2564 MCKENZIE AND CLOUGH: EXPERIMENTS ON THECCLXII. -Experiments on the Walden Irwersion. PalatVI. Conversion of the Optically Active a-Hydroxy-a-phenylpopionic Acids into a- Chloro-a-phenyl-propionic Acids.By ALEX. MCKENZIE and GEORGE WILLIAM CLOUGH.THE authors showed in a former paper (Trans., 1910, 97, 1016) thatwhen I-a-hydroxy-a-phenylpro-pionic acid was acted on by thionylchloride at the ordinary temperature there was no evidence thatthe carboxyl group had been attacked, whereas the hydroxy-groupin direct attachment to the asymmetric carbon atom was displacedby the chlorine atom with great readiness.* By this method I- andd-a-chloro-a-phenylpropionic acids were easily obtained in a stateof opt,ical purity, and the interconversion of the optically activea-hydroxy-a-phenylpropionic acids was brought about by aid of theWalden inversion, thus :(by SOC12) Ph c1p h ~ > C < O ~ Me’ C02H + h*/C<CO,H3% ;:i?:,t J%$%r)d- AX 2 c - G &H t- (by SOCI2) d-E>C<E213PhThe results obtained by a further study of this subject are placedon record in the present communication.When the action of thionyl chloride on I-a-hydroxy-a-phenyl-propionic acid is conducted in such a manner that both hydroxy-groups are displaced by chlorine, the resulting chloro-acid chlorideis Zaevorotatory, and on decomposition with acetone containing alittle water it gives I-a-chloro-a-phenylpropionic acid.The inter-* Later, Stoll6 (Ber., 1910, 43, 2471) observed that benzilic acid is convertedinto diphenylchloroacetic acid in a similar matlner when its solution in carbontetrachloride is acted on by thionyl chloride a t the ordinary temperatureWALDEN INVERSION. PAKT VJ.2565conversion of the active hydrosy-acids can be carried out accordingto the scheme: z-g>c<g:2H (by SOC12) + I-Ph>"&, Me$(by iteetonrand I\ ater)+--- Ph\C<C1 d-g>c<EE~ (by AgN03 and water) z - ~ e / cO,H.On the other hand, when I-a-hydroxy-a-phenylpropionic acid isacted on by phosphorus peatachloride, a change of sign of rotationoccurs ; the chloro-acid chloride is in this case dexti-orotatol-y, andit gives d-a-chloro-a-phenylpropionic acid when it is decomposed byaqueous acetone. When the chlorine in the latter product isdisplaced by the hydroxy-group by aqueous silver nitrate, the parenta-hydroxy-a-phenylpropionic acid is regenerated. The followingcycle can thus be effected :P hI-l\le?C<&TThionyl chloride andI ( b y ncetoircphosphorus pentachloride thus behavediff ereniy towards 2-a-hydroxy-a-phenylpropionic acid, inasmuchas the former gives the I-chloro-acid, whereas the latter gives thed-chloro-acid.The action of phosphorus pentachloride is accom-panied with a large amount of racemisation.I n the course of his work on the Walden inversion, E. Fischer(Ber., 1907, 40, 489) found t'hat Ebromopropionic acid is formedfrom d-alanine by the action of nitrosyl bromide, this action beingprobably abnormal; when the ester of d-alanine, however, is actedon by nitrosyl bromide, the resulting compound is t,he ester ofd-bromopropionic acid, thus :cl-alanine -+ I-bromopropionic acidd-alnnine ester -+ J-bromopropionic ester.The conclusion is drawn that the same reagent (iiitrosyl bromide)can act on closely-related substances (for example, acid and ester),in one case normally and in the other case abnormally.This con-clusion was borne out when the behaviour of nitrosyl chloridetowards active leucine (and its ester) and active aspartic acid (andits ester) was examined. Again, silver oxide, which apparentlybehaves abnormally in Walden's experiments with the malic acids,can be used to produce either d- or I-lactic acid from Z-bromo-propionic acid, thus :Lbromopropionic acid -+ d-lactic acid.I-bromopropionylglycine -+ I-lactic acid.VOL. XCVII.8 2566 McKENZIE AND CLOUQH: EXPERIMENTS ON THEBearing these results of Fischer in mind, we decided to investigatethe action of phosphorus pentachloride and thionyl chloriderespectively on the ethyl ester of active a-hydroxy-a-phenylpropionicacid in the hope of throwing some light on the problem as t owhich of these agents acts abnormally on the hydroxy-acid. Thefollowing change, which is accompanied by a, large amount ofracemisation, takes place when ethyl I-a-hydrosy-a-plienylpropionateinteracts with phosphorus pentachloride :(IJ)’ PCI5) d- P h >c<& Et. I- 3 c < g Et ---+ Me“g>C<gp > d-M,>C<:b,Eb.