2062 TUTIN AND CATON : SYNTHESIS OF 2 : 4 : 6-TRIMETROXY-CCXVIL-Tlze Synthesis of 2 : 4 : 6-TrimethoxyphenylA Methyl Derivative 3 : $-Dimethoxystyr,$ Ketone.of Eyiodictyol, Homoeiioclict yol, and Hesperitin.By FRANK TUTIN and FREDERIC WILLIAM CATON.SOME time ago Power and Tutin (Proc. Amer. Pharm. Assoc., 1906,54, 352) isolated from the leaves of Em’odictyon Californicurn(Hooker and Arnott), Greene, two crystalline substances of phenolicnature, which were designated eriodictyol and homoeriodictyolrespectively. Eriodictyol was shown to possess the formula,CI5Hl2O6, whilst homoeriodictyol was found to be isomeric withhesperitin, having the formula Cl,Hl,O,.I n a subsequent communication (Power and Tutin, Trans., 1907,9 1, 887) results were recorded which indicated that homoeriodictyolpossesse the following constitutional formula, :HOHO/-\CH: CH-CO/-\OH.\-/ \-/hleO HOThe amount of eriodictyol at that time available did not permitof many experiments being conducted with this substance, but theconclusion was drawn that it differed from its homologue only byhaving a hydroxyl in the place of the methoxyl group. It wasfurthermore suggested from certain similarities in the properties ofhomoeriodictyol and hesperitin t.hat these two substances differedonly in the relative positions of the hydroxyl and methoxyl groupsin the catechol part of the molecule, and that, consequently, theconstitutional formula, hitherto assigned to hesperitin was incorrect(Tiemann and Will, Ber., 1881, 14, 970; Perkin, Trans., 1898,73, 1037):HOHesperi tin(Power and Tntin)OHHesperitinThe above conclusions have now all been confirmed, inasmuchas, in the preceding paper, it is shown that eriodictyol, homeeriodictyol, and hesperit.in each yield the same product when fullymethylated.Final and conclusive proof of the structure of thesecompounds was, however, desirable, and it appeared that this couldbest be obtained by the synthesis of the fully methylated product.If the views previously expressed regarding the constitution oferiodictyol, homoeriodictyol, and hesperitin be correct (Power andTutin, Trans, Zoc. cit.), then these substances are 2 : 4 : 6-trihydroxy-phenyl 3 : 4-dihydroxystyryl ketone, 2 : 4 : 6-trihydroxyphenyl4-hydroxyl-3-methoxystyryl ketone, and 2 : 4 : 6-trihydroxyphenyl3-hydroxy-4-methoxystyryl ketone respectively. It was thereforesought to synthesise 2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxystyrylketone in order that it might be compared with the compoundobtained on fully methylating the naturally occurring substancesunder consideration. This has been accomplished, and the syn-thetical compound has been found to be in all respects identicalwith the product obtained from the three substances occurring innature, thus affording conclusive proof of the constitution of thelatter compounds.2 : 4 : 6-Trimethoxyacetophenone was prepared by the interactionof acetyl chloride and phloroglucinol trimethyl ether in the presenceof anhydrous ferric chloride, and this ketone was condensed withvanillin methyl ether by means of I' molecular " sodium in a mannersimilar to that employed by Perkin and Weizmann (Trans., 1906,89, 1649), when 2 : 4 : 6-trimethoxyphenyl 3 : 4-dimethoxysty ytketone resulted in good yield:Me0(Tiernnnn and Will).Me0MeO<-)CH :CH-CO/-\OMe + H20.