292 GHOSH CURCUMIN. XX X.-Curcumin. By PRAPHULLA CHANDRA GIIOSH. THE work described in this paper was in progress and in fact' ccm-pleted before an account of the synthesis of curcumin by Lampe (Ber. 1918 51 1347) appeared. A portion of the work was devoted to proving the presence of .the CO*CH,*CO* group which in any case is now clear from the OMe OMe synthesis confirming the formula, HO/ \CH:CH~CO.C~~,~CO~CH:\-/ \-/ previously put forward by Milobendzki Kostanecki and Lampe (Ber. 1910 43 2163). The results bearing on tnhis point' are there-fore given in an exceedingly abbreviated form. With benzaldehyde curcumin forms a benzylidene derivative and i t also forms a condensation product with nitrosodimethylaniline. The action of bromine on curcumin and some of its derivati-Jes was also studied with the object of testing the presence' of two double bonds.There were obtained mono b romoc urcumin C,,H,,O,Br mono-bromodicarb e thox y curcumin C2,H1704Br (0- CO,E t)2 monob romo-dicarb ethoxycurcumin tetra bromide C2,H,704Br,(O*C02Et)2 cli-bromdicurb e t Jboxy curcumin t c trabromide C2,H,,04Br,( O*CO,Et) , monobromodiacetylcurcumin C21,H,70,Br(OAc)2 dibromodiucetyl-curcumin tetra b romide C2,H1604Br6 ( O h ) mono b romodime t hyl-curcumirc C2,H,,04Br(OMe), and clibromocFimethylcurc~~mi~ tetra-bromide C21H&&k6(OMe)2. In the course of this investigation it was discovered that dicarb-ethoxycurcumin could be converted into dicarbethoxyisocurcumin simply by boiling with acetic anhydride and pyridine. to the view that one of the two known diacetyl derivatives of curcumin was probably diacetylisocurcumin.Ciamician and Silber described a diacetylcurcumin melting a t 169-1 70° whilst JackJon prepared an isomeric substance melting a t 1 5 4 O by acetylating curcumin with acetic anhydride and sodium acetate. It vas suspected that the latter was probably diacetylisocurcumin and this suspicion was found to be correct. Jackson's compound which can also be formed by acetylating aureumin with acetic anhydride and pyridine gives Heller's iso-curcumin (Her. 1918 47 887) on hydrolysis. It is therefore quite Thi GHOSH (YURQUMIN. 293 easy t o convert curcumin into isocurcumin. represented graphically thus : These reactions can be Dicarbethoxycurcumin + Curcumin Boiling with I ace.Dicarbethoxyisocurcumin +- isoCureurnin. This simple conversion of curcumin into isocurcumin confirms Heller's view that the two substances are geometrical isomerides. Frqm the ferric chloride reaction Heller considers that curcumin exists in the enolic form (I) and as there is considerable OMe CM~-CH:CH/-\OH c (0 H ) CH c H ~ \ O H CO-CH:CH/-\OH CO- CH :CH<>OH \-/ \-/ 1, OMe ""\ OMe L/ (1.1 (11.1 resemblance between this formula and that of divanillylidenc-mesityl oxide (11) the latter has been prepared in the course of this investigation by condensing two molecular proportions of vanillin with one of mesityl oxide by means of hydrochloric acid and it is intended ta compare the absorption spectra of this substance with that of curcumin; similarity of absorption spectra would speak in favour of similarity of molecular configuration.The condensation of mesityl oxide with some other aromatic aldehydes has been studied and p-hydroxyb enaylidenemesityl oxide, 0-hydroxybenzylidememesityl oxide piperonylidenemesityl oxide, and cinnamylidenemesityl oxide have been isolated. I n connexion with the geometrical isomerism of curcumin and isocurcumin there is some evidence that cinnamylidenemesityl