A New Electrochemical Method for Studying Organic Compounds and Their Mixtures

B. V. TRONOV and A. P. KULEV

PART ONE

METHOD OF WORK

Introduction

In the study of organic compounds, an important place is occupied by the determination of functional groups, i.e., the most active atomic groups in the molecule—for example, carboxyl, alcoholic hydroxyl, aldehyde and ketone carbonyl, nitro group, amino group, carbon–carbon double and triple bonds, etc. For each of these groups there are characteristic reactions by means of which the presence of the group in a molecule can be established. In this direction, quite a great deal has already been done in organic chemistry.

Considerably fewer results have so far been obtained on the question of determining the chemical activity of functional groups. Meanwhile, the activity of a given atom or group depends to a high degree on the composition and structure of the remaining parts of the molecule and on the position of the given group in the molecule. From the activity of a functional group we can judge its position and thus take one more step in determining the structure of the substance under study. Moreover, the composition and structure of the entire molecule and the position within it of the group of interest to us sometimes influence the activity of the latter so strongly that the given functional group may give none, or almost none, of the reactions characteristic of it. Thus, the amine nitrogen of triphenylamine does not add many, even strong, acids:

\[ (\mathrm{C}_6\mathrm{H}_5)_3\mathrm{N} + \mathrm{HCl}\ldots \text{does not react.} \]

There are ketones that do not react with bisulfite:

\[ \mathrm{C}_6\mathrm{H}_5 - \mathrm{CO} - \mathrm{C}_6\mathrm{H}_5 + \mathrm{NaHSO}_3 \ldots \text{does not react.} \]

A carbon–carbon double bond sometimes does not add bromine, but reacts with metallic sodium, i.e., behaves differently from the same bond in simple hydrocarbons of the ethylene series:

\[ (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C} = \mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2 + \mathrm{Br}_2 \ldots \text{does not react;} \]

\[ (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C} = \mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2 + 2\mathrm{Na} \rightarrow (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C} - \mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2 \]

\[ \phantom{(\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C} - \mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2} \begin{matrix} | & |\\ \mathrm{Na} & \mathrm{Na} \end{matrix} \]

Such anomalies in the properties of atoms and groups become understandable if a consecutive study is made of changes in activity as a function of the composition and structure of the entire molecule.

To determine the chemical activity of functional groups, a whole series of methods is employed. These include: determination of the rates of reactions in which the given group participates; determination of the electrolytic dissociation constant, which is most often derived from the results of determining the electrical conductivity of the substance in a suitable solvent; determination of the amount of heat liberated in reactions, etc. Each method has its own advantages and shortcomings and its own field of application.

The measurement of the electromotive force arising in the reaction of organic compounds with metals deserves considerable attention. Metals, especially the most active among them, such as the alkali metals, can act on the following groups of organic substances.

1. Compounds with active hydrogen. These include acids (carboxylic, sulfonic, etc.), alcohols and phenols, thiols, primary and secondary amines, amides and imides, oximes, etc., hydrocarbons having hydrogen on a carbon with a triple bond, and hydrocarbons with a cyclopentadiene ring, carbonyl compounds passing into the enol form, many nitro compounds and nitriles. All these compounds react more or less readily with metals, with transfer of an electron from the metal to hydrogen:

\[ \mathrm{R'H} + \mathrm{Na} \rightarrow \mathrm{R'Na} + \mathrm{H}. \]

1. Organo-halogen compounds. In the reaction of organic halides with metals, the metal likewise loses an electron:

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

then

\[ \mathrm{CH_3} + \mathrm{Na} \rightarrow (\mathrm{CH_3})\mathrm{Na} \quad \text{etc.} \]

This too should lead to the appearance of an electromotive force.

1. Compounds with double or triple bonds. Carbonyl compounds, on interaction with sufficiently active metals, either pass into the enol form with readily replaceable hydrogen:

\[ \begin{array}{ccc} \mathrm{CH_3} & \mathrm{CH_2} & \mathrm{CH_2} \\ | & \| & \| \\ \mathrm{C{=}O} & \rightleftarrows \mathrm{COH} + \mathrm{Na} \rightarrow \mathrm{CONa} + \mathrm{H} \\ | & | & | \\ \mathrm{CH_3} & \mathrm{CH_3} & \mathrm{CH_3} \end{array} \]

or directly add the metal to the oxygen atom:

\[ (\mathrm{CH_3})_2\mathrm{C{=}O} + \mathrm{Mg} + \mathrm{O{=}C}(\mathrm{CH_3})_2 \rightarrow \begin{array}{c} (\mathrm{CH_3})_2\mathrm{C} - \mathrm{C}(\mathrm{CH_3})_2 \\ | \qquad | \\ \mathrm{O{-}Mg{-}O} \end{array} \]

\[ \begin{array}{ccccc} \mathrm{C_6H_5} - \mathrm{C} - \mathrm{C_6H_5} + \mathrm{Na} & \rightarrow & \mathrm{C_6H_5} - \mathrm{C} - \mathrm{C_6H_5} \\ \| && | \\ \mathrm{O} && \mathrm{ONa} \end{array} \]

Everywhere the metal passes into the state of a positive ion, losing an electron (or electrons), i.e., here too an electromotive force should arise.

Addition of the metal also occurs with certain unsaturated hydrocarbons. Sometimes a simple addition product is formed:

\[ (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C}=\mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2+2\mathrm{Na} \rightarrow (\mathrm{C}_6\mathrm{H}_5)_2\mathrm{C}-\mathrm{C}(\mathrm{C}_6\mathrm{H}_5)_2 \]

\[ \begin{array}{cc} | & |\\ \mathrm{Na} & \mathrm{Na} \end{array} \]

In other cases the initial product cannot be isolated, and the process ends in polymerization:

\[ n\,\mathrm{CH}_2=\mathrm{CH}-\mathrm{CH}=\mathrm{CH}_2(+\mathrm{Na}) \rightarrow (-\mathrm{CH}_2-\mathrm{CH}=\mathrm{CH}-\mathrm{CH}_2-)_n \]

Nitriles are capable of the same kind of polymerization:

\[ 3\mathrm{C}_6\mathrm{H}_5-\mathrm{C}\equiv\mathrm{N}\,(+\mathrm{Na}) \rightarrow \begin{array}{c} \mathrm{C}_6\mathrm{H}_5\\ |\\ \mathrm{C}\\ /\ \backslash\\ \mathrm{N}\quad \mathrm{N}\\ |\quad \|\\ \mathrm{C}\quad \mathrm{C}\\ \backslash\ /\ \backslash\\ \mathrm{N}\quad \mathrm{C}_6\mathrm{H}_5\\ /\\ \mathrm{H}_5\mathrm{C}_6 \end{array} \]

probably also with intermediate addition of sodium.

Choice of Method and Arrangement of the Experiments

The electromotive force, as a measure of the free energy of a chemical process, should in principle be recognized as a good criterion of the activity of a given organic substance with respect to the metal acting upon it. The reason why this criterion has so far been rarely used is apparently the difficulty of experimentation. Accurate measurement of EMF by the potentiometric method requires a very complicated apparatus. Therefore, it must be supposed, we could find in the literature preceding our investigations only one paper on this question—by Denner [49], who studied the reaction of sodium amalgam with one ethyl alcohol.

Setting ourselves the goal of determining the activity of many organic compounds, we decided to use a simplified, though less accurate, method, taking a voltmeter, i.e., an instrument employed for technical measurements of the voltage of various installations producing electrical energy.

With a voltmeter based on the principle of a galvanometer, we determined the electromotive force of an element composed of the substance under investigation and a pair of metallic electrodes. One electrode was made of the active metal (most often sodium), which reacted directly with the organic compound; the second metal was, in most of our experiments, platinum. A galvanometer, even with a large resistance, directly indicates the strength of the current passing through it, and the current strength depends both on the electromotive force and on the resistance of the system, which in the present case depended chiefly on the electrical conductivity of the substance taken. Here, however, this circumstance only increased the value of the method, since for many compounds the electrical conductivity and the EMF arising upon reaction with metals change depending on

structural changes are parallel. A solution of a stronger acid conducts current better, and with the same acid the galvanic cell described by us should give a larger EMF. Consequently, under the given conditions, the difference between two substances will show up more sharply in the galvanometer readings than if we were to determine the EMF by the classical compensation method, or determine the electrical conductivity with one and the same current source. As the experiments showed (see below), these expectations were fully justified, while the simplicity of the apparatus did indeed make it possible, in a short time, to investigate a very large number of organic compounds.

In the first design of our apparatus [33], the substance under examination, in an amount of 5–6 ml*), was poured into a glass test tube. Into the liquid

Fig. 1. Diagram of the apparatus for the electrochemical investigation of organic compounds.

1. Metal cylinder. 2. Platinum crucible. 3. Ebonite cover. 4. Metal bushing. 5. Contact plate. 6. Metal rod. 7. Sodium electrode. 8. Glass cartridge. 9. Movable contact. 10. Lead conductors.

there was immersed, almost touching the wall of the test tube, a plate of platinum sheet measuring \(40 \times 66\) mm. The plate was soldered to a brass rod, which was connected to one of the poles of the galvanometer. In the apparatus of the second design [11], the electrode is a platinum crucible (Fig. 1), into which the liquid under investigation is poured. The crucible is lowered into an ebonite socket located in a metal cylinder. One of the galvanometer contacts is brought to the bottom of the crucible. The other contact is connected to a rod passing through a glass tube, which is inserted into

* ) The amount of liquid poured in affected the galvanometer readings, but not very significantly, as specially arranged experiments showed (see the experimental part).

closing the metal cylinder with an ebonite stopper. In experiments with sodium we took (this was done also in the first works) a brass rod with a sharp end, thrust into a plate of metallic sodium 9 mm thick. By means of the sharpened end of a tube (internal diameter 6 mm), a cylinder of the appropriate size is cut from the sodium, which is pressed out, at the very moment the experiment begins, directly into the liquid contained in the crucible. In this way an entirely fresh surface of the metal reacts.

Organic substances were for the most part taken for investigation either without solvent, or in benzene, and sometimes in toluene.

In experiments with carboxylic acids we usually used aqueous or aqueous-acetone solutions, replacing sodium with other metals: magnesium, iron, copper, zinc, tin, lead, etc. These metals were most often taken in the form of rods 4 mm or 8 mm in diameter, and sometimes in the form of cylinders, which were screwed onto a brass rod. Experiments were also made with crucibles of copper, lead, and antimony.

The measuring instrument was a galvanometer with high resistance. We had galvanometer-voltmeters and millivoltmeters with a resistance of up to 2800 ohms. It was necessary to use different instruments, since the measured values of the EMF were very different—from tenths of a millivolt to almost 2 volts.

The experiments were almost always carried out at ordinary temperature (16–20°). The influence of changes in temperature was traced separately for several substances and mixtures. It was different for different substances, but with small changes in temperature, in any case, it was slight [44, 16].

The duration of the experiments varied depending on the motion of the galvanometer needle. Sometimes the deflection of the needle was immediately maximal, and this was followed by a rapid or slow fall. In other cases a gradual increase in the observed value of the EMF was observed, sometimes over the course of 30 minutes, followed likewise by a fall. In some cases, finally, the galvanometer showed a constant value of the EMF for a prolonged time, up to several tens of minutes. Such a difference in the course of the process may be due to a change in the electromotive force itself depending on the composition of the solution (dissolution of the reacting metal in the liquid), polarization, or a change in electrical conductivity.

Aiming above all to cover, by the investigation, as large a number as possible of compounds and mixtures, even if only in a reconnaissance manner, we have not yet been concerned with the details of the processes occurring at the electrodes and in the solution. In our experiments a series of readings was always taken at definite time intervals (for example, 5 sec., 10 sec., 20 sec., 30 sec., 1 min., 2 min., 5 min., 10 min., etc.), and for comparison with the results of other experiments the maximum values were taken.

For investigation by the method described here, the following were taken:

1. Individual organic compounds, namely: a) alcohols and phenols; b) carboxylic acids; c) hydrocarbons with a mobile (metal-replaceable) hydrogen atom; d) amines; e) carbonyl compounds, nitro compounds, nitriles, which can pass into a tautomeric form with mobile hydrogen, as well as those for which such a transition is impossible; in many cases it turned out that our method makes it possible to determine quickly and accurately the structure of the substance being studied.

2. Mixtures of organic substances, where this method makes it possible to detect the formation of complex compounds and to determine the composition of the latter.

3. Substances containing small amounts of impurities; some of these can be very easily and accurately determined, and thus the purity of the product can be controlled.

PART TWO

STUDY OF INDIVIDUAL ORGANIC COMPOUNDS

I. Alcohols and Phenols

Review of Earlier Studies

The activity of hydroxyl hydrogen in alcohols has most often been determined from the rate of reactions in which this hydrogen is split off.

By far the greatest number of studies has been devoted to the esterification of alcohols by acids:

\[ \mathrm{ROH + R'COOH \rightleftharpoons R'COOR + H_2O}. \]

We find a very large number of determinations of the rate of this reaction in N. A. Menshutkin [18, 19, 20]. He heated alcohols with acetic acid and, at definite intervals of time, titrated the unreacted acid; he then calculated the rate constants, the amounts of substances that had reacted after 1 hour, and the state of equilibrium. Menshutkin arrived at the following general conclusions.

For primary alcohols, especially methyl alcohol, the rate and the limit of esterification prove to be the greatest. With equimolecular amounts of the reacting components, 65–70% is esterified at equilibrium. In the course of one hour of heating to 155° with acetic acid, primary alcohols react by approximately 50%. For secondary alcohols, the hourly rate under the same conditions was from 5 to 20%, and the limit of esterification reached only 45–50%. Tertiary alcohols gave a rate of no more than 5%, and a limit of up to 20%. Branching of the chain somewhat lowers the rate, while having little effect on the state of equilibrium. Unsaturated (allyl) and aromatic (benzyl) alcohols proved to be somewhat less active than the corresponding alcohols of the saturated series. Phenols behaved similarly to tertiary alcohols.

Other authors studied, in this respect, a large number of alcohols and the most varied acids. The results, however, did not always agree with what was observed in experiments with acetic acid and carbonic acids close to it; for example, trichloroacetic acid gave the following rate constants (at 25°):

Table 1

Alcohols	\(K\,10^6\)
\(\mathrm{CH_3OH}\)	37
\(\mathrm{CH_3 \cdot CH_2 \cdot CH_2OH}\)	7.25
\(\mathrm{CH_3 \cdot (CH_2)_6 \cdot CH_2OH}\)	18.6
\(\mathrm{CH_3 \cdot CHOH \cdot CH_3}\)	0.98
\(\mathrm{CH_3 \cdot CH_2CHOH \cdot CH_3}\)	0.90
\(\mathrm{(CH_3)_3COH}\)	1.18
\(\mathrm{(CH_3)_2C(OH)-CH_2 \cdot CH_3}\)	2.48

With respect to this acid, tertiary alcohols are more active than secondary ones, although they are greatly inferior to primary alcohols.

As was later clarified in a detailed study of the esterification process, these fluctuating results are explained by differences in the mechanism of the reaction. Many believed that esterification proceeds analogously

neutralization of the acid and the base, i.e., the alcohol in this case loses hydroxyl, while the acid gives up its hydrogen:

\[ \mathrm{KOH + HCl \rightarrow KCl + H_2O}; \]

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

However, if this were so, then the rate of reaction with all acids should have changed in one direction depending on the structure of the alcohol. Meanwhile, Petrenko-Kritchenko P. I., Bogatskii, and Lubman [23] showed that hydrogen chloride acts considerably more readily on tertiary alcohols than on primary ones, while secondary alcohols occupy an intermediate position. The following table compares the relative values of the rate constants of reactions of several aromatic alcohols with acetic, trichloroacetic, and hydrochloric acids, the constant for benzyl alcohol being taken as 100.

Table 2

Alcohols	CH₃COOH	CCl₃COOH	HCl
\(\mathrm{C_6H_5CH_2OH}\)	100	100	100
\(\mathrm{(C_6H_5)_2CHOH}\)	58	15.09	1600
\(\mathrm{(C_6H_5)_3COH}\)	16.4	24.2	7656

With hydrogen chloride, alcohols undoubtedly give up the entire hydroxyl:

\[ \mathrm{CH_3OH + HCl \rightleftarrows CH_3Cl + H_2O} \]

If tertiary alcohols react here faster than all others, this means that in them the hydroxyl is more readily split off. With respect to acetic acid, these alcohols are the least active. Evidently, under the action of \(\mathrm{CH_3COOH}\) and similar acids, some other bond in the alcohol is broken, namely the bond of oxygen with hydrogen (since the hydrocarbon radical of the alcohol remains unaffected).

These facts led [29] to the adoption, for the reaction of alcohols with carboxylic acids, of a different reaction mechanism:

\[ \mathrm{CH_3OH + CH_3C(=O)OH \rightleftarrows CH_3C(OH)(OH)OCH_3 \rightleftarrows CH_3C(=O)OCH_3 + H_2O} \]

i.e., the hydroxyl hydrogen leaves the alcohol; in the acid the double bond in the carbonyl group is broken. Hydrogen adds to the oxygen, and the \(\mathrm{OCH_3}\) group to the carbon of the carbonyl. Further, a molecule of water is eliminated from two hydroxyl groups attached to the same carbon atom.

This explanation of esterification found confirmation in the results of experiments on the cleavage of simple [33] and compound esters [33] by hydrogen halides:

\[ \begin{gathered} \mathrm{CH_3-C(=O)-O-R + HBr \rightarrow CH_3-C(=O)-OH + RBr;}\\ \mathrm{C_6H_5O-R + HBr \rightarrow C_6H_5OH + RBr.} \end{gathered} \]

Everywhere, the bond with oxygen is broken more readily in tertiary radicals and with difficulty in primary ones, i.e., the same bond as in the action of hydrohalic acids on alcohols. Finally, the most direct proof of the carbonyl scheme of esterification was obtained by using the heavy isotope of oxygen with atomic weight 18. Esterification of benzoic acid was carried out with methyl alcohol containing an increased amount of heavy oxygen. Let us write both possible reaction schemes:

\[ \mathrm{ C_6H_5C(=\,^{16}O)-O^{16}H + H^{18}OCH_3 \rightleftharpoons C_6H_5C(=\,^{16}O)-O^{16}CH_3 + H_2^{18}O } \]

\[ \mathrm{ C_6H_5C(=\,^{16}O)-OH + H^{18}OCH_3 \rightleftharpoons C_6H_5C(OH)(O^{18}CH_3)-O^{16}H \rightleftharpoons C_6H_5C(=\,^{18}O)-OCH_3 + H_2^{16}O } \]

According to the old neutralization scheme, water enriched with the heavy isotope of oxygen should have been liberated, which can readily be determined from its specific gravity. In fact, water with a normal isotope content was obtained, i.e., the oxygen of the alcohol remained in the molecule of the ester formed.

Thus, in esterification the alcohol gives up its hydrogen atom, and the rate of this reaction can evidently serve as a criterion of the activity of the hydroxyl hydrogen of alcohols, provided that an acid that is not too strong is taken. Acids with a very large dissociation constant, for example trichloroacetic acid, can react both according to the carbonyl scheme and by their hydrogen (of the HCl type), which explains the peculiar ratios of the reaction rates with different alcohols indicated above.

Results analogous to esterification by acids are given by determination of the rate of reaction of alcohols with acetic anhydride [21, 22]:

\[ \mathrm{ CH_3OH + CH_3C(=O)-O-C(=O)CH_3 \rightarrow CH_3C(OCH_3)(OH)-O-C(=O)CH_3 } \]

\[ \mathrm{ \rightarrow CH_3C(=O)OCH_3 + HO-C(=O)CH_3 } \]

with the ester of benzenesulfonic acid [4]:

\[ \mathrm{CH_3OH + C_2H_5OSO_2C_6H_5 \rightarrow CH_3OC_2H_5 + HOSO_2C_6H_5,} \]

with acid chlorides, where, according to the general equation of the process, it is clear that only hydroxyl hydrogen is evolved:

\[ \mathrm{CH_3OH + Cl{-}C(=O){-}CH_3 \to CH_3O{-}C(=O){-}CH_3 + HCl.} \]

(more precisely, one should write:

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

A major study was carried out by Norris and Ashdown [51] with p-nitrobenzoyl chloride. The results they obtained, for most alcohols, agree well with N. A. Menshutkin’s data. A new feature is the periodicity they observed among aromatic alcohols between the reaction rate and the distance of the hydroxyl from the benzene ring. Phenylethyl alcohol proved more active than benzyl and phenylpropyl alcohols.

Similar rate ratios were found for the reaction with phenyl isocyanate:

\[ \mathrm{ROH + C_6H_5{-}N{=}C{=}O \to C_6H_5{-}NH{-}C(=O){-}OR} \]

If the rate of reaction of methyl alcohol is taken as 1, then for primary alcohols values from 0.7 to 1 are obtained here, for secondary alcohols about 0.3, and for tertiary alcohols 0.01.

Rate of the Reaction of Alcohols with Sodium and the Electromotive Force Arising in This Reaction

One of us, together with L. A. Igonina [5], studied the reaction of methyl, ethyl, isopropyl, and tert-butyl alcohols with sodium, taken in the form of a two-phase amalgam.

Chemically pure and completely anhydrous alcohols were used for the experiments. Thus, ethyl alcohol was distilled with dilute sulfuric acid, boiled with silver nitrate and caustic soda, distilled off again, then boiled twice with pure calcium oxide, and after final distillation was stored in an atmosphere of dry nitrogen. The other alcohols, taken as the purest preparations, were dehydrated by boiling and distilling with calcium oxide.

Kinetic measurements. Measurement of the rate of the heterogeneous reaction between solid sodium amalgam and an alcohol was carried out by the volume of hydrogen evolved in an apparatus whose construction is shown schematically in the drawing (Fig. 2).

The streamlined shape of the reaction vessel promoted uniform evolution of hydrogen. The tubes (m) and (n) served for removal of the evolved gas and for filling the vessel. The mercury seal (b), fixed in the neck of the vessel with a rubber stopper, provided the necessary tightness of the apparatus. The movable part of the seal (C) was fastened to a porcelain rod (S), the end of which was fused into a cylinder of amalgam (a). The other end of the rod was connected to a motor.

The hydrogen evolved was collected in an apparatus connected by rubber tubing with the reaction vessel and with a measuring burette. The measurement consisted of the following manipulations: at the initial moment the levels in the cylinder and the burette were equalized, and a reading was taken, after which the burette

was lowered by several divisions. This produced a slight vacuum in the reaction vessel. The hydrogen evolved now displaced the corresponding volume of water from the cylinder, which caused the meniscus in the burette to rise. After a definite time had elapsed, the levels were again equalized and a new reading was taken. The difference between two readings corresponded to the volume of hydrogen evolved at the given temperature and atmospheric pressure. The change in level in the cylinder could be neglected because of its large volume. Readings on the burette were made with an accuracy of up to 0.05 ml. In some cases a microburette was used.

To determine the influence of the rate of stirring of the liquid on the rate of dissolution, the rate of rotation of the cylinder in specially arranged experiments was varied from 120 to 500 revolutions per minute. It turned out that such a change has no noticeable effect on the rate of dissolution. This circumstance indicates the non-diffusional character of the process. All subsequent measurements were made at a constant rotation speed equal to 120 revolutions per minute.

Fig. 2. Apparatus for measuring the reaction rate of alcohols with sodium amalgam.

a—amalgam cylinder, b—mercury seal, c—movable part of the seal, s—porcelain rod, m and n—tubes, A—vessel for collecting hydrogen.

The apparatus was placed in an oil thermostat. For the experiment, pure alcohols without solvents were taken. Sodium amalgam was prepared by dissolving metallic sodium in mercury. The still-liquid amalgam was poured into a metal mold in which a rod had previously been fixed, serving subsequently as the axis for rotating the cylinder. The piece of amalgam thus prepared was stored under dry xylene. Its sodium content was in all cases 7%, and the total surface area was 15.2 sq. cm. Before the experiment, the surface of the cylinder was cleaned with a piece of glass and then polished with emery paper.

After the apparatus had been assembled and placed in the thermostat, alcohol was introduced into the reaction vessel through tube (m) and the motor was started. At the same time, the first reading of the level in the burette was taken. Subsequent readings were taken every 2 minutes (by stopwatch) for 15–20 minutes. Then the thermostat was set to a new temperature, and readings were again taken. Throughout the experiment the rate of hydrogen evolution remained constant within 0.05 ml. Only isolated readings showed deviations of 0.1–0.15 ml.

The results obtained are presented in Table 3, where the temperatures, the rates expressed in ml of hydrogen per second, and the values of the temperature coefficient of the reaction rate (per 10°) are given.

Table 4 gives the rate values calculated for 25° in millimeters of hydrogen evolved per second, the values of the relative reaction rate (the rate for \( \mathrm{CH_3OH} \) being taken as 100), the values of the activation energy (\(E\)) and the action constant (\(A\)) according to the Arrhenius equation:

\[ K = A \cdot e^{-E/RT} \]

The electromotive force arising in the reaction of ethyl alcohol with sodium amalgam was determined by Denner [49], who

Table 3

Alcohols	\(t^\circ\text{C}\)	\(K_v\) (ml/sec)	\(\dfrac{K_{t+10}}{K_t}\)	Mean temperature coefficient
\(\mathrm{CH_3OH}\)	15.3	0.8		3.41
\(\mathrm{CH_3OH}\)	10.0	1.24	3.44	3.41
\(\mathrm{CH_3OH}\)	14.5	1.95	3.55	3.41
\(\mathrm{CH_3OH}\)	28.7	2.82	3.30	3.41
\(\mathrm{CH_3CH_2OH}\)	13.6	0.25		3.05
\(\mathrm{CH_3CH_2OH}\)	18.0	0.33	3.20	3.05
\(\mathrm{CH_3CH_2OH}\)	22.5	0.44	2.96	3.05
\(\mathrm{CH_3CH_2OH}\)	27.0	0.655		3.05
\(\mathrm{CH_3CHOHCH_3}\)	14.0	0.07		2.84
\(\mathrm{CH_3CHOHCH_3}\)	18.3	0.086	2.85	2.84
\(\mathrm{CH_3CHOHCH_3}\)	23.0	0.113	2.80	2.84
\(\mathrm{CH_3CHOHCH_3}\)	28.0	0.161	2.86	2.84
\((\mathrm{CH_3})_3\mathrm{COH}\)	25.0	0.01		2.44
\((\mathrm{CH_3})_3\mathrm{COH}\)	30.2	0.0124		2.44
\((\mathrm{CH_3})_3\mathrm{COH}\)	35.5	0.0162		2.44

Table 4

Alcohols	\(K_v\) (ml/sec)	Relative reaction rate	\(E\)	\(A\)
\(\mathrm{CH_3OH}\)	2.0	100	24 400	\(6 \cdot 10^{13}\)
\(\mathrm{CH_3CH_2OH}\)	0.56	28	13 500	\(8.5 \cdot 10^{3}\)
\(\mathrm{CH_3CHOHCH_3}\)	0.13	6.5	18 300	\(2.3 \cdot 10^{7}\)
\((\mathrm{CH_3})_3\mathrm{COH}\)	0.01	0.5	15 600	\(3.3 \cdot 10^{2}\)

used the compensation method. A determination of the potential of lithium amalgam in isoamyl alcohol was also carried out [27].

We have studied in this respect the alcohols methyl, isopropyl, and trimethylcarbinol.

The apparatus (Fig. 3), modeled on that used by Danner, consisted of an ordinary hydrogen electrode (a), i.e., a platinum electrode saturated with hydrogen (passed for drying through a tube with phosphorus anhydride and a wash bottle with phosphoric acid), and a dropping amalgam electrode (b), held by a ground joint (c), placed as close as possible to one another. The dropping amalgam electrode (Fig. 3, b) was a small vessel (B), holding about 15 ml of liquid, which was introduced through the opening (o). The space above the ground joint was filled with mercury in order to facilitate heat exchange between the amalgam and the thermostat, in which the entire apparatus was immersed. By means of a platinum wire the amalgam was stirred to ensure equilibrium with the solid phase. A rod (e) made of thick-walled capillary tubing, ground at its lower part, closed the opening of the capillary (f) with a small cone (m) projecting from the side. The latter was a small cup with a small opening in the middle, through which the cup communicated with the capillary.

The amalgam (the rate of its outflow was regulated by a ground-glass rod) filled the capillary and formed on the cup a droplet about 2–3 mm in diameter. The excess amalgam beyond what could fit on the cup overflowed over the edge of the latter and rolled through the narrow opening \((n)\) into the reservoir \((k)\). The opening of the reservoir was closed by a ground-glass rod \((i)\), by means of which the size of this opening could be regulated so that the spent amalgam passed freely into the reservoir without the electrode liquid entering it. This construction ensured complete removal of the spent amalgam. Electrical contact with the droplet was provided by a platinum wire \((p)\), which passed through the end of the rod into its capillary filled with mercury. The end of the conductor from the potentiometer terminal was immersed in the mercury.

Fig. 3. Apparatus for determining the EMF arising in the reaction of alcohols with sodium amalgam.

\(a\)—hydrogen electrode, \(b\)—dropping amalgam electrode, \(c\)—ground joint, \(e\)—rod, \(f\)—capillary, \(m\)—capillary tip, \(k\)—reservoir, \(n\)—opening for amalgam outflow, \(i\)—rod, \(p\)—platinum wire.

Between the electrodes there was alcohol in which a small amount of sodium had previously been dissolved. The initial concentration of alcoholates in the different experiments was from 0.0089 to 0.0870 g-mol per liter.

The entire apparatus was immersed in an oil thermostat with a thermoregulator that made it possible to maintain the temperature within several thousandths of a degree.

In determining the EMF, a normal Weston cell was used as the standard cell. The maximum deviations of three such cells serving as standards did not exceed 0.0001 V. The EMF was read by the high-resistance potentiometric method of the “Etalon” plant with the aid of a mirror galvanometer.

Measurement of EMF. A weighed piece of metallic sodium was added to a weighed quantity of alcohol. After it had dissolved, the liquid was poured into the electrode vessel. The amount of solution was chosen so that the platinum electrode was immersed up to half its height, and so that, when the dropping electrode was set in place, all the air was displaced from its cell. The vessel was immersed in the thermostat and, by means of the ground joint, connected to the hydrogen generator. Hydrogen (obtained by electrolysis of water) was passed for 3–3.5 hours to establish complete equilibrium between the solution and the platinum electrode. Immediately before the measurement, the dropping electrode was put into operation. After the thermostat had been set to a new temperature, hydrogen was again passed for 1.5–2 hours before the measurement.

Results of the determination of \(E\) and calculation of \(E_0\). In recalculating the EMF values found to \(E_0\), i.e., to the potential of the cell at a thermodynamic concentration of the electrolyte equal to unity, a major difficulty was the almost complete absence in the literature of data on the values of activity coefficients in nonaqueous solutions. As Denner has shown, satisfactory results are obtained if equality of the coeff—

coefficients of activity of sodium ethylate and hydrochloric acid in ethyl alcohol. The activity values of hydrogen chloride in alcoholic solution were obtained by this author by measuring the EMF of the cell:

\[ \mathrm{H_2}\,|\,\mathrm{HCl}\,|\,\mathrm{HgCl}\,|\,\mathrm{Hg} \quad \text{in alcohol.} \]

Table 5 gives: the concentrations of alcoholate used in calculating the activity coefficients \((\gamma)\), the found EMF values \((E)\), and the values of \(E_0\) calculated from the equation:

\[ E = E_0 + \frac{RT}{nF}\ln C^2 \gamma^2. \]

All refer to a temperature of 25°. For ethyl alcohol, the results of Denner’s determinations were used. Several experiments performed gave EMF values very close to them.

Table 5

Alcohols	Alcoholate conc.	Activity coeff. \(\gamma\)	\(E\), recalculated to 760 mm \(\mathrm{H_2}\)	\(E_0\), volts
\(\mathrm{CH_3OH}\)	0.0870	0.374	1.180	0.960
\(\mathrm{CH_3OH}\)	0.0629	0.426	1.208	0.963
\(\mathrm{CH_3OH}\)	0.0426	0.475	1.221	0.961
\(\mathrm{CH_3OH}\)	0.0159	0.960	1.255	0.960
\(\mathrm{CH_3OH}\)				Average \(E_0 = 0.961\)
\(\mathrm{CH_3CH_2OH}\)	From Denner’s determinations	From Denner’s determinations	From Denner’s determinations	average \(E_0 = 0.746\)
\(\mathrm{CH_3CHOH.CH_3}\)	0.0521	0.445	0.822	0.570
\(\mathrm{CH_3CHOH.CH_3}\)	0.0248	0.525	0.843	0.553
\(\mathrm{CH_3CHOH.CH_3}\)	0.0135	0.629	0.868	0.565
\(\mathrm{CH_3CHOH.CH_3}\)	0.0115	0.655	0.874	0.526
\(\mathrm{CH_3CHOH.CH_3}\)				Average \(E_0 = 0.566\)
\((\mathrm{CH_3})_3\mathrm{COH}\)	0.0194	0.579	0.780	0.491
\((\mathrm{CH_3})_3\mathrm{COH}\)	0.0137	0.629	0.787	0.484
\((\mathrm{CH_3})_3\mathrm{COH}\)	0.0089	0.687	0.706	0.487
\((\mathrm{CH_3})_3\mathrm{COH}\)				Average \(E_0 = 0.487\)

The table shows that the EMF values of the cell determined by the potentiometric method

\[ \text{sodium amalgam}\, \left|\, \frac{\mathrm{ROH}}{+\mathrm{RONa}} \,\right|\, \mathrm{H_2} \]

vary, depending on the nature of the alcohol, in the same direction as the reaction rate of the same alcohols with sodium, but less sharply. The EMF of methyl alcohol exceeds the value of this quantity for trimethylcarbinol by only a factor of two, whereas in reaction rate \(\mathrm{CH_3OH}\) proves to be 100 times more active with respect to sodium.

