Proceedings of the Tomsk Technological Institute of Emperor Nicholas II.

Vol. XX, 1910, No. 4.

On a New Isomer of Heptane, Symmetrical Tetramethylpropane or Methyl-2-Methyl-4-Pentane, and on Its Presence in Caucasian Petroleum

(Diploma work).

Student G. V. Khonina.

The present article is a detailed description of diploma work¹ carried out by me in the Laboratory of Organic Chemistry of the Tomsk Technological Institute at the suggestion and under the direction of Prof. N. M. Kizhner.

The work described here, according to the time of its execution, may be divided into two parts. In carrying out the first of them, the aim was, above all, to obtain symmetrical tetramethylpropane and to study its properties. As a result it turned out that the boiling point of this hydrocarbon (83°–84°) comes very close to the boiling point of petroleum cyclohexane (80°–82°). This circumstance, in connection with the fact of the formation of tertiary nitro compounds upon the nitration of petroleum cyclohexane, immediately suggested the possibility of the presence of symmetrical tetramethylpropane in Caucasian petroleum and thus prompted an attempt to investigate the latter in the indicated direction. It is precisely to the execution of this investigation that the second half of the present work is devoted.

In setting forth the article presented here, I shall, as far as possible, adhere to such an order of presentation as would correspond to the course of the experimental side of the investigation.

1. On Symmetrical Tetramethylpropane.

In carrying out the synthesis of symmetrical tetramethylpropane, the starting material used was the tertiary alcohol dimethylisobutylcarbinol. This alcohol was converted into the corresponding—

…corresponding iodide; the latter, by reduction according to the method of Berthelot, is then converted into the hydrocarbon. A more detailed description of the indicated experiments follows below.

Preparation of the tertiary alcohol dimethylisobutylcarbinol

\[ \begin{array}{ccccc} \mathrm{CH_3}—\mathrm{CH}—\mathrm{CH_2}—\mathrm{C(OH)}—\mathrm{CH_3}\\ & | && |\\ & \mathrm{CH_3} && \mathrm{CH_3} \end{array} \]

This alcohol was first obtained by D. Pavlov¹ in 1874 with the aid of the Butlerov method, customary at that time, for obtaining tertiary alcohols: in particular, by the action of isovaleryl chloride on zinc methyl.

Almost simultaneously with D. Pavlov, the very same alcohol was also obtained by V. V. Markovnikov² from oxyisocaproic acid.

The synthesis of the same dimethylisobutylcarbinol was carried out by me with the aid of Masson’s method,³ which is more convenient and more advantageous (in the sense of the yield of the final product); for this purpose ethyl isovalerate and magnesium methyl iodide were introduced into the reaction.

Ethyl isovalerate, \((\mathrm{CH_3})_2\mathrm{CH.CH_2COOC_2H_5}\), was prepared in the following way: a mixture of 200 g of isovaleric acid, 96 g of 95% ethyl alcohol, and 30 g of concentrated sulfuric acid (to remove the elements of water) was heated in a flask with a reflux condenser to boiling for \(1 \tfrac{1}{2}\) hours. After the reaction was complete, water was poured into the flask. The ether that floated to its surface was separated and washed several times with a solution of soda until the acid reaction disappeared. The ether thus washed was then dried with finely divided calcium chloride, after which it was distilled with a dephlegmator. Its boiling temperature proved to be \(134.3^\circ\).⁴

Magnesium methyl iodide \(\mathrm{MgCH_3J}\) was obtained by Grignard’s method,⁵ whereby, calculated for 57 g of powdered metallic magnesium, up to 327 g of dry, pure methyl iodide were taken. The reaction proceeded in an ethereal medium. Into a flask connected with a reflux condenser was placed a mixture of absolute

ether and magnesium; then methyl iodide was added thereto little by little through a dropping funnel. At first the reaction proceeded very violently, accompanied by strong heating of the mixture. To avoid excessive loss of the readily volatile methyl iodide, and also of the ether, the flask had to be cooled with snow. After some time had elapsed, the mass in the flask noticeably thickened and turned white.

To obtain the tertiary alcohol, ethyl isovalerate was gradually introduced through a dropping funnel into the same flask, now containing magnesium iodomethyl; in total, according to the calculation, about 150 g were taken (reckoning 1 molecule of ester for 2 molecules of magnesium iodomethyl). The reaction in this case, as in the case indicated above, proceeded so energetically that the flask had to be cooled with snow. To complete the reaction, the mixture was then boiled with a reflux condenser for 2 hours, after which it was left undisturbed for 15 hours at room temperature. Subsequent decomposition of the reaction product with cold water yielded the desired dimethylisobutylcarbinol.

The process itself of formation of the above-mentioned tertiary alcohol may be represented by the following series of equations:

\[ \text{1) }\quad \mathrm{CH_3J + Mg = CH_3MgJ} \]

\[ \text{2) }\quad \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{O}\\ | \qquad\qquad \|\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}OC_2H_5 + CH_3MgJ} \\ |\\ \mathrm{CH_3} \end{array} = \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{OMgJ}\\ | \qquad\qquad |\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}OC_2H_5}\\ | \qquad\qquad |\\ \mathrm{CH_3}\qquad\mathrm{CH_3} \end{array} \]

\[ \text{3) }\quad \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{OMgJ}\\ | \qquad\qquad |\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}OC_2H_5 + CH_3MgJ}\\ | \qquad\qquad |\\ \mathrm{CH_3}\qquad\mathrm{CH_3} \end{array} = \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{OMgJ}\\ | \qquad\qquad |\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}CH_3 +}\\ | \qquad\qquad |\\ \mathrm{CH_3}\qquad\mathrm{CH_3} \end{array} \mathrm{MgJOC_2H_5} \]

\[ \text{4) }\quad \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{OMgJ}\\ | \qquad\qquad |\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}CH_3 + H_2O}\\ | \qquad\qquad |\\ \mathrm{CH_3}\qquad\mathrm{CH_3} \end{array} = \begin{array}{c} \mathrm{H}\qquad\qquad\mathrm{OH}\\ | \qquad\qquad |\\ \mathrm{CH_3{-}C{-}CH_2{-}C{-}CH_3 + MgJOH.}\\ | \qquad\qquad |\\ \mathrm{CH_3}\qquad\mathrm{CH_3} \end{array} \]

The tertiary alcohol thus obtained passed into the ethereal solution. The excess solvent was removed from the latter by distillation on a water bath. The thick white mass that then floated to the surface was already a saturated ethereal solution of dime-

...dimethylisobutylcarbinol was to be distilled off with steam, salted out from the distillate obtained with potash, and finally separated from it by means of a long funnel. A very small portion of this alcohol, which remained in the aqueous solution, was extracted from the latter by extracting it with ether.

