63 SCIIUNCK AND MARCHLEWSKI : V 11. -So me Derivatives of Anthrquinonc . By EDWARD SCHUNCK, Ph.D., F.R.S., and LEON MARCHLEWSKI, Ph.D. IN our previous communications, we gave a description of three differ- ent artificial methSlparpuroxantllins ; one of these was obtained by the condelisation of p.trat,oluic acid with metndihydroxybenzoic acid and was found by Hummcl and Perkin (Trans., 1894, 65, 854) to be iden- tical with a substarxe obtained by them from the $fang-lioudu roat. Another one was prepared by the action of benzoic acid on met:s- dihydroxj pnratoluic acid ; a comparison of this substance with rubiadin enabled us to point out the constitution of the latter. Finally we prepared a fourth niethylpu~puroxnntliin by the action of orthotoluic acid on metadih37droxybenzoic acid.The confititution i j i these four substances is represented in the following graphic formule obtained by the action of benzoic acid on I. 1 I I loH 9 niet~dii~yr~roxypalrstoluic acid (Trans., co OH / \ / \ / \ c H ~ \A$/ 1894, 65, p. 184). CO OH '""' I I I 10, ' 973, and 1894, 65, 183). V\/'/ rnhiadin from madder (Trans., 1893, 63, CO CH3 CO OH from paratoluic acid and metadihydroxy- 'I1* 1 I I IC)H ' benzdc acid (Trans., 1893, 63, 1137). A/V\ cH3 \A/\/ CO CH3C0 OH /"\'\ from orthotoluic acid and. metadihjdroxy- IV. 1 I I IoH ' benzoic acid. \Ac<\/SOME DERIVATIVES OF ANTH R h QUIXONE. 69 This substance has been prepared by us already a year ago. Up to the present the followirg two methylpurpuroxant~i~is were unknown. CO OH CO OH ~ H 3 A / \ / \ /\/\/\ *' ( I 1 IOH and I I 1 10, \/\/\/ c13:3 co /\Q/ It was to be expected that both these substances would be obtain- able by the action o€ metatoluic acid on metadihydroxybenzoic acid, since the former possessing an asymmetiical structure should be able to react wit,h the latter in two different ways, as shown by the follow- ing equatio:is.OH CO OH /\ CD3/\/\/\. CH/\COOH I 1 I 1 /OH = 2HsO -+ I A* L I +coorI\!oa v\c/o\/ Metrttoluic acid. Metadihydroxp- benzoic acid. OH CO OH B. ('"OH+ COOH, Me tat,oluic Me tadih y dro xy- acid. benzoic acid. The process of condensation was brought about in the following way : metadihydroxpbenzoic acid (1 part), and metatoluic acid (4 parts), were heated with solphuric acid (25 parts) for 10 hours at 110'. The dark brown solntion contaiuing the above mentioned methyl purpuroxan t hins, together with unaltered met ad ihg drox y benzo I c and metatoluic acids as well as anthrachrysone, the latter formed by the condensation of two molecules of the dihydroxy-acid, was poured into water and the whole extracted with ether; the ethereal extract was then evaporated and the residue treated with steam in order to expel unaltered metatoluic acid.Tlie remaining liquid mas again evaporated, and by extracting with boiling benzene, the methylpurpuroxanthins were dissolved ; the separation of the substances contained in this benzene solution is very difficult since the two methylpurpuroxanthins which are foi*med, like the methyipurpuroxanthins in general, have very similar pro- perties. Two substances differing from one another in their melting points may however be obtained in the following manner.The benzene is evaporated and the residue, which melts at ahont 225-235", is dissolved in boiling alcohol ; to this solution, an equal70 SCHUNCR AND JIARCHLEWSKI : volume oE benzene is added, and the mixture, if left, for some time, deposits orange coloured needles ; these are collected and subjccted to the same treatment several times, whilst the first filtrate must be evaporated and the residue treated again in a similar manner. Finally two substances with different melting points are obtained, of which the more easily soluble has the lower melting point; but we think that it would be useless to mention their respective melting points, as we were unable to completely separate the two.(I.) Analysis of the substance with the higher melting point. ([I.) That of the lowela melting point. I . 0.1021 gave 0.2642 CO, and 0.0379 H,O. C = 70.66 ; H = 4.12. 11. 