CERTAIN ACIDIC OXIDES ON SALTS OF HYDROXY-ACIDS. 1457 XCV.-Some Derivatives of Pyopionic acid, o f Acrylic ucid, ccnd of Gluturic acid. By WILLIAM HENRY PEXBIN, jun. ! ~ E S E researches originated in an attempt to prepare synthetically an acid of the formula CH*COOH \/- C(CHS)*COOH which Bredt (Ber., 1893, 26, 3049) first proposed as being a very probable expression of the constitution of camphoric acid. The starting point selected was a~~-trirneth~lprop.;Mzic acid (methyl- isopropylacetic acid), and this was prepared by a method already described by Romburgh (Iiec. ITrav. Chim,., 1886,6,236), namely, the distillation of methylisopropylmalonic acid. CH(CH3)2*C(CH,) (COOH)2 = CH(CH~)2~CH(CH~)*COOH + Cop The very carefully purified acid was then converted into ethylic 2-brmotrimethyZpropionate, CH(CH,),*CBr( CHr,)*COOC,H5, by first treating it with phosphorus and bromine (Hell, Ber., 1881,14, 891 ; Volhard, Annalen, 1887, 242, 161), and subsequently pouring the product into absolute alcohol.If now this ethereal salt were treated with alcoholic potash, i t should yield an unsaturated acid, CsBl0O2, and in the formation of this acid it was anticipated that the elimination of hydrogen VOL. LXIX, 5 F1458 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, bromide would take place between the bromine atom and the B-hydrogen atom marked * thus : C&(CH3)z*CBr(CH3)GOOC2H, + 2 K 0 H = C(CH,)2:C(CH3)*COOK + C,H5*OH + KBr + HzO, with formation of apB-triiitethylacl.yl~~ acid, and not between the bromine atom and a hydrogen atom of the a-met,hyl group, to form a-isopropylacrylic acid, CH(CH3),*CBr(CH3)*COOC2H, + 2KOH = CH(CH,)2*C(:CH,)*COOK + C2H5*0H + KBr + H20, because i t is generally found that the hydrogen atom in the C H of the isopropyl group reacts much more readily than the hydrogen atoms i n methyl or ethyl groups.Now, as Auwers and many others have shown, the ethereal salts of ap-unsaturated acids readily condense with the sodium derivative of ethylic malonate, forming additive products ; for example, in the case where ethylic acrylate is employed, the substance formed is ethylic propanetricarboxylate : (COOC2H5)2CH2 + CH,:CH*COOC,H5 = (COOC,H,)2CH*CH2*CH,.C 00 C2H5. As this reaction takes place in all the cases which have so far been investigated, it seemed probable that if etlhylic tri methylacrylate were digested in alcoholic solution with the sodium derivative of ethylic mttlonate, a similar condensation would take place, and that ethylic trimethyl;propanetricarboxylate would be formed, thus : (GO 0 C2Ha),CH2 + C (CH3) 2:C (CH,).COO C2H6 = (COOC2€3,)2CH*C (CH,)2*CH(CH,) *COOC2H5.This ethereal salt would be a most interesting substance in many ways, as, apart from the capabilit,y, which it would possess, of form- ing a sodium compound, ( C 0 0 C2H,),CNa*C ( CK,) ,*CH( CH,) COO C2H6, and its consequent value in synthetical work, it would, 011 hydrolysis, yield a tribasic acid, which, when heated, would lose 1 mol. of carbon dioxide, with formation of ap/3-t,.iineth?llgZutaric acid. ( C 0 OH) zCH* C ( C Ha) 2. C H ( CH3) * C 0 0 H = C 0 0 H* C H2* C ( C H,) 2-C H (C H3) *C 0 OH + C 0 2 .It would be very interesting to synthesise this aeid, in order to compare the synthetical acid with an acid which L. Balbiano (Rer., 1895, 28, 1507) obtained from a product of the oxidation of cam- phoric acid with cold potassium permanganate, and which this chemist considers to be c+@trirnethylglutaric acid.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1459 I n using ethylic trimethylpropaneti~icarboxylate for the synthesis of camphoric acid, i t was next proposed to treat its sodium derivative with ethylene chlorhydrin, in order in this way to prepare ethylic hydroxyet h y ltrimeth y lpropanetricarbozy late, thus : (C 0 OC2H5)2CNa.C( CH,),*C H( CR,) G O OCaHj + C H,Cl*C H,*OH = (CO OCsHj),*C (CH,*CH2OH).C(CH3),*C H (CH3) *CO OC2H5 + NaCl, a decomposition which might be expected to proceed in this way, since a strictly analogous reaction, namely, the synthesis of ethylic hydr- oxyethy lacetoacetate, CH3*CO*C H( CH,*C H2.0H)*C0 0 C2H5, by the interaction of ethylene chlorhydrin with the sodiam compound OE ethylic acetoacctate, has been accomplished by Chanlarow (Annalen, 1884, 226, 326).On hydrolping the ethereal salt thus obtained, the corresponding tribasic acid would be formed, and this, as it is a derivative of malonic acid, should, when carefully heated, lose 1 mol. of carbon dioxicie, with formation of It ydroxyetl~yItrtmeth?llglutaric ucid. ( C 0 0 H ) 2C ( C: H2*C H,*O H ) C ( C H3) ,*C H ( CH,) C 00 H = Lastly, as Bredt’s formula for camphoric acid represents this sub- stance as containing a 5-carbon ring, and as this ring is generally produced with great ease, it seemed possible that, by treatment with dehydrating agents, or by other means, an acid of this formula might be formed by the simple elimination of mater from hydroxyethyl- drimethylglutaric acid, thus : C 0 0 H* C H ( C H 2.CH2.0 H) C (C H3) 2. C H (C Ha) C 0 0 H + C 0 2 . CH-COOH CHCOOH /\ H z y )C(CH3), + H,O. j H 9*CH3 COOH In investigating the various reactions described above, very un- expected difficulties were met x i t h , necessitating a very careful examination of the methods of preparation and behaviour of a large number of interesting substances which had not previously been pre- pared, and I beg now t o lay before the Society an account of some of the experiments which have, so far, been carried out, reserving the description of the others for a future occasion. A s preliminary experiments soon showed that the preparation of pure a,@,&trirnethylpropionic acid, CH(CH,),*CH( CHJ*COOH, in quaotity required much time, and was a matter of considerable diffi- 5 ~ 21460 PERKIN : SOME DERIVATITES OF PROPIONIC ACID, culty, and as the preparation is a very expensive one, it was thought best, in the first place, to experiment with isoraleric acid (PP-&- methylpropionic acid), CH(CH,)2*CH2*COOH, in order to discover the conditions which would probably be most favourable for the sub- sequent work with the trimethyl acid.Pure isovaleric acid was brominated in the presence of phosphorus in the usual way, and the product converted into ethylic a-bromo- valerate, CH(CHJ2*CHBr*COOC2H5, by treatment with alcohol, the conditions for obtaining the best yield being carefully determined.In order now to eliminate hydrogen bromide from this brominated ethereal salt, if was either hydrolysed by means of alcoholic potash, CH(CH,)~*CHBI**COOC~H~ + 2KOH = C(CHs)2:CHCOOIC + CzH5*OH + KBr + H20, or digested with quinoline (compare Weinig, Annaten, 1894, 280). In the firat case, dimethylacrylic acid is at once obtained, whereas,. in the second case, its ethereal salt is produced, and as in the subse- quent experiments the latter waR nearly always required, the second method was most frequently employed. Dimethylacrylic acid has already been described by various investigators (see p.1469) : i t is a beautifully crystalline acid, which melts at 70' ; ethylic dimethyl- acrylate is a colourless oil and boils at 155'. The condensation of ethylic dimethylacrylate with the sodium derivative of ethylic malonate was next invest+ ted, and found t o proceed normally, ethylic dimethylpropnnetricarboxylate being formed,, thus : (COOC,H,),CH, -+ (CH,),C:CH*COOC2H, = (C 00 CZH,) ,CH*C( C H,) 2. CH2*C 00 C2H5- But the determination of the conditions for obtaining thc best yield of this condensation product gave a considerable amount of trouble, as, in working under the conditions usually employed, only a very small yield of the substance is obtained ; * ultimately, however, a method was devised by which it is posbible to prepare considerable quantities of this ethereal salt.Etlhylic dimethylpropanetricarboxy- late is a colourless oil, which boils at. 203' (60 mm.) ; on hydrolysis, * An abstract of this work on ethylic dimethylprop~netricarboxylate containing a description of @&dimethylglutaric acid and some of its derivatives, appeared some time since in the Proceediitgs (W. Goodwin and W. H. Perkin, jun., 189$, 64). Auwers (Ber., 1895, 28, 1130 ; Annaleia, 1896,292, 145), not knowing of this publication, eubsequently investigated the same subject, and obtained results which confirm, in a welcome way, those described here. Auwers also noticed that the yield of ethylic dimethylpropsnetricarboxylate, obtained by condensing ethylic dimethylacrylate with the sodium derivative of ethSlic malonate under the ordinary conditions, was very small, being only about 8 per cent.of the theoretical.OF ACRYLIC ACID, AND OF GLUTARlC ACID. 1461 it yields the corresponding tribasic acid (m. p. 173'), and this, when heated at 200°, loses carbon dioxide, with formation of /3,3-dimethyl- glutaric acid. ( C 0 OH) ,CH*C ( CH,),*CH2- C 0 0 H = This melts at lolo, and, when treated with acetic anhydride, is converted into the anhydride, C ( C H 3 ) 2 < ~ ~ $ ~ > 0 , which melts at 124O, or 23' higher than the acid itself, a very unusual thing, and especially interesting when it is remembered that camphoric acid, whicb is supposed by Bredt t o be a derivative of dimethylglutaric acid, yields an anhydride mhich melts 29-30' higher than the acid itself does. Dimethylacry lic acid yields a well cha.racterised anilic acid, COOH*CH2*C(CH3)2*CH2*CO*NH-C6Hs (m.p. 134O), and this, at its boiling point, is converted into the corresponding anil, C 0% + C 0 0 H. C H2* C (C H3) 2.C H 2.CO OH. which melts at 156-157'. One of the most remarkable points in connection with this acid is its abnormally high dissociation constant, the value found by Dr. James Walker being K = 0.0200. Dr. Pfaff subsequently examined the acid obtained by Auwera and Avery (Annalen, 1896, 232, 147), and confirmed tvhe above result, his determination giving the value K = 0.0220. The dissociation constants in the glutaric series gener- ally vary between 0*0050 and 0*0060, the only other exception to this rule, which has so far been observed, being, as Auwers points out, the aaa,-triruethylglutaric acid, C 0 OH* C (C H3) 2.C H2* C H ( CH,)*C 0 0 H, which has the low dissociation constant K = 0.0035.This point is again interesting in view of the possible connec- tion between camphoric acid and ljp-dimethylglutaric acid, but in this respect the values for the two acids are widely different, the dissociation constant of camphoric