance of the voltmeter itself.

We had no possibility of determining, with any degree of accuracy, the electrical conductivity of the substances investigated, since this electrical conductivity was not constant during the experiment, changing greatly (increasing) as the sodium dissolved. This reduced the accuracy of measurements of the electromotive force. Indeed, different voltmeters (with different resistance) sometimes gave, for one and the same reaction, somewhat (though not greatly) different values of the potential difference. However, this very fact indicates that in any case the resistance of the substance did not greatly exceed the resistance of the voltmeter, i.e., in any case the order of magnitude obtained was correct.

Moreover, the aim of our work is not, strictly speaking, to determine the electromotive force as such, with the greatest possible accuracy, but to solve practical problems of organic chemistry. The figures obtained with our apparatus may not express exact values of the electromotive force arising in a given reaction; but these figures depend very sharply on the structure of the substance investigated, on impurities in it, etc.; that is, the readings of the voltmeter under the given conditions can provide definite indications as to the nature of the compound introduced into the reaction. What we find here are not directly the values of the electromotive force, but perhaps certain derived quantities depending both on the electromotive force and on the electrical conductivity. Nevertheless, the results of these measurements, in the worst case, can still be used in organic chemistry in the same way as, for example, the parachor is used, whose physical meaning is also unclear. Experience shows, however, that the figures we obtain for the electromotive force often point to finer features in the structure of compounds than those that can be detected by means of molecular refraction or the parachor. The method we propose for determining the electromotive force should be regarded chiefly as a practical method, which may find application above all for determining the position of certain functional groups and, in general, for determining the structure of many organic compounds. As we shall see below, this method gives good results in the study of complex compounds which, because of their instability, often cannot be isolated and investigated by ordinary methods. Finally, the determination of various impurities (e.g., water) in organic substances, and in general the analysis of mixtures, can in many cases also be carried out by this method very rapidly and with sufficient accuracy.

The electromotive force during the experiments, as a rule, never remains constant. Most often, at the beginning it rises rapidly (within several seconds and, in any case, during the first minute), and then falls more slowly. More rarely, a slow increase in the magnitude of the EMF is observed (sometimes up to 30 min.—1 hour). In some cases it remains at one level for 20—10 minutes. In all experiments we recorded the course of change in the value indicated by the voltmeter, but we draw our conclusions chiefly from the maximum figures.

So far we have studied the following types of organic compounds:

I. Compounds with active hydrogen.

The following were investigated here:

1) Alcohols:

The constant of electrolytic dissociation of alcohols is very small and, for most of them, does not lend itself to direct determination from electrical conductivity. It is more convenient here to determine the activity of hydrogen from the rate of those

reactions in which this hydrogen is removed. Such reactions include the esterification of alcohols by carboxylic acids, their anhydrides, acid chlorides, and the preparation of simple ethers by the action of benzenesulfonic acid esters. In all the reactions listed it has been sufficiently proved\(^{1}\) that in the alcohol it is primarily the hydroxyl hydrogen that is split off, e.g.:

\[ \mathrm{CH_3{-}C(=O){-}OH + CH_3OH} \rightleftarrows \mathrm{CH_3{-}C(OCH_3)(OH){-}OH} \rightleftarrows \mathrm{CH_3{-}C(=O){-}OCH_3 + H_2O} \]

The results of all investigations in this field reduce to the following general rules:

Methyl alcohol reacts most readily. It is followed by primary alcohols, and side chains, benzene rings, etc., somewhat retard esterification. Secondary alcohols react more difficultly than primary ones, tertiary alcohols and phenols still more difficultly. Very characteristic are the figures for the relative rate of esterification obtained for the reaction with acid chlorides, which is the most reliable in this respect, since it cannot in any way be interpreted without the splitting off of hydrogen from the oxygen atom. The interpretation of this process that best agrees with modern data is\(^{2}\):

\[ \mathrm{CH_3{-}C(=O){-}Cl + CH_3OH} \rightarrow \mathrm{CH_3{-}C(OCH_3)(Cl){-}OH} \rightarrow \mathrm{CH_3{-}C(=O){-}OCH_3 + HCl} \]

Determination of the rate of interaction of alcohols with para-\(\mathrm{NO_2{-}C_6H_4{-}COCl}\) gave the following values for the relative rates\(^{3}\).

Alcohols	Relative rate of reaction	Alcohols	Relative rate of reaction
\(\mathrm{CH_3OH}\)	100	\(\mathrm{C_6H_5CH_2OH}\)	9.3
\(\mathrm{C_2H_5OH}\)	45.7	\(\mathrm{C_6H_5CH_2CH_2OH}\)	21.6
\(\mathrm{n\text{-}C_3H_7OH}\)	35.9	\(\mathrm{C_6H_5(CH_2)_2CH_2OH}\)	10.7
\(\mathrm{n\text{-}C_4H_9OH}\)	40.2	\(\mathrm{CH_3CHOH\ CH_3}\)	5.4
\(\mathrm{(CH_3)_2CH\ CH_2OH}\)	16.7	\(\mathrm{CH_3CHOH.CH_2.CH_3}\)	4.0
		\(\mathrm{(CH_3)_3COH}\)	1.5

We determined\(^{4}\) the electromotive force of the reaction of alcohols with sodium without solvent and in benzene solution. As will be seen from the results of the experiments, clearer regularities are obtained when the experiments are carried out with solutions. Here a large rise in temperature is avoided and, moreover, it is possible to have everywhere the same concentration of hydro-

\(^{1}\) B. V. Tronov, L. N. Dyakonova-Shults, O. I. Gulyaeva, N. M. Nikiforova, ZhRKhO 59, 545–62, 1927. Ch. Prevost, Cr. 185, 1283–5, 1927. Chm. Zbl. 1928. 1. 672 and a number of other works. B. W. Tronow, N. Ch. Ssibgatullin. Ber. D. Ch. Ges. 62 2850 (1929).

\(^{2}\) H. Staudinger K. Clar, E. Czako. Ber. 44, 1640–1647. 1911.

\(^{3}\) I. F. Norris, A. A. Ashdown J. Am. Ch. S. 47. 873.

\(^{4}\) B. V. Tronov and L. P. Kulev. Zh. O. Kh. 4, 197–202. 1934.

acid. Octyl alcohols were taken in pure form; the others were diluted with benzene to an equimolecular concentration with them.

The results are given in the following table:

No.	Alcohols	EMF without solvent (in millivolts)	EMF in benzene
1.	CH₃OH	1850	850
2.	CH₃·CH₂OH	1250	550
3.	CH₃·CH₂·CH₂OH	1050	120
4.	CH₃·CHOH·CH₃	260	20
5.	CH₃·CH₂·CH₂·CH₂OH	260	60
6.	(CH₃)₂·CH·CH₂OH	—	30
7.	CH₃·CHOH·CH₂CH₃	45	2,8
8.	(CH₃)₃COH	4,8	0
9.	CH₃(CH₂)₃·CH₂OH	—	45
10.	(CH₃)₂CH·CH₂·CH₂OH	100	15,2
11.	(CH₃)₂COH—CH₂CH₃	0	0
12.	CH₃·(CH₂)₄·CH₂OH	—	20
13.	(CH₃)₂CH·CH₂·CHOH·CH₃	—	0,20
14.	CH₃·(CH₂)₅·CH₂OH	—	8,25
15.	(CH₃)₂CH·CH₂·CHOH·CH₂—CH₃	—	0,10
16.	CH₃·(CH₂)₆·CH₂OH	—	3,6
17.	CH₃·(CH₂)₅·CHOH·CH₃	—	0,12
18.	CH₂=CH·CH₂OH	—	160
19.	C₆H₅·CH₂OH	—	8,3
20.	C₆H₅·CH₂·CH₂·OH	—	15,2
21.	C₆H₅·CH₂·CH₂·CH₂OH	—	3,85
22.	C₆H₅OH	—	30,0
23.	o—CH₃·C₆H₄OH	—	2,8
24.	m—CH₃·C₆H₄OH	—	8,7
25.	p—CH₃·C₆H₄OH	—	3,9
26.	C₆H₅·CH=CH·CH₂OH	—	0,5
27.	Cyclohexanol	—	0,35
28.	CH₂OH·CH₂OH	770	—
29.	CH₂Cl·CH₂OH	1100	—
30.	CH₃CHOH·COOC₂H₅	120	—

The numbers in this table show that the activity of alcohols with respect to sodium, as determined from the electromotive force, depends on the structure of the alcohols in the same way as the rate of esterification. If we compare our results with those obtained, for example, by determining the rate of reaction with chlorinated p-nitrobenzoyl, we find that, for the hydroxyl hydrogen of secondary alcohols, the activity is several times lower than that of primary alcohols; tertiary alcohols are still several times less active. The comparison of trimethylcarbinol with tertiary amyl alcohol is very interesting. The former, without a solvent, nevertheless gave some electromotive force; for the latter, EMF = 0. As for the rate of esterification of these alcohols, tertiary butyl alcohol, for example, with acetic anhydride, although slowly, does form an ester, whereas dimethylethylcarbinol, even after 720 h, gave nothing. Alcohols with a branched carbon skeleton also proved in our experiments to be somewhat less active than their straight-chain isomers.

