XX I V. --Studies in the Phenylsucci.ruk Acid Swies. Pcwt VI. Racerriiscction Phewmenu Observed dwing the 1nz.esfigation uf the Optically ActZ'ue Yhenyl- c c ~ d ~~~~~.7/,eia~~~-s.2ICC.ilz.ic A cicJs c m d theicia -Zlek.w ti P P S . By ITENRY ~ V R E N . IT has been recently shown (Wren and Still T. 1917 111 1019) t h a t t h e esters of T- and meao-diphenylsuccinic acids are inter-convertible under the influence of alkali. To quote a specific ex-ample ethyl r-diphenylsuccinate is almost quantitatively tram-formed into the ester of the mcso-acid when its alcoholic solution is treated with a solution of sodium ethoxide in ethyl alcohol; also when the ester is hydrolysed with an insufficientl amount of aqueous ethyl-alcoholic potassium hydroxide solution the non-hydrolysed portion is found t o have suffered partial isomerisation, with the production of ethyl mesodiphenylsuccinate.Since in solution a t any rate ethyl 7.-diphenylsuccinate may be regarded as a molecular mixture of ethyl d-dipheaylsuccinate and ethyl / diphenylsuccinate it was to be expected t h a t the esters of the optically active acids would separately undergo similar trans-formation and would therefore become racemised the possibility of a balanced action being excluded in this case since any further isomerisation of the naeso-ester which might be formed would merely lead t o the production of the r.-ester. The theoretical expectations have been fully confirmed by a series of experiments with the methyl and ethyl esters of the optically active acids. A typical example is furnished by the formation of ethyl nzeso-diphenylsuccinate as a product' of the action of ethyl-alcoholic sodium ethoxide on ethyl I-diphenylsuccinate.Similarly when ethyl d-diphenylsuccinate is hydrolysed with an insufficient amount of aqueous ethyl-alcoholic potassium hydroxide solution the non-hydrolysed portion is found t o be almost inactive towards polarised light and t o consist practically entirely of ethyl mesodiphenyl-succinate. It would thus appear t h a t the racemisation previously observed t o occur during the hydrolysis of esters of the optically active diphenylsuccinic acids (Wren and Still T. 1915 107 1454) takes place in part. a t any rate previous t o the actual hydrolysis, and is thus due t o rearrangement of the ester and notl of the acid.The greater liabilitry of esters as compared with the corresponding acids t o racemisation under t h e influence of alcoholic alkali has been pointed out by Gr",(l:micr jc//cnt. Z r i t . 1910 34 3004 ; J . p r AclD SEEIES PAET VI. 211 Chem. 1913 [ii] 87 312) who found that whilst t'he esters of tropic acid are readily racemised the acid itself is not racemised by hydroxyl ions. This difference in behaviour is ascribed to the ionisation of the acid a t the carboxyl group whilst the ester which behaves as a pseudo-acid only becomes ionised in the presence of hydroxyl ions thus : Tropic acid. .1 Esters of tropic acid. .1 It has however been pointed out by Frankland (T. 1913 103, 725) that the racemisation may be due to loss of asymmetry by fautomeric change previous to ionisation thus : The only instance of racemisation of an ester previous to lvsis which aDDears to have been noted in the literature caused hydro-is the okervation 0; lMcKenzie and Widdows (T, 1915 107 713) that I-menthyl d-phenyl-p-tolylacetate is in part catalytically racemised by alkali prior to its undergoing hydrolysis.I n a case which pre-sents some analogies Wren (T. 1909 95 1596) found that the portion of the ester which escapes conversion into amide is partly inactivated when methyl Z-mandelate is acted on by cold alcoholic ammonia. Ths observations have been extended to the optically active phenylsuccinic acids and their esters since the latter substances are much more readily soluble than the derivatives of the diphenyl-succinic acids and therefore allow the course of racemisation to he followed directly by polarimetric measurement.It is thus found that ethyl and