Determination of the Activity of Alcohols and Phenols by Our Method

a) Experiments with sodium and platinum

Most of the determinations by our method were made using electrodes of sodium and platinum, in an apparatus of the old design [33] (platinum in the form of sheet metal). The alcohols were taken either without solvent or in benzene solution. The concentration of the solution was selected so that the content of hydroxyls in it was the same as in octyl alcohol, which had the largest molecular volume of all the alcohols we investigated. For this purpose, weighed portions

\[ 1.12\ \text{g } \mathrm{CH_3OH};\quad 1.61\ \text{g } \mathrm{C_2H_5OH};\quad 2.10\ \text{g } \mathrm{C_3H_7OH} \]

and so on were diluted with benzene to 5 ml.

The motion of the galvanometer needle in these series was as follows: with the lower alcohols the deflection of the needle immediately reached its greatest value and then rapidly decreased; with substances of higher molecular weight the whole process proceeded more slowly, and the maximum of the observed electromotive force was sometimes reached only after several minutes.

The results obtained are presented in Table 6, which gives the maximum deflections of the galvanometer needle, expressed in millivolts (as the instrument had been calibrated).

Table 6

Alcohols	EMF of reaction with Na, without solvent	EMF of reaction with Na, in $\mathrm{C_6H_6}$
$\mathrm{CH_3OH}$	1850	850
$\mathrm{CH_3.CH_2OH}$	1250	550
$\mathrm{CH_3.CH_2.CH_2OH}$	1050	120
$\mathrm{CH_3.CHOH.CH_3}$	125	20
$\mathrm{CH_3.CH_2.CH_2.CH_2OH}$	260	60
$\mathrm{CH_3.CH_2.CHOH.CH_3}$	45	2.8
$\mathrm{(CH_3)_2CH.CH_2OH}$	—	30
$\mathrm{(CH_3)_3COH}$	4.8	0
$\mathrm{CH_3.CH_2.CH_2.CH_2.CH_2OH}$	—	45
$\mathrm{(CH_3)_2CH.CH_2CH_2OH}$	100	15.2
$\mathrm{(CH_3)_2COH.CH_2CH_3}$	0	0
$\mathrm{C_8H_{17}CH_2CH_2CH_2CH_2CH_2OH}$	—	20
$\mathrm{(CH_3)_2CH.CH_2CHOHCH_3}$	—	0.2
$\mathrm{CH_3CH_2CH_2CH_2CH_2CH_2CH_2OH}$	—	8.25
$\mathrm{(CH_3)_2CH.CH_2CHOH.CH_2CH_3}$	0.2	0.1
$\mathrm{CH_3.CH_2.CH_2.CH_2.CH_2.CH_2CH_2.CH_2OH}$	—	3.6
$\mathrm{CH_3.CH_2CH_2.CH_2CH_2CH_2CHOHCH_3}$	—	0.12
$\mathrm{CH_2{=}CH.CH_2OH}$	—	160
Cyclohexanol	—	0.35
$\mathrm{C_6H_5CH_2OH}$	—	15.2
$\mathrm{C_6H_5CH_2CH_2OH}$	—	8.3
$\mathrm{C_6H_5CH_2CH_2CH_2OH}$	—	3.85
$\mathrm{C_6H_5CH{=}CH.CH_2OH}$	—	0.50
$\mathrm{C_6H_5OH}$	—	30
$o$-$\mathrm{CH_3C_6H_4OH}$	—	2.8
$m$-$\mathrm{CH_3C_6H_4OH}$	—	8.7
$p$-$\mathrm{CH_3C_6H_4OH}$	—	3.9
$\mathrm{CH_2OH.CH_2OH}$	770	—
$\mathrm{CH_2Cl.CH_2OH}$	1100	—
$\mathrm{CH_3CHOH.COOC_2H_5}$	120	—

b) Experiments with Other Metals

This series of experiments was carried out only with three alcohols: ethyl, normal butyl, and primary isobutyl, but with the most varied pairs of metals. The following were taken: sodium, aluminum, calcium, iron, copper, zinc, antimony, platinum, and lead in various combinations. The metals were taken in the form of crucibles and cylinders 10 mm high and 8 mm in diameter (crucibles: Al, Fe, Cu, Zn, Sb, Pt, Pb; cylinders: Na, Ca, Cu, Zn, Sb).

Owing to the great activity of ethyl alcohol with respect to some metals, the galvanometer needle sometimes went beyond the limits of the scale; therefore experiments with C₂H₅OH were set up using benzene as solvent. Butyl and isobutyl alcohols were used without solvent.

The test substance was poured into the crucible in an amount of 5 ml. The temperature was from 18 to 20°. Each experiment was performed 2–3 times with good agreement. For example, for Fe electrodes (crucible) and Na (cylinder) with a solution of ethyl alcohol at a concentration of 3 g-mol per liter, the following EMF values were obtained (in millivolts).

Table 7

No. of experiment	Beginning of experiment	After 0.25 min	After 0.5 min
1	11.20	18.37	3.14
2	11.22	18.345	3.09
3	11.11	18.35	3.11
Average	11.2	18.2	3.1

In the following tables (8, 9, 10, 11) only the mean values of the EMF maxima are given.

Table 8

C₂H₅OH in C₆H₆, conc. 3 g-mol/liter.

Crucible	Fe	Cu	Cu	Cu	Zn	Zn	Zn
Cylinder	Na	Ca	Zn	Sb	Ca	Cu	Sb
EMF in mV	18.4	2.9	1.0	0.45	2.02	0.45	0.4

Table 9

C₂H₅OH in C₆H₆, conc. 5 g-mol/liter

Crucible	Fe	Fe	Fe	Cu	Cu	Zn	Zn	Zn	Sb	Sb
Cylinder	Zn	Sb	Cu	Ca	Sb	Ca	Sb	Cu	Cu	Zn
EMF in mV	6.0	0.9	0.8	20	2.9	14.6	1.1	2.5	2.5	4.7

Crucible	Pt	Pt	Pt	Pb	Pb	Pb	Al	Al	Al
Cylinder	Cu	Zn	Sb	Cu	Zn	Sb	Cu	Zn	Sb
EMF in mV	0.7	4.0	1.1	2.5	4.7	0.4	2.8	2.7	0.23

Table 10

CH₃CH₂CH₂CH₂OH without solvent

Crucible	Fe	Fe	Fe	Cu	Cu	Zn	Zn	Sb	Sb
Cylinder	Cu	Zn	Sb	Zn	Sb	Cu	Sb	Cu	Zn
EMF, mV	0.72	1.23	0.67	1.68	1.57	1.50	0.45	1.4	0.9

Crucible	Pb	Pb	Pb	Pt	Pt	Pt	Al	Al
Cylinder	Cu	Zn	Sb	Cu	Zn	Sb	Cu	Sb
EMF, mV	1.57	0.36	0.67	0.45	2.13	2.02	1.79	1.12

Table 11

\((\mathrm{CH_3})_2\mathrm{CH}\cdot\mathrm{CH_2OH}\) without solvent

Crucible	Fe	Fe	Fe	Cu	Cu	Zn	Zn	Sb	Sb
Cylinder	Cu	Zn	Sb	Zn	Sb	Cu	Sb	Cu	Zn
EMF, mV	0.64	3.65	1.01	4.5	0.9	1.86	0.52	1.57	2.69
Crucible	Pb	Pb	Pb	Pt	Pt	Pt	Al	Al	Al
Cylinder	Cu	Zn	Sb	Cu	Zn	Sb	Cu	Zn	Sb
EMF, mV	2.35	2.58	0.49	1.12	2.13	1.84	2.69	1.84	1.34

Influence of the solvent, solution concentration, temperature, and electrode surface area

A series of determinations [24] of the EMF with phenol, sodium, and platinum (crucible) was carried out in various solvents, which included aromatic hydrocarbons, their halogen derivatives, simple and complex ethers, nitro compounds, and aliphatic and aromatic amines. The phenol concentration was varied from 1 g-mol per liter to 8 g-mol per liter (from 0.47 to 3.76 g per 5 ml of solution). The temperature was 18–20°. The maximum observed EMF values are given in Table 12.

Table 12

EMF in mV at phenol content per 5 ml

Solvents	0.47	0.94	1.41	1.88	2.35	2.82	3.29	3.76
\(\mathrm{C_6H_6}\)	—	—	—	0.4	1.6	4.2	9.6	18.4
\(\mathrm{C_6H_5CH_3}\)	—	0.1	0.4	1.2	2.8	4.2	11	20
\(\mathrm{C_6H_6(CH_3)_2}\) (mixture)	—	0.2	0.7	1.8	4.4	10	18.4	—
\(1,3,5\text{-}\mathrm{C_6H_3(CH_3)_3}\)	0.1	0.15	0.5	3.6	—	—	—	—
Indene	—	0.1	—	1.3	—	—	—	—
Tetralin	—	0.1	0.6	5.9	—	—	—	—
\(\mathrm{C_6H_5Cl}\)	0.15	4.4	12.8	35	—	—	—	—
\(\mathrm{C_6H_5Br}\)	—	3.8	12.2	20	—	—	—	—
\(o\text{-}\mathrm{C_6H_4Cl_2}\)	1.5	10	28	—	—	—	—	—
\(o\text{-}\mathrm{CH_3C_6H_4Cl}\)	0.2	3.0	11.6	19.8	—	—	—	—
\(m\text{-}\mathrm{CH_3C_6H_4Cl}\)	0.4	4.4	17.1	—	—	—	—	—
\(p\text{-}\mathrm{CH_3C_6H_4Cl}\)	0.7	6.1	21	—	—	—	—	—
\(\mathrm{C_6H_5CH_2Cl}\)	2.1	5.6	11.6	—	—	—	—	—
\(\mathrm{C_6H_5CCl_3}\)	30	—	—	—	—	—	—	—
\(\mathrm{C_6H_5Cl}\)	0.3	1.1	2.1	10.3	—	—	—	—
\(\mathrm{C_6H_5OCH_3}\)	1.7	6.7	13.7	—	—	—	—	—
\(\mathrm{C_6H_5OC_2H_5}\)	1.1	2.3	4.5	8.6	18	—	—	—
\(\mathrm{C_6H_5COOC_2H_5}\)	30	—	—	—	—	—	—	—
\(\mathrm{CH_3COOCH_2C_6H_5}\)	2.7	12.2	30.5	—	—	—	—	—
\(\mathrm{C_6H_5NO_2}\)	30	—	—	—	—	—	—	—
\((\mathrm{C_2H_5})_3\mathrm{N}\)	—	3.8	25.7	—	—	—	—	—
\((\mathrm{C_3H_7})_2\mathrm{NH}\)	0.2	3.3	26.8	—	—	—	—	—
\((\mathrm{C_3H_7})_3\mathrm{N}\)	0.1	2.1	30.1	—	—	—	—	—
\((i\text{-}\mathrm{C_5H_{11}})_2\mathrm{NH}\)	0.15	0.9	9.7	—	—	—	—	—
\((i\text{-}\mathrm{C_5H_{11}})_3\mathrm{N}\)	0.1	2.6	22.8	—	—	—	—	—
\(1,3\text{-}\mathrm{CH_3C_6H_{10}NH_2}\) (cyclo)	5.2	24	30.4	—	—	—	—	—
\(\mathrm{C_6H_5NH_2}\)	30	—	—	—	—	—	—	—
\(\mathrm{C_6H_5CH_2NH_2}\)	—	—	—	—	—	—	—	—
\(\mathrm{C_6H_5N(CH_3)_2}\)	0.15	0.4	2.6	4.1	—	—	—	—

With the solvents benzyl chloride and benzotrichloride, certain anomalies were observed in the course of the process. Thus, with \(\mathrm{C_6H_5CH_2Cl}\), the movement of the galvanometer needle gave two maxima, evidently corresponding to two reactions:

\[ \mathrm{C_6H_5OH + Na \rightarrow C_6H_5ONa + H,} \]

\[ \mathrm{C_6H_5CH_2Cl + Na \rightarrow C_6H_5CH_2 + NaCl,\ etc.} \]

With \( \mathrm{C_6H_5CCl_3} \), a brightly colored (red) solution was obtained. Condensation must have occurred with the formation of a quinoid derivative of triphenylmethane.

As further investigation showed (see the third part), the sharp increase in the observed EMF value in such solvents as nitro compounds, amines, and esters is explained by complex formation. With these substances, phenol gives molecular compounds in which the activity of the hydroxyl hydrogen with respect to metals is considerably greater. In the case of halogen derivatives there is no complex formation, and their influence (especially pronounced in o-dichlorobenzene) must be attributed to a large dipole moment. Such, too, is probably the action of anisole and phenetole. With phenol, the influence of temperature was also traced [16]. With an antimony crucible and a zinc cylinder, the galvanometer (resistance \(2800\ \Omega\)) showed:

Table 13

Temperature	\(85^\circ\)	\(100^\circ\)	\(140^\circ\)	\(163^\circ\)
EMF in mV	0.40	0.40	0.90	0.96

For comparison, we also give the results of experiments with nitrobenzene using the same electrodes:

Table 14

Temperature	\(20^\circ\)	\(30^\circ\)	\(57^\circ\)
EMF in mV	0.22	0.22	0.45

Finally, a mixture (complex-forming) of 60% (molecular) \( \mathrm{C_6H_5NO_2} \) and 40% \( \mathrm{C_6H_5OH} \) showed (the same electrodes):

Table 15a

Temperature	\(17^\circ\)	\(18.5^\circ\)	\(30^\circ\)	\(50^\circ\)	\(70^\circ\)	\(103^\circ\)	\(173^\circ\)
EMF in mV	4.48	4.70	4.93	5.82	6.72	8.29	13.44

These results show that the EMF value observed in our apparatus generally increases with increasing temperature, although without definite quantitative regularities.

The magnitude of the electrode surface and the cross-sectional area of the liquid dependent on it also have a noticeable effect [16]. These experiments were carried out with isopropyl alcohol. The electrodes were an antimony crucible and cylinders of zinc and copper. The galvanometer was the same. Temperature \(18\text{–}20^\circ\).

Table 15b

	2.5	2.6	2.7	2.8	2.9	3.1	3.25	3.5	4.0
Electrodes Sb and Zn; EMF in mV	—	—	—	—	—	14.5	16.04	18.0	20.0
Electrodes Sb and Cu; EMF in mV	7.45	8.06	8.74	9.18	10.02	—	—	—	—

The increase in the observed magnitude of the EMF with an increase in the amount of liquid poured into the apparatus must evidently be explained by a decrease in the resistance of the system.

Comparison of the results of determining the EMF and the reaction rate

Table 16 compares the results of determinations of the activity of hydrogen in alcohols according to the reaction rate with various reagents, according to the EMF of the reaction with sodium amalgam, and according to the current strength in a closed circuit in which the EMF arises owing to the reaction of the alcohol with sodium. The table gives the data of N. A. Menshutkin on the reaction rate of alcohols with acetic acid and with acetic anhydride; of V. Zagrebin, who studied the reaction with benzenesulfonic acid ester; and of Norris and Ashdown on the reaction rate with para-nitrobenzoyl chloride. Everywhere the reaction rate or the EMF for methyl alcohol is taken as 100.

The results of the determinations by N. A. Menshutkin with acetic acid must be regarded as less accurate, since a reversible reaction was studied, and the author took not rate constants but the quantities of substances that had reacted in 1 hour; moreover, these values were sometimes fairly close to the reaction limit. This smoothed out the difference between the individual alcohols.

Table 16

Alcohols	CH₃COOH	CH₃C(=O)–O–C(=O)CH₃	C₆H₅SO₂OC₂H₅	NO₂C₆H₄COCl	Sodium amalgam	EMF of reaction with Na	Current strength in closed circuit with Na and Pt, without solvent	Current strength in closed circuit with Na and Pt, in C₆H₆
CH₃OH	100	100	100	100	100	100	100	100
CH₃CH₂OH	81	48.4	48.9	45.7	35	77	67	65
CH₃CH₂CH₂OH	81	42.9	35.2	35.9	—	—	57	15
CH₃CHOH.CH₃	45	13.2	16.1	5.4	6	59	6.7	2.4
CH₃CH₂CH₂CH₂OH	—	41.6	—	40.2	—	—	—	7
(CH₃)₂CH.CH₂OH	80	35.9	21.4	16.7	—	—	—	3.5
CH₃CH₂CHOH.CH₃	39.5	11.0	—	4.0	—	—	2.4	0.3
(CH₃)₃COH	2.5	0.8	0.8	1.46	1	45	0.25	0
CH₃(CH₂)₃.CH₂OH	—	—	—	—	—	—	—	5.3
(CH₃)₂CH.CH₂CH₂OH	—	—	23.6	—	—	—	13.8	1.8
(CH₃)₂COH.CH₂CH₃	1.4	—	2.6	—	—	—	0	0
CH₃(CH₂)₄.CH₂OH	—	—	—	—	—	—	—	2.4
CH₃(CH₂)₅.CH₂OH	—	35.1	—	—	—	—	—	0.97
CH₃(CH₂)₆.CH₂OH	81	33.7	16.2	—	—	—	0.42	—
C₆H₁₃CHOH.CH₃	32	8.1	4.6	—	—	—	0.014	—
CH₂=CH.CH₂OH	64	25.9	28.1	—	—	—	—	18.8
Cyclohexanol	—	12.5	—	—	—	—	—	0.04
C₆H₅OH	2.5	21.6	—	—	—	—	—	3.5
o-CH₃C₆H₄OH	—	—	—	—	—	—	—	0.33
m-CH₃C₆H₄OH	—	—	—	—	—	—	—	1.0
p-CH₃C₆H₄OH	2.4	—	—	—	—	—	—	0.46
C₆H₅CH₂OH	66	25	15	9.3	—	—	—	0.95
C₆H₅CH₂CH₂OH	—	—	—	21.6	—	—	—	1.4
C₆H₅CH₂CH₂CH₂OH	—	—	—	10.7	—	—	—	0.45

From the figures in this table, and from all the factual material of our work pertaining to alcohols, the following conclusions may be drawn.

1) The rate of action upon alcohols of the most diverse reagents that split off the hydroxyl hydrogen changes, depending on the structure of the alcohol, almost identically. If, for example, one takes the (best studied) alcohols,

having from 1 to 4–5 carbon atoms, then for primary alcohols the rate of reaction with sodium, with acid chlorohydrides and anhydrides, and with benzenesulfonic acid ester proves to be 2–3 times less than for methyl alcohol. In an alcohol with a branched chain the reaction rate is still somewhat lower; in secondary alcohols it is ordinarily 4–5 times lower than in primary ones; in tertiary butyl and amyl alcohols it is 30–100 times lower than in methyl alcohol. As already noted above, similar results were also given by determination of the reaction rate with phenyl isocyanate (from the relative amounts of products in mixtures of alcohols).

2) The free energy of reaction with sodium, determined from the electromotive force, changes in the same direction, but much less markedly, than the rate. For trimethylcarbinol the EMF proved to be only half as great as for methyl alcohol, whereas in reaction rate one alcohol is inferior to the other by a factor of 100.

Of the two quantities of which, according to the Arrhenius equation, the rate constant is a function, one—namely the activation energy—changes without any definite regularity. The action constant decreases sharply from methyl alcohol to the tertiary alcohol, i.e., as the alcohol group becomes surrounded by hydrocarbon radicals (Table 5).

3) The current strength measured by the galvanometer in a closed circuit in which the electrodes are sodium and platinum, and in which the EMF arises as a result of the reaction of sodium with the alcohol, gave, for primary, secondary, and tertiary alcohols, results that agree better with the rate constants than with the magnitudes of the true electromotive force. The difference is noticeable when comparing alcohols with different chain lengths. The EMF values observed on the galvanometer in our apparatus fall sharply with increasing alcohol molecule size. However, the regularity in the ratio of the activities of primary and secondary alcohol groups is retained in octyl alcohols just as in propyl and butyl alcohols.

Alcohols with a branched chain in our work showed appreciably lower activity than their isomers of normal structure, again in complete agreement with what was shown by measuring the rate of reaction with acid anhydrides and chlorohydrides. In aromatic alcohols we observed the same periodicity in the change of activity with the length of the side chain as that noted by Norris and Ashdown in experiments with p-nitrobenzoyl chloride. Thus the current strength in our apparatus, which is a complex function of several physical quantities (true EMF, electrical conductivity), at the same time gives a good characterization of the chemical properties of alcohols. It is also of interest that the constants often used in recent times—the dipole moment and bond energy—in the present case give chemists nothing. For the dipole moments of monohydric alcohols of the limiting series the following values have been found:

\[ \begin{aligned} \mathrm{CH_3OH} &\ldots\ldots\ldots\ldots 1.73\cdot 10^{-18}; & \mathrm{CH_3CH_2OH} &\ldots\ldots\ldots\ldots 1.64\cdot 10^{-18} \\ \mathrm{CH_3CH_2CH_2OH} &\ldots\ldots 1.66\cdot 10^{-18}; & \mathrm{CH_3CH_2CH_2CH_2OH} &\ldots\ldots 1.65\cdot 10^{-18} \\ \mathrm{CH_3CH_2CH_2CH_2CH_2OH} &\ldots 1.65\cdot 10^{-18}; & \mathrm{CH_3CH_2CH_2CH_2CH_2CH_2OH} &\ldots 1.66\cdot 10^{-18} \\ \mathrm{(CH_3)_2CH.CH_2OH} &\ldots 1.65\cdot 10^{-18}; & \mathrm{(CH_3)_2CH.CH_2CH_2OH} &\ldots 1.64\cdot 10^{-18} \\ \mathrm{CH_3CH_2CHOH.CH_3} &\ldots 1.65\cdot 10^{-18}; & \mathrm{(CH_3)_2COH.CH_2CH_3} &\ldots 1.65\cdot 10^{-18} \end{aligned} \]

For the bond energy \( \mathrm{C{-}OH} \), calculated (from ultraviolet absorption spectra) in cal. per gram-molecule:

\[ \begin{gathered} \mathrm{CH_3OH} — 133.6;\qquad \mathrm{C_2H_5OH} — 137.5;\qquad \mathrm{n{-}C_3H_7OH} — 140.9; \\ \mathrm{n{-}C_4H_9OH} — 141.9;\qquad \mathrm{n{-}C_5H_{11}OH} — 142.3;\qquad \mathrm{n{-}C_6H_{13}OH} — 142.7; \\ \mathrm{n{-}C_7H_{15}OH} — 143.4;\qquad \mathrm{n{-}C_8H_{17}OH} — 144.5;\qquad \mathrm{(CH_3)_2CH.CH_2OH} — 141.2; \\ \mathrm{CH_3CH_2CHOH.CH_3} — 140.2;\qquad \mathrm{(CH_3)_3COH} — 141.2. \end{gathered} \]

These quantities relate, however, to the bond of carbon with oxygen, and not of oxygen with hydrogen; it is possible that the dipole moment also has a greater bearing on this same bond. However, in terms of the ease of cleavage of the hydroxyl group (for example, under the action of hydrogen halides or during dehydration), primary, secondary, and tertiary alcohols also differ quite noticeably from one another.

In order, when working by our method, to obtain values of the activity of individual compounds, in particular alcohols, that are comparable with one another, it is evidently necessary to carry out the experiments under identical conditions: at the same concentration, temperature, electrode surface, and the same distance between them. As for the electrode material, as we have seen, sodium and platinum give EMF values that agree well with the results of determining the reaction rate. The use of other pairs of metals, as is evident from the figures given above, gives the same relationships of activity for alcohols that differ sufficiently from one another, for example when comparing ethyl alcohol with butyl alcohol. But for normal and isobutyl alcohols fluctuations are observed. Isobutyl alcohol more often gave a higher EMF than the normal alcohol, which contradicts all known data on reaction rates.

Such fluctuations are probably connected with the fact that the electromotive series of metals with different reagents prove to be different. This has also been noted by other authors. Let us cite, for example, the work of N. V. Kondyrev and V. I. Yur’ev [8], who determined the potentials of certain metals in ethereal solutions of magnesium iodide and obtained a series that differs considerably from the usual electromotive series in aqueous solutions.

According to the readings of the galvanometer in our apparatus, for example, copper, zinc, and antimony in ethyl and butyl alcohols are arranged according to their activity as follows:

\[ \mathrm{Zn} > \mathrm{Cu} > \mathrm{Sb} \]

In a nitrobenzene solution of phenol, with 12.67% of the latter, the following was obtained:

\[ \mathrm{Cu} > \mathrm{Sb} > \mathrm{Zn}, \]

and with 24.65% phenol:

\[ \mathrm{Cu} > \mathrm{Zn} > \mathrm{Sb}. \]

The same different electromotive series (with respect to the metals) may also be obtained for alcohols. For example, with electrodes of copper, zinc, and antimony in a platinum crucible, butyl and isobutyl alcohols gave EMF values in millivolts:

	Cu	Zn	Sb
\(n\text{-}\mathrm{C_4H_9OH}\)	0.45	2.13	2.02
\(i\text{-}\mathrm{C_4H_9OH}\)	1.12	2.13	1.84

It follows that, with respect to copper, isobutyl alcohol is more active; with respect to antimony, normal butyl alcohol is more active; and with zinc they are equal in activity.

The same fluctuations, however, are also found when comparing reaction rates. Methyl alcohol in all substitution reactions of hydroxyl hydrogen proves more active than trimethylcarbinol, but for trimethylcarbinol and phenol the reaction rate with acetic acid is the same, whereas with acetic anhydride phenol reacts almost 30 times faster.

For the present, therefore, one may consider established the general regularities in the dependence of chemical activity on structure (whatever method we use to determine this activity), although in individual cases exceptions to the rules may be observed. When the difference in activity of the substances being compared is small, the general regularity may be obscured by certain individual features, the elucidation of which should be the subject of further investigations.

II. Carboxylic Acids

Experiments without solvents and in benzene solution

Only a few of the carboxylic acids could be studied without the use of solvents or in such a neutral solvent as benzene. In the case of the lower acids—formic and acetic—the experiments with sodium were impeded by the fact that the metal quickly became covered with a layer of salt, and this interrupted the current. The more complex acids are often solids with a rather high melting point, and in benzene they are by no means always readily soluble.

We therefore decided to use here mainly aqueous solutions, as well as aqueous-alcoholic and aqueous-acetone solutions, employing, of course, not sodium but less active metals. Only a very small number of experiments were carried out without a solvent and in benzene, with sodium and platinum as the metals, mainly for the purpose of comparing under these conditions the acidity of carboxylic acids and of the corresponding primary alcohols [34]. The acids taken were propionic, butyric, and isovaleric; the alcohols were propyl, butyl, and isoamyl. The measurements were made at a temperature of 18–20°. The concentration of the acids in benzene was equimolecular to the concentration of the alcohols, i.e. 0.035 gram-mole per liter. Determination of the EMF with the same voltmeter gave the following result (Table 17a).

Table 17a

Investigated compounds	Observed values of EMF, mV: without solvent	Observed values of EMF, mV: in benzene
$\mathrm{CH_3-CH_2-COOH}$	570	20
$\mathrm{CH_3-CH_2-CH_2OH}$	1050	120
$\mathrm{CH_3-CH_2-CH_2-COOH}$	150	16
$\mathrm{CH_3-CH_2-CH_2-CH_2OH}$	260	60
$\mathrm{(CH_3)_2CH-CH_2-COOH}$	70	—
$\mathrm{(CH_3)_2CH-CH_2-CH_2OH}$	100	—

It thus turns out that the hydrogen of the hydroxyl group in the acid itself is less active with respect to the metal than the hydrogen of the hydroxyl in an alcohol with the same carbon skeleton. Strictly speaking, the acid is not the molecule

$$ \mathrm{R-C(=O)-OH} $$

or (according to Hantzsch)

$$ \mathrm{R-C\begin{matrix}O\\[-2pt]O\end{matrix}H}, $$

nor the dimer $\mathrm{(RCOOH)_2}$, but a complex of $\mathrm{RCOOH}$ with water. Our data fully support the hydrate theory of acids and bases and are in complete agreement with the results, cited below, of the study of other complex-forming systems (for example, carboxylic acids with nitro compounds).

Experiments with aqueous solutions (11, 44)

Solubility conditions make water a more suitable medium (than benzene) for studying the reaction of acids with metals. Still better in this respect proved to be mixtures of water with alcohol and especially with acetone.

As active metals in these experiments we used magnesium, aluminum, iron, copper, zinc, cadmium, tin, and lead, with the second electrode being a platinum crucible. The water was distilled twice over permanganate, and its purity was checked by its EMF and by its electrical conductivity. The water we used gave, with zinc, an EMF of 1.5–2 mV and had an electrical conductivity of \(3 \div 3.3 \cdot 10^{-6}\). Further distillations only worsened its quality (owing to leaching from the glass vessels). In the principal series of experiments with acids, the substances under investigation were taken at a concentration of 0.1%. In a water–acetone solution, a 0.1-g portion of the acid was first dissolved in 10 ml of acetone, and then the solution was diluted with water to 100 ml. Five milliliters of liquid were poured into the crucible of the apparatus. The temperature was 15–17°. Separate measurements showed that the dependence of the EMF observed in our apparatus on temperature was slight here. For example, for acetic acid, on going from 8 to 23°, the EMF values changed: with iron, from 48 to 49 mV; with tin, from 5.5 to 6.2 mV; with lead, from 45 to 46 mV. With zinc and copper, and for oxalic acid also with other metals, the fluctuations were still smaller.

In all experiments with acids the maximum EMF was observed immediately upon immersion of the metal in the solution. This was followed by a fairly rapid fall to a certain value, which remained constant for a comparatively long interval of time (tens of minutes). This value (reached after approximately 2 minutes) was the one taken into account in comparing acids. Repeat experiments usually gave almost exactly coinciding values for the majority of the metals taken. Somewhat larger fluctuations were observed for copper and tin. No evolution of hydrogen bubbles and no noticeable rise in temperature upon immersion of the metal in the liquid occurred, even with the strongest carboxylic acids, for example with oxalic acid.

Tables 18 and 19 give the results of experiments with a series of monocarboxylic saturated acids, as well as with several dicarboxylic saturated and unsaturated acids and with phthalic acid.

The tables also give the dissociation constants of the same acids in water at 25°.

Table 17 b

Aqueous solutions

No.	Acids	Mg	Al	Fe	Cu	Zn	Cd	Sn	Pb	\(K \cdot 10^5\)
1	HCOOH	460	45	125	4	250	140	35	125	21.4
2	CH\(_3\)COOH	240	35	75	10	110	90	40	66	1.86
3	CH\(_3\)CH\(_2\)COOH	215	35	45	2	90	85	15	45	1.32
4	CH\(_3\)(CH\(_2\))\(_2\)COOH	175	20	60	7	100	75	16	60	1.48
5	(CH\(_3\))\(_2\)CHCOOH	175	20	50	6	90	75	14	50	1.5
6	CH\(_3\)(CH\(_2\))\(_3\)COOH	175	15	55	2	110	75	10	55	1.6
7	(CH\(_3\))\(_2\)CHCH\(_2\)COOH	165	18	55	5.2	95	55	14	40	1.7
8	(CH\(_3\))\(_3\)CCOOH	180	24	31	1.5	95	70	7.8	25	0.96
9	C\(_7\)H\(_{15}\)COOH	165	10	18	0.75	45	37	7	27	1.41
10	CH\(_3\)(CH\(_2\))\(_8\)COOH	145	11	55.5	1	32.5	56	4.8	14	1.41
11	COOH COOH	810	23	120	19	240	175	135	98	3800
12	COOH (CH\(_2\))\(_2\) COOH	570	50	96	7	145	130	12	75	6.7
13	COOH.CH=CHCOOH cis	700	125	110	2.3	240	190	25	100	1370
14	COOHCH=CHCOOH trans.	640	20	115	20	240	140	7	160	91
15	O—C\(_6\)H\(_4\)(COOH)\(_2\)	500	35	110	10	205	175	1.7	95	126

Table 18

Aqueous–acetone solutions

No.	Acids	Mg	Al	Fe	Cu	Zn	Cd	Sn	Pb	K · 10⁵
1	HCOOH	460	30	80	2	190	140	5	50	21.4
2	CH₃COOH	240	25	48	2	90	85	2	35	1.86
3	CH₃CH₂COOH	210	14	28	2	80	70	4	24	1.32
4	CH₃(CH₂)₂COOH	208	13	23	1	75	70	2	20	1.48
5	(CH₃)₂CHCOOH	200	8	28	2	80	80	5	20	1.5
6	CH₃(CH₂)₃COOH	160	7	30	1	55	50	2	16	1.6
7	(CH₃)₂CHCH₂COOH	130	9	30	3	50	35	4	15	1.7
8	(CH₃)₃CCOOH	145	12	17.5	1	55	48	2.5	14	0.96
9	CH₃(CH₂)₄COOH	80	6	0.5	0.9	9	1.8	1.2	5	1.41
10	CH₃(CH₂)₆COOH	72.5	6.5	0.7	1	7.5	2.2	1.4	6	1.41
11	COOH COOH	530	20	75	4	190	190	95	68	3800
12	COOH(CH₂)₂COOH	240	10	60	2	125	95	8	45	6.7
13	COOH CH=CH COOH cis	682	12	72	9	225	160	2	70	1370
14	COOHCH=CH COOH trans	650	4	75	3	175	160	1	60	91
15	O—C₆H₄(COOH)₂	440	7	95	3	170	115	1	55	125

A large number of acids had been studied only with the metals iron, copper, zinc, tin, and lead. Here many other oxy acids of the saturated and aromatic series were also taken, as well as some acids with three and six carboxyls and several new representatives of the groups already considered. The apparatus was somewhat different, with metal rods 8 mm in diameter. Therefore, in Tables 19 and 20, for those acids that appear in the preceding tables, we find other values of the EMF.