oxide and piperonylidenernesityl oxide exist in two forms. Most of the compounds examined in the course of this investiga-tion are fluorescent. They are arranged in the following list in the order of their intensity of fluorescence: VOL. cxv. 294 GHOSH aURUUMIN. Substances arranged in gradually diminishing order of fluorescence.Curcumin . . . . . . . . . . . . . . . . . . Benzylidenecurcumin, CHPh:C( CO'CH:CH'C6H,[OMe]'OH), pHydroxybenzylidenemesity1 oxide, OE'C,B;CH:CH'CO'CH:CMe, .................. o . H y droxyben z ylidenemesit yl oxide Dicarbthoxycurcumin, CH,(CO'CH:CH*C,H,[OMe]*O*CO,Et), Diacetylcurcumin, CR,(CO*CH:CH'C,H,[OMe]*OAc), Divanillylidenemesityl oxiclo .............................. Solvent in which the greatest intensity is observed. Chloroform. Acetone. Toluene Toluene. Ace tone. Acetone. Acetone. Monobromocurcumin, Moaobrorno&aarbethoxycuwumin tetrabromide. Dibromo dicarbe t hox ycur cumin tetrabromide, CHBr(CO'CH:CH'C,H,[OMe]'OH) Acetone. CHBr( CO'CHBr'CHSr'C sH,[OMe]'O'CO,Et) Chloroform. CBr,( CO'CHBr'CHBr'C ,H ,[OMe~O*CO,Et) Chloroform.From these observations it may be ipferred that (1) auxochromic groups (2) the labile hydrogen atoms (3) the double bonds and (4) the symmetry of the molecule increase fluorescence butt are not the invariable factors of fluorescence. EXPERIMENTAL. RenzyZide.neczirczcmi,l. C,H,*CH:C,,H,,O,.-A current of dry hydrogen chloride was passed through a mixture of benzaldehyde (0.5 gram) curcumin (1.8 gram) and alcohol (40 c.c.) cooled with ice. The colour became dark violet probably owing to the form-aticm of curcumin hydrochloride and in about a day the whole of the curcumin dissolveld and the colour changed to orange-red. After forty-eight houm the mixture was poured into ice-cold water, when a greenish-yellow precipitate was obtained which could not be crystallised.It was purified by solution in acetic acid and frac-tional pxecipitation by sodium. acetate (sample E) and subsequently in the sitme way by precipitation from an alcoholic solution by water (sample 11). E t is soluble in alcohol acetic acid acetone, or chloroform shrinks slightly a t 170° and melts a t ZOOo: C=73*5; € 3 ~ 5 . 4 . I. 0-1000 gave 0.2695 CO and 0.0486 H,Q. 11. 0.1104 , 0.2965 CO , 04.0527 B,O. C=73.25; H=5*3. CzsH2*0 requires C=73*68; H=5.26 per cent. Condensation Product of Curcumin md Nitrosodime t hylaniline , N( CH3)2,*C6H,'N:C2,Hl,06.-o.ne part of curcumin and one part of nitrosodimethylaniline were dissolved in alcohol two parts of zin chloride were added and the mixture was warmed for a few minutes on the water-bath.When the colour became brown the product was precipitated by water. A good deal of tarry matter was dis-solved by treatment with 30 per cent. acetic acid. The residue could not be crystallised but was purified by fractional precipita-tion from acetic acid solution by water. The substance obtained in this way .is fairly readily soluble in alcohol acetic acid or acetone shrinks a t 98O and melts and decomposes at 144-145O. Two fractions obtained by precipitation with water from acetic acid solution had the same melting point and gave the following results on analysis : I. 0.1502 gave 7.6 C.C. N a t 30° and 760 mm. N=5.7. 11. 0-1430 , 7.1 C.C. N , 29" , 758 mm. N=5*61. C,,H,O,N requires N = 5.8 per cent. Monobromocurczcmh C,,H,,O,Br.