Freed from ether by distillation on a water bath, the tertiary alcohol was dried with fused potash and then distilled. The first fraction of the distillate, which passed over at a temperature below \(115^\circ\) and contained an admixture of ether, was quite insignificant. The greater part of the alcohol entered the second fraction of the distillate, which distilled within the limits \(115^\circ\)—\(139^\circ\). Since the tertiary alcohol obtained is almost insoluble in water, in order finally to remove traces of ethyl alcohol from it, it was washed several times with water. After this, dried with fused potash, it was subjected to distillation with a dephlegmator. As a result, up to 107 g of a completely pure preparation were obtained, which passed over almost entirely at \(133^\circ\) ¹. Its actual yield amounts to about 80% of the theoretical.

The dimethylisobutylcarbinol obtained is a colorless mobile liquid; it has a burning taste and a pleasant camphor odor; it does not freeze in a cooling mixture of snow and calcium chloride, but in doing so becomes thicker and less mobile. Its boiling point is \(133^\circ\) at 749 mm pressure.

Apart from the boiling point of this alcohol, all its other physical constants were not indicated by D. Pavlov. The determination of them which I carried out in order to fill this gap gave the following results.

Specific gravity:

\[ \begin{aligned} \text{Weight of alcohol at }0^\circ &\ .\ .\ .\ .\ 2.7466\ \text{g}.\\ \text{“}\quad\text{“}\quad\text{“}\quad20^\circ &\ .\ .\ .\ .\ 2.6912\ \text{g}.\\ \text{Weight of water at }0^\circ &\ .\ .\ .\ .\ 3.2987\ \text{g}. \end{aligned} \]

\[ d^0_0 = 0.8326;\qquad d^{20}_0 = 0.8158. \]

Refractive index \(n_{20}=1.4172\).

Molecular refraction:

\[ \begin{array}{cc} \text{Found:} & \text{Calculated for } C_7H_{15}OH: \\ R^2 = 35.77. & R^2 = 35.84. \end{array} \]

Conversion of the Tertiary Alcohol into the Iodide

\[ \begin{gathered} \mathrm{CH_3-CH-CH_2-CJ-CH_3}\\ \phantom{\mathrm{CH_3-}}\vert\qquad\qquad\vert\\ \mathrm{CH_3}\qquad\quad \mathrm{CH_3} \end{gathered} \]

The transition from dimethylisobutylcarbinol to the corresponding hydrocarbon, as I have already mentioned earlier, was carried out through the iodide. The latter was obtained from the tertiary alcohol by saturating it with gaseous hydrogen iodide. The iodination of one and the same portion of alcohol was repeated up to four times, and after each of them the liquid, for completeness of the reaction, was left at rest for at least twenty-four hours; during this time it was kept in a glass vessel with a ground-glass stopper and, to avoid the decomposing action of sunlight upon it, was stored in a dark place. In order to protect the liquid from considerable heating during the reaction, the latter was conducted almost under continuous cooling with snow; however, toward the end of the saturation, in order to attain the greatest possible completeness of the reaction, the liquid was heated on a water bath to approximately \(58^\circ\).

The reaction of formation of the iodide may be represented by the following scheme:

\[ \begin{gathered} \mathrm{CH_3-CH-CH_2-C(OH)-CH_3 + HJ} = \mathrm{CH_3-CH-CH_2-CJ-CH_3 + H_2O}\\ \phantom{\mathrm{CH_3-}}\vert\qquad\qquad\vert \qquad\qquad\qquad\quad \phantom{\mathrm{CH_3-}}\vert\qquad\qquad\vert\\ \mathrm{CH_3}\qquad\quad \mathrm{CH_3} \qquad\qquad\qquad\qquad \mathrm{CH_3}\qquad\quad \mathrm{CH_3} \end{gathered} \]

The crude iodide obtained was noticeably colored by iodine to a dark-brown color. In order to purify it from admixture of free iodine, and also hydrogen iodide, it was shaken several times in a separatory funnel—first with water, then with a \(5\%\) solution of soda until decolorization. The iodide washed in this way was dried with fused pieces of calcium chloride and weighed in that form. From 95 g of the tertiary alcohol, 177 g of dry pure iodide were obtained. Its actual yield is equal to \(95.6\%\) of the theoretical.

The boiling point of the iodide, determined by distillation of a separate small portion of it, proved to be \(140^\circ\)–\(142^\circ\) at 756 mm pressure. It should be noted that, with respect to temperature, this iodide proved to be quite stable; the distillation indicated above was accompanied by only a very slight decomposition of it.

Determination of the percentage content of iodine in the undistilled iodide, carried out by Carius’ method, gave the following results:

0.2259 g of iodide gave 0.2217 g of AgJ and 0.0058 g of Ag.

Obtained:  Calculated for $C_7H_{15}J$:

$J\quad 56.00\%$    $J\quad 56.17\%$

Conversion of the iodide into a hydrocarbon

\[ \begin{array}{ccccc} \mathrm{CH_3}—\mathrm{CH}—\mathrm{CH_2}—\mathrm{CH}—\mathrm{CH_3}\\ &|&&|\\ &\mathrm{CH_3}&&\mathrm{CH_3} \end{array} \]

The initial attempt to reduce a separate small portion of the iodide with the aid of a zinc–copper couple gave a negative result: in this case an almost exclusively unsaturated hydrocarbon was obtained.

In order to obtain from the iodide the corresponding saturated hydrocarbon, Berthelot’s method was therefore employed, i.e. its reduction with hydriodic acid in sealed tubes.

