504 BARNETT COOK AND NIXON THE SYNTHESIS OF mes0-m~~ LXXV.-The Synthesis of meso- AZkyZ and meso-AryZ Anthracene Derivatives. Part I. By EDWARD DE BARRYBARNETT,JAMESWILFREDCOOK and IVORGRAY NIXON. THE interest which attaches to the three-carbon transannular tautomerism exhibited by certain meso-alkyl anthracenes (Barnett Cook Rlatthews Ber. 1926 59 1429 2863) has rendered it desirable to review the methods available for the preparation of these compounds. Such methods fall roughly into two classes according as the meso-substituent is attached prior or subsequent to the closing of the central ring of the anthracene complex. The present communication deals with the former class of reaction. (A).By treating methyl trjphenylmethane-o-carboxylatewith magnesium phenyl bromide Haller and Guyot (Bull.SOC.chim. 1904,31,979) obtained a product which when treated with hydrogen chloride in glacial acetic acid passed into 9 :9 :10-triphenyl-9 :10-dihydroanthracene (11). They described their product as a methyl ether (I),but offered no evidence of the presence of the methoxyl group other than ultimate ana,lysis which in this case is totally insufficient to differentiate between the free hydroxy-compound and its methyl ether (Found C 89.7; H 6.2. C,,H,,O requires C 90.1; H 6.1. C,H,,O requires C 90.0; H 6.4%) and it has now been found that the compound is as would be expected the free hydroxy-compound since identical products are obtained from the methyl and the ethyl ester. With diphenylmethane-o-carboxylicacid (0-benzylbenzoic acid) which is easily accessible a synthesis similar to the above might be expected to take place.To a limited extent this is true as the esters of this acid on treatment with magnesium phenyl bromide yield an oily product (undoubtedly 111) which is easily and almost quantitatively dehydrated to 9 :9-diphenyl-9:10-dihydroan-thracene (IV). Attempts to extend this synthesis by replacing the AND 'WLCSO-ARYL ANTHRACENE DERIVATIVES. PART I. 505 magnesium phenyl bromide by magnesium ethyl isopropyl or naphthyl bromide or by magnesium benzyl chloride were not successful as no crystalline substance could be obtained from the viscid products of the reaction. The hydroxy-compound (111)is not the sole product of the action of magnesium phenyl bromide on the esters of o-benzylbenzoic acid a second product being formed at the same time to which formula (V) is given this formula being based on ana,lysis and molecular- weight determinations the oxidation of the substance to o-dibenzoyl- benzene and on the fact that its formation is independent of the alcoholic radical (methyl ethyl or phenyl) in the ester.This pinacol obviously originates by the reduction of the ketone (0-benzylbenzophenone) which must be the first product of the action of magnesium phenyl bromide on the esters of o-benzyl-benzoic acid and although the reducing action of Grignard solu- tions is well known this appears to be the first case recorded in which they have been found to cause pinacol condensation.(B).By treating diphenylphthalide with magnesium phenyl bromide Guyot and Catel (BUZZ.Xoc. chim. 1906 35 551 562) obtained a compound which they described as a hydroxyphthalan (VI R = Ph) tetraphenylphthalan (VIII) being obtained by employing a large excess of the Grignard solution (Schlenk and Brauns Ber. 1915 48,727). (VI.) (VII.) (VIII.) Guyot and Catel's compound cannot be dehydrated to give diphenylanthrone by treatment either with hydrogen chloride in glacial acetic acid or with concentrated sulphuric acid but it does give diphenylanthrone on heating to 300". Similar compounds (R = Me or Et) are obtained by the action of magnesium phenyl bromide on dimethyl- and diethyl-phthalide ; these however on heating do not pass into the corresponding dialkylanthrones but lose water in a different direction to give unsaturated substances which are probably represented by for-muh (IX) and (X) although the possibility that they may be indene derivatives (e.g.XI) is not excluded. CMe:CH <CEt:CHMe C6H4<cMe(oH)>CH C~H~<COP~~ 6 4 COPh CPh (XI.) S 506 BARNETT COOK AND NIXON THE SYNTHESIS OF ‘W1eSO-ALKYL NOproof of the hydroxyphthalan structure (VI) is at present available and the hydroxy-ketonic structure (VII) is suggested by Beilstein (“ Handbuch,” 4th Ed. Vol. VIII p. 223). It is possible that the two formulae may represent phases of an isodynamic change and it is proposed to investigate this point at a later date and at the same time to examine the dehydration products more closely.In the meantime the hydroxyphthalan structure is retained in the following although the dehydration which takes place on heating is better explained by formula (VII). (C). Arylhydroxyphthalans (VI or VII) have been obtained from the esters of phthalic acid by Guyot and Catel (bc. cit.) and by Weiss and Heidrich (Monatsh. 