PURFURAN DERIVATIVES FROM BENZOIN AND PHENOLS. 1035 CII.-Fu?$wan Der.iz?atives jaom Bemoin and Pheizols. By FRANCIS R. JAPP, F.R.S., and ANDREW N, MELDRUN, B.Sc. BY the action of cold, concentrated sulphuric acid on a mixture of benzoin and phenol, Japp and Wadsmorth (Trans, 1890, 58, 965) ob- tained paradesylphenolsulphonic acid, from which, by hydrolysing it with strong hydrochloric acid at 1 50°, paradesylphenol, (1 :4), CoH5*QH*Co ' H *OH C6€15*C0 was prepared. We now find that by heating a mixture of benzoin and phenol with rsulphuric acid of 73 per cent. strength,* the foregoing condensation occurs mithont sulphonation of the resulting compound, and an * Corresponding with the hydrate H,SOJ,2H,0. A sulphuric acid of this strength was first used for coildensations by Bistrzycki n u 1 Oclilert (Be?.., 1894, 27, 2632). 3 2 21036 JAPP AND MELDRUM: FURFURAN excellent yield of paradesylphenol is obtained. The various condensa- tions about to be described were effected by means of sulphuric acid of this strength. If the desyl group were to take up the ortho-position towards the hydroxpl of the phenol, it is evident that, by a further elimination of water, a furfuran derivative might be formed : With phenol and benzoin, no such reaction occurs : the pam-com- pound alone is formed. With thymol and benzoin, however, a mixture of desylthymol and cymodiphenylfurfursn is obtained. I n the case of resorcinol and of quinol, either one or two diphenylfurfuran groups may be introduced. With phloroglucinol, only the compound contain- ing three such groups was obtained.Up to a certain point these reactions resemble those studied by Hantzsch and his pupils (Bey., 1SS6, 19, 1290, 2987, and 2934), in which furfuran derivatives were obtained by the action of ethylic a-chlorscetoacetate on the sodium compounds of phenols. Starting with phenol, a benzomonof urfuran derivative was thus prepared ; whilst from di- and tri-hydroxybenzeneP, compounds containing either one, or two, or three furfuran groups were obtained, just as in the reactions described in the present paper, But there is an important distinction t o be drawn between the mechanisms of the two sets of reactions. I n Hantzsch’s reactions, the first stage is the formation of an ether of the phenol in question-thus of ethylic a-phenoxyl- acetoacetate in the first of the syntheses referred to-and the linking of carbon to earbon is a subsequent process.I n the reactions here de- scribed, the first stage is tho linking of the desyl group to the nucleus of a phenol, whilst the closing of the furfuran ring by oxygen follows. Were this otherwise, phenol and benzoin mould yield only diphenyl- benzofurfuran, instead of, as actually happens, only parsdesylphenol. We may mention that the present research was completed last year, before the publication of Graebe’s investigation of benzoin yellow (Ber., 1898, 31, 2975), in which it is shown that, by the action of sulphuric acid on a mixture of benzoin and gallic acid, a compound containing a diphenylfurfuran group is formed.DERIVATIVES FROM BENZOIN AND PHENOLS.1037 EXPERIMENTAL. 1. Benzoin and Phenol. 20 grams of benzoin, 9 grams of phenol," and 80 grams of 73 per cent. sulphuric acid mefe heated by means of an oil-bath at 120-150° for 20 minutes, frequently shaking the flask. The product, which was dark-coloured, was allowed to cool, the aqueous portion was poured off, and the organic substance was boiled, first with water and then with a. solution of sodium carbonate, after which it was re- crystallised from a mixture of benzene and light petroleum. It melted constantly at 1333 and exhibited all the other properties of para- desjlphenol (Trans., 1890, 57, 966). The yield was good, and the method is a great improvement on that previously described (Zoc. cit.). Afterwards, when me had ascertained that various other phenols yielded furf a r b derivatives, a second preparation of the foregoing compound was made in order to ascertain whether any benzodiphenyl- furfuran mas formed at the same time.For this purpose, the operation was conducted as just described, except that the crude product, after extraction with sodium carbonate, was dissolved in ether, and the solution shaken with dilute caustic soda as long as the latter removed anything. Any benzodiphenylfurfuran would thus remain behind in the ether; but, on evaporating the ethereal liquid, only a small quantity of an uncrystallisable resin was obtained, closely resembling the product of the action of sulphuric acid on benzoin alone. 2. Benzoin and TILynaol. 