The most substantial difference from the preceding data is observed in the dependence of the electromotive force on the length of the alcohol chain. Determination of the rate of esterification of many primary alcohols of the limiting series gives rather close values for them. In our experiments a sharp decrease was observed. Primary hexyl alcohol gave the same EMF as secondary propyl alcohol, and primary octyl almost equaled secondary butyl

A clearly expressed periodicity in these alcohols was not observed by us. There is periodicity only among alcohols of the benzene series with the hydroxyl in the side chain.

In our results, phenols, as in esterification reactions, proved to be considerably inferior in hydrogen activity to the lower (although primary) alcohols of the saturated series. This contradicts the opinion established in science concerning the greater acidity of phenols, which is based on the behavior of phenols in aqueous solution (they dissolve in aqueous solutions of alkalis).

With two hydroxyls (glycol), the EMF of the reaction with sodium is even smaller than that of the corresponding alcohol with one hydroxyl. The activity of hydrogen is also somewhat weakened by chlorine standing in the neighboring link and by an ester group (chloroethanol and ethanol, isopropyl alcohol and the methyl ester of lactic acid).

2) Carboxylic acids.

The activity of hydrogen in organic acids has been studied in great detail. For every more or less investigated acid we find in handbooks the constant of electrolytic dissociation, calculated according to Ostwald’s equation:

\[ \frac{\alpha^{2}}{(1-\alpha)v}=K, \]

where \(\alpha\) is the degree of dissociation of the acid at a concentration of 1 gram-equivalent in \(v\) liters. The degree of dissociation is almost always determined from the electrical conductivity.

Many empirical rules have been derived expressing the dependence of the dissociation constant on the composition and structure of the acid molecule. Monobasic acids of the saturated series are weakly dissociated (except for the sharply distinguished formic acid). Electronegative atoms and radicals increase the degree of dissociation: halogens, OH (apart from some cases still not amenable to explanation, e.g., the fact that \(p\)-oxybenzoic acid is weaker than benzoic acid), CN, NO\(_2\), CO, a second COOH. Moreover, the closer the electronegative substituent stands to the carboxyl (along the chain or in space), the more strongly it increases acidity. The degree of dissociation is also increased by a double bond, by a triple bond (much more strongly), and by a benzene ring. Branching of the chain has very little effect.

We have determined the electromotive force of the reaction with sodium for only a few monocarboxylic acids of the saturated series\(^1\). Our method is not suitable for solid compounds, and in benzene dicarboxylic acids, oxy acids, and even monocarboxylic acids of the benzene series are not sufficiently soluble.

Experiments with pure acids and with benzene solutions of the same concentration as was used for the alcohols gave the following values of electromotive force:

No.	Acids	EMF of pure acid	EMF in benzene solution	For the primary alcohol—of the same structure	For the primary alcohol—of the same structure
1.	CH\(_3\)·COOH	—	—	1250	550
2.	CH\(_3\)·CH\(_2\)·COOH	570	—	1050	120
3.	CH\(_3\)·CH\(_2\)·CH\(_2\)·COOH	150	20	260	60
4.	(CH\(_3\))\(_2\)CH·CH\(_2\)·COOH	70	16	100	15.2

\(^1\) B. V. Tronov and L. P. Kulev. Zh. O. Kh. 4, 783–789. 1934.

Of particular note is the fact that carboxylic acids, both in the pure state and in benzene solution, give with sodium a considerably smaller electromotive force than alcohols with the same structure of the carbon skeleton.

This is in complete contradiction with the generally accepted notion of the acidity of these classes of compounds, but it agrees with the fact that acids in the individual state and in benzene also possess only negligible electrical conductivity. Evidently, the dimers observed in benzene solution (from the lowering of the freezing point) and in the acid itself (even from the vapor density at temperatures somewhat above the boiling point) have almost no acidic properties. In aqueous solutions, according to modern ideas on the nature of electrolytes¹, complexes of the dissolved substance with water are formed, and these already dissociate as acids. As will be seen below, this is fully confirmed by determinations of the electromotive force of the same acids in mixture with certain organic compounds (nitro compounds, amines, etc.).

3) Amines and hydrocarbons.

Primary and secondary amines, as is known, can react with the most energetic metals, giving up their hydrogen. They react still more readily with organometallic, e.g., organomagnesium, compounds. The heterocyclic amine pyrrole can even exchange its hydrogen for potassium on heating with KOH,

\[ \begin{array}{c} \text{CH}\quad \text{CH}\\[-0.2em] \begin{array}{c} \boxed{\phantom{XXXXX}} \end{array} \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!>\text{NH} + \text{KOH} \;\longrightarrow\; \text{CH}\quad \text{CH}\\[-0.2em] \begin{array}{c} \boxed{\phantom{XXXXX}} \end{array} \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!>\text{NK} + \text{H}_2\text{O} \end{array} \]

i.e., it definitely behaves as a weak acid.

The same is observed for certain hydrocarbons. In hydrocarbons of the acetylene series, the hydrogen attached to the carbon with the triple bond is also replaced rather readily by metals. The same is true in the cyclopentadiene series. Cyclopentadiene, indene, fluorene,

\[ \text{cyclopentadiene};\quad \text{indene};\quad \text{fluorene} \]

react both with metallic sodium and with organomagnesium compounds. In fluorene it has been possible to replace H by K also under the action of KOH (in pyridine solution).

Our experiments, however, gave with these classes of compounds only negligible—

measurable values of the electromotive force, and sometimes gave no measurable EMF at all.

Nos.	Compounds investigated	EMF (without solvent)
1.	$(\mathrm{C_3H_7})_2\mathrm{NH}$	$0$ mv.
2.	$\mathrm{C_6H_5NH_2}$	$0.27$ ”
3.	$\mathrm{C_5H_{10}NH}$ (piperidine)	$0$ ”
4.	$\mathrm{C_4H_4NH}$ (pyrrole)	$2.7$ ”
5.	Indene	$0$ ”

The method, therefore (at least in its simple form), must be recognized as of little use for the study of amines and hydrocarbons. Evidently, in these compounds the activity of hydrogen is too small.

4) Enolizable (and similar) compounds.