The reaction for the formation of the hydrocarbon indicated above should in this case proceed according to the following equation:

\[ \begin{array}{ccccccccc} \mathrm{CH_3}—\mathrm{CH}—\mathrm{CH_2}—\mathrm{CJ}—\mathrm{CH_3} +\mathrm{HJ} = \mathrm{CH_3}—\mathrm{CH}—\mathrm{CH_2}—\mathrm{CH}—\mathrm{CH_3} +\mathrm{J_2}\\ &|&&|&&&|&&|\\ &\mathrm{CH_3}&&\mathrm{CH_3}&&&\mathrm{CH_3}&&\mathrm{CH_3} \end{array} \]

In view of the fact that the degree of purity of the iodide, judging from the analytical data, proved quite sufficient for the purpose of the reaction, it was used without preliminary distillation. For carrying out the reaction, 5 g of iodide and 20 cu. cm of fuming hydriodic acid of specific gravity 1.9 were sealed into each tube. The tubes were heated for 5 hours at $220^\circ$ and then for another 2 hours at $230^\circ$. On opening the tubes, a fairly strong pressure of gases was detected inside them. The liquid contained in each of them was diluted with water. The hydrocarbon that floated to its surface was drawn off by means of a pipette into a separatory funnel, shaken in the latter with a solution of caustic potash until decolorized, and then washed several times in the same funnel with more water. From 151 g of iodide, 63 g of crude, undistilled hydrocarbon were obtained.

The hydrocarbon, dried over fused calcium chloride, was subjected to fractional distillation with a dephlegmator, as a result of which two fractions were obtained: 1) with b.p. $65^\circ$–$80^\circ$, in the amount of 6 g, and 2) with b.p. $80^\circ$–$85^\circ$, in the amount of 53 g. The reaction

Beilstein’s test revealed the presence of halogen in both fractions of the hydrocarbon. The latter, in order to purify it from iodine, was boiled over metallic potassium. This operation was carried out in a small flask with a reflux condenser and continued for 8 hours. As a result, from the first fraction, owing to the decomposition of its impurities on the one hand and loss through volatilization on the other, only a few drops remained. The second fraction, after being boiled over metallic potassium, was subjected to a second distillation with a dephlegmator. In this operation the hydrocarbon passed over almost entirely within the limits \(80^\circ\)—\(86^\circ\); the greater part of it distilled at \(82^\circ\)—\(84^\circ\). The hydrocarbon thus purified, judging by Beilstein’s test, proved to be completely free from iodine. As a result of all the operations indicated above, 47 g of perfectly pure hydrocarbon were obtained.

The latter is a colorless and very mobile liquid, the odor of which very strongly recalls the peculiar odor of kerosene.

The hydrocarbon begins to dissolve in fuming nitric acid of specific gravity 1.52, this reaction being accompanied by a rather strong evolution of heat and brown oxides of nitrogen. This behavior of the hydrocarbon toward strong nitric acid clearly indicates the presence in it of a tertiary group—CH¹.

The boiling point of the pure hydrocarbon is \(83^\circ\)—\(84^\circ\) at 764 mm pressure.

Specific gravity.—

Weight of hydrocarbon at \(0^\circ\) . . . . . 2.2994 g.
" " " \(20^\circ\) . . . . . 2.2446 g.
Weight of water " \(0^\circ\) . . . . . 3.2984 g.

\[ d^0_0 = 0.6971;\qquad d^{20}_0 = 0.6805 \]

Refractive index \(n_{20}=1.3825\).

Molecular refraction.—

Obtained:	Calculated for \(C_7H_{16}\):
\(R^2=34.24.\)	\(R^2=34.32.\)

Analysis of the hydrocarbon:

0.2133 g of substance gave 0.3074 g \(H_2O\) and 0.6563 g \(CO_2\)

Obtained:	Calculated for \(C_7H_{16}\):
C 83.91%	C 84.00%
H 16.12%	H 16.00%

In conclusion, I must note that the hydrocarbon described above was subsequently also obtained by the late Prof. M. I. Konovalov¹. The symmetrical tetramethylpropane synthesized by him had the following physical constants: b.p. \(83^\circ\)—\(84^\circ\) at 749 mm; \(d^0_0 = 0.7022\); \(d^{22}_0 = 0.6879\); \(n^{22}_{\mathrm D} = 1.38477\); \(MR = 34.05\). As is evident from the comparison, the properties of the two hydrocarbons agree very closely with one another.

Taking into account the close connection that exists between the symmetrical tetramethylpropane described above, on the one hand, and all the other isomers of heptane, on the other, and also, in general, wishing to give the present work a more complete form, I allow myself to present here, if only briefly, the most important information about these hydrocarbons, comparing them with one another in the form of the table appended below².

Theory admits the possibility of the existence of nine different isomers of heptane. Of these, only six have actually been obtained and investigated so far.

2. On the Presence of Symmetrical Tetramethylpropane in Caucasian Petroleum

In passing to the question of the presence of symmetrical tetramethylpropane in Caucasian petroleum, one cannot at the same time avoid touching upon the question of the occurrence of different isomers of heptane in petroleum in general. Although the latter does not constitute the direct task of the present investigation, it nevertheless stands in a certain mutual relation to it. From the point of view of elucidating the chemical composition of petroleum, this question is far from uninteresting; nevertheless, to this day it remains very little investigated. Leaving aside the more

a detailed examination of it, I consider it not superfluous to cite here, if only in very brief form, the few positive data concerning it that are available in the relevant chemical literature.

Of the two most widespread kinds of petroleum—American and Caucasian—only the first has been more or less studied in this respect. Thus, as early as 1865 Warren¹, studying American (Pennsylvanian) petroleum, first discovered in it the presence of two isomeric heptanes: the normal, with b.p. \(98.1^\circ\), and the iso-normal, with b.p. \(90.4^\circ\). In 1871 Schorlemmer² isolated from the very same petroleum an isomer of heptane with b.p. about \(90^\circ\). On the basis of his investigations, and also relying on Grimshaw’s work³, he subsequently expressed his view on the structure of this hydrocarbon, supposing it to be analogous to ethylisoamyl⁴. Considerably later Francis and Sidney Young⁵ succeeded in obtaining from the same American petroleum nearly entirely pure normal heptane with b.p. \(98.2^\circ\)—\(98.5^\circ\). They also then isolated from it ethylisoamyl with b.p. \(89.9^\circ\)—\(90.4^\circ\)⁶. In addition, Pelouze and Cahours⁷, while studying American petroleum, found in it, among other things, isoheptane with b.p. \(92^\circ\)—\(94^\circ\). The broad limits within which the boiling temperature of this hydrocarbon varies give rise to some doubt as to its individuality; in any case, its structure could not be precisely established. Finally, Morgan⁸ isolated from Pennsylvanian petroleum not fully pure normal heptane with b.p. \(96^\circ\)—\(99^\circ\). In Schorlemmer’s opinion⁹, it must have contained an admixture of isoheptane with b.p. \(90^\circ\). These last indications exhaust almost all the data concerning American petroleum.