1924 45,210). Shibata (J.? 1909 95,1449) treated diethyl phthalate with magnesium phenyl bromide and after purifying his product by distillation under reduced pressure (b. p. 280-295’/8 mm.) isolated a colourless crystalline substance which he described as diphenylphenylene-phthalan (XII) and he based certain conclusions concerning the structure of the benzene ring on its formation. The formula given by Shibata appeared in the highest degree improbable and repetition of his experiments has proved that this supposed phenylene com- pound is really diphenylanthrone (XIII) the formation of which is undoubtedly due to dehydration of the hydroxyphthalan (VI or VII) first formed.Shibata’s conclusions as regards the structure of the benzene ring are therefore valueless. The action of magnesium phenyl bromide on o-dibenzoylbenzene also leads to the hydroxyphthalan (VI or VII) and from this to diphenylanthrone. (D).Liebermann and Lindebaum (Ber. 1905 38 1804) by heating 9-bromo-10:10-diphenyl-9:10-dihydroanthracene (XIV) alone or with naphthalene obtained a product which they described CPh as tetraphenylheptacyclene and to which they gave formula (XV) without however bringing forward any experimental evidence in AND -0-ARYL ANTHRACENE DERIVATIVES.PART I. 507 its favour. The formation of a compound of this structure by the above reaction is not very probable and an exactly similar insoluble and highly stable substance is formed when diphenylanthrone is submitted to the pinacol reduction with zinc and hydrochloric acid in acetic acid solution and also when 9 9-diphenyl-9 10-dihydro-anthracene is heated to 250" in oxygen or is heated with benzal chloride or benzophenone chloride. Owing to the extreme insolu- bility of these substances and to the fact that they do not melt it is impossible to state with certainty that they are identical but it seems extremely probable fhat they are identical with the so-called tetraphenylheptacyclene and have formula (XVI).The imposei- bility of purifying them by recrystallisation renders it difEicult to be certain that the molecule does not contain two hydrogen atoms less than the number required by the formula given (C,,H, requires C 94.3; H 5-7. C52H36requires C 94.6; H 5.45%). (E). Padova (Ann. Chim.,1910 19,400)found that dihydro- anthracene condensed with benzophenone chloride to give bisdi- phenylmethylenedihydroanthracene (XVII),the compound obtained by Staudinger (Ber.,1908 41 1362) by the action of diphenylketen on anthraquinone. 9 9-Diphenyldihydroanthracenebehaves in the same way with benzophenone chloride and gives 9 :9-diphenyl-10 :10-diphenylmethylene-9 10-dihydroanthracene (XVIII) in which it has not been found possible to reduce the double bond.(XVII.) C H <C(?Ph2)>C H C H <-c!h2->C6H4 (XVIII.) * C(.CPh,) 6 4 6 4 C(.CPh,) Padova (loc. cit.) claimed t,o have reduced his compound by means of sodium and boiling benzyl alcohol but we have not been able to repeat this. EXPERIMENTAL. oo-Diphenyl-o-tolyldiphenylcurbinol.-The methyl and ethyl esters of triphenylmethane-o-carboxylic acid were prepared in the usual way by esterification with the alcohol and hydrochloric acid. The former (m. p. 98") has been prepared by Haller and Guyot (loc.cit.) by a different method. The latter after being twice recrystallised from alcohol formed a colourless crystalline powder m. p. 69" (Found C 83-2; H 6.6.C2,H2,0 requires C 83.5; H 6.3%). In each case an ethereal solution of the ester was added to an ethereal solution of magnesium phenyl bromide (2 mols.) with cooling in a freezing mixture and the whole was boiled for 4 hours and then poured on to a mixture of ice and solid ammonium chloride. The solid which separated and that which was obtained by distilling off the ether from the ethereal solution were identical and in each case after recrystallisation from toluene-light petroleum melted 508 BARNETT COOK AND NMON THE SYNTHESIS OF mo-ALKYL at 218" and a mixture of the specimens made from each ester showed no depression of melting point. o-Benzylbenzoic Acid.