methyl d-phenylsuccinates are rapidly and coxl?.pletely racemised by the action of a solution of the requisite sodium alkyloxide in the corresponding alcohol and also that when the latter ester is treated with potassium hydroxide dissolved in slightly aqueous ethyl alcohol in quantity insufficient for complete hydrolysis the non-hydrolysed portion is almost completely racemised under the conditions described in the experimental section of the paper. Some time ago it was found by Wren arid Williams (T. 1916, 109 579) that a partly racemised acid was obtained when methyl d-phenylsuccinate is hydrolysed by an excess of aqueous ethyl 212 WREN STUDIES IN THE YHENYLSUCCINIC alcoholic potassium hydroxide solution.Thus in one experiment, the ester was heated with about six tdrnes the theoretically neces-sary quantiby of potassium hydroxide dissolved in aqueous ethyl alcohol (water alcohol 1 vol. 2 vols.) when the recovered acid was found t o have [a]= + 1 1 0 * 2 O in acetone solution whilst the parent acid had [ a ] i f 5 -t- 175.4O. The considerable activity of this acid became somewhat remarkable when considered in conjunction with the almost complete racemisation of the non-hydrolysed ester in the experiment' just quoted and pointed t o a profound influence of the varying factors namely the amount of potassium hydroxide and the water content of the solution.The effect of variation in the latter was therefore studied and was found to exercise a very marked influence on the optical activity of the recovered acid. I n three strictly comparable experiments in which the only differ-ences consisted in the use of absolute approximately 80 per cent. and 50 per cent alcohol respectively the acid formed by hydrolysis was found to have [aII + 3*1° + 59'1° and + 100*3O respectively in ethyl-alcoholic solution in which the parent acid had [a] +147.4*; under closely similar conditions the latter acid was found to be unaffected by the action of an excess of alcoholic or aqueous-alcoholic sodium hydroxide solution. During the last few years a considerable amount of evidence has been accumulated which indicates that racemisation under the influence of alkali may frequently be attributed to keto-enolic desmotropy with consequent disappearance of the asymmetry of the compound.I n the present instance the inactivation of methyl d-phenylsuccinate when acted on by sodium methoxide may be due to the forniation of the enolic form thus: S H,* C0,M.e -+ StI,*CO,;\le -+ YH,*C'O,Me c KPh*CO,Me CPh f- CH Ph- C0,Me a<;% Active. Inactive. Inactive. The evidence which has been brought forward in favour of such an explanation of racemisation in analogous case is so far of a purely qualitative nature; in no single instance has the enolic form been isolated or its amount determined. The actual separa-tion of the enolic modification would doubtless be a matter of great difficulty in many cases since the proportion of enolised molecules need be but infinitesimal to accoiint for the phenomena of racemisation.On the other hand it was hoped to be able to bring quantitative evidence of the presence of the enolic modifica-tion by utilising the method adopted by K. H. Meyer (Ber. 1912, $5 2864) in his investigations on ethyl malonate and ethy ACID SERIES. PART VI. 213 methanetricarboxylate. Direct titration of a methyl-alcoholic solution of methyl d- or r-phenylsuccinate with bromine failed t o disclose the presence of any enol and a similar result was obtained when experiments were performed on the esters dissolved in methyl-alcoholic sodium methoxide solution although under the latter conditions ethyl malonate was found to be largely enolised.The experimental conditions were not greatly varied in this case how-ever since the objection could possibly be raised that in order to demonstrate the point a t issue it would be necessary to show definitely thatl enolisation occurs in such a manner as to form compound I and not 11: It is hoped to apply the method however t o