acid (K = 0~00225) being abnormally low . While the above experiments were i n progress and Dearly com- pleted, a similar series of reactions, starting with aPP-trimethylpropi- onic acid, CH(CH,),.CH (CX&COOH, in the place of isovaleric acid, were being pushed forward. In the first place, this acid was treated with bromine in the pres- ence of phosphorus, under the same conditions as those employed in1462 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, the case of isovaleric acid, and the product was decomposed by the addition of alcohol.In this way n very good yield of ethylic a- brom 0- app- trimeth y lp ropion a t e, C H (C H,) ,* C R r (C H3) CO 0 H, was o b- tained as a colourless oil boiling at 130° (100 mm.) without decom- position. This ethereal salt, on hydrolysis with alcoholic potash, or when digested with quinoline, behaved, to all appearance, exactly as described above in t'he case of ethylic bromisovalerate, yielding, in the first instance, an oily acid boiling at 204-205', and, in the second case, an ethereal salt boiling, not very constantly, at 162-167' ; and for a long time it was believed t h a t these reactions proceeded in the following manner : (1) CH(CH3),*CBr(CH3)*C:OOCZH, + 2KOH = (2) CH(CH3),*CBr(CH3)*COOC2H6 = HBr + C(CH3)2X(CH3)*COOK + KBr + H,O + C,H,*OH; C (CH3) 2: C( CH3)*COO C,H5 ; that the acid was, in fact, trimethylacyylic acid, and the ethereal salt ethy lic trimethylawylate. As the work progressed it was, however, soon seen that it was most important to be perfectly sure that the constitution of these substances is that given above, and this was found to be a dificult matter, and entailed many months' work.A quantity of the acid boiling a t 200--205° was prepared by t h e hydrolysis of ethylic bromotrimethylpropionate, and also by t h e hydrolysis of the ethereal salt produced by digeating this bromo- ethereal salt with quinoline, the oily acid obtained in both cases appa- rently having the same composition. When this oily acid is allowed to stand for a long time in a cool place, it gradually deposits thick prismatic crystals ; these were collected and purified by recrystallisa- tion; they then melted at 70-'71', and were subsequently proved to consist of pure a@/?-trimethy Zacry Zic acid, C (CH,),:C(CH,)*COOH. This acid combines with bromine to form a/3-dibro~~ao-a~~-trimethyl- propionic acid, CBr(CH,),.CBr(CH,)*COOH (m.p. 1 9 1 O ) ; with hydrogen bromide it yields p- bromotrimethylpyopionic acid, CB 1. (CH3) 2.C H ( C H3)' C 0 0 H (m. p. 8S0), and with hydrogen iodide, p-iodotrimethylpropionic acid (m. p. €32'); the formation of these substances, which takes place almost quantitatively in each case, was used as a means of characterising and identifying the acid.In order to prove that the above acid was in reality trimethyl- acrylic acid, it was necessary to prepare it by some method which left no doubt as to its constitution, and this was ultimately accom-OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1463 plished by employing a reaction which was first devised by Frank- land and Duppa, and which has lately come into some prominence owing to the researches of Reformatsky and others. Frankland and Duppa showed that a-hydroxy-derivatives of the fatty ac.ids may be prepared synthetically by the action of zinc on a mixture of methylic oxalate and an alkylic balo'id ; methylic hydroxy- isobutyrate, for example, is formed when zinc acts on a mixture of methylic oxalate and methylic iodide (AnnnZen, 1865, 133, 80) ; this reaction may be conveniently represented thus : Q(CH,),*OZnI COOCH, (?(CH3'2'0H + ZI~(OH)~ + HI.COOCH, Ir. I + 2H20 = + Zn(OCH3)I. + 2Zn + SUH,I = COOCH, C(CH3),*OZnI COOCH3 1. I COOCH, Reformatsky (.Journal of the Russian Chemical Society, 1890, 22, 49) subsequently extended this reaction to ketones aEd ethereal salts of a-halogen fatty acids, and succeeded in this way in spthesising P-hydroxy-fatty acids. As an example of this important method, the action of zinc on a mixture of acetone and ethylic monochlor- acetate may be given. I. (CH3)2C0 + CH2CI*COOC2H5 + Zn = C(CH,)2(OZnC1)*CH2.COOC2H5. OH*C(CH3)2*CH2*COOC2H5 + Zn(OH)* + HCI. With the aid of this reaction, trimethylacrylic acid may be pre- pared in a way which clearly proves its constitution, and in carrying out the experiment on this point, I was fortunate in having the assistance of Dr.J. F. Thorpe. When a mixture of acetone and ethylic a-bromopropionate is treated with zinc, under snitable conditions, condensation readily takes place, with formation of a peculiar zinc compound, the reaction evidently proceeding in the following way : (CH3)2C0 + CHBr(CH3)*COOC2H, = C ( CH3) 2 ( 0 ZnB r) C H ( C H,) C 0 0 C2H,. This zinc compound is decomposed, on treatment with water and dilute sulphuric acid, with formation of efhyZic t3-hydrozy-aPP-lri- inethyZpopionate, a thick, colourless oil which distils at 105' (30 mm.) without decomposing. 11. C(CH,)~(OZUC~).CH~.COOC~E~ + 2H20 = C(CH3)2(OZnBr)*CH(CH3)*COOC2H6 + 2H20 = OH*C(CH3)2*CH(CH3)*COOC2H5 + Zn(OH)2 + HBr.This ethereal salt, on hydrolysis, yields the corresponding hydroxy- acid, a colourless oil which distils at 160' (35 ram.), and which is1464 PERRIN : SOME DERIVATIVES OF PROPIONIC ACID, very readily acted on by concentrated aqueous hydrobromic and hydriodic acids, yielding P-brornot,rimethylpropionic acid, CBr( CH&.CH( CH,) GOOH, and /3-iodotrimethylpropionic acid respectively, substances which are identical with the acids already mentioned as being produced by the addition of hydrogen bromide and hydrogen iodide to trimet hylacrylic acid. Lastly, trimethylacrylic acid (m. p. 70') is obtained when [3-bromotrimethylpropionic acid is treated with alcoholic potash, CBr(CH3)2*CH(CH3)*COOH = C(CH3)2:C(CH3)*COOH + HBr; and as this same acid is also formed by the elimination of hydrogen bromide from a-bromotriniethylpropionic acid, it can only Lave the constitution represented by the formula C(CH3)2:C(CH,)-COOH.These experiments proved conclusively that the ethereal salt formed by the action of quinoline on et hylic a-bromotrimethylpropionate as described above, consists, certainly in part, of ethylic trimethylacrlp- late, and for a long time it was thought that it, was composed wholly of this compound. On this assumption, the experiments on the con. densation of this ethereal salt with the sodium compound of ethylic malonate were proceeded with, and a condensation product was obtained, which was naturally supposed t o be ethylic trinzethylrpro- pa netricar boxytate, (COOG;?H,)2CH2 + C(CH3)2:C(CH,)*COOC*H6 = (COOC,H,)2CH*C (CH&*CH( CH,)*COOC&H,.Unfortunately the yield of this new substance is very small, milch smaller, indeed, than the yield of ethylic dime thylpropanetricnrb- oxylate from ethjlic dimethylacrylate, and a large number of experi- ments carried out under the most varyiug conditions failed to increase the yield to more than about 10 per cent. of the theoretical. The substances used in this reaction are so difficult to prepare that it, was almost impossible to continue the experiments unless the yield of condensation prod uct could be considerably increased, and in order bo get over the difficulty, experiments on the action of ethylic a- bromotrimethy1propicinate on the sodium derivative of ethylic malon- ate were next instituted. In this decomposition, which takes place readily, it was, of course, possible that the two substances would react normally, with formation of ethylic methylisopropglethanetri- carboxylate, thus- (COOC2H5),CHNa + COOC2Hb*CBr( CH,)*CH(CH,), = (COOC2H,),CH*C(COOC2H,)(CH,)*CH(CH,), + NsBr.but, judging from the behnviour of other a-bromo-ethereal salts, such as, for example, e th y lic a-bmmoisobuty ra t e, CB r (CH,) ,*COO C2H5, nnder similar Gircumstances, it seemed probable that the reactionOF ACRYLIC ACID, AND OF GLUTARIC ACID. 1465 would take place in two stages, that is, that ethylic trimethylacry- late, C(CH3),:C(CH3)*COOC2H5, would be first formed by the elimi- nation of hydrogen bromide, and that this unsaturated ethereal salt would then condense with the ethylic malonate present to form ethylic trimetl~ylpropanetricarboxylate, as explained above.It was thought likely that, at the moment of formation, the ethylic trimethylacrylate might condense more readily with the ethylic malonate, and give a larger yield of condensation product. Both these assumptions were found to be correct., as not only did a careful comparison or" the product obtained prom that it was identical with the condensation product obtained on digesting the sodium derivative of ethylic mrtlonnte with the supposed ethglic trimethylacrylate as described above ;* but the yield was also considerably larger, being, indeed, sometimes considerably over 20 per cent. of the theoretical, and as this method of preparation involves fewer operations, and therefore requires much less time, it was used in all subsequent preparations of this ethereal salt.In the course of a careful investigation of the condensation pro- duct obtained by either of the above methods, several facts came to Jight which made i t doubtful whether, after all, the substance was really eth ylic trime th ylpropane tricerboxy iat e. The condensation product, on hydrolysis, yields a beautifully crystalline tribasic acid, which, when heated at 200°, readily loses l. mol. of carbon dioxide, with formation of a crystalline dibasic acid, which should be a&%trimethyl- glutaric acid. (COOH),CH*C( CH,),*CH( CH,)*COOH = CO, + COOH*CH2*C(CH~),*CH(CH~)*COOH. The acid thus produced is very similar to the acid which Balbiano (Bey., 1895, 28, 1507) obtained from camphoric acid, and which is very probably ap/3-trimethylglutaric acid; the former melts at 94-95O, and gives an anilic acid, meltingat 258-159', whereas the latter melts at 89", and yields an anilic acid melting at 150°, but the anhydrides of the two acids differ consideqably, that from the acid obtained by me melting a t 53", whereas the anhydride of Balbi_a_no's acid melts at But, because the acid is not identical with Balbiano's acid, it does not prove that it is not apl/3-trimelhylglutaric acid, for the reason that the constitution of Balbiano's acid is not known with certainty ; 80-a 10.