\-/MeO- Me0This substance gave, on heating with aluminium chloride, ahydroxytetramethoxy-compound, which melted at practically thesame temperature as 2-hydroxy-4 : 6-dimethoxyphenyl 3 : 4-di-methoxystyryl ketone described by Eostanecki (Ber., 1904, 37, 793)2064 YUTIN AND CATON : SYNTHESIS OF 2 : 4 : G-TRIMETHOXY-The present authors, however, were unable to obtain from thesubstance prepared by them the acetyl derivative described byKostanecki.Nevertheless, it is considered certain that the hydroxy-tetramethoxy-compound obtained by the present authors is identicalwith that prepared by Kostanecki, inasmuch as the latter authorbias shown that a methoxyl group in the phloroglucinol nucleussituated in the 2-position with respect to the side-chain is mosteasily hydrolysed by aluminium chloride.2 : 4 : 6-Trimethoxyacetophenone was prepared by Kostanecki(Ber., 1899, 32, 2262) in the manner already mentioned, but it hadpreviously been stated by Friedlander and Schnell (Ber., 1897,30, 2150) to result from the interaction of phloroglucinol trimethylether and acetyl chloride in the presence of aluminium chroride.The present authors, however, could succeed in preparing it onlyby Kostanecki's method, and, when employing that of theearlier investigators, obtained a hgdroxpdk ce t yldim e t hox y b enz ene,C18H1405-EXPERIMENTAL.Met hylation of Phloroglucinol.Will (Ber., 1888, 21, 603) obtained phloroglucinol trimethylether by the action of sodium and methyl iodide on the dimethylether, the latter being prepared by passing anhydrous hydrogenchloride into a methyl-alcoholic solution of phloroglucinol.Thepresent authors, with the endeavour to simplify this process, soughtto obtain the trimethyl ether by the direct methylation of phloro-glucinol with methyl sulphate and potassium hydroxide. Thismethod, however, did not give satisfactory yields of the desiredsubstance, a considerable part of the material being converted intoa neutral oily product.The latter distilled at 140--145O/13 mm.,or at 258-266O under the ordinary pressure. It was unchanged byfurther treatment with methyl sulphate and potassium hydroxide,but was not further investigated. The method employed by Will(Zoc. cit.) was therefore adopted, but with the employment ofmethyl sulphate instead of methyl iodide. I n this way satisfactoryyields of phloroglucinol trimethyl ether (m. p. 52O) were obtained.Action of Acetyl Chloride and Aluminium Chloride onPhloroglucinoi Trimethgt Ether.Friedllnder and Schnell (Zoc. cit.) treated phloroglucinol tri-methyl ether with acetyl chloride and aluminium chloride, andobtained thereby 2-hydroxy-4 : 6-dimethoxyacetophenone and thecorresponding trimethoxy-ketone.The present authors, however,when employing the method of these investigators, could isolatPHENYL 3 : 4-DIMETHOXYSTPRYL KErONE. 2065from the reaction mixture only a small amount of 2-hydroxy-4 : 6-di-me thoxyacet op henone, unchanged p hloroglucinol trimet h yl ether,and a htydroxydiacetyldimethoxybenzene, CI2Hl4O5, no matter howthe conditions of the experiment were varied.A quantity (11.5 grams) of phloroglucinol trimethyl ether wasdissolved in carbon disulphide, 6 grams of acetyl chloride added,and then 10 grams of powdered aluminium chloride graduallyintroduced. After heating the mixture f o r one hour the solventwas decanted, and the residue decomposed with ice and hydrochloricacid. The product so obtained was dissolved in ether and shakenwith a solution of potassium hydroxide, which removed the greaterpart of the material.The neutral product remaining in the etherconsisted almost entirely of unchanged phloroglucinol trimethylether. The alkaline extracts were acidified, and the precipitatedproduct was crystallised, first from ether, and then from alcohol,when it formed long, slender prisms, melting a t 127-128O :0.1090 gave 0.2407 (20, and 0.0583 H,O. C = 60.2 ; H = 5.9.0.1130 ,, 0.2491 CO, ,, 0.0595 H,O. G=60'1; H=5.9.0.1316 ,, 0.2924 CO, ,, 0'0714 H,O. C=60*6; H=6.0.OHo~6H2(O~e),(OO*CH3) requires c"= 61.2 ; H = 6.1 per cent.OH*C,H(OMe),(C'O*CH,), ,, C = 60.5 ; H = 5.9 ,,OH-@6(OMe)2(CO*CH3)3 ,, C = 60.0 ; H = 5.7 ,,These analyses indicate that the substance under considerationwas a hydroxydiacetyldimethoxy6 enaene, OH°C6H(OMe),(CO*CH3)2,and this conclusion was subsequently confirmed by the analysis ofits benzoyl derivative.No direct proof of the constitution of thediketone was obtained, but it would appear most probable that itis represented as