We have tested a series of compounds which in their usual form should not contain hydrogen directly replaceable by metals, but in which such hydrogen appears in one of the tautomeric forms. These include, above all, carbonyl compounds: aldehydes, ketones, and esters.

\[ \begin{array}{ccc} \mathrm{CH_3{-}C{=}O} & \rightleftarrows & \mathrm{CH_2{=}C{-}OH} \\ \ \ \ \ \ \ \vert & & \ \ \ \vert \\ \mathrm{H} & & \mathrm{H} \end{array} \]

\[ \begin{array}{ccc} \mathrm{CH_3{-}C{-}CH_3} & \rightleftarrows & \mathrm{CH_2{=}C{-}CH_3} \\ \ \ \ \ \ \ \Vert & & \ \ \ \vert \\ \mathrm{O} & & \mathrm{OH} \end{array} \]

\[ \begin{array}{ccccc} \mathrm{C_2H_5O{-}C{-}CH_2{-}C{-}OC_2H_5} & \rightleftarrows & \mathrm{C_2H_5{-}O{-}C{-}CH{=}C{-}OC_2H_5} \\ \ \ \ \ \ \ \Vert & \ \ \ \ \ \Vert & \ \ \ \ \ \Vert \ \ \ \ \ \ \vert \\ \mathrm{O} & \mathrm{O} & \mathrm{O} \ \ \ \ \ \mathrm{OH} \end{array} \]

For aldehydes and esters of formic acid, another tautomerism is also possible:

\[ \begin{array}{ccc} \mathrm{CH_3{-}C{=}O} & \rightleftarrows & \mathrm{CH_3{-}C{-}OH} \quad \text{(with divalent carbon)}\\ \ \ \ \ \ \ \vert & & \\ \mathrm{H} & & \end{array} \]

Among nitrogen-containing compounds, the best known in this respect are nitro compounds:

\[ \begin{array}{ccc} \mathrm{CH_3{-}N{=}O} & \rightleftarrows & \mathrm{CH_2{=}N{-}OH} \\ \ \ \ \ \ \vert & & \ \ \ \ \Vert \\ \mathrm{O} & & \mathrm{O} \end{array} \]

Acid amides also exist in two tautomeric forms, but in them

\[ \begin{array}{ccc} \mathrm{CH_3{-}C{=}O} & \rightleftarrows & \mathrm{CH_3{-}C{-}OH} \\ \ \ \ \ \ \vert & & \ \ \ \ \Vert \\ \mathrm{NH_2} & & \mathrm{NH} \end{array} \]

both forms should have sufficiently active hydrogen.

For nitriles, the possibility is not excluded of transition into the form with a hydrogen atom attached to nitrogen:

\[ \mathrm{CH_3-C\equiv N \rightleftarrows CH_2=C=NH.} \]

Of course, for all these compounds a reaction is also possible according to the type of unsaturated compounds, i.e., direct addition of the metal to the more electronegative of the atoms connected by the double bond, perhaps with subsequent liberation of hydrogen. Apparently, in some cases this is what occurs. However, certain facts indicate that there are substances which react at once in the enolic (or analogous) form. Thus, in the molecule of acetoacetic ester the carbonyl, according to all the data, should be less active than in acetone, whereas the electromotive force is considerably greater; acetonitrile gives, with sodium, an electromotive force too large for an unsaturated compound. In the next chapter methods will be indicated that make it possible to distinguish the types of activity. For the present we shall give only the magnitudes of the EMF of enolizable compounds, obtained either with the pure compounds or in benzene solution of the same concentration as in the preceding experiments.

No.	Compounds investigated	EMF of pure compound	EMF in benzene solution
1.	\(\mathrm{CH_3.CH_2.C(=O)H}\)	110	0,59
2.	\(\mathrm{C_6H_5.C(=O)H}\)	4,1	—
3.	\(\mathrm{CH_3.CO.CH_3}\)	390	23
4.	\(\mathrm{C_6H_5.CO.CH_3}\)	15,6	4,8
5.	Cyclohexanone	18	2,5
6.	\(\mathrm{H.COOC_2H_5}\)	30	1,8
7.	\(\mathrm{COOC_2H_5.CH_2COOC_2H_5}\)	2,75	—
8.	\(\mathrm{COOC_2H_5.CHCl.COOC_2H_5}\)	—	—
9.	\(\mathrm{CH_3.CO.CH_2.COOC_2H_5}\)	470	250
10.	\(\mathrm{CH_3.C\equiv N}\)	20	30
11.	\(\mathrm{C_6H_5.CH_2.CN}\)	1400	12
12.	\(\mathrm{H.CONH_2}\)

The results of these few experiments lead to the following conclusions:

a) Aldehydes pass into the enolic form less readily than ketones (comparison of acetone with its isomeric propionic aldehyde).

b) The isomerization of aldehydes into the tautomeric form with divalent carbon is in any case rather weak.

c) Acetoacetic ester enolizes to a greater degree than malonic ester.

d) Nitriles having hydrogen atoms on the carbon adjacent to the CN group apparently isomerize readily into the form with hydrogen on the nitrogen atom.

Some of the conclusions drawn may prove inaccurate, since here, after all, one has to deal with two possible directions of the process.

II. Complex Compounds with Active Hydrogen

As was already mentioned above, determining the activity of hydrogen in alcohols and acids from the electromotive force of the reaction with sodium leads to an entirely different view of the relative strength of acidic properties than had been accepted on the basis of measurements of the electrical conductivity of aqueous solutions. Since this difference is undoubtedly explained by the formation in water of a complex of the substance under investigation with the solvent, it was of interest to follow how such complexes would behave in our apparatus.

It made no sense to determine the electromotive force of the reaction with sodium in aqueous solutions. We therefore decided to use such organic solvents as could form complexes with hydroxyl compounds, while themselves being as little active as possible with respect to metals. According to the information available in the literature, nitro compounds and amines should have proved to be the most suitable solvents. Acids with amines can form salts, which cannot be classified as complex compounds; but, as the experimental results suggest, this does not always occur. Moreover, salt formation also has a definite influence on the activity of hydrogen, so that salts, no less than complexes, deserved attention in our work.

We have so far studied the following systems:

1) Alcohols with nitro compounds and amines

1. \(\mathrm{CH_3OH + C_6H_5NO_2}\)

Alcohol	\(\mathrm{C_6H_5NO_2}\)	EMF in \(mv\)
0	100	0,9
10	90	80,0
20	80	110
50	50	850
100	0	1850

2. \(\mathrm{CH_3OH + C_5H_5N}\) (pyridine)

\(\mathrm{CH_3OH}\)	\(\mathrm{C_5H_5N}\)	EMF in \(mv\)
0	100	4,6
10	90	105
20	80	125
50	50	1050
100	0	1850

3. \(\mathrm{CH_3CH_2 \cdot CH_2 \cdot CH_2OH + C_5H_5N}\)

\(\mathrm{C_4H_9OH}\)	\(\mathrm{C_5H_5N}\)	EMF in \(mv\)
0	100	4,6
10	90	12,6
50	50	125
70	30	155
100	0	260

\[ \mathrm{(CH_3)_3COH + C_5H_5N.} \]

Maximum EMF \(= 6,1\ mv\) (at 20% alcohol).

4. \(\mathrm{(CH_3)_3COH + C_6H_5NO_2}\)

Maximum EMF \(= 4,3\ mv\) (at 20% alcohol).

The results of these experiments show that the alcohols we used either do not form complexes with amines and nitro compounds or, if they do give complexes, then in the activity of hydrogen they differ little from the alcohols themselves. Only in tertiary alcohols and in some approaching them in…

in certain ratios of cyclohexanol, a small maximum of EMF was observed in mixtures with pyridine and nitrobenzene¹).

2. Phenols with nitro compounds²).

Phenols gave an entirely different picture.
We shall give the values of the maximum electromotive force for several mixtures of nitrobenzene with phenol, the three cresols, and naphthols:

Phenols	Max.³) EMF in mv.	Molecular % of phenol	Molecular % of C₆H₅NO₂
C₆H₅OH	1400	25	75
o—CH₃C₆H₄OH	820	—	—
m—CH₃C₆H₄OH	1100	—	—
p—CH₃C₆H₄OH	900	—	—
α—C₁₀H₇OH	1350	10	90
β—C₁₀H₇OH	1200	10	90

3) Phenols with amines⁴).

Mixtures of phenols with pyridine have been studied in greatest detail. We give the figures for the system C₆H₅OH + C₅H₅N.

