Absorption Spectra and Chemical Constitution of Organic Substances.— Fifth Interim Report of the Committee, consisting of Professor W. NOEL HARTLEY (Chairman and Secretary), Professor F. R. JAPP, Professor J. J. DOBBIE, and Mr. ALEXANDER LAUDER, appointed to investigate the Relation between the Absorption Spectra and Chemical Constitution of Organic Substances.

THE work of two of the members of the Committee, Dr. Dobbie and Mr. Lauder, has been exclusively devoted to the investigation of certain alkaloids, and the connection between their chemical constitution and their absorption spectra, and the results they have obtained since the last meeting at Belfast constitute the substance of this report.1

NOTE. As sometimes the nitrates of the alkaloids are well-crystallised salts, the examination has been in certain cases made with nitrates. It is necessary to observe, however, that unlike chlorides, sulphates, and acetates, which are very diactinic and exert only a general absorption, nitric acid and the nitrates give characteristic absorption bands.2 This does not affect the spectra here referred to, but it might happen that if the effect of the nitric acid were not taken into account, erroneous conclusions could be drawn from the absorption band of the nitric acid being attributed to the organic base.

In a paper communicated to the Royal Society eighteen years ago by Hartley, it was proved that the principal alkaloids give highly characteristic absorption spectra which can be used for their identification and for ascertaining their purity. Furthermore, that alkaloids closely related to one another, like quinine and quinidine, cinchonine and cinchonidine, all contained a similar nucleus, which was probably formed by the conjugation of four pyridine or two quinoline groups, and that the opium alkaloids had also a characteristically constituted nucleus which is either a benzene or a pyridine derivative. The effect of alkyl and acetyl substitutions on the curve of absorption was demonstrated, the increased intensity of absorption of the apo-derivatives was shown and accounted for, and the occurrence of several oxidised radicals-hydroxyl, methoxyl, carbonyl, or carboxyl-in the constitution of an alkaloid was shown to be capable of causing remarkable differences in the absorption curves of the original nucleus. At the time at which this paper was published, however, little progress had been made with the investigation of the alkaloids, and it was not possible, therefore, to trace any closer connection between their structure and their spectra. In this connection, however, the relationship of the absorption curves to the differences in constitution of quinoline, dihydroquinoline, and tetrahydroquinoline, was determined by Hartley.

The Absorption Spectra of Corydaline, Berberine, and the Opium Alkaloids.

The constitution of the principal members of the group of alkaloids to which corydaline and berberine belong-namely, papaverine, hydrastine,

Dobbie and Lauder, Chem. Soc. Trans., 1903, 83, pp. 605, 626.
Hartley, Chem. Soc. Trans., 1902, 81, and 1903, 83.

3 Phil. Trans., 1885, Part II., p. 471.

narcotine, and narceine-has now been definitely determined, and the examination of this group furnishes a good basis for the study of the relationship between the absorption spectra and the constitution of the alkaloids.

Since papaverine is, in some respects, more simply constituted than the other members of the group, it will be convenient to consider each of the others with reference to it. According to Goldschmiedt, the structure of papaverine is represented by the following formula :

The absorption curve of papaverine shows two absorption bands, the first lying between 1/A 2998 (3335) and 1/A 3295 (λ=3035), and the second between 1/λ 3956 (λ=2528) and 1λ 4555 (λ=2195).