follows:OH/\CO*CH,CO*CH,CH,*O\/ I O*CH, .This formula is preferred to the other possible one, as it is knownthat 2 : 4 : 6-trimethoxyacetophenone, which must first be formed,readily yields 2-hydroxy-4 : 6-dimethoxyacetophenone in the presenceof aluminium chloride, and a second acetyl group, when enteringthe nucleus of the latter ketone, would assuredly enter the par%position with respect to the hydroxyl group.Hydroxydiacetyldimethoxybenzene is sparingly soluble in ether,and moderately so in alcohol.It did not lose either of the methylgroups when treated with anhydrous ferric chloride or withaluminium chloride, and attempts to introduce a third acetyl groupinto the nucleus of it were unsuccessful2066 TUTIN AND CATON : SYNTEIESIS OF 2 : 4 : 6-TRIMETHOXY-Acetoxydiacetyldirnethoxybenzene, OAC-C~H(OM~)~(CO*CH~)~.-The acetyl derivative of the above-described compound was preparedby means of acetic anhydride. It crystallised from its concentratedsolution in this solvent, and was purified by recrystcallisation frombenzene. A cetoxydkcetyldimethoxy b enzene forms well-definedprisms, melting at 150-5O :0.1132 gave 0.2409 CO, and 0.0572 H,O.C14Hl606 requires C = 60.0 ; H = 5.7 per cent.This subst4ance cannot be crystallised from ordinary alcohol orother hydrous solvents, as it is rapidly hydrolysed when dissolvedin such liquids.Benzoyloxydkcetyldmethoxybenzene, OBz*C6H(OMe),(CO.CH3)z.-This compound was prepared by the Schotten-Baumann reaction.It crystallised from benzene in cubes, and from absolute alcohol inslender prisms, melting at 179O:C = 60.0 ; H = 5.6.0.1132 gave 0.2768 C'O, and 0.0551 H,O.C19H2206 requires C = 66.7 ; H = 5.3 per cent.The analysis of this compound conclusively confirms the resultsobtained by the analysis of the original hydroxy-diketone, inasmuchas the related compounds containing one and three acetyl groupswould require C= 65.6 and C = 68.0 per cent.respectively.Hydroxydiacet yldirne t hoxy b enz enephenylh ydrazone.- One gramof hydroxydiacetyldimethoxybenzene was dissolved in a smallamount of alcohol, and slightly more than two molecular proportionsof freshly distilled phenylhydrazine dissolved in acetic acid added.The mixture was heated on a water-bath for a quarter of an hour,when, after concentration, a crystalline product separated. Thiswas collected, and crystallised many times, first from acetic acid,and subsequently from a mixture of ethyl acetate and alcohol, whenthe melting point gradually rose from 215O to 230°, the productfinally obtained forming hexagonal prisms melting sharply at thelatter temperature :C=66*7; H=5*4.0.0927 gave 0-2248 CO, and 0.0529 H,O.0*1012 ,, 8.4 C.C. N, (moist) at 19O and 753 mm. N=9*4.Cl,H,,04N, requires C = 65.8 ; H = 6-1 ; N = 8.5 per cent.This compound was therefore a mowphenylhydrazone of hydroxy-diacetyldimethoxybenzene.The original ethereal mother liquors from the hydroxydiacetyl-dimethoxybenzene yielded a small amount of a substance which,when crystallised from alcohol, melted at 80°:0.0754 gave 0.1685 CO, and 0.0425 H20.C= 66.1 ; H = 6.3.C = 61.2 ; H = 6-3.CloH,,O, requires C = 61.2 ; H = 6.1 per centPHENYL 3 : 4-DIMETHOXYSTYRYL KETONE.2067This substance was evidently 2-hydroxy-4 : 6-dimethoxyaceto-phenone, described by FriedlSnder and Schnell (loc. cit.).Since no trimethoxyacetophenone could be obtained by means ofthe aluminium chloride reaction, recourse was had to the use ofsublimed ferric chloride (Eostanecki, Ber., 1899, 32, 2261).Bythis means it good yield of product was obtained, which consistedlargely of 2 : 4 : 6-t.rimethoxyacetophenone (m. p. 99-loo6), butcontained a little of the previously described hydroxydiacetyl-dimethoxybenzene.Condensation of 2 : 4 : 6-Trimethoxyacetophenone with VanillinMethyl Ether. Formation of 2 : 4 : 6-T~imethoxyphenyZ3 : 4-Dimethoxystyryl getone.Vanillin was methylated by means of methyl sulphate in themanner described by Perkin and Weizmann (Trans., 1906, 89,1649), when vanillin methyl ether was readily obtained. Onemolecular proportion of the latter compound was then heated in dryethereal solution with equivalent amounts of 2 : 4 : 6-trimethoxy-acetophenone and finely divided sodium.