No.	Weighed portion C₆H₅OH (in g)	Amount C₅H₅N (in cu. cm)	Molecular ratio	EMF	No.	Weighed portion C₆H₅OH (in g)	Amount C₅H₅N (in cu. cm)	Molecular ratio	EMF
1.	0.23	5	—	355	8.	1.87	5	—	565
2.	0.35	5	—	495	9.	2.35	5	1 : 2.5	1000
3.	0.46	5	—	585	10.	2.88	5	—	170
4.	0.58	5	—	610	11.	3.29	5	—	155
5.	0.69	5	1 : 8	615	12.	3.76	5	—	115
6.	0.92	5	—	500	13.	4.7	0	—	90
7.	1.38	5	1 : 4	300

Here the presence of two maxima—615 and 1000 mv.—with a sharply marked minimum (300 mv.) between them is especially interesting. It must be supposed that this corresponds to the existence of different⁵) complexes. It is difficult for the present to say whether the maximum or minimum value of the EMF corresponds to a definite compound. The fact that the special points of the curve are shifted toward pyridine indicates the participation in the complexes of several molecules of pyridine per one molecule of phenol. Determination of the electromotive force in benzene solution (1.41 g C₆H₅OH + 2.8 cu. cm C₅H₅N + C₆H₆ to 6 cu. cm) gave a maximum (350) at the same ratio of components (approximately 1 C₆H₅OH + 2.5 C₅H₅N) as corresponds to the main maximum in experiments without solvent.

Cresols behave, in general, in the same way as phenol. EMF maxima were observed at 20–30%. Likewise for phenol with quinoline.

¹ B. V. Tronov and L. P. Kulev, “Activity of hydrogen in complexes of alcohols with unsaturated organic compounds” (in press).

² As yet unpublished work of L. P. Kulev.

³ In some cases the course of the change in EMF as a function of the relative amounts of the components was not traced; rather, only 1–2 experiments were made with mixtures close to the optimum composition.

⁴ L. P. Kulev, “On the nature of complexes of phenols with amines” (in press).

⁵ Zh. Kh. P. 7, No. 8–9 (1930).

In mixtures of phenol with aromatic amines (aniline, dimethylaniline), the electromotive force gradually increases with increasing content of C₆H₅OH. Triethylamine gave a large increase in EMF (up to 850 mv) at 90% phenol.

It may be thought that pyridine molecules attach not to the hydroxyl of phenol, but to the carbon atoms of the benzene ring. In this case, electronic considerations make it probable that it is not the nitrogen atom that attaches, but the carbon adjacent to it, to the negatively charged carbons (ortho and para) of phenol. Then the complex would have the composition C₆H₅OH·3C₅H₅N, which does not contradict the results of our experiments. With triethylamine, whose nitrogen is more capable of addition reactions, simply a salt may be obtained, and the possibility of association of phenol molecules is not excluded.

4. Acids with nitro compounds.

There are indications in the literature that nitrobenzene gives complex compounds with certain mineral acids. Thus, with H₂SO₄ a crystalline substance of composition C₆H₅NO₂·H₂SO₄ is formed.¹)

As the study of the kinetics of nitration of nitrobenzene has shown,²) it must give two complexes with nitric acid:

\[ \mathrm{C_6H_5NO_2 \cdot HNO_3} \quad \text{and} \quad \mathrm{C_6H_5NO_2 \cdot 2HNO_3}. \]

The formation of such compounds accounts for the very slow action on C₆H₅NO₂ of small amounts of HNO₃—up to 1 gram-molecule, the considerably faster nitration by the second molecule, and an even greater acceleration upon addition of the third molecule of acid. Only the third molecule proves not to be bound by the NO₂ group and can react freely with the benzene nucleus.

All the facts mentioned led one to expect the formation of similar compounds with carboxylic acids as well. This was fully confirmed by determining the electromotive force of the reaction of the corresponding mixtures with sodium.³) We took the nitro compounds: C₆H₅NO₂, m-C₆H₄(NO₂)₂, and α-C₁₀H₇NO₂, in combination with the acids: CH₃COOH, CH₃·CH₂·COOH, CH₃·CH₂·CH₂·COOH, (CH₃)₂CH·COOH, (CH₃)₂CH·CH₂·COOH. Since such mixtures by themselves reacted too vigorously with Na, it was necessary to dilute them with benzene: \(n/200\) gram-mol. RCOOH + \(^{1}/_{200}(10-n)\) gram-mol. RNO₂ + C₆H₆ to 6 cubic cm.

Owing to the rather poor solubility of the solid nitro compounds, it was not always possible to follow the system under study completely. The experiments gave the following values of the electromotive force (see the table on p. 386):

The EMF values found sometimes exceed by many times those given by the acids themselves at the corresponding concentration, i.e., 0.05 gram-mol. + C₆H₆ to 6 cubic cm. Thus propionic acid without nitro compounds gave 0.98 mv, butyric acid—0.8 mv, isovaleric acid even 0.13 mv, i.e., 700 times less than its mixture with nitrobenzene. Such a difference can be explained only by the formation of some compound which, in hydrogen activity, considerably surpasses the acid taken. The position of the maximum of the electromotive force indicates a somewhat different composition of the complexes obtained. In many cases it is apparently close to 2 mol. of acid per 1 mol. of nitro derivative, but sometimes, as in the most active complex of isovaleric acid with nitrobenzene, we find a ratio—

¹) E. Chembuliez, Chem. Zbl. 1923 I–II, 1491.
²) B. V. Tronov and A. P. Grigor’eva, “Izv. SFTI” 1932.
³) B. V. Tronov and L. P. Kulev. Zh. O. Kh. 5.

ratio 1:1. It is possible that some acids add to the nitro group as dimeric molecules, others as simple molecules, which of course may affect the acidity of the complex.

Nos.	Acids	Nitro compounds	Acids	Nitro compounds	EMF in %
1.	$\mathrm{CH_3COOH}$	$\mathrm{C_6H_5NO_2}$	60	40	5.4
2.	$\mathrm{CH_3COOH}$	$\mathrm{C_6H_4(NO_2)_2}$	70	30	(12.7)
3.	$\mathrm{CH_3COOH}$	$\alpha$-$\mathrm{C_{10}H_7NO_2}$	70	30	15.5
4.	$\mathrm{C_2H_5COOH}$	$\mathrm{C_6H_5NO_2}$	60	40	8.4
5.	$\mathrm{C_2H_5COOH}$	$\mathrm{C_6H_4(NO_2)_2}$	70	30	(6.0)
6.	$\mathrm{C_2H_5COOH}$	$\mathrm{C_{10}H_7NO_2}$	70	30	16.2
7.	$\mathrm{C_3H_7COOH}$	$\mathrm{C_6H_5NO_2}$	60	40	11.9
8.	$\mathrm{C_3H_7COOH}$	$\mathrm{C_6H_4(NO_2)_2}$	80	20	(4.2)
9.	$\mathrm{C_3H_7COOH}$	$\mathrm{C_{10}H_7NO_2}$	70	30	11.9
10.	$i$-$\mathrm{C_3H_7COOH}$	$\mathrm{C_6H_5NO_2}$	60	40	55
11.	$i$-$\mathrm{C_3H_7COOH}$	$\mathrm{C_6H_4(NO_2)_2}$	80	20	(13.3)
12.	$i$-$\mathrm{C_3H_7COOH}$	$\mathrm{C_{10}H_7NO_2}$	70	30	(35)
13.	$i$-$\mathrm{C_4H_9COOH}$	$\mathrm{C_6H_5NO_2}$	50	50	90
14.	$i$-$\mathrm{C_4H_9COOH}$	$\mathrm{C_6H_4(NO_2)_2}$	80	20	(1.9)
15.	$i$-$\mathrm{C_4H_9COOH}$	$\mathrm{C_{10}H_7NO_2}$	60	40	(35)

Note: the numbers in parentheses may not be maximum EMF values; a larger amount of the nitro compound could not be taken because of poor solubility.

The electromotive force of the reaction with sodium of the complex of an acid with a nitro compound bears no relation to the degree of dissociation of the same acid in aqueous solution. For the same five acids we find the following constants of electrolytic dissociation:
$\mathrm{CH_3COOH}$ — $K = 1.8 \cdot 10^{-5}$;
$\mathrm{CH_3CH_2COOH}$ — $1.34 \cdot 10^{-5}$;
$\mathrm{CH_3CH_2CH_2COOH}$ — $1.52 \cdot 10^{-5}$;
$\mathrm{(CH_3)_2CHCOOH}$ — $1.5 \cdot 10^{-5}$;
$\mathrm{(CH_3)_2CHCH_2COOH}$ — $1.7 \cdot 10^{-5}$.