There are far fewer data on the same question with regard to Caucasian petroleum. Thus, in 1899 K. V. Kharichkov¹⁰, while investigating Caucasian (Grozny) gasoline, isolated from its fraction at \(84^\circ\)—\(100^\circ\) an isomeric heptane with b.p. \(90.5^\circ\)—\(91.5^\circ\). The structure of this hydrocarbon remained undetermined,

Table

isomers

Name and structural formula of the hydrocarbon \(C_7H_{16}\)	Origin of the hydrocarbon	Boiling temperature	Specific gravity
1) Normal heptane \(CH_3—(CH_2)_5—CH_3\)	Obtained in the distillation of azelaic acid \(C_9H_{16}O_4\) with baryta. \(^{1}\) Also obtained in the distillation of turpentine from Pinus sabiniana (Dough., Californien) with water \(^{2}\) . . . Found in resin from the [[unclear: Kennel]] coal \(^{3}\) . . . . . . Found in American petroleum. Discovered in Caucasian petroleum \(^{6}\)	\(100,5^\circ\) \(984^\circ\) \(98^\circ—99^\circ\) \(98,1^\circ\) \(98,2^\circ—98,5^\circ\) \(97,5^\circ—99,5^\circ\)	\(d^{20,5}=0,6840\) \(d^0=0,7006;\ d^{14,2}=0,6885\) \(d^{17,5}=0,709\) \(d^0=0,7085.\ ^{4}\) \(d^0_4=0,7013.\ ^{5}\) —
2) Triethylmethane (ethyl-3-pentane) \(CH_3—CH_2—CH—CH_2—CH_3\) \(\qquad\qquad\ \ \ |\) \(\qquad\qquad\ CH_2\) \(\qquad\qquad\ \ \ |\) \(\qquad\qquad\ CH_3\)	Obtained from ortho-formic ether by the action on it of zinc ethyl and sodium \(^{7}\) . . . . . . .	\(95^\circ—98^\circ\)	\(d^{27}=0,689\)
3) Methylethylpropylmethane (ethyl-4-pentane) \(CH_3—CH_2—CH_2—CH—CH_3\) \(\qquad\qquad\qquad\ \ \ |\) \(\qquad\qquad\qquad\ CH_2\) \(\qquad\qquad\qquad\ \ \ |\) \(\qquad\qquad\qquad\ CH_3\)	Obtained by the action of sodium on a mixture of active amyl iodide and ethyl iodide . . . . . . . . . .	\(91^\circ\) \(80^\circ—88^\circ\)	\(d^{20}=0,6895\ ^{8}\) \(d^{17}=0,7806;\ d^{60}=0,7299\)
4) Ethylisoamyl (methyl-5-hexane) \(CH_3—(CH_2)_3—CH—CH_3\) \(\qquad\qquad\qquad\ \ \ |\) \(\qquad\qquad\qquad\ CH_3\)	Obtained by the action of sodium on a mixture of ethyl iodide and isoamyl iodide \(^{10}\) . . . . . . Also by the action of sodium on a mixture of ethyl bromide and isoamyl bromide \(^{11}\) . . . . . . Obtained also by the action of zinc and hydrochloric acid on methylisoamylcarbinol iodide \(^{12}\) . . . . . . Formed in the electrolysis of a mixture of potassium butyrate and potassium zinc-acetate \(^{13}\) . . . . . . Found in American petroleum	— \(90^\circ\) — — \(90,5^\circ\) \(89,9^\circ—90,4^\circ\)	— \(d^{18,4}=0,6833\) — — \(d^{17,5}=0,6819\ ^{14}\) \(d^0_4=0,7067\ ^{15}\)

1) Dale.—Liebig’s Annalen der Chemie [und Pharmacie], 132, 247.
2) Thorpe.—Lieb. Ann. d. Chem. 198, 361. Schorlemmer u Thorpe.—Lieb. Ann. d. Chem. 217, 150.
3) Schorlemmer.—Lieb. Ann. d. Chem. 125, 109.
4) Warren.—Zeitschrift für Chemie (1865 г.), 668.
5) Francis, Sidney Young.—Journal of the chem. soc. 73, 916, 921.
6) Charichkov.—Journal of the Russian Physico-Chemical Society, 31, 554, (1899 г.).
7) Ladenburg.—Berichte der Deutschen chemischen Gesellschaft, 5, 752.
8) Just.—Lieb. Ann. d. Chem. 220, 154.
9) Welt.—Bullet. de la soc. chim. de Paris (3) 11, 1188; Ann. de chim. et de physique (7) 6, 121.
10) Wurtz.—Ann. de chim. et de physique (3) 44, 275.

