EXPERIMENTS ON THE WALDEN INVERSION. PART 111. 121XIV.-Experiments on, the Walden Inverszon. Part .HI.Optically Active 6-Hydroxy-p-phenylpropionic Acidsand the Corresponding P- Bromo-B-phenylpropionicAcids.By ALEX. MCKENZIE and HERBERT BROOKE PERREN HUMPHRIES.THE effect, which in certain cases is associated with the electro-negative phenyl group, has been brought out clearly in previouswork on the Walden inversion (McKenzie and Clough, Trans., 1908,93, 811 ; 1909, 95, 777). For example, I-phenylchloroacetic acid,C6H,*CHCl*C0,H, is converted into a mixture of r- and Emandelicacids when aqueous sodium hydroxide is used t o displace the chlorineatom by the hydroxy-group; a mixture of r- and d-mandelic acidsis, however, produced when silver carbonate is substituted for sodiumhydroxide.This behaviour makes the problem of the Walden inversionmore complicated than before, for this reason that, by analogy withprevious work of Walden and others, it was to have been expectedthat sodium hydroxide should have caused the formation of a deztro-rotatory mandelic acid mixture from the laevorotatory chloro-acid, andthat silver carbonate should have caused the formation of a Zaewo-rotatory mandelic acid mixture. The contrast between the inter-conversion of the active lactic acids, CH,*CH(OH)-CO,H, on theone hand, and the interconversion of the active mandelic acids,C,H,-CH(OH)*CO,H, on the other, by the Walden inversion, isvery striking, and must be taken into account in any interpretationregarding the mechanism of the action which may be advanced.The present research is concerned with changes undergone by theoptically active P-hydroxy-P-phenylpropionic acids,C,H,*CH(OH)*CH,*CO,H.One of the objects was to find out if any Walden inversion could bedetected in the course of changes undergone by a compound where thecarboxyl group is not attached directly to the asymmetric carbonatom. Meanwhile this problem has been investigated by E.Fischer andScheibler (Ber., 1909, 42, 1219), who studied the displacement of thehydroxy-group in Z-P-hydroxybutyric acid, CH,*CH(OH)*CH,*CO,H.By the action of phosphorus pentachloride, this :acid was convertedinto d-/3-chlorobutyric acid, from which the original I-hydroxy-acidwas regenerated by displacing the chlorine by means of a number ofdifferent agents.Bearing in mind the possibility of the phenyl group exerting aninfluence entirely diff went from that of the methyl group, we preparedthe optically active P-hydroxy-P-phenylprpionic acids by the resolu122 McKEEjZIE AND HUMPHKIES :tion of the inactive acid with morphine in aqueous solution.Whenthe I-acid is acted on by hydrobromic acid, the resulting bromo-acidis dextrorotatory, and the change is accompanied by a certain amountof racemisation, which is less pronounced when the temperature a twhich the displacement occurs is kept low. The behaviour of thed-hydroxy-acid towards hydrobromic acid is, of course, similar.Phosphorus pentabromide also brings about a change of sign ofrotation by its action on the active hydroxy-acids.When the brominein the active bromo-acids is displaced by the hydroxy-group, either bymeans of silver oxide and water, or by sodium carbonate and wateror by water alone, a change of sign of rotation again occurs. Theparent acid is accordingly regenerated :by HBrl-C6H,*CH(OH)*CH2*C0,H ;y& d-C6H’,*CHBr*CH2*C02Hby waterAg20 or Xu‘azCOs--- -+ l-C,H5*CH(OH)*CH2*C02H.The displacement of the hydroxy-group in the active P-hydroxy-P-phenylpropionic acids by the bromine atom appears t o be a normalaction, since both phosphorus pentabromide and hydrobromic acid actin a similar manner, and since a change of sign also accompanies theaction of hydrobromic acid on the methyl d-ester.There is, therefore, no evidence of the occurrence of a Waldeninversion in any of the changes studied.EXPERIMENTAL.Resolution of Inactive /3-Hydroxy-/3 phenylpropionic Acid.Inactive /3-hydroxy-P-phenylpropionic acid was prepared by theaction of boiling water on inactive P-bromo-P-phenylpropionic acid,which is readily obtained from hydrobromic acid and cinnamic acid(Fittig and Binder, AnnaZen, 1879, 195, 131).The resolution by means of morphine proceeds with exceptionalease.So far as we are able to judge, it is immaterial whethersynthetic or storax cinnamic acid is used as the starting point for thepreparation of the inactive acid.Powdered morphine (61 grams) was added t o a solution of 36 gramsof the hy’droxy-acid in 750 C.C. of boiling water. Crystallieationbegan after the solution was allowed to cool at the ordinary tempera-ture for one hour ; the solution was then stirred occasionally, and leftovernight at the ordinary temperature.About half of the totalmorphine salt separated. The crystals, which melted and decomposedat about 2 0 6 O , were suspended in 50 C.C. of water and the morphineprecipitated by means of a slight excess of ammonia. The addition ofan excess of hydrochloric acid to the filtrate, from which the morphinEXPERIMENTS ON THE WALDEN INVERSION. PART 111. 