Table 19

No.	Acids	Fe	Cu	Zn	Sn	Pb	K · 10⁵
1	HCOOH	225	20	425	95	150	21.4
2	CH₃COOH	150	20	225	55	105	1.86
3	CH₃CH₂COOH	125	18	200	50	75	1.32
4	CH₃CH₂CH₂COOH	115	13	175	50	65	1.48
5	(CH₃)₂CHCOOH	110	10	175	45	60	1.5
6	CH₃(CH₂)₃COOH	95	8.8	105	25	55	1.6
7	(CH₃)₂CHCH₂COOH	105	12	160	45	55	1.7
8	COOH.COOH	150	25	250	155	105	3800

The results of experiments, even those carried out with aqueous solutions, do not always coincide with what is given by determination of the electrolytic dissociation constant from electrical conductivity and with the electromotive series of the metals used. There is a certain general parallelism, but considerable deviations are also observed. These deviations make the method employed not especially suitable under the given conditions for determining the activity of hydrogen in acids, but at the same time they increase its practical value in determining the structure of organic substances. Of the metals at our disposal, tin deviated most from the general rules; it proved to be extremely sensitive to structural differences in acid molecules. It gives, for example, sharply lowered EMF values when the chain of dicarboxylic acids is branched. Thus, ethylmalonic

Table 20

Aqueous–acetone solutions

No.	Acids	Fe	Cu	Zn	Sn	Pb	K · 10⁵
1	HCOOH	125	6.8	250	50	100	21.4
2	CH₃COOH	60	5.0	115	15	50	1.86
3	CH₃CH₂COOH	55	6.0	100	13.8	45	1.3
4	CH₃CH₂CH₂COOH	50	5.0	95	12.7	45	1.48
5	(CH₃)₂CH·COOH	50	5.6	95	15	45	1.5
6	CH₃·(CH₂)₃·COOH	50	4.8	55	11.9	50	1.6
7	(CH₃)₂CH·CH₂·COOH	55	5.2	95	11.2	40	1.7
8	(CH₃)₃C·COOH	45	2	90	5	45	0.96
9	CH₃·(CH₂)₄·COOH	19	0.9	45	6	27	1.41
10	CH₃·(CH₂)₆·COOH	5.5	0.9	30	4.5	16	1.41
11	C₆H₅COOH	55	5.8	75	16.2	60	6.6
12	o—CH₃·C₆H₄COOH	35	4	55	5.8	50
13	p—(CH₃)₂CH·C₆H₄·COOH	25	4	55	4	35	5.0
14	C₆H₅CH=CH·COOH	50	0.8	100	1.5	22
15	COOH·COOH	250	65	400	255	173	3800
16	COOH·CH₂·COOH	200	40	350	160	145	160
17	COOH·CH(C₂H₅)COOH	110	9.2	250	3.2	100	125
18	COOH·(CH₂)₂·COOH	110	14	180	11	60	6.7
19	COOH·CH(CH₃)CH₂COOH	60	2.9	140	2.8	50	8.6
20	COOH(CH₂)₃·COOH	100	7.8	150	11	50	4.67
21	COOH(CH₂)₆·COOH	70	5.6	105	8.8	45	2.98
22	COOH(CH₂)₇·COOH	55	5.9	75	7.6	40	2.88
23	COOH·CH=CH·COOH cis	170	25	350	5.8	140	1370
24	COOHCH=CH·COOH trans	150	20	306	3.8	110	91
25	CH₃·C·COOH ∥ H·C·COOH	155	20	305	4	110	333
26	CH₃C·COOH ∥ HOOC·C·H	150	14	225	3.5	100	77.5
27	CH₂=C·COOH | H₂C·COOH	105	10.3	200	2.6	75	118
28	o—C₆H₄(COOH)₂	115	8.0	250	2.2	100	126
29	COOH·CH(COOH)CH₂COOH	90	4.5	150	2.0	50	22
30	COOHCH₂C(COOH)=CHCOOH	100	6.8	240	2.2	50	1.6
31	C₆(COOH)₆	150	14.4	330	14	90
32	CH₂OH·COOH	125	10.8	210	45	95	15
33	CH₃CHOH·COOH	100	8.0	150	40	60	13.8
34	(CH₃)₂COH·COOH	80	7.8	150	30	60	10.6
35	o—HO·C₆H₄·COOH	100	4.4	175	2.5	90	106
36	m—HO·C₆H₄·COOH	55	4.0	125	9	55	8.33
37	p—HO·C₆H₄·COOH	40	1.8	80	6.5	50	2.9
38	C₆H₄CHOH·COOH	75	4.6	155	2.8	75	43.8
39	COOH·CHOH·CH₂COOH	100	5.4	190	2.5	60
40	Levorotatory tartaric acid			200			97
41	Mesotartaric acid	105	5.0	200	3.0	60	97
42	Tartaric acid	100	4.6	200	2.0	50	97
43	COOH₂C(OH)CH₂COOH | COOH	100	6.0	175	2.2	40	82
44	C₆H₂(NO₂)₃OH	45	3	110	2	55

acid, in terms of its dissociation constant (K·10⁻⁵ = 125), is somewhat inferior to malonic acid (160); the difference in EMF for these acids is of the same order with the metals Fe (110 and 200 mV), Zn (350 and 250 mV), Pb (145–100 mV); in experiments with copper these acids showed a sharper difference (40 mV and 9.2 mV),

Tin with malonic acid showed a disproportionately large EMF (160 mV), much greater than iron (40 mV) and even greater than lead (145 mV), whereas for ethylmalonic acid, on the contrary, an extremely small value was obtained (3.2 mV). Dicarboxylic acids with a straight chain, beginning with succinic acid, showed EMF values that were not so small, but still even smaller than might have been expected (smaller, for example, than monocarboxylic acids with the same chain length). The double bond in fumaric and especially in maleic acids noticeably increases the dissociation constant in comparison with succinic acid; it also increases the EMF with the metals Mg, Fe, Cu, Zn, Cd, Pb, but lowers it with tin in an aqueous-acetone medium.

Abnormally small EMF values with tin were given by many hydroxy acids, especially salicylic, mandelic, malic, tartaric, and citric acids.

Other metals show their own regularities. Thus, for aluminum it may be noted that maleic acid, both in aqueous and in aqueous-acetone solution, gives larger EMF values (125 mV and 12 mV) than succinic acid (50 mV and 10 mV), while fumaric acid gives smaller values (20 mV and 4 mV), etc. All these features make it possible, with the aid of our method, to identify acids and to determine their structure fairly accurately—for example, to determine the number of carboxyls, alcohol groups, side chains, double bonds, etc. With a larger number of metals, the accuracy of the determinations should increase still further.

We took the acids at an arbitrary concentration of 0.1%. At other concentrations, of course, other EMF values are obtained, but the ratios are generally preserved, as is shown by the results of experiments with oxalic acid (Table 21). Together with the EMF, the electrical conductivity of the solutions was determined.

Table 21

No.	Amount of acid per 100 ml of water	Fe	Cu	Zn	Sn	Pb	Specific electrical conductivity
1	100 mg	250	65	400	255	175	\(2.3 \cdot 10^{-3}\)
2	50 ″	110	10	155	90	85	\(1.15 \cdot 10^{-3}\)
3	25 ″	70	2	105	45	50	\(8.88 \cdot 10^{-4}\)
4	12.5	45	2	80	30	45	\(4.00 \cdot 10^{-4}\)
5	6.25	19	1.5	55	9	19	\(1.25 \cdot 10^{-4}\)
6	3.125	7	1	40	6	13	\(1.00 \cdot 10^{-4}\)
7	1.62	3.5	1	12	2	9	\(8.33 \cdot 10^{-5}\)
8	0.781	2	0	10.2	2	8	\(8.33 \cdot 10^{-5}\)
9	0.39	1.5	1	8	1	6	\(9.43 \cdot 10^{-6}\)
10	0.195	1.5	0	8	1	6	\(9.43 \cdot 10^{-6}\)

Amides of acids from formic to butyric acid were also studied. With them, somewhat smaller EMF values were observed (in aqueous-acetone solution). It is interesting that amides, in contrast to acids, gave a larger EMF with lead than with iron.

III. Other classes of organic compounds

(Hydrocarbons, halogen-organic compounds, esters, aldehydes and ketones, nitro compounds, amines and nitrogenous heterocycles, nitriles)

Literature data

On the question of the electromotive force arising in reactions of organic substances with metals, there are only a few individual articles in the literature dealing with compounds of various classes.

In addition to the above-mentioned article by Denner on the reaction of ethyl alcohol with sodium amalgam, one may also point to the following works, which appeared after the beginning of our investigations:

Brun [48], in 1935, studied a galvanic cell composed of magnesium, platinum, and an ethereal solution of ethyl bromide. The maximum electromotive force of this cell proved to be equal to 2.2 V.

Bent and Keevil [46, 47] dealt with the determination of the EMF of the reaction of sodium, likewise taken in the form of an amalgam, with ethereal solutions of free radicals of the triphenylmethyl type, aromatic ketones, and hydrocarbons capable of adding sodium. Among the hydrocarbons the following were taken:

\[ \text{[[structural formula of a fused tricyclic aromatic hydrocarbon shown in the original]]} \]

\[ \mathrm{C_6H_5CH{=}CHC_6H_5};\qquad (\mathrm{C_6H_5})_2\mathrm{C{=}C}(\mathrm{C_6H_5})_2; \]

from the ketones:

\[ \mathrm{C_6H_5C(=O)C_6H_5};\qquad \mathrm{C_6H_5C_6H_4C(=O)C_6H_5}; \]

\[ \text{[[structural formula of a fused-ring ketone shown in the original]]} \qquad \text{[[structural formula of a tetraphenylcyclopentadienone-type ketone shown in the original]]} \]

The electromotive force was determined by the potentiometric method.

Experiments with individual compounds and in benzene solution

In the first experiments [33] we took representatives of the most diverse classes of organic compounds capable of reacting with metals. The hydrocarbon indene:

\[ \text{[[structural formula of indene shown in the original]]} \]

aldehydes: propionic and benzoic; ketones: acetone, methyl propyl ketone, diethyl ketone, methyl butyl ketone, methyl isobutyl ketone, mesityl oxide, acetophenone, ethyl phenyl ketone, cyclohexanone; esters: ethyl acetate, malonic ester; also esters of hydroxy acids and keto acids, ethyl lactate

and acetoacetic ester; nitrobenzene; amines of the saturated series and aromatic amines: dipropylamine, tripropylamine, aniline; nitrogen heterocycles: pyrrole and pyridine

nitriles: acetonitrile and benzyl cyanide. The substances studied were taken either without solvent or in benzene solution, at the same concentration as the alcohols in our first work, i.e., 7 g-mol per liter (0.035 g-mol in 5 ml), in the same apparatus with sodium and platinum. The temperature was 18–20°. The results are given in Table 22.

Table 22

No.	Compounds studied	Observed EMF values, mV, without solvent	Observed EMF values, mV, in $C_6H_6$
1	Indene	0	—
2	$CH_3CH_2CH{=}O$	110	0.59
3	$C_6H_5CH{=}O$	—	4.10
4	$CH_3CO\cdot CH_3$	390	23
5	$CH_3CO\cdot CH_2CH_2CH_3$	13	7.4
6	$CH_3CH_2CO\cdot CH_2CH_3$	80	9.8
7	$CH_3CO\cdot CH_2CH\cdot (CH_3)_2$	20	—
8	$(CH_3)_2C{=}CH\cdot CO\cdot CH_3$	30	—
9	Cyclohexanone	15.6	2.5
10	$C_6H_5\cdot CO\cdot CH_3$	—	4.8
11	$C_6H_5\cdot CO\cdot CH_2CH_3$	3.2	—
12	$H\cdot COO\cdot C_2H_5$	18	1.8
13	$CH_2(COOC_2H_5)_2$	30	—
14	$CHLi(COOC_2H_5)_2$	2.75	—
15	$CH_3\cdot CHOH\cdot COOC_2H_5$	—	120
16	$CH_3CO\cdot CH_2COOC_2H_5$	—	250
17	$C_6H_5NO_2$	1.4	1.0
18	$(C_3H_7)_2NH$	0	—
19	$(C_3H_7)_3N$	0	—
20	$C_6H_5NH_2$	0.27	0.1
21	Pyrrole	2.7	4.6
22	Pyridine	0.15	0.8
23	$CH_3CN$	470	20
24	$C_6H_5CH_2CN$	30	12

The figures in this table show that the greatest EMF is given by compounds that can pass into the enol or a similar form with especially active hydrogen. Some ketones proved to be even more active than the corresponding secondary alcohols (cyclohexanone and cyclohexanol). It is interesting that acetone gave a larger EMF than the isomeric propionaldehyde. It should also be noted that, with increasing complexity of the molecule, the activity of ketones falls more slowly than that of alcohols.

Acetoacetic ester, in activity, greatly surpasses all the secondary alcohols and the ester of lactic acid. Even in benzene solution it gave an EMF value far exceeding the EMF of malonic ester without solvent. This is entirely consistent with the fact that in acetoacetic ester the eno-

ization and by other signs appears more clearly; for example, it gives a coloration with \(\mathrm{FeCl_3}\), which cannot be observed with malonic ester.

Benzaldehyde and formic ester, in which transition to a form with divalent carbon is possible:

\[ \mathrm{H{-}C\!\left<\begin{matrix}O\\ OC_2H_5\end{matrix}\right.} \;\rightleftarrows\; \mathrm{C\!\left<\begin{matrix}OH\\ OC_2H_5\end{matrix}\right.}, \]

proved to be only slightly active.

A fairly considerable EMF was observed for acetonitrile and benzyl cyanide, while it remained unclear whether here sodium adds to nitrogen or whether hydrogen is replaced in the tautomeric form

\[ \mathrm{R{-}CH{=}C{=}NH}. \]

Indene gave nothing, as did amines of the aliphatic series (which could add the metal to nitrogen). Amines showed only a very small EMF, pyrrole somewhat more, but less than pyridine.

Experiments with the Addition of Sodium Iodide

As is evident from the figures in the preceding table, compounds with a double bond that are incapable of enolization gave rather small EMF values in our apparatus. In order to make the method more suitable for determining the degree of unsaturation of these compounds, we began to seek ways of increasing their activity toward sodium. For this purpose we decided to make use of the following circumstance.

The metal adds to the more electronegative of the atoms between which a double or triple bond is located. If, for example, for benzophenone (a ketone for which transition to the enol form is impossible) one adopts an electronic formula with a semipolar double bond,

\[ \mathrm{C_6H_5\cdot C\!\cdot C_6H_5} \]
\[ \mathrm{\ \ \ \ddot O:} \]

then the positive sodium ion must attach to one of the free (unshared) electron pairs of oxygen. The electron of the metal in this case either passes directly into another part of the molecule, or (as should be the case in our apparatus) goes into the external circuit, giving a current. If the excess positive charge of the carbonyl carbon is in some way weakened beforehand, then this effect should be transmitted to the oxygen, and the free electron pairs of the latter should become more active. Consequently, the positive sodium ion will then be attracted with greater force, and its electron will be more easily driven outward.

Neutralization of the positive carbon charge and activation of oxygen with respect to the metal should be facilitated by the addition of some negative ion. As such we chose the iodide ion, since it is especially capable of complex formation. It gives, for example, the complex ion \(\mathrm{I'_3}\). Owing to the same ability of the \(\mathrm{I'}\) ion to form complexes, iodide salts, for example sodium iodide, dissolve appreciably in many organic solvents: ketones, esters, amines, nitro compounds, and others. Sometimes \(\mathrm{NaI}\) crystallizes with solvents, for example with acetone in the form \(\mathrm{NaI\cdot 3(CH_3)_2C{=}O}\).

In one of the works of our laboratory [32], ternary compounds are described whose composition includes: an iodide salt, a polynitro compound, and a compound with a double or triple bond \(\mathrm{>C{=}O}\); \(\mathrm{>C{=}N{-}}\); \(\mathrm{-C{\equiv}N}\)

(carbonyl compounds, nitriles, etc.). In these ternary complexes, too, iodine must be assumed to be bound to an electropositive carbon.

The EMF values shown by the voltmeter in our experiments could also increase upon addition of sodium iodide because of the increase in electrical conductivity. To what extent, in each given case, there was an actual increase in EMF and to what extent the electrical conductivity had an effect is difficult to decide. However, this circumstance does not diminish—as in the case of other organic substances—the practical value of the method for studying this group of compounds as well.

The experiments [35] were carried out as follows: into the reaction vessel (of the first design, i.e., a glass test tube in which a cylinder of platinum foil was placed) a definite amount of the substance under study was introduced. A small weighed portion of sodium iodide (which almost always first dissolved) was added to the same vessel, the mixture was stirred; benzene was then added to a total volume of 6 ml, and EMF determinations with sodium were made in the same way as in the preceding studies. For comparison, an experiment was also performed under the same conditions without NaI.

Some compounds were studied at different concentrations and with different amounts of sodium iodide. The results are given in Tables 24, 25, and 26, which list: the amounts of the test compound and sodium iodide per 6 ml of solution, the maximum observed EMF values, the time required to reach the maximum, and the coefficients of increase in EMF with NaI.

The same determinations were made with alcohols and acids (Table 23).

Table 23

Alcohols and Carboxylic Acids

No.	Compounds studied	Amount of substance, ml	Time to reach max.	Max. EMF, mV	Amount of NaI, g	Time to reach max.	Max. EMF, mV	Coefficient
1	$\mathrm{CH_3\cdot CH_2\cdot CH_2OH}$	3	$^{1)}$	200	0.01	—	290	1.45
2	$\mathrm{CH_3\cdot CH_2\cdot CH_2OH}$	3	—	200	0.025	—	480	2.4
3	$\mathrm{CH_3\cdot CH_2CH_2OH}$	3	—	200	0.05	—	650	3.25
4	$\mathrm{CH_3\cdot CH_2CH_2OH}$	2.1	—	7.8	0.01	—	50	6.4
5	$\mathrm{CH_3CH_2CH_2OH}$	1.5	—	0.35	0.01	—	2.0	5.7
6	$\mathrm{CH_3CHOH\cdot CH_3}$	3	—	40	0.01	—	150	3.75
7	$\mathrm{CH_3\cdot CHOH\cdot CH_3}$	2.5	—	2.5	0.01	—	40	1.6
8	$\mathrm{CH_3\cdot CHOH\cdot CH_3}$	1.9	—	0.3	0.01	—	3.5	11.7
9	$n\text{-}\mathrm{C_4H_9OH}$	3	—	30	0.01	—	50	1.7
10	$n\text{-}\mathrm{C_4H_9OH}$	2.6	—	4.9	0.01	—	30	6.1
11	$n\text{-}\mathrm{C_4H_9OH}$	1.8	—	1.0	0.01	—	1.2	1.2
12	$\mathrm{(CH_3)_3COH}$	3	—	0	0.01	—	0.3	—
13	$\mathrm{(CH_3)_3COH}$	1.9	—	0	0.01	—	0.05	—
14	$\mathrm{CH_2{=}CH\cdot CH_2OH}$	3	—	310	0.01	—	400	1.3
15	$\mathrm{CH_2{=}CH\cdot CH_2OH}$	2.3	—	85	0.01	—	100	1.2
16	$\mathrm{CH_2{=}CH\cdot CH_2OH}$	1.4	—	1.3	0.01	—	1.7	1.2
17	Cyclohexanol	3	—	0.1	0.01	—	0.1	1.0
18	$\mathrm{C_6H_5\cdot CH_2OH}$	3	1 min	0.7	0.01	—	7.3	10.4
19	$\mathrm{C_6H_5\cdot CH_2OH}$	2.1	1 min	0.15	0.01	—	1.6	10.7
20	$\mathrm{C_6H_5CH{=}CH\cdot CH_2OH}$	3	1 min	0.6	0.01	—	2.4	4
21	$\mathrm{C_6H_5CH{=}CH\cdot CH_2OH}$	2.6	0.5	0.2	0.01	—	1.7	8.5
22	$\mathrm{C_6H_5CH{=}CH\cdot CH_2OH}$	2.1	0.5	0.15	0.01	—	0.4	2.7
23	$\mathrm{CH_3COOH}$	3	30	3.2	0.01	50 min	6.6	2.1
24	$\mathrm{CH_3COOH}$	2.3	35	0.2	0.01	30 min	0.25	2.25
25	$\mathrm{CH_3CH_2COOH}$	2.1	2	0.95	0.01	2 min	2.3	1.15
26	$\mathrm{(CH_3)_2CHCH_2COOH}$	3	0.5	0.1	0.01	1 min	0.2	2.0
27	$\mathrm{(CH_3)_2CH\cdot CH_2COOH}$	2.2	10	0.1	0.01	12 min	0.15	0.5

$^{1)}$ The maximum EMF is observed immediately after immersion of the metal.

Table 24

No.	Compounds studied	Conc., mol/L	Time to max.	Max. EDS, mV	NaI conc., g	Time to max.	Max. EDS, mV	Coefficient
1	\(C_6H_5\cdot O\cdot CH_3\)	3	0.5 min	0.2	0.01	0.5 min	1.4	7
2	\(CH_3\cdot ONO_2\)	1.2	—	0.05	0.01	—	0.05	1
3	\(C_2H_5\cdot ONO_2\)	3	—	0.7	0.01	—	1.2	1.7
4	\(C_3H_7\cdot ONO_2\)	1.6	—	0.05	0.01	—	0.2	4
5	\(C_4H_9\cdot ONO_2\)	1.5	—	0.05	0.01	—	0.05	1
6	\(C_2H_5\cdot ONO_2\)	3	—	0.25	0.01	—	0.35	1.4
7	\(i\)-\(C_5H_{11}\cdot ONO_2\)	3	—	0.3	0.01	—	3.5	11.7
8	\(i\)-\(C_5H_{11}\cdot ONO\)	2.2	—	0.05	0.01	8 min	1.35	27
9	\(CH_3COOCH_3\)	3	—	0.35	0.01	22 min	1.9	5.4
10	\(C_2H_5COOC_2H_5\)	3	0.5 min	0.1	0.01	—	0.7	7
11	\((COOC_2H_5)_2\)	3	—	0	0.01	—	0.1	—
12	\(C_2H_5OOC\cdot CH_2COOC_2H_5\)	3	—	0.1	0.01	—	2.3	23
13	\(CH_3\cdot CHO\)	3	—	30	0.01	—	800	27
14	\(C_2H_5\cdot CHO\)	1.1	38 min	0.4	0.01	—	1.5	3.75
15	\(C_6H_5\cdot CHO\)	3	6 min	2.3	0.01	—	480	280
16	\(C_6H_5\cdot CH=O\)	2	0.5	0.45	0.01	—	110	1
17	\(C_6H_5CH=O\)	1.3	—	0.1	0.01	0.5	50	350

Ketones and ketoesters

Table 25

No.	Compounds studied	Conc., mol/L	Time to max.	Max. EDS, mV	NaI conc., g	Time to max.	Max. EDS, mV	Coefficient
1	\(CH_3CO\cdot CH_3\)	3	—	45	0.005	—	425	9.4
2	\(CH_3CO\cdot CH_3\)	3	—	"	0.01	—	600	13.4
3	\(CH_3CO\cdot CH_3\)	3	—	"	0.025	—	650	14.4
4	\(CH_3CO\cdot CH_3\)	3	—	"	0.05	—	700	17.8
5	\(CH_3CO\cdot CH_3\)	2.91	—	30	0.01	—	390	13
6	\(CH_3CO\cdot C_7H_8\)	2.4	0.5 min	8.2	0.01	—	190	23
7	\(CH_3CO\cdot CH_3\)	1	1 min	0.7	0.01	—	40	57
8	\(CH_3CO\cdot CH_2CH_3\)	3.5	0.5 min	15.8	0.01	—	290	18.3
9	\(CH_3CO\cdot CH_2CH_3\)	2.9	0.5 min	14.4	0.01	—	225	15.6
10	\(CH_3CO\cdot CH_2CH_3\)	2.2	—	9.4	0.01	—	95	10.1
11	\(CH_3CO\cdot CH_2CH_3\)	1.8	2 min	1.8	0.01	—	45	64
12	\(CH_3CO\cdot CH_2CH_3\)	3	2.5	2.8	0.01	—	300	107
13	\(CH_3CO\cdot CH_2CH_2CH_3\)	1.8	1 min	0.3	0.01	—	25	83
14	\(CH_3CH_2COCH_2CH_3\)	3	0.5	3.4	0.01	0.5	25	8
15	\(CH_3CH_2COCH_2CH_3\)	2.1	1 min	0.65	0.01	—	5.9	7
16	\(CH_3CH_2CH_2COCH_2CH_2CH_3\)	3	1.5	0.35	0.01	—	9.5	27
17	\((CH_3)_2C=CH\cdot CO\cdot CH_3\)	3	1.5 min	5.6	0.005	—	460	82
18	\((CH_3)_2C=CH\cdot CO\cdot CH_3\)	2.8	0.5	5.2	0.01	—	695	194
19	\((CH_3)_2C=CH\cdot CO\cdot CH_3\)	2.3	0.5	2.9	0.01	—	490	169
20	\((CH_3)_2C=CH\cdot CO\cdot CH_3\)	1.6	2 min	0.6	0.01	—	110	183
21	Cyclohexanone	3	—	2.7	0.01	—	600	222
22	\(C_6H_5\cdot CO\cdot CH_3\)	3.1	3 min	0.7	0.01	0.5	310	444
23	\(C_6H_5\cdot CO\cdot CH_3\)	3	12 min	1.2	0.01	—	65	60
24	\(C_6H_5\cdot CO\cdot CH_3\)	2.6	20 min	1.3	0.01	—	55	42
25	\(C_6H_5\cdot CO\cdot CH_2CH_3\)	1.9	18 min	0.4	0.01	0.5	4.8	12
26	\(C_6H_5\cdot CO\cdot CH_2C_6H_5\)	3	—	1.1	0.01	—	90	82
27	\(C_6H_5CO\cdot CH_2CH_3\)	2.6	—	0.6	0.01	—	50	83
28	\(C_6H_5CO\cdot C_6H_5\)	3.64	1 min	0.1	0.01	—	25	250
29	\(CH_3CO\cdot CH_2COOC_2H_5\)	3	2.5	25	0.01	—	350	14
30	\(CH_3COCH_2COOC_2H_5\)	2.5	2.5	5.7	0.01	—	175	31
31	\(CH_3COCH_2COOC_2H_5\)	1.4	1.5	2.5	0.01	—	30	12
32	\(CH_3COCH(CH_3)COOC_2H_5\)	3	0.5	6.2	0.01	—	70	11.3
33	\(CH_3COCH(CH_3)COOC_2H_5\)	2.8	0.5	4.2	0.01	—	45	10.7
34	\(CH_3COCH(CH_3)COOC_2H_5\)	2.2	0.5	3.2	0.01	—	30	9.4
35	\(CH_3COCH(CH_3)COOC_2H_5\)	1.6	0.5	0.25	0.01	—	0.85	3.4

Table 26

Nitro compounds, amines, nitrogen heterocycles, nitriles

No.	Compounds studied	Amount of substance, ml	Time to max. increase	Max. EMF, mV	Amount of NaI, g	Time to max. increase	Max. EMF, mV	Coefficient
1	$\mathrm{CH_3NO_2}$	1.1		0.1	0.01		0.1	1
2	$\mathrm{C_6H_5NO_2}$	3		0.3	0.01		1.8	6
3	$\mathrm{C_6H_5NO_2}$	2	0.5 min	0.05	0.01		0.5	10
4	$(\mathrm{C_2H_5})_3\mathrm{N}$	6		0	0.01		0
5	Pyrrole	3		0.25	0.01		0.45	1.8
6	Pyrrole	2		0.05	0.01		0.45	9
7	Pyridine	3		0.9	0.01		85	95
8	Pyridine	2		0.2	0.01		20	100
9	$\alpha$-Picoline	3		0.3	0.01		40	133
10	$\alpha$-Picoline	2.1		0.05	0.01		40	800
11	$\beta$-Picoline	3		0.3	0.01		40	133
12	$\beta$-Picoline	2.1		0.05	0.01		40	800
13	Quinoline	3		0.4	0.01		35	88
14	Quinoline	2.1		0	0.01		3.4
15	$\mathrm{CH_3CN}$	3		500	0.01		650	1.3
16	$\mathrm{CH_3CN}$	1		55	0.01		60	1.1
17	$\mathrm{C_6H_5CN}$	3		3.4	0.01		180	53
18	$\mathrm{C_6H_5CN}$	2		1.1	0.01		75	68
19	$\mathrm{C_6H_5CN}$	1.6		0.7	0.01		35	50
20	$\mathrm{C_6H_5COCN}$	6		12.3	0.01		60	4.9

We see that: 1) In hydroxy compounds—alcohols and acids—the magnitude of the electromotive force indicated by our instrument increases only slightly upon addition of sodium iodide.

2) Aldehydes, ketones, and nitriles for which, by virtue of their structure, rearrangement into a form with active hydrogen is excluded—for example, benzaldehyde, benzophenone, benzonitrile—by themselves give a very small EMF, but its magnitude increases many tens of times, and sometimes hundreds of times, under the influence of NaI. Esters of carboxylic acids (benzoethyl ester) and amyl nitrite, as well as nitrobenzene, gave a small increase in EMF. For nitrates the EMF remained almost unchanged.

3) Carbonyl and other compounds that can pass into the enol or a similar form (in nitro compounds and nitriles), sometimes (for example, mesityl oxide and cyclohexanone) sharply increase their activity with respect to sodium under the influence of sodium iodide; sometimes their EMF with NaI changes almost not at all.

4) Amines of the saturated series, neither by themselves nor in the presence of sodium iodide, give an appreciable EMF. Pyrrole gave a small increase in EMF; with pyridine, picolines, and quinoline an increase of approximately 100-fold was observed.

Thus, by introducing a definite modification into our method—namely, by adding sodium iodide—it is possible to extend the investigation also to atomic groups with double and triple bonds. Like alcoholic hydroxyls, the carbonyl, nitro group, and others behave differently depending on the other parts of the molecule bound to them.

At the same time, insofar as can be judged by comparing the data presented here with those available in the literature, the results of determining activity prove similar when the most varied methods are applied. Thus, the EMF values found in our instrument indicate a greater activity of the carbonyl in aldehydes and ketones than in esters; and we come to the same conclusion when determining the strength of the complexes formed by these compounds.

It may be thought that the proposed method can be further improved if, instead of sodium iodide, other complexing agents are taken—for example, iodide salts of organic ammonium bases, which are still more soluble in many organic solvents. It is also possible to determine the activity of the complexes of the carbonyl compounds under study, and of related compounds, with phenols or carboxylic acids (on these complexes see below), etc.

PART THREE

STUDY OF COMPLEX-FORMING SYSTEMS IN ORGANIC CHEMISTRY

I. Methods for Studying Complex-Forming Systems

Complex Compounds in Organic Chemistry

From the structural point of view, complex or molecular compounds (the former name is more appropriate, since it is more general: not only different molecules, but also parts of one and the same molecule may be joined by a complex bond) include such compounds in which certain atoms are bound not by principal valences, but by so-called secondary valences. The electronic theory refines the concepts of valence and complex bonding. In the formation of a valence bond, the electrons of both atoms joining one another take part. A complex bond is formed through the attraction of the electrons of one atom to the positive nucleus of another:

\[ \mathrm{ \begin{array}{c} \mathrm{H}\\[-2pt] \mathrm{H{:}N{:}}\\[-2pt] \mathrm{H} \end{array} + H{:}\ddot{Cl} \;-\; \left[ \begin{array}{c} \mathrm{H}\\[-2pt] \mathrm{H{:}N{:}H}\\[-2pt] \mathrm{H} \end{array} \right]^+ \left[ {:}\ddot{Cl}{:} \right]^- } \]

In inorganic chemistry this type of compound has long been known. It includes numerous crystallohydrates, ammoniates, complex and double salts, etc. Subsequently there were added to them similarly constructed complexes of mineral substances with organic substances similar to \(\mathrm{H_2O}\) and \(\mathrm{NH_3}\), for example, salts with alcohol of crystallization, ammoniates in which ammonia is completely or partially replaced by amines:

\[ \mathrm{CaCl_2 \cdot 4C_2H_5OH,} \qquad \left[ \begin{array}{c} \begin{array}{ccccc} & \mathrm{CH_2} & & \mathrm{CH_2} & \\ / & & \backslash & / & \backslash \\ \mathrm{CH_2} & & \mathrm{NH_2} & \mathrm{H_2N} & \mathrm{CH_2} \\ \backslash & & / & \backslash & / \\ & \mathrm{NH_2} & \cdots \mathrm{Co} \cdots & \mathrm{NH_2} & \\ & & / & \backslash & & \\ & \mathrm{NH_2} & & \mathrm{NH_2} & \\ & | & & | & \\ & \mathrm{CH_2} & - & \mathrm{CH_2} & \end{array} \end{array} \right]\mathrm{Cl_3} \]

By the present time, such mixed mineral-organic complexes have been obtained for almost all classes of organic compounds, even for unsaturated and aromatic hydrocarbons, for example:

\[ \mathrm{FeCl_2 \cdot 2C_2H_4 \cdot H_2O;\quad PtCl_2 \cdot C_2H_4;\quad \left[\,\mathrm{Pt}\begin{matrix}\mathrm{C_2H_4}\\ \mathrm{Cl}\end{matrix}\right]\mathrm{Me};} \]

\[ \mathrm{C_6H_6 \cdot 3SbCl_3;\quad Ni(CN)_2 \cdot NH_3 \cdot C_6H_6.} \]

Less studied are complexes whose molecules are composed only of organic substances. The best-known groups of such substances are salts of amines with organic acids, the group of picrates, and phenoquinones:

\[ \begin{array}{ccc} \mathrm{C_{10}H_8 \cdot C_3H_2(NO_2)_3OH;} & \mathrm{C_6H_4O_2 \cdot 2C_6H_5OH;} & \mathrm{C_6H_4O_2 \cdot C_6H_4(OH)_2.} \\ \text{naphthalene picrate} & \text{phenoquinone} & \text{quinhydrone} \end{array} \]

Among compounds with internal complex bonds, the copper salts of amino acids and a number of other complexes that contain polyvalent metals have long been studied:

[[unclear: two displayed chemical structural diagrams of chelate complexes containing Cu and Ni]]

Among natural organic substances, these include, for example, chlorophyll and hemoglobin, etc.