Isomers of Heptane

Name and structural formula of the hydrocarbon \(C_7H_{16}\)	Origin of the hydrocarbon	Boiling temperature	Specific gravity
5) Dimethyldiethylmethane (dimethyl-3-pentane) \(\mathrm{\ \ \ \ \ \ CH_3}\) \(\mathrm{\ \ \ \ \ \ \vert}\) \(\mathrm{CH_3-CH_2-C-CH_2-CH_3}\) \(\mathrm{\ \ \ \ \ \ \vert}\) \(\mathrm{\ \ \ \ \ \ CH_3}\)	Obtained from acetochloride and zinc ethyl \(^{15}\) . . . . . . . .	\(86^\circ\text{–}87^\circ\)	\(d^\circ=0.7111\)
6) Symmetrical tetramethylpropane (methyl-2-methyl-4-pentane) \(\mathrm{CH_3-CH-CH_2-CH-CH_3}\) \(\mathrm{\ \ \ \vert\ \ \ \ \ \ \vert}\) \(\mathrm{\ \ CH_3\ \ \ \ CH_3}\)	Obtained from dimethylisobutylcarbinol by converting it into the iodide and reducing the latter with hydroiodic acid in sealed tubes . . . . . . . . . Its presence in Caucasian petroleum has also been established \(^{19}\).	\(83^\circ\text{–}84^\circ\)	\(\left\{\begin{array}{l} d^0_4=0.6971;\ d^{20}_4=0.6805\ ^{17)} \\ d^0_4=0.7022;\ d^{22}_4=0.6879\ ^{18)} \end{array}\right.\)
7) Methylethylisopropylmethane (methyl-3-methyl-4-pentane) \(\mathrm{CH_3-CH_2-CH-CH-CH_3}\) \(\mathrm{\ \ \ \ \ \ \vert\ \ \vert}\) \(\mathrm{\ \ \ \ \ CH_3\ CH_3}\)	Not obtained.	—	—
8) Dimethylisobutylmethane (methyl-2-dimethyl-3-butane) \(\mathrm{\ \ \ \ \ \ CH_3}\) \(\mathrm{\ \ \ \ \ \ \vert}\) \(\mathrm{CH_3-CH-C-CH_3}\) \(\mathrm{\ \ \vert\ \ \vert}\) \(\mathrm{\ CH_3\ CH_3}\)	Not obtained.	—	—
9) Ethylisobutylmethane (dimethyl-4-pentane) \(\mathrm{\ \ \ \ \ \ \ \ \ \ CH_3}\) \(\mathrm{\ \ \ \ \ \ \ \ \ \ \vert}\) \(\mathrm{CH_3-CH_2-CH_2-C-CH_3}\) \(\mathrm{\ \ \ \ \ \ \ \ \ \ \vert}\) \(\mathrm{\ \ \ \ \ \ \ \ \ \ CH_3}\)	Not obtained.	—	—

11) H. Grimshaw.—Lieb. Ann. d. Chem. 166, 163.
12) Purdie.—Journal of the Chem. Soc. 39, 467.
13) Wurtz.—Lieb. Ann. d. Chem. 96, 272.
14) Schorlemmer.—Chemisches Central-Blatt (1871), 888. Lieb. Ann. d. Chem. 136, 259.
15) Francis, Sidney Young.—Journ. of the Chem. Soc. 73, 906, 922.
16) Friedel, Ladenburg.—Lieb. Ann. d. Chem. 142, 310.
17) Khonin.—Journ. Russ. Phys.-Chem. Soc. 37, 521 (1905).
18) Konovalov.—Journ. Russ. Phys.-Chem. Soc. 37, 910 (1905).
19) Khonin.—Journ. Russ. Phys.-Chem. Soc. 40, 731 (1908).

although in boiling point it is very close to ethylisoamyl (89.9°—90.5°) on the one hand, and to methylethylpropylmethane (91°) on the other. In the Grozny gasoline K. V. Kharchikov subsequently discovered the presence of normal heptane¹, isolating it in the form of a fraction with b.p. 97.5°—99.5°. K. V. Kharchikov did not, however, succeed in obtaining this hydrocarbon in a pure form, i.e., in fully individualizing it.

As is evident from the facts set forth above, the question of the presence of various isomers of heptane in Caucasian petroleum has been investigated extremely little. Any attempt directed toward its further study should therefore be regarded, from the theoretical point of view, as far from superfluous. Below is given precisely a description of work carried out with the aim of clarifying the question of the presence of symmetrical tetramethylpropane in Caucasian petroleum.

The boiling point of symmetrical tetramethylpropane (83°—84°) is very close to the boiling point of petroleum cyclohexane (80°—82°). This circumstance, as I have already mentioned earlier, gave occasion and warrant at the time to express the supposition concerning the possible presence of the former of these in Caucasian petroleum². True, this supposition, taken merely by itself, would be problematic to a considerable degree. However, the fact that, upon nitration of petroleum cyclohexane, tertiary nitro compounds are formed gave it already a high degree of probability and, having clarified the possibility of attaining more or less positive results from the investigation begun, thus contributed to the successful completion of the latter.

In view of the very great significance of the fact indicated above for the entire subsequent course of the present work, it is necessary to dwell on it in somewhat greater detail. As early as 1898 Prof. V. V. Markovnikov, while investigating the cyclic hydrocarbons of Caucasian (Baku) petroleum, called attention to the impossibility of obtaining from it perfectly pure cyclohexane by fractionation. Even preliminary treatment of the corresponding petroleum fraction with a nitrating mixture under cooling, and with nitric acid, did not in this case give the expected results³. Further, studying the cyclohexane isolated by him from petroleum, V. V. Markovnikov observed that, as a result of nitration of the latter, in addition to the product formed in the main

the mass of the secondary nitro product (nitrocyclohexane), there is obtained, in an insignificant quantity, also some tertiary nitro compound¹. The tertiary character of the latter had already been detected by the very method of separating it from the secondary nitro product—by treating the nitration products with a strong solution of caustic potash². Describing his work in the pages of the Journal of the Russian Physico-Chemical Society³, V. V. Markovnikov, among other things, says:

“Of the nature of the impurity not removed by nitric acid, one can now judge only conjecturally. It may perhaps be elucidated by a further investigation of the tertiary nitro products formed in small quantity during the nitration of petroleum hexanaphthene. This question has not yet formed part of our task,” etc. Having established only the very fact of the formation of the indicated tertiary nitro compounds, V. V. Markovnikov, however, did not subject them to further investigation, and the question of the composition of the latter has therefore remained open to this day.

In 1900 Prof. N. M. Kizhner⁴ converted the above-mentioned tertiary nitro compound into the corresponding amine. The latter was not at that time studied in detail, and its composition, as well as its structure, has remained unexplained up to the present. Here I confine myself for the time being to this single mention of that work; fuller information about it will be given by me below.

Undoubtedly, the above-mentioned tertiary nitro compounds could have been formed only at the expense of those impurities of petroleum cyclohexane which contain in their composition a tertiary group —CH

With the existence of such data, it was quite natural to admit the supposition of the possible presence of symmetrical tetramethylpropane in petroleum—namely as a companion of petroleum cyclohexane. This question could finally be resolved only by comparing fully individualized derivatives of the synthetic hydrocarbon with the corresponding derivatives obtained from petroleum. A consecutive description of the work on obtaining these and other derivatives and on studying their properties is given below:

Synthetic amine
\[ \begin{array}{c} \mathrm{CH_3{-}CH{-}CH_2{-}C(NH_2){-}CH_3}\\ \phantom{\mathrm{CH_3{-}}}\vert \qquad\qquad \vert\\ \phantom{\mathrm{CH_3{-}}}\mathrm{CH_3}\qquad\ \ \mathrm{CH_3} \end{array} \]
and its derivatives.