* These experiments were conducted, in the usual manner, in alcoholic solution ; in x2lene solution, aleo, the reaction proceeds in the same way, which is rather remarkable, as usually the uneaturated intermediate substance is not so readily formed from a-bromo-ethereal salts under these conditions, the product consisting, a t all events to a large extent, of the normal ethane dqivative.1466 PERKIN: SOME DERIVATIVES OF PROPIONIC ACID, the latter might, for example, quite well be aap-trimethyl~lutaric acid, COOH*C(CH3),*CH(CH,)*C&*COOH, nevertheless it seemed desirable to further investigate the subject, and, as the result, it was ultimately conclusively proved that the synthetical acid is not tri- met h ylglu t aric acid, but a-isop ropy1 g lu t aric acid, C H (CH3) 2*CH (C 0 0 H) *CH2* CH, C 0 0 H.On oxidation with chromic acid the acid yields, besides acetic acid, only succinic acid, whereas from trimethylglutaric acid under these circumstances trimethylsuccinic acid should be formed.C H ( C H3) C H (C 0 0 H) CI&-C H2* C 0 0 H. ' Isopropylgluhric acid. C 0 OH*/ CH2- C (CH3),*C H ( CH,) C 0 0 H. ' Trimethylglutaric acid. Subsequently it was shown that the acid is identical with the isopropylglutaric acid, which, for the sake of comparison, was syn- thesised by Mr. Heinke and myself from the product of the action of ethylic /%iodopropionate on the sodium compound of ethylic isc- propylmalonate, by hydrolysis and subsequent elimination of carbon dioxide. I. C H (C H3) 2.CN a ( C 0 0 C2H5) + I* CHz*C Hz* C 0 0 C,H, = NaI + CH (C H&*C (C 00 C2H5)z*CHz*C H2.C 00 C2H3. I I. C H (C H3) 2. C ( C 0 0 H) Z*CHz*CH2*C 0 OH = C02 + CH(CHa)2*CH(COOH)*CHz*CHz*COOH* A careful re-investigation of the whole matter showed that when ethylic a-bromotrimethylpropionate is treated with quinoline, the product does not consist entirely of ethylic trimethylacrglate, but is x mixture of this subdance with ethylic a-isopropylacrylate, the elimination of hydrogen bromide from the brom-ethereal salt having taken place in two directions. CH (CHs),*C (C 0 0 C2H5) : C H2, ...... - . I:. (CH,) 'L GH- C B r (C 0 0 C2H5) *C H3 = C(CH3)2:C(COOC2H',)-CH3 + HBr. . ........................... I....... . 11. (CH&CH*CBr (C 00 C2H,) CH3 = (CH&J3H*C(COOC2H5):CH2 + HBr- When these mixed ethereal salts are digested with the sodium derivatire of ethylic malonate, the ethylic isopropylacrylate alone enteis into the reaction.OF ACRYLIC ACID, AND OF GLUTARIC ACID.1467 (COOC2H5)zCHz + CH2:C(COOC2&)*CR(CH,), = (CO OC2H,),C H a C H,* C H (C 0 OC,H,) *C H (CH,) 2.. Ethylic isopropylpropanetricarboxylate. The ethylic trimethylacrylate takes apparently" .no part in the con- densation, and, indeed, the latter snbstance may be, at all events, partially recovered from the product of the action. This behaviour of ethylic trimethylacrylate is very remarkable, and, so far, without parallel ; very probably t.he further investigation of the condensation of a/3-unsaturated ethereal salts with the sodium, derivatives of ethylic malonate and allied compounds mill show the nature and number of the groups, which, when attached to the double band, render the condensation a matter of difficulty, or, in some cases, prevent it altogether.During the course of these experiments, and while it was thought that the condensation product described above had the constitution, (CC)O C2H5) ,CHt* C( C H,),*CH( C H)3* CO 0 C,H5, experiments were being made on the introduction of the group -CH,-CH,*OH, at the point marked t, for the reasons given at the commencement of this paper. These experiments were not in the first instance actually made with this condensation product, owing to the great difficulty of obtaining it in any quantity, but with somewhat similarly constitnted sub- stances which could be more readily prepared. Arguing from t h e results of Chanlarow's experiments on the action of ethylene chlor- hydrin on the sodium derivative of ethylic acetoacetate (Annulen, 1884, 226, 326), it seemed probable that the best way of achieving the object in view would be to treat the sodium compound of this condensation product with ethylene chlorhydrin, but although many experiments were made on the action of this substance on the sodium derivatives of ethylic methylmalonate, ethylic isopropylmalonate, and on the corresponding mono-substitution derivatives of ethylic aceto- acetate, in no case could more than traces of the hydroxyethyl sub- stitution product be isolated, the unaltered ethereal salt being in all cases recovered almost quantitatively.This want of success led to the investigation of the action of yphenoxyethylic bromide, C6H5-O-CH2*CH2Br, on the sodium deri- vatives of ethylic malonat e, ethylic methylmalonate, and similarly constituted compounds ; a description of tbese experiments, all of which gave satisfactory results, has already appeared in the Transac- * A careful examination of the mother liquors from the crystallisation of the isopropy lpropanetricarboxylic acid resulting from the hydrolysis of the ethylic salt produced by the condensation failed to reveal thc presence of even traces of tri- methylpropanetricarbox~lic acid.This acid may, neverthelees, of course have been ,present in small quantity and have escaped detection.1468 PEBKIN : SOME DERIVATWES OF PROPIONIC ACID, tions of this Society (W. H. Bentley, E. Hawortb, and W. H. Perkin, jun., 1896, 69, 161). While this preliminary work was in progress, the action of phenoxyethylic bromide and sodium ethoxide on ethylic dimethyl- propanetricsrboxylate wa8 being investigated, became this ethereal salt is much more readily prepared than the substance now known to be ethylic isopropylpropanctricarboxylate. In this case it was expected that the reaction would proceed thus.( COOC2H,),CNa*C ( CH3)2*CH2*COOC2H5 + CH2Br*CHz*O*C6H5 = (C 0 0C2H6) 2C (C H2*CHz*OC6Ht,)*C ( CHJ2*CH2*C OOC2H5 + NaBr. But on hydrolysing the product, an acid was obtained, which was not derived from an ethereal salt of this constitution ; this was sub- sequently shown to be diphenoxyethylmaloliic acid, (C6H,*O*CH2CH2)2C( COOH),, and identical with the acid obtained by the action of phenoxyethylic bromide and sodium ethoxide on ethglic malonate (Trans., 1896, 69, 169), a reaction which was, in fact, studied in order to prove the identity of the two acids.It seems probable that during the course of the above reaction the ethylic dimethylpropanetricaboxylate undergoes, i n the first instance, partial decomposition into ethylic malonate and ethylic dimethyl- ncrylate, (COOC2H,)2CH*C( CH,)2*CH,*COOC2H, = (COOC,H,),CH, + C(CH3),:CH*COOC,HS, a kind of decomposition which has been observed in other cases, but which is particularly remarkable in the present instance as being the exact reverse of the process which results .in the formation of the ethereal salt, namely, by the condensatlion OE ethylic malonate with ethylic dimethy lacry late in the presence of sodium ethoxide. The regenerated ethylio malonate then reacts with the phenoxyethylic bromide and sodium ethoxide, with formation of ethylic diphenoxy- ethylmalonate, ( c6H5*o*cn2-CH2)2C( COOC2H5)2.At first sight it would seem more likely that ethylic phenoxy- eth 9 lmalonate, C,H,O* C H2*CH2- CH (C 0 0 C2 H5)2 (this vol., p. 16 7), would be formed in this way, but probably the decomposition of the ethylic dimethylpropanetricarboxylate into ethylic malonate and ethylic dimet'hylacrylate is a gradual one, and thus the ethylic maloriate would always be in the presence of excess of sodium ethoxide and phenoxyethylic bromide. Although these experiments did not seem to indicate that there was much chance of the study of the action of phenoxyethylic bro- mide on the sodium compound of ethylic isopropylpropanetricarb-OF ACRYLIO ACID, AND OF QLUTARIC ACID. 1469 oxylate" giving the desired results, it was nevertheless decided to investigate this point, and the result was certainly unexpected, as the reaction in this case was fonnd to proceed i n a perfectly normal manner, thus.( COOCzH5)aCNa*CH,*CH( COOCzH5)*CH (C H3)2 4- C H:,Br*C H2*0*C6H5 + NaRr- The yield of ethylic phenoxyethy lisopropy lpropanetricarboxylute thus obtained is certainly not good, but there was no indication of i~ decomposition of the molecule similar to that observed i n the case of ethylic dimethylpropanetricarboxylate. On hydrolysing the product of the above reaction, the correspond- i n g tribasic acid is obtained as a beautifully crystalline substance, which melt? at 180°, decomposing a t the same time into carbon dioxide and phenoxyethylisopropylglutaric acid (m. p. 93"). COOH*Q H*CH,-QH*COOH C6H6*O*CH2-CH2 CH(CH& (COOC,H,)zC( CHz*CHz*OC6H,) *CH:,*CH( COOCzH,)*CH( CH,), When digested with hydrobromic acid, this acid yields quantities of phenol, and a new crystalline acid, which is probably bromethyl- isoprop yl gI u taric acid. COOH*$IH*CH,*qH*COOH CEzBr*CHz CH(CH3), This acid and the corresponding hydroxy-acid, which is formed by boiling the bromo-acid with sodium carbonate solution, appear to have very interesting properties, and will be made the snbjectl of a further investigation, the results of which I hope to lay before the Society in a short time.Some of the experimeuts described ilr this paper were carried out with the assistance of Dr. J. F. Thorpe and Mr. W. Goodwin; whew this has been the case I have stated the fact at the commencement of the section in question.I beg to thank these gentlemen, and also Messrs. W. H. Bentley, E. Haworth, and J. L. Heinke, for their valuable assistance during the whole course of this investigation. E x P E R I M P N T A L. j?,3-l)imethylacrylic acid, C(CH3),:CH.COOH. [With W. G o o ~ w r ~ . ] This acid has already been prepared by Miller (Annulen, 1881, 206, 261) by the elimination of water from /?-hydroxyisobntylformic Q This was of CouPee suppojed at the time to bs ethylic trimethylpropanetricarb- ox J late.1470 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID. acid, ( CH3)2C(OH)*CH2*COOH, by boiling it with dilute snlphuric acid, and by Duvillier ( A m . Clzim. Phys., [ 5 ] , 19, 428), by the action of sodium ethoxide on ethylic a-bromovalerate.Subsequently Weinig (Ai:nale?h, 1894, 280, 253) showed that a good yield of ethylic dimethylacrylate may be obtained by boiling ethylic a-bromovalerate with diet hylaniline. In our experiments, we used one of the two following processes, according as we required the free acid or its ethereal salt. 1. The hydrolysis OC ethplic a-bromovalerate with alcoholic potash. 