In a number of cases, the stability of the hydrates of chloral, oxalic acid, and the like must be explained by the closure of internal complex bonds:

\[ \begin{array}{cc} \mathrm{CCl_3} & \mathrm{C(OH)_3} \\ \vert & \vert \\ \mathrm{HC(OH)_2} & \mathrm{C(OH)_3} \end{array} \]

Complex bonds, both external and internal, are of great importance in the structure of proteins, complex polysaccharides, and many other very important organic substances with very complex molecules. Complex formation is undoubtedly an intermediate stage in a large number of syntheses (such as the Friedel–Crafts reaction, halogenation, nitration, esterification, etc.). Complexes are formed in such important technological processes as dyeing. There can be no doubt that complex bonds also arise first of all under the action of many drugs, poisons, etc., on living organisms, and so forth.

In general, it may be stated with certainty that in organic chemistry complex compounds are no less widespread and have no less significance than in inorganic chemistry. This has recently compelled chemists to pay great attention to the study of organic complexes and to the development of methods for investigating them.

Methods for Investigating Organic Complex Compounds

The most reliable method for investigating complex-forming systems is the isolation of the compounds formed in these systems. This is easiest to accomplish in those cases where the molecular compound differs sharply from its constituent substances in solubility.

...in some solvent. Thus, aqueous solutions of quinone and hydroquinone give the very sparingly soluble quinhydrone; when alcoholic solutions of naphthalene and picric acid are mixed, crystals of naphthalene picrate precipitate; aqueous solutions of right- and left-tartaric acids give the complex racemate—racemic acid—which dissolves much less readily than either of the optical antipodes.

Such complexes can readily be purified by crystallization, analyzed, and in general studied as ordinary valence compounds.

However, most molecular compounds are too unstable for them to be isolated and purified. Sometimes the results of molecular-weight determinations—for example, by vapor density (dimers of carboxylic acids) or by the cryoscopic method—definitely indicate the formation of complexes; but for the most part, in order to recognize such compounds and determine their composition, one must be content with various methods of physicochemical analysis. The substances capable of forming a complex are mixed in different ratios, and for these mixtures the values of some constant are determined—for example, specific gravity, refractive index, viscosity, surface tension, etc. If the substances taken do not react with one another, the measured constant changes with the composition of the mixture according to a definite law, most often additively, sometimes in a more complex way. For example, on the melting diagram a depression (eutectic) is obtained. Deviations from the regularities established for chemically inert substances with respect to one another indicate chemical interaction between the components of the system under study. Of the physical constants, melting temperature and electrical conductivity, viscosity, and certain others are used for this purpose. Sometimes good results are given by observation of the color of the compounds and measurement of the absorption spectrum. Chemical activity, measured, for example, by the rate constant of one reaction or another, may also indicate the formation of complexes. Thus, by determining the rate of nitration of nitrobenzene by nitric acid at different quantitative ratios between them, we \[32\] found two compounds:

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

Application of Our Method

The current strength determined in our apparatus in a closed circuit, in which the source of energy is the reaction of the compound under study with the metal, changes sharply, as we have seen, depending on structural changes in the molecule. This led us to test the given method on a variety of complex-forming systems of organic chemistry. The immediate occasion for such a test was provided by the comparison of alcohols, phenols, and carboxylic acids.

The EMF values observed with a voltmeter for the lower alcohols of the limiting series were much greater than for phenols, and even somewhat greater than for acids with the same carbon skeleton. Meanwhile, it is well known that, in the strength of their acidic properties, methyl, ethyl, and other alcohols are inferior to phenols. The latter give phenolates with alkalis, which are decomposed by water only to a comparatively small extent, whereas methylates and ethylates under the same conditions are hydrolyzed almost completely. As for acetic, propionic, and other acids, in electrical conductivity, in their action on indicators, in reactions with alkalis and with carbonates, and finally in sour taste, they are true, although not especially strong, acids.

Such a diversity in the readings given by different methods for determining acidity could be explained only by differences in the conditions under which the experiments were carried out. Phenol or a carboxylic acid show stronger acidic properties in water or alcohol, i.e., in solvents that readily form complexes. Without a solvent, or in benzene, which does not form complex compounds with them, alcohols prove to be the stronger acids.

This explanation is also confirmed by the fact that, for example, acetic acid by itself or in benzene has very low electrical conductivity.

We therefore decided to investigate mixtures of phenols and carboxylic acids, and also, for comparison, alcohols, with such compounds as are known for their ability to form complexes with substances of an acidic character. Such complex-forming agents include:

a) Amines, which give ammonium salts with acids, as well as more complex molecular compounds, for example,

\[ \mathrm{C_6H_5NH_2 \cdot 2CH_3 \cdot COOH.} \]

b) Nitro compounds—nitrobenzene, for example, adds sulfuric acid, giving a crystalline product of the composition:

\[ \mathrm{C_6H_5NO_2 \cdot H_2SO_4.} \]

By the kinetic method we detected the existence of two compounds of \(\mathrm{C_6H_5NO_2}\) with nitric acid (see above).

c) Carbonyl compounds, which are likewise capable of adding mineral acids:

\[ \begin{matrix} \mathrm{CH_3}\\ \mathrm{C{=}O \cdots HX}\\ \mathrm{CH_3} \end{matrix} \]

d) Simple ethers, for which the formation of oxonium compounds was first observed, for example:

\[ \begin{matrix} \mathrm{CH_3} && \mathrm{CH_3}\\ | && |\\ \mathrm{O} + \mathrm{HCl} \longrightarrow \mathrm{O}\cdots\mathrm{HCl}\\ | && |\\ \mathrm{CH_3} && \mathrm{CH_3} \end{matrix} \]

etc. We have so far touched only upon carbonyl compounds, nitro compounds, and amines.

The investigation was carried out according to the general method of physicochemical analysis. In studying each system of interest to us, a successive series of mixtures was taken, differing from one another in composition by 10 molecular percent of one or the other constituent. These mixtures, in most cases with the further addition of the solvent benzene or toluene, were poured into the reaction vessel, and several readings were taken on the voltmeter until a noticeable decrease in the EMF indicated by it occurred.

The experiments were almost always carried out at ordinary temperature (16–20°). For some systems (phenol–nitrobenzene, phenol–benzophenone) the influence of temperature was followed over wide limits (up to 173°).

II. Complex-forming systems that include alcohols and phenols

Alcohols with nitro compounds and amines [36]

Of the alcohols of the saturated series, the following were taken for experiments on complex formation: primary alcohols—propyl and butyl; secondary—isopropyl; tertiary—trimethylcarbinol and amylene hydrate; in addition, the secondary cyclic alcohol cyclohexanol and the aromatic β-phenylethyl alcohol. These alcohols were studied in mixtures with nitrobenzene, triethylamine, aniline, dimethylaniline, and pyridine, using the sodium–platinum pair of metals. The apparatus was of the first design. The experiments were carried out in benzene solution. The sum of the molecular concentrations of both complex-forming agents was 8.33 g-mol per liter. For each determination, 0.05 g-mol was taken and benzene was added to 6 ml. The results are given in Tables 27 and 28.

Table 27

No.	Alcohol	Percent alcohol	Percent of other complex-former	$C_6H_5NO_2$	$(C_2H_5)_3N$	$C_6H_5NH_2$	$C_6H_5N(CH_3)_2$	$C_5H_5N$
1	$CH_3CH_2CH_2OH$	20	80	—	—	—	—	20
1	$CH_3CH_2CH_2OH$	50	50	—	4.6	6.4	10	85
1	$CH_3CH_2CH_2OH$	60	40	—	30	50	40	85
1	$CH_3CH_2CH_2OH$	70	30	—	125	120	150	45
1	$CH_3CH_2CH_2OH$	80	20	—	225	300	325	280
1	$CH_3CH_2CH_2OH$	100	0	—	350	350	350	350
2	$CH_3CH_2CH_2CH_2OH$	20	80	—	—	—	—	20
2	$CH_3CH_2CH_2CH_2OH$	50	50	—	1.1	5.3	3.2	85
2	$CH_3CH_2CH_2CH_2OH$	70	30	—	7.7	20	20	85
2	$CH_3CH_2CH_2CH_2OH$	80	20	—	35	65	55	—
2	$CH_3CH_2CH_2CH_2OH$	90	10	—	65	125	100	130
2	$CH_3CH_2CH_2CH_2OH$	100	0	—	145	145	145	145
3	$CH_3CHOHCH_3$	20	80	—	—	—	—	3.9
3	$CH_3CHOHCH_3$	40	60	—	—	—	—	40
3	$CH_3CHOHCH_3$	50	50	—	—	2.2	2.5	55
3	$CH_3CHOHCH_3$	60	40	—	—	—	—	50
3	$CH_3CHOHCH_3$	70	30	—	3.3	7.2	15.7	11.4
3	$CH_3CHOHCH_3$	80	20	—	2.1	25	25	9.6
3	$CH_3CHOHCH_3$	90	10	—	10.3	40	60	—
3	$CH_3CHOHCH_3$	100	0	—	85	85	85	85

Thus, the primary alcohols of the saturated series—propyl and butyl—showed almost no increase in EMF in our apparatus either with nitrobenzene or with amines. Methyl alcohol, not included in the table, behaved in the same way. For isopropyl alcohol, a maximum was observed in experiments with pyridine. The aromatic phenylethyl alcohol showed the same thing (it was not tested with other amines or with nitrobenzene). Cyclohexanol showed an increase in EMF with pyridine and with nitrobenzene. Tertiary butyl alcohol gave very noticeable maxima with nitrobenzene and with pyridine, and small ones with aniline and dimethylaniline (with

Table 28

Nos.	Percent alcohol	Percent of other complex-forming substance	\(C_6H_5NO_2\)	\((C_2H_5)_3N\)	\(C_6H_5NH_2\)	\(C_6H_5N(CH_3)_2\)	\(C_5H_5N\)
3			\((CH_3)_3COH\)	\((CH_3)_3COH\)	\((CH_3)_3COH\)	\((CH_3)_3COH\)	\((CH_3)_3COH\)
3	0	100	0.9	0	0	0	3.5
3	10	90	3.4	0	—	—	3.0
3	20	80	4.3	0	—	—	6.1
3	30	70	4.0	0	—	—	2.7
3	40	60	3.8	0	—	—	2.6
3	50	50	2.4	0	0.3	0.25	3.0
3	60	40	—	0	0.35	—	3.3
3	70	30	—	0	0.35	0.35	—
3	80	20	—	—	0.2	0.4	2.0
3	90	10	—	0.1	0.2	0.45	—
3	100	0	0.35	0.35	0.35	0.35	0.35
4			\((CH_3)_2COH \cdot CH_2 \cdot CH_3\)	\((CH_3)_2COH \cdot CH_2 \cdot CH_3\)	\((CH_3)_2COH \cdot CH_2 \cdot CH_3\)	\((CH_3)_2COH \cdot CH_2 \cdot CH_3\)	\((CH_3)_2COH \cdot CH_2 \cdot CH_3\)
4	50	50	—	—	—	—	\(1.6^1\)
4	70	30	—	—	—	—	—
4	100	0	—	—	—	—	0
5			Cyclohexanol	Cyclohexanol	Cyclohexanol	Cyclohexanol	Cyclohexanol
5	30	70	3.6	—	—	—	5.5
5	40	60	3.1	—	—	—	5.8
5	70	30	1.5	—	—	—	—
5	100	0	0.4	—	—	—	—
6			\(C_6H_5 \cdot CH_2 \cdot CH_2OH\)	\(C_6H_5 \cdot CH_2 \cdot CH_2OH\)	\(C_6H_5 \cdot CH_2 \cdot CH_2OH\)	\(C_6H_5 \cdot CH_2 \cdot CH_2OH\)	\(C_6H_5 \cdot CH_2 \cdot CH_2OH\)
6	20	80	—	—	—	—	12.8
6	30	70	—	—	—	—	90
6	40	60	—	—	—	—	55
6	50	50	—	—	—	—	49
6	70	30	—	—	—	—	12.1
6	100	0	—	—	—	—	65

\(C_6H_5NH_2\): after this maximum there follows a decrease, then again an increase for the alcohol itself, and it gave no maximum with triethylamine. Tertiary amyl alcohol, finally, for which only an experiment with pyridine was made, also showed under these conditions a noticeable increase in EMF. From this it can definitely be seen that tertiary alcohols form complex compounds with nitrobenzene and pyridine. With less certainty this can be said of the alcohol–aromatic amine systems. Primary alcohols show no signs of complex formation. Secondary alcohols occupy an intermediate position. Phenylethyl alcohol gives a complex with pyridine.

Regarding the mechanism of formation here of molecular compounds, the following considerations may be expressed.

Alcohols, if there are no other active parts in their molecules, must form complexes either through oxygen or through hydroxyl hydrogen. According to the electronic formula

\[ \mathrm{R:\ddot O:H} \]

this means that the substance combining with them must be attracted either by one of its atomic nuclei to the free pair of electrons of oxygen, or by some unshared pair of electrons to the hydrogen nucleus.

\(^1\) Pyridine itself at the same concentration, i.e., 0.025 g-mol per liter, gave 0.1 mV.

Other compounds entering into the composition of complexes can be divided into two groups:

a) Some of them, for example amines of the fatty series and ethers:

\[ \begin{array}{c} \ \ \ R\\[-2mm] \ \ \vdots\\[-1mm] R:N:\ ;\ R:O:R\\[-1mm] \ \ \vdots\\[-2mm] \ \ \ R \end{array} \]

have only one complex-forming site—the pair of electrons at N—or two such pairs at O.

b) Nitro compounds, carbonyl compounds, heterocycles of the pyridine type, etc.

\[ \begin{array}{ccc} \ddot{O}: & \ddot{O}: & \begin{array}{c} H\\[-1mm] \vdots\\[-1mm] H:C:C:C:H\\[-1mm] \ \ \ \vdots\ \ \vdots\\[-1mm] H:C:N:C:H \end{array}\\[-1mm] R:N:O & R:C:R & \end{array} \]

can add complexly both through the free electron pairs of oxygen or nitrogen and through the positive nuclei N or C via the free corners of the tetrahedra.

Consequently, fatty amines and ethers can give complexes only by adding to hydroxyl hydrogen; compounds of the second group can add both to hydrogen and to oxygen. As for alcohols, in primary alcohols the hydroxyl-group hydrogen is more active and, one must think, more capable of forming a complex bond. In tertiary alcohols, on the contrary, the hydrogen atom is held more firmly; hence the electrons of its bond with oxygen are little neutralized by the latter. The free electron pairs of oxygen must likewise be comparatively little neutralized, and therefore complex compounds should be formed precisely at the expense of these electrons.

How the formation of complexes involving hydroxyl hydrogen should affect the activity with respect to sodium can be judged from the sharp increase in EMF for carboxylic acids in mixtures with amines, and especially with triethylamine (see below), where ammonium salts should be obtained:

\[ \begin{array}{c} R\\[-1mm] \vdots\\[-1mm] R:N:H:O:C:R\\[-1mm] \vdots\ \ \ \ \ \vdots\\[-1mm] R\ \ \ \ \ \ :O: \end{array} \]

And in such a case we should expect the greatest increase in EMF not for tertiary, but for primary alcohols, and with triethylamine rather than with other amines taken.

Since in both cases we observed precisely the opposite results, it is obvious that alcohols (not phenols, which will be discussed below) form complexes not through hydrogen, but through oxygen; reacting—

compounds with alcohol are attracted to this oxygen by the most active of the positive nuclei of their molecule, according to the schemes:

\[ \begin{array}{ccccc} & \ddot{\mathrm O}: & & & \ddot{\mathrm O}:\\ & \ddot{} & & & \ddot{}\\ \mathrm R:\mathrm N:\ddot{\mathrm O}: & & ; & & \mathrm R:\mathrm C:\mathrm R\\ \mathrm R':\ddot{\mathrm O}: & & & & \mathrm R':\ddot{\mathrm O}:\\ \mathrm H & & & & \mathrm H \end{array} \]

The possibility is not excluded that two molecules of the substance reacting with it are added to the alcohol, or that the complex is further complicated through association of one of the complex-forming substances, i.e., by the formation of complex bonds among its molecules themselves.

Alcohols and amines of the benzene series may also participate in the formation of complexes by means of their benzene rings.

Complexes of phenols with other organic compounds

Of the phenols, we studied \( \mathrm{C_6H_5OH} \) in the greatest detail. For it, a series of experiments was carried out in various solvents (see above). Those solvents that especially strongly increased the electromotive force shown by our apparatus were then tested as complex-forming agents, toluene being taken as the solvent. Several other analogous compounds were also added to the list of these complex-forming agents.

This series of experiments [24] was carried out in apparatus of design No. 2 with a platinum crucible and a movable sodium electrode. The measurements were made with a galvanometer having a resistance of 2336 ohms, system C V (from the experimental workshops of Leningrad University); the concentrations of the reacting substances were chosen with regard to convenience of determination, so that, on the one hand, a noticeable deflection of the pointer was obtained, and, on the other, so that, as far as possible, it did not go off the scale.

The temperature was 16–20°. It should be said that with phenols in general, even without a solvent, when the observed EMF reached a value of the order of 1 volt, no noticeable self-heating occurred. In this respect phenols differ sharply from acids, for which the reaction with sodium was sometimes accompanied by ignition.

The data obtained are given in Table 28, where for each compound tested for its ability to form complexes with phenol, as well as for phenol itself, the highest concentrations used in the experiments (corresponding to 100%) are indicated.

No formation of complexes with phenol was observed for the following compounds: ortho-dichlorobenzene, although in a solution of \( o\text{-}\mathrm{C_6H_4Cl_2} \) phenol showed a considerable increase in EMF; this, however, must evidently be explained by the electrical influence of a solvent with a large dipole moment.

Ortho-nitrophenol. Probably in it the nitro group is too firmly bound (complexed) with its hydroxyl:

\[ \begin{array}{c} \mathrm{CH}\\[-2pt] \begin{array}{ccccc} & & & & \\ \mathrm{HC} & \diagup & & \backslash & \mathrm{C}-\mathrm{O}-\mathrm{H}\\ & & & & \vdots\\ \mathrm{HC} & \diagdown & & \diagup & \mathrm{C}-\mathrm{N}=\mathrm{O}\\ & & & & \|\\ & & \mathrm{CH} & & \mathrm{O} \end{array} \end{array} \]

Therefore the formation of a bond with a molecule of \( \mathrm{C_6H_5OH} \) becomes impossible.

Aniline. Here no complex formation was detected; because the experiments were carried out in solution, and \( \mathrm{C_6H_5NH_2} \), as an amine with weak basic properties, gives with phenol an unstable compound which readily decomposes in solutions (see below the experiments without solvents).

Table 28

No.	Compound tested	Amount, g or ml, of complex-former	0	10	20	30	40	50	60	70	80	90	100
1	\( \mathrm{C_6H_5\cdot CH{=}O} \)	1	0		0.15		1.5	1.8	2.6		0.9		0.1
2	\( \mathrm{C_6H_5\cdot CO\cdot C_6H_5} \)	2	0	14	23.4	3.5	19.8	18.2	23.7	10	3.6	0.9	0.1
3	\( \mathrm{CH_3COO\cdot CH_2\cdot C_6H_5} \)	4	0	0	0.1	0.35	0.7	1.8	4.0	3.8	1.6	1.4	0.6
4	\( \mathrm{C_6H_5\cdot COO\cdot C_2H_5} \)	4	0	0.3	0.7	1.1	4.5	9.6	13	11	4.2	2.4	0.6
5	\( \mathrm{C_6H_5\cdot NO_2} \)	2	0.17	4.0	19.7	20.7	21.3	15.6	11.2	2.7	1.5	0.4	0.1
6	\( \alpha\text{-}\mathrm{C_{10}H_7NO_2} \)	2	0.17	1.9	14.5	24.6	16.6	12.4	8.1	3.8	2.2	0.2	0.1
7	\( m\text{-}\mathrm{C_6H_4(NO_2)_2} \)	2	0.16	0.23	0.26	0.5	0.24	0.3	0.6	0.74	0.6	0.6	0.1
8	\( m\text{-}\mathrm{ClC_6H_4NO_2} \)	2	0.3	0.4	0.1	0.9	0.6	0.55	0.3	0.3	0.2	0.1	0.1
9	\( p\text{-}\mathrm{CH_3OC_6H_4NO_2} \)	2	0.3	5.5	5.2	30	30	78	17	11.6	2.4	0.6	0.1
10	\( \mathrm{(C_3H_7)_2NH} \)	4	0	0	0	0.1	0.3	0.7	2.4	15	28	15.2	0.6
11	\( \mathrm{(C_3H_7)_3N} \)	4	0	0	0.1	0.15	0.5	1.8	9.7	25	44	34	0.6
12	\( \mathrm{(iC_5H_{11})_2NH} \)	4	0								10		0.6
13	\( \mathrm{(i\text{-}C_5H_{11})_3N} \)	4	0								36.5		0.6
14	Piperidine	4	0	0	0.15	0.6	1.5	14.3	36.5	39.5	34.9	31.6	0.6
15	Pyridine	4	0	16.4	31.9	35	38.3	33.4	21	6.8	3.2	1.6	0.6

Observed EMF values in mV at the molecular percentage of phenol in the mixtures.

In addition to the large series of experiments described, we made a number of determinations of EMF under other conditions and with other complex-forming agents. Other measuring instruments were also used.

a) Experiments with \( \mathrm{C_6H_5OH} \) without solvent [12].

With pyridine at the metals Na and Pt, the following was obtained (Table 29):

Table 29

No.	Amount of phenol, g	Amount of pyridine, ml	Molecular percent \( \mathrm{C_6H_5OH} \)	Molecular percent \( \mathrm{C_5H_5N} \)	Maximum observed EMF, mV
1	0.23	5	3.9	96.1	355
2	0.35	5	5.3	94.7	495
3	0.46	5	7.3	92.7	585
4	0.58	5	9.0	91	610
5	0.69	5	10.5	89.5	615
6	0.92	5	13.6	86.4	500
7	1.38	5	18	82	300
8	1.87	5	24.5	75.5	565
9	2.35	5	28.6	71.4	1000
10	2.88	5	34.0	66.0	170
11	3.29	5	35.9	64.1	155
12	3.76	5	39	61	115

b) Experiments with \( \mathrm{C_6H_5OH} \) and various amines in benzene solution [12].

Benzene was added to mixtures of phenol with amines to a total volume of 6 ml.

c) Experiments with other phenols [12].

In addition to \( \mathrm{C_6H_5OH} \), we studied all three isomeric cresols in mixtures with pyridine, and meta-cresol also with triethylamine; \(\alpha\)- and \(\beta\)-naphthols with pyridine; and one experiment each was carried out with the naphthols in nitrobenzene. Benzene was added to the amounts of phenols and amines indicated in the table to 6 ml.

Table 30

No.	Amine	Weighed amount of \(C_6H_5OH\), g	Amount of amine, ml	Molecular percent of phenol	Molecular percent of amine	Maximum observed EMF, mV
1	Pyridine	0.235	3.8	5	95	110
2	Pyridine	0.47	3.6	10	90	255
3	Pyridine	0.94	3.2	20	80	260
4	Pyridine	1.175	3.0	25	75	255
5	Pyridine	1.41	2.8	30	70	350
6	Pyridine	1.88	2.4	40	60	340
7	Pyridine	2.35	2.0	50	50	210
8	Pyridine	3.29	1.2	70	30	105
9	Pyridine	4.7	0	100	0	90
10	Pyridine	0.16	1.17	10	90	0.4
11	Pyridine	0.32	1.04	20	80	0.6
12	Pyridine	0.40	0.97	25	75	1.2
13	Pyridine	0.48	0.91	30	70	1.3
14	Pyridine	0.64	0.78	40	60	3.6
15	Pyridine	0.80	0.65	50	50	6.8
16	Pyridine	0.96	0.52	60	40	6.7
17	Pyridine	1.12	0.39	70	30	1.3
18	Pyridine	1.60	0	100	0	0.1
19	\(C_6H_5NH_2\)	1.41	3.2	30	70	2.6
20	\(C_6H_5NH_2\)	2.35	2.3	50	50	5.2
21	\(C_6H_5NH_2\)	3.29	1.4	70	30	11.4
22	\(C_6H_5NH_2\)	4.23	0.5	90	10	40
23	\(C_6H_5NH_2\)	4.7	0	100	0	90
24	\(C_6H_5N(CH_3)_2\)	0.94	5.1	20	80	0.5
25	\(C_6H_5N(CH_3)_2\)	2.35	3.1	50	50	1.6
26	\(C_6H_5N(CH_3)_2\)	3.29	1.9	70	30	5
27	\(C_6H_5N(CH_3)_2\)	3.76	1.3	80	20	13.8
28	\(C_6H_5N(CH_3)_2\)	4.23	0.6	90	10	40

Table 31

No.	Phenols studied	Amine	Amount of phenol	Amount of amine	Molecules of phenol, %	Percent amine	Maximum observed EMF for cresol, mV: \(o\)	Maximum observed EMF for cresol, mV: \(m\)	Maximum observed EMF for cresol, mV: \(p\)
1	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	0.54	3.6	10	90	155	195	175
2	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	1.08	3.2	20	80	100	210	315
3	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	1.35	3.0	25	75	215	310	250
4	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	1.62	2.8	30	70	200	320	225
5	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	2.16	2.4	40	60		300
6	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	2.72	2.0	50	50			105
7	\(CH_3 \cdot C_6H_4OH\)	\(C_5H_5N\)	5.4	0	100	0	11.2	40	15.6
8	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	1 ml	5.6	20	80		0
9	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	2.1	4.2	40	60		5.2
10	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	2.6	3.5	50	50		4.4
11	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	3.1	2.8	60	40		100
12	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	3.7	2.1	70	30		400
13	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	4.2	1.4	80	20		850
14	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	4.7	0.7	90	10		1000
15	\(CH_3 \cdot C_6H_4OH\)	\((C_2H_5)_3N\)	5.0	0.3	95	5		500

mV for naphthols:

No.	Phenols studied	Amine	Amount of phenol	Amount of amine	Molecules of phenol, %	Percent amine	\(\alpha\)	\(\beta\)
16	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.12	1.23	5	95	0.1
17	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.24	1.17	10	90	2.2	0.6
18	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.36	1.10	15	85		1.3
19	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.48	1.04	20	80	1.4	1.6
20	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.6	0.97	25	75	0.8	6.0
21	\(C_{10}H_7OH\)	\(C_5H_5N\)	0.72	0.91	30	70	0.6	2.10

Various pairs of metals were used with nitrobenzene (Table 33).

Table 32

No.	Substance and metals investigated	0	10	20	30	40	50	60	70	80	90	100
1	\(C_6H_5OH +\) \(+ C_6H_5CO\cdot C_6H_5\) Cu—Zn			4.8	6.3	5.8	6.8	7.6	7.2	6.3		\(14^{1)}\)
2	\(C_6H_5OH +\) \(+ CH_3COOC_5H_{11}\) Sb—Al	0	0.2	0.4	1.0	1.4	2.3		3.8	4.5	7.8
3	\(C_6H_5OH +\) \(+ CH_3\cdot CH_2COO\cdot C_2H_5\)	0	0.5	1.1	2.0	3.1	4.0	5.6	3.6	3.8	5	\(1.8^{2)}\)
4	\(C_6H_5OH^{3)} +\) \(C_6H_5NO_2\) Cu—Zn	20.7	400	635	897	1173	1035	799	621	552	483
5	\(C_6H_5OH +\) \(+ C_6H_5NH_2\) Cu—Zn	0.36	0.39	0.5	0.9	1.3	2.4	2.9	3.8	4.4	2.8

Observed EMF values in mV for mixtures with molecular percent of phenol.

In nitrobenzene and pyridine solution, with a charge of 0.72 g per 5 ml of solvent, naphthols gave:

\(\alpha\)-naphthol in \(C_6H_5NO_2\)—1200 mV, in \(C_5H_5N\)—685 mV,
\(\beta\)-naphthol in \(C_6H_5NO_2\)—1350 mV, in \(C_5H_5N\)—710 mV.

г) Application of other metals [15].

With the application of other metals, the following systems containing phenol were studied (Table 32):

1) phenol—benzophenone,
2) phenol—isoamyl acetate,
3) phenol—ethyl propionate,
4) phenol—nitrobenzene,
5) phenol—aniline.

The experiments were carried out without solvent. In this connection we were interested, among other things, in the question whether it might not be possible in this way to detect the existence of such complex compounds which are decomposed by solvents; for example, whether phenol and aniline would give a complex.

For some of the systems listed we used the pair of metals zinc (cylinder)—copper (crucible), and for others zinc (cylinder)—antimony (crucible). Finally, for two mixtures of phenol with nitrobenzene, various pairs of metals were used (Table 33).

Table 33

Composition of mixture	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)	24.67% \(C_6H_5OH\) + 87.33% \(C_6H_5NO_2\)
Metals: crucible / cylinder	Al / Cu	Al / Zn	Al / Sb	Fe / Cu	Fe / Zn	Fe / Sb	Zn / Cu	Zn / Sb	Sb / Cu	Pb / Cu	Pb / Zn	Pb / Sb
Observed EMF value	7.8	2.9	3.6	1.6	2	2.5	1.7	1.8	1.3	13.6	3.7	10.6

Composition of mixture	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)	24.65% \(C_6H_5OH\) + 75.35% \(C_6H_5NO_2\)
Crucible / cylinder	Al / Cu	Al / Zn	Al / Sb	Fe / Cu	Fe / Zn	Fe / Sb	Zn / Cu	Sn / Sb	Sb / Cu	Sb / Zn	Pb / Zn	Pb / Sb
EMF, mV	8.5	3.1	2.2	1.9	5.2	4.3	3.8	1.1	2	4.4	4.7	5.4

\(^{1)}\) at \(73^\circ\)
\(^{2)}\) at \(45^\circ\)
\(^{3)}\) a different galvanometer; therefore these data can be compared only with one another.

Thus, phenols give complexes, detectable by our method, with all the compounds examined that have sufficiently active trivalent nitrogen or a double bond of carbon or nitrogen with oxygen. As will be shown below, the maximum EMF values observed in our apparatus correspond (how exactly must be clarified by further investigations) to the composition of the complex compounds. Consequently, from the results of experiments with phenols the following complexes can be detected:

1. \(\mathrm{C_6H_5CH{=}O \cdot C_6H_5OH}\) or \(\mathrm{C_6H_5CH{=}O \cdot 2C_6H_5OH}\)
2. \(\mathrm{C_6H_5OH \cdot 2C_6H_5COC_6H_5}\)
3. \(\mathrm{C_6H_5COC_6H_5 \cdot 2C_6H_5OH}\)
4. \(\mathrm{CH_3COOCH_2C_6H_5 \cdot 2C_6H_5OH}\)
5. \(\mathrm{C_6H_5COOC_2H_5 \cdot 2C_6H_5OH}\)
6. \(\mathrm{C_6H_5OH \cdot 2C_6H_5NO_2}\)
7. \(\mathrm{C_6H_5OH \cdot 2\alpha\text{-}C_{10}H_7NO_2}\)
8. \(\mathrm{C_6H_5OH \cdot 2m\text{-}C_6H_4(NO_2)_2}\)
9. \(\mathrm{m\text{-}C_6H_4(NO_2)_2 \cdot 2C_6H_5OH}\)
10. \(\mathrm{C_6H_5OH \cdot 2ClC_6H_4NO_2}\)
11. \(\mathrm{C_6H_5OH \cdot 2p\text{-}CH_3OC_6H_4NO_2}\)
12. \(\mathrm{(C_2H_5)_2NH \cdot 4C_6H_5OH}\)
13. \(\mathrm{(C_2H_5)_3N \cdot 4C_6H_5OH}\)
14. \(\mathrm{(i\text{-}C_5H_{11})_2NH \cdot 4C_6H_5OH}\)
15. \(\mathrm{(i\text{-}C_5H_{11})_3N \cdot 4C_6H_5OH}\)
16. \(\mathrm{C_5H_{11}N \cdot 2C_6H_5OH}\) (with piperidine)
17. \(\mathrm{C_6H_5NH_2 \cdot 4C_6H_5OH}\)
18. \(\mathrm{C_6H_5OH \cdot 2C_5H_5N}\) (pyridine)
19. \(\mathrm{C_6H_5OH \cdot 4C_5H_5N}\) ¹)
20. \(\mathrm{C_6H_5OH \cdot C_5H_5N}\) ²)
21. \(\mathrm{o\text{-}CH_3C_6H_4OH \cdot 3C_5H_5N}\)
22. \(\mathrm{m\text{-}CH_3C_6H_4OH \cdot 2C_5H_5N}\)
23. \(\mathrm{p\text{-}CH_3C_6H_4OH \cdot 4C_5H_5N}\)
24. \(\mathrm{\beta\text{-}C_{10}H_7OH \cdot 3C_5H_5N}\)

In two cases, namely in the systems meta-cresol–triethylamine and \(\alpha\)-naphthol–pyridine, the EMF maxima fell at a mixture composition at which it is difficult to admit the existence of a definite compound: 10 molecular percent of one component and 90% of the other. These cases deserve additional study. Probably, we are dealing here with some complications connected with the physical complexity of the phenomena occurring in our apparatus.