In practical terms, of particular interest is the sharp difference in the electromotive force of complexes given by acids with a straight and a branched chain. We know of no other readily determinable constant whose value, for example, for butyric and isobutyric acids would differ by a factor of 5. Even the rate of oxidation by permanganate for isobutyric acid proved to be only 2.5 times greater¹. It is possible that good results for distinguishing acids with a straight or branched chain could be given by chlorination with antimony pentachloride, which is recommended as a method for determining hydrocarbons of iso-structure², but our method is considerably simpler.

EMF of systems: carboxylic acids in a mixture with nitrobenzene in benzene solution.

Scale: along the abscissa \(1 - 0.8\ \mathrm{mV}\); along the ordinate \(1\ \mathrm{mV} - 0.8\ \mathrm{mm}\).

\(C_2H_4O_2\), valeric acid

Curve 1.

\(C_7H_5O_2\)

Curve 2.

\(C_2H_6O_2\)

Curve 3.

\(i\ C_6H_{12}O_2\)

Curve 4.

\(i\ C_5H_{10}O_2\)

Curve 5.

\(C_2H_4O_2 \cdot 2H_2O\)

Curve 6.

Maximum EMF of carboxylic acids in nitrobenzene solution.

Scale: \(1\ \mathrm{cm} - 100\%\ \mathrm{k\text{-}moles}\); \(1\ \mathrm{mV} - 0.8\ \mathrm{mm}\).

\(C_2H_4O_2\)

\(C_3H_6O_2\)

\(C_4H_8O_2\)

\(i\ C_5H_{10}O_2\)

\(i\ C_6H_{12}O_2\)

Curve 7.

Maximum EMF of carboxylic acids in mixtures with \(m\text{-}C_6H_4(NO_2)_2\).

Scale: \(1\ \mathrm{cm} - 100\%\ \mathrm{k\text{-}moles}\); \(1\ \mathrm{mV} - 0.8\ \mathrm{mm}\).

\(C_2H_4O_2\) — 12.7

\(C_3H_6O_2\) — 6

\(C_4H_8O_2\) — 4.2

\(i\ C_5H_{10}O_2\) — 3.3

\(i\ C_6H_{12}O_2\) — 1.8

Curve 8.

5) Acids and amines.

Acids with amines should give salts:

\[ \mathrm{CH_3—C(=O)—OH + CH_3NH_2 \rightarrow CH_3—C(=O)—ONH_3—CH_3.} \]

It is not impossible that sometimes a dimeric molecule will take part in the formation of the salt: \((\mathrm{CH_3COOH})_2 + \mathrm{CH_3NH_2}\). According to previously generally accepted ideas, one would have expected a decrease in the activity of the hydrogen in salts. The modern interpretation of the mechanism of electrolytic dissociation made it possible to think, on the contrary, that the activity of the carboxyl hydrogen upon salt formation may increase.

Experiments were carried out¹ with propionic, butyric, isovaleric, and isobutyric acids and with the amines triethylamine, pyridine, and aniline. The determination of the electromotive force of the reaction with sodium was carried out sometimes without solvent, but for the most part in benzene solution of the same concentration as in the preceding works.

Experiments with \(\mathrm{C_5H_5N}\) without solvent.

% acids	EMF with acids: \(\mathrm{C_3H_6O_2}\)	EMF with acids: \(n\,\mathrm{C_4H_8O_2}\)	EMF with acids: \(i\,\mathrm{C_5H_{10}O_2}\)
10	70 mv.	40 mv.
30	100 "
50	560 "	175 "
60	910 "	550 "
70	1700 "	1050 "	750 mv.
80	1300 "	700 "
100		150 "	70 "

Experiments in benzene solution.

Acids	EMF of the acid itself	\((\mathrm{C_2H_5})_3\mathrm{N}\) max. EMF	% acid	\(\mathrm{C_5H_5N}\) EMF	% acid	\(\mathrm{C_6H_5NH_2}\) EMF	% acid
\(\mathrm{C_3H_6O_2}\) . . .	0.98	500	80	1.6	70	no maximum	no maximum
\(n\,\mathrm{C_4H_8O_2}\) . .	0.8	600	70	1.45	70	no maximum	no maximum
\(i\,\mathrm{C_5H_{10}O_2}\) . .	0.13	300	80	50	50	2.2	80

We see that compounds of acids with amines give, with sodium, an electromotive force many times greater than the values observed for free acids. The clear parallelism between EMF and the ability of the amine toward salt formation speaks in favor of the salt-like nature of the compounds obtained. Salts, apparently, are formed more rarely with simple acids, and more often with associated acid molecules. In any case this may be said of salts with the most active hydrogen.

All our experiments with compounds possessing active hydrogen give good confirmation of the view now becoming ever more firmly rooted in science, that propo-

...that acidic properties are possessed not by the molecule itself of acetic, propionic, nitric, and other acids, but by its complex with some other substance.

Comparing various compounds with one another in terms of the strength of their acidic properties, we arrive at different results depending on the nature of the solvent, the concentration of the solution, etc. For example, water, propyl alcohol, propionic acid, and phenol, in their acid activity, are arranged in the following series:

Aqueous medium: \( \mathrm{C_2H_5COOH > C_6H_5OH > H_2O > C_3H_7OH} \).

Pure substances: \( \mathrm{HOH > C_3H_7OH > C_2H_5COOH} \) (\(\mathrm{C_6H_5OH}\) is excluded, as a solid substance at the temperature of the experiments).

Benzene solution: \( \mathrm{C_3H_7OH > C_2H_5COOH > C_6H_5OH} \) (\(\mathrm{H_2O}\) is excluded because of insolubility).

Concentrated solution in pyridine or nitrobenzene:

\[ \mathrm{C_2H_5COOH > C_6H_5OH > C_3H_7OH}. \]

Dilute solution in pyridine or nitrobenzene:

\[ \mathrm{C_6H_5OH > C_2H_5COOH > C_3H_7OH}. \]

Consequently, one must speak not of one compound being a stronger acid and another a weaker one, but of such-and-such a compound having stronger acidic properties under such-and-such conditions.

III. Unsaturated organic compounds and the complexes formed by them.

Direct determination of the activity of unsaturated compounds from the electromotive force of the reaction with sodium yielded few results. For substances that cannot pass into a form with active hydrogen, the electromotive force was always very small. Even without solvent, \(\mathrm{C_6H_5NO_2}\) gave \(0.7\)—\(0.8\) mv. At a concentration of \(1/40\) gram-mole per 6 cubic cm (diluted with benzene), the following EMF values were obtained:
\(\mathrm{C_6H_5NO_2}\) ... \(0.3\) mv, \(\mathrm{C_6H_5CN}\) ... \(3.4\) mv, \(\mathrm{C_6H_5CO \cdot CH_3}\) ... \(0.1\) mv. For indene the EMF \(= 0\) (also without solvent).

The question arose: could one not, similarly to what had been done with compounds possessing active hydrogen, increase the electromotive force by complex formation? As the other component of the complex it was necessary to choose a substance that would itself not react with sodium.

The very process causing the appearance of electromotive force in our experiments must consist in the fact that the positive part of the atom is attracted to the negatively charged atom among those between which the double bond is located (i.e., to the oxygen of ketones, the nitrogen of nitriles, etc.), while the valence electron opened from the sodium atom goes into the external circuit. Hence one could suppose that, if it were possible in some way to impart a negative charge to the molecule of the unsaturated compound under investigation, then the molecule would begin to attract the cation more strongly and repel the electron, i.e., the electromotive force of the reaction should increase under these conditions. Best of all, therefore, would be to add some anion as a complex.

Among the simplest anions, iodine is known to form complexes especially well. Aqueous solutions of iodide salts dissolve free iodine with formation of the ion \(\mathrm{J'_3}\):

\[ \mathrm{K^{\bullet} + J' + J_2 \rightarrow K^{\bullet} + J'_3} \]

Sodium iodide forms complex compounds with certain organic substances: for example, with acetone there is obtained¹ \( \mathrm{NaJ}\cdot 3\mathrm{C}_3\mathrm{H}_6\mathrm{O} \).