2. The removal of hydrogen bromide from ethylic a-bromovalerate by means of quinoline. lMethod I. Prqaration of Dimethylacrylic acid.-Potash (100 grams) is dissolved in a small quantity of boiling 80 per cent. alcohol, the Eolution heated on a water bath in a flask connected with a reflux condenser, and then pure ethylic a-bromoisovalerate (100 grams) added as rapidly as possible through the condenser, As soon as the very rigorous action has siibsided, the heating is continued for about half an hour ; water is then added, the clear brownish solution evaporated unlil free from- alcohol, acidified, and extracted three times with.pure ether.The ethereal extract is dried over calcium chloride, the ether dis- tilled off, and the residue fractionated under the ordinary pressure, when nearly the whole passes over between 185' and 195' as a colour- less oil which, on standing, becomes completely filled with long, colourless crystals. These were freed from adhering oil by means of the pump, and purified by recrystallisation from light petroleum (b. p. 50-60') ; the oily mother liquor from the crystals, on careful fractionation, yields more of the crystalline acid ; ultimately a con- siderable quantity of aa unpleasant smelling, oily liquid is left which does not crystallise at ordinary temperatures, and the nature of which we have not investigated.Dimethylacrylic acid crystallises from light petroleum in long, colourless needles which melt at 69", and are readily soluble in alcohol, ether, and hot, light petroleum, sparingly soluble in water. The analpis of the acid gave the following results. C = 59.97 ; H = 8.02. I. 0.1439 gave 0.3164 C02 and 0.1040 H20. Ir. 0.1600 ,, 0.3495 ,, ,, 0.1157 ,, c' = 59.56; H = 8.03. C,H802 requires C: = 60.00 ; H = 8 0 per cent. Preparation of Ethylic DirnethylacryZate.-In the pre- paratiori of this ethereal salt, ethylic a-bromisovalerate wag digested with quinoline instead of with diethylaniline as proposed by Weinig (Anwalen, 1894, 280, 253), because i t was found that the former reacted more vigorously than the latter, and yielded a product con- Nethod II.OF ACRYLlC ACID, AND OF GLUTARTC ACID. 1471 +aining only traces of bromine ; the whole method requiring much less time than when diethylaniline is employed.Pnre ethylic a-bromisovalerate, in quantities of 50 grams, is heated i n a reflux apparatus with freshly distilled cod tar quinoline (75-100 grams), a thermometer being placed in the Iiquid to allow of the temperature being observed. At about 170-175', and as soon as the reaction sets in, the flame is removed ; the whole boi!s gently for some time, the temperature rising spontaneously to 190'. As soon as the .action has subsided, the liquid is heated at 185-190' for 10 minntes, the dark brown product poured into excess of dilute hydrochloric acid, extracted with ether, the ethereal solution washed with hydro- chloric acid, dried over calcium chloride, and the ether distilled off.If the ethereal salt is required free from bromine, it is again heated with about hslf its weight of quinoline in the same manner as before, and ultimately carefully fractionated ; after three fractiona- tions, almost the whole passes over between 154' and 155' as a colonr- less oil of penetrating odour. An analysis gave the following results. 0.1530 gave 0.3660 GO, and 0.1341 H20. When digested with excess of potash in methyl alcoholicz solution, ethylic dimethylacrylate is completely hydrolysed in less than one hour; it is not necessary to digest for 12-14 hours, as Weinig (Zoc.cit., p. 254) states. In order to isolate the dimethylacrylic acid, the product of the hydrolysis is evaporated with water until free from alcohol, acidified, extracted with ether, and purified as described under Method 1. The ethglic dimethylacrylate used in this research was prepared partly by the action of quinoline on ethylic a-bromisovalerste, as described above, and partly by the etherification of pure dimethyl- acrylic acid. I n the latter case, the pure acid (50 grams) was dissolved in absolute alcohol (100 grams), concentrated sulphuric acid (50 grams) added, and the whole allowed to staud over night. Water was then added, the ethereal salt extracted with ether, the ethereal solution washed with water and dilute sodium carbonate solution, dried over anhydrous potassium carbonate, and the ether distilled off.The residual ekhereal salt, on fractionation, distilled almost constantly at 254-155', but it did not give very good results on analysis. C = 65.24; H = 9-43. C7HI2OZ requires C = 65.62; H = 9.37 per cent. Found. I--- 7 Theory. I. 11. 111. CiHlZOZ. 33 ... .. .. . . 9.17 9.21 9.22 ,, 9.37 ,, C.. ........ 65.01 64.91 64-75 p. c. 65.62 p. c.1472 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, Ethylic L)imethylpropanetricarboxylate, ( C 0 0 C2H5) 2CH* C (C Ha) .L*C €I2* C 0 OC2H5. I n order to prepare this substance, pure ethylic dirnethylacrylat,e was digested i n alcoholic solution with excess of the sodium derivative of ethylic malonate, the quantities usually employed being the following.Ethylic malonate.. ........ 160 grams Sodium ,, .......... 23 ,, Alcohol ,, .......... 250 ,, Ethylic dimethylacrylate , . , 80 ,, [ Wi tli W. GOODW IN. 1 .The sodium was dissolved in the alcohol, the slightly warm soln- tion mixed with the ethylic malonste, the ethylic dimethylacrylate then added, and the whole heated in a reflus apparatus (or, in some cases, in soda-water bottles) for eight hours at 60', and then for eight hours on ft water bath. The opaque, slightly brownish pro- duct was mixed with water and dilute hydrochloric acid, extracted twice with ether, the ethereal solution well washed with water and dilute hydrochloric acid, dried over calcium chloride, evaporated, and the residual oil purified by fractionation under reduced pressure (60 mm.).After a considerable quantity of unchanged ethylic dimethylacrylate and ethylic malonate had passed over, the thermometer rose rapidly to 195O, almost the whole of the residue distilling between this tem- perature and 210'. Pure ethylic dimethylpropanetricarboxylate distils at 203' (60 mm.). Auwers (Zoc. cit., p. 113) found the boiling point to he 194' (43 mm.) ; he does not appear, however, to have analysed his product. The analyses of the ethereal salt prepared by us gave the following results. C = 58-00 ; H = 8-30, 11. 0.158 ,, 03375 ,, ,, 0.1183 ,, C = 58.25; H = 8-32. C,H,,Oe requires C = 58.33 ; H = 8-33 per cent. Ethylic dimethylpropanetricarboxylate is a moderately thick, colourless oil, which, when heated in small qnsntities, distils under the ordinary pressure almost without decomposition.The yield of this ethereal salt, obtained under the above conditions, Taried considerably in different experiments, amounting, as it did in more than one case, to over 40 per cent. of theory, whilst in one or two cases the yield was as low as 21 per cent. Auwers, in his experi- ments on the condensation of ethylic dimethylacrylate with etliylic malonate in alcoholic solution, obtained only 8 per cent. of the theoretical yield of this ethereal salt ; bst he worked under different eonditions from those we employed. I. 0.132 gave 0-2809 CO, and 0.0986 H,O.OF ACRYLIC ACID, AND OF QLUTARIC ACID. 1473 Dimethylpropanet?-carbo~yZ~c acid, (C OOH),CH*C (CH,),*CH,*COOH. In order to prepare this acid, 20 grams of pure ethylic dimethyl- propane t ricarbox ylate, in me thy1 alcoholic solution, was digested with excess of potash (20 grams) for two hoursin a reflux apparatus. The solution was then mixed with water, concentrated on a water bath, boiled until quite free from alcohol, cooled well, acidified, and extracted 20 times with ether.The ethereal solution, after drying over calcium chloride, was evaporated, and the residue allowed to stand over sulphuric acid in a vacuum desiccator until it had almost completely solidified. After standing for six days in contact with porous porcelain, the crystals, which were colourless and free from oil, were dissolved in a little water, and the solution filtered and saturated with hydrogen chloride, the whole being well cooled during the operation. The acid soon commenced t o separate in sandy crystals, which, after 24 hours, were collected, drained on a porous plake, dried at looo, aud analysed with the following results.0.1501 gave 0,2585 COa and 0.0815 H,O. C = 46.99; H = 6.04. Dimethylpropanetricarbooylic acid softens at 168", and decomposes a t 173' with rapid evolution of carbon dioxide. It is readily soluble in water and alcohol, much less SO in ether, and only very sparingly in hydrochloric acid, so that when its concentrated aqueous solution is saturated with hydrogen chloride the acid separates almost completely. The pure substance is very stable, and may be heated at 100' for a considerable time without appreciable decomposition. Salts.-A neutral solution of the ammonium saltpf the acid gives no precipitate with coppey sdphate or with barium or calcium chlovides ; but on the addition of lead acetate, a voluminous gelatin- ous precipitate is produced. The silver salt, CeHsOsAg3, was prepared by adding silver nitrate to a slightly alkaline solution of the ammonium salt ; it is a white, curdy precipitate, insoluble in water.After well washing and drying, first over sulphuric acid and then at loo', the following results were obtained on analysis. C8H,,0s requires C = 47.06 ; H = 5.88 per cent. I. 0-1809 gave on ignition 0*1108 Ag. Ag = 61-26. 11. 0-2159 ,, ,, 0.1325 ,, Ag = 61.37. C8HgOsAg3 requires Ag = 61-71 per cent. pp-Dimethylglutaric acid, COOH*CH,*C(CH,),*CH,*COOH. [With W. GOODWIN.] This acid is formed when dime thylpropnnetricarboxylic acid is VOL.LXIX. 5 G1474 PERKIN : SOME DERJVATIVES OF PROPIONIC ACID, heated above its melting point,. In carrying out this decomposition quantitatively, the following belinvionr was observed. 