On the other hand, a fact that is very interesting and permits the developed method to be used with great confidence is that experiments carried out under quite different conditions—in solvent and without solvent, using sodium and platinum or copper and zinc—give identical results in determining the composition of complex compounds. Thus, for the phenol–nitrobenzene composition, the maximum EMF value was observed, independently of the experimental conditions, for a mixture of 60 molecular percent \(\mathrm{C_6H_5NO_2}\) and 40% \(\mathrm{C_6H_5OH}\). The phenol–benzophenone system gave two maxima, and in all cases at the same points:

\[ \mathrm{C_6H_5OH \cdot 2C_6H_5COC_6H_5} \quad \text{and} \quad \mathrm{C_6H_5COC_6H_5 \cdot 2C_6H_5OH} \]

The only difference was that in benzene solution and with sodium and platinum metals the first maximum was greater; without solvent, and

¹) EMF maximum, but only slightly noticeable.
²) The maximum was observed in only one of four experiments with pyridine, whereas three series of determinations made under different conditions indicate the composition \(\mathrm{C_6H_5OH \cdot 2C_5H_5N}\).

with the metals zinc and copper—the second, but this latter circumstance cannot be considered strange. It is to be explained, one must think, by the fact that the voltage series of metals in reactions with different substances are different. This was observed both in experiments with individual compounds (for example, with alcohols) and in experiments with complex-forming systems, and even with mixtures of the same complexing agents, but in different ratios. As is seen from Table 33, when zinc and antimony rods are immersed in a liquid of composition 12.67 mol.% \(C_6H_5OH\) and 87.33% \(C_6H_5NO_2\), contained in an aluminum or iron crucible, antimony gives the greater EMF; whereas in a liquid of composition 24.65% \(C_6H_5OH\) and 75.35% \(C_6H_5NO_2\), we obtain a larger EMF value for zinc than for antimony (at the same time, with a lead crucible, in both cases the larger EMF was observed with antimony).

As for the structure of the complexes found, in general they must be composed according to the type of ammonium salts, in which phenol reacts by its hydroxyl hydrogen, i.e., as an acid, while in the other component the complexing agent is a base. This is supported above all by the similarity in the behavior of phenols and carboxylic acids. Both of these combine with amines, nitro compounds, and carbonyl compounds according to the general scheme:

\[ R_3N \ldots HX;\qquad R-N=O \ldots HX;\qquad R-C=O \ldots HX \]
\[ \phantom{R_3N \ldots HX;\qquad} \begin{array}{c} \| \\ O \end{array} \qquad\qquad\qquad \begin{array}{c} | \\ R \end{array} \]

This explanation of the reaction is also confirmed by the fact that precisely amines with the strongest basic properties—for example, amines of the aliphatic series—give, with phenols, more stable molecular compounds, whereas aniline and dimethylaniline, whose basic properties are weakened by the benzene ring, gave no EMF maximum at all with phenol in benzene solution; the formation of a complex of phenol with aniline could be detected only in an experiment without solvent.

Many of the complex compounds that we have discovered have a more complicated composition. For one molecule of one of the complexing agents they contain two, three, and even four molecules of the other. In this connection certain regularities may be noted. Thus, one molecule of a carbonyl compound (aldehyde, ketone, ester) most often corresponds to two molecules of phenol (exception—the second complex of phenol with benzophenone, \(C_6H_5OH \cdot 2C_6H_5CO \cdot C_6H_5\)). Nitro compounds with one nitro group form complexes of the type \(C_6H_5OH \cdot 2RNO_2\). Aliphatic amines and aniline add 4 molecules of phenol. Pyridine differs sharply from them in complex formation with phenol, cresols, and naphthols, giving compounds of the composition:

\[ ROH \cdot 2C_5H_5N,\qquad ROH \cdot 3C_5H_5N,\qquad ROH \cdot 4C_5H_5N. \]

Concerning the structure of all these complex complexes, certain assumptions may be made. It may be thought, for example, that the carbonyl oxygen can attach two hydroxyl hydrogens (by two unshared electron pairs):

\[ R_2C \begin{matrix} \nearrow & HOC_6H_5 \\ \searrow & HOC_6H_5 \end{matrix} \]

or that phenol forms a dimeric molecule, which enters into the composition of the complex as a whole.

Phenol can attach one nitro group by means of its hydrogen. A second molecule of the nitro compound may be attracted either to one of the hydrocarbon units of \(C_6H_5OH\), or (more probably) to the nitrogen of the first \(RNO_2\) molecule, owing to the strengthening of its positive charge:

\[ \begin{array}{c} \text{[[structural formula: phenol–nitro-compound complex with two aromatic rings and }O^-\,N^+\,O^-\text{ groups]]} \end{array} \]

Amines should give salt-like compounds with one molecule of phenol. For other molecules of the latter, attachment to the hydrocarbon units of the first molecule in the ortho and para positions is most probable, where, under the influence of the nitrogen atom, the negative charges are intensified:

\[ \begin{array}{c} \text{[[structural formula: amine complex with phenol molecules attached at ortho and para positions]]} \end{array} \]

At complete saturation, precisely \(R_3H \cdot 4C_6H_5OH\) was obtained. Pyridine reacts differently, one may suppose, because it also acts through its hydrocarbon units, which can, for example, attract new molecules of the same pyridine. The considerations expressed here are as yet preliminary in character and require verification on more extensive experimental material.

III. Complexes of Carboxylic Acids

Acids with Nitro Compounds

As was mentioned above, the literature data and the results of our experiments showed that nitro compounds, for example nitrobenzene, can attach strong mineral acids, giving

\[ C_6H_5NO_2 \cdot H_2SO_4 \]

\[ C_6H_5NO_2 \cdot HNO_3 \]

\[ C_6H_5NO_2 \cdot 2HNO_3. \]

The products formed may have a valence or a complex structure, but in any case they must dissociate strongly electrolytically:

\[ \mathrm{C_6H_5N{=}O + HOSO_2OH \rightarrow C_6H_5{-}N\!\begin{matrix} /OH \\ \| \\ \backslash OSO_2OH \end{matrix} \rightleftarrows \left[ \mathrm{C_6H_5N{-}OH} \right]^{\cdot} + \mathrm{HSO_3}^{\prime}} \]

\[ \mathrm{\qquad\qquad\quad \|O} \]

and it may be:

\[ \left[\mathrm{C_6H_5NOH}\right]^{\cdot\cdot} + \mathrm{SO_4}^{\prime\prime} \]

\[ \mathrm{\qquad\quad |} \]

\[ \mathrm{\qquad\quad OH} \]

In the case of complex addition:

\[ \mathrm{C_6H_5N{=}O \ldots HOSO_2HO} \]

\[ \mathrm{\quad \|O} \]

and so on, to the same ions

\[ \left[\mathrm{C_6H_5N{-}OH}\right]^{\cdot} \text{ and } \mathrm{HSO_3}^{\prime} \quad \text{or} \quad \left[\mathrm{C_6H_5N{-}OH}\right]^{\cdot\cdot} \text{ and } \mathrm{SO_4}^{\prime\prime}. \]

\[ \mathrm{\qquad\|O \qquad\qquad\qquad\qquad \|O} \]

With nitric acid we obtain:

\[ \mathrm{C_6H_5N{-}OH \rightleftarrows \left[C_6H_5N{-}OH\right]^{\cdot} + NO_3^{\prime}} \]

\[ \mathrm{\qquad \|O \quad \backslash ONO_2 \qquad\qquad \|O} \]

or

\[ \mathrm{C_6H_5N{=}O \ldots HONO_2 \rightleftarrows \left[C_6H_5N{-}OH\right]^{\cdot} + NO_3^{\prime}} \]

\[ \mathrm{\qquad \|O \qquad\qquad\qquad\qquad \|O} \]

and

\[ \mathrm{C_6H_5N\!\begin{matrix} /ONO_2 \\ {-}OH \\ \backslash ONO_2 \\ |OH \end{matrix} \rightleftarrows \left[C_6H_5N\!\begin{matrix} /OH \\ \backslash OH \end{matrix}\right]^{\cdot\cdot} + 2NO_3^{\prime}} \]

or

\[ \mathrm{C_6H_5N{=}O \ldots HONO_2 \rightleftarrows \left[C_6H_5N(OH)_2\right]^{\cdot\cdot} + 2NO_3^{\prime}.} \]

\[ \mathrm{\qquad \|O \ldots HONO_2} \]

With carboxylic acids the same process of addition and electrolytic dissociation is possible; moreover, here the difference in the degree of dissociation between the acid and its complex with the nitroso compound must be very great, and this should lead to a strong increase in electrical conductivity and in the electromotive force arising in the reaction with metals, i.e., to a large deflection of the galvanometer needle in our apparatus.

We carried out experiments [37] with acetic, propionic, butyric, isobutyric, and isovaleric acids, and with three nitroso compounds of the benzene series: nitrobenzene, α-nitronaphthalene, and metadinitrobenzene. The apparatus was of the first design: the metals were sodium and platinum; a 3-volt voltmeter and a millivoltmeter with a resistance of 374 ohms. The experiments were conducted at a temperature of 16–20°, and almost always it was necessary to dilute the complex-forming mixture with benzene, since without a solvent there was observed

strong self-heating. For nitronaphthalene and dinitrobenzene, the poor solubility of these nitro compounds made it necessary, even with benzene solutions, to confine ourselves to a far from complete series of mixtures (beginning with 60–70% acid). The sum of the amounts of the reacting compounds in each experiment was 0.025 g-mol. The mixture was diluted with benzene to 6 ml. The results of this series of experiments are given in Table 34.

Table 34

Compound	Molecular percent: \( \mathrm{C_6H_5-NO_2} \)	Molecular percent: acid	Maximum observed EMF values, mV, with acid \( \mathrm{C_2H_4O_2} \)	Maximum observed EMF values, mV, with acid \( \mathrm{C_3H_6O_2} \)	Maximum observed EMF values, mV, with acid \( \mathrm{C_4H_8O_2} \)	Maximum observed EMF values, mV, with acid iso-\(\mathrm{C_4H_8O_2}\)	Maximum observed EMF values, mV, with acid \( \mathrm{C_5H_{10}O_2} \)
\( \mathrm{C_6H_5-NO_2} \)	90	10	1.15				25
\( \mathrm{C_6H_5-NO_2} \)	80	20
\( \mathrm{C_6H_5-NO_2} \)	70	30					40
\( \mathrm{C_6H_5-NO_2} \)	60	40					40
\( \mathrm{C_6H_5-NO_2} \)	50	50	0.4	3.6	5.4	50	90
\( \mathrm{C_6H_5-NO_2} \)	40	60	0.7	8.4	11.9	5.5	65
\( \mathrm{C_6H_5-NO_2} \)	30	70	0.2	7.6	7.0	5.5	55
\( \mathrm{C_6H_5-NO_2} \)	20	80		2.7	6.7	35	48
\( \mathrm{C_6H_5-NO_2} \)	10	90				0.4
\( \mathrm{C_6H_5-NO_2} \)	0	100	0	1.0	0.8		0.13
\(\alpha\)-\( \mathrm{C_{10}H_7NO_2} \)	40	60	—	12.3	7.1	—	35
\(\alpha\)-\( \mathrm{C_{10}H_7NO_2} \)	30	70	—	16.2	11.9	35	25
\(\alpha\)-\( \mathrm{C_{10}H_7NO_2} \)	20	80	—	9.0	8.2	—	10
\(m\)-\( \mathrm{C_6H_4(NO_2)_2} \)	30	70	12.7	6.0	—	—	1.8
\(m\)-\( \mathrm{C_6H_4(NO_2)_2} \)	20	80	11.8	4.15	42	3.3	1.9
\(m\)-\( \mathrm{C_6H_4(NO_2)_2} \)	10	90	10.3	—	—	—	—

For comparison, several determinations were made with acetic acid in benzene (Table 35).

Table 35

Amount of \( \mathrm{CH_3COOH} \), ml	0.29	1.45	2.03	2.32	2.61	2.9
Max. observed EMF, mV	0	0	0.2	0.8	1.25	3.2

With isovaleric acid and nitrobenzene, additional experiments were also carried out without solvent [16], with the metals zinc and copper and with another galvanometer (resistance 361 ohms).

Table 36

Molecular percent \( \mathrm{C_6H_5NO_2} \)	100	90	80	70	60	50	40	30	20	10	0
Molecular percent \( \mathrm{C_5H_{10}O_2} \)	0	10	20	30	40	50	60	70	80	90	100
Observed value, EMF mV	4	1484	1780	2042	1939	2456	1745	690	30	29	0

Thus, all the carboxylic acids that we used give, with nitro compounds, complexes that are considerably more active with respect to

metals than the acids themselves.*) In this case acids with a branched chain, especially isovaleric acid, gave, in complexes with nitrobenzene and nitronaphthalene, a much greater EMF than acids of normal structure. The difference observed in our experiments, as compared with acids of normal structure, exceeds, it seems, the difference in the values of all the other constants of these acids. This feature can be used for determining the structure of acids.

The addition of a small amount of water to the acid sharply increases the EMF indicated by the instrument. Thus, with 1.74 ml of acetic acid and 1.02 ml of nitrobenzene in 6 ml of benzene solution, we obtained an EMF of 5.4 mV. With a content of 1% water in the acid, the EMF rose to 55 mV; with 2% \(H_2O\)—to 110 mV.

In view of this, we once again distilled isovaleric acid and separately tested in our instrument two of its fractions with b.p. 174–175° and 175–177°. With 1.89 ml of acid and 0.76 ml of nitrobenzene per 6 ml of solution, EMFs of 55 mV and 60 mV were obtained, i.e., under the conditions of our experiment the large value of the electromotive force was given by the isovaleric acid itself, without the participation of water or any other impurities.

In the system \(C_6H_5NO_2—C_5H_{10}O_2\), the sharp maximum of EMF observed corresponds to an equimolecular mixture. According to the foregoing, this should be a complex compound:

\[ \mathrm{ C_6H_5{-}N(=O){-}O\cdots HOC(=O){-}CH_2CH(CH_3)CH_3 \rightleftharpoons \left[\mathrm{C_6H_5NOH(=O)}\right]^{\cdot} \left[\mathrm{O.C(=O){-}C_4H_9}\right]^{\prime} } \]

For other acids the maxima corresponded to 60–65 molecular percent, i.e., about \(C_6H_5NO_2 \cdot 2RCOOH\) or, possibly, \(C_6H_5NO_2 \cdot (RCOOH)_2\), since the simplest monocarboxylic acids of the saturated series (for example, the very well studied acetic acid) give dimers that persist even in the vapors of the boiling acid and in its benzene solution. The very strong influence of water on the EMF given by acetic acid is explained, perhaps, by the fact that water splits the dimer into separate \(CH_3COOH\) molecules, which then add to nitrobenzene according to the scheme written above. However, here too the maximum remains at 60% (and not at 50%). This may occur as a consequence of incomplete dissociation of the dimer, or \(C_6H_5NO_2\) adds two molecules of acid to its two oxygen atoms.

\[ \mathrm{ C_6H_5N(=O){-}O\cdots HO{-}C(=O){-}CH_3 \quad O\cdots HO{-}C(=O){-}CH_3 \rightleftharpoons \left[\mathrm{C_6H_5NOH(=O)}\right]^{\cdot} \left[\mathrm{O{-}C(=O){-}CH_3\ \ O\cdots HO{-}C(=O){-}CH_3}\right]^{\prime} } \]

\[ \mathrm{ C_6H_5N(=O){-}O\cdots HO{-}C(=O){-}CH_3 \quad O\cdots HO{-}C(=O){-}CH_3 \rightleftharpoons \left[\mathrm{C_6H_5N(OH)_2}\right]^{\cdot\cdot} \left[\mathrm{OC(=O){-}CH_3}\right]^{\prime}_{2} } \]

*) The increase in the activity of the acid upon addition of nitrobenzene can be demonstrated in a very simple experiment. Pour an acid (preferably isovaleric acid) into one test tube, nitrobenzene into another, and a mixture of both compounds into a third, and drop into each a piece of metallic sodium. In the test tube with the mixture of \(C_5H_{10}O_2\) and \(C_6H_5NO_2\), the reaction proceeds most vigorously; with the acid alone sodium reacts considerably more slowly; with nitrobenzene the reaction is imperceptible.

Dimeric molecules, apparently, are formed more readily by acids of normal structure. With branched chains this may be hindered by steric obstacles.

Acids with amines

Acids with amines give salts. It is known, however, that the radicals entering into their composition have a great influence on the basic properties of amines. In aniline the basic properties are weaker than in ammonia. Triphenylamine is already incapable of forming a salt even with hydrochloric acid. If weak acids are taken, for example monobasic acids of the saturated series, then one may suppose that with them salt formation will be observed all the more not with all amines.

The strength of acid properties is usually determined from the constant of electrolytic dissociation, which is calculated from the results of conductivity measurements. For the basic properties of amines the criterion is the dissociation constant of the hydroxides formed by them with water:

\[ \mathrm{R_3N} + \mathrm{HOH} \rightleftarrows \mathrm{R_3N} \begin{matrix} \diagup \mathrm{H}\\[-2pt] \diagdown \mathrm{OH} \end{matrix} \rightleftarrows [\mathrm{R_3NH}]^{\cdot} + (\mathrm{OH})' \]

or, more precisely:

\[ \mathrm{R_3N} + \mathrm{HOH} \rightleftarrows \mathrm{R_3N}\ldots \mathrm{HOH} \rightleftarrows [\mathrm{R_3NH}]^{\cdot} + (\mathrm{OH})' \]

We decided [38] to use, in the study of acid–amine systems, the electrochemical method described in the present work. The apparatus was the same as in the study of complexes of acids with nitro compounds. The metals were sodium and platinum. The reacting substances were taken without solvent or in benzene solution of the same concentration as with nitro compounds, i.e. the sum of the amounts of both complex formers was 0.025 gram-mole and the mixture was diluted with benzene to 6 ml. The temperature at the beginning of the experiments was 18–20°, but in the course of the reaction with sodium self-heating sometimes occurred, especially considerable in the absence of solvent.

Of the acids, propionic, butyric, isobutyric, and isovaleric acids were taken. With formic and acetic acids it was not possible to carry out experiments, since the sodium rapidly became covered with a layer of insoluble salt, which interrupted the current. As complex formers of basic character, triethylamine, aniline, dimethylaniline, and pyridine were taken.

The results of the determinations are given in Tables 37 and 37a.

As was to be expected, the magnitude of the electromotive force shown by our instrument, arising in the reaction with sodium of a salt of a carboxylic acid with an amine, is noticeably greater than for the free acid. Free amines with sodium either do not react at all or (for example, pyridine) give only a negligible EMF.

At the same time, a clear parallelism is observed between the electrolytic dissociation constant of the hydroxide formed by the amine in aqueous solution and the EMF value determined by us. Of the amines we took, triethylamine has the most basic character; it has a dissociation constant \(k = 6.4 \cdot 10^{-4}\) at 25°, and precisely \((\mathrm{C_2H_5})_3\mathrm{N}\) gave a large increase in EMF with all acids. Pyridine (\(k = 2.3 \cdot 10^{-9}\)) also gave an increase in EMF with all four acids, but with propionic and butyric acids the maxima were small. Aniline (\(k = 4.6 \cdot 10^{-10}\) at 45°) and dimethylaniline (\(k = 2.42 \cdot 10^{-10}\) at 18°) markedly increase the EMF for isovaleric acid, but with propionic and butyric acids there is no longer a maximum. It is possible that here there is no strict additivity (for the solution of this

Table 37

Experiments without solvent

No.	Acids	Amines	40	50	60	70	80	90
1	$\mathrm{C_3H_6O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		150		460	550	150
2	$\mathrm{C_3H_6O_2}$	$\mathrm{C_6H_5NH_2}$				4.6	8.2	12.6
3	$\mathrm{C_3H_6O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				5.2	6.0	30
4	$\mathrm{C_3H_6O_2}$	$\mathrm{C_5H_5N}$				120.4	100
5	$\mathrm{C_4H_8O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		700		1100	1600
6	$\mathrm{C_4H_8O_2}$	$\mathrm{C_6H_5NH_2}$				7.5	10.4	10.2
7	$\mathrm{C_4H_8O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				5.4	7.6	10.0
8	$\mathrm{C_4H_8O_2}$	$\mathrm{C_5H_5N}$				50	60.5
9	$i$-$\mathrm{C_4H_8O_2}$	$\mathrm{C_5H_5N}$	92	200	120	180.5	80
10	$i$-$\mathrm{C_5H_{10}O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		125		450	580
11	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_6H_5NH_2}$				55.5	72.4	52.0
12	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				75	50	12.4
13	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_5H_5N}$				400

Maximum observed values of EMF in mV at the molecular percent of acid indicated in the column headings.

Table 37a

Experiments with benzene solutions

No.	Acids	Amines	40	50	60	70	80	90
1	$\mathrm{C_3H_6O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		70		350	500	180
2	$\mathrm{C_3H_6O_2}$	$\mathrm{C_6H_5NH_2}$				0.1	0.38	0.8
3	$\mathrm{C_3H_6O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				0.18	0.38	0.98
4	$\mathrm{C_3H_6O_2}$	$\mathrm{C_5H_5N}$		1.2		4.6	3.4
5	$\mathrm{C_4H_8O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		450		650	1350
6	$\mathrm{C_4H_8O_2}$	$\mathrm{C_6H_5NH_2}$	—	—	—	0.2	0.5	0.65
7	$\mathrm{C_4H_8O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				0.18	0.2	0.6
8	$\mathrm{C_4H_8O_2}$	$\mathrm{C_5H_5N}$				1.4	1.42
9	$i\mathrm{C_4H_8O_2}$	$\mathrm{C_5H_5N}$	30	50	38	18	4.4
10	$i$-$\mathrm{C_5H_{10}O_2}$	$(\mathrm{C_2H_5})_3\mathrm{N}$		60	250	300	275
11	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_6H_5NH_2}$				1.8	2.2	1.4
12	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_6H_5N(CH_3)_2}$				2.4	1.7	0.8
13	$i$-$\mathrm{C_5H_{10}O_2}$	$\mathrm{C_5H_5N}$				205

Maximum observed values of EMF in mV at the molecular percent of acid indicated in the column headings.

…of the question, these systems must be investigated in greater detail, but in any case either the reaction between the acid and the amine in these systems scarcely proceeds at all, or else certain complexes are formed that act only weakly on the metal.

The acids studied by us—propionic $(k_{25}=1.3\cdot10^{-5})$, butyric $(k_{25}=1.48\cdot10^{-5})$, isobutyric $(k_{25}=1.5\cdot10^{-5})$, and isovaleric $(k_{25}=1.7\cdot10^{-5})$—differ little from one another in their electrolytic-dissociation constants in aqueous solution. In our experiments, a sharp difference was observed between propionic and butyric acids on the one hand, and isobutyric and isovaleric acids on the other. The first two acids show a very considerable increase of EMF in compounds with triethylamine, a small increase with pyridine, and no maximum with aniline and dimethylaniline. Isovaleric acid likewise gave the greatest EMF value in experiments with triethylamine, a somewhat smaller value (but much greater than propion-

... (the new and the oily acids) with pyridine, and small but quite distinct maxima with aromatic amines. A somewhat smaller EMF in mixtures with pyridine (no experiments were performed with other amines) was obtained for isobutyric acid. Thus, carboxylic acids with amines behave in the same way as with nitro compounds: acids with a branched chain prove to be considerably more active than acids of normal structure.

As for the composition of the mixtures most active with respect to sodium, it almost never corresponded to the simple ammonium salt \(R\mathrm{COOH}\cdot NR_3\). The EMF maximum in most cases was shifted toward an excess of acid. It may be thought that if an acid forms a salt with an amine, then it enters this salt in the form of a dimer, and perhaps also in the form of an even more highly associated molecule. For example, at 80 molecular percent acid we have the compound \(NR_3\cdot 4R\mathrm{COOH}\).

IV. Other Complex Organic Compounds

Complexes Formed by Substances Passing into a Tautomeric Form with Active Hydrogen

Of the compounds that can pass into the enol or an analogous enol tautomeric form, we studied acetoacetic ester and nitromethane:

\[ \begin{array}{ccccc} \mathrm{CH_3} && \mathrm{CH_3} && \mathrm{CH_3} \qquad \mathrm{CH_2} \\ | && | && | \qquad \| \\ \mathrm{C{=}O} && \mathrm{COH} && \mathrm{N{=}O} \quad \rightleftarrows \quad \mathrm{N{=}O} \\ | && \| && \| \qquad | \\ \mathrm{CH_2} & \rightleftarrows & \mathrm{CH} && \mathrm{O} \qquad \mathrm{OH} \\ | && | && \\ \mathrm{C{=}O} && \mathrm{C{=}O} && \\ | && | && \\ \mathrm{O.C_2H_5} && \mathrm{O.C_2H_5} && \end{array} \]

Both of these substances, even by themselves, showed considerable activity with respect to metallic sodium. It was of interest to test them also in mixtures with complex-formers of the basic type, for example with nitrobenzene, and with “acidic” complex-formers (carboxylic acids and phenols). These systems were studied without solvent, with Al and Sb electrodes and with a galvanometer having a resistance of 2800 ohms (18).

Table 38

	100	90	80	70	60	50	40	30	20	10	0
Mol. percent \( \mathrm{CH_3COCH_2COO_2C_2H_5} \)	100	90	80	70	60	50	40	30	20	10	0
\( \mathrm{C_6H_5NO_2} \)	0	10	20	30	40	50	60	70	80	90	100
EMF in mV	11	—	—	3.8	—	2.7	—	—	—	—	0.9
Mol. percent \( \mathrm{CH_3COCH_2COOC_2H_5} \)	100	90	80	70	60	50	40	30	20	10	0
\( \mathrm{C_5H_{10}O_2} \)	0	10	20	30	40	50	60	70	80	90	100
EMF in mV	11	7.2	—	6.05	—	2.5	—	—	—	—	0
Mol. percent \( \mathrm{CH_3COCH_2COOC_2H_5} \)	100	90	80	70	60	50	40	30	20	10	0
\( \mathrm{C_6H_5OH} \)	0	10	20	30	40	50	60	70	80	90	100
EMF in mV	11	12.1	13	15.9	13.9	18.6	12.5	13	10.3	13.4	—

The figures in this table show that the formation of complexes of acetoacetic ester with other compounds can by no means always be

was observed. Thus, with phenol it gives definite EMF maxima (the largest—at 50 molecular percent), but with isovaleric acid and nitrobenzene no maxima were found. The picture here is obscured, it must be supposed, by the fact that the molecules of acetoacetic ester itself, in its different forms, must be capable of mutual complex formation. This must also explain the very great relative activity of pure acetoacetic ester, for which, with sodium and platinum in a benzene solution of concentration 7 g-mol per liter, we observed an EMF of 250 mV, whereas ethyl lactate under the same conditions gave 120 mV. The concentration of hydroxyl groups in ethyl lactate is greater, since all 100% of it have an alcoholic group. In acetoacetic ester the concentration of the hydroxyl-containing form is smaller, but its activity is greatly increased by complex formation.

For nitromethane, experiments were carried out [41] with nitrobenzene, triethylamine, and pyridine. \(CH_3NO_2\), in an amount of 1.3 ml, diluted with benzene to 6 ml, gave (with sodium and platinum) 0.1 mV. On addition of an equimolecular amount of \(C_6H_5NO_2\), 6.6 mV was obtained; with \((C_2H_5)_3N\), under the same conditions, 4.0 mV; with \(C_5H_5N\), 16 mV. There is no doubt that nitromethane reacts here in its acidic form.

Complexes of compounds not containing active hydrogen

We studied [13] mixtures of para-benzoquinone

\[ \begin{array}{c} \begin{array}{ccc} & CH=CH & \\ O=C & \hexagon & C=O \\ & CH\quad CH & \end{array} \end{array} \]

with all isomeric nitrotoluenes (ortho, meta, and para), nitrophenols, nitroanilines, nitrobenzoic acids, and with paranitrodimethylaniline. Nitrotoluenes and nitrodimethylaniline do not have hydrogens that could be replaced by metals. In \(NO_2C_6H_4NH_2\), the hydrogens of the amino group might react, but the activity of amino hydrogen is in general much weaker than that of the hydrogen atom of a hydroxyl group; in our experiments the nitroanilines behaved analogously to nitrodimethylaniline. All this indicates that they did not react by hydrogen. We consider nitrophenols and nitrobenzoic acids here, and not together with other phenols and acids, because the experiments were carried out under quite different conditions, exactly as with nitrotoluenes, etc. For comparison, \(C_6H_5OH\) was also taken, which with quinone gives the long-known phenoquinone \(C_6H_4O_2 \cdot 2C_6H_5OH\). The substances under study were taken in nitrobenzene solution at a total concentration of 0.02 g-mol per liter. As is evident from the preceding, nitrobenzene itself is a good complex former, but the results of the experiments showed that in it, too, one can clearly detect the formation of complexes between quinone and the other nitro compounds listed.

In the work, a platinum crucible and sodium were used as the active metal. The experiments were conducted at 16–20°.

All the compounds taken with quinone gave an increase in the electromotive force observed on the millivoltmeter. The results of some experiments are given in Tables 39–43. The tables also trace the change of EMF with time.

The compounds of quinone with nitrophenols are evidently built according to the general phenoquinone–quinhydrone type. It is believed [45] that here, on the side of the phenols, the carbon atoms participate in complex formation.

Table 39

Phenol + quinone

Molecular ratios of components	Molecular ratios of components	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after
\(C_6H_4O_2\)	\(C_6H_5OH\)	0 min.	1 min.	5 min.	10 min.	15 min.
0	3	9.7	13.9	15.15	21.2	28.0
0.5	2.5	12.7	14.95	15.15	17.3	26.0
1	2	15.15	17.0	18.8	20.6	25.45
1	1	12.1	14.85	17.0	18.4	21.2
2	1	8.1	9.4	9.7	8.5	8.9
2.5	0.5	5.45	6.3	8.2	9.1	7.3
3	0	5.15	4.8	4.6	4.5	—
Ready-made component 1:2	Ready-made component 1:2	18.2	23.9	>30	>30	>30

Table 40

Para-nitrophenol + quinone

Molecular ratio of components	Molecular ratio of components	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after
\(C_6H_4O_2\)	\(NO_2C_6H_4OH\)	0 min.	1 min.	5 min.	10 min.	15 min.
0	3	14.5	29.7	>30	>30	>30
0.5	0.5	13.6	24.6	29.1	>30	—
1	2	15.15	>30	—	—	—
1	1	8.5	14.9	18.2	>30	—
2	1	6.7	10.3	12.7	22.7	>30
2.5	0.5	5.8	5.95	6.0	7.5	8.7
Comp. 1:2	Comp. 1:2	21.8	29.0	>30	—	—

Table 41

Para-nitrotoluol + quinone

Molecular ratio of components	Molecular ratio of components	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after
\(C_6H_4O_2\)	\(p.\,NO_2C_6H_4CH_3\)	0 min.	1 min.	5 min.	10 min.	15 min.
0	3	3.0	2.7	2.4	1.8	1.5
0.5	2.5	3.15	3.0	2.6	2.3	2.1
1	2	3.6	3.3	2.7	2.4	1.8
1	1	4.8	4.2	3.6	3.0	2.7
2	1	4.2	3.9	3.3	3.0	2.9
2.5	0.5	4.2	3.9	3.0	2.4	2.1

Table 42

Para-nitraniline + quinone

Molecular ratio of components	Molecular ratio of components	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after	Observed EMF values, in mV, after
\(C_6H_4O_2\)	\(p.-NO_2C_6H_4NH_2\)	0 min.	1 min.	5 min.	10 min.	15 min.
0	3	2.3	2.3	2.1	1.8	1.8
0.5	2.5	3.6	3.3	3.0	2.4	2.1
1	2	3.9	3.0	2.4	2.1	2.1
1	1	4.85	4.5	3.6	3.0	2.7
2	1	3.9	3.6	3.0	2.7	2.4
2.5	0.5	4.2	3.9	3.3	2.4	2.1

Table 43

para-Nitrodimethylaniline + quinone

Molecular ratio of components	Molecular ratio of components	Observed EMF values after	Observed EMF values after	Observed EMF values after	Observed EMF values after	Observed EMF values after
$C_6H_4O_2$	$pNO_2C_6H_4\cdot N(CH_3)_2$	0 min.	1 min.	5 min.	10 min.	15 min.
0	3	0.75	5.4	4.8	4.5	3.6
0.5	2.5	5.4	5.5	4.5	3.6	3.3
1	2	5.4	4.8	4.5	3.6	3.0
1	1	5.7	4.8	4.2	3.3	3.0
2	1	5.4	4.5	3.6	3.0	2.7
2.5	0.5	4.85	4.2	3.3	2.7	2.7

of the benzene ring, which are attracted by the quinone oxygen. This is contradicted by the coincidence, in the complex, of the number of phenolic hydroxyls and oxygen atoms of quinone:

\[ \begin{array}{cc} C_6H_4O_2\cdot 2C_6H_5OH; & C_6H_4O_2\cdot C_6H_4(OH)_2 \\ \text{phenoquinone} & \text{quinhydrone} \end{array} \]

This indicates the possibility that hydroxyl hydrogens participate in the reaction; these could quite well attach to the oxygens of the carbonyl groups.

\[ C_6H_5OH \cdots O=C \left\langle \begin{array}{c} CH\quad CH\\[-2pt] \Vert\quad\Vert\\[-2pt] CH\quad CH \end{array} \right\rangle C=O \cdots HOC_6H_5 \]

It is still difficult to say anything definite about the structure of the complex compounds we have found, in which both constituent parts are devoid of active hydrogens. In nitrotoluenes, the nitro group may possibly attach by oxygen to the carbon, or by nitrogen to the oxygen, of the quinone carbonyl; or perhaps the complexes are formed through the carbon atoms of the benzene nucleus. In nitro derivatives of aromatic amines, there is also the additional possibility of attraction of the nitrogen of the amino group to the carbon of quinone. In order to choose among several conceivable directions of complex formation, further investigations are necessary.