Hydrastine differs structurally from papaverine in the following particulars: (i) The isoquinoline nucleus is partially reduced; (ii) The two methoxyl radicals of the isoquinoline nucleus are replaced by a dioxymethylene group; (iii) A methyl group is attached to the nitrogen atom; (iv) A carbonyl group is attached to the carbon atom (4), and through the medium of an oxygen atom is also linked to carbon atom (2), which has only one atom of hydrogen attached to it. From this comparison, it is obvious that the two substances differ considerably in their constitution. On comparing the curve of the absorption spectra of hydrastine (fig. 4) with that of papaverine (fig. 2), it will be seen that there is a correspondingly wide difference between them; hydrastine exhibits slightly less general absorption than papaverine, and shows only one absorption band which is wider and much more persistent than either of the absorption bands of papaverine. Narcotine only differs from hydrastine in containing an additional methoxyl group attached to ring IV, and the two alkaloids give practically identical absorption spectra (figs. 4 and 5). Assuming the constitution of corydaline, as determined by Dobbie and Lauder, to be correct, it is represented by the second of the following formulæ :

On comparing this formula with that of papaverine, the differences will be seen to consist in the partial reduction of the isoquinoline nucleus and in the presence of carbon atom (5), which, with its associated methyl group, is linked on the one hand to carbon atom (4), and on the other to the nitrogen atom, thus forming a fourth closed chain in the molecule. Here, again, the difference between the absorption spectra and those of papaverine is very marked. The amount of general absorption is less, and there is only one absorption band, which is, however, better defined and more persistent than the papaverine bands (figs. 2 and 6).

In discussing the relations between corydaline and berberine, it is to be remembered that corydaline corresponds to tetrahydroberberine, and berberine to dehydrocorydaline. The constitutional connection between corydaline and tetrahydroberberine is undoubtedly very close,' as a comparison of the above formulæ will show, and between the spectra of the two substances there is also a very close relation (figs. 6 and 7), the only difference being that the general absorption of tetrahydroberberine is slightly greater than that of corydaline.

When papaverine is reduced to tetrahydropapaverine, it is brought structurally very near to corydaline. A comparison of the formulæ of the two substances shows that the former substance differs from the latter in the absence of carbon atom (5) of ring II with its associated hydrogen atom and methyl group. The spectra of the two compounds are almost identical (figs. 3 and 6). Viewing corydaline as derived from tetrahydropapaverine by the addition of CH, forming a fourth closed chain in the molecule, it might have been anticipated that the difference between the absorption spectra of the two substances would be greater than is found to be the case. It should be noted, however, that ring II in corydaline is a reduced ring, and would not therefore exert the same influence on the absorption spectra as the formation of a pyridine ring. It might be expected to produce an effect comparable with that produced by the substitution of a dioxymethylene for two methoxyl groups, which, we shall show later, is slight in compounds of high molecular weight.2

Narceine is the extreme member of this group. It has two benzene nuclei, but no pyridine ring, and in other particulars differs considerably in constitution from papaverine. The absence of any absorption band differentiates the spectra widely from those of the other members of the group (fig. 22).

NOTE. This was accounted for by Hartley in the following explanation: Carbonyl, carboxyl, hydroxyl, and methoxyl on side-chains, or as forming a portion of the substituted benzene nuclei, exhibit great absorptive power, and the occurrence of several oxidised radicals may cause the following variations in spectra: (a) the absorption band becomes so widened as to extend into the region of rays affected by naphthalene, quinoline, and their derivatives; (b) or the absorption is so powerful that it extends to rays less refrangible than those in which the band is situated, and continues so far down the curve that the selective absorption is not made manifest. Narceine appears to be a good example of this; its absorptive power is very great, extending into the

1 Chem. Soc. Trans., 1902, 81, 145.

Hartley, Chem. Soc. Trans., 1885, 47, 691; Hartley and Dobbie, Chem. Soc. Trans. 1900, 77, 846.

1000

region of such low refrangibility as A 3000 when 1 mm. of liquid is examined containing only Tooth of substance, so that no band is visible. The remarks on narceine are also applicable to papaverine in every particular.'1

Dobbie and Lauder 2 have shown that corydaline and berberine give rise to parallel series of derivatives. The absorption spectra of the corresponding derivatives are related to one another in the same way as the spectra of the parent substances. When corydaline is acted on with mild oxidising agents, four atoms of hydrogen are removed, and a yellow substance is obtained, which stands in the same relation to corydaline as berberine to tetrahydroberberine.3

Oxidation with dilute nitric acid converts corydaline and berberine respectively into the dibasic corydic and berberidic acids:

C13H6(CH3)(OCH3)2°N(CO2H)2,
Corydic acid.