esters of certain monobasic acids t o which this objection cannot be urged. Lastly if racernisation be attributable to the formation of an enolic derivative i t would be expected that the nature of the basic atom or group of the alkali would have an influence on the result. Experiments on the hydrolysis of methyl d-phenylsuccinate with aqueous ethyl-alcoholic tetramethylammonium hydroxide solution were therefore performed; racemisation was found t o bel even more pronounced t h m with potassium hydroxide solution under nearly similar conditions.E x P E R I M E N T A L. Forniution of Ethyl mesoDi~henylsucciriute from Ethyl d- and 1-Di~henylsuccinates. A. By t h e Action of Ethyl-alcoholic Sodimn Ethoxide Solution. -Ethyl 1-diphenylsuccinate (1 gram) was warmed with a solution of sodium ethoxide in ethyl alcohol (0-666N; 20 c.c.) in a closed vessel to about 50° when it dissolved completely; the solution was allowed t o cool t o the ordinary temperature when it became almost solid. After remaining during three days a t the temperature of the laboratory the precipitate was removed and washed successively with alcohol warm water and finally alcohol. It melted a t 139-1403 and was optically inactive in acetone solution.After being crystallised from rectified spirit ethyl mesodiphenylsuccinate was obtained in well-defined needles melting a t 140-141O. R . B y Partial Hydrolysis.-Ethyl d-diphenylsuccinate (1. grams) was heated during four hours with aqueous ethyl-alcoholic potassium hydroxide solution (0.25N ; 15 c.c.) this being about three-fifths of the quantity required for complete hydrolysis. The solution was diluted with water and the unattacked esters were extracted with chloroform. The residue obtained after removal of the latter melted at’ 138-140O; it had [a]= + 5 . 4 O in acetone solution (1 = 2 c = 1.113 a + 0 . 1 2 O ) whereas the pure &ester has [a] + 279.4O under similar conditions. After being crystallised from rectified spirit it yielded pure ethyl nzesodiphenylsuccinate, which was optically inactive in acetone solution and melted a t 1 4 I)- 1 4 1 O.A. By the Action of u Nethyl-alcoholic Solution of Sodium Methoxide.-Methyl I-diphenylsuccinate (1 gram) was heated in a stoppered flask during four hours a t 55-65O with a solutioii of sodium methoxide in absolute methyl alcohol (0*704N ; 60 c.c.). When the solution was cooled a quantity of fine prismatic crystals separated which were removed washed with water and methyl alcohol and dried. After being crystallised from acetone the pro-duct was obtained in well-defined needles which were optically inactive when dissolved in chloroform (1=2 c =0*5115) and were identified as methyl m esodiphenylsuccinate by their crystalline form melting point (218.5-219.5O) alone and when mixed with a n approximately equal quantity of the synthetic meso-ester.B. By Partial Hydrolysis.-A solution of methyl I-diphenyl-siwcinate (1.5 grams) in boiling methyl alcohol (100 c.c.) was heated during four hours with aqueous methyl-alcoholic potassium hydr-oxide solution (0-593T; 8 c.c.) this being approximately half the quantity o€ alkali required for the complete hydrolysis of this amount of the ester. The alcohol was removed on the water-bath and the residue warmed with water. The undissolved portion was removed and dried. It was optically inactive when dissolved i n chloroform (I = 2 c = 0*8365) and after being crystallised from acetone yielded methyl mesodiphenylsuccinate which melted a t m*5-219.5O ; the melting point was unaltered by admixture with the synthetic meso-ester ACT14 SERIES.PART VI. 215 A B y Sodium A lky loxide .