To determine the activity of unsaturated organic compounds, in one of the studies of our laboratory² triple complexes of sodium iodide, an unsaturated compound (ketone, ester, nitrile, etc.), and a trinitro derivative of the benzene series (trinitrobenzene, trinitrotoluene, trinitrophenol) were used. In all these complexes NaJ participates, probably, through the iodine ion³.

Therefore, in the work on electromotive force we decided to use, as the second complex-forming agent, the same sodium iodide, which, incidentally, dissolves rather well in many organic substances (evidently precisely because of complex formation). With NaJ the compound under investigation, for example \((\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C}=\mathrm{O}\), should give a complex anion:

\[ \left[ \begin{array}{c} (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C}-\mathrm{O} \\ \vdots \\ \mathrm{J} \end{array} \right]^{\prime} \]

(possibly of a somewhat different composition, but constructed according to the same scheme) with the properties we need.

However, with sodium iodide, not only ketones but also other compounds with double (or triple) bonds can form complexes. This salt dissolves in alcohols, acids, and amines. It was quite possible that, when it dissolves in hydroxyl-containing compounds, the electromotive force given by them would also increase. Perhaps the hydrogen of the hydroxyl group in the complex with NaJ becomes more active. The effect of a change in electrical conductivity is also not excluded. All this had to be tested experimentally.

The following table gives the results of determinations of the electromotive force for representatives of various classes of organic substances, with and without addition of NaJ. The experiments were carried out in benzene solution at various concentrations and sometimes with various amounts of sodium iodide. The salt was not always completely dissolved.

Tested unsaturated compounds	Amount of substance (diluted to 6 cu. cm. with benzene)	EMF in mv	Weighed portion in g.	EMF with NaJ
Indene	3 cu. cm.	—	0.01	—
Indene	6 “	—	0.01	0.25
Indene	(without solvent)	—
\(\mathrm{CH}_3\cdot\mathrm{CH}=\mathrm{O}\)	3 cu. cm.	30	0.01	800
\(\mathrm{CH}_3\cdot\mathrm{CH}=\mathrm{O}\)	1.1 “	0.4	0.01	1.5
\(\mathrm{C}_6\mathrm{H}_5\cdot\mathrm{CH}=\mathrm{O}\)	3 “	2.3	0.01	480
\(\mathrm{C}_6\mathrm{H}_5\cdot\mathrm{CH}=\mathrm{O}\)	2 “	0.45	0.01	50
\(\mathrm{C}_6\mathrm{H}_5\cdot\mathrm{CH}=\mathrm{O}\)	1.3 “	0.1	0.01	35
\(\mathrm{CH}_3\cdot\mathrm{CO}\cdot\mathrm{CH}_3\)	3 “	45	0.005	425
\(\mathrm{CH}_3\cdot\mathrm{CO}\cdot\mathrm{CH}_3\)	3 “	45	0.01	600
\(\mathrm{CH}_3\cdot\mathrm{CO}\cdot\mathrm{CH}_3\)	3 “	45	0.05	700
\(\mathrm{CH}_3\cdot\mathrm{CO}\cdot\mathrm{CH}_3\)	2.4 “	8.2	0.01	190

Unsaturated compounds tested	Amount of substance (diluted to 6 cu. cm with benzene)	EMF in mv	Sample weight in g	EMF with NaJ
$\mathrm{CH_3.CO.CH_3}$	1,5	0,7	0,01	1,3
$\mathrm{CH_3.CO.C_2H_5}$	3	15,8	0,01	290
$\mathrm{CH_3.CO.C_2H_5}$	1,8	0,7	0,01	45
$\mathrm{CH_3.CO.C_3H_7}$	3	2,8	0,01	300
$\mathrm{CH_3.CO.C_3H_7}$	1,8	0,3	0,01	25
$\mathrm{C_2H_5.CO.C_2H_5}$	3	3,4	0,01	25
$\mathrm{C_2H_5.CO.C_2H_5}$	2,1	0,85	0,01	5,9
$\mathrm{C_3H_7CO.C_3H_7}$	3	0,35	0,01	9,5
Cyclohexanone	3	2,7	0,01	600
Cyclohexanone	2,1	0,7	0,01	310
$\mathrm{C_6H_5.CO.CH_3}$	3	2,2	0,01	65
$\mathrm{C_6H_5.CO.CH_3}$	1,9	0,4	0,01	4,8
$\mathrm{C_6H_5.CO.C_2H_5}$	3	1,1	0,01	90
$\mathrm{C_6H_5.CO.C_2H_5}$	2,6	0,6	0,01	50
$\mathrm{C_6H_5.CO.C_6H_5}$	3,64 g	0,1	0,01	25
$\mathrm{(CH_3)_2C{=}CH.CO.CH_3}$	3 cu. cm	5,6	0,005	460
$\mathrm{(CH_3)_2C{=}CH.CO.CH_3}$	2,3	2,9	0,01	490
$\mathrm{CH_3.COOCH_3}$	3	0,35	0,01	1,9
$\mathrm{(COOC_2H_5)_2}$	3	0,1	0,01	2,3
$\mathrm{C_6H_5.COOC_2H_5}$	3	0,1	0,01	0,7
$\mathrm{CO(OCH_3)_2}$	3	—	0,01	0,1
$\mathrm{CH_3.CO.CH_2.COOC_2H_5}$	3	25	0,01	350
$\mathrm{CH_3.CO.CH_2.COOC_2H_5}$	2,5	5,7	0,01	175
$\mathrm{C_6H_5OCH_3}$	3	0,2	0,01	1,4
$\mathrm{C_2H_5.ONO_2}$	3	0,7	0,01	1,2
$\mathrm{C_5H_{11}.ONO_2}$	3	0,25	0,01	0,35
$\mathrm{C_5H_{11}.ONO}$	3	0,3	0,01	3,5
Pyrrole	3	0,25	0,01	0,45
$\mathrm{(C_2H_5)_3N}$	6	—	0,01	—
$\mathrm{C_5H_5N}$ (pyridine)	3	0,9	0,01	85
$\mathrm{C_5H_5N}$ (pyridine)	2	0,2	0,01	20
$\alpha\text{-}\mathrm{CH_3C_5H_4N}$	3	0,3	0,01	40
$\beta\text{-}\mathrm{CH_3.C_5H_4N}$	3	0,3	0,01	40
Quinoline	3	0,4	0,01	35
$\mathrm{CH_3NO_2}$	1,1	0,1	0,01	0,1
$\mathrm{C_6H_5NO_2}$	3	0,3	0,01	1,8
$\mathrm{C_6H_5NO_2}$	2	0,05	0,01	0,5
$\mathrm{CH_3CN}$	3	500	0,01	650
$\mathrm{CH_3CN}$	1	55	0,01	60
$\mathrm{C_6H_5CN}$	3	3,4	0,01	180
$\mathrm{C_6H_5CN}$	2	1,1	0,01	75
$\mathrm{C_6H_5COCN}$	2,6	12,3	0,01	60
$\mathrm{CH_3CH_2CH_2OH}$	3	200	0,01	290
$\mathrm{CH_3CH_2CH_2OH}$	1,5	0,35	0,01	2,0
$\mathrm{CH_3.CHOH.CH_3}$	3	40	0,01	150
$\mathrm{CH_3.CHOH.CH_3}$	1,9	0,3	0,01	3,5
$\mathrm{CH_3CH_2CH_2CH_2OH}$	3	30	0,01	50
$\mathrm{CH_3CH_2CH_2CH_2OH}$	1,8	1,0	0,01	1,2
$\mathrm{(CH_3)_3COH}$	3	—	0,01	0,3
$\mathrm{(CH_3)_3COH}$	1,9	—	0,01	0,05
$\mathrm{CH_2{=}CHCH_2OH}$	3	310	0,01	400
$\mathrm{CH_2{=}CHCH_2OH}$	1,4	1,3	0,01	1,7
Cyclohexanol	3	0,1	0,01	0,1
$\mathrm{C_6H_5{-}CH_2OH}$	3	0,7	0,01	7,3
$\mathrm{C_6H_5{-}CH_2OH}$	2,1	0,15	0,01	1,6
$\mathrm{CH_3COOH}$	3	3,2	0,01	6,6
$\mathrm{CH_3COOH}$	2,3	0,2	0,01	0,25
$\mathrm{CH_3CH_2COOH}$	3	4,5	0,01	4,5
$\mathrm{CH_3CH_2COOH}$	2,1	0,95	0,01	2,3
$\mathrm{(CH_3)_2CH.CH_2.COOH}$	3	0,1	0,01	0,2
$\mathrm{(CH_3)_2CH.CH_2.COOH}$	2,2	0,1	0,01	0,15

From the figures in this table it is evident that the electromotive force of the reaction with sodium of compounds containing a double or triple bond, and incapable of passing into the enol (or analogous) form, sometimes increases manyfold. Such are ketones (benzophenone), nitriles (benzonitrile), and heterocycles of the pyridine group. In some esters (nitrates) the EMF does not change; other esters, e.g., esters of benzoic and carbonic acids, occupy an intermediate position. The difference from ketones is quite understandable, since the application of other methods¹ has also shown that the carbonyl group in ketones is more active. Nitro compounds occupy an intermediate position.