Symmetrical tetramethylpropane, for the purpose of its further conversion into a series of derivatives, had first to be nitrated. The nitration of this hydrocarbon was first carried out by Prof. M. I. Konovalov, as he had already reported at the time in the Journal of the Russian Physico-Chemical Society¹. In order to achieve the above-mentioned aim, I therefore had, in turn, to reproduce this work, changing only some of the experimental conditions to a slight degree. The nitration process itself of symmetrical tetramethylpropane was carried out by M. I. Konovalov’s now generally accepted method, i.e. by heating it with dilute nitric acid in sealed tubes. The reaction taking place in this case may be represented by the following scheme:

\[ \begin{array}{c} \mathrm{CH_3{-}CH{-}CH_2{-}CH{-}CH_3 + NO_2OH} = \mathrm{CH_3{-}CH{-}CH_2{-}C(NO_2){-}CH_3 +}\\ \phantom{\mathrm{CH_3{-}}}\vert \qquad\qquad \vert \qquad\qquad\qquad\quad \phantom{\mathrm{CH_3{-}}}\vert \qquad\qquad \vert\\ \phantom{\mathrm{CH_3{-}}}\mathrm{CH_3}\qquad\ \ \mathrm{CH_3} \qquad\qquad\qquad\quad \phantom{\mathrm{CH_3{-}}}\mathrm{CH_3}\qquad\ \ \mathrm{CH_3}\\[6pt] \mathrm{+H_2O} \end{array} \]

To carry out the reaction, 5 cubic centimeters of hydrocarbon and 25 cubic centimeters of nitric acid of specific gravity 1.075 were sealed in each tube. The tubes were heated for 9 hours at \(100^\circ\)–\(105^\circ\), and only toward the end of the experiment, in order to accelerate the reaction, was the heating temperature raised to \(110^\circ\). On opening the tubes, a rather strong pressure of gases was found inside them. From the reaction product, previously washed with water and with a solution of soda, the un-nitrated free hydrocarbon that remained was distilled off on a salt bath and then again subjected to the nitration process. As a result, from 40 g of symmetrical tetramethylpropane taken for the reaction, 19 g of crude nitro product were obtained, and, in addition, 4 g of free hydrocarbon remained, which on subsequent examination proved to be completely unchanged.

The crude nitro product thus obtained, in order to isolate from it the tertiary nitro compound, was treated in a long funnel with an excess of a strong alcoholic solution of caustic potash. The floating—

…rising to the surface of the orange-red liquid, the almost colorless tertiary nitro compound was separated and then washed with water until the alkaline reaction disappeared. From 19 g of the crude nitro product, only 17 g of the tertiary nitro compound were isolated; the remaining 2 g of it, which passed into solution, apparently represent a mixture of primary and secondary nitro compounds. The formation of the latter in such a small quantity is fully explained by the conditions of the experiment¹. The tertiary mononitro compound obtained, \((\mathrm{CH}_3)_2\mathrm{CHCH}_2\mathrm{C}(\mathrm{NO}_2)(\mathrm{CH}_3)_2\), is an almost colorless liquid, possessing a pungent taste and a pleasant camphor-like odor.

As will be seen below, the corresponding tertiary nitro compound obtained from petroleum was not individualized to a sufficient degree and, consequently, its properties could not have great significance for purposes of comparison. A detailed study of the synthetic tertiary nitro compound therefore seemed unnecessary, all the more so since its properties had already been determined at one time by M. I. Konovalov. In view of the circumstances indicated above, the tertiary nitro compound obtained was, without further treatment or investigation, converted directly into the amine.

The reaction of reducing the tertiary mononitro compound to the corresponding amine was carried out by the method of M. I. Konovalov, by whom this amine was first obtained and described². To achieve the indicated aim, a mixture of the tertiary nitro compound and metallic granulated tin was placed in a small round-bottomed flask connected with a reflux condenser; concentrated hydrochloric acid was then gradually added to this mixture through a dropping funnel. The reaction was conducted with frequent shaking of the mixture. At first it proceeded so vigorously that the flask had to be cooled with snow for at least an hour; over the next four hours the reaction was then conducted while heating the mixture on a water bath. Toward the end of it one could observe the gradual disappearance of the camphor-like odor of the nitro product and the abundant evolution of hydrogen bubbles from the liquid.

The process itself of formation of the amine may be expressed in the form of the following equation:

\[ \begin{gathered} 2\mathrm{CH_3}-\mathrm{CH}-\mathrm{CH_2}-\mathrm{C(NO_2)}-\mathrm{CH_3}+3\mathrm{Sn}+12\mathrm{HCl}=\\ \begin{array}{cc} \vert & \vert\\[-2mm] \mathrm{CH_3} & \mathrm{CH_3} \end{array}\\[-1mm] =2\mathrm{CH_3}-\mathrm{CH}-\mathrm{CH_2}-\mathrm{C(NH_2)}-\mathrm{CH_3}+3\mathrm{SnCl_4}+4\mathrm{H_2O}. \\[-1mm] \begin{array}{cc} \vert & \vert\\[-2mm] \mathrm{CH_3} & \mathrm{CH_3} \end{array} \end{gathered} \]

From the reaction product the amine was separated by an excess of a strong aqueous solution of caustic soda and then distilled off with steam. The distillate was saturated with hydrochloric acid until an acid reaction appeared, after which the solution of the hydrochloride salt of the amine thus formed was evaporated to dryness. The concentrated aqueous solution of this salt was again decomposed by an excess of caustic alkali, with formation of the free amine. The latter was distilled off with steam and salted out from the distillate with pieces of caustic soda. The amine isolated in this way was dried with freshly powdered caustic potash, after which it was distilled. To attain more complete dehydration it was then boiled for 5 hours over anhydrous barium oxide. Having been separated from the latter by distillation on an oil bath, the amine was finally distilled once more over fresh pieces of anhydrous barium oxide. At the beginning of the distillation the temperature at once rose to \(121^\circ\), after which almost all the amine passed over within the limits \(121^\circ\)—\(122.2^\circ\).