-Meth yl d-phenylsu ccin a te (0.8 15 7 gram) was dissolved in methyl-alcoholic sodium methoxide solution (20.5 c.c.) and transferred as rapidly as possible to a 2-dcm. tnbe. Approximately three minutes after being prepared the solution had a +to but the field was somewhat cloudy and the activity altered too rapidly to allow of accurate me'asurement. After about fifteen minutes the value had decreased to - t 3 * 5 O after forty minutes to + 0 * 5 O after seventy minutes to +0*17O after one hundred minutes to +0*05O. After two hours the solution was poured into a slight excess of dilute hydrochloric acid. The ester was extracted with ether the ethereal solution washed with sodium carbonate and dried over calcium chloride.The residue obtained after removal OF the solvent solidified readily and was optically inactive in acetone solution ( I = 2 c = 1.555). After being crystal-lised from light petroleum it yielded characteristic clusters of prisms of methyl r-phenylsuccinate which melted at 57*5-58.5O. The melting point remained unchanged when the substance was mixed with the synthetic r-ester. In a similar manner ethyl d-phenylsuccinate (0.9964 gram) was dissolved in ethyl-alcoholic sodium ethoxide solution (1.069iV ; 20 c.c.). I n this case racemisation proceeded so rapidly arj t o be almost complete before polarimetric measurements could be made, and the solution had become quite inactive within ten minutes.Since the liquid nature of the ethyl r-ester rendered the identifica-tion of a small quantity of it a matter of difficulty water was added to the solution and the ester hydrolysed. The result-ing f*-phenylsuccinic ac'id melted a t 167-1 68O whereas Higson and Thorpe (T. 1906 89 1470) give 168O as the melting point of this acid. B. By Yartiul Hyddysis.-Methyl d-phenylsuccinate (3 grams) was heated to boiling during two and a-half hours with slightly aqueous methyl-alcoholic potassium hydroxide solution (0.395N ; 25 c.c.) this being rather more than one-third of the quantity theoretically necessary for complete hydrolysis. The alcohol was removed on the water-bath and the residue mixed with water and shaken with ether. The residue (about 0-3 gram) obtained after removal of the ether solidified readily and had [a]= +2.12* when dissolved in acetone (Z== 2 c = 1.178 aD + O*OSo) whereas the pure ester has [a]:,O + 152.2" under similar conditions.After being crystallised from light petroleum (b. p. 40-60°) it yielded pur 216 WREN STUDIES IN THE PHENYLSUCCINIC methyl r-phenylsuccinate which melted a t 57-58'5O ; this value remained unchanged after admixture with the synthetic ester. A. ction of Alkali o n 1-Phenylsuccinic Acid. The I-phenylsuccinic acid used in these experiments had [ a ] -148-3O in ethyl-alcoholic solution a value which agrees well with the data oE Wren and Williams (Zoc. cit.). Three comparative experiments were performed in which the acid (1 gram) was treated in a closed vessel with (a) sodium ethoxide solution (1.059N ; 50 c.c.) and absolute ethyl alcohol (20 c.c.); ( b ) sodium ethoxide solution (50 c.c.) alcohol (20 c.c.), and water (0.95 c.c.) and (c) sodium ethoxide solution (50 c.c.), alcohol (10 c.c.) and water (10 c.c.).I n each case a.certain amount ol precipitate separated. The mixtures were heated with frequent agitation during five hours at 70° then neutralised with hydrochloric acid and evaporated t o remove alcohol ; the acids were isolated by extradion of the acidified solutions with ether. The dried acids were polarimetrically examined in ethyl-alcoholic solution the valu'es for the specific rotations being (a) -146.6O aD -4*92O) and (c) -146'2O (Z=2 c=1*7035 a -4.98O). In no case theref ore was any evidence of racemisation obtained.(1=2 ~=2.8815 CZ, -S*45') (6) -145.5' (Z=2 ~ = 1 * 6 9 1 , Complete Hydmlysis of Methyl d-Pherzylsuccinate b y Alcoholic Sodium Hydroxide containiiig differing Proportions of Water. I n these experiments the weight of ester and alkali the total volume of the solution the temperature and duration were main-tained uniform the only variation consisting in the replacement of differing volumes of alcohol by water. Stock solutions of ester were prepared by dissolving 2 grams of the latter in 30 C.C. of ethyl alcohol and of alkali by dissolving sodium in absolute ethyl alcohol; the latter solution was l * l N . Ten C.C. of the ester