-Ha~f a gram of curcumin was dissolved in 35 C.C.of warn chloroform the solution cooled until curcumin began to separate and 4.5 C.C. of a 5 per cent. solu-tion of bromine in chloroform were then gradually added avoid-ing rise of temperature. The solution became quite dear on the first addition of the bromine solution (1 c.c.) and hydrogen bromide was evolved. After ten minutes air was blown through the mix-ture to drive off the chloroform and the residual emulsion was rubbed with a little alcohol. Needle-shaped crystals of mbno-bromocurcumin separated which were soluble in alcohol chloro-form acetic acid acetone or toluene shrank a t 131° and melted a t 136": 0.1187 gave 0:0510 AgBr. Br=18.2. C2,H,,0,Br requires Br = 17.9 per cent. Monobromodicnrbeth~xyczrrczLmii C,,H,,O,Br(O*CO,Et), pre-pared in a similar way from dicnrbethoxycurcumin cryst'allises from much alcohol in needles melting a t 165-170°: 0.1172 gave 0.0382 AgBr.Non80 b romodicarb e t h ox y curcu min t e trab romide, Br = 13.87. C2,H2,010Br requires Br= 13.47 per cent. C,,H1,0,Br,( O*CO,E t)2. -Dicarbethoxycurcumin (1 gram) was dissolved in the minimum quantiby of mld chloroform and treated with 40 e.c. of a 2.5 per cent. solution of bromine in the same solvent. After remaining at the ordinary temperature for one and a-half hours the chlrorobrm was driven off by a current of air. The product was extracted succeasiveliy with boiling ethyl and methyl alcohols leaving a residue which could not be crystallised but was purified by dissolving in chloroform and fractionally precipitating with alcohol.Et is N 296 GHOSH ClURCUMIN. sparingly soluble in alcohol or light petroleum and melts and decomposes a t 209-210°. Different fractions gave the same analytical results : 0.1397 gave 0-1438 AgBr. Br=43*8. Dibromodicarb et hoxycurcumin tetrabromide, C;7H2701,Br requires Br = 43.9 per cent. C21H1,04Br6( Co,E t)2, was obtained by broininating dicarbethoxycurcumin in cold chloro-form solution with excess of bromine. It separated from alcohol in colourless crystals softening a t 148O and melting a t 152O. It remains unchanged on boiling with water : 0.1278 gave 0.1460 AgBr. Br=48-61. C27H,010Br6 requires Br = 48.48 per cent. Monob rom diace t yl curcumin C2,H170,Br (OAc),.-Ciamician and Silber’s diacetylcurcumin was brominated in cold chloroform solu-tion with a little more than the theoretical quantity of a chloro-form solution of bromine (as in the preparation of monobromodi-carbethoxycurcumin) .It crystallises from alcohol in needles melt-ing a t 173-174O and is brighter in colour than diacetylcurcumin itself : 0.1162 gave 0.042 AgBr. Br=15*38. C,H,08Br requires Br= 15-03 per cent. Dibromodiacetylc~rcumilz tetrabromide C2,H,,0,Br6(OAc), was obtained by brominating diacetylcurcumin in cold chlorofom solu-tion with excess of bromine and was crystallised from alcohol. It melts and decomposes a t 80-83O: 0.1477 gave 0.1785 AgBr. Br=51.43. C2,H,,0,Br6 requires Br = 51-45 per cent. Mono b romdm e t h y l cu rczcmin C,,H170,Br (OMe) was obtained by treating dimethylcurcumin in cold chloroform solution with a little more than the theoretical quantity of bromine.It crystallises from alcohol in needles melting a t 140-141° : 0-1200 gave 0.0468 AgBr. Br=16*6. cBH,06Br requires Br = 16-49 per cent. D.i.brolmodimethylczLrculrzi7-L tetrabromide c21H1604Br6(oM0)~.-Dimethylcurcumin dissolved in chlorof o m was treated with excess of bromine when hydrogen bromide was evolved and after half an hour the chloroform was evaporated. This substance could not be crystallised but was purified by precipitation from its hot alcoholic solution with water. Distinct fractions possessed the same melting point (softens a t 96O melts a t 102-104°) and gav QHOSH CURCUMIN. 