The activity of double bonds in pyrrole is sharply lowered. For indene a noticeable increase in EMF was observed, which, apparently, should be attributed to its unsaturation.

In compounds with mobile hydrogen, the electromotive force of the reaction with sodium either hardly increases, or increases not so strongly. Substances which can, in one way or another, pass into the enol form behave differently, and this can be used (though with certain reservations) to decide the question of which form predominates in them. It is interesting, for example, that nitromethane in our experiments proves to be more similar to acids than to nitro compounds, i.e., it reacts in the iso-form:

\[ \begin{array}{c} \mathrm{CH_2=N=O}\\ \ \ \ \vert\\ \mathrm{OH} \end{array} \]

It even turned out that, like carboxylic acids, it markedly increased its activity toward sodium in mixtures with nitrobenzene and pyridine.

Analysis of mixtures and standardization of organic substances.

If one takes a mixture of two compounds, one reacting with sodium and one not reacting, then it is natural that the electromotive force given by this mixture with metallic Na will change more or less sharply depending on the content of the active component. We shall cite the results of several experiments with benzene solutions of alcohols and acids.

No. Na, sequential	Substances investigated	Amount of substance in cc (in 6 cc benzene)	EMF	No. Na, sequential	Substances investigated	Amount of substance in cc (in 6 cc benzene)	EMF
1	$\mathrm{C_2H_5OH}$	0.29	—	5	$\mathrm{CH_3COOH}$	0.29	—
1	$\mathrm{C_2H_5OH}$	0.58	—	5	$\mathrm{CH_3COOH}$	0.58	—
1	$\mathrm{C_2H_5OH}$	0.87	1.0	5	$\mathrm{CH_3COOH}$	0.87	—
1	$\mathrm{C_2H_5OH}$	1.45	12.0	5	$\mathrm{CH_3COOH}$	1.74	2.5
2	$\mathrm{C_3H_7OH}$	0.37	—	6	$\mathrm{CH_3CH_2COOH}$	0.37	—
2	$\mathrm{C_3H_7OH}$	0.74	—	6	$\mathrm{CH_3CH_2COOH}$	0.74	—
2	$\mathrm{C_3H_7OH}$	1.11	0.5	6	$\mathrm{CH_3CH_2COOH}$	1.48	—
2	$\mathrm{C_3H_7OH}$	1.84	10.0	6	$\mathrm{CH_3CH_2COOH}$	2.22	10.0
3	i-$\mathrm{C_3H_7OH}$	0.37	—	7	$\mathrm{C_3H_7COOH}$	0.45	—
3	i-$\mathrm{C_3H_7OH}$	0.74	—	7	$\mathrm{C_3H_7COOH}$	0.90	—
3	i-$\mathrm{C_3H_7OH}$	1.11	—	7	$\mathrm{C_3H_7COOH}$	1.35	—
3	i-$\mathrm{C_3H_7OH}$	1.85	2.0	7	$\mathrm{C_3H_7COOH}$	2.7	2.0
4	$\mathrm{C_4H_9OH}$	0.44	—	8	$\mathrm{C_6H_5OH}$	Weight in g	—
4	$\mathrm{C_4H_9OH}$	0.90	—	8	$\mathrm{C_6H_5OH}$	0.23	—
4	$\mathrm{C_4H_9OH}$	1.34	—	8	$\mathrm{C_6H_5OH}$	0.35	—
4	$\mathrm{C_4H_9OH}$	2.22	3.0	8	$\mathrm{C_6H_5OH}$	0.46	4.2
				8	$\mathrm{C_6H_5OH}$	—	—

Thus, from the electromotive force of the reaction with sodium it is possible to determine the content of alcohol or acid in certain solutions.

However, in this simplest form the method cannot find wide application. The obstacles are:

1) The fact that very many substances of diverse classes react with sodium.
2) The insignificance of the EMF for many compounds, even in pure form.
3) The shape of the curve of EMF change with concentration. For example, comparatively substantial concentrations (up to 10%) of alcohol, phenol, or acid in benzene are not detected at all.

The limits of application of our method are expanded many times if complex formation is used. Let us give several examples:

1) Determination of acids and phenols.

We have seen that the electromotive force given by some compound with mobile hydrogen—for example, some carboxylic acid—sometimes increases hundreds of times upon the addition of another substance (nitrobenzene, etc.), which by itself is very little active toward sodium. This greatly increases the accuracy of determining the content of acids and similar substances. In addition, we can very easily distinguish whether the solution contains alcohol, phenol, or acid. Alcohol can be distinguished because it gives no increase in EMF either with nitrobenzene or with amines; phenols and acids are also similar to one another, but with amines they nevertheless behave somewhat differently. We have so far examined in greater detail the system benzene + phenol, since the determination of phenol content in benzene hydrocarbons is often encountered at carbonization plants (fractions of coal tar).

Solutions containing from 1% to 6% phenol were taken. Such solutions with sodium gave no electromotive force. The addition of \(C_6H_5NO_2\) sharply increased the EMF. We first established, using the example of a 1% solution, what quantity of nitrobenzene gives the greatest effect. It was found:

Amount of 1% solution of \(C_6H_5OH\) in \(C_6H_6\)	Added \(C_6H_5NO\)	Maximum EMF in mv
5 cu. cm.	1 cu. cm.	0.3
4 »	2 »	4.0
3 »	3 »	14.0
2 »	4 »	14.8
1 »	5 »	5.8

Since the ratios 3:3 and 2:4 gave almost identical results, we next settled on the simplest practical procedure: adding to the solution under investigation an equal volume of nitrobenzene. With different phenol concentrations the following was obtained:

1% ... max. EMF = 14.0; 2%—40 mv; 45 mv; 40 mv (different experiments); 3% ... 65 mv; 60 mv. 4%—100; 100 mv; 100 mv.
5%—210 mv; 200 mv
6%—345 mv; 275 mv

Pure \(C_6H_5NO_2\) (used for the experiments): 6.2 mv.

Each percent of \(C_6H_5OH\) gives, as we see, a very noticeable difference in EMF. Taking experimental errors into account, it may be considered that the determination of phenol in benzene solution can be carried out with an accuracy up to 0.1–0.2% (at any rate when the phenol content is up to 5%).

—still unpublished work of L. P. Kulev and Vasil’eva.

2) Determination of Water.

A very frequently encountered impurity in organic compounds, from which it is sometimes rather difficult to free oneself, is water.

The discovery of rapid and convenient methods for determining the amount of water in alcohols (e.g., ethyl alcohol), acids (e.g., acetic acid), etc., is of great importance both in the practice of laboratory work and in industry. Experiments have shown that in certain compounds an admixture of even very small amounts of water can easily be recognized by determining the electromotive force of the reaction with sodium, given by the compound itself (i.e., probably by the water contained in it). Thus, with pyridine the following values of the EMF were obtained:

Pure pyridine 4 cub. cm + 2 cub. cm \(C_6H_6\) . . . 4 mv.
4 cub. cm pyr. + 2 cub. cm \(C_6H_6\)—1 drop \(H_2O\) (about 1%) 8 mv.
" " " " " " " " 2 " " ( " 2%) 45 mv.
" " " " " " " " 3 " " ( " 3%) 1150 mv.