0.0988 gave 0.2561 CO, and 0.0339 H,O. C = 70.60; H = 3.80. I n order to determine the constitution of these two substances, we stndied their products of oxidation. The substance of lower melting point, of which we had st some- what larger quantity, when mixed with dilute nitric acid (1 : 10) and heated on a water bath, gradually dissolved giving a bright yellow solution, which left a yellowish residue on evaporation. This was redissolqed in hot water and lead acetate added ; the white precipi- tate thus produced, being finally decomposed with dilute sulphuric acid, and the mixture extracted with ether.The ethereal solution after evaporation gave nearly white crystals, which on recrystalliua- tion melted a t 216". The acid was evidently a tricarboxylic derivative of benzene, and, judging by the melting point and its general properties, murt have been I : 2 : 4-trimellithic acid, C6H,(COOH),, and the constitution of the lower melting methylpurpuroxanthin is represented therefore by the formula A (p. 69). The body witli higher melting point, when treated in asirnilar way, gave a trace of a substance which melted at about 180°, and in all prohabiiity was hsmimellitl~ic acid, a circunistance which shows that the methylp urpnroxanthin of higher melting point has the constitu- tion represented by formula B. C15Hlo04 requires C = iO.86 ; H = 3.95 per cent.Methy~urpuroxaizthin, [OH : OH : CH, = 1 : 3 : 1'1. A mixtmre of 4 grams of dihydroxgbenzoic acid, 15 grams of ortho- toluic acid, and 200 grams of sulphuric acid was heatsd for 15 hours at 110-120°; the product was t,hen poured into water and extracted with ether; the residue left on evaporating the ethereal extract was suspended in water, and the excess of orthotoluic acid present driven off by means of a current of steam. This was followed by treatment with boiling benzene, arid the solution, on cooling, depo-SOME DERIVATIVES OF ANTI3RAQUINONE. 7 1 sited crystalline needles, which, after being recrystallised three times from hot benzene, melted a t 246O, and yielded the following results on analysis. 0.1254 gave 0.3249 CO, and 0.0484 H,O.It dissolves in alkalis, forming a red liquid. C = 7O 66 ; H = 4.29. Cl5.H,,,O, requires C = 70.86 ; H = 3-93 per cent. The solution in concen- trated sulphuric acid is brownish-red, and shows a narrow absorption band in the red. The substance is easily soluble in alcohol o r ether, and is obtained in orange-coloured, pointed prisms on evaporating the ethereal solution. Acetyl Uwicatire of ~~ethy123iiriourozanthin, [OAc: OAc:CH, = 1:3:1']. In order t o acetlglate the above methylpurpuroxanthin, i t was digested with acetic anhydride and a small fragment of stannous chloride ; the excess of acetic anhydride was then decomposed w i t h alcohol, and, after driving off the ethylic acetate, the residue was crystallised from alcohol. In this way a product was obtained crys- tnllising i n lustrous, silky needles, and melting at 195".It is not changed by cold alkalis, but is decomposed on heating with them. Ethers of ,41iza~in. We proposed some time ago the following formula €or the mon- ethers of alizarin, obtainable by heating alizarin with caustic potash arid an alkylic iodide, CO OH and we pointed out (Trans, 1894, 65, 186) that the natural methyl ether discovered by Humtnel and Perkin (Trans., 1893, 63, 1174) ill the Chay root contains, probably, t'he methyl groiip attached to the oxygen of the a-hydroxy-group. Thia view has been lately adopted by the above-mentioned chemists (Trans., 1895, 67, S19). We think that our suggestion finds experimental support in a work of Lagodzinski (Ber., 1895, 28, 1427) lately published.ThiH chemist obtained a monomethyl ether of alizarin by using hemipinic acid, OCHa72 SCHUNCR AND MARCHLEWSRI : which, considering the kind of reaction adopted, and the properties of the methoxy group adjoining one of the carboxylic groups, must have the same constitution as the ether formed direct by alkylisation of alizarin. The properties of both substances agree in every par- ticular, except as regards the melting point. Lagodzinski finds this to be 201°, whilst we determined it as 228", a new determination giving the same value. This differerice might be due to one of the substances being impurs ; we consider our preparation to be pure. Ethcrs of Aiathraquinoneoxime. I n a preliminary communication (Ber., 