{Unfortunately, the action of thionyl chloride on ethyl d-a-hydroxy-a-phenylpropionate does not proceed smoothly, but we have evidencethat the following change takes place :(by EOCI‘J YhWhen the changes-chloro-acid (by PC15)(by PCIS) d-cbloro-esterI-hydroxy-acid -+-+ I - hydroxy-esterE-hydroxy-acid 3 I-chloro-acid { I-hydroxy-ester (by SOC1-J I-chloro-esterare contrasted, it will be seen that no conclusion can be drawnfrom them as to which of the agents, thionyl chloride or phosphoruspentachloride, causes the Walden inversion when it acts on thehydroxy-acid.It was obviously desirable to prove that the I-chloro-acid corre-sponds in configuration withh the I-chloro-ester.This was done byshowing that the I-chloro-acid chloride (which gives the E-chloro-acid on treatment with aqueous acetone) is converted into theI-chloro-ester on addition to ethyl alcohol, the sign of rotationremaining unchanged.It is accordingly possible to convertI-a-hydroxy-a-phenylpropionic acid into either ethyl I- or ethyld-a-c hlorea-p hen ylpropionat e, thus :(by SOCI.,)+soclz) $>C<C’ COCl ~ (by EtOH) + 7- Mle>C<%,Et P’1EXPERIMENTAL.A ction. of ThLionyZ Chloride on 1-a-Hydroxy-a-phenylpropionic A cid.A mixture of I-a-hydroxy-a-phenylpropionic acid (6 grams) ,prepared by the resolution of the r-acid with morphine (loc. cit.)WALDEN INVERSION. PART VI. 2567and thionyl chloride (20 grams) was kept at the ordinary temperature for two days, and then heated at 80° for one hour. Five gramsof a Zaevorotatory acid chloride boiling at 10&-llOo/10 mm. wereobtained.It was added to a mixture of acetone (20 c.c.) and water(1 c.c.), and after a day the solvent was removed at the ordinarytemperature under diminished pressure. The crude chloro-acid(3 grams) was crystallised once from light petroleum, when the pureZ-acid (Zoc. cit.) was obtained, melting a t 71-7Z0, and having thefollowing specific rotation in benzene solution :This acid gave a dextrorotatory mixture of a-hydroxy-a-phenyl-propionic acids when the displacement of chlorine was effected byaid of aqueous silver nitrate.It is thus possible t o prepare optically pure I-a-chlorea-phenyl-propionic acid even when the temperature of the action with thionylchloride is raised so that t'he carboxyl group of the hydroxy-acid isdisplaced.The laevorotation of the chloro-chloride, prepared from G gramsof the I-hydroxy-acid and 28 grams of thionyl chloride under tem-perature conditions somewhat different from the preceding, wasconfirmed.One gram of the chloro-chloride was added to 7 C.C. ofacetone, and the solution gave a, -4.5O in a 1-dcm. tube.I = 2, c = 7.188, a, - 3*78O, [a] , - 26.3O.Action of Phosphorus Pentachloride on 1- and d-a-Hydroxy-a-phenylpropionic Acids.A mixture of I-a-hydroxy-a-phenylpropionic acid (12 grams) andt'he calculated amount of phosphorus pentachloride (30 grams) wasmaintained at 15O for thirty minutes, by which time the vigorousevolution of hydrogerr chloride had subsided. On heating to 70°, abrisk action again took place ; the temperature was raised from 70°to looo within thirty minutes, and finally maintained at looo forone hour.The action of the phosphorus pentachloride apparentlytook place in two distinct stages. The phosphoryl chloride wasremoved, and the residual oil fractionated under diminishedpressure. Six grams of a colourless oil, boiling at 113-115°/15 mm.,were obtained. This chloro-acid chloride was dextrorotatory, givinga,+3.1O0 in a 1-dcm. tube, and its solution in acetone was alsodextrorotatory. It was added to a mixture of acetone (20 c.c.) andwater (1 c.c.). Next day, the solvent was removed at the ordinarytemperature under diminished pressure, and the residual soIiddrained on porous earthenware. It amounted to 4 grams, meltedat 70--75O, and gave the following value in benzene solut'ion :I = 2, c = 12.7, a, + 0*56O, [a] + 2 * 2 O .Found, C1= 19.2.Calc. C1= 19.3 per cent.8 3 2568 McKENZlE AND CIAUUH: EXPERIMENTS Oh' THEThe resulting acid was thus a mixture of d- and r-a-chloro-a-phenylpropionic acids, and a change of sign of rotation accordinglytook place during the displacement of the chlorine by the hydroxy-group under the following conditions. The acid was shaken for fourhours a t the ordinary temperature with a solution of 5 grams ofsilver nitrate in 20 C.C. of water, and the mixture was then boiledfor one minute and filtered. The hydroxy-acid was extracted withether. It melted at 88-90°, was free from chlorine, and was lzevo-rotatory. I n ethyl-alcoholic solution :Z =4, c = 9-58, a, - 0*67', [a] - 1.7'.The action of phosphorus pentachloride (25 grams) ond-a-hydroxy-a-phenylpropionic acid (10 grams) was also studiedunder conditions similar to the preceding.The chlore-acid chloridewas in this case lzevorotatory. On decomposition with aqueousacetone it