COMPOUNDS OF NATURAL YELLOW COLOURIXG MATTERS. 1439 XCI1.-Acid Compounds of Natuml Yellozv Colouring Mcittew. Part IT. BY a. G. PERKIN, F.R.S.E. IN a previous communication (Perkin and Pate, Trans., 1895,67), it was shown that certain yellow colouring matters, when treated with mineral acids in the presence of boiling acetic acid, yield crystalline compounds, the composition of xhich may be generally represented as an additive product of 1 mol. of the colouring matter with 1 mol. of acid ; for inst'ance, the quercetin compounds have the formula, C,,H1,O~,H2SO~, C15H,007,HBr, and Cl,H,oO,,HCl. These substances in the presence of water are decomposed quantitatively into the colouring matter and free acid. I t wqs t4here pointed out that the natural yeZ2ow mordant dye-stuffs, which contain carbon, hydrogeu, and oxygen only, belong as far as their constitution is known principally to three classes, namely, the ketones, xanthones, and phenylated pheno-y-pyrones.Closely allied also to these is the axithraquinone series, which, howeyer, contains no yellow dje-stuff coming under this head.1440 PERKIN : ACID COMPOUNDS co co co co /’\/\/\ A/\A A\/\ /\/\/\ 1 1 1 1 I H I I I I I - \ l l l l \/ \/ yo/\/ \,y-L \A/\/ co Ketone. Xanthone. Plienylated pheno-y-pyrone. Anthraquinone. Of the substances which gave this reaction with acids, quercetin and fisetin were known from the work of Herzig to belong to the third group, whilst it has since been shown that there is little doubt that niorin (Bablich and Perkia, this vol., 792) also belongs to this class.Previously, no member of the second or xanthone group could be obtained for examination, and but one member of the ketones was available, and it was, therefore, pointed out that before attempting to assign a constitution for these compounds with acids, it was ad- visable to study the behaviour of members of all these groups in this respect. With this object, the investigation has been continued from time to time, the isolation or purification of many of the substances employed being necessarily slow. Though some of the results are of a negative character, not only are they interesting theoretically, but they tend to show that the property of reacting with acids is peculiar t o and distinguishes the quercetin group from those other classes of yellow mordant dye-stuffs at present known.Ketone Group-Of this class, “ alizarin yellow A ” (trihydroxybenzo- phenone), ~ 6 ~ , * c o o ~ 6 ~ , ( o H ) , [(OH), = 1 : 2 : 31, and “ alizarin yellow c ” (gallacetophenone), C,H,(OH),*CO*CH,, were cxamined, but no compounds with an acid could be obtained from them. I n the previous paper (Zoc. c i t . ) , i t was shown that maclurin (pentahydroxp- benzophenone) does not combine with acids, and as these three sub- stances may be considered as typical dye-stuffs of this group, it appears evident that the ketone coloixring matters do not yield acid compounds. It is of interest to mention here that catechin from Catechu, kino‘in from Malabar Kiiio, and cyanomaclurin from Artocarpus integrifolia, also do not yield these compounds with acids.The constitution of these substances is not known, but in some respects their properties are similar to those of the ketone group. Xanthone Group.-T hose selected for examination were gentisin, the colouring matter of gentian root (Kostanecki and Tambor, Monatsh., 1895, 15, 1) ; datiscetin obtained from the glucoside datiscin which exists in Datisca cannabina (Schunck and Marchlewski, AnnuZen., 1894, 277, 261), and euxanthone. The constitution of these is given in the former paper. As a result it was found that they yielded no compounds with mineral acids. Various members of the Awthi-aquinone group also gave nega- tive results. The compound of purpuroxanthin with acetic acidOF NATURAL YELLOW COLOURING MATTERS. 1441 :3CI4H8O~,2CzH4O2, (Plath, Ber., 1877, 10, 615), is interesting, but this is rather a case of acetic acid of crystallisation, than of an acid coinpound such as those described in this paper.A somewhat similar product is that from rhamnrtzin (Trans., 1895, 67, 498), C17M1407,C2H102, which I obtained in a similar way by crystallising this substance from acetic acid. Quercetin Grozy.