* Hydrogen was slowlyevolved, and, after six hours, a yellow solid commenced to separate.When the reaction was complete, which was after about twenty-fourhours heating, the ether was decanted, and the residue dissolved inbenzene and washed with water.Considerable difficulty wasoriginally experienced in causing the yellow, viscid residue obtainedon evaporating the benzene to crystallise, but this was eventuallyeffected from alcoholic solution, after which subsequent batches ofmaterial quickly crystallised on inoculating with the solid firstobtained. 2 : 4 : 6-Trimethoxyphenyl 3 : 4-dimethoxystyryt? ketoneforms stout, pale yellow prisms, which contain one molecule ofalcohol of crystallisation, and melt at 85O. The anhydrous substancemelts a t 117.5O :0.1158 t gave 0.2778 CO, and 0.1158 H20.C=65-4; H=7*1.0.1436 ,, 0.3512 CO, ,, 0'0796 H20. C = 66.7; H = 6'1.l*OOlO,t on heating at looo, lost 0.1140 EtOH- EtOH = 11.4.C2,H,0,,EtOH requires C = 65.3 ; H = 6.9 ; EtOH = 11.4 per cent.C,,HB0, requires C = 67.0 ; H = 6.1 per cent.2 : 4 : 6-Trimethoxyphenyl 3 : 4-dimethoxystyryl ketone is veryreadily soluble in benzene, xylene, chloroform, or acetic acid,moderately so in alcohol or carbon disulphide, and almost insolublein ether or petroleum. It also dissolves readily in moderatelyconcentrated hydrochloric or sulphuric acids, yielding deep red* Prepared by violently shaking molten sodium under xylene, the latter beingsubsequently removed by the addition of pure ether, and decantation.t Air-dried.$ Anhydrous substance2068 TRIMETHOXYPHENYL DIMETHOXYSTYRYL KETONE,solutions. It distils without decomposition at 325y 13 mm., solidify-ing to a yellow, vitreous mass.A quantity of the ketone was heated for five hours with a largeexcess of a concentrated alcoholic solution of potassium hydroxide.The greater part of the material was recovered unchanged afterthis treatment, but some veratric acid was isolated from the reactionmixture. It thus appears that, under the influence of potassiumhydroxide, the ketone in question is attacked at the double linking,just as has been shown, in the preceding paper, to be the case withmonomethylhomoeriodictyol. The first product of this change is,doubtless, vanillin methyl ether, whicb yields the veratric acid bythe further action of the alkali.Action of Aluminium Chloride on 2 : 4 : 6-Trimethoxyphenyl3 : 4-Dim.ethoxystyry2 Ketone.With the endeavour to remove methyl groups from the above-described ketone, in order to see if, in this way, homoeriodictyolcould be obtained, the effect of ferric and aluminium chlorides onthe pentamet.hoxy-compound was investigated.Ferric chloride wasquite without action on it, but when heated with aluminium chlorideit slowly lost one met-hyl group. The change was by no meanscomplete, even after many hours’ heating, but the hydrolysedproduct was readily separated from the unchanged material byboiling the mixture with successive portions of dilute aqueouspot,assium hydroxide. The alkaline liquids so obtained yielded, onacidification, a yellow compound, which crystallised very readilyfrom alcohol in deep golden-coloured plates, melting at 154O :0.1122 gave 0.2724 CO, and 0.0584 H20.C19H2006 requires C = 66.3 ; H = 5.8 per cent.This compound is doubtless identical with the 2-hydroxy-4 : 6-di-methoxyphenyl 3 : 4-dimethoxystyryl ketone (m. p. 157O) preparedin another way by Kostanecki (Ber., 1904, 37, 793), but the presentauthors were unable to obtain the crystalline acetyl derivativedescribed by him, although several attempts were made.C = 66.2 ; H = 5.8.THE \ITELLCOME CHEMICAL RESEARCH LABORATORIES,LONDON, E.C