20 grams of benzoin and 40 grams of thymol were melted to- gether, 80 grams of 73 per cent. sulphuric acid were added, and the mixture was heated, with shaking, for 20 minutes a t 150-170".The viscid product mas washed with water, and then steam-distilled as long as any thymol passed over. The dark-coloured solid which remained was dissolved in boiling alcohol. The filtered solution deposited, on cooling, ft resinous substance; the liquid, poured off from this, gave, on standing, cauliflower-like aggregates of crystals (A). The filtrate from these contained, along with much resin, 8 very soluble, crystalline substance : this was obtained, by crystallisation from a mixture of alcohol and light petrolcum, in colourless lamins melting at 126'. It is readily soluble in alcohol and benzene, sparingly soluble in light petroleum.Analysis gave figures agreeing with the C,H,* YH* C,H,(CH,)( C1,HII,)OH C,H,* CO formula of a desyZthpaol, * From our expericncc with other phenols, it is probable that the employmcnt ot an excess of phenol in this experiment would have given a still better result,1038 JAPP AND MELDRUM : FURFURAN 0.1591 gave 0.4882 (30, and 0.0999 H,O. C = 83.69 ; H = 6.98. 0.1778 ,, 0.5462 CO, ,, 0.1121 H20. C=83*78; H=7*01. C,,H,,O, requires C =: 83.72 ; H = 6.98 per cent. The position of the desyl group is uncertain, excopt that it is doubt- less not contiguous t o the hydroxyl, otherwise the compound would be uonverted, with elimination of water, into cymodiphenylfurfuran. Although desylthymol contains a phenolic hydroxyl group, it is almost insoluble in aqueous caustic alkali.A similar phenomenon has been observed in the case of Mazzara's benzylthymol and dibenzyl- thymol (Abstr., 1882, 42, 171), which are, however, quite insoluble in aqueous caustic alkali, although the benzyl groups replace hydrogen of the nucleus, I n the case of desylthymol, the addition of a very little alcohol to the aqueous caustic alkali causes the substance to dissolve -a property which might possibly be utilised in separating it from the cymodiphenylfurfuran (vide infra) formed along with it in the process of preparation. The presence of a hydroxyl group in desylthymol may be readily proved by boiling the compound with acetic anhydride, when it forms a monacetyl derivative which cry stallises from alcohol in colourless needles, melting at 110".On analysis, i t gave figures agreeing with H-c,H,(CH,)(C~H~)O.C,H,O the formula of cccetyZdes?/Zthy?)toZ, . 6 5 0.1503 gave 0.4431 CO, and 0.0918 H,O. C,,H,,O, requires C = 80.83 ; H = 6-74 per cent. The cauliflower-like aggregates of crystals (substance A, wide SU~TCG), obtained in the crystallisation of the crude product of the condensation, were washed with cold, and then recrystallised from hot, alcohol. As they were found to contain alcohol of crystallisntion, which was diffi- cult to expel completely, they were recrystallised from light petroleum. In this way, the substance was obtained in tufts of slender needles melting at 115-116". I t sublimes slowly when heated on the water- bath. It gave figures agreeing with the formula of a cymodiphnt&'h- CH, C=80*40; H=6*79.0.1895 gave 0.6109 CO, and 0.1171 H20. C = 87-95! ; H = 6.87. 0.1614 ,, 0*5308 00, ,, 0.1007 H,O. C =88.00 ; H = 6.93, C24'H,,0 requires C = 58.34 ; H = 6.74 per cent,DERIVATIVES FROM BENZOIN AND PHENOLS. 1039 3. Benzoin and Pyrocatechol. The dihydric phenols interact either with one 01' with two molecules of benzoin, yielding respectively mono- and di-furf uran derivatives. The former are soluble, the latter insoluble, in caustic alkalis. The result obtained with pyrocatechol was so unsatisfactory that me refrain from describing in detail the process employed. The greater part of the product dissolved in caustic soda, but the reprecipitated substance was so readily oxidisable that we were unable to purify it. A small portion, insoluble in caustic soda, was recrystallised from benzene, and was thus obtained in tufts of slender needles melting at 237O.The results of analysis left no doubt that it had the expected formula C34H2202, but the values for carbon were low, and the quantity of substance was insufficient for further purification. Found: C, 87.69, 87-53; H, 4.83, 4.80. Judging by the results obtained with the other dihydric phenols, the compound is o~t~obenxotetrc6p~en~~d~urfuri~n, Calculated : C, 88-31 ; H, 4-76 per cent. /--\ C,H,* C()--f>*C6H5. C,H,\C- 0 0 -C-C,H, 4. Benzoin and Eesorcinol. I n the reactions with the dihydric phenols, it was found advan- tageous to employ different proportions of the interacting substances according as the mono- or the di-furfuran derivative was required. In the former case, a large excess of the phenol-3 mols.of the latter to 1 of benzoin-was used ; in the latter, equimolecular proportions of the two substances-this being again an excess of the phenol, from the point of view of the desired reaction-gave the best result. 