C13H7(CH2O2) N(CO2H)2,

Berberidic acid.

whilst oxidation with permanganate gives rise, amongst other products, to corydaldine in the former case, and to w-aminoethylpiperonylcarboxylic anhydride in the latter. The corresponding derivatives differ structurally from one another in the same way as corydaline and tetrahydroberberine, excepting that, in the case of corydaldine and w-aminoethylpiperonylcarboxylic anhydride, ring II having disappeared, the difference between the two compounds is confined to the replacement of the two methoxyl groups of the former by dioxymethylene in the latter. The spectra of the corresponding derivatives (figs. 10 and 11, and 14 and 15), exhibit the same close relationship as those of the alkaloids themselves. The general absorption of the berberine derivatives is, however, always slightly greater than that of the corresponding corydaline derivatives. This is probably due to the influence of the dioxymethylene group, and the correctness of this inference is supported by the fact that piperonylic acid, CH3(CH2O2) CO,H, shows slightly greater general absorption than veratric acid, CH3(OCH3)2 CO2H (figs. 12 and 13).

Whilst the spectra of corydaldine and w-aminoethylpiperonylcarboxylic anhydride approach one another closely, they differ widely from those of cotarnine and hydrastinine (figs. 14, 15, and 16), the corresponding oxidation products of narcotine and hydrastine respectively. The difference finds a sufficient explanation in the fact that whilst all four substances are nearly related, the chain containing the nitrogen atom, which is closed in the two former, is open in the two latter. When, however, hydrastinine is oxidised by means of an aqueous solution of potassium hydroxide, the open chain is closed, and oxyhydrastinine results, the absorption spectra of which substance are almost identical with those of corydaldine and w-aminoethylpiperonylcarboxylic

1 Phil. Trans., 1885.

1903.

2 Chem. Soc. Trans., 1902, 81, 145.

Ibid., 1902, 81, i45.

K

anhydride (figs. 14, 15, and 17). The relationship between these compounds is shown by the following formula:

Though Dobbie and Lauder have found that cotarnine and hydrastinine in alcoholic solution do not possess the constitution commonly assigned to them, this in no way affects the argument, since there is an important constitutional difference between oxycotarnine and oxyhydrastinine on the one hand, and cotarnine and hydrastinine on the other, whatever formulæ be accepted for the two latter.

Again, when the pyridine ring of cotarnine and hydrastinine is closed by the conversion of these substances into their salts or by their reduction to hydro-derivatives, the changes of structure are reproduced in a striking manner in the spectra.

Relationships established between differences in Constitution and Absorption Spectra, which may be applied to the study of Alkaloids of unknown Constitution.

It is now known that many alkaloids which possess the same formula are stereoisomerides. Alkaloids which are related in this way give, like other stereoisomerides, identical spectra. Illustrations of this are afforded by d-corydaline and i-corydaline (fig. 16), narcotine and gnoscopine (fig. 15), tetrahydroberberine and canadine (fig. 17). Quinidine (conquinine) and cinchonidine also give absorption spectra identical with those of quinine and cinchonine respectively, of which substances they are probably stereoisomeric forms (figs. 18 and 19). This relationship might sometimes be used to assist the investigation of cases of suspected stereoisomerism. Where, for example, two compounds of the same formula are known, one active and the other inactive, it may be inferred that they are not optical isomerides if they have different absorption spectra.

A case in point is afforded by canadine and papaverine, which possess the same molecular formulæ but give widely different absorption spectra. Even if it were not known otherwise that these two substances

1 Hartley and Dobbie, Chem. Soc. Trans., 1900, 77, 498 and 509.