-The statement by Ullmann (Annalen 1896,291,23)that o-benzoylbenzoic acid cannot be reduced beyond the phthalide stage by means of ammonia and zinc dust is incorrect the following method having been found to be the most convenient one for the preparation of o-benzylbenzoic acid.Technical o-benzoylbenzoic acid (150 g.) was dissolved in 1 1. of concentrated ammonia (d 0-880)diluted with 300 C.C. of water and a few C.C. of copper sulphate solution were added followed by 300 g. of zinc dust in several portions with cooling. The whole was heated on the water-bath until a filtered sample when acidified gave a precipitate which was completely soluble in sodium carbonate (about 2 days). The filtrate and washings were then acidified with hydrochloric acid and the precipitate was dissolved in sodium carbonate solution reprecipitated with hydrochloric acid and finally recrystallised from aqueous acetic acid.The yield was 117 g. of snow-white material m. p. 118". Gresley (Annulen 1886 234 234) states that o-benzylbenzoic acid on treatment with sulphuric acid does not give anthrone but is oxidised to anthraqujnone. This is erroneous for when a solution of benzylbenzoic acid in concentrated sulphuric acid is kept for 2 hours at the ordinary temperature an almost quantitative yield of anthrone is obtained and indeed this is a convenient method of preparing anthrone. The methyl and ethyl esters were prepared in the usual way by means of the alcohol and hydrochloric acid and formed oils b. p. 320" and 325" respectively. The phenyl ester was obtained from the acid phenol and thionyl chloride in pyridine solution (Barnett and Nixon Chem.News 1924 129 190) and was used in the crude state. 9 :9-Diphenyl-9:10-dihydroanthracene (IV).-To an ethereal solution of magnesium phenyl bromide prepared from 8 g. of magnesium and 48 g. of bromobenzene and cooled in a freezing mixture an ethereal solution of 32 g. of ethyl o-benzylbenzoate was added slowly. After boiling for an hour or being kept at the ordinary temperature over-night the whole was poured into ice and dilute hydrochloric acid and the solid (15 g.; see below) collected. After washing and drying the ether was removed from the ethereal solution. The viscid residue which refused to crystal- lise was heated for an hour on the water-bath with a glacial acetic acid solution of hydrogen chloride.The resulting diphenyl- dihydroanthracene (18 g.) m. p. 200° was quantitatively oxidised to diphenylanthrone by chromic acid in glacial acetic acid. AND meso-mn ANTHRACENE DERJVATIVES. PART I. 509 Similar results were obtained when the methyl and the phenyl ester were used in place of the ethyl ester. s-oo'-Dibenzylbenz~inawl(V).-The solid by-product insoluble in ether obtained as described above after recrystallisation from benzene-light petroleum chloroform-light petroleum and methyl ethyl ketone formed a colourless crystalline powder m. p. 175". Its hot benzene solution showed a strong green fluorescence which vanished on cooling (Found C 87.8 ; H 6.5 ; M in ethylene requires C 87.9; H 6.2%; M 546). dibromide 540.C40H3402 When 1 g. of the above pinacol in 20 C.C.of boiling glacial acetic acid was treated slowly Nith a concentrated aqueous solution of 3 g. of chromic acid oxidation took place rapidly and on cooling and dilution a solid was obtained which after recrystallisation was identified as o-dibenzoylbenzene by direct comparison with an authentic sample. Attempts to convert the pinacol into a dianthranyl derivative an acetate and a benzoate gave oily products only. The pinacol was not regenerated from these oils on boiling with alcoholic alkali but they were converted into o-dibenzoylbenzene by oxidation with chromic acid. It is therefore likely that loss of water with formation of an a-pinacolin had taken place but owing to the uninviting nature of the products the matter was not further investigated.DimethyZphenylhydroxyphdhhn (VI or VII ; R = Me).