In other cases it is useful to add still another compound, forming with the substance under investigation a more active complex. For example, acetic acid, which by itself under the conditions of our experiments (see above), in a mixture with nitrobenzene in benzene solution, gave with sodium a maximal EMF . . . 5 mv, upon addition of 1% \(H_2O\) gave 40 mv; 2% \(H_2O\) raised the EMF to 75; 3% . . . to 112 mv.

In this way the content of a small amount of water in an acid can be determined no less accurately and in any case more rapidly than by other known methods, especially since determination by specific gravity is hindered by the contraction occurring when \(CH_3COOH\) is mixed with \(H_2O\); therefore, one and the same specific gravity sometimes corresponds to two compositions of the mixture.

3) Determination of Various Other Impurities.

In our experiments it sometimes happened that one or another organic compound behaved toward sodium not as would be expected from the composition and structure of that compound.

For example, cyclohexanol by itself gave a small EMF (as it should), but upon addition of nitrobenzene or pyridine the electromotive force increased manyfold (from 0.15 mv to 24 mv with \(C_6H_5NO_2\), and from 0.15 mv to 6.5 mv with \(C_5H_5N\)), i.e., cyclohexanol proved to resemble phenol. Since the preparation we had was of the grade “gereinigt,” i.e., not particularly pure, and since it is produced technically from phenol, it was natural to explain the anomalous properties by an impurity of the latter. Indeed, after washing with an alkali solution (removal of phenol) and redistillation, cyclohexanol only slightly increased the EMF upon addition of complex-forming agents.

Benzyl alcohol, which had stood for several years and had probably partly oxidized to benzoic acid, gave a very large increase in EMF upon addition of triethylamine. This increase disappeared after distillation of the alcohol over lime.

A more complex example, but at the same time one very promising for the practical application of the method, is found with nitrobenzene. \(C_6H_5NO_2\), as is clear from all the preceding, is one of the best auxiliary reagents (complex-forming agents) by means of which one can determine the activity, and from it the structure and purity, of many compounds with mobile hydrogen. For this purpose we at first used the best Kahlbaum nitrobenzene. Later we had to switch to a preparation of our own manufacture. However, by itself it gave with sodium an EMF of about 4 mv.

(instead of 0.7–0.8 mv for Kalbaum’s). By more thorough drying of the substance it was possible to reduce the electromotive force only slightly. In the end our nitrobenzene was brought up to the Kalbaum standard only by fractional distillation, the smallest EMF being given by the second half of the distillate. It may be supposed that the influence here was exerted not only by water, but also by certain impurities.

It was of interest to follow how the very method of obtaining the product affects its quality, since differences in the method of preparation could quite well lead to different impurity contents. A series of experiments was carried out, nitrating benzene with the usual nitrating mixture (nitric acid, sp. gr. 1.4, and sulfuric acid) under several different conditions, for example, with simple shaking for 5, 10, 20 min., etc., or with heating on a water bath, again for different intervals of time. In all experiments the product was subjected to exactly the same treatment: washing with water and soda, drying with calcium chloride, and distillation within a range of two degrees. It turned out that even after such treatment all the preparations gave such different values of the electromotive force of the reaction with sodium that it was possible to determine unmistakably under what conditions the nitration had been carried out:

1) $C_6H_5NO_2$, obtained with heating on a water bath (5 min.) EMF = 8 mv.
2) " " " " " " " (10 min.) EMF = 6.4 mv.
3) " " " " " " " (20 min.) EMF = 4.0 mv.
4) " " " " " " " (30 min.) EMF = 2.2 mv.
5) $C_6H_5NO_2$, obtained without heating, but with shaking (20 min.) EMF = 25.0 mv.
6) " " " " " " (30 min.) EMF = 55.0 mv.
7) " " " " " " (60 min.) EMF = 25.0 mv.
8) $C_6H_5NO_2$ obtained without heating and shaking, and distilled without treatment with water and soda EMF = 11.4 mv.

Under identical nitration conditions, by this method it is possible to distinguish well within what temperature limits the distillation was carried out.

$C_6H_5NO_2$ b.p. 204°–206°  EMF = 6.4 mv. Sample of $C_6H_5NO_2$
" " " 206°–208°  EMF = 3.0 mv. prepared with
                  heating on a water
                  bath (10 min.).

Which impurities here exert an influence on the electromotive force has not yet been clarified. The resolution of this question will be the subject of further work¹. In any case, however, determination of the EMF of the reaction with sodium is evidently a very convenient method for recognizing even small differences in the preparation of a substance and may be proposed as one of the most convenient means of controlling the manufacture of certain organic compounds. The determination of the EMF itself takes, in the case of nitrobenzene, only about 1 min.

In the industry for the production of pure organic reagents, it would be useful to establish, for a number of the most important compounds, definite standard EMF values, the attainment of which should serve as a guarantee of the purity of the preparation. Sometimes this standard may be established from the EMF of the reaction with sodium (or another metal) of the pure substance; sometimes, perhaps, it will be useful to add some compound that increases the activity (as with nitrobenzene, or pyridine in the case of phenols and acids), as we have seen in the example of acetic acid.

Conclusion.

From all that has been said it is evident that the method we propose for determining the electromotive force of reaction with sodium, while perhaps not entirely clear in some respects from the physical point of view, can bring definite practical benefit in organic chemistry. This method:

1) makes it possible to analyze such fine features of molecular structure as, with other methods of investigation, require considerably more work;

2) is very helpful for finding one’s bearings in the large and extremely little-studied field of complex organic compounds;

3) is one of the simplest and most sensitive methods for analyzing mixtures and determining the purity of certain organic preparations.

In the further development of the method the following is envisaged:

1) To subject to investigation a whole series of other compounds, not yet touched upon, or almost not touched upon, in our work, in order to reveal in detail the relationships between the composition and structure of the molecule, on the one hand, and the activity of the compound with respect to metals, on the other. It would be very interesting, for example, to study a larger number of acids with straight and branched chains, containing simultaneously both active hydrogen and unsaturated groups (e.g., nitrophenols), sulfur-containing organic compounds, and to compare the activity of ordinary and heavy hydrogen in acids, alcohols, etc.

2) To reconsider, with the application of our method, the question of the structure of certain complex organic compounds (chinhydrones, picrates, and complexes analogous to them, etc.).

3) To elucidate the influence of solvents on the activity of substances dissolved in them and reacting with sodium. Until now we have used benzene almost exclusively as the neutral solvent; meanwhile, some preliminary experiments show that it is not entirely neutral.

4) To apply the electromotive-force method to the determination of the activity of halides in organic compounds, and also to the determination of the activity of esters that especially readily give up the alcohol radical (such as sulfuric-acid esters).

5) To use other metals instead of sodium (for example, magnesium, which may prove more suitable for halogen compounds).

6) To improve the procedure so that negligible amounts of substance would suffice for the investigation. For this purpose it is planned to use certain mixtures (in which complex compounds have formed) that are especially sensitive to the addition of foreign substances. Thus, a mixture of phenol with pyridine in the molecular ratio 1:4 (1.38 g of $\mathrm{C_6H_5OH}$ in 5 cc of pyridine), which with sodium gave 300 mv, proved to change its activity sharply from a drop of some nitro compound. With a drop of nitrobenzene 625 mv was obtained; with a drop of ortho-nitrotoluene, 275 mv; with meta-nitrotoluene, 450 mv; methods for distinguishing these nitro compounds in such small quantities have not hitherto been indicated in the literature.

It may be hoped that, with proper selection of the mixtures used in the work, this drop method for determining structure may find broad application in the investigation of organic substances.

1. As yet unpublished work of L. P. Kulev and Vasil’eva. ↩↩↩↩↩↩↩

2. B. V. Tronov, L. N. Dyakonova-Shults, E. A. Zonova. Zh.R.Kh.O. 59, pp. 333–345, 1927. ↩↩

3. This is especially proved by certain new, as yet unpublished experiments. ↩