Collected in an amount of 4 grams, the pure amine is a colorless liquid, oily in appearance, with a special sharp odor generally characteristic of the higher amines. It avidly attracts water and carbon dioxide from the air; with the former it mixes in all proportions, while with the latter it gives a crystalline salt.

The boiling point of the pure amine is \(121^\circ\)—\(122.2^\circ\) at 753 mm pressure.

Specific gravity.—

Weight of amine at \(0^\circ\)		0.7242 g
“ “ “ \(20^\circ\)		0.7089 g
Weight of water at \(0^\circ\)		0.9182 g

\[ d^{0}_{0}=0.7887;\quad d^{20}_{0}=0.7720. \]

Refractive index \(n_{14.5}=1.4199\).

Molecular refraction.—

Obtained:	Calculated for \(\mathrm{C_7H_{15}NH_2}\):
\(R^2=37.73\)	\(R^2=37.82\).

The properties of the amine obtained by M. I. Konovalov are as follows: b.p. 121°–122° at 747 mm; \(d^{30}_{0}=0.7549\); \(n^{30}_{D}=1.4009\); \(MR=37.83\). As is evident from the comparison, the physical properties of both amines are very close to one another.

The hydrochloride of the amine \(C_{7}H_{15}NH_{3}Cl\) was obtained by adding dilute hydrochloric acid to its aqueous solution. The solution of the salt thus formed was evaporated to dryness on a water bath; the crystals that separated were dried in a desiccator over sulfuric acid, in which, in order to remove from the salt the residues of free hydrochloric acid, separate pieces of caustic potash were also placed. The salt dissolves very readily in water, crystallizing from it in the form of colorless prismatic needles gathered in bundles. Its crystals greedily attract moisture from the air, fairly quickly deliquescing in the process. The salt begins to melt, with noticeable decomposition, at about 208°–209°; however, the melting point is not characteristic for it, since it is extremely variable.

The chloroplatinate \((C_{7}H_{15}NH_{3}HCl)_{2}PtCl_{4}\) was obtained from the amine hydrochloride by adding to its aqueous solution, weakly acidified with hydrochloric acid, an aqueous solution of chloroplatinic acid. The chloroplatinate formed remained in solution, from which, after concentration by evaporation on a water bath, it separated after some time in the form of crystals. The latter proved not entirely pure and therefore were recrystallized twice more. The chloroplatinate is readily soluble in water, separating from it in the form of beautifully formed polyhedral crystals. A microphotograph taken from one of these crystals is presented in Fig. 1 (Plate I). The chloroplatinate is also soluble in alcohol, though less readily than in water; from 95% wine alcohol it also separates in the form of polyhedral orange-red crystals, which are only smaller and flatter in comparison with those from water. The chloroplatinate dried at 100° does not melt, but, upon being heated very rapidly, already at about 180° begins to curl up, and then at 183° blackens, decomposing in the process.

Tetramethylpropylphenylthiourea

\[ \begin{array}{c} \mathrm{NH.(CH_{3})_{2}C.CH_{2}.CH(CH_{3})_{2}}\\ \vert\\ \mathrm{CS}\\ \vert\\ \mathrm{NH\,C_{6}H_{5}} \end{array} \]

was obtained from the amine by adding phenyl mustard oil to its solution in petroleum ether. The reaction was accompanied by considerable warming of the liquid. The fine crystalline precipitate of the substituted phenylthiourea that separated out was filtered off with the aid of a pump on a Büchner funnel; in order to remove from it the excess phenyl mustard oil, it was washed several times with petroleum ether. The precipitate, pressed and then dried on a porous clay plate, was recrystallized from hot methyl alcohol. However, the crystals that separated in this way proved not to be entirely pure; only after repeated crystallization from methyl alcohol was it finally possible to obtain completely pure substituted phenylthiourea. The latter had the appearance of colorless, regular rhombic plates with a constant melting point of \(111^\circ\)—\(112^\circ\). It also dissolves in ethyl alcohol, although with more difficulty than in methyl alcohol. From 95% ethyl alcohol, with very slow crystallization, it separates in the form of colorless plates of an elongated quadrangular shape, with blunted angles. A microphotograph taken from these plates is shown in Fig. 2 (Plate I). From absolute ethyl alcohol the above-mentioned thiourea crystallizes in rhombic plates. In petroleum ether, even on heating, it is rather difficultly soluble, separating from it in the form of very small scales of rhombic shape. In benzene, even in the cold, it dissolves very readily, subsequently separating in the same rhombic, indistinctly expressed scales.

Analysis of the substituted phenylthiourea:

I. 0.1292 g of substance gave 0.1014 g \(H_2O\) and 0.3190 g \(CO_2\);

II. 0.1531 g of substance gave 14.7 cubic cm of moist nitrogen at \(16^\circ\) and 770 mm.

Obtained:         Calculated for \[ \begin{array}{c} \mathrm{NH.C_7H_{15}}\\ \vdots\\ \mathrm{CS}\\ |\\ \mathrm{NH.C_6H_5} \end{array} \]

	I	II		Calculated
C	67.35%	—	C	67.10%
H	8.78%	—	H	8.87%
N	—	11.35%	N	11.21%

Petroleum Amine and Its Derivatives.

Speaking above of the tertiary nitro compounds formed in the nitration of petroleum cyclohexane, I briefly mentioned the work of Prof. N. M. Kizhner relating to their study. In view of the direct mutual connection existing between the present work, on the one hand, and the above-mentioned work of N. M. Kizhner, on the other, I consider it necessary to present the latter here in a more detailed form¹.

In 1900, while studying the action of silver oxide on bromamines with a tertiary position of the amide group, N. M. Kizhner set himself, among other things, the goal of tracing the course of this reaction with respect to the brominated amine obtained by him from the above-mentioned tertiary nitro compound. The latter, in its turn, had been obtained as a by-product in the nitration, by M. I. Konovalov’s method, of that fraction isolated from Caucasian (Baku) petroleum gasoline in which hexanaphthene is usually found (b.p. 80°–82°). The tertiary nitro compound isolated from the reaction product, when distilled at 40 mm pressure, gave two fractions: 1) 93°–95° with specific gravity \(d^{20}_{0}=0.9722\), and 2) 95°–100° with specific gravity \(d^{20}_{0}=0.9840\). It was then reduced with tin and hydrochloric acid to the amine. From 31 g of nitro product taken for the reaction, 20 g of crude amine were obtained. After drying, the amine was distilled. Its boiling began at 118°, but then it quickly passed over almost entirely within the range 122°–124° at 753 mm pressure.