V. Determination of the Composition of Complexes and Comparison of Our Method with Other Methods for Studying Complex-Forming Systems

Isolation of Complex Compounds

After, by means of one or another method of physicochemical analysis, we have found that the compounds studied form a complex, the task arises of determining the composition of this complex. This composition may directly correspond to some special point on the composition–property diagram, for example, to a maximum or minimum of the measured physical constant; sometimes it may deviate from special points. As will be shown when applying the new method of analysis, we can approach the solution of this question either theoretically or find the answer experimentally.

In the present case, the theoretical elucidation of the dependence of the measured quantity on the composition of the mixture is made difficult by the complexity of the phenomenon under study.

Experimental verification is simpler. It can be carried out either by isolating the molecular compounds formed, or by comparing the results obtained with those given by old, well-tested methods.

It has so far been possible to isolate complex compounds in the following systems: quinone–phenol (phenoquinone is obtained), quinone–p-nitrophenol, quinone–p-nitrotoluene, quinone–o-nitroaniline, quinone–m-nitroaniline, quinone–p-nitroaniline. In all these cases the reacting components are solid, readily crystallizing substances. When they were fused and crystallized from suitable solvents, crystalline molecular compounds were also obtained, which were analyzed. The quinone content was determined from the liberation of iodine from an acidified potassium iodide solution.

Thus the following new compounds were obtained:

\[ \mathrm{C_6H_4O_2 \cdot 2p{-}NO_2 \cdot C_6H_4OH;} \qquad \mathrm{C_6H_4O_2 \cdot p{-}NO_2 \cdot C_6H_4CH_3;} \]

\[ \mathrm{C_6H_4O_2 \cdot o{-}NO_2 \cdot C_6H_4NH_2;} \quad \mathrm{C_6H_4O_2 \cdot m{-}NO_2 \cdot C_6H_4NH_2;} \quad \mathrm{C_6H_4O_2 \cdot p{-}NO_2 \cdot C_6H_4NH_2.} \]

In all these cases, as with the already known phenoquinone, the composition of the complex corresponded to the maximum deflection of the millivoltmeter needle in our apparatus.

Moreover, whenever the complex compound could be isolated, a sharp increase in the observed EMF was noted upon mixing the components, although its value was different from that for the finished complex. In systems with a less pronounced maximum (for example, quinone–p-nitrodimethylaniline), attempts to isolate molecular compounds gave no results. Thus, from the course of the change in EMF as a function of the composition of the mixture, one can not only determine the composition of the complex, but also judge approximately its stability (Tables 39–43).

Comparison of Our Method with Other Methods of Physicochemical Analysis

A large part of the complex-forming systems we have studied has not yet been investigated at all by other methods. Only for a few systems is it possible to make a comparison with the results obtained by applying other methods for recognizing molecular compounds.

1. Thermal analysis. Grinakowski, Stashevsky, and Shmitovna [50], from the melting diagram, found that phenol with nitrobenzene forms only a eutectic with a melting temperature of −16.4° (at 58.3 mol.% \(\mathrm{C_6H_5NO_2}\)). We found a large increase in EMF not far from this point.

One of the complexes of phenol with benzophenone, namely \(\mathrm{C_6H_5OH \cdot 2C_6H_5CO \cdot C_6H_5}\), was also detected [24] by the melting temperature, which, for a mixture of this composition, proved to be higher than for mixtures of similar composition. A complete thermal analysis of this system was not carried out.

1. Electrical conductivity. D. P. Konovalov [9] determined the electrical conductivity of certain mixtures of carboxylic acids with amines. Thus, for propionic and butyric acids with aniline he obtained the following data (Table 44).

When recalculated into molecular percentages, the maxima of electrical conductivity occur at approximately 30% aniline. According to the experiments of D. P. Konovalov [10], the maxima of the heat of mixing, the heat capacities of the solutions, and the heat of dilution are also shifted toward a higher acid content. All these results are therefore similar to those given by our method. M. I. Usanovich and A. Tenenbaum [43], on the contrary, found no special points on the electrical-conductivity and viscosity diagrams of the system—acetic

Table 44

\(\mathrm{CH_3\cdot CH_2COOH + C_6H_5NH_2}\)		\(\mathrm{CH_3\cdot CH_2\cdot CH_2COOH + C_6H_5NH_2}\)
Weight percent of aniline	\(K_{21}\)	Weight percent of aniline	\(K_{31}\)
22.30	3.56	15.65	0.01
29.68	5.98	27.10	0.67
36.46	6.73	36.80	1.03
42.61	6.35	44.73	0.86
49.40	4.99	51.35	0.61
55.51	3.40	58.30	0.33
60.76	2.05	62.00	0.22
65.70	1.07	67.44	0.12
70.49	0.53	70.96	0.08

acid—nitrobenzene, whereas in our case this system also gave a definite maximum.

A comparison was also made [14] between the EMF (i.e., the current strength in our instrument, measured with a millivoltmeter) and the electrical conductivity for nitrobenzene solutions: (a) of hydroquinone, quinone, and quinhydrone; (b) of naphthalene, picric acid, and naphthalene picrate.

The electromotive force was determined with sodium and a platinum crucible. The electrical conductivity was measured by means of a Wheatstone bridge. The temperature in these and the other experiments was 17–18°. The concentration of the reacting substances was 0.02 g-mol per liter. The nitrobenzene solvent was brought, by purification, to the standard EMF value \(=0.8\) mV under the same conditions (with the same millivoltmeter having a resistance of 2167 ohms).

The EMF and conductivity readings were taken immediately after the sodium was immersed in the liquid and at definite intervals of time, up to 15 min, when a certain amount of metal had had time to pass into the solution (which markedly increased the electrical conductivity).

The results of the measurements are given in Table 45.

Table 45

Test substances	Observed EMF values in mV after	Observed EMF values in mV after	Observed EMF values in mV after	Observed EMF values in mV after	Specific electrical conductivity after	Specific electrical conductivity after	Specific electrical conductivity after	Specific electrical conductivity after	Specific electrical conductivity after
	0 min.	5 min.	10 min.	15 min.	0 min.	1 min.	5 min.	10 min.	15 min.
Hydroquinone	2.5	2.4	2.4	2.35	0.09	0.137	0.176	0.181	0.190
Quinone	7.3	5.5	5.0	4.4	0.09	0.09	0.09	0.09	0.09
Hydroquinone + quinone	5.45	4.7	4.3	3.9	0.09	0.152	0.191	0.222	0.270
Quinhydrone	5.45	5.15	4.8	4.5	0.09	0.09	0.127	0.131	0.142
Picric acid	25.15	23.6	21.7	21.6	0.554	0.640	0.723	0.800	0.900
Naphthalene	1.2	1.2	1.2	1.1	0.09	0.09	0.09	0.09	0.09
Picric acid + naphthalene	28.5	23.3	21.2	20	0.549	0.690	0.725	0.810	0.905
Naphthalene picrate	29.8	29.7	29.4	29.1	0.555	0.693	0.750	0.810	0.890

From the figures in this table it is evident that the EMF value shown by the millivoltmeter in our instrument characterizes the molecular compound better than does the electrical conductivity.

With quinhydrone this is obscured by the fact that the complex gives a smaller EMF than one of its constituent parts (quinone). In the picric acid—

naphthalene—a quite definite complex compound—gave a higher EMF than the component active with respect to metals, picric acid, while at the same time the difference in electrical conductivity lies within the limits of measurement error.

1. Viscosity. Some authors note that good results in the study of complex-forming systems are obtained by determining internal friction. Thus A. M. Tikhomirova and N. N. Efremov [26] used this method to find singular points in mixtures of phenol with amines, for example with ethylaniline and diethylaniline, where the melting diagram gave nothing. In the works of V. V. Udovenko [41] the method of internal friction was applied to systems composed of carboxylic acids and ketones. Chemical interaction was observed, sometimes quite distinct (acetone and formic acid), sometimes barely noticeable. The same was confirmed by density curves.

For phenol with amines, V. V. Udovenko and M. I. Usanovich confirmed the presence of complex compounds by cryoscopic determination of molecular weight [42].

1. Surface tension can also be used to recognize molecular compounds. N. A. Trifonov and R. V. Mertslin [28] in this way found the formation of complexes in some of the systems we studied (e.g., phenol, piperidine).

2. Refractive index [16]. Determination of the refractive index was carried out with the systems: ethyl acetate–phenol, nitrobenzene–phenol, nitrobenzene–isovaleric acid, nitrobenzene—ethyl acetate.

No deviations from additivity were observed. Meanwhile, by our method complexes are clearly detected in some of these systems.

Thus, in studying complex-forming systems in organic chemistry, the method we have proposed for determining activity with respect to metals either gives the same results as those obtained by applying other methods of physicochemical analysis, or proves to be more sensitive than such methods as thermal analysis, measurement of electrical conductivity, and determination of the refractive index.

As is evident from the factual material presented above, our method can be used in studying mixtures of the most diverse substances, since metals can replace hydrogen or add to atoms of oxygen, nitrogen, carbon, and other elements in very many compounds. Of course, in some cases the reaction with the metal may not proceed to any noticeable extent, or complex formation (somewhere else in the molecule) will have little influence on this reaction. Thus, when mixing nitrobenzene with aniline or dimethylaniline, we cannot detect the formation of a molecular compound by EMF, whereas the appearance of color clearly indicates complex formation. However, there are few such examples, and in general the method provides the chemist with a good means of easily and rapidly finding complex compounds and determining their composition.

PART FOUR

SOME PRACTICAL APPLICATIONS OF THE NEW ELECTROCHEMICAL METHOD

1. General Considerations

Electrochemical methods are of great importance in determining the purity of certain substances. Thus, the purity of water is best determined by electrical conductivity. The magnitude of the electrical conductivity is sharply affected by such negligible traces of impurities that cannot be detected by any other methods. The extreme sensitivity of electrical conductivity to all contaminations makes it necessary, in electrochemical investigations, to purify substances especially carefully and to keep the preparations intended for experiments in sealed vessels.

The current strength measured in the closed circuit by the method we have developed depends both on the electrical conductivity and on the electromotive force that produces the current. At the same time, both structural factors and the addition of other compounds to the given substance act, as a rule, on the electromotive force and on the electrical conductivity in the same direction. Hence the deflection of the galvanometer needle in our apparatus should change still more sharply depending on the nature of the substance under investigation itself and on the impurities contained in it. This is fully confirmed by the results of the experiments described above with complex-forming systems. We have seen that the method we have proposed makes it possible to detect such complexes whose formation cannot be observed from electrical conductivity.

This led us to expect that, in determining the purity of many substances as well, the method would prove to be very sensitive. If a compound that does not react, or reacts only with difficulty, with metals—for example, with sodium—contains a small impurity of another substance that is very active with respect to the metal, then the amount of this impurity should be determined easily and accurately.

In organic substances, water is especially often such an active impurity. Alcohols, phenols, and acids are sufficiently active with respect to metals, and their content, for example, in hydrocarbon mixtures, should be readily detected qualitatively and determined quantitatively in this way. If there is too little of the hydroxyl-containing compound in the mixture under investigation for it, by reacting with the metal, to produce a noticeable current strength in the apparatus, then sometimes the determination of such an impurity can be carried out after adding a reagent that gives a more active complex with it. Thus, the magnitude of the electromotive force, shown by a voltmeter in our apparatus, arising in the reaction of phenols or carboxylic acids with sodium, sometimes increases hundreds of times upon addition of nitro compounds or amines. In this way we can further increase the sensitivity of the method.

So far, in the electrochemical determination of the purity of organic compounds, we have done the following: (a) the influence of the content of small amounts of water in pyridine, acetic acid, and phenol has been determined; (b) a procedure has been developed for determining some of the most active compounds, or compounds that sharply increase the activity in complex formation, in less active substances—for example, phenol in benzene or cyclohexanol; (c) the influence of various impurities on the properties of nitrobenzene and phenol has been studied in detail. In connection with this, methods have been developed

purification of these compounds, making it possible to obtain them in the most electrochemically pure form.

The proposed electrochemical method of investigation may also find application in kinetic measurements (for example, determination of the rate of formation or disappearance of a substance active with respect to metals). It is also useful in elucidating the direction of a reaction (examples will be given below).

Finally, in inorganic chemistry the same method can be used to determine the purity of water, to analyze certain alloys, etc.

II. Determination of Water in Organic Substances

For the qualitative and quantitative determination of water in organic compounds and their mixtures, a wide variety of methods are used, depending on the nature of the substance under examination.

Thus, in the study of high-boiling compounds and mixtures, for example lubricating oils, the water can be distilled off with the vapors of an added hydrocarbon of medium volatility, for example xylene (according to Dean and Stark). The water content in various substances that do not react with metals can be determined from the volume of hydrogen evolved upon addition of sufficiently active metals, for example sodium, and still better with metal hydrides, for example calcium hydride, since in this case twice as much gas is evolved:

\[ 2\mathrm{H}_2\mathrm{O} + 2\mathrm{Na} \rightarrow 2\mathrm{NaOH} + \mathrm{H}_2; \]

\[ 2\mathrm{H}_2\mathrm{O} + \mathrm{CaH}_2 \rightarrow \mathrm{Ca(OH)}_2 + 2\mathrm{H}_2. \]

For the qualitative determination of water in alcohol, the well-known reaction of formation of the blue hydrate of copper sulfate is used,

\[ \mathrm{CuSO}_4 \cdot 5\mathrm{H}_2\mathrm{O} \]

upon addition of anhydrous sulfate to the alcohol being tested.

Sometimes the substance is mixed with a liquid that dissolves almost no water. For example, the water content in phenol can be determined with the aid of chloroform.

For the quantitative determination of water in liquids miscible with it, for example in ethyl alcohol or acetic acid, physical constants are often used, especially specific gravity.

Our electrochemical method, as will be seen from the examples given below, gives very good results in the determination of small amounts of water in certain substances.

A preparation of pyridine that we had in anhydrous form, when diluted with half its volume of benzene, gave 4 mV with sodium and platinum [41]; with 1% \(\mathrm{H}_2\mathrm{O}\) we obtained 8 mV, with 2%—45 mV, and with 3%—1150 mV.

In other cases, direct determination of water proved insufficiently accurate, but it was possible to make the method more sensitive by means of a third substance capable of giving an active complex. For example, acetic acid in an equimolecular mixture with nitrobenzene (0.025 g-mol each, diluted with benzene to 6 ml; electrodes—sodium and platinum) gave 5 mV. Upon addition of 1% water (relative to the acid), 40 mV was obtained; with 2%—75 mV, and with 3%—112 mV.

In Table 46, using phenol with various water contents as an example, the results obtained by our method are compared with those given by determination of the freezing point—the constant being very sensitive to impurities of low molecular weight, i.e., also to water. In work by our method, the phenol was taken in nitrobenzene solution (0.0047 g per 5 ml, i.e., 0.01 g-mol per liter).

Table 46

Percent water in $C_6H_5OH$	Freezing temperature	Observed EMF values, mV, after 20 sec.	Observed EMF values, mV, after 1 min.	Observed EMF values, mV, after 2 min.	Observed EMF values, mV, after 5 min.	Observed EMF values, mV, after 10 min.
0%	40.92°	3.9	4.8	6.1	14.2	23.0
0.0001	—	4.1	4.8	6.5	14.4	23.4
0.0002	—	4.6	5.0	8.0	15.1	23.8
0.0005	40.85	5.0	7.6	8.8	15.3	23.9
0.001	40.00	5.9	8.2	9.4	15.9	23.7
0.002	—	6.1	8.9	9.7	16.6	25.0
0.005	39.8	6.4	9.2	10.7	18.0	25.5
0.01	39.2	7.3	9.6	11.2	18.9	26.7
0.1	38.75	8.4	10.0	12.5	26.6	—
1	36.90	18.3	21.9	26.4	—	—

Thus even so moisture-sensitive a constant as the melting point changes less sharply (at any rate for phenol) than the electromotive force indicated by the millivoltmeter in our apparatus. It should also be noted that determining the freezing temperature with an accuracy to hundredths of a degree requires a rather considerable amount of substance (on the order of 25 g), whereas in each experiment we take 0.0047 g of phenol.

III. Determination of the content of organic compounds distinguished by special activity with respect to metals

Determination of phenol in benzene [35]

An admixture of phenol to benzene in an amount up to 6% could not be detected by direct determination in our apparatus (even with sodium and platinum). After the addition of an equal volume of nitrobenzene, the activity of the solution with respect to sodium increased sharply, despite the twofold decrease in the concentration of phenol. The nitrobenzene used by us showed 6.2 mV. With various solutions of $C_6H_5OH$ in $C_6H_6$ under these conditions the following was obtained: 1%—14 mV, 2%—40 mV, 3%—60 mV, 4%—100 mV, 5%—200 mV.

Determination of phenol in cyclohexanol [39, 24]

Cyclohexanol is prepared industrially by reduction of phenol:

$$ C_6H_5OH + 3H_2 \longrightarrow C_6H_{11}OH $$

Insufficiently purified cyclohexanol therefore always contains some admixture of phenol. The industry consuming cyclohexanol is interested in finding a rapid and convenient method for determining its purity, in particular for determining the phenol content.

We decided to make use of the circumstance that only phenol, but not cyclohexanol, gives with nitrobenzene a complex that is very active with respect to metals. Mixing 1 ml of cyclohexanol (pure or with phe-

with 4 ml of nitrobenzene, the following readings were obtained in our apparatus with sodium and platinum.

Pure cyclohexanol—4.5 mV; with 0.05% phenol—48 mV; with 0.1% phenol—6.7 mV; with 0.2% phenol—11.8 mV; with 0.3%—20.4 mV; with 0.4%—23.4 mV; with 0.5%—26.45 mV.

The accuracy is no less than that given by the colorimetric method proposed for this purpose, and the determination is simpler.

Determination of the purity of benzyl alcohol [39]

In experiments with mixtures of alcohols and amines, we obtained an abnormally large EMF value for the reaction of benzyl alcohol with sodium in the presence of triethylamine. Since the alcohol had stood for several years and could have partly oxidized to benzoic acid, we distilled it over lime. The anomaly in its properties disappeared.

IV. Determination of the purity and standardization of nitrobenzene

Determination of the purity of nitrobenzene

We had to study nitrobenzene in especially great detail, since it is often necessary to use it in investigations of other organic compounds by our method. The nitrobenzene preparations at our disposal gave the following EMF values with sodium and platinum [15].

Table 47

No.	Firm or organization	Grade	Color of product	Observed EMF values, mV
1	Kahlbaum	Purest	Light yellow	0.3
2	Kahlbaum	Pure	"	2.6
3	Kahlbaum	Purified	Light brown	7.2
4	Kahlbaum	Technical	Dark brown	13.8
5	Goslaborreaktiv	ch.d.a	Light yellow	2.7
6	Institute of Chemically Pure Reagents		Light yellow	0.8

From the figures in this table it is evident that the purer the nitrobenzene preparation, the less active it is with respect to sodium. The purest is evidently Kahlbaum’s \( \mathrm{C_6H_5NO_2} \) of the “purest” grade. It gave the smallest EMF value, which moreover did not change after purification by the most varied methods. In second place was the nitrobenzene of the Institute of Chemically Pure Reagents; next came the almost identical preparations Kahlbaum “pure” and Goslaborreaktiv “ch.d.a,” and then Kahlbaum “purified” and, finally, “technical.”

The further investigation that was undertaken had the aim of determining: 1) with what impurities the nitrobenzene preparations were mainly contaminated; 2) how the method of its preparation affects the purity of nitrobenzene; 3) how methods of purification affect it and how purification must be carried out in order to obtain a product of the highest quality; 4) how storage conditions affect it.

Nature of the impurities contaminating nitrobenzene

Water is a fairly common impurity in organic substances. However, water can easily be removed with the aid of drying agents, for example calcium chloride, and even without drying by simple re-

thinly, discarding the turbid first portion of the distillate. Experiments showed that one and the same preparation of \(C_6H_5NO_2\) (a state-standard reagent, “pure for analysis”), after standing for 24 hours with water and simple distillation, gave 1.7 mV, while after drying with calcium chloride and subsequent distillation it showed 1.6 mV.

Among other impurities in nitrobenzene prepared from pure benzene there may be: a) products of further nitration, chiefly metadinitrobenzene; b) products of oxidation of benzene with subsequent nitration and nitrosation.

The normal course of the reaction of benzene with nitric acid is expressed (if one does not speak of certain details, for example of complex addition) by the scheme [31, 42]:

\[ C_6H_6 + \underset{OH}{\overset{O}{\mathrm{N}}{=}O} \;\longrightarrow\; C_6H_5\underset{OH}{\overset{OH}{\mathrm{N}}}{=}O \;\longrightarrow\; C_6H_5\underset{\phantom{OH}}{\overset{O}{\mathrm{N}}}{=}O + H_2O . \]

However, the very appearance of brown vapors indicates the presence of a side process of decomposition of nitric acid and, consequently, oxidation of benzene. This process probably proceeds as follows:

\[ C_6H_6 + \underset{OH}{\overset{O}{\mathrm{N}}{=}O} \;\longrightarrow\; C_6H_5{-}O{-}\underset{OH}{\overset{H}{\mathrm{N}}}{=}O \;\longrightarrow\; C_6H_5OH + HNO_2 \]

and further:

\[ 2\,HNO_2 \longrightarrow H_2O + NO + NO_2 . \]

The phenol that is formed will undergo nitration much more readily than benzene itself, to

\[ NO_2.C_6H_4OH,\quad (NO_2)_2C_6H_3OH \ \text{and}\ (NO_2)_3C_6H_2OH, \]

and also nitrosation. Nitrosophenols, especially the para compound, may also be formed through phenyl nitrite:

\[ C_6H_5{-}O{-}\underset{OH}{\overset{H}{\mathrm{N}}}{=}O \;\longrightarrow\; H_2O + C_6H_5{-}O{-}N{=}O \;\longrightarrow\; O{=}N.C_6H_4OH . \]

Tables 48 and 49 present the results of experiments with the addition of dinitrobenzene, phenol, various nitrophenols, and nitrosophenols. The data in the first of these tables were obtained with the purest nitrobenzene (EMF — 0.3 mV); the other refers to a preparation with \( \mathrm{EMF}=0.8 \) mV [15, 14]. For EMF, the values observed at various intervals after immersion of sodium in the solution are given. For comparison, the EMF values of pure \(C_6H_5NO_2\) are also included; moreover, an untreated preparation and the same nitrobenzene dried with \(CaCl_2\) and distilled were taken. The latter was also used in the experiments with impurities.

We see that even negligible concentrations of phenol and of some nitro- and nitrosophenols in nitrobenzene can be detected by our method. The EMF values observed in our apparatus do not, however, agree with the electrolytic dissociation constants of the compounds taken. Thus 2,6-dinitrophenol has a larger dissociation constant (\(1.7 \cdot 10^{-4}\)) than 2,4-dinitrophenol (\(8 \cdot 10^{-5}\)), but gives a smaller EMF.

Table 48

No.	Impurities	Percent	Observed EMF values, mV, after 0.5 min	Observed EMF values, mV, after 5 min	Observed EMF values, mV, after 10 min
1	Without impurities (unpurif.)	—	0.5	0.5	0.4
2	Without impurities (purif.)	—	0.3	0.3	0.3
3	1,3—C₆H₄(NO₂)₂	0.2	0.3	0.3	0.3
4	1,2—C₆H₄(NO₂)OH	0.2	0.6	0.6	0.5
5	1,4—C₆H₄(NO₂)OH	0.025	1.2	1.5	2.2
6	1,4—C₆H₄(NO₂)OH	0.01	0.8	1.0	1.2
7	1,4—C₆H₄(NO)OH	0.02	1.0	1.8	2.4

Table 49

No.	Impurities	Percent	Observed EMF values, mV, at the moment of immersion	Observed EMF values, mV, after 5 min	Observed EMF values, mV, after 10 min
1	Without impurities	—	0.8	0.6	0.6
2	C₆H₅OH	0.094	13.6	19.1	21.6
3	“	0.0019	1.8	3.0	3.6
4	“	0.00019	0.9	0.9	0.9
5	“	0.000038	0.9	0.8	0.8
6	“	0.000019	0.8	0.6	0.6
7	1,2—C₆H₄(NO)OH	0.005	0.9	0.6	0.6
8	1,3—C₆H₄(NO)OH	0.005	1.5	2.4	3.0
9	1,4—C₆H₄(NO)OH	0.246	24.2	34	95
10	“	0.049	3.6	4.8	6.1
11	“	0.0049	1.5	1.8	2.0
12	“	0.00049	1.1	0.9	0.9
13	“	0.000049	0.8	0.6	0.6
14	1,2—C₆H₄(NO₂)OH	0.278	1.5	1.2	1.2
15	“	0.028	0.8	0.8	0.6
16	1,3—C₆H₄(NO₂)OH	0.278	13.6	20	30
17	“	0.028	3.6	7.6	9.0
18	“	0.00055	1.2	1.1	1.1
19	1,4—C₆H₄(NO₂)OH	0.278	14.3	19.1	25.5
20	“	0.028	1.5	2.2	3.0
21	“	0.0011	0.8	0.8	0.6
22	C₆H₃(NO₂)₂OH(2,4,1)	0.368	21.7	30	30
23	“	0.0037	5.5	6.6	8.8
24	C₆H₃(NO₂)₂OH(2,4,1)	0.00037	2.0	2.5	2.5
25	“	0.000074	0.8	0.7	0.7
26	C₆H₃(NO₂)₂OH(2,6,1)	0.368	21	29	30
27	“	0.0037	5.1	6.0	7.0
28	“	0.00037	1.2	1.4	1.6
29	“	0.000074	0.8	0.7	0.7
30	C₆H₂(NO₂)₃OH(2,4,6,1)	0.0391	25.2	23.6	21.7
31	“	0.0076	2.4	2.2	2.0
32	“	0.000091	1.3	1.2	1.2
33	“	0.0000046	0.8	0.8	0.8

In a non-complex-forming solvent—benzene—phenol and nitrophenols, even at a concentration of 0.5 gram-mole per liter, showed only a very small EMF (Table 50), the values of which likewise did not at all agree with the dissociation constants.

Table 50

No.	Phenols	Observed EMF values, mV, at immersion	Observed EMF values, mV, after 5 min	Observed EMF values, mV, after 10 min	Observed EMF values, mV, after 15 min	Electrolytic dissociation constant
1	\( \mathrm{C_6H_5OH} \)	0.5	0.6	0.6	0.6	\(0.8\cdot10^{-10}\)
2	\(o\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.6	0.6	0.4	0.4	\(5.6\cdot10^{-8}\)
3	\(m\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.95	1.5	1.5	1.2	\(3.9\cdot10^{-9}\)
4	\(p\text{—}\mathrm{C_6H_4(NO_2)OH}\)	2.2	2.6	3.0	3.0	\(5.6\cdot10^{-8}\)
5	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.3	0.35	0.35	0.3	\(1.6\cdot10^{-1}\)

For the same phenols, the values of the specific electrical conductivity in a nitrobenzene solution at \(18^\circ\) were determined (Table 51). The measurements were carried out under the same conditions as the determination of the EMF, i.e., in the presence of sodium, the ions of which gradually increased the electrical conductivity.

Table 51

No. of experiment	Substance studied	Content of substance, percent	Specific electrical conductivity after 2 min	Specific electrical conductivity after 5 min	Specific electrical conductivity after 10 min	Specific electrical conductivity after 16 min
1	\( \mathrm{C_6H_5OH} \)	0.188	0.355	0.415	0.430	0.450
2	\( \mathrm{C_6H_5OH} \)	0.038	\(<0.09\)	\(<0.09\)	0.09	0.09
3	\(p\text{—}\mathrm{C_6H_4(NO)OH}\)	0.246	0.36	0.39	0.44	0.45
4	\(p\text{—}\mathrm{C_6H_4(NO)OH}\)	0.049	0.20	0.26	0.34	0.36
5	\(p\text{—}\mathrm{C_6H_4(NO)OH}\)	0.025	\(<0.09\)	\(<0.09\)	\(<0.09\)	\(<0.09\)
6	\(o\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.278	\(<0.09\)	\(<0.09\)	\(<0.09\)	\(<0.09\)
7	\(o\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.056	\(<0.09\)	\(<0.09\)	\(<0.09\)	\(<0.09\)
8	\(m\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.278	0.285	0.69	1.19	1.20
9	\(m\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.056	0.167	0.570	0.66	0.705
10	\(m\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.028	0.09	0.16	0.25	0.40
11	\(m\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.0056	\(<0.09\)	\(<0.09\)	\(<0.09\)	\(<0.09\)
12	\(p\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.278	0.40	0.505	0.68	0.71
13	\(p\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.056	0.28	0.42	0.50	0.535
14	\(p\text{—}\mathrm{C_6H_4(NO_2)OH}\)	0.028	\(<0.09\)	0.10	0.10	0.10
15	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.046	1.06	1.31	1.53	1.69
16	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.0092	0.87	0.89	0.92	0.99
17	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.0046	0.43	0.485	0.55	0.565
18	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.00092	0.156	0.148	0.145	0.140
19	\( \mathrm{C_6H_2(NO_2)_3OH} \)	0.000042	\(<0.09\)	\(<0.09\)	\(<0.09\)	\(<0.09\)

For the determination of impurities incapable of forming complexes with nitrobenzene that are active with respect to metals (for example, dinitrobenzene), our method is not suitable; in fact, unpurified nitrobenzene may contain impurities of both kinds. There is reason to think that even those by-products which increase the EMF are different in different cases. We compared nitrobenzene preparations obtained by nitration at \(40\text{–}45^\circ\) and \(80\text{–}110^\circ\), and separated into two fractions with boiling points of \(208\text{–}209^\circ\) and \(209\text{–}210^\circ\). In the first preparation the lower-boiling fraction gave \(\mathrm{EMF}=2.6\) mV, the higher-boiling one 2.2 mV; in the second, 2.4 mV and 2.8 mV, respectively, were obtained, i.e., an impurity active with respect to metals distilled over in one case below the nitrobenzene itself, and in the other above it. However, removing these impurities by simple distillation is almost impossible. Thus, to nitrobenzene with \(\mathrm{EMF}=0.8\) mV, para-nitrosophenol was added in an amount of 0.002 mole per liter. The electromotive force rose to 2.73 mV and even after five distillations was equal to 2 mV.

Effect of the Method of Preparing Nitrobenzene

A series of experiments was carried out on the nitration of benzene with a mixture of nitric and sulfuric acids, with identical ratios of the components. For 100 g of thoroughly purified benzene (cryoscopic grade “ch. d. a.”), 120 g of nitric acid of specific gravity 1.456 and 143 g of concentrated sulfuric acid of specific gravity 1.842 were taken. The external reaction conditions were varied: the temperature, the duration of the experiment, and the intensity of stirring (number of revolutions of the stirrer). Purification of the products in all experiments was carried out identically: washing with water, with soda solution, and again with water, drying with calcium chloride, and distillation within the range 208–210°.

Experiments carried out without heating on a water bath (owing to the heat of reaction the temperature rose to 50–55°) and with identical stirring (250–300 revolutions per minute), but with different nitration times, gave the following result (Table 52).

Table 52

No.	Duration of experiment	Yield of \(C_6H_5NO_2\), %	Boiling temperature of fraction studied	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1	10 min.	85.5	208–209°	1.2	1.0	0.9
2	30 min.	91.2	"	1.8	1.6	1.4
3	1 hour	92.7	"	3.2	3.25	3.3
4	2 hours	93.3	"	0.5	0.8	0.5
5	4 hours	92.0	"	1.2	1.0	0.8
6	5 hours	90.7	"	1.4	1.2	1.0

Tables 53 and 54 give the results of experiments carried out with heating on a water bath without shaking, but for different intervals of time, and for the same time (30 minutes), but with different stirring intensities.

Table 53

No.	Duration of heating on water bath	Yield, %	Boiling temperature of fraction studied	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1	5 min.	84.4	208–209°	3.0	2.5	2.5
2	10 min.	86.0	"	3.5	2.8	2.6
3	15 min.	87.8	"	4.8	6.2	6.8
4	20 min.	90.2	"	5.8	8.4	10.8
5	25 min.	91.5	"	5.4	7.8	8.4
6	30 min.	90.8	"	4.6	4.8	5.2
7	60 min.	87.9	"	4.2	4.3	4.5

Table 54

No.	Reaction temperature	Number of stirrer revolutions per minute	Boiling temperature of fraction studied	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1	40–45°	100–150	208–209°	4.2	4.6	4.8
2	40–45°	250–300	"	3.4	3.6	3.2
3	55–60°	100–150	"	4.0	4.4	4.5
4	55–60°	200	"	3.2	3.0	2.7
5	55–60°	300	"	2.0	1.8	1.4

Finally, Table 55 gives EMF values for preparations obtained with identical stirring (200–250 revolutions per minute) and with the same duration of heating, but under different temperature conditions. For the investigation, the fraction of the product with a boiling point of 208–209° was used throughout.

Table 55

No.	Nitration temperature	Yield of C₆H₅NO₂, percent	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1	40–45°	85.8	2.6	2.8	2.6
2	45–50°	90.5	2.5	2.2	1.8
3	50–55°	92.2	1.8	1.4	1.2
4	55–60°	94.7	1.3	1.1	0.8
5	60–65°	96.4	1.2	1.0	0.85
6	75–80°	96.8	1.6	1.3	1.2
7	80–110°	88.2	2.4	2.0	1.8

From the data in these four tables it follows that impurities which increase the EMF of the reaction of nitrobenzene with sodium are formed in particularly large amounts when the reacting substances are insufficiently stirred, or at temperatures below 50° and above 80°. The duration of the experiment also has an effect. Prolonged heating on a water bath, which is usually recommended for achieving complete nitration, has a harmful effect on the quality of the nitrobenzene, and, when stirring is insufficient, also decreases its yield.