-An ethereal solution of 8 g. of dimethylphthalide was added with water-cooling to an ethereal solution of magnesium phenyl bromide prepared from 2 g. of magnesium and 12 g. of bromobenzene. After boiling for 2 hours the whole was poured on to ice and hydrochloric acid and the ethereal layer was washed with sodium hydroxide solution and with water dried over sodium sulphate concentrated and diluted with light petroleum. The m. p. 118" of the resulting crystals (9.7 g.) was not altered by further recrystallisation from benzene-light petroleum (Found C 80.0; H 6.8. c$& requires C 80.0; H 6.7%). o-isoPropenyZbenxophenone(IX).-On distillation at atmospheric pressure the above phthalan (10 g.) gave water and a ,product b.p. 305" which was dried in ethereal solution and distilled under reduced pressure; the yellow oil b. p. 208-209"/45 mm. obtained became solid on cooling and after recrystallisation from aqueous alcohol formed colourless needles m. p. 44" (Found C 86.2; H 6-4. C,,H,,O requires C 86.5; H 6.3%). o-isoPropenylbenzophenone at once decolorised a solution of bromine in carbon tetrachloride. After 1 g. in 15 C.C.of glacial acetic acid had been boiled for an hour with addition of a con- centrated aqueous solution of 5 g. of chromic acid a solid was s2 510 BARNETT COOK AND NIXON THE SYNTHESIS OF meso-ALKYL obtained on cooling and dilution which was reprecipitated from its solution in aqueous sodium carbonate and then identified as o-benzoylbenzoic acid by direct comparison with an authentic sample.Diethylphenylhydroxyphthalun (VI or VII; R = Et) was pre-pared from diethylphthalide and magnesium phenyl bromide in exactly the same way as the corresponding dimethyl derivative described above and formed colourless glistening crystals m. p. 94-95' (Found C 80-2; H 7.5. C,,H,,O requires C 80.5; H 76%). On heating it lost water and passed into an unsaturated oil which was not further examined. Attempts to prepare the dibenzyl analogue from dibenzyl-phthalide were not successful most of the dibenzylphthalide being recovered unchanged. Action of Magnesium Phenyl Bromide on Ethyl Phthalute.-The experiment was carried out essentially as described by Shibata (loc.cit.),and the product identified as diphenylanthrone by direct comparison with an authentic sample. Action of Magnesium Phenyl Bromide on o-Dibenzoylbenxene .-Finely powdered o-dibenzoylbenzene (14.5g.) was added with cooling to an ethereal solution of magnesium phenyl bromide prepared from 1.25 g. of magnesium and 8 g. of bromobenzene. After boiling for 6 hours the whole was poured into ice and dilute sulphuric acid unchanged o-dibenzoylbenzene (2.5 g.) removed by filtration and the ether distilled off. The residue after purification by distillation in steam and drying in ethereal solution was heated at 300" for 4 hour and after recrystallisation was identified as diphenyl-anthrone by direct comparison with an authentic sample.In both the above preparations of diphenylanthrone it is probable that triphenylhydroxyphthalan (VI or VII; R = Ph) intervenes as it was found that this substance passed into diphenylanthrone when heated at 300"for hour. Dehydration to diphenylanthrone however could not be effected either with concentrated sulphuric acid or with hydrogen chloride in acetic acid. Tetraphenyltetrahydrodianthrunyl (XVI).-(a) Diphenylanthrone (4.4 g.) and 2.5 g. of zinc dust were boiled with 50 C.C.of glacial acetic acid and 5 C.C. of fuming hydrochloric acid were added slowly. After 5 hours the solid was collected extracted with hydrochloric acid and with boiling glacial acetic acid and then recrystallised from a very large volume of boiling nitrobenzene in which however it was almost insoluble.The resulting cream-coloured powder did not melt at 320" (Found C 93-9; H 5.75. C,,H, requires C 94-3; H 5.7%). (b) Diphenyldihydroanthcacene was heated for 2 hours in a AND WW80-ARYL ARTHRACENE DERIVATIVES. PART I. 511 current of oxygen at 250" and the product repeatedly extracted with boiling xylene to remove impurities. After being washed with ether and dried it formed a colourless crystalline powder which did not melt at 340" (Found C 94.3; H 5*9y0). (c) An insoluble product similar to the above was obtained when diphenyldihydroanthracene was heated at 250' with benzal chloride. 