7.1825 grams of the tribasic acid heated in a wide test tube at 185-190' for 3-4 minutes lost 1.6070 gram = 22.4 per cent., but on continuing the heating for half an hour, the loss was 2.92 gram or 26.7 per cent. Theoretically, for the elimination of one molecule of carbon dioxide, the loss should he! 21.5 per cent. ; this agrees with the results obtained on heating the acid for a short time at 185- 190'. The further loss on prolonged heating is due to anhydride formation, and, to a small extent, to sublimation. The residue from this experiment solidified completely on cooling.It was powdered, dissolved in water, and the solntion saturated with hydrogen chloride. On standing, beautiful, colourless crystals sepa- rated, which, after drying at looo, gave the following results on analysis. C = 52.42 ; H = 7.72. PI. 0.1538 ,, 0.2962 ,, ,, 0.1071 ,, C = 52.53 : H = 7.73. I. 0.1694 gave 0.3256 C02 and 0.1177 H20. C7HI2Oc requires C = 52.50. H = 7.50 per cent. ,pp- Diinethylg lutaric acid is a colourless, crystalline substance which melts at 10Lo. It is very readily soluble in water, alcohol, and ether, but only sparingly in hydrochloric acid, benzene, and light petroleum ; i t crystsllises well froni water in colourless needles, but is most readily and economically obtained pure by recrystallisation from hot hydro- chloric acid, as described above, as in this way very little remains in the mother liquor.S i l v e ~ Salt, C7H1,Ag2O4.--This salt mas prepared by adding silver nitrate to the slightly alk a 1' me solution of the ammoninm salt. It is a white, insoluble precipitate, which, after well washing with water and drying a t looo, gave the following results on analysis. C = 22-38 ; Salts of pp-Di?nethylgZzlt~.ric acid. I. 0.1608 gave 0.1320 GO2, 0.0403 H20, and 0.093 Ag. H = 2.79; Ag = 57.83. 11. 0.201 on ignition gave 0.1159 Ag. C,H,,Ag2O4 requires C = 22.46 ; H = 2.66; Ag = 57.76 per cent. Ag = 57.66 A neixtral solution of the ammonium salt shows the following behaviour with reagents. Calcium cliloi-ide and Barium chloride give no precipitate even on boiling.Copper acetate gives no precipitate in the cold, but, on boiling, a light blue, apparently crystalline precipi- tate separates. Lead acetate also gives no precipitate in the cold, but, on warming, the solution rapidly becomes turbid, and an amorphous lead salt separates.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1475 Ethylic Dimethy Zglutnrate, cOOc,H,.Ca,*C(GH,>,*CH~*~OOC,H~. This was prepared by dissolving the pure acid i n absolute alcohol, adding concentrated sulphuric acid (+ vol.), arid allowing the mixture t o staud for two clays ; it was then heated in the water bath for two hours, mixed with water, and the oily ethereal salt extracted with ether. The ethereal extract was washed well with dilute sodium carbonate, dried over anhydrous potassium carbonate, evaporated, and the resid ual oil purified by distillation under reduced pressure, when almost the whole quantity passed over a t 170-172" (100 mm.).C = 60.89 ; I€ = 9.28. 11. 0.1407 ,, 0.3126 ,, ,, 0.1184 ,, C = 60.59 ; H = 9.35. C,,H,,04 requires C = 61.11 ; H = 9.26 per cent. Ethylic dirriethylglutarate distils, apparently without decomposition, at 241-243' (755 mm.), or only slightly higher than efhylic glutarate, which boils at 237" ; it has a faint odour somewhat resembling that of ethylic succinate. I. 0.2011 gave 0.4488 CO, and 0.1680 H,O. Dimethy Zg Zutar ic An?) y dr ide, C (C H3) < CHZ*CO CH,. > 0. [With W. Goouwr~.] Dimethylglutaric acid dissolves readily in hot acetic anhydride, and if the solution is heated in a test-tube in a sulphuric acid bath until the excess of acetic anhydride has distilled oE, the residue solidifies on cooling to a hard, crystalline cake.The crystals were spread on porous porcelain until free from mother liquor, and dried at 100' ; they then melted a t 124', and, after recrystallisation from acetic anhydride, a t 124-125", C = 58.72 ; H = 7-20. 11. 0.1249 ,, 0.2674 ,, 0.0803 ,, c' = 58.79; H = 7.19. C~H1003 requires C = 59.16; H = 7-04 per cent. Dimethylglutaric a&$ride cry s t a k e s from acetic anhydride, in which i t is very soluble, in thin plates; these, aftor drying well between filter paper, smell of acetic anhydride, and become quite sticky in the water oven, and remain so for at least an hour, so that it is possible that these crystals consist of a mixed anhydride of acetic and dimetbylglutaric acids, which, on heating, decompose into acetic anhydride and dimethylglutaric anhydride ; this point has, however, not been further investigated.Dimethylglutaric anhydride is readily soluble in benzene, but only sparingly in light petroleum (b. p. 40-50'), it dissolves, however, in petroleum boiling a t 110-120°, and on cooling ci.ystallises out in plates. It is insoluble in cold water, b u t on heating it melts under water, and rapidly dissolves. 5 ~ 2 I. 0.1376 gave 0.2960 CO, and 0.0892 H,O.1476 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, Dimethy7gZutaraniZic acid, C 00H*CH2*C (CH,),*C H,*U O*N EI*C6H5, and Dintet77yZglutaranZ1, C(CB,),<cHf.co>N*C,H,. CH *CO If a solution of dimethylglutaric anhydride i n a little benzene is mixed with the calculated quantity of freshly distilled aniline, the mixture gets hot and rapidly begins to deposit crystals ; these, after being collechd and drained on porous porcelain, are crptailised two or three times from dilute alcohol.0.1461 gave 7.7 C.C. moist nitrogen at 17' and 7.56 mm. N = 6.12. Di~ethylglzctarnnilic acid melts at about 134O, but if slowly heated it softens at about 128' and melts at 131'. It is very readily soluble in alcohol and chloroform, but only sparingly in water, benzene, and light petroleum; it crystallises from dilute alcohol in small, hard, glistening prisms, and very beautifully, in similar form, when its solution in chloroform is allowed to evaporate spontaneously. DimethyZgZutarnniZ is obtained by heating the anilic acid to boiling for 10 minutes, in a test-tube; water is given off, and, on cooling, the residue solidifies completely.After recrystallising t mice from dilute alcohol, the anil is obtained pure in colourless, glistening plates, which melt at 156-157'. 0.1380 gave 7.8 C.C. moist nitrogen at 18" and 760 mm. N = 6.52. C,,H,JO, requires N = 6.45 per cent. Dimethylglutaranil is readily soluble in alcohol, benzene, and CL3H17N03 requires N = 5.96 per cent. chloroform, but only sparingly in water or light petroleum. a/3/3- Trimethylpropionic acid, C H (C H3),*C H (CH,) *C 0 0 H. This acid has already been prepared by Markownikoff (Zeit. f. C H (C H3) 2*C H ( C HJ C N,, by hydrolysis, and also by Romburgh (Rec. Trav. Chinz., 1886,5, 231) by the hydrolysis of ethylic methyl isopropylacetoacetate, and by the distillation of methylisopropylmalonic acid, Chem., 1866, 502) from methylisopropylcarbinyl cyanide, CH3.C 0 *C ( CH3) (CSHT) *C 0 0 CZHs, C ( C H3) ( C3H7) ( C 0 0 H) 2- In preparing large quantities of this acid, I have used exclusively the latter method: as, however, the isolation of the pure acid was found t o be a matter of considerable difficulty, I give the method of preparation in full.In the first place, ethylic methylmalonate was prepared by the action of methylic iodide on the sodium derivative ofOF ACRYLIC ACID, AND OF GLUTARIC ACID. 1477 ethylic malonates in alcoholic solution, great care being taken by using a slight excess (about 5 per cent'.) of the calculated quantity of sodium, and a considerable excess of methylic iodide, that the product should be as free from ethylic malonate as possible. I n subseqiiently introducing the isopropyl group into the carefully fractioned ethylic methylmalonate, the following method was found convenient.Ethylic methylmalonnte (345 grams) is mixed with sodium (46 grams) dissolved in alcohol (500 grams) in a large flask con- nected with a very long reflux condenser ; isopropylic bromide (260 grams) is then added, and the whole very gently warmed, the action being kept well under control, otherwise the whole begins to boil very vigorously, and much isopropylic bromide escapes through the condenser. As soon as the principal reaction is ovm, the mass is heated to boiling for six hours, the bulk of the alcohol distilled off, water added, and the oil which separates extracted three times with ether; the ethereal solution, after washing well with water and eraporating, deposits a nearly colonrless oil, from which, by repeated fractiona- tion, 330 grams, or about 75 per cent.of the theoretical yield of nearly pure ethylic methylisopropylrnalonate, boiling at 21 7-222', can be obtained. 0.1451 gave 0.3190 CO, and 0.1162 H,O. C,,H,,O, requires C = 61.11; H = 9.26 per cent. This ethereal salt was hydrolysed by digesting it for two hours with potash (260 grams) dissolved in purified methylated spirit ; water was then added, and the product heated first on a water bath and finally boiled over a free flame till quite free from alcohol ; the concentrated solution was then acidified and extracted six times with pure ether.The ethereal solution, after drying over calcium chloride and evaporating, deposited about 280 grams of an oily acid, which still contained some ether, and, on standing, rapidly began t o crystal- lise. This crude product was now heated at 180-200' until the evolution of carbon dioxide had ceased, and the residue distilled with- out, in the first imtance, attempting to fractionate. The crude distillate always contains traces of propionic and iso- valeric acids, and it is remarkable that pure methylisopropylacetic acid can only be isolated from this with very considerable difficulty, Q I n carrying out thie prepamtion with large quantities of material, the rnethylic iodide must be added very cautiously in small quantities to the alcoholic solution of the sodium derivative of ethylic malonate, the whole being well cooled during the operation, otherwise the action is very violent, and if care be not taken, the liquid boils most vigorously, with loss of methylic iodide from volatilisation, and frequently from frothing over.C = 60.80; H = 9.03.1478 PERKIN : SOXE DERIVATIVES OF PROPIONIC ACIII, repeated fractionation from a flask fitted with a long colonna being necessary, in order to effect as complete a separation a s possible. Ultimately, however, about 180 grams of an oil boiling constantly at 189-190' were obtained, which consisted of nearly pure methyl- isopropylacetic acid, as the following analyses show. I. 0.1606 gave 0.3622 CO, and 0.3.440 H,O. C = 61-50 ; H = 9.96. I I .0.1314 ,, 0.2952 ,, ,, 0.1220 ,, C: = 61.27; H = 10.34. C,H120, requires C = 62-07 ; H = 10.34 per cent. Tq-imethyZpmpionic acid is a most disagreeably smelling mobile Rornbzcrgh (Zoc. cit.) found prac- liquid, which boils at 189-190'. tically the same boiling point, namely, 189-191 '. Ethy Zic a-Bromotrimethy~~oloio?Lute, C H (CH&* C Br (C H,)*C 0 0 C,H,. The bromination of methylisopropylacetic acid is very readily carried out by treating the acid bromide with bromine according to Volhard's method (Anizalen, 1887, 242, ISl), and as the result of a number of experiments, the followiiig quantities were found to give the best results. I n a flask, into the neck of which along condensing Cube is ground, methylisopropylacetic acid (50 grams) is mixed with phosphorus pentabromide (75 grams) and allowed to remain for about one houP, until the formation of the acid bromide is complete.Dry bromine (85 grams) is then added in about six portions, the flask being cooled between each addition ; the liquid is then heated gradually on a water bath to about 50-60°, until the action is nearly complete, and finally at 100' for two hours. The well cooied product, which is usually coloured, owing to the presence of a slight excess of bromine, is poured in a thin stream into three times its volume of absolute alcohol, the vigorous action being allowed to proceed without cooling ; after standing overnight, water is added, and the bromo-ethereal salt is extracted with et*her. The ethereal solution, after washing with water and dilute sodium carbon- ate solution, drying over calcium chloride, and evaporating, deposits a somewhat brownish oil, which, on fractionating twice under reduced pressure (100 mm.), boils for the most part constantly at 130', the yield being usual17 good, although, in one or two cases, not inconsiderable quantities of some higher boiling Substance was obtained.The analysis of the pure substance gave the following result. 0.2035 gram substance, heated with fuming nitric acid and silver nitrate at 180' for four hours, gave 0.1722 gram AgBr = 36.0t per cent. of bromine. C,HI5BrO, reqnires Br = 35.89 per cent.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1479 Ethylic a-bromotrinaeth~~r~p~onate is a heavy, very pungent smell- ing, colourless oil, very similar to ethylic a-bromisovalerate in its general properties.Trimeth y Zac~y Zic acid, C (C H,) 2: C (C H3) *C 0 0 H. As stated in the introduction, this acid was first prepared by the hydrolysis of ethylic methylisopropylbromacetate with alcoholic potash. Potash (100 grams) was dissolved i n the smallest possible quantity of SO per cent. alcohol in a flask connected with a long reflux con- denser, the solution heated on a boiling water bath, and hhen the pure bromo-ethereal salt (100 grams) added as rapidly as possible through the condenser tube. After the very vigorous action had subsided, the whole was heated to boiling €or about an hour, water was added, t'he brown solution evaporated until quite free from alcohol, aciditied, and extracted five times with pure ether. The ethereal solution was then dried over calcium chloride, evaporated, and the residue fractionated two or three times under reduced pressure, and subsequently &.the ordinary pressure, in order to separate the trimethylacrylic acid as far as possible from a considerable quantity oE a-hydroxytrimethyl- propionic acid (p.1486), which is formed at the same time. In this way, an oily acid was obtained, boiling constantly at 204-205' (under 100 mm. pressure it boils at 150'), and consisting of a mixture of the two isomeric a.cids, isopropylacrylic acid and tri- methylacrylic acid. It gave the following results on analysis. 0,1641 gave 0.3760 CO, and 0.1320 H,O. C6H,,0, requires C = 65-15; H = 8.77 per cent. During the winter months, this oily acid gradually deposited a quantity of beautiful, prismatic, four sided crystals ; these were col- lected with the aid of the pump, drained on porous porcelain, and recrystallised from l i g h t petroleum. The new substance was thus readily obtained in a pure state, and gave the following numbers on analysis.C = 62-64 ; H = 8.94. 11. 0.1441 ,, 0.3334 ,, ,, 0.1155 ,, C = 63.10; H = 8-90. C(CH3),:C(CH3)COOH requires C.= 63.15; H = 8-77 per cent. TrinaethyZac~ylic acid, when moderately rapidly heated in a capillary tube, softens at 67', and melts at 79-71'. I t is readily soluble in hot light petroleum, benzene, alcohol, ether, and chloroform, sparingly in hot water. The hot concentrated aqueous solution becomes milky on standing, and then gradually deposits crystals, but, on slowly cool- ing a hot dilute aqueous solution, the acid crystallises well in long, colourless needles.C = 62.5; H = 8.93, I. 0.1312 gave 0 3014 CO, and 0.1056 H,O.1480 PERKIN: SOME DERIVATIVES OF PROPIONIC ACID, The solution of the pure acid in sodium carbonate decolorises potassium permanganate, but not nearly so rapidly as many other unsaturated acids. Triinet h3ZacryZic Anilide, C( C H3) z:C ( CHJ *C 0 *NH*C6H5. This substance was prepared in the following way. Pure tii- methylacrylic acid was mixed with excess of freshly distilled phos- phorus trichloride, and, as very little action appeared to take place in the cold, the whole was heated on the water bath for about 15 minutes. On fractionating the product, a considerable quantity distilled a t 145-150', and evidently consisted of trimethylacrylic chloride, C(CH3)2:C(CH3)*COC1.This was dissolved in pure dry ether, mixed with excess of aniline, and, as soon as the very vigorous action had subsided, the whole was treated with water and extracted with ether. The ethereal solution was washed first with dilute hydrochloric acid and then with sodium carbonate, dried over anhydrous potassium carbonate, and the ether distilled off, when a syrupy residue was obtained, which, on standing, gradually solidified, In order t o purify this crude substance, it was ground up with cold light petroleum, collected on a filter, and recrystallised twice from boiling light petroleum (b. p. 70-80"). 0,2133 gave 14.0 C.C. moist nitrogen at 20" and 765 mm. N = 7.55. T~imethylacrylic anilide melts at 93-94', and crystallises from hot light petroleum in beautiful, glistening plates.It is readily soluble in alcohol, ether, and benzene, sparingly in light petroleum, and almost insoluble in water. ClzHl5NO requires N = 7.41 per cent. Dibromotrimethyl33ropionic acid, CBr(CH,),*CBr(CH,)*COOH. A solution of trimethylacrylic acid in chloroform decolorises bro- mine rapidly, and if, after the addition of the calculatled quantity of bromine, the liquid is allowed to evaporate spontaneously, a semi- solid residue is obtained, consisting of impure dibromo trimethyl- propionic acid. The crude product was left in contact with porous porcelain for some days until the oily impurities had been entirely absorbed, and the solid residue purified by recrystallisation from light petroleum (b.p. 100-105'). 0.1547 gram substance, heated with nitric acid and nitrate of silver at 180°, gave 0.2128 gram AgBr ; Br = 58.54. C6H,,,Br20, requires Br = 58.40 per cent. Dibromotrimethy kropionic acid is very sparingly soluble in cold light petroleum, but dissolves fairly readily on warming, and separatesOF ACRYLIC ACID, AND OF GLUTARIC ACID. 1481 again, on standing, as a heavy, white, crystalline powder. When heated in a capillary tube, it sinters at 185', and melts at 190-L9lo wit,h rapid evolution of gas. P- Bromotrimethy ~ r o p i o n i c acid, CBr (CH,),*C H (CH,) *COOH. In order t o prepare this substance, finely powdered trimethylacrylic acid (1 gram) was mixed in a test tube with 5 C.C. of fuming hydro- bromic acid saturated at 0'. On gently warming the mixture, the acid completely dissolved, but very soon an oily layer formed on the surface.After standing for 10 minutes, an equal volume of water was added, and the crystals which separated were collected, well washed with water, drained on a porous plate, and dried over sulphuric acid in a vacuum. As the substance could not be satisfac- torily recrystallised, the colourless, crystalline mass obtained in this way was directly analysed. 0.1431 gram substance, heated at 200' with nitric acid and nitrate of silver, gave 0.1375 gram AgBr. Br = 40.92. CsHl,BrOz requires Br = 41.02 per cent. p-BromotrimetjLylpro~ionic acid, when heated in a capillary tube, softens at 83' and melts at about 87-88'. It is readily soluble in most, organic solvents, but practically insoluble in water.When boiled with water, it is rapidly decomposed with separation of hydrogen bromide. p- Iodotrimeth yZpropionic acid, CI (CH, ),*CH( C H,) GOOH. Trimethylacrylic acid dissolves readily i n f nming hydriodic acid (sp. gr. 1-96> in the cold, and, on standing, the solution gradually deposits beautiful crystals of P-iodotrimethylpropionic acid. After diluting with water, the crystals were collected, washed well with water, and dried on a piece of porous porcelain over sulphuric acid in a vacuum. The analysis of the glistening, crystalline mass thus obtained gave the following results. 0,1439 gram substmce, heated with silver nitrate and nitric acid at 170°, gave 0.1390 gram AgI. I = 52.23. CsH,,IO2 requires I = 52.48 per cent. p-Iodotrimethylpropiolzic acid melts at 80--82O, and, when heated in a test tube, i t rapidly decomposes wifh separation of iodine and evolution of hydrogen iodide; it also decomposes on exposure to strong sunlight, with separation of iodine.It is readily soluble in alcohol, light petroleum, chloroform, and ether, but almost insoluble in cold water. Cold concentrated nitric acid decomposes it instantly, with separation of crystals of iodine.1482 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, Condensation. 