Malyarevskii [17], studying the nitration process under plant conditions, also found that good stirring of the reacting mixture is decisive for the purity of the product, while the temperature may be raised at times to 80°. He further indicates that, when these conditions are observed and an excess of benzene is present, the formation of dinitrobenzene is very slight.

Purification of Nitrobenzene

Wishing to obtain for our work a preparation of the best possible quality, and having at our disposal Goslaborreaktiv nitrobenzene with an EMF = 8 mV, i.e., a rather poor one*), we first tried to purify it by simple distillation. We also tried drying with calcium chloride, washing with water, and, finally, washing with a two-normal solution of caustic soda. After each purification the EMF of the reaction with sodium was determined. The results of a series of experiments with this nitrobenzene sample are given in Table 56.

Table 56

No.	Method of treatment	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1	Without treatment	8.0	6.2	4.6
2	Distilled once	4.5	3.8	2.6
3	Dried with CaCl₂ and distilled a second time	2.2	1.8	1.75
4	Stood with CaCl₂ for 24 h, but not distilled	12.6	10.6	9.8
5	Stood with water for 24 h and distilled	1.7	1.2	1.0
6	Stood with water for 24 h, dried and distilled	1.6	1.4	1.7
7	Stood with water for 46 h, dried and distilled	1.2	0.8	0.6
8	Stood with water for 12 days, dried and distilled	2.9	2.2	1.7
9	Stood with water for 6 months, dried and distilled	1.8	1.4	1.2
10	Treatment with NaOH, dried and distilled	0.6	0.55	0.55
11	Repeated the same treatment	0.3	0.3	0.3

*) Preparations obtained later gave smaller EMF values.

As was to be expected, the best results were obtained by washing with alkali. It is interesting that drying with calcium chloride without subsequent distillation increased the EMF. Formation of a complex is possible. By repeated washing with alkali, drying, and distillation we were able to bring all the available samples of nitrobenzene to the standard EMF = 0.3 mv. The most contaminated preparation, of “technical” grade, required four treatments for this (EMF = 13.8 → 8.0 → 2.6 → 0.8 → 0.3 mv); the purified preparation required three (7.2 → 3.0 → 0.7 → 0.3); the pure preparation—two (2.6 → 0.8 → 0.3); likewise the nitrobenzene from Goslaborreaktiv (2.7 → 0.8 → 0.3); for the preparation from the Institute of Chemically Pure Reagents a single purification proved sufficient.

In the work of M. V. Klyuchko, “Physicochemical Analysis of Non-Aqueous Solutions” [7], nitrobenzene was used as the solvent, and the author believed it to be especially well dehydrated by phosphorus anhydride. In the cited work it is noted, however, that the preparation was pale green in color. Assuming the possibility that some compound might form here as well, we determined the EMF of samples of nitrobenzene purified by our method and then kept for a definite time in sealed test tubes over \(P_2O_5\) (Table 57).

Table 57

No.	Drying time	Subsequent treatment	EMF in mv after 0.5 min.	EMF in mv after 5 min.	EMF in mv after 10 min.
1	5 min.	without distillation	10	12	12.5
2	5 min.	distilled	1.5	1.5	2
3	3 hours	”	2	2.2	2.2
4	6 hours	”	2.4	2.5	2.7
5	12 hours	”	6.1	15	16.4

The electrical conductivity of all these preparations remained constant. It is evident that, although phosphorus anhydride is a good desiccant, in the present case its use must be considered undesirable, since nitrobenzene undergoes some change during drying. It is better to treat the preparation with alkali, dry it with calcium chloride, and remove the last traces of water by careful fractional distillation.

Change of Nitrobenzene during Storage

On standing, especially in the light and in the presence of air, nitrobenzene, like many other organic compounds, darkens. We followed the change under various conditions of pure \(C_6H_5NO_2\) (a preparation with EMF = 0.8 mv was taken) and \(C_6H_5NO_2\) containing an admixture of nitro- and nitrosophenols. We also took a preparation dried with phosphorus anhydride. Table 58 shows how the EMF of the preparations studied by us changed after standing for 4 or 6 months.

It is apparent from the table that the purest sample of nitrobenzene changed its properties less during storage than did nitrobenzene containing small amounts of nitrophenols and especially para-nitrosophenol. This is reflected not only in the EMF but also in the color of the preparations. The same was observed for \(C_6H_5NO_2\) prepared at temperatures below 50° and above 80°, or with insufficient stirring. Resistance to the action of light is greatly increased if the substance is stored under water.

From all that has been said about nitrobenzene it follows that, by the magnitude of the EMF arising in the reaction with sodium, it is indeed possible to rapidly and

Table 58

Impurities	Storage conditions	EMF in mV after 0.5 min.	EMF in mV after 5 min.	EMF in mV after 10 min.
1. Without impurities	In a sealed tube, without air, in the light for 6 months	1.35	1.35	1.35
2. "	With a small amount of air in the light for 6 months	1.4	1.35	1.35
3. 0.005% \(o\)-\(NO_2C_6H_4OH\)	Freshly prepared	0.9	0.6	0.6
4. "	After 4 months	1.5	1.7	1.7
5. 0.005% \(m\)-\(NO_2C_6H_4OH\)	Freshly prepared	1.5	2.4	3.0
6. "	After 4 months	13.5	13.8	14
7. 0.005% \(p\)-\(NO_2C_6H_4OH\)	Freshly prepared	1.06	1.06	1.06
8. "	After 4 months	10.9	10	9.7
9. Kept 5 months over \(P_2O_5\)	Freshly prepared	6.5	6	5.4
10. "	After 6 months	9.9	7.6	6.9
11. 0.005% \(p\)-\(NOC_6H_4OH\)	Freshly prepared	1.5	1.8	1.9
12. "	After 4 months	16.2	12.2	11.9

simply determine the purity of this compound. The proposed method made it possible to develop the most advantageous procedure for obtaining the product and for its purification, and it can quite well be applied for production control under factory conditions.

V. Determination of the Purity and Standardization of Phenol

Synthetic and Coal-Tar Phenol

The principal sources for the industrial production of phenol are:

a) Carbolic oil from coal-tar distillation. From this oil the phenols are extracted with alkali and then isolated in the free state by acids.

b) Benzene, from which phenol can be prepared synthetically via sulfonic acid, via aniline, and via chlorobenzene:

1. \(C_6H_6 + H_2SO_4 \rightarrow C_6H_5SO_3H + H_2O;\quad C_6H_5SO_3H + 2KOH \rightarrow C_6H_5OH + H_2O + K_2SO_3.\)

2. \(C_6H_6 \rightarrow C_6H_5NO_2 \rightarrow C_6H_5NH_2 \rightarrow C_6H_5N_2Cl \rightarrow C_6H_5OH.\)

3. \(C_6H_6 \rightarrow C_6H_5Cl \rightarrow C_6H_5OH.\)

The first of these synthetic methods is the one most widely used.

The production of coal-tar phenol is simpler, but the purification of the product is considerably more difficult. If a preparation extracted from coal tar is subjected to the most thorough purification by the usual methods—among which distillation occupies the principal place—we obtain a phenol that is completely colorless and, by all generally accepted constants (above all by melting point), entirely identical with the synthetic product. However, this phenol has a more unpleasant odor and changes color more rapidly on storage. Evidently traces of some impurities still remain in it. Synthetic phenol is prepared from benzene, which in itself can more easily be obtained in a purer form; moreover, in the course of the further conversion of benzene into phenol an additional purification also takes place.

The difference, not detected by the usual methods of testing, between phenol preparations of different origin is quite noticeable in industrial use of the product. Thus, certain kinds of plastics prepared from phenol are obtained more colorless and transparent if a purer phenol is used. From coal-tar phenol, picric acid of poorer quality is obtained, and so on. This compels one to pay atten-

attention to the development of purification methods that would make it possible to prepare, from coal-tar phenol, a product not inferior in quality to the synthetic one, and also to the discovery of more accurate methods for determining purity.*)

Wishing to ascertain how sensitive our method would prove to be to impurities contaminating coal-tar phenol, we decided to compare, in our apparatus, samples of coal-tar phenol of different degrees of purification with a synthetic product obtained from benzene via sulfonic acid and from aniline by diazotization followed by heating in aqueous solution.

The coal-tar phenol preparation taken for study was a light-pink crystalline mass, colored light brown at the bottom. With aqueous caustic soda it gave a clear brownish solution; after a single distillation over ignited sodium sulfate it had the appearance of colorless crystals with a melting point of 39°. The distillation was carried out 5 times, and all the products obtained were tested for EMF.

Phenol from aniline was obtained in the form of colorless crystals. Its purification was carried out by drying over \(\mathrm{Na_2SO_4}\) and by double distillation. The product had a melting point of 39.6°. From the sodium salt of benzenesulfonic acid, phenol was prepared (also colorless crystals) with a melting point of 40°.

For the electrochemical study, nitrobenzene solutions containing 0.0047 phenol in 5 ml were taken, i.e., 0.094% or 0.01 gram-mole per liter. \(\mathrm{C_6H_5NO_2}\), taken as the solvent, was washed twice with caustic soda and, after drying, distilled (b.p. 210.5°). The experiments were carried out in an apparatus with sodium and platinum electrodes. Temperature 16–18°. The results are given in Table 59.

Table 59

No.	Phenol samples	0.5 min.	1 min.	2 min.	5 min.	10 min.
		\multicolumn{5}{c}{Observed EMF values in mV after}
1	Phenol from sulfonic acid	3.9	4.8	6.1	14.8	23
2	Phenol from aniline	4.6	6.1	9.1	18.9	27
3	Coal-tar phenol without distillation	20.2	27.6	29.8	\(>30\)	—
4	The same after 1 distillation	8.7	12.8	17.9	27.3
5	The same after the 2nd	6.3	12.2	16.4	26.1
6	The same after the 3rd	5.4	10.9	13.0	25.7
7	The same after the 4th	5.2	8	9.9	25.5
8	The same after the 5th	5.2	7.9	9.7	25.4

The table shows that, by our method, synthetic phenol can immediately be distinguished from coal-tar phenol. At the same time, in phenol extracted from coal-tar tar, there evidently are impurities that cannot be removed by simple distillation. The fifth distillation no longer changed the EMF value given by the sample taken, and this EMF value considerably exceeded that observed for synthetic phenol. The two synthetic preparations also differed noticeably from one another. The one obtained from sulfonic acid was purer than that prepared from aniline. Perhaps this difference arose from insufficient purity of the aniline taken, and it is possible that during the diazotization reactions themselves and the decomposition of diazo compounds, some by-products appear which cannot be removed by the subsequent purification employed.

*) The work was carried out by B. A. Grandilevskii under the direction of one of us.

Influence of Various Impurities on the Properties of Phenol

A very characteristic feature distinguishing coal-tar phenol from synthetic phenol is that the former becomes colored more rapidly during storage. The immediate cause of the appearance of color is undoubtedly oxidation by air. In this process phenol first passes into hydroquinone and pyrocatechol; these are further oxidized to para- and ortho-quinones, and the latter form brightly colored phenoquinones with phenol.

\[ \begin{gathered} \text{phenol} \longrightarrow \text{hydroquinone} \longrightarrow p\text{-quinone},\\[4pt] \text{phenol} \longrightarrow \text{pyrocatechol} \longrightarrow o\text{-quinone},\\[6pt] \mathrm{C_6H_4O_2 + 2C_6H_5OH \longrightarrow C_6H_4O_2 \cdot 2C_6H_5OH} \end{gathered} \]

\[ \text{phenoquinone }(o \text{ or } p) \]

The fact that oxidation proceeds more rapidly in coal-tar phenol is explained by the influence of impurities. Some authors believe that the decisive role in this respect belongs to metals—for example copper, lead, and iron—which enter either from the factory apparatus or from the glass. According to other investigators, phenol first absorbs ammonia from the air, forming ammonium phenolate; then a metal phenolate is formed, and, with an admixture of hydrogen, the latter gives the coloring substance—pheneitrene. It is possible that the deterioration of phenol depends on compounds of aldehydic character, on indole, and finally on sulfur compounds of the thiophene group, which very possibly also account for the unpleasant odor of coal-tar phenol. Coal-tar phenol may well contain cresols, dihydric phenols, and diphenyl ether. Some of these compounds are also formed as by-products in the synthetic preparation of phenol.

A test was carried out of the influence, on the electromotive force shown by phenol in our apparatus, of adding small amounts of ortho-, meta-, and para-cresols, pyrocatechol, resorcinol, diphenyl ether, as well as phenoquinone present in oxidized phenol, and water. The experiments were conducted under the former conditions (see Table 59), i.e., in nitrobenzene solution, with sodium and platinum as the metals. Samples of phenol were taken from sulfonic acid and of coal-tar phenol after two, three, and four distillations. The results are given in Tables 60 and 61.

Experiments with synthetic phenol

Table 60

Nos.	Impurities	EMF in mV after 0.5 min	1 min	2 min	5 min	10 min
1	Without impurities	3.9	4.9	6.1	14.2	23
2	0.1% H₂O	6.1	7	9.6	21.4	28.5
3	0.1% phenoxynone	7.1	10	12.2	20	27
4	0.1% H₂O + 0.1% phenoxynone	6.9	10.6	20	24.5	>30
5	0.1% o-cresol	6.0	11.2	14.7	23.6	>30
6	0.1% m-cresol	4.4	10.0	13.3	19.8	>30
7	0.1% p-cresol	4.15	5.4	12.1	18	26.8
8	0.1% pyrocatechin	5.0	11.0	14.1	24	>30
9	0.1% resorcinol	5.3	11.4	12.6	28.4	>30
10	0.1% diphenyl ether	5.0	7.0	10.0	16.7	26.7

Experiments with coal-tar phenol

Table 61

Nos.	Treatment and impurities	EMF in mV after 0.5 min	1 min	2 min	5 min	10 min
1	Double distillation, without impurities	6.3	12.2	17.9	26.1	>30
2	0.1% H₂O	8.7	20.6	24.2	28.7
3	0.1% phenoxynone	7.5	18.1	14.5	27
4	0.1% H₂O + 0.1% phenoxynone	9.1	10.9	19.8	27
5	0.1% o-cresol	7.1	10.7	21.4	29
6	0.1% m-cresol	6.9	16.2	20	28
7	0.1% p-cresol	5.2	16	18.4	26
8	0.1% pyrocatechin	6.8	19	21.9	29
9	0.1% resorcinol	7.4	20.3	22.4	29
10	0.1% diphenyl ether	6.0	13	16.3	26.8
11	Triple distillation, without impurities	5.4	10.9	13.0	25.7
12	0.1% H₂O	9.0	18.4	22.7	28.7
13	0.1% phenoxynone	8.4	9.4	12.7	29
14	0.1% H₂O + 0.1% phenoxynone	9.0	10.1	20.1	—
15	0.1% o-cresol	7.0	18.0	19.4	27
16	0.1% m-cresol	6.4	15.1	16.3	26.4
17	0.1% p-cresol	5.4	14.0	16.3	27.0
18	0.1% pyrocatechin	6.1	19.0	20.4	29
19	0.1% resorcinol	7.0	19.4	22	29
20	0.1% diphenyl ether	6.0	12	16	27
21	Fourfold distillation, without impurities	5.2	8.9	9.9	25.4
22	0.1% H₂O	8.4	10.0	12.5	26.6
23	0.1% phenoxynone	7.3	8.3	13.3	23.3
24	0.1% H₂O + 0.1% phenoxynone	9.0	11.0	14.1	20.3
25	0.1% o-cresol	7.0	17.4	19.2	26.4
26	0.1% m-cresol	6.8	15.0	16.9	26
27	0.1% p-cresol	5.4	13.6	16.0	25
28	0.1% pyrocatechin	7.1	19.1	21.6	29
29	0.1% resorcinol	5.9	18.6	20.4	28
30	0.1% diphenyl ether	5.6	12.0	16.2	27

From comparison of the data of the last two tables it follows: 1) that all the substances taken by us as impurities to phenol increased the electromotive force produced by the latter; 2) that coal-tar phenol, with all impurities, always showed a greater EMF than the synthetic phenol.

Influence of storage conditions

To determine the influence on phenol of external factors, namely light and air, the following experiments were carried out with synthetic phenol (from sulfonic acid) and coal-tar phenol of various degrees of purification: 1) phenol samples were stored in sealed ampoules in the dark; 2) storage in the same sealed ampoules, but in the light; 3) storage in open vessels in a desiccator with calcium chloride, in the dark; 4) the same, but in the light.

Each sample was tested on the EMF (under the conditions described above) after 1, 2, 3, and 6 months.

The experiments gave the following result (Table 62).

Table 62

No.	Quality of phenol and storage conditions	0.5 min, 1 mo.	0.5 min, 2 mo.	0.5 min, 3 mo.	0.5 min, 6 mo.	1 min, 1 mo.	1 min, 2 mo.	1 min, 3 mo.	1 min, 6 mo.	2 min, 1 mo.	2 min, 2 mo.	2 min, 3 mo.	2 min, 6 mo.
1	Synthetic phenol, sealed, in the dark	5.0	5.0	5.2	5.4	8.0	8.4	8.9	9.0	9.3	10.0	10.4	10.6
2	Same, open, in the dark	5.2	5.4	6.8	7.0	9.6	9.8	10.0	10.3	12.0	13.0	13.2	13.5
3	Same, sealed, in the light	5.0	6.0	6.0	6.0	8.5	8.6	8.7	9.0	13.0	13.6	13.7	13.8
4	Same, open, in the light	6.0	6.9	7.1	7.2	10.0	9.9	10.7	11.2	15.8	16.2	17.3	17.6
5	Coal-tar phenol after 2 distillations, sealed, in the dark	7.8	8.2	9.4	10	12.6	15.6	16.3	16.5	17.2	18.6	19.1	20
6	Same, open, in the dark	7.9	9.6	10.1	11.3	13	15	17	17.8	17.9	19.9	21.1	21.8
7	Same, sealed, in the light	10.0	12.6	13.8	14.1	14.8	16.9	17.4	18	17.9	20.6	20.9	21.4
8	Same, open, in the light	12.4	14.5	15.2	16.2	18.1	19.2	21	21.9	19.6	20	24	25
9	Coal-tar phenol after 3 distillations, sealed, in the dark	6.6	9.8	10.3	11.3	11.5	15.3	16.0	16.8	13.6	19.2	20	20
10	Same, open, in the dark	7.8	9.9	11	12	12	14.4	17	17.3	14.7	19.3	20	21.3
11	Same, sealed, in the light	7.8	10.9	11.4	12	11.2	16.3	17	17.6	13.9	20	20.4	—
12	Same, open, in the light	11.2	13	15.6	16.3	12.4	16.7	20.6	21.4	14.8	18.4	21.3	24.1
13	Coal-tar phenol after 4 distillations, sealed, in the dark	6.0	9.1	11.6	13	9.1	10.2	15	16.7	15.7	18	19.3	20.6
14	Same, open, in the dark	7.8	10	11.1	13	10.3	12.3	16	16.3	16.8	19.1	20	21
15	Same, sealed, in the light	6.9	10.5	11.4	12	8.7	16	16.3	17	13.5	20	20.6	20.6
16	Same, open, in the light	10.3	11.9	12.7	14	12.1	14.6	18.4	20.1	14.45	20	21	21.4

The color of the synthetic phenol remained unchanged in all experiments. All samples of coal-tar phenol became noticeably colored during storage, and the intensity of the coloration decreased noticeably with an increase in the number of distillations. For example, the samples of experiments 5 and 6, even after 6 months, became only slightly pink.

The data given in the table show that all samples of coal-tar phenol, even that which had been subjected to fourfold distillation, change during storage more strongly than synthetic phenol, although the latter nevertheless markedly increases the EMF it gives even when stored in a sealed ampoule and in the dark. Storage in the light or with access of air intensifies this change.

As for the influence of the individual external factors, comparison of the action of air and light does not make it possible to draw a definite conclusion as to which of them changes phenol more strongly during storage. Together, both act considerably more energetically. In this case the difference between synthetic and coal-tar phenol becomes still more noticeable.

The change during storage under various conditions was also followed for preparations of phenol containing small amounts of water, phenoxyquinone, meta-cresol, partly also with ortho-cresol, para-cresol, pyro-

with catechol and resorcinol. These experiments, like the preceding ones, were carried out with phenol from sulfonic acid and with coal-tar phenol distilled two, three, and four times. Storage was tested in sealed ampoules in the dark and in the light, and in open vessels likewise in the dark and in the light.

The results of these experiments showed that, among the impurities added to synthetic phenol, the emf given by it during storage is increased most by meta-cresol, pyrocatechol, and resorcinol. Water, phenohinone, ortho-cresol, para-cresol, and diphenyl ether act considerably more weakly, although a mixture of water with phenohinone also has a strong effect. Synthetic phenol with water, phenohinone, ortho- and para-cresols, and diphenyl ether increased the emf less than coal-tar phenol even after four distillations. The influence of storage conditions is in general the same as for phenol samples without impurities, i.e., light and air together act more strongly than each of these factors separately.

Observations were also made of changes in the color of the preparations studied. The results, in general, were analogous to those given by determination of the emf.

Purification of Coal-Tar Phenol

For purifying coal-tar phenol from impurities that have a harmful effect on it, reagents of various types were proposed: a) oxidizing agents, for example, chromic mixture; b) reducing agents, such as zinc dust, stannous chloride, hydrogen sulfide; c) condensing catalysts, for example, aluminum chloride, zinc chloride, boron fluoride.

Since the nature of the harmful impurities present in phenol has not yet been reliably elucidated, in choosing a purification method it was necessary to proceed by a purely empirical route. Coal-tar phenol was treated with reagents of different types, sometimes under different conditions. Then the electromotive force given by the products obtained was determined for freshly prepared preparations, and for some of these preparations also after several months of storage.

Treatment with zinc dust and iron filings was carried out by simple distillation over these substances. After the action of acidic reagents (chromic mixture, phosphoric anhydride) and chlorides of aluminum, iron, and tin, the phenol was distilled with soda or treated with caustic soda. Finally, in necessary cases it was dried with ignited sodium sulfate and only then distilled.

The criterion of purity was the magnitude of the electromotive force, determined under the conditions described above, i.e., in nitrobenzene solution with the metals sodium and platinum.

Table 63 gives the results of several preliminary experiments (see p. 76).

The most satisfactory results were given by treatment of coal-tar phenol with chromic mixture. The chromic mixture contained 3 g of potassium dichromate and 4 g of sulfuric acid in 100 ml and was taken in an amount of 5 g per 100 g of phenol. Other ratios gave phenol differing more in emf from synthetic phenol.

Rather good purification was also obtained with ferric chloride. The action of the latter was studied in more detail, with the influence of water (anhydrous FeCl₃ or a 50% solution), the amount of reagent taken, subsequent treatment, and other conditions being followed. Table 64 gives the results of experiments with ferric chloride, aluminum chloride, zinc chloride, and stannous chloride in the anhydrous state and in solution. Table 65 compares data obtained only with anhydrous FeCl₃ under different conditions. In all cases the phenol was taken as coal-tar phenol, double dis—

Table 63

No.	Quality of the preparation or method of purification	Observed EMF values, mV, after 0.5 min	1 min	2 min	5 min	10 min
1	Synthetic phenol	3.9	4.8	6.9	14.9	18.9
2	Coal-tar phenol before purification	8.7	12.8	17.9	27.3	>30
3	Distillation over Zn dust	5.9	10.9	15.8	26.1	30
4	Distillation over Fe	6.1	11.1	16.2	26.3	>30
5	Solid AlCl₃ and NaOH	6.0	10.9	18.3	26.8	30
6	P₂O₅ and NaOH	7.5	9.1	14.2	18.2	29
7	Chromic mixture and NaOH	4.7	5.6	11.6	17.1	22
8	Solid FeCl₃ (1%) and distillation over calcined Na₂CO₃	6.0	7.8	8.9	20.6	27
9	Solid FeCl₃ (2%) and Na₂CO₃, calcined	5.0	6.0	8.7	18	22
10	Solid SnCl₂ and Na₂CO₃	6.4	11	10.6	22.4	25

Table 64

No.	Quality of the preparation or method of purification	Observed EMF values, mV, after 0.5 min	1 min	2 min	5 min	10 min
1	Synthetic phenol	3.9	4.8	6.06	14.3	23
2	Coal-tar phenol without purification	6.3	12.2	16.35	26.1	[[unclear: blank]]
3	2% anhydrous FeCl₃	5.9	7.0	9.6	15.0	24.9
4	2% aqueous solution of FeCl₃ (50%)	9.0	8.9	12.9	25	27
5	2% anhydrous AlCl₃	6.0	10	15	20	30
6	2% aqueous AlCl₃ (50%)	8.4	11.6	18.3	22.3	29.6
7	2% anhydrous ZnCl₂	6.0	7.9	11.2	23.6	29.7
8	2% aqueous ZnCl₂ (50%)	9.0	12.0	19.9	23.0	29.8
9	2% anhydrous SnCl₂	6.5	8.9	11.6	22.6	29.1
10	2% aqueous SnCl₂	8.0	11.0	13	24	29.6

Table 65

No.	Method of purification of coal-tar phenol	Observed EMF values, mV, after 0.5 min	1 min	2 min	5 min	10 min
1	1% anhydrous FeCl₃ at 50° and distillation over Na₂CO₃	6.0	7.5	11.8	20.3	27.4
2	1% anhydrous FeCl₃ at 50° and treatment with NaOH and H₂SO₄, followed by distillation	5.6	7.3	10.9	19.9	26.3
3	2% anhydrous FeCl₃ at 50°, distillation with Na₂CO₃	6.2	7.3	9.6	15	24.9
4	2% anhydrous FeCl₃ at 50°, treatment with NaOH and H₂SO₄	5.9	7.0	9.6	18	25
5	1% anhydrous FeCl₃, secondary distillation with Na₂CO₃	6.0	6.3	9.3	16.9	24.1
6	1% anhydrous FeCl₃, secondary treatment with NaOH and H₂SO₄	5.6	6.1	9.4	17	25
7	2% anhydrous FeCl₃, secondary distillation with Na₂CO₃	6.0	7.2	9.0	15	24
8	2% anhydrous FeCl₃, secondary treatment with NaOH and H₂SO₄	5.3	6.1	9.0	15	24.2
9	2% anhydrous FeCl₃ once at 70° (3 h), distillation with Na₂CO₃ twice	4.1	5.0	8.0	14.3	23
10	2% anhydrous FeCl₃ (70°, 3 h), distillation with Na₂CO₃, then treatment with chromic mixture	4.1	4.6	7.4	14.3	23
11	3% anhydrous FeCl₃ (50°), distillation with Na₂CO₃	7.4	9.6	11.2	24.6	29

Storage in sealed test tubes and in the dark

Table 66

Purification method	Fresh preparation, 0.5 min	Fresh preparation, 1 min	Fresh preparation, 5 min	1 month, 0.5 min	1 month, 1 min	1 month, 5 min	3 months, 0.5 min	3 months, 1 min	3 months, 5 min	6 months, 0.5 min	6 months, 1 min	6 months, 5 min
1. Shaking with zinc dust	5.9	10.9	26	7.3	11	26	7.2	11.7	27	10.2	24	25
2. 1% solution of FeCl₃ and Na₂CO₃	5.9	7.9	22	6.3	8	21	6.3	8	22	6.2	21.3	21.4
3. 2% solution of FeCl₃ and Na₂CO₃	5.7	7.9	18	6.2	7.8	19	6.3	9.3	18	19	21	20
4. 1% anhydrous FeCl₃ and Na₂CO₃	6.0	7.5	20	7.0	8.0	21	7.5	8.2	28	8	9.3	18.6
5. 1% anhydrous FeCl₃, NaOH, H₂SO₄	5.6	7.2	20	6.0	8.0	20	7.2	8.0	18	7.6	8.6	19
6. 2% anhydrous FeCl₃ and Na₂CO₃	6.2	7.3	18.6	6.5	8.0	20	7.0	8.0	19	6.9	10	18
7. 2% anhydrous FeCl₃ and twice treated with Na₂CO₃	4.1	5.1	14.3	4.3	5.0	15	5.0	6.0	15	5.1	7	17
8. 2% anhydrous FeCl₃, Na₂CO₃, chromium mixture	4.1	4.6	14.2	4.2	5.0	15	4.7	5.9	15.2	5.2	7	17

Storage in sealed test tubes in the light

Table 67

Purification method	Fresh preparation, 0.5 min	Fresh preparation, 1 min	Fresh preparation, 5 min	1 month, 0.5 min	1 month, 1 min	1 month, 5 min	3 months, 0.5 min	3 months, 1 min	3 months, 5 min	6 months, 0.5 min	6 months, 1 min	6 months, 5 min
1. Shaking with zinc dust	5.9	10.9	26	6.9	13.8	27	7.1	11.2	28	7.3	12	28
2. 1% solution of FeCl₃, Na₂CO₃	6.1	7.9	22	7.2	8.6	24	7.9	8.9	24	8.9	24	24
3. 2% solution of FeCl₃, Na₂CO₃	5.7	7.9	18	6.2	8.3	20	6.9	9.3	25	7.4	10.2	25
4. 1% anhydrous FeCl₃, Na₂CO₃	6.0	7.5	21	7.1	8.6	21	7.6	9	21	8.0	21.5	22
5. 1% anhydrous FeCl₃, NaOH, H₂SO₄	5.6	7.2	20	5.9	8.1	20	7.2	10	21	8.1	21	26
6. 2% anhydrous FeCl₃, Na₂CO₃	6.2	7.3	19	6.6	7.8	19	7.1	8.3	20	7.4	8.6	20
7. 2% anhydrous FeCl₃, twice Na₂CO₃	4.1	5.0	14.3	5.3	6.2	15	6.0	7.2	16	6.2	7.3	17
8. 2% anhydrous FeCl₃, Na₂CO₃, chromium mixture	4.1	4.6	14.2	5.2	6.4	15	5.6	7.0	17	5.9	7.3	19

Table 68

Storage in air in the dark

Purification method	Fresh preparation, 0.5 min	Fresh preparation, 1 min	Fresh preparation, 5 min	1 month, 0.5 min	1 month, 1 min	1 month, 5 min	3 months, 0.5 min	3 months, 1 min	3 months, 5 min	6 months, 0.5 min	6 months, 1 min	6 months, 5 min
1. Recrystallization with zinc dust	5.9 mV	7.9	22	7.3	11.2	27	8.0	11.3	27	10.4	24	28
2. 1% solution of FeCl₃, Na₂CO₃	5.9	10.9	26	6.5	9.0	21	7.0	8.2	23	7.8	13.8	25
3. 2% solution of FeCl₃, Na₂CO₃	5.7	7.9	18	6.0	7.0	20	6.8	9.4	21	7.2	9.8	21
4. 1% anhydrous FeCl₃, Na₂CO₃	6.0	7.5	20	6.0	6.4	18	7.9	12.6	21	7.0	9.7	21
5. 1% anhydrous FeCl₃, NaOH, H₂SO₄	5.6	7.2	20	6.4	8.1	21	7.4	9	19	7.8	8.9	18
6. 2% anhydrous FeCl₃, Na₂CO₃	6.2	7.3	19	6.5	8.3	21	7.5	9.2	19	7.9	8.7	19
7. 2% anhydrous FeCl₃, twice Na₂CO₃	4.1	5.0	14.3	4.2	5.2	15	6.0	7.9	15	5.2	7.0	17
8. 2% anhydrous FeCl₃, Na₂CO₃, chromium mixture	4.1	4.6	14.2	4.3	5.0	16	4.8	6.2	16	6.0	8.3	19

Table 69

Storage in air and in the light

Purification method	Fresh product, 0.5 min	Fresh product, 1 min	Fresh product, 5 min	1 month, 0.5 min	1 month, 1 min	1 month, 5 min	3 months, 0.5 min	3 months, 1 min	3 months, 5 min	6 months, 0.5 min	6 months, 1 min	6 months, 5 min
1. Recrystallization with zinc dust	5.9 mV	7.9	22	6.9	9.0	24	8.1	9.1	25	8.3	10	25
2. 1% solution of FeCl₃, Na₂CO₃	5.9	10.9	26	6.9	12	28	8.2	11.3	28	9.9	13.6	28
3. 2% solution of FeCl₃, Na₂CO₃	5.7	7.9	18	6.2	8.2	21	6.4	9.7	25	6.9	11.6	27
4. 1% anhydrous FeCl₃, Na₂CO₃	6.0	7.5	20	7.4	8.3	21	7.9	8.9	25	10	12	26
5. 1% anhydrous FeCl₃, NaOH, H₂SO₄	5.6	7.2	20	6.3	8.6	21	8.0	8.9	22	8.6	11	24
6. 2% anhydrous FeCl₃, Na₂CO₃	6.2	7.3	19	6.9	8.4	21	8.0	16	24	8.0	13	22
7. 2% anhydrous FeCl₃, twice Na₂CO₃	4.1	5.0	14	5.4	6.9	15	5.9	8.0	15	6.3	8.2	18
8. 2% anhydrous FeCl₃, Na₂CO₃, chromium mixture	4.1	4.6	14	5.6	7.2	15	6.0	8.0	16	7.9	10	17

distillation. When comparing the action of anhydrous and aqueous reagents, purification was continued for 2 hours at 50°, and then the phenol was distilled over anhydrous Na₂CO₃.