1894, 27, 2125), we showed that the oxime of anthraquinone can be easily etherified.It suffices to boil an alkaline solution of anthraquinonegxime with an alkylic iodide; after boiling for two o r three hours, the solution is poured into water, the deposit i s collected, and washed, firstly, with a weak solution of sodium hydroxide. and finally with water. We propose hers to describe these substances more accurately, but before doing so, some remarks coucerning the constitution of these compounds might be of use. The comtitution of the anthraquinoneoxime ethers may be in accord- ance with either of the two following formule, co co ' >o N*R The first formula would show that these ethers are so called oxygen ethers, whilst the second suggests that they are nitrogen ethers ; the method of preparation, however, would in all probability lend to the formation of oxygen ethers, as it is well known that the alkali salts of oximes are generally derivatives of t'rue isonitroso-compounds.Therefore the substitution of the metal by an alkyl group would cause the formation of an oxygen ether. This is supported by the following facts. If the berizyl ether of anthraquinoneoxime is boiled with strong hydriodic acid, benzylic iodide is formed, and is easily detected by its characteristic odour. Accordingly, if the methyl ether is heated with hydriodic acid i n an apparatus simiIar to Zeisel'P, metliylic iodide is given off. Were those compounds nitrogen ethers, benzylamine and methylamine respectively mould be formed.SO3fE DERIVATIYES OF ANTHRAQUINONE. 73 Jfeth yl Ether of ,4nt?wapuinoneoxime.This was obtained by the above-mentioned method. The product, after washing with a dilute watery solution of sodium hydroxide, was crystallised from dilute alcohol, the crystallisation being repeated three times. The methyl ether melts at 147", crystallises in yellowish needles and is easily soluble in organic solvents, especially ether ; the latter solution, when treated with hydrogen chloride, did not give any precipitate, and the residue left on evaporation melted at exactly the same temperature as the pure methyl ether. This fact again con- firms the above-mentioned formula, as it is known that nitrogen ethers of the oximes under these circumstances give hydrochlorides. 0.2091 gave 0.5836 CO, and 0.0899. The molecular weight determined i n the ethereal solution by meam k = 2110.C = 76.12; H = 4.77. Cl5H,,NO2 requires C = 75.94 ; H = 4.64 per cent. of the ebulioscopic method mas found to be normal. 0.2290 gave 46-14 Et. 4 = 9.045". Mol. wt. calculated = 237. 3101. wt. found = 234. Ethyl Ether 01. Anthraqninoneoxirne. The oxime was dissolved in absolute alcohol, a few drops of a cou- centrated solution of potassium hydroxide added, and the mixture boiled for three hours, after adding some ehhylic iodide. The solu- tion soon loses its dark brown colour, becoming pale yellow, and if the tint does not change after adding more potassium hydroxide, the treatment may be stopped, and the solution poured into water. The whole is now extracted with ether, and the residue left on dis- tilling off the ether is crgstallised three times from dilute alcohol. The ethyl ether melts a t a much lower temperature than the methyl ether, namely at 97". 02073 gave 10.99 C.C. of moist nitrogen a t 24' and 762 mm. N = 5.90. C,,H,,N02 require8 N = 5.57 per cent. Benzyl Ether of Bnthrapinoneoxime. The alcoholic alkaline solution of the oxime was boiled with the calculated quantity of beneylic chloride, the solution poured into water, and the benxyl ether extracted by agitation with ether; the residue left; on evaporating the ether was crystallised from alcohol. If the alcoholic solution is too concentrated, an oil seprates, which becomes crystalline after a time ; as these crystals are never quite pure, i t is advisable to prepare! a dilute solution, and induce crystal- VOL. T'XIX. G74 MARSH AND OARDNER: lisation by adding a crystalline fragment of the benzyl ether. The benzyl ether crystallises very well, and is easily soluble i n alcohol, ether, benzene, and chloroform. It melts at 82'. 0.1258 gave 0.3709 CO, and 0.0569 H,O. C = 80.40; H = 5.01. C21H,5N02 requires C = 80.51 ; H = 4-79 per cent.