gave a mixture of Z- and r-a-chloro-a-phenylpropionicacids.Conuersion of 1-a-Chloro-a-phenylpropio?zyZ Chloride into Eth$1-a-ChZoro-a-ph enylpro pionat e.The lzvorotatory a-chlorcla-phenylpropionyl chloride, preparedfrom I-a-hydroxy-a-phenylpropionic acid (10 grams) and thionylchloride (28 grams), was added to ethyl alcohol and the productfractionated. The ester was a Zmuorotatory oil; it amounted to4 grams, boiled at 131-132°/18 mm., and gave the followingvalues :I = 1, a1,7 -- 6.30°, D:' 1.124, [a]: -5.6'.Found, C1= 16.8.We have no evidence that this ester was optically pure, and itis necessary to bear in mind that partial racemisation occurs veryfrequently when the alcoholic hydroxy-group of an optically activehydroxy-carboxylic acid is displaced by halogen.It appeared from the following experiment.that this ester corre-sponds in configuration with I-a-chlorcm-phenylpropionic acid. Theester (3 grams) was added to 25 C.C. of 1V-aqueous silver nitrate, andmaintained at the ordinary temperature for twenty-four hours. Thesilver chloride was removed, and the filtrate boiled with aqueouspotassium hydroxide until all the oil had disappeared. The result-ing a-hydroxy-a-phenylpropionic acid obtained by acidification andextraction with ether amounted to 2 grams, melted a t 84-89',and was dextrorotatory in ethyl-alcoholic solution :CllH,,O,C1 requires C1= 16.7 per cent.!= 2, c = 18.48, a, + 1'33O, [ a ] + 3'6OWALDEN INVERSION.PART VI. 2569Action of PhospJ~oi~us Pentachloride and of Thionyl Chloride on theOptically A ctive Ethyl a-Hydroxy-a-phenylpropionates.Ethyl l-a-hydroxy-a-phtnyZpro@onate, OH-CMePh-CO,Et, pre-pared by the esterification of I-a-hydroxy-a-phenylpropionic acid byethyl alcohol and sulphuric acid, is an oil boiling at 127O/12 mm. :0.1330 gave 0.3322 CO, and 0.0865 H,O.It has Did 1.097 and C Z ~ - 29.24O ( I = 1) ; whenceWhen this ester is acted on by phosphorus pentachloride, theresulting ethyl a-chloro-a-phenylpropionate is dextrorotatory, butit is impossible to prevent a large amount of racemisation takingplace in this change.Powdered phosphorus pentachloride(7 grams) was added gradually within an interval of one hour t o theI-hydroxy-ester (6 grams). After three hours a t the ordinary tem-perature the product was warmed at 30° for a few minutes, whenno further action was perceptible. It was then treated with etherand ice, and the ethyl a-chloro-a-phenylpropionate isolated as anoil, boiling at 138-139°/28 mm.:C = 68.1 ; H =7'3.- 26.7'.C,,H,,O, requires C = 68-0 ; H = 7.3 per cent.Found, C1= 16.5.The ester had a , + l * l O o in a 1-dcm tube.On decomposing it with a slight excess of 1.03iV-aqueous sodiumhydroxide, the displacement of the chlorine by the hydroxy-groupof course accompanied the saponification; the product of the actionwas r-a- h y dr oxy-a-p hen ylpr opionic acid.Although attempts to prepare ethyl dZ-a-chloro-a-phenylpropionatein a state of even approximate purity from ethyl dl-a-hydroxy-a-phenylpropionate (b.p. 124-125O/ 10 mm.) and thionyl chloridemet with no success, it was nevertheless decided to examine theaction of thionyl chloride on the optically active hydroxy-esters.The material used for this purpose was the dextrorotatory mixtureof d- and r-a-hydroxy-a-phenylpropionic acids obtained from themother liquors resulting from the morphine resolution of the r-acid.It was esterified by ethyl alcohol and sulphuric acid, and the esterobtained had a,+ 14O in a 1-dcm. tube. Twelve grams of thisproduct remained for seven days at the ordinary temperature afterhaving been mixed with 14 grams of thionyl chloride. The liquidwas then heated for twenty hours at 50-60°, and for five hourslonger a t 60-70°, but the action was apparently not quite completeeven under these conditions. The product was fractionated, and anoil boiling at 138-141°/25 mrn. was obtained; it was dextrorotatory,giving a,, + 1*50° in a 1-dcm. tube. It contained, however, only 11-7per cent. of chlorine. We were able to draw the conclusion,C,,H,,O,Cl requires C1= 16.7 per cent2570 FORSTER AND NEWMAN: THE TRIAZO-GROUP. PART xv.however, from the following experiment Chat d-a-chlorea-phenyl-propionate was actually present in this impure product. Thechlorine in it was first displaced by aqueous silver nitrate, and theproduct then saponified with aqueous potassium hydroxide. Theresulting potassium salt was Zaevorotutory in aqueous solution, andthe a-hydroxy-a-phenylpropionic acid obtained from it was alsoZaevorotatory. This result was confirmed by a second experiment.BIRKBECK COLLEGE,LONDON, E.C