1%ha3 been separated, caused the gradual separation, in the formof needles, of the Z-acid, which is sparingly soluble in water. Theacid was drained off and, after crystallisation from 300 C.C. of benzene,was pure. The yield amounted to 9 grams.Its melting point and itsspecific rotation did not alter after it had been recrystallised severaltimes from benzene.l-P-Hydrox~-P-pl~elzy~?.opionic acid, C,H,*CH( OH)* CH,. CO,H, issparingly soluble in water and in benzene. It separates in colourlessneedles and melts at 115-1 16' :001925 gave 0.4579 CO, and 0.1032 H,O.C9H,,08 requires C = 65.0 ; H = 6.1 per cent.Its rotation was determined in ethyl-alcoholic solution :C = 64 *9 ; H = 6.0.I = 2, c = 5.153, a: - 1-95', [u]: - 18.99I n order to obtain the enantiomorphously related isomeride, themother liquor, from which the morphine I-salt had been separated, wasconcentrated by evaporation to 150 c.c., when no separation of salttook place. The dextro-acid was then separated in the mannerdescribed above and crystallised from benzene.The yield amountedto 10 grams.d-p-Hydroxy-P-phelzylpropionic acid melts a t 115-1 16O, andresembles its Eisomeride in other particulars :0.194 gave 0.4638 CO, and 0.1051 H,O.C,€€.,,O, requires C = 65.0 ; H= 6.1 per cent.A determination of its specific rotation in ethyl-alcoholic solutionC = 64.9 ; H = 6.0.gave a value in agreement with that of the 1-acid :Z = 2, c=5.194, a? + 1-99', [a]: + 19.2'.Although the inactive acid had not been resolved previously, theactive acids have been obtained by Barkow (Inaug. Diss., Strasburg,1906), working in Erlenmeyer's laboratory, in the course of aninvestigation dealing with the a-halogen-P-hydroxy-/3-phenylpropionicacids. Barkow found that when d-u-bromo-/3-hydroxy-/3-phenyl-propionic acid, C,H,*CH(OH)*CHBr-CO2H, was reduced by sodiumamalgam, it was converted into d-/3-hydroxy-P-phenylpropionic acidwith [a]= + 19' in ethyl-alcoholic solution.Action of H9drobromic Acid on the Active P-H?/drox~-P-p~enyZpropionicAcids.Attempts to resolve inactive P-bromo-P-phenylpropionic acid into itsoptically active isomerides were not promising.The bases employedcaused some decomposition of the bromo-acid into styrene. The activebromo-acids were accordingly obtained from the correspondin124 McKENZIE AND HUMPHRlES :hydroxy-acids, but they underwent partial racemisation in the processof their formation by this method.The I-hydroxy-acid ( 2 5 grams) was covered with aqueous hydro-bromic acid, which had previously been saturated at 0'.The hydroxy-acid dissolved, and the bromo-acid separated. After one hour, waterwas added, the sparingly soluble acid drained off, washed with water,and dried over soda-lime under diminished pressure. The product had[a], + 16.8' for c = 2.059 in ethyl-alcoholic solution. It was a mixtureof the T- and d-bromo-acids, since its melting point was indefinite andits rotation changed on crystallisation. The effect of crystallisingthree times from carbon tetrachloride was to give an acid mixture,which contained more of the inactive form than before, the value for itsrotation in ethyl-alcoholic solution being [aID + 8.5' for c = 2-05.If the fuming hydrobromic acid is shaken with the I-hydroxy-acidfor a few minutes only at the laboratory temperature, the racemisationis less pronounced.In one experiment, for example, the crude bromo-acid, obtained from the I-hydroxy-acid, was crystallised once fromcarbon tetrachloride, and then gave [aJD + 20.6' for c = 2.204 in ethyl-alcoholic solution.When the d-hydroxy-acid was shaken with aqueous hydrobromicacid, saturated at Oo, for two or three minutes at Oo, the crudebromo-acid which separated had [a], - 23.9O for c = 2.974 in ethyl-alcoholic solution.Obviously, therefore, the amount of racemisation could be lessenedby maintaining the temperature low during the action of the hydro-bromic