9 9-Diphenyl-10 10-diphenylmethylene-9 10-dihydracnthracene (XVIII).-9 9-Diphenyl-9 10-dihydroanthracene (6.6 g.) and benzophenone chloride (5 g.) were heated for 2 hours at 250° aiid coloured resinous by-products removed by washing with ether.The residual solid (7.8 g.) was dissolved in boiling xylene and a small amount of an insoluble substance (probably tetraphenyl- tetrahydrodianthranyl) removed by filtration. On cooling a colourless crystalline powder separated which after further recrys- tallisation melted at 286" (Found C 94.0; H 5.8. C,SH, requires C 94-4; H 5.6%). The above compound was recovered unchanged after attempts to reduce it with zinc and hydrochloric acid in boiling glacial acetic acid with sodium in boiling amyl and benzyl alcohols and with hydriodic acid and red phosphorus at 180" for 7 hours.10-Chloro-9 9 :10-triphenyl-9 10-dihydr0anthracene.-Five g. of triphenyldihydroanthranol (m. p. 204" prepared from dipheiiyl- anthrone and magnesium phenyl bromide as described by Haller and Guyot Compt. rend. 1904,139,9 but without the use of benzene as a solvent) were dissolved in 25 C.C. of benzene and dry hydrogen chloride passed through the boiling solution for 20 minutes. On cooling and addition of ether a solid was obtained which after recrystallisation from benzene-ether formed a colourless powder m. p. 193-194". It contained solvent of crystallisation which it lost at 120" (Found (2 8.0. C,,H,,Cl requires C1 8.0y0). This chloro-compound when boiled in benzene solution with copper powder in an atmosphere of carbon dioxide gave a yellow solution which became more deeply coloured on heating.This solution probably contained a free radical but further investigation has been postponed pending the publication of the work of another investigator. 9 :9 :10-Triphenyl-9 10-dihydroanthracene (II).-The preparation of this compound by the reduction of triphenyldihydroanthranol with zinc dust and acetic acid as described in the literature gave poor results and it was found much better to reduce 15 g. of triphenyldihydroanthranolin 150 C.C. of boiling amyl alcohol by the addition of 7.5 g. of sodium. The solid which separated when the hot liquid was poured into water was washed with alcohol 512 BUCKNAIL CARTER AND WARDIAW and after recrystallisation from benzene-alcohol (yield 12 g.) melted at 230" (the m.p. given in the literature is 220"). 9 :10-Diphnyl-9:10-dihydroanthracene was obtained most con- veniently by reducing 1 g. of 9 :10-diphenylanthracene in 25 C.C. of boiling amyl alcohol by the addition of 1g. of sodium. The solution at first became yellow but the yellow colour subsequently vanished with the fluorescence. The mixture was poured into water the amyl alcohol removed with steam and the product recrystallised from benzenelight petroleum. Action of Aluminium Chloride on 9 :9-Diphenyl-9:10-dihydro-unthrucene.-Diphenyldihydroanthracene (5 g.) and aluminium chloride (10 g.) were boiled for 6 hours with 50 C.C. of carbon disulphide. After decomposition with dilute hydrochloric acid the carbon disulphide solution was washed with water the solvent removed by distillation and the product recrystallised from acetic acid and from alcohol.It then melted at 155-157" alone or when mixed with an authentic sample of 9-phenylanthracene. By the action of aluminium chloride on 9 :10-diphenyl-9 10-di-hydroanthracene a product was obtained m. p. 215-220" which appeared to consist chiefly of 9 :10-diphenylanthracene since it did not depress the melting point of this substance; repeated recrystal- lisation however failed to raise the melting point. The action of aluminium chloride on 9 9 10-triphenyl-9:10-di-hydroanthracene in the cold led only to the recovery of the unchanged material although the fluorescence of the solution and the low melt- ing point of the crude recovered material indicated that re-establish- ment of the "bridge " had taken place to a slight extent.When the experiment was carried out in boiling carbon disulphide solution only resinous products were obtained. The authors desire to express their thanks to the Research Fund Committee of the Chemical Society for a grant which has defrayed some of the expenses of this research. SIR JOHNCASS TECHNICAL INSTITUTE JEWRYSTREET, E.C.3. [Received December 29th 1926.1