solu-tion were mixed with (a) alkali solution (25 c.c.) alcohol (25 c.c.), and water (0.45 c.c.); ( b ) alkali solution (25 c.c.) alcohol (15 c.c.), water (10 c.c.) and (c) alkali solution (25 c.c.) water (25 c.c.).The solutions were heated in closed flasks during four hours a t 60-65O; precipitates speedily separated in ( a ) and ( b ) but (c) remained homogeneous throughout. The resulting mixtures were nearly neutralised with hydrochloric acid and evaporated to remove alcohol; the aqueous solutions were extracted with ether after acidification with mineral acid. The residual phenylsuccinic acids were polarimetrically examined in ethyl-alcoholic solution whe ACID SERIES. PART VI. 217 the following values were observed for the specific rotation: (G) +3.1' (2=2 ~ ~ 2 . 2 3 9 1 +0*14'); ( b ) +59*1° (Z=2, c = 1.913 a + 2-26') ; (c) + 100.3O ( I = 2 C = 2.073'1 a + 4.16').Complete Hydrolysis of Methyl d-Phenylsuccinate b y A queozis-A 1 coho lic T e tram e thy lam m o nium Hydroxide Solu t ion. The solution of the alkali was prepared by warming an aqueous solution of tetramethylainmonium iodide with a slight excess of silver oxide and removal of silver iodide and unchanged oxide. The filtrate was concentrated t o 14 c.c. and then diluted with ethyl alcohol t o 55 C.C. An approximately iY-solution was thus obtained. The methyl d-ester (1 gram) was heated during two and a-half hours with the solution described above and the corresponding acid isolated in the usual manner; it' melted a t 164-168*5° and had [a] + 10*lo in ethyl-alcoholic solution (I = 2 c = 3-329, a +0.67O). A ctiolt of Ferric Chloride ota Methyl cl-Phertylszcccinate.It has been shown by Meyer (Ber. 1911 44 2725) in the case of ethyl acetoacetate that! ferric chloride exerts a direct enolising action. The behaviour of an ethyl-alcoholic solution of methyl dphenylsuccinate towards anhydrous ferric chloride has therefore been polarimetrically investigated in the expectation that enolisa-tion if induced a t the asymmetric carbon atom would betray itself by racemisation. 'The solutions however were found to be optically stable under these conditions. Methyl d-phenylsuccinate (0.4869 gram) was dissolved in ethyl alcohol and the solution made up t o 20 c.c.; a portion of this solix-tion had a + 6-73' when examined in a 2-dcm. tube and this value had not changed a t the end of forty hours after the addition of a small quantity of ferric chloride.A further portion pf the latter substance was added and the solution allowed t o remain a t the temperature of the laboratory during nine days a t the end of which period the ester was isolated and examined in ethyl-alcoholic solution; it had [aID + 129'8' whereas the value + 138.2O had been determined for the original specimen. Possih7e Enolisntion of Methyl Phenylstccciiicctc in So7utiotr. Methyl 6-phenylsuccinate (0.5369 gram) and methyl T-phenyl-snccinate (0.4496 gram) were separately dissolved in methyl alcoho (20 c.c.) and titrated with an lV/lO-solution of bromine in the same solvent until a faint permanent yellow coloration was pro-duced; 0.55 C.C. of bromine was required in each case whilst in a blank experiment 0.60 C.C. was necessary. Methyl d-phenylsuccinate (0.3292 gram) was dissolved in a well-cooled methyl-alcoholic solution of sodium methoxide and the product’ poured into an excess of a solution of bromine in methyl alcohol containing hydrogen chloride. Excess of bromine was removed by the addition of &naphthol dissolved in methyl alcohol, and the resulting solution warmed after addition of aqueous potassium iodide (10 per cent .). The liberated iodine required 0.3 C.C. of AT/ 10-sodium thiosulphate solution this quantity being the same as that required in a blank experiment. The author desires t o express his thanks t o the Research Fund Committee of the Chemical Society for a grant which has defrayed a part of the cost of the investigation. ?vXUNICIPAL TECHNICAL k T l ” T U T E , BELFAST. [Received February 20i% 1918