-The colours of this class previously examined were quercetin, rhamnetin, rhamnazin, fisetin, and morin. Since ihen, however, the acid compounds of luteolin (this vol., 208), and rnyricetin (ibid., l287), have been described, and at the same time i t was shown that these colouring matters bear a close relationship t o qnercetin and fisetin. The various compounds of these colouring niatters with acids which have been obtained are given in the follow- ing table.Myricetin, Cl,Bl~Os Qaercetin, CI5Hl0O1 Rliamnetin, C,6H&j Rlismnazin, Cl,H,,Oi Morin, CI,HloOi Luteol in, C1,H1006 Fisetiii, C15H,006 I Those marked with an asterisk have not previously been analysed, ilnd this R as due to the fact that with few exceptions these substances when heated to 100' are decomposed with evolution of acid. As previously shown (Zoc. cit.), this difficulty was avoided in the case of a hydrochloric acid compound (morin hydrochloride) by using a special method of analysis, aud it now appeared desirable to examine, if' possible, another of these unstable componnds in a similar way. Quercetin hydrochZoride, freshly prepared, was strongly pressed in order to remove as much adhering acetic acid as possible, suspended in water, and after standing 12 hours, the regenerated quercetin was collected ; it weighed 0.5552 gram ; the chlorine in the filtrate corresponding to 10.72 per cent. ClJ€1007,HC1 requires C1= 10.48 per cent.Attempts were now made to analyse the hydriodic acid compound of quercetin, b u t with such unsatisfactory results that this was abandoned. This compound is rapidly decomposed on exposure to the air, evolving hydriodic acid even at 40", and is probably very slowly attacked even when washed with acetic acid. Mo&z Hydriodide.-The hydrobromic and hydrochloric acid com- pounds of this colouring matter resemble those of quercetin, but the sulphuric acid compound cliff ers from those of the other colouring1442 PERKIN : ACID COMPOUNDS matters of this group in that during its formation 1 mol.of water is eliminated ; thus its formula is C15H806,H2S01, and not Cl5HlOO,,H2SO4, AS one would be lead to expect. It was interesting, therefore, to examine its bydriodic acid compound, and here again n further dis- tinction between the behaviour of niorin and quercetin was observed, -the cornpound being perfectly stable a t 100". 0.1243 dried a t 100' gave 0,1915 CO, and 0.0295 H,O. C = 42-01 ; H = 2.63. C,,H,,O,,HI requires C = 41-86 ; H = 2.55 per cent. Liiteolin Hydriodide.-In contradistiction to the principal haloid acid compoonds of the colouring matters shown in the above table, those of luteoliri have been previously found (Zoc. cit.) to be exceed- ingly stable, undergoing no alteration even a t the temperature of boiling aniline, and appearing to crystallise with 1H20.It appeared probable, therefore, that its hydriodic acid compound would also be stable, and this was found to be the case. 0.1273 gave 0.2043 CO, and 0.0292 K,O. C = 43.59 ; H = 2.53. 0.1125 ,, 0.1790 CO, ,, 0.0285 H,O. C = 43.39 ; H = 2.79. This compound contains, therefore, no water of crystallisation, and differs in this respect from the hydrobromic and hydrochloric acid compounds. Friedlander and Riidt (Bey., 1896, 29, 878) have lately obtained ct colouring matter which they consider to be the first artificid nieniber of the quercetin series, dihydroxyflavone, CI5Hl0O6,H1 requires C = 43.48. H = 2.65 per cent. by the interaction of chlorogallacetophenone with benzaldehydc i l l the presence of alkali.Discussing the constitution of this dihydroxy- flavone i n a second paper (&id., 1753), they state that like the iiatural colouring mattem it yields with mineral acids beautifully crystalline salts readily decomposed by water ; it, therefore, behaves in this respect like the members of the quercetin group above studied.* The only rernaining member of this class which had not * Kesselkaul and Kostanecki in a later paper (Bey., 1896, 29, 1886) consider that the reaction proceeds in a different manner, the colouring matter being in realit,y not dihydroxyflavone but a benzylideneanhydroglycogallol, c,H~(oH)~<~;>o:cHc~H~. To this Friedlaander and Riidt have not yet replied.OF NATURAL YELLOW COLOURING MATTERS. 