40 grams of benzoin and 62 grams of resorcinol were melted to- gether, 160 grams of 73 per cent. sulphuric acid were added, and the mixture was heated, with shaking, for 15 minutes at 120-130°. The dark coloured product, which solidified on cooling, was washed with water, and boiled with dilute caustic soda, in which it nearly all dis- solved. The filtered solution deposited R sodium salt in six-sided laminae with a satiny lustre ; these, when filtered off, cohered, forming a greasy mass. As the salt was pink coloured, it was recrystallised from boiling water, with the addition of a little caustic soda t o pre- vent the hydrolysis which otherwise occurs, until it was practically colourless.It gives a colourless, aqueous solution with a yellowish- green fluoresceuce, resembling that of uranium glass, It contains1040 JAPP ARD MELDRUM: FURFURAN water of crystallisation, which is entirely driven off at 110'. The anhydrous salt gave, on analysis : Na = 7.86, Required for C,,H,,O,Na : Na=7*47 per cent, The excess was doubtless due to the fact that the substance was deposited from a solution containing caustic soda. The substance could not be contaminated with the sodium compound of resorcinol, as that compound is very readily soluble in cold water, whereas the new sodium salt is only sparingly soluble.The potassium salt, on the contrary, is readily soluble. I n order to obtain the phenolic substance, the sodium salt was dis- solved in hot, dilute alcohol, and the boiling solution was acidified and poured into .excess of boiling water. The organic substance separated as a dark coloured, thick oil, which did not solidify on standing. I t s purification presented difficulties, owing to the tendency of the substance t o separate in an amorphous form from its solutions. It mas finally purified, although not without considerable loss, by dis- solving i t in chloroform and adding light petroleum gradually, so as to precipitate resinous impurities ; the colourless solution thus ob- tained yielded tufts of white needles with a satiny lustre, melting con- stantly at 117.5'.It is readily soluble in most orgmic solvents ; the alcoholic solution shows a violet fluorescence. Analysis gave figures agreeing with the formula of metahydroxybenaocliphenylfurfurnn, c,H,(oH)<~ C(C 0 H 5 )'C*C,H5. 0.2084 gave 0.6389 CO, and 0.0969 H,O. C = 83.61 ; H = 5-16. 0.1516 ,, 0.4655 CO, ,, 0.0687 H,O. C=83*74 ; H=5*03. C,,H,,O, requires C = 83-92 ; H = 4.89 per cent. OE the two possible isomeric hydroxybenzodiphenylfurfurans de- Judging rivable from resorcinol, only the !oregoing was observed. from analogy, its constitution is probably 0 Acetyt Derivative.-4 grams of the foregoing substance were boiled with 6 grams of acetic anhydride for 2 hours, and the resulting acetyl derivative was purified by recrystallisation, first from a mixture of benzene and light petroleum, then from alcohol, from which it was deposited in slender prisms melting at 117'.The practical coinci- dence of the melting point with that of the original compound is anomalous, inasmuch as acetyl derivatives in which acetyl replaces hydrogen of a hydroxyl group generally melt much lower than the hydrox y-coin pounds. Analysis of the substance, dried at loo', gave figures agreeing with the for mula of metacstox~beizzo~~p~~n y Zjhfuran,DERIVATIVES FROM BENZOIN AND PHENOLS, 1041 C,H,(OC,H,0)<C(C6H5~: 0- c! 'eH5. 0.2314 gave 0.6821 CO, and 0.1078 H,O. C = 80.39 ; H = 5.18. C,,HIGO, requires C = 80.49 ; H = 4.88 per cent. Metc~benxotets.a~,l~n~Zdiq~u~,c~n, C,H, grams of benzoin and 10 grams of resorcinol were melted together, 80 grams of 73 per cent.sulphuric acid were added, and the mixture was heated at 160' for 15 minutes, shaking it from time to time, and gradually adding a further 10 grams of benzoin. The melt, which became verytarry, was cooled, washed with water, and boiled with strong caustic soda; the mixture was then diluted with water,and filtered hot, so as to remove phenolic compounds. The insoluble residue was boiled up with a little alcohol, allowed to cool, and then filtered ; this removed a considerable amount of colouring matter. The substance was then crystallised several times from hot benzene, from which it was deposited in short, oblique prisms or plates, containing benzene of crystallisation, and finally from hot glacial acetic acid, from which i t separated in tufts of colourless, slender needles, melting at 217-219O.It is fairly soluble in hot benzene or glacial acetic acid, readily soluble in chloroform, only sparingly soluble in alcohol. All the solutions showed a violet fluorescence. Analysis gave figures agreeing with the formula of metabenzotetrc~~~en?