297 identical analytical results, acid : It is soluble in chloroform or acetic 0.1120 gave 0.1437 AgBr.Br=54.6. c,H,o,Br requires Br = 54.87 per cent. Conversion of Dicarbethoxycurcumk into Heller's Dicadethoxy-isocurcumin. A mixture of 1 gram of dicarbethoxycurcumin 10 C.C. of acetic anhydride and 1 C.C. of pyridine was digested a t the boiling point for forty-five minutes. Alcohol (6 c.c.) and a little water were added to the cooled mixture and tho supernatant liquid was decanted from the viscous mass which thus separated. From the lat.ter by stirring with 2 C.C. of glacial acetic acid a solid product was obtained and this when crystallised from alcohol melted a t 1 4 2 O which is identical with the melting point given by Heller for dicarbethoxyisocurcumin . Found C= 63-0; H =5-5 C,,H,O, requires GI= 63.28 ; H = 5.46 per cent.Acetylation of Curcumin with Acetic ,49thydride and Pyridine. Curcumin was digested with acetic anhydride and pyridine under the same conditions as described in the last paragraph and a solid product isolated. The yellow solid obtained in this way was frac-tionally crystallised from acetic acid. The first fraction (which was ofily a minor portion) melted a t 169-170° and was found to be identical with Ciamician and Silber's diacetylcurcumin (mixed melt-ing point). The second fraction (major portion) melted a t 1 5 4 O , and this it is interesting to note is identical with that given by Jackson for his diacetyl compound obtained by means of acetic anhydride and sodium acetate. Found C=66-4; H=5*23. C,,H,,O requires C = 66.37 ; H,= 5.3 per cent.Dleacetyla'tion of Diacetylczwcumin (m. p. 1 5 4 O ) and Isolation of Heller's isoCurcumin. One gram of this diacetiyl compound was dissolved hi 15 C.C. of acetic acid 1 C.C. of sulphuric acid (D 1-84) added the mixture warmed for a minute cooled and poured into water. The yellow precipit,ate was dried on porous porcelain and extracted with hot benzene. The benzene solution on cooling deposited the substance as a yellow amorphous powder soluble in cold alcohol acet.ic acid, acetone ethyl acetate or chloroform sintering a t 140° and meltin a b u t 2$Oo. identical with Heller's isocurcumin : There could be little doubt that this compound was 0.1130 gave 0.2822 CO and 0.0571 H20. p-Hydroxybenzylidene~esityt oxide, C-68.1; H=5*6.C21Rm06 requires C=68*4; H=5*43 per cent. OH*C&H,*CH CH COO @H C (CH,) 2. -To an alcoholic solution of 2.4 grains of phydroxybenzaldehyde and 1 gram of mesityl oxide 1.5 grams of zinc chloride were added, and the liquid was boiled for two hours. The solution after concentration was poured into water causing the deposition of oily drops which became crystalline on agitation. On recrystal-lisation from alcohol yellow crystals were obtained which melted a t 120° and dyed yelIow shades on chrome-mordanted wool : 0.1210 gave 0.3590 CO a i d 0.0784 H,O. o-HydroxybenzylideiiemesityZ olm'de was prepared in a similar way to the corresponding y-hydroxy-compound. The crude pro-duct separated as an oil which was obtained crystalline' by dissolv-ing in aqueous potassium hydroxide allowing t o remain for a few days and precipitating with hydrochloric acid.The substance dis-solves with a very beauti€ul orange-red colour in potassium hydroxide solution : C=80-91; H-7-2. C13H1402 requires C= 81.25 ; H = 7.36 per cent. 0'1000 gave 0.2970 CO and 0.0667 H,O. Piperony