1443 yet been examined was chrgsin, the colouring matter of poplar buds, to which the following constitution has been assigned 0 Some of this was prepared according to Piccard's method ( B e y ., 1873, 6, 884) but curiously enough was found to be iizcapable of combining with acids in this way. I n concluding these experiments, it appeared of interest to examine the behaviour toward acids of some substitution products of these colouring matters, and for this purpose quercetin tetramethyl ether, dibromoquercetin, an& tetrabromomorin were selected. The former yields with sulphuric acid, although with dificulty, an exceedingly unstable compound, c r p tallising in orange-red needles, b u t is not acted 011 by hgdrobromic or hydrochloric acid. Its be- haviour is, therefore, identical with that of rhainnetin (quercctin monomethyl ether), described in the previous paper (Zoc.cit.). From di bromoquercetin and fetrabromomorin, acid compounds could not be obtaiiied. In the previous paper (Zoc. cit.), it; Ras stated that these acid coni- pounds on treatment wikh boiling acetic anhydride are decomposed apparently with production of the acetyl compound of the colouriiig matter. To be sure of this, quercetin enlphate was digested with boiling acetic anhydride without the addition of sodium acetate. The substance was quickly attacked, a colourless solution being almost im- mediately formed ; after heating for one hour, this was poured into water, and allowed to stand several days. As the acetylised product separated from the mixture with difficulty, a, little alcohol was added and this greatly facilitated its deposition.The colourless precipitate was collected, and after crystallisation from alcohol, formed colourless needles, melting at 189-191°, and having the properties of acetyl- quercetin. Theoretics 1 Consid era t i o m The results of this investigation show that members of the ketone and anthraquinone group do not yield cornpounds with mineral acids, and that all members of the xztnthone class here examined are also devoid of this property. On the other hand, all members of the quercetin or phenylated pheno-y-pyrone group, with the exception of chrysin, combine with acids, and, further, whereas the methyl ethers of quercetin react only with sulphuric acid, the bromine substitution products of quercetin and morin yield no compounds of this class.1444 PERKIN : ACID COMPOUNDS This property of combining with acids is possessed, as previously described, by hsemate'in and brazilein (Eoc.c i t . ) ; from resacetejin also, Nencki and Sieber ( J . pr. Chem., [2J, 23, 54) have obtained the corn- pounds Cl6Hl,O4,HCl + 2H20 and (CI~H&~)Z,&SO~, but as the true constitution of the substances themselves is yet unknown, they possess at present but little interest. More interesting are the acid compounds of the phthale'ins, of which fluorescein (Baeyer, Annalen, 18'76, 183, 1) yields a sulphate, CzoH1205 + H2SO4 = CzuH,20,sOa + H2Q9 and orcinphthdein (E. Fischer, AnnaEelz, 1876, 183, 63) a hydro- chloride, C22H,sOa,HCl. In a paper on the constitution of fluorescein, Nietzki and Schroter (Bey., 1895, 28,50) describe a hydrochloric acid compound of colourless fluorescein diethyl ether, (I) which salt crjs- tallises in intensely yellow needles, readily decomposed by water.To this they assign the constitution (111, based on the theory that whereas in the free state fluorescein possesses a lactone group, that in the form of its salts and other coloured compounds it has il quinonoid structure, 0 \/ /'roo= A close resemblance in fact can be traced between these compounds and the salts of the triphenylmethane and allied bases. In considering quercetin group, the constitution of the acid compounds of theOF NATURAL YELLOW COLOURING MATTERS. 1445 i t appears to me that their formation may be represented in two ways: (1) either similarly to these phthalein compounds, or (2) by the saturatioii of the ethylene bond in the pyrone ring.Thus, taking fisetin hydrochloride as an example, its constitution mould then be 0 OH 0 OH Fisetin (stable moclification) . Unstable mcidificntion (not known), 0 H OH Hydrochloride. Hydrochloride. In favour of this first formula is the intense orange to orange- red colour of these acid salts, compared with the pale yellow colour of the original substances.