/Zda~urful.an, 0.1380 gave 0.4462 CO, and 0.0610 H,O. C = 88.00 ; H = 4.91. 0.1443 ,, 0.4669 CO, ,, 0.0661 H,O. C=88.24; H=5*10. CY4H2302 requires C = 88.31 ; H = 4-76 per cent. 6 . Benzoin und Quinol. HOi\ C*C,H, 0 I J-8.C H .--20 \/\/ Paral~?/dro~ybencoc2;~~rien~lfu?~ur~n, grams of berizoin and 3 1 grams of quinol were melted together, SO grams of 73 per cent. sulphuric acid were added, and tho whole was heated, with shaking, a t 120-150° for about 20 minutes, gradually raising the temperature.The solid product was powdered and then thoroughly extracted with boiling water to remove quinol. It was then treated exactly like the corresponding resorcinol derivative, by dissolving it in caustic soda (when a considerable residue of parabenzotetraphenyl- difurf uran remained undissolved), purifying the sodium salt by recrys-1042 FURFURAN DERIVATIVES PROM BENZOIN AND PHENOLS. tallisation, liberating the phenolic compound, and pnrifying the latter by recrystallisation, first from a mixture of benzene and light petroleum, then from chloroform and light petroleum. It was thus obtained in rosettes of colourless, flat needles melting a t 158-160'. It crystal- lises much more readily than the meta-componnd.Analysis gave figures agreeing with the formula of l)C11.aji?JcEi.ox?lBenzod~~?~~~~- furfuvan, 0.1724 gave 0.5285 CO, and 0.0802 H,O. c! = 83.61 ; H = 5-17. 0.1505 ,? 0,4613 GO, ,, 0.0671 H,O. C=S3*59 ; H=4*95. C,oHI,O, requires C! = S3.92 ; 11 = 4.89 per cont. Jcetgl Derivative.-Half R gram of the foregoing compound was boiled with 2 grams of acetic anhydride for 3 hours, and the product recrystallised from alcohol. It formed laminx!, melting a t 137'. Analysis showed that i t was a monacetyl compound 0.1247 gave 0.3669 CO, and 0.0558 H,O. C = 80.24 ; H = 4.97. C,,H,,O, requires C = 80.49 ; H = 4.48 per cent. of benzoin and 10 grams of quinol were melted together, 80 grams of 73 per cent. sulphuric acid were added, and the mixture was heated at 150° for 15 minutes with frequent shaking.The resinous mass thus obtained was powdered and boiled with alcohol; this gave a deep blue solution, from which nothing definite was obtained, and a white residue. The latter was purified by recrystnllising it several times from benzene. It mas thus obtained in clusters of needles melting at 278'. The benzene solution was strongly fluorescent, the crystals slightly so. Analysis gave fignres agreeing with the formula of ~~c~r~ben~oteti~ccl,hen?/klifurficrcc?z. 0.1828 gave 05897 CO, and 0.0821 H,O. C = 87-98 ; H = 4-99. 0.1430 ,, 0.4619 CO, ,, 0.0623 H,O. C=8Se09; H=4*S4. C,,H,,O, requires C = SS.31 ; H = 4.76 per cent. Only the foregoing compound could be detected in the product of the reaction, although two isomerides are possible. 6. Benzoin mad PlJoroglzccinol. 50 grams of benzoin, 10 grams of phloroglucinol, and 80 grams of 73 per cent. sulphuricacid mere taken, and the process was coaciucted as in the previous experiments. The heating was continued for 15 minutes at 150'. It is best to add the whole of the sulphuric acid at once to the mslted mixture of benzoin and phloroglucinol ; in one experinlent in which only a, little was added at first, a vigorousINTERACTION OF BENZOIN WITE PHENY LENEDIAMINES. 1043 reaction took place, steam was given off, and the substance turned very dark. The dark brown product was boiled with water, with alcoholic sodium hydrate, again with water, then with alcohol alone, and finally with a small quantity of benzene, I n this may, an almost white residue was obtained, which mas recrystallised several times from solvent naphtha, and afterwards from benzene. It forms tufts of colourless, slender needles melting at 360'. Iluring the purification, fluorescent solutions were obtained ; but solutions of thc pure sub- stance are non-fluorescent. It is almost insoluble in alcohol, glacial acetic acid, and ethylic acetate ; but chloroform dissolves it readily. Analysis gave figures agreeing with the formula of bensohexnp?henyZ- 0.1747 gave 0.5642 CO, and 0.0751 H,O. C = 88.08 ; H = 4.75, 0.1488 ,, 0.4787 CO, ,, 0.0642 H,O. C=S7*74; 13=4.79, C,,H,,O, requires C = SS.07 ; H = 4-51) per cent. UNIVERSITY OF ABERDEEN. CHEMICAL DEPARTMEKT,