lidenemesit y 1 oxide, C=81*0; H=7*42. C13H14O2 requires C-81.25 ; R='i*36 per cent. CH,~0,:C,H3-CH:C13*CO*CI~:C(CH,),. -Piperonal (2.6 grams) and mesityl oxide (2 grams) were dis-solved in alcohol and alcoholic potassium hydroxide was added t o render the solution fairly alkaline. The1 mixture was boiled for a few minutes when an orange-coloured substance began to separate. On cooling this solid was collected dissolved in acetic acid and reprecipitated with water when i t melted a t 130-135O ( A ) .This proved to be a mixture of two substances melting at' 148-153O and 175-1 78O respectively which can be separated either by fractional precipitation by alcohol from acetic acid or by extraction with alcohol in which the substance of higher melting point is scarcely soluble. The hot alcoholic extract on cooling deposits the sub-stance which melts a t 148-153O ( B ) . The insoluble residue dis-solved in chloroform and fractionally precipitated by alcohol (the first fraction being rejected) melted a t 175-178O (C): (A) 0.1000 gave 0.2662 CO and 0.0558 H,O. c'=72*6; H=6.2. ( B ) 0.1150 , 0'3040 COY , 0.0580 H20. C=72*1; H=5.6. ( C ) 0.1021 , 0'2710 CO , 0.0533 H20. C=72.4; H=6*8. CI4Hl4O3 requires C = 73.0 ; H = 6-08 per cent Ciil.na~~lidene,mesityl oxide, C,H,*CH:CH-CH:CH*CO*CH:C(CH,),, was prepared in the same ~ b .7 as the abom piperobylideae m-p u n d . On coding the reaction mixture some viscous substance was deposited from which the supernatant liquid was decanted and poured into water. The semi-solid mass was dissolved in acetic acid and precipitated with alcohol. It melts and decomposes a t 180-182O (a). Addition of water to the filtrate caused the separa-tion of a second substance which welted and decomposed a t 8 8 O (8). Both the a- and P-cmnpounds were soluble in alcohol chloroform, acetone or toluene but could not be cryst8allised from any of these solvents. Experiment showed that them -pounds were inter-convertible ; the former (a) on boiling with acetic acid for two to three minutes and adding water to the solution gave the &corn-pound melting a t 88O whereas the' latter on boiling with alcohol for about five minutes gave on cooling the solution t h ~ s obtained, a deposit of the a-compound melting a t 180-182°.The ocoompound is less readily soluble in alcohol than the 8 : (a) 0-1082 gave 0.3365 CO and 0.0670 H20. C=84-4; H=6*8. (p) 0.1494 , 0'4610 W2 , 0*'0808 B20. C=84*1; H=6.0. CT,,H,,O requires C=84.4; H=6.4 pet cent. Divainittytidenemesityt oxide (11 p . 293).-A current o€ dry hydrogen chloride was passed into an ice-cooled alcoholic solution of 2.6 grams of vanillin and 1 gram of mesityl d d e the liquid became deep blue. After two days the solution was potired into cold wate? the precipitate collected and triturated with sodium acetate solution. The brown amorphous powder obtaihed in this way 2ouid not be crystallised and was purified by fractional pre-cipitation from acetic acid %ith water solution. It is soluble in alcohol chloroform or acetone softens a t 1 6 5 O and melts a t 178O. Distinct fractions obtained by the above described method gave identical analytical results : I. 0.1200 gave 0.316 CO and 0.067 H20. C=71.8; H=6*2. 11. 0.1065 , 0.282 CO , 0.059 €&O. C=72*21; H=6*1. C22n2205 requires C=72*1; H=6 per cent. I take this opportunity of thanking Prof. B. N. b a s for his kind CHEMICAL LABORATORY, help and encouragement during the progress of the work. DACCA COLLEGE BBNQAL INDIA. [Received December 30th 1 9 1 ~