* Further, it is possible in this way to account for the non-production OE acid salts from the substituted bromine derivatives. The quercetin compound studied by Herzig was fo:iiid to be peculiar, in that by frequent recrystallisation it was slowly decomposed, and that by reducing agents it could be recon- verted into quercetin. This is no doubt due to the fact that tho bro- mine substitution takes place in the phloroglucinol nucleus, for, as is well known, tribromophIoroglucino1 possesse,s somewhat similar properties.It is very probable that, on bromination, the hydroxyls adjacent to the bromine i n dibromoquercetin assume the ketonic con- dition, the formula in this case being unsusceptible of change under the influence of acids, Moreover, by this formula (I) the decomposition of these acid com- pounds by acetic anhydride can be understood, in that the acetyl compounds of these colonring matters exist only in the stable, colourless, or non-quinonoyd form. The acid compounds of fluoresce’in and orcinphthalein (Zoc, cit.) are, however, more stable thau those of the quercetin group, for the former is not altered by washing with cold water and crystalli- sation from alcohol, and both require hot water i;o effect their decom- * Armstrong, “ Theory of Coloured Carbon Compounds ” (Proc.Chem. Soc.). VOL. LXIX. 5 E1446 COMPOUNDS OF NATURAL YELLOW COLOURINC MATTERS. position. In general properties, however, they appear to be very similar, and the above distinction appears easy to understand when the difference in constitution of these two groups of colouring matters is considered. Fo~rnzcla ZI.---The difficulty with which the ethers of quercetin react with sulphuric acid, and tlieir non-reaction with the halond acids, distinguishes more markedly this class of colouring matter fi-om that of the phthale'in group, in so far as the latter has been examined in this respect.Fluorescein diethyl ether (Zoc. cit.), for instance, readily combines with hydrochloric acid. Taking the con- stitution at present assigned to luteolin, quei cetiii, and rhamnetin as correct, the effect on the stability of the halo'id acid compounds pro- duced by the substitution of the hydrogen in the a-position in the pyrone ring by hydroxyl or niethoxyl is remarkable. 0 OH Luteolin. Quercetin. 0 OH Rhamnetin. For instance (1) all haloid acid cornpoutids of luteolin are stable when heated to 180°, moreover, the hydrochloride and hydrobromide appear to crystallise with lHzO. (2) Quercetin hydrobrotnide is stable at looo, but, when heated to this temperature, the hydrochloride and hydriodide are decomposed with evolution of acid.(3) Rhamnetin does not combine with the haloTd acids. On this account the second constitution above given for these acid compounds suggested itself, namely, that which depends on the satu- ration of the ethylene bond in the pyrone .ring. Experiments carried out in this way on the behaviour of chelidonic and meconic acids, . co COOHAOH COOH,, I II ' 0 towards mineral acids added no support to this theory, for no addi- tive products were t h u s obtained. It is also not'eworthy that neit'herSELL STUDIES ON CITRAZINIC: ACID. 1447 these acids nor the members of the quercetin group form additive compounds with bromine. The strong colour of t'he acid salts of the quercetin group appears so cogent an argument in favour of the quinono'id formula, that of the two it must bc considered by far the mom preferable. I am actively engaged in the study of natural yellow dye-stuffs, with the hope of isolating new members of this class. If such be obtained, the study of their behaviour with mineral acids should throw further light on the nature of this interesting reaction. From the above experiments it appears probable that the colour- ing matters of the xantlione class do not react with acids; but this cannot be absolutely decided until other members of varied constitu- tion are available for examination. Should this be the case, how- ever, this property of forming compounds with acids will be of value in that it can be employed as a means of distinguishing the members of the quercetin group from those other natural classes of ?:on-nitrogenoti,s, yellow, mordant d y e - s h p which are at present known to exist. Clothacorkem' Besearch Laboratory, Yodishire College. Djeing Department,