19 Aceto?be with Ethylic a-B1.onzopl.opiorante. Pormation of Eth y lic p- H~di.oxy-alj~-trimeth~~p~op~o~~at e, 0H.C (CH3)2*CH(CH,)*COOC,H,. [With J. F. THOKPE.] This condensation was studied, as explained in the introduction, with the object of synthesising trimethylacrylic acid, C ( CH,) ,:C (C H3) *C 0 0 H, and its derivatives by a method which could leave no doubt as t o their constitution, and the results obtained served to prove conclu- sively that the solid acid (melting at 71') produced, as already described, by the hydrolysis of ethylic bromotrimethylpropionate,. was in fact this acid, and not isopropylacq-lic acid, C H (C H3) 2.C (:CH2) C 0 OH, a s was at one time thought to be the case.The first step was to prepare ethylic P-hydroxy-app-trimethylpro- pionate, and this is accomplished as follows. Acetone (90 grams) is mixed with ethylic a-bromopropionate (182 grams), ar?d the mixture poured on to an excess of dry and carefully cleaned zinc in a moderately large flask connected with a reffux a,ppa- ratus. The zinc, which should be in the form of turnings, must be quite free from oil, and it is also essential that it should be quite dry.We found it necessary to wash it several times with hot caustic soda, then with dilute acid to remove any oxide, and finally to wash i t well with water and dry it with alcohol and ether. Experi- ments conducted with zinc which had not been treated in this way were always unsuccessful. The flask containing a little of the zinc in contact with the mixture is now gently warmed on the water bath, when, as boon as the temperature has risen to about 50°, the action usually commences and continues very energetically ; the flask is removed from the water bath if necessary, but as a rule the excess of acetone employed serves to prevent the temperature from rising too high.When the reaction has subsided, if i t is found that nearly all the zinc has disappeared, more is added, and the whole ultimately heated on the water bath for about 2-3 hours ; a t the end of this time the contents of the flask will have become quite thick and slightly colonred. The syrupy liquid is poured off as far as possible from the un- changed zinc, and mixed with water ; this causes the mass to become nearly solid, owing to the separation of a white zinc compound, which, however, disappears on adding dilute sulphuric acid and shaking vigorously. The oil which separates is extracted by means of ether,OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1483 the ethereal solution is washed three times with dilute sulphuric acid in order to remove the last traces of zinc, dried over calcium chloride, the ether distilled off, and the residual oil purified by careful fractionation under reduced pressure (30 mm.).A considerable portion of the product distils below looo, and appears tolconsist essentially of ethylic propionate and ethylic acry- late, the former produced by the reduetion of, and the latter by t,he elimination of hydiaogen bromide from, the et hylic a-bromopropiona te employed in this synthesis. About 50 per cent. of the whole distils at 105' (30 mm.) as a moderately thick, colourless oil of very faint odour, which after redistillation gave the follomiiig I-esults on analysis. C = 60.16; H = 10.08, 11. 0.1309 ?, 0.2872 ,, ,, 0.1198 ,, C = 59.82; H = 10.16. C,H,,O, requires C = 60-00; H = 10 00 per cent. This substance consists, therefore, of pure ethylic P-Jydroxy-a/3/3-tri- 1.0.1186 gtve 0.2616 C02 and 0.1076 H20. meth ylpropionate. /3-Hydrox y-app- t ~ i m ei hy Zpropionic acid, 0H.C ( C H3) 2*C H (CH,) * C 0 0 H . [With J. F. THORPE.] I n order to prepare this acid, the corresponding ethereal salt just described was mixed with excess of a solution of potash (14 mols.) in pure alcohol, and heated on the water bath for two hours. Water was then added, the solution evaporated until quite free from alcohol, acidified, and extracted several times with pure ether, the extraction being much facilitated by saturating the solution with ammonium snlphate. The ethereal solution was carefully dried over calcium chloride, filtered, evaporated, and the thick, syrupy residue purified by distillation in small quantities under reduced pressure ; almost the whole passed over at 160' (35 mm.) as a very thick, viscid, colour- less oil which, on analysis, gave the following numbers.I. 0.1888 gave 0.3814 CO, and 0.1592 H20. C = 55.08 ; H = 9.37. If. 0.1318 ,, 0.2650 ,, ,, 0.1106 ,, C = 54.83; H = 9.32. C,H,,O, requires C = 54.54 ; H = 9.09 per cent. /J-Hyd?.ozy-aPS-trimethylpi-opionicc acid is a very thick syrup, which even on long standing does not show any signs of crystallising. I t is readily soluble in water, the solution having a strongly acid reac- tion. When distilled in large quantities (100 grams), it decomposes with elimination of water and formation of trimethylacrylic acid. The dissociation constant for the electrical conductivity of this acid at different concentrations was determined by Dr.Ewan, who found K = 0.00333.1484 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, /3- Bromo-apB- t rim eth y Zpropionic acid, CBr ( CHJ 2 * C H ( CH,) C 0 OH. [With J. P. THORPE.] This substance, which is identical with the bromo-acid obtained by the addition of hydrogen bromide to trimethglacrylic acid (p. l48l), is obtained by dissolving /3-hyclroxytrimethylpropionic acid in fuming hydrobromic acid (saturated at -10'). The mixture becomes warm, and, on standing, the bromo-acid rapidly separates in lustrous plates, which, after washing with water and drying on a porous plate oyer sulphuric acid in a vacuum, melt at 86-88'. 0.2012 gram substance, heated at 180' with silver nitrate and nitric acid, gave 0.1919 gram AgBr.Br = $0.98 per cent, C,H,,BrO2 requires Br = 41.02 per cent. E th y 1 ic /I- bromo-a Pp-t rimeth y lpropionaie, C(CH3)2Br*CH(CH3)*C:OOC2H5.-[With J. F. THORPE.] In order t o prepare this compound for some synthetical experi- ments described below, the etcherification of the corresponding acid by means of alcohol and sulphuric acid, and by saturating the solu- tion of the acid in alcohol with hydrogen chloride was tried, but with very unsatisfactory results. Ultimately, however, the ethereal salt was prepared in considerable quantity, with ease, i n the following way. The pure dry bromo-acid (10 grams) was mixed with phosphorus pentachloride (12 grams) and allowed t o stand until the somewhat energetic action had subsided: the whole was t,hen heated on the water bath for 15-20 minutes, and the product poured into absolute alcohol and allowed to stand over night.On adding water, a heavy, brownish oil was precipitated; this was extracted with ether, the ethereal solution well washed with water and dilute sodium carbonate, dried over calcium chloride, and the ether distilled off. The residual, somewhat brownish oil did not appear to distil without decomposi- tion ; but after standing over sulphuric acid in a vacuum for a few days, i t gave the following results on analysis, showing it to consist of the nearly pure ethereal salt. 0.2103 gave 0.1760 AgBr. Br = 35*96. C,H1502Br requires Br = 36.16 per cent. Many experiments with this substance mere made with the object of synthesising aPP-trimethglglutaric acid so as to be able to compare it with the acid Balbiano obtained by the oxidationof camphoric acid with cold potassium permanganate (Ber., 1895, 28, 1507) ; the method of procedure will be well understood fi-om the following equations.OF ACRYLIC ACID, AND OF GLUTARIC ACID.1485 I. (COO C2H,),CHNa + CBr (C H,) ,*C H (C H,)*COOC,H, = (CO OC,H,),CH*C (CH,),*CH (CH,) *CO OC,H, + NaBr. rr. (COOH),CH~~:(CH,),~CH(CH,)~COOH = C 0 OH.CH2.C (CH3)z.C H (CH,)*C OOH + C02. The action of the brom-ethereal salt on the sodium derivative of ethylic malonate was tried under various conditions in alcoholic and xylene solution, but only very small quantities of a high boiling con- densation product were formed in any case, and similar results were obtained when ettylic cyanoacetate was employed.Apparently in all these experimeuts the bromo-ethereal salt is to some extent decom- posed with elimination of hydrogen bromide and formation of ethylic trimethylacrylate ; the reaction, however, is evidently a complicated one, as is shown by the fact that in all the above experiments the pro- duct was found to contain considerable quantities of ethylic ethane- tetracarboxyl ate (COO C2H5),CH.C H (COOC,H,) 2. p-Iodo-app-trilnethyZlrro~'ionic acid, C ( CH,),I*CH(CH3)-COOH. /3-Hydroxytrimethylpropionic acid dissolves readily in fuming hydriodic acid, and, on standing, crystals of the above substance separate rapidly. The crystals were collected, washed with water, dried on a porous plate over sulphuric acid in the dark, and analysed.0.1296 gave 0.1252 AgI. I = 52.39. C6H,,T-O2 requires I = 52.48 per cent. This acid melts at 80' and is identical with the compound obtained by the addition of hydrogen iodide to trimethylacrylic acid (p. 1481). When it is treated with phosphorous pentachloride, and the product is poured into alcohol under exactly the same conditions as described in the case of the corresponding bromo-acid, it yields a heavy ethereal salt, which cannot be distilled without decomposition; it was not analysed. Trimethylucrytic acid, C(CH3),:C (CH,)*COOH.-[With J. F. THORPE.] This a3d is produced when alcoholic potash, or weak alkalis, such as the sodium derivative of ethylic malonate, act on ethylic P-bromo- trimethylpropionate. It was first obtained from this ethereal salt, in the following manner.The low boiling fractions from the product of the action of the bromo-ethereal salt on the sodium derivative of ethylic malonate (see above) were hydrolysed by boiling for three hours with an excess of alcoholic potash. The alkaline solution was then evaporated with water until free from alcohol, acidified, and extracted with ether ;1486 PERKIN : SOME DERIVATIVES O F PROPIONIC ACID, the ethereal solution after being dried over calcium chloride was evaporated, and the residual oil fractionated under reduced pressnre. The fraction boiling at 135-137' (50 mm.) solidified on cooling, and the crystals, after being drained on a porous plate and crystallised from water, consisted of pure triniethylacrylict acid melting at 70-71'. 0.2201 gave 0.5085 CO, and 0.1737 H,O.C6H,,O2 requires C = 63.15 ; H = 8.77 per cent. The silver salt, C6H9Ag02, was obtained as a white precipitate sparingly soluble in water, on adding silver nitrate t o the nentral solution of the ammonium salt. C = 63.00 ; H = 8.79. 02002, on ignition, gave 0.0970 dg. C6HgAg02 requires A g = 48.77 per cent. After these experiments had been carried out, it was found that the same acid was obtained by the hydrolysis of ethylic p-bromo- trimethylpropionate with alcoholic potash ; and the acid obtained in both cases was found to be identical with the trimethyacrylic acid obtained by the hydrolysis of ethylic a-bromotrimethylpro- pionate (p. 1479). As this was an important point, a sample of the tri- methylacrylic acid obtained from the /%brom-ether.eal salt by means of alcoholic potash was dissolved in chloroform and treated with bromine under the conditions described on p.1480. The resulting dibromide melted at 190-191", and was identical in all respects with the dibromotrimethylpropionic acid previously obtained. Br = 58.10. C6H,,Br0, requires Br = 58.40 per cent. Ag = 48.73. 