A portion of the amine thus obtained was then subjected to bromination. Upon the subsequent action of silver oxide on the brominated amine there was obtained, among other things, a bromide with b.p. 160°–170°. Judging by the boiling temperature, the latter could belong only to the bromides of the heptane series.

These last indications exhaust almost all the data obtained as a result of N. M. Kizhner’s investigation of petroleum amine. In accordance with the plan of the present work outlined earlier, I was thus faced with the task of supplementing the characterization of this amine by determining all its remaining physical properties, as well as by preparing and studying its corresponding derivatives. Thanks to the kind offer of N. M. Kizhner, I made use—

...for this purpose the petroleum amine preparation that had remained with him, I avoided the necessity of obtaining the latter anew.

The amine handed over to me by N. M. Kizhner was dried by boiling with anhydrous barium oxide for two hours, after which it was distilled. The distillation yielded an entirely pure amine. Both in its external appearance and in odor, the latter is completely analogous to the synthetic amine described above. The additionally found physical constants of the petroleum amine are given below.

Specific gravity.—

Weight of amine at \(0^\circ\)		0.7405 g
“ “ “ \(20^\circ\)		0.7252 g
Weight of water at \(0^\circ\)		0.9182 g

\[ d_{0}^{0}=0.8064;\qquad d_{0}^{20}=0.7898. \]

Refractive index \(n_{14.5}=1.4301\).

Molecular refraction.—

Obtained:	Calculated for \(C_7H_{15}NH_2\):
\(R^{2}=37.67\)	\(R^{2}=37.82\).

As may be seen from the comparison, the physical properties of this amine very closely approach the properties of the synthetic amine.

To obtain derivatives of the petroleum amine, the same methods were used by means of which the corresponding derivatives of the synthetic amine were obtained. The properties of the obtained derivatives needed for comparison are given below.

The hydrochloride of the petroleum amine is readily soluble in water, from which it crystallizes in the form of needles gathered in bundles, entirely analogous to the needles of the corresponding salt of the synthetic amine. It greedily attracts moisture from the air, as a result of which it quickly deliquesces. It melts at a temperature above \(155^\circ\), with noticeable decomposition.

The chloroplatinate of the petroleum amine dissolves very readily in water, separating from it in the form of regularly formed polyhedral crystals. The latter have an orange-red color and, in their external appearance, are entirely analogous to the crystals of the synthetic chloroplatinate described above. A photomicrograph taken of one of the crystals of the petroleum chloroplatinate is presented in Fig. 3 (Table I). In 95% wine spirit the chloroplatinate dissolves with more difficulty than in water; it crystallizes from it in the form of the same polyhedral orange-red crystals as from water, but only smaller and flatter in comparison with

the latter. The chloroplatinate dried at \(100^\circ\) does not melt; when heated to approximately \(180^\circ\), it first cakes together and then decomposes, turning black in the process.

Substituted phenylthiourea is readily soluble in methyl alcohol, crystallizing from it in the form of colorless rhombic plates. Purified by repeated crystallization from this alcohol, it melts at \(111^\circ\)—\(112^\circ\). It dissolves with some difficulty in absolute ethyl alcohol, crystallizing from it likewise in rhombic plates. On very slow crystallization from 95% wine spirit it separates in the form of colorless quadrangular plates with blunted corners. A microphotograph taken of these plates is shown in Fig. 4 (Table I). As is seen from comparison, their external form coincides completely with the form of the plates of synthetic substituted phenylthiourea. In petroleum ether, the petroleum-derived substituted phenylthiourea dissolves with difficulty even on heating; it crystallizes from it in the form of small scales having a rhombic shape. It dissolves very readily in benzene, crystallizing from it in the form of the same rhombic scales, only their shape is not quite distinctly expressed.

From the foregoing it is evident that the properties of the derivatives of the one and the other amine are very similar to one another; with respect to the substituted phenylthioureas this similarity reaches the degree of actual identity. In the presence of such data, the fact of the presence of symmetrical tetramethylpropane in Caucasian petroleum should be considered fully proven.

The presence of the very same hydrocarbon in American petroleum, which in general is very rich in paraffins, can in its turn hardly be subject to great doubt; although, for lack of factual data, this question remains open for the time being.

In conclusion, I consider it my duty to express my sincere and profound gratitude to Professor N. M. Kizhner, at whose suggestion and under whose direction the present work was carried out.

Tomsk

17 February 1909.

Table I

Fig. 1.
Crystal of synthetic chloroplatinate.

Fig. 2.
Crystals of synthetic substituted phenylthiourea.

Fig. 3.
Crystal of petroleum chloroplatinate.

Fig. 4.
Crystals of petroleum substituted phenylthiourea.

From the article by G. V. Hoppe.

1. The work of N. M. Kizhner cited below was published by him at the time in the form of a brief note in the Journal of the Russian Physico-Chemical Society [32, 384 (1900)]. Certain data not included in that note, however, are given here for the first time from the words of N. M. Kizhner and with his kind permission. ↩↩↩↩↩↩↩↩↩↩↩

2. M. I. Konovalov, Journal of the Russian Physico-Chemical Society, 37, 1122 (1905). ↩↩↩↩↩↩

3. V. V. Markovnikov, Journal of the Russian Physico-Chemical Society, 30, 158 (1898). ↩↩↩↩

4. N. M. Kizhner.—Journal of the Russian Physico-Chemical Society, 32, 384 (1900). ↩↩↩

5. Francis, Sidney Young. Chemisches Central-Blatt (1899), B. I, S. 178. ↩↩

6. See “Ergänzungsbände zur dritten Auflage des Handbuchs der organischen Chemie von F. Beilstein,” B. 1, S. 13. ↩

7. Pelouze u. Cahours.—Liebig’s Annalen der Chemie 127, 196 (1863). ↩

8. Th. M. Morgan.—Liebig’s Annalen der Chemie 177, 307 (1875). ↩

9. Schorlemmer.—Liebig’s Annalen der Chemie 177, 212 (1875). ↩

10. K. V. Kharichkov—Journal of the Russian Physico-Chemical Society 31, 553, 554 (1899). ↩