Since a very important criterion of the purity of phenol is the preservation of its properties during prolonged storage, it was necessary to carry out appropriate experiments with preparations of coal-tar phenol purified with one or another reagent. In this respect, the preparations purified with ferric chloride were studied primarily, but one more sample of phenol, distilled over zinc dust, was also taken. This sample was of interest because it retained its color very well even over several months.

The preparations under study were stored for up to 6 months under various conditions, namely in sealed test tubes and in open vessels, in the dark and in the light. The electromotive force was determined in the same way as in all the preceding cases, i.e., in the same apparatus with sodium and platinum, in a nitrobenzene solution of the same concentration.

The results are given in Tables 66, 67, 68, 69.

As the figures given in the last tables show, preparations with the lowest EMF and with little change during storage are obtained by combined purification of coal-tar phenol: with anhydrous ferric chloride, soda, and chromic mixture. Ferric chloride is a weak oxidizing agent and, in addition, a reagent with a condensing action. The chromic mixture acts as an oxidizing agent. Soda removes acids, for example the hydrogen chloride liberated during treatment with FeCl₃. In terms of the EMF for fresh samples and the change in EMF upon storage under different conditions, the preparations obtained are almost indistinguishable from synthetic phenol from sulfonic acid. Such preparations also retain their color well. Purification with zinc dust effectively delays the appearance of coloration, but gives phenol that differs sharply from the synthetic product in EMF.

Thus, for phenol, just as was shown for nitrobenzene, our method is also a good means for rapidly determining the purity of the product. With the aid of this control method it was easy to choose among the many methods proposed for purifying coal-tar phenol. The developed procedure, using inexpensive reagents and very small amounts of them, certainly deserves attention for trial in production.

VI. Study of Reaction Products

The high sensitivity of our electrochemical method led us to expect that it would prove useful in studying mixtures formed in certain reactions of organic chemistry, especially oxidation reactions. Oxidative processes very often lead to the formation of acids, and acids in particular, even in very small quantities, can be readily individualized by our method.

So far, three reactions have been studied in this respect:

1) Oxidation of benzene with a solution of manganese anhydride in carbon tetrachloride.* The reaction proceeded vigorously. As a result, the greater part of the benzene consumed was oxidized to CO₂. In addition, a very small quantity (about 0.2 g) of colorless crystals was obtained, which were presumably some carbonic acid, or more probably a mixture of acids (as indicated by a very indefinite melting point). There were too few of these crystals

* Done by one of us jointly with Lomakina.

too little for investigation by ordinary methods. Therefore we decided to apply our method, which makes it possible to operate with negligible portions of a substance.

In the oxidation of benzene one might have expected the formation of the following acids:

Oxalic

$$ \mathrm{COOH{-}COOH}. $$

Malonic

$$ \mathrm{COOH{-}CH_2{-}COOH}. $$

Maleic

$$ \begin{aligned} \mathrm{H{-}C{-}COOH}\\ \mathrm{\ \ \Vert}\\ \mathrm{H{-}C{-}COOH}. \end{aligned} $$

Fumaric

$$ \begin{aligned} \mathrm{H{-}C{-}COOH}\\ \mathrm{\ \ \Vert}\\ \mathrm{HOOC{-}C{-}H} \end{aligned} $$

Muconic

$$ \mathrm{COOH{-}CH{=}CH{-}CH{=}CH{-}COOH}. $$

The presence of significant quantities of the last two acids was excluded, since the crystals dissolved well in water. Of the remaining acids, oxalic acid sublimes most readily. Therefore the mixture was placed in a long, thin-walled glass tube, open at one end, and immersed in a small beaker with glycerin, previously heated to \(100\text{--}105^\circ\).

After colorless crystals had appeared on the cold walls of the tube, the tube was sealed off in the flame, and its sealed end was heated for about another two hours at \(150^\circ\).

The crystals subsequently collected from the walls of the tubes were examined separately. It was noted that the sublimed substance was distinguished by good solubility in water. The crystals from the sealed part dissolved considerably worse than before heating. This could have depended on the conversion of maleic acid into fumaric acid.

Next, both portions of crystals were dissolved in a mixture of acetone and water (1:5 by volume; concentration of the substance under investigation in the solution, 0.1%) and tested in our apparatus. The results are given in Table 70, where the EMF values are also given for the pure acids oxalic, maleic, and fumaric.

Table 70

Nos.	Substance tested	Fe	Cu	Zn	Sn	Pb
1.	Crystals studied before heating	225	40	385	200	150
2.	Sublimed crystals	240	60	400	250	186
3.	Remaining crystals before washing with water	175	40	325	50	125
4.	The same crystals after washing with water	150	25	310	5.0	110
5.	Oxalic acid	250	65	400	255	175
6.	Maleic acid	170	25	350	5.8	140
7.	Fumaric acid	150	20	300	3.8	110

Comparison of the figures presented in the table clearly shows that the crystals sublimed in the tube represented almost pure oxalic acid; what remained in the tube (after heating) was fumaric acid with an admixture of maleic acid, from which it was purified by washing with water. Finally, the mixture originally obtained consisted chiefly of

of maleic and oxalic acids, i.e., the general course of the oxidation of benzene by manganese anhydride is as follows:

\[ \begin{gathered} \text{benzene} \;\longrightarrow\; \mathrm{HOOC{-}CH{=}CH{-}COOH} + \mathrm{HOOC{-}COOH}. \end{gathered} \]

2) Oxidation of ethylene by potassium permanganate (44).
3) Oxidation of acetylene by potassium permanganate (44).

In both experiments a 0.1% solution of \(\mathrm{KMnO_4}\) in water was taken, through which ethylene and acetylene were passed until the permanganate color had completely disappeared. The brown precipitate of manganese dioxide was filtered off, and the filtrate was examined by the EDS method with platinum and several active metals.

Table 71

Test liquid	Mg	Al	Fe	Cu	Zn	Cd	Sn	Pb
1. \(\mathrm{KMnO_4}\) solution	120	2.5	0.5	0.2	40	10	0.5	40
2. Solution after passing \(\mathrm{C_2H_4}\)	120	2	0.4	0.2	40	8	0.4	36
3. Solution after passing \(\mathrm{C_2H_2}\)	267	12	5	6	180	45	7	65

Observed EDS in mV with metals.

In the oxidation of ethylene, not only is there no increase in the EDS, but, on the contrary, for some metals it even becomes somewhat smaller. This can be explained by the fact that, under these conditions, oxidation proceeds only as far as the formation of glycol:

\[ \begin{gathered} \mathrm{CH_2}\\[-2mm] \Vert\\[-2mm] \mathrm{CH_2} \end{gathered} \;\longrightarrow\; \begin{gathered} \mathrm{CH_2OH}\\[-2mm] \vert\\[-2mm] \mathrm{CH_2OH} \end{gathered} \]

After acetylene was passed through, a sharp increase in the EDS was observed for all metals. Assuming that formic acid could have formed here—of course, in the form of the potassium salt—

\[ \begin{gathered} \mathrm{CH}\\[-2mm] \Vert\!\Vert\\[-2mm] \mathrm{CH} \end{gathered} + 2\,\mathrm{KMnO_4} \longrightarrow 2\,\mathrm{HCOOK} + 2\,\mathrm{MnO_2}, \]

the following experiment was carried out: the theoretically calculated amount of potassium formate (on the assumption that the reaction proceeds completely in the indicated direction) was dissolved in water, and the solution was tested for EDS.

Table 72

Test substance	Mg	Al	Fe	Cu	Zn	Cd	Sn	Pb
\(\mathrm{HCOOK}\)	255	11	5	4	180	45	5	64

Observed EDS in mV with metals.

The agreement is quite good. Thus, in the simplest cases, even direct determination of the EDS for the solution obtained upon oxidation says much about the direction.

process. Sometimes it is necessary to subject the obtained product to some additional treatment and only then use our method; moreover, the application of this method makes it possible to shorten the time considerably, simplifies the very procedure of investigation, and makes it possible to determine the direction of the reaction with a very small amount of substance. In the latter respect, the method we propose should be placed in the same category as other methods of microchemical analysis.

VII. Some applications of the new electrochemical method in inorganic chemistry

4) Determination of the purity of water

The problem of rapidly determining the purity of drinking water, water for various technical needs, etc., is of great interest both to public health and to all branches of the national economy. In the literature of the past few years alone one can find a whole series of works aimed at solving this problem. Thus, it has been proposed to distinguish boiled water from raw water by the drop method, using free and bound carbon dioxide [6], or by titration with hydrochloric acid in the presence of rosolic acid as an indicator [3]. The alkalinity of river water is also determined by electrical conductivity, passing current from a Daniell cell through a tube with the liquid under test and a galvanometer [1]. By the same electrical conductivity, using a lamp bridge, it has been proposed to monitor the acidity of factory wastewater [2].

Our electrochemical method is simpler, since here, instead of an external galvanic cell, a cell is used into which the liquid under investigation itself has been poured.

The method should also be more sensitive: the EMF of the cell we use also depends on the purity of the water being investigated.

Preliminary experiments carried out with samples of raw water (tap water), water heated to 80°, and boiled water fully confirmed these expectations.

Table 73

Water samples	Fe	Cu	Zn	Sn	Pb
1. Raw water	80	15	200	25	40
2. The same water, heated to 80°	60	10	120	20	35
3. Boiled water	30	8	80	15	30

Observed EMF (at 20°), mV, with metals: Fe, Cu, Zn, Sn, Pb.

In Table 74 we present a comparison of the results of determinations in our apparatus with those obtained by measuring electrical conductivity in the usual way, with the aid of a Wheatstone bridge. The experiments were carried out with preparations of distilled water of different degrees of purification.

With the most active of the metals taken—zinc—the two preparations of water at the extremes of purity showed a sixfold difference, while in electrical conductivity they differed by less than a factor of 3. The use of a copper or lead electrode gave a very large difference. It should also be noted, in addition, that there is a difference in the behavior of individual metals toward various impurities that may be present in water. For example, with zinc, water purified by a single distillation with permanganate gave a larger EMF value than

Table 74

Water samples	Fe	Cu	Zn	Sn	Pb	Electrical conductivity
1. Ordinary distilled water	4	3	12	5	7	\(9.43 \cdot 10^{-6}\)
2. The same water after a single distillation	2	0.4	7	2.5	1	\(5.56 \cdot 10^{-6}\)
3. The same after a second distillation	1.5	0.3	4	1.5	0	\(4 \cdot 10^{-6}\)
4. The same after a third distillation	1.5	0.4	4	2	0	\(4 \cdot 10^{-6}\)
5. Ordinary distilled water, purified by a single distillation with \(\mathrm{KMnO_4}\)	2	0.2	5	2	0	\(4 \cdot 10^{-6}\)
6. The same water after a second distillation with \(\mathrm{KMnO_4}\)	1.5	0	2	1.5	0	\(3.3 \cdot 10^{-6}\)

water purified simply by double or triple distillation, while with copper it is the reverse. This difference, of course, can be used to determine the nature of the impurities present in the water. It must be supposed that in this way it will be possible, for example, to distinguish alkaline water from acidic water, perhaps to determine the presence of ammonia or decomposing organic residues, etc. One of us has constructed a portable pocket instrument with which the quality of water can be determined under any conditions—in laboratories, pharmacies, hospitals, and the testing of drinking water can be carried out during field work, etc.

Study of Salt Solutions

a) Study of salts of carboxylic acids [44]. So far, by our method, only salts of acetic acid with various cations have been studied. Solutions of salts in distilled water were taken. The concentration was 0.1%, the temperature 18–20°, and the apparatus was the same as that used in the experiments with free acids (in all cases the second electrode was a platinum crucible).

The results of the measurements are given in Table 75, where the EMF values are also given for acetic acid of the same concentration and for the water that served as solvent.

Table 75

Compounds studied	Fe	Cu	Zn	Sn	Pb
1. \(\mathrm{H_2O}\)	4	3	12	5	7
2. \(\mathrm{CH_3COOH}\)	60	5	115	15	50
3. \(\mathrm{CH_3COONa}\)	1	0	160	3	55
4. \(\mathrm{CH_3COOK}\)	0	0	110	1	70
5. \(\mathrm{CH_3COONH_4}\)	100	2	193	30	75
6. \((\mathrm{CH_3COO})_2\mathrm{Mg}\)	1	0.8	145	6	55
7. \((\mathrm{CH_3COO})_2\mathrm{Ca}\)	1	0.3	185	5	50
8. \((\mathrm{CH_3COO})_2\mathrm{Co}\)	0	0	105	1	25
9. \((\mathrm{CH_3COO})_2\mathrm{Cu}\)	2	1	300	5	99
10. \((\mathrm{CH_3COO})_2\mathrm{Zn}\)	1	1	110	2	36
11. \((\mathrm{CH_3COO})_2\mathrm{Ba}\)	1	1	155	2	60
12. \((\mathrm{CH_3COO})_2\mathrm{Pb}\)	1	1	170	5	55
13. \((\mathrm{CH_3COO})_3\mathrm{Cr}\)	0.5	0	20	4	11

As the figures obtained show, it is not yet possible to detect any regularity in the dependence of the EMF values observed in our instrument on the nature and properties of the compounds taken for study. The elec-

the conductivity of the solution in any case is not of decisive importance here. Water has an electrical conductivity many times lower than that of solutions, for example, of sodium or potassium acetate, and the value of the EMF for zinc and lead shown by the voltmeter of our instrument is many times greater in solutions of these salts than in pure water; but with iron, copper, and tin, on the contrary, we have a sharp decrease in electromotive force for the salts as compared with water. Displacement by a metal of hydrogen or of another metal also does not play the leading role. Zinc can displace hydrogen from acetic acid, but it cannot displace sodium or calcium from their acetates, whereas the EMF for Zn is greater with CH₃COONa and (CH₃COO)₂Ca than with CH₃COOH, etc.

The processes occurring in these experiments are undoubtedly complex. Additional investigations are necessary in order to clarify them. However, it can definitely be said that sharp fluctuations in EMF values with different electrode metals and different salt cations may find application in analytical practice. Thus, with ammonium acetate, as also with acetic acid, iron gave a considerably greater EMF than tin, and somewhat greater than lead; with salts of potassium, sodium, calcium, magnesium, zinc, copper (although copper is readily displaced from its salts by it), lead, barium, cobalt, and chromium—iron everywhere proves to be much less active than lead, somewhat less active than tin, and sometimes even is comparable in EMF magnitude with the least active metal, copper. Tin with ammonium acetate gives an EMF less than 3 times lower than lead, while with potassium acetate—70 times lower. In pure water these two metals have almost identical activity.

It is of interest to study, by this same method, salts of other acids, both organic and mineral, and moreover with a large series of cations, in various solvents and at different concentrations. What has been done in this vast field so far must be regarded only as preliminary experiments, showing that these investigations deserve attention both theoretically and practically.

b) Determination of small concentrations of electrolytes. In two papers, B. A. Rashkovan (25) describes experiments applying essentially the same method as ours to the determination of small concentrations of certain inorganic salts. The apparatus constructed by the author of these articles consisted of the following parts: 1) electrodes—mercury and amalgamated copper; 2) a weak aqueous solution of the salt under investigation poured between them; 3) a galvanometer of low resistance. The experiments were conducted first of all with solutions of silver nitrate at concentrations from 0.0002 N to 0.0043 N. The results obtained were quite well expressed by the equation:

\[ \lg y = a \lg x + C, \]

where \(y\) is the galvanometer reading, \(x\) is the normality of the solution, and \(a\) and \(C\) are constants for the given conditions. Deviations ranged from \(-5\%\) to \(+5\%\). The author finds that by this method one can determine concentrations down to 0.00001 N, and perhaps even lower. The same procedure was applied to nitrates of sodium, potassium, ammonium, calcium, strontium, barium, copper, zinc, aluminum, and iron. It was found that salts of iron and copper deviate from the general rule.

Analysis of alloys and determination of the purity of metals

If one takes an alloy of two metals sharply differing from each other in the electromotive force observed in our instrument (upon contact with some acid or another suitable reagent), then

one may expect to determine the composition of an alloy by this method. In this direction, experiments have so far been carried out only with the most readily available alloys of tin with lead [44].

The alloys taken were prepared from the purest available preparations of the indicated metals (“Kahlbaum” grade; analytical data: tin 99.9%, lead 99.99%). The metals were taken in definite ratios, melted in a porcelain crucible, and a glass tube with an internal diameter of 4 mm was immersed in the liquid. By means of a rubber bulb, a certain amount of the liquid alloy was drawn into the tube, and the latter, while still hot, was immersed in water. The metal rods obtained in this way were homogeneous and had a mirror-smooth surface. These rods were immersed in an aqueous or water-acetone solution of the acid, poured into the platinum crucible of our apparatus. The experimental conditions were exactly the same as in the study of carboxylic acids with the use of pure metals: the same galvanometer, acid concentration 0.1%, room temperature.

With alloys, 6 monocarboxylic and 5 dicarboxylic acids have so far been tested. The results are given in Tables 76 and 77; in the first of them are placed the EMF values obtained with alloys differing from one another in composition successively by 10% of one or the other metal. In the second table, the figures relating to alloys containing 90 percent or more of tin or lead are singled out.

Table 76

Medium	Acids	Sn, %	Pb, %	100 / 0	90 / 10	80 / 20	70 / 30	60 / 40	50 / 50	40 / 60	30 / 70	20 / 80	10 / 90	0 / 100
Water	HCOOH	Sn	Pb	35	45	53	58	60	75	75	80	90	90	0
Water	CH₃·COOH	Sn	Pb	15	40	50	55	55	55	60	60	65	65	50
Water	CH₃·CH₂·COOH	Sn	Pb	14	10	30	30	30	30	30	35	45	45	45
Water	CH₃·(CH₂)₂·COOH	Sn	Pb	13	12	25	45	50	50	50	50	55	55	45
Water	(CH₃)₂CH·COOH	Sn	Pb	14	7	18	30	40	45	45	45	50	50	45
Water	CH₃·(CH₂)₃·COOH	Sn	Pb	10	9	22	33	35	40	40	45	50	50	50
Water	COOH·COOH	Sn	Pb	135	120	120	120	120	120	110	110	110	110	175
Water	Tartaric	Sn	Pb	12	20	30	43	50	55	55	55	55	55	60
Water	Maleic	Sn	Pb	25	13	45	60	90	95	95	95	95	95	140
Water	Fumaric	Sn	Pb	7	5	10	20	45	90	90	90	105	110	110
Water	Phthalic	Sn	Pb	1.7	3	8	15	43	73	75	70	90	90	95
Water-acetone	HCOOH	Sn	Pb	5	18	20	20	35	40	40	40	45	50	50
Water-acetone	CH₃·COOH	Sn	Pb	2	4	9	11	12	15	15	18	18	18	35
Water-acetone	CH₃·CH₂·COOH	Sn	Pb	4	4	10	12	14	18	19	23	23	23	24
Water-acetone	CH₃·(CH₂)₂·COOH	Sn	Pb	2	4	9	10	13	15	15	16	18	18	20
Water-acetone	(CH₃)₂CH·COOH	Sn	Pb	5	4	9	11	13	15	16	17	20	21	20
Water-acetone	CH₃·(CH₂)₃·COOH	Sn	Pb	2	2	5	7	8	10	10	13	13	14	16
Water-acetone	COOH·COOH	Sn	Pb	1.5	95	95	95	95	95	95	90	80	65	85
Water-acetone	Tartaric	Sn	Pb	8	6	14	17	17	23	35	35	40	45	45
Water-acetone	Maleic	Sn	Pb	2	9	55	55	55	65	65	65	70	70	60
Water-acetone	Fumaric	Sn	Pb	1	10	23	43	43	50	50	50	55	60	60
Water-acetone	Phthalic	Sn	Pb	1.5	5	15	38	40	45	45	49	55	55	55

The figures in these tables show that the EMF values recorded by our apparatus, as the composition of the alloy changes, sometimes change without any definite regularity. In some cases the alloys gave a greater or smaller EMF than both of the metals taken in pure form. Therefore, for each type of alloy the method of determination must first be developed on specimens of known composition. It is necessary to establish which electrode is best to use, in what solvent, and at what concentration.

Table 77

Solvent	Acids	Sn/Pb 100/0	99/1	98/2	95/5	90/10	10/90	5/95	2/98	1/99	0/100
Water	H.COOH	35	30	35	45	45	90	95	95	95	100
Water	CH₃.COOH	15	40	40	40	40	65	65	65	65	66
Water	CH₃.CH₂.COOH	14	10	10	10	10	45	45	45	45	45
Water	CH₃.(CH₂)₂.COOH	13	7	8	12	12	55	55	55	55	45
Water	(CH₃)₂CH.COOH	14	4	5	7	7	50	50	50	50	45
Water	CH₃.(CH₂)₃.COOH	10	6	6	8	9	50	50	50	50	50
Water	COOH.COOH	135	120	120	120	120	110	110	105	100	98
Water	Succinic	12	8	8	8	8	55	55	55	55	60
Water + acetone	H.COOH	5	8	10	15	18	50	55	60	60	50
Water + acetone	CH₃.COOH	2	4	4	4	4	18	35	35	35	25
Water + acetone	CH₃.CH₂.COOH	4	3	3	4	4	23	24	24	24	24
Water + acetone	CH₃(CH₂)₂.COOH	2	4	4	4	4	18	40	40	40	20
Water + acetone	(CH₃)₂CH.COOH	5	3	3	3	4	21	40	40	40	20
Water + acetone	CH₃.(CH₂)₃.COOH	2	2	2	2	2	14	16	16	16	16
Water + acetone	COOH.COOH	95	95	95	95	95	65	67	67	67	68
Water + acetone	Succinic	8	2	2	3	6	45	45	45	45	45
Water + acetone	Maleic	2	2	2	2	9	70	80	80	80	60
Water + acetone	Fumaric	1	2	3	3	10	60	60	60	60	60
Water + acetone	Phthalic	1	1	2	3	3	55	55	55	55	55

Having carried out this preliminary work, we shall have a very simple and rapid method of analysis, entirely suitable for use in production laboratories.

Thus, the canning industry is interested in the rapid determination of lead impurity in tin. By our method this determination takes only a few minutes, and the sensitivity of the method is no less than that of ordinary analytical methods.

What has been said applies, of course, only to binary alloys. An impurity of a third metal will greatly complicate the picture. However, in cases where there is no need to determine the composition of the alloy or the nature of the impurities contained in the metal sample taken, but it is necessary only to know the degree of its purity, the problem can be solved very simply. As one electrode—conveniently in the form of a crucible or cylinder—we take a sample of metal of standard purity; the other electrode we prepare from the metal under investigation. We select the electrolyte so that it is most sensitive to the impurities ordinarily occurring in the metal at the given enterprise. If the EMF observed under these conditions is very small (less than the established standard value), then the sample is sufficiently pure. Experiments carried out with electrolytic and technical copper [16] showed that copper electrodes of different quality give a quite appreciable EMF in our apparatus. Even in such a weakly active electrolyte as butyl alcohol, the EMF was 0.65 mV.

It may be hoped that by our method it will be possible easily and rapidly to determine the resistance of metals to corrosion. It is also of interest to trace the influence on the EMF of physical and mechanical treatment of the metal: hardening, forging, rolling, etc.

Conclusions

1. The authors have developed a new electrochemical method for determining the activity of organic compounds with respect to metals. The method is based on measuring the current strength in a closed circuit, the source of energy in which is the reaction of the compound under study with the metal.

1. The proposed method gives good results in determining the structure of many organic substances. Thus, in the reaction of alcohols with sodium, primary, secondary, and tertiary alcohols differ sharply from one another in the readings of the galvanometer in our apparatus. A good parallelism is observed with what is given by determining the reaction rate of the same alcohols with sodium amalgam, carboxylic acids, their anhydrides, acid chlorides, etc. The parallelism proved to be considerably better than in comparison with the rate constants of the values of the potential difference upon contact of alcohols with sodium amalgam.

In carboxylic acids it is possible (with a suitable selection of metals) to determine the number of carboxyl groups, the presence of side chains, alcoholic groups, etc. It is interesting that acids in pure form or in benzene solution showed (by our method) a lower activity toward metallic sodium than did primary alcohols with the same skeleton.

Sharp differences were noted in the activity of carbonyl compounds depending on the composition and structure of the molecule (aldehydes, ketones and esters, enolizable compounds and compounds incapable of enolization). Enolizable and non-enolizable nitro compounds and nitriles, fatty or aromatic amines, and pyridine bases behaved quite differently. Thus, the method is also suitable for determining the structure of certain compounds with a double or triple bond.

1. The current strength indicated by the galvanometer in our apparatus for certain compounds with active hydrogen increases sharply upon addition of aldehydes, ketones, esters, nitro compounds, and amines. Among the compounds of the first group, carboxylic acids and phenols give this effect to a greater degree than alcohols. An increase in current strength was sometimes also observed upon mixing substances that do not have active hydrogen, for example, quinone with aromatic amines or nitro compounds. The phenomenon described can be explained only by complex formation.

2. It has been shown in several examples that the maxima of the current strength correspond to the composition of molecular compounds. Some of the complexes formed (among them 5 new ones) were successfully isolated in crystalline form. More than 40 new compounds were detected but not isolated.

3. It has been found that this method makes it possible to find molecular compounds even where thermal analysis and determination of electrical conductivity do not give a positive result.

4. Application of different pairs of metals (sodium–platinum, zinc–copper) to complex-forming systems gave maxima for mixtures of identical composition, which makes this method more reliable.

5. The method proposed by us also gives good results in determining the purity of organic compounds and may be recommended as a rapid method for monitoring the production of products of the organic chemical industry. Determination of purity takes only a few minutes and can be carried out anywhere. The apparatus is extremely simple and portable, and working with it requires neither heating nor external electrical power.

6. The same method can be applied in the study and technical testing of inorganic substances. Thus, the purity of water can be determined no less rapidly, simply, and accurately than by electrical conductivity. It is possible to analyze certain alloys, determine the purity of metals, etc.

Literature

1. Vainshtein G. M.—Remote determination of river water from galvanometer readings. Water Supply and Sanitary Engineering, No. 8–9, p. 57, 1938.

2. Gel'grev E. Kh. and Larionov Yu. A.—Control of waste and wash waters by means of a lamp voltmeter. Zavodskaya laboratoriya, 9, No. 2, p. 191, 1940.

3. Gol'dshtein M. V.—A method for distinguishing boiled water from raw water. Laboratory Practice, No. 2–3, p. 26, 1939.

4. Zagrebin V.—Rate constants for the formation of simple ethers. Journal of the Russian Physico-Chemical Society—Chemical Section, 31, p. 19, 1899.

5. Igonin L. A.—Thermodynamics of equilibrium and the rate of the chemical reaction of alcohols with sodium amalgam. Dissertation defended before the Council of Tomsk State University, 1940.

6. Iokhel'son D. V.—A drop method for distinguishing boiled water from raw water. Hygiene and Sanitation, No. 2, p. 35, 1937.

7. Klochko M. A.—Physicochemical analysis of nonaqueous solutions. Proceedings of the Academy of Sciences, chemical series, No. 3, p. 664, 1937.

8. Kondyrev N. V. and Yur'ev V. I.—On the potentials of certain metals in ether solutions. $MgJ_2$. ZhOKh, 7, p. 2903, 1937.

9. Konovalov D. P.—On the electrical conductivity of solutions. Journal of the Russian Physico-Chemical Society—Chemical Section, 24, 440, 1892; 25, 192, 1893.

10. The same author—Thermal phenomena accompanying the mixing of amines with acids. Journal of the Russian Physico-Chemical Society—Chemical Section, 25, 5, 211, 1893.

11. Kulev L. P.—On the methodology of analysis of carbonic acids. Proceedings of the Tomsk Industrial Institute, 59, No. 3, p. 103, 1942.

12. The same author—On the nature of complexes of phenols with amines. ZhOKh, 5, 1566, 1935.

13. Kulev L. P. and Grandilevskii B. A.—Investigation of quinone complexes with other aromatic compounds.

14. The same authors—Electrical conductivity and electromotive force of solutions of nitro compounds. (In press.)

15. Kulev L. P. and Ivanov B. V.—Determination of the purity and standardization of nitrobenzene. Proceedings of the Tomsk Pedagogical Institute, 1, 1939.

16. Kulikova R. M.—Application of various metals in the investigation of organic compounds by the electromotive-force method. Dissertation defended before the Academic Council of Tomsk State University, 1944.

17. Malyarevskii—Journal of Chemical Industry, IV, No. 5, p. 399, 1927.

18. Menshutkin N. A.—Investigation of the formation of acetic esters of primary alcohols. Journal of the Russian Physico-Chemical Society—Chemical Section, 9, p. 326, 1877.

19. The same author—Formation of acetic esters of secondary alcohols. Journal of the Russian Physico-Chemical Society—Chemical Section, 9, p. 346, 1877.

20. The same author—Investigation of the formation of acetic esters of tertiary alcohols. Journal of the Russian Physico-Chemical Society—Chemical Section, 10, p. 367, 1878.

21. The same author—Influence of the composition of alcohols on the rate of esterification. Journal of the Russian Physico-Chemical Society—Chemical Section, 23, p. 262, 1891.

22. The same author—On the rate of chemical transformations in polymethylene series. Journal of the Russian Physico-Chemical Society—Chemical Section, 38, p. 980, 1906.

23. Petrenko-Kritchenko P. I., Bogatskii V., and Lubman M.—Journal of the Russian Physico-Chemical Society—Chemical Section, 58, p. 215, 1926.

24. Predtechenskaya N. A.—Investigation of complexes of phenol with other organic compounds. Dissertation defended before the Academic Council of the Tomsk Industrial Institute, 1941.

25. Rashkova B. A.—Determination of low electrolyte concentrations by the current strength in a closed circuit. ZhOKh, 9, No. 3, pp. 209–212–220, 1939.

26. Tikhomirova A. M. and Efremov N. N.—Internal friction of the systems: phenol–ethylamine and phenol–diethylamine. Proceedings of the Academy of Sciences, chemical series, No. 1, p. 137, 1937.

27. Trakhtenberg F. M.—Potentiometric determination of lithium ion activity in isoamyl alcohol. Visti of the Institute of Physical Chemistry, 11, 99, 1938. Chemical Abstracts Journal, No. 5, p. 13, 1939.

28. Trifonov N. A. and Mertslin R. V.—Surface tension of irrational systems. Proceedings of Physicochemical Analysis, 12, p. 139, 1940.

29. Tronov B. V.—On reactions of rupture and formation of double bonds. Proceedings of the Tomsk Technological Institute, 45, No. 9, p. 1, 1924.

30. The same author and Grigor'eva A. P.—Nitration of nitrobenzene with nitric acid in the presence of catalysts. Proceedings of the Siberian Physico-Technical Institute, 1, No. 4, p. 7, 1932.

31. Tronov B. V., D'yakonova L. N., Gulyaeva O. I., and Nikiforova N. S.—On the strength of the bond of hydrocarbon radicals with oxygen in simple and complex ethers and on the mechanism of esterification and hydrolysis reactions of complex esters. Journal of the Russian Physico-Chemical Society—Chemical Section, 59, p. 545, 1927.

32. Tronov B. V., D'yakonova L. N., and Zonova E. A.—Complexes of aromatic polynitro compounds with unsaturated compounds and certain salts. Journal of the Russian Physico-Chemical Society—Chemical Section, 59, p. 339, 1927.

33. Tronov B. V. and Kulev L. P.—Determination of the activity of hydrogen in organic compounds by the electromotive force of the reaction with sodium. ZhOKh, 4, p. 197, 1934.

1. The same authors.—Comparison of the activity of alcohols and carboxylic acids under various conditions. ZhOKh, 4, p. 783. 1934.

2. The same authors.—Determination of the activity of unsaturated organic compounds by the electromotive force of the reaction with sodium. ZhOKh, 5, p. 1007. 1935.

3. The same authors.—The activity of hydrogen in complexes of alcohols with unsaturated organic compounds. ZhOKh, 5, p. 1239. 1935.

4. The same authors.—Complexes of nitro compounds of the benzene series with acids. ZhOKh, 5, p. 401. 1935.

5. The same authors.—On the interaction of fatty, aromatic, and heterocyclic amines with carboxylic acids. Izv. TKI, 59, No. 3, p. 118. 1942.

6. The same authors.—Determination of the electromotive force of the reaction with metals as a method for studying organic compounds. Izv. GPI, 55, p. 335. 1936.

7. Tronov B. V. and Ladygina L. V.—On the mechanism of benzene nitration. Ukr. Chem. J. 1932.

8. Udovenko V. V. et al.—Viscosity of systems: ketones—organic acids. ZhOKh, pp. 1162, 1512, 1564, 1731, 1796. 1939.

9. The same author and Usanovich M. I.—Cryoscopic study of the systems: phenol–aniline, phenol–dimethylaniline, and phenol–diethylaniline. ZhOKh, 10, p. 17. 1940.

10. Usanovich M. I. and Tenenbaum A.—Viscosity of the system: acetic acid—nitrobenzene. ZhOKh, 5, p. 709. 1935.

11. Fevralev K. D.—Methods for the analysis of carboxylic acids. Dissertation defended at the Academic Council of the KhTF TII. 1941.

12. Hückel W.—Theoretical Foundations of Organic Chemistry. 2nd ed., vol. I, p. 113. 1935.

13. Bent H. E., Keevil N. B.—Journ. Amer. Chem. Soc. 5, 1228, 1936.

14. Bent H. E., Keevil N. B.—Journ. Amer. Chem. Soc. 58, 1371. 1936.

15. Brun P.—Compt. rend. d l’Ac. Sc. 200, 1932. 1935.

16. Danner—Ph. S. Journ. Amer. Chem. Soc. 44, 2832. 1922.

17. Hrinakowski K., Staszewski H., Szmytowna W. Roczn. Chemii 17, 189. 1937. Chem. Zbl II, 2196 1937.

18. Norris J. F., Ashdown A. A.—Journ. Amer. Chem. Soc. 47, 873. 1925.