0.1398 gave 0,191 AgBr. ol-Rydroxytr~methylpropionic acid, CH(CH,),*C (OH) (CH,)*COOH. I n distilling t'he crude acid obtained by the hydrolysis of ethylic a-bromotrimethylpropionate with alcoholic potash, a mixture of isopropylacrylic acid and trimethylacrylic acid at first passed over as explained on p. 1479 ; the thermometer then rose rapidly, and a thick colourless oil distilled over which after standing for some months deposited prismatic crystals.These were collected, drained on a porous plate, and purified by recrystallisation from light petroleum (b. p. 60-80'), when they were obtained in the form of glistening, colourless needles which, on analysis, gave the folio wing numbers. C = 54.73 ; H = 9.20. 11. 0.1601 ,, 0.3225 ,) ,, 0.1305 ,, C = 54.90; H = 9.06. 111. 0.1765 ,, 0.3511 ), ,, 0.1451 ,) C = 54.25; H = 9 12. C6H,,03 requires = 54.54 ; H = 9-09 per cent. I. 0.1931 gave 0.3874 CO, and 0.1598 H,O.OF ACRYLIC ACID: AND OF GLUTARIC ACID. 1487 cr-Hydrozytri~,2efhy~,.opio?ii~ acid melts a t 75-77', but not very sharply. It is very readily soluble in water, the solution possessing a very strong acid reaction. It is readily soluble in alcohol, ether, OY .chloroform, moderately in benzene, and sparingly in cold, light petroleum.When thrown on to the surface of water, the crystals of the acid rotate in a very vigorous manner and dissolve rapidly. The dissociation coilstant for the electrical condnctivitly of this acid a t different concentrations a t 2 5 O , was found b y Dr. Ewan to be K = 0.01135. It is remarkable that this value should be so high in comparison with the constant (K = 0.00333) for the corresponding p-hydroxy-acid (p. 1483). Salts of a- Hydrox ytrinaethy Zpropionic acid. The silver salt, C,H,,AgO,, obtained as a white precipitate on t h e addition of silver nitrate to a strong neutral solution of the ammonium salt of the acid, is sparingly soluble in cold water, but dissolves readily on boiling, and the hot solution on cooling deposits the salt in a beautifully crystalline condition.f Ag = 45.19. 0.2445 gave 0.1105 Ag. CsH,,Ag03 requires Ag = 45.18 per cent. This silver salt is not Feadily acted on by light, and may be dried a t 100' without decomposition ; its solution in dilute ammouia is only very slowly decomposed on boiling. A neutral solution of the ammonium salt of the acid shows the following behaviour with reagents. Copper acetate no precipitate in the cold, but, ,on boiling, a pale blue crystalline copper salt sepal-ates ; this gradually decomposes on prolonged boiling, especially if a little alkali be present, and cuprous oxide is precipitated. Barium and calcium chlorides, and lead acetate give no precipitate. Action of Hydrobroinic acid and Hydriodic acid O?L a-Hydroaytri- methyZpp.1.opionic acid.-This reaction was investigated in the hope of easily obtaining a-bron~otrimethylpropionic acid, C H (CH3) 2.C BI*( CH,) *C OOH, and a-iodotrimethylpropionic acid, CH(CH,),*CI(CH,)*COOH, for comparison with the cor~esponding @halogen derivatives which had been prepared from a-hydroxytrimethylpropionic acid and trimethyl- acrylic acid by the methods described on pp.1484-1485. It was found that p-hydroxytrimethylpropionic acid reacts very readily with con- centrated hydrobromic and hydriodic acids in the cold, formiiig the corresponding 8- halogen derivatives, but the behaviour of the tori-e- sponding a-hydroxy-acid is very different in this respect. The latter dissolves very readily in the concentrated halo'id acids, but no action1488 PERKIN : SOME DERIVATZVES OF PROPIONIC ACID, appears to take place, as was shown by the fact that, after remaining a t the ordinary temperature for 12 hours, no crystals had separated, and a drop of the liqiiid dissolved in water formed a clear solution.The solutions were then sealed up in tubes and heated in boiling water, but even then the action seemed to take place very slowly, the liquids only becomiiig turbid after two hours' boiling. After eight hours, the tubes were allowed to cool, when, on opening them, a considerable pressure was noticed, which was found to be due to carbon dioxide, and, on pouring the contents of the tubes into water, a heavy oil was precipitated in each case. This was extracted with ether, the ethereal solution washed well with water and dilate sodium carbonate (and, in the case of the hydriodic acid experiment, also with sodium thiosulphate to remove iodine), evaporated, and the nearly colourless, oily residues fractionated.After twice fractionating, in the case of the hydro- bromic acid experiment a small quantity of a heavy oil was obtained, which boiled at 115-120', and on analysis gave the following result. 0.1420 gram heated with silver nitrate and fuming nitric acid at Br = 51.10 per cent. C5HllBr requires Br = 58.98 per cent, Although, owing to the difficulty of purifying the small amount of material at my disposal, this analysis does not agree well with the formula C5HIIBr, i t seems probable that this bromide is identical with methylisopropylcarbinyl bromide, CH(CH,),*CHBi**CH,, which Wischnegradsky (Annalen, 1878, 190, 357) obtained by the action of hydrogen bromide on isopropylethylene, C H(CH,),*CH:CH,, and which boils at 114-116'.The formation of this substance would be readily explained by the following equation : CH(C€33)2*C(OH)(CHs)*COOH + HBr = 200' for 4 honrs gave 0.1711 grain AgBr. CH(CH,),*CHBr*CH, + C02 + H,O, but it is certainly remarkable that an a-bromo-fatty acid, which one ~ o u l d expect to be formed in the first instance, should lose carbon dioxide in this way, as this is a decomposition usually only shown by p-bromo-derivstivea. The oil obtained from the hydriodic acid experiment distilled roughly a t 125-130°, and is possibly methyliao- propylcarbinyl iodide, CH(CH&CHI*CH,, which, according to Wischnegradsky (A.nnnlsn, 1878, 190, 337), boils at 127-128' ; i t was not analysed.Preparation of the mixed Ethereal Salts of Isopropylacrylic acid and of Trimet h y lncry lic acid. The greater portion of these mixed ethereal salts used in the first experimeuts in this research was prepared by the etherification ofOF ACRYLlC ACID, AND OF GLUTARIC ACID. 1489 the mixture of isopropylacrylic and trimethylacrylic acids obtained by the hydrolysis of ethjlic a-bromotrimethylpropionate by means of alcoholic potash, as explained on p. 1462 (Method I). The action of quinoline on the brominated ethereal salt (Method I T ) was not inves- tigated until mosh of the experiments described in this paper had been completed. Method 1.-The mixed ethereal salts were prepared by dissolving the mixed acids (1 vol.) in absolute alcohol (3 vols.), and adding concentrated sulphnric acid (1 ~01.1.After standing overnight, water was added, the oily layer extracted with ether, the ethereal solution washed with sodium carbonate solution, dried over anhydrous potassium carbonate, evaporated, and the residual crude ethereal salt purified by repeated fractional distillation. Although this ethereal salt has been prepared in large quantities (500 grams) on several occasions, it has always been found very difficult to obtain a product of anything like constant boiling point, and, for this reason, in the experiments described in this paper the fraction 162-1'75' was usually employed. It is, of course, possible that the ethereal salts of the two acrylic acids may boil at different temperatures, but it is .more probable that the product contains small quantities of ethylic a-hydroxytrirnethylpropionate, produced .by the action of the sul- phuric acid on the unsaturated acids during the process of etherifica- tion, and this is borne out by the results obtained by the analysis of different fractions, which were always too low.I (b. p. 163-165') ; I. 0.1240 gave 03020 CO, and 0.1094 H20. C = 66.43 ; H = 9.80. 11. 0.1561 ,, 0.3758 ,, ,, 0.2360 ,, C = 65.66; H = 9.67. 111. 0.1404 ,, 0.3379 ,, ,, 0.1218 ,, C = 65.64; H = 9.64. CBHlaO2 requires C = 6'7.60; H = 9.86 per cent. The presence of the impurity, whatever it may be, in the fraction 162-175', does not interfere with the experiment for which it was employed (condensation with ethylic malonate), except, of course, in so far as it may affect the yield of the condensation product.Method II.-A mixture of ethylic a-bromotrimethylpropionate (100 grams) and quinoline (200 grams) was heated in a metal bath in a flask connected with a retlnx apparatus, the temperature of the mixture being observed by placing a thermometer in the liquid. When the temperature had risen to 175-180°, the reaction set in, and on removing the flame the liquid boiled, though not violently,for some minutes, the thermometer rising to about 19.5'. As soon as the action had subsided, the mixture was heated to boiling for 10 minutes, the well-cooled, dark reddish-brown product poured into an excess of dilute hydrochloric acid, and the oily layer extracted twice with ether.I1 (b. p. 166-168') ; I11 (b. p. 168-172"). VOL. LXIX. 5 H1490 PERKIN : SOME DERIVATIVES OF PROPLONIC ACID, After well washing with dilute hydrochloric acid and drying over calcium chloride, the ether was distilled off and the residual oil pnrified by repeated fractionation, when the greater part was found to distil between 162' and 175O, the quautity obtained being about 60-70 per cent. of the theoretical yield. This fraction, on analysis, gave only approxirna,tely correct results. 0.1364 gave 0.3344 C02 and 0.1209 H20. C8H,,02 requires C = 67.60; H = 9.85 per cent. In order to prove that thia ethereal salt was a mixture of the etbereal salts of isopropylacrylic and trimethylacrylic acids, it was hydrolysed by boiling with alcoholic potash in the usual manner.The acid obtained distilled at 148-150' (100 mm.), and, 011 standing in a freezing mixture, deposited crystals of trimethylacrylic acid. The filtrate from these crystals did not solidify on cooling, and was proved to contain isopropylacrylic acid. Apparently, then, the elimination of hydrogen bromide from ethylic a-bromotrimethylpropionate proceeds in the same way, whether alcoholic potaah or quinoline be employed, a mixture of iso- propylacrylic acid and trimethylacrylic acid, or their ethereal salts, being formed in both cases. C = