from "Pharmaceutical Journal" (27 April 1889)
Dr. O. Hesse (1889)

Chemistry of the Coca Bases

The somewhat vague discussion that appeared about two years ago in pharmaceutical journals in regard to "amorphous cocaine" induced me to publish my observations on that subject in Pharmaceutische Zeitung, while expressly reserving to myself the further investigation of the bases cocamine and hygrine then described very briefly. Liebermann has, however, without waiting for the completion of that inquiry, lately published his investigation of isotropylcocaine, which is for the most part nothing else than th cocamine discovered by me in 1886. Though the investigation of cocamine was at that time all but complete so far as to relates to cocamine, I have thought it desirable to continue the inquiry, more especially since the crude bases I operated upon had not been subjected to the action of permanganate, as was the case with the material on which Liebermann operated. However, the subject of my investigation has also been touched by Frankfeld, and therefore in order to avoid further collision it appears desirable to make known at once the results I have arrived at so far.

In the first place I must mention that two sorts of coca leaves are now met with in commerce. The one sort is derived from Erythroxylon Coca; it was formerly investigated by Erdmann and Lossen, and was for some time the only material employed in the preparation of cocaine. Subsequently a second sort came into the market, originating from a variety of Erythroxylon growing in Jamaica and St. Lucia, which has been regarded, though perhaps incorrectly, as a variety of the well-known coca plant and termed Novagranatense. It is chiefly this latter kind of coca that has hitherto been used in North Germany for making cocaine.

Cocamine was obtained by me from the last mentioned variety of coca leaves, and from a parcel that contained chiefly that base to the extent of nearly 0.6 percent. In operating upon the crude base with petroleum spirit, some part of the cocamine remained undissolved and by repeatedly dissolving that in dilute acetic acid and precipitating with ammonia the base could soon be rendered sufficiently pure for analysis. In the following statement of the results thus obtained the figures marked A relate to the base precipitated from a dilute solution and perfectly crystalline. Those marked B relate to the hydrochloride dried at the ordinary temperature over sulphuric acid.

These results agree well with the formula C₁₇H₂₁NO₄ and C₁₇H₂₁NO₄HCl, by which I consider the composition of cocamine was represented, especially since platinum determinations gave corresponding results. However, a subsequent determination of chlorine in the hydrochloride gave only 9.11 percent. This discordant result induced me to examine more closely the products of decomposition, and in that way I ascertained that the true formula of cocamine is C₁₇H₂₃NO₄. Hence the substances represented by A and B were really C₁₇H₂₃NO₄ + 0.5 H₂O and C₁₇H₂₁NO₄HCl + H₂O.

The attempt to determine the water in the pure base was unsuccessful. When heated to 90 to 95 °C, it melted, but there was no loss of weight. On heating to 120 °C, the loss of weight was very near that corresponding to 0.5 H₂O, but the examination of the residue showed that some methyl alcohol must have been given off as well as water. By long continued heating at this temperature a point is reached at which the decomposition of the base goes on rapidly and a refreshing odour is given off. This is probably due to the formation of the methyl ester of cocaic acid.

The hydrochloride undergoes decomposition even more readily than the base itself. Even at 100 °C, it gives off, together with one molecule of water, a certain amount of methyl chloride; if the salt be heated for some hours to 120 °C, the residue contains only traces of chlorine. This decomposition takes place according to the following equation:--
C₁₉H₂₃NO₄HCl + H₂O = C₁₈H₂₁NO₄ + CH₃Cl + H₂O.

A similar decomposition takes place when the water solution of the hydrochloride is boiled for some time, but then the methyl is given off chiefly in the state of methyl alcohol, and on cooling the cocamylecgonine separates as a colourless or yellowish oil. When the solution of the salt is mixed with hydrochloric acid, the decomposition is not only promoted but the cocmylecgonine produced is also split up into ecgonine and cocaic acid.
C₁₈H₂₁NO₄HCl + H₂O = C₉H₁₅NO₃ + C₉H₈O₂.

I have given the name of cocaic acid to the product thus formed, because it is characteristic of coca. To purify this acid the crude product is boiled with lime and water. From a dilute solution of the lime salt hydrochloric acid precipitates the acid in a crystalline form; by recrystallization from boiling glacial acetic acid it is obtained in the form of small colourless laminae which melt at 266 °C. Cocaic acid is sparingly soluble in ether, but it dissolves more readily in hot alcohol, and on cooling crystallizes out in needles. The composition of cocaic acid is represented by the formula C₉H₈O₂. The lead salt (C₉H₈O₂)₂Pb + H₂O crystallizes in small needles.

Cocaic acid is in fact nothing else than the substance that has been meanwhile described by Liebermann as γ-isatropic acid, but the melting point assigned to it by that chemist is incorrect and too high.

In the residue left after treating the crude cocaic acid with lime there was a lime salt of another acid that dissolved readily in ether, and crystallized from the ether solution in long flat prisms. The quantity obtained was too small for analysis.

After removing this acid, the hydrochloric solution left on evaporation a residue of ecgonine hydrochloride. It melted at 234-238 °C with considerable frothing. Liebermann gives the melting point of ecgonine hydrochloride as 246 °C, but the whole of the alkaloid would have been decomposed long before reaching that temperature. This decomposition begins in fact far below 234 °C, about 200 °C, and its commencement is recognizable by the crackling of the crystals. The platinum salt, which is obtainable in several forms as laminæ, prisms and grannular crystals, contained 25.26 percent platinum, and the formula (C₉H₁₅NO₂)₃PtCl₅H₂ requires 24.93 percent.

Professor Stockman informs me that the physiological action of cocamine is similar to that of cocaine, but somewhat weaker, so that it must be given in larger doses; its anæsthetic action is especially weak. Experiments that I made upon myself confirm Stockman's observations, and I wish to point out that I was unable to detect any poisonous effect of cocamine. This is the more remarkable, since Liebermann states that the so-called γ-isatropyl-cocaine does not act like cocaine but as a heart poison. An explanation of this discrepancy will perhaps to be found further on.

Attempts were made to obtain cocamine from the crude bases that remained dissolved by petroleum spirit after the cocaine had been separated from that solution, but without success. It was only after the mass dissolved by dilute sulphuric acid had been treated with permanganate until no further reduction took place that a product was obtained by fractional precipitation, which when boiled with dilute hydrocholoric acid gave for the most part ecgonine and cocaic acid, together with methyl chloride. At the same time there was obtained a minute quantity of a brownish-yellow oil, which resisted the action of the boiling acid. The same result was obtained in operating upon a specimen of isatropyl-cocaine sent to me by Professor Stockman.

When the crude bases obtained by fractional precipitation were dissolved in ether, and the solution evaporated, there remained a syrupy mass, which when heated for time to 60 °C, no longer dissolved completely in dilute acids. But the portion insoluble in dilute acid was dissolved by ammonia. The solution gave white flocculent precipitates with barium chloride and dilute acids. The substance could be obtained by pure repeated solution in ammonia and fractional precipitation with hydrochloric acid. When dried in the air it could be rubbed to a yellowish-white powder that melted near 120 °C, and was readily soluble in ether, ammonia and caustic soda, but not in dilute acids. When heated with with concentrated hydrochloric acid this substance did not give off methyl, but when boiled with dilute acid it melted to a yellowish oil, and after some long time was completely decomposed, yielding ecgonine and an acid that was not more closely examined. The composition of this substance is represented by the formula C₁₈H₁₉NO₄.

It is probable, therefore, that the acid product contained two atoms less hydrogen than cocaic acid, and that its relation to that acid is the same as that of acrylic acid to propionic acid. For that reason I give the acid to the name of cocrylic acid, the substance it is derived from that of cocrylecgonine, and the methylated cocrylecgonine, which would correspond with cocamine, that of cocrylamine. The brownish-yellow oil observed in the decomposition of the so-called isatropylcocaine by hydrochloric acid may be for the most part only cocrylecgonine, and its production may be due to the presence of some cocrylamine in the isatropylcocaine. That moreover is probably the reason that the product obtained by Liebermann from the crude bases possessed poisonous characters.

The liquid from which cocamine had been separated by fractional precipitation by ammonia still contained substances that could be extracted by petroleum spirit and obtained as a tolerably liquid oil that was readily soluble in dilute hydrochloric acid. Platinic chloride added to the solution gave a pale yellow precipitate which contained 19.44 percent platinum after drying at 110 °C. The last fraction gave only 18.22 percent platinum, and therefore the substance was a mixture.

This mixed substance, when boiled with dilute hydrochloric acid, yielded some ecgonine as hydrochloride and an oily non-volatile base, together with some hygrine. The latter was separated by adding baryta and shaking out with ether. There were also obtained cocaic acid, apparently corresponding in quantity to ecgonine, much cinnamic acid and some benzoic acid. The cinnamic acid was separated in the form of a lime salt, crystallizing in small needles very sparingly soluble in cold water and consisting of (C₉H₇O₂)₂Ca + 3H₂O.

The melting point of the acid was 131 °C, or rather less than that observed by Kraut, Liebermann and Frankfeld. With permanganate it yielded benzoic aldehyde. In regard to the acid said to be benzoic it may be mentioned that its melting point was 121 °C, and that it gave the reactions characteristic of benzoic acid.

These observations show distinctly that the oil in question did not contain any cinnamylcocaine. It has already been stated that judging from the platinum salt of hygrine its formula is C₁₂H₁₃N, and that it may be a trimethylchinoline, corresponding to methylated piperdine; but I shall return to this subject on a future occasion.

From the bases of true coca there was obtained, after separating cocaine, a fraction which had the same percentage composition as cocamine, but differed from that substance by its solubility in petroleum spirit and in other particulars. That substance I temed cocaidine. On decomposing it with dilute hydrochloric acid it was however ascertained that a sensible quantity of cocaic acid was formed, as well as other acids and an oily product that was not readily decomposed by hydrochloric acid. Consequently cocaidine is a mixture, but its further investigation must be deferred for want of material. From the solution out of which that mixture had been precipitated by ammonia, oily bases could be obtained as above described, which on boiling with dilute hydrochloric acid behaved in the same manner as those obtained from the other kind of coca, with this difference, however, that the principal part of the acid product in this case was benzoic acid.

From my investigation I consider I am justified in concluding that the "amorphous" bases from true coca consist chiefly of benzoyl compounds of an oily non-volatile base, together with some cocamine, while, on the contrary, those obtained from the other kind of coca - the so-called Truxillo - consist essentially of cocamine and the cinnamyl compounds of that oily base; also that cocamine is in both cases accompanied by another base containing two atoms less hydrogen, which I have named cocrylamine. In both cases the amorphous bases yield some hygrine, but whether it be a product or an educt I have not been able to decide.