acid. Forty C.C. of aqueous hydrobromic acid (saturated a t 0')were accordingly cooled to - lo', and 8.5 grams of the I-hydroxy-acidadded.The rapid solution of the hydroxy-acid was succeeded by theseparation of a voluminous crop of the bromo-acid. After five minutes,the crystals mere separated, washed with a little water, and dried. Theproduct melted indefinitely at 126-133O, and had [aID +32*2' forc = 2.125 in ethyl-alcoholic solution. It mas crystallised from 65 C.C.of chloroform, and the crop which separated (6 grams) had [a]= f21'.From the mother liquor, two successive crops were withdrawn, thesecond of which (1.1 gram) had [a], +58*3 for c=1.03 in ethyl-alcoholic solution, whilst the residual mother liquor yielded 2 grams ofacid with [alD +96.2" for c=1*107 in ethyl-alcoholic solution. Anestimation of bromine in the latter acid indicated the presence ofcinnamic acid together with the bromo-acid.The pure active bromo-acids have accordingly a value for theirspecific rotation higher than 96*2O, and appear to be more readilysoluble in most solvents than the inactive isomeride.Furtherattempts ko isolate them were not made, since the points of interestEXPERIMENTS ON THE WALDEN INVERSION. PART 111. 125from the point of view of this investigation, could be established by aidof the partly-racemised acids.The d-hydroxy-acid gave similar results t o the above when it wastreated with fuming hydrobromic acid at - 10'.Action of Phosphorus Pentabromide on the 1- Zlydroxy-mid.The E-acid WRS dissolved in a mixture of chloroform and ether andacted on with an excess of phosphorus pentnbromide, the temperaturebeing maintained low.The bromo-acid, obtained after decompositionof the acid bromide with water, gave [.ID + 14.4" fcr c = 2-39 in ethyl-alcoholic solution.Thus phosphorus pentabromide behaves like hydrobromic acid incausing a change of sign of rotation when it acts on the activehydroxy-acid, As is usually the case when a phosphorus halide actson an active hydroxy-acid, the formation of the balogen acid isaccompanied by partial racemieation.Tormation of 1-Bromo-ester from d- Pydroxy-ester.d-P-Hydroxy-P-phenylpropionic acid was converted into its methyester by the Fischer-Speier method. This ester, which had [.ID + 14.1"for c = 4-717 in ethyl-alcoholic solution, was added to an excess offuming hydrobromic acid at - lo", and, after shaking for five minutes,the bromo-ester was separated.It was laevorotatory, giving[.ID - 28.5' for c = 6.654 in ethyl-alcoholic solution.The action of phosphorus pentabromide on the d-hydroxy-ester wasalso examined, the bromination being effected in dry chloroform andat a low temperature. The resulting bromo-ester was again Irevo-rotatory, giving [.ID - 4.6" for c = 3.03 in ethyl-alcoholic solution.Digplncemelit of Bromine in the d-Bromo-acid by t ? ~ IIydrozy-group.A mixture of the dextro- and inactive bromo-acids (0-4 gram) with[.ID +58.3* was added t o water (10 c.c.), and, after five days at theordinary temperature, the solution was heated for a few minutes untilthe odour of styrene had disappeared. The pi-oduct was thenevaporated to dryness at the ordinary temperature under diminishedpressure.The residue gave the value [.ID - 7.7" for c = 2.27 in ethyl-alcoholic solution.A mixture of the dextro- and inactive bromo-acids (0.7 gram) with+ 58.5' was added t o a solution of 0.4 gram of sodium carbonateAfter five days, the small amount of styreneOn acidification with hydrochloric acid, therein 10 C.C. of water.present was removed126 FORSTER AND MULLER: THE TRIAZO-GROUP. PART XI.was no appreciable separation of cinnamic acid. The hydroxy-acidwas extracted with ether, and had the specific rotation [.ID - 5.5' forc = 1 *451 in ethyl-alcoholic solution.Silver oxide, obtained from 1 gram of silver nitrate, was added tothe dextrorotatory bromo-acid (0.7 gram) with [a], + 96.2' and 10 C.C.of water. After twenty-four hours, with occasional shaking, theproduct was treated with hydrochloric acid, filtered, and the filtrateextracted with ether. The resulting hydroxy-acid gave the value[a]= - 1 3 0 7 ~ with c= 1.17 in ethyl-alcoholic solution.In these cases, therefore, using either water alone, sodium carbonateand water, or silver oxide and water, the hydroxy-acid recovered isopposite in sign to t h a t of the bromo-acid used, and of the samesign as the parent hydroxy-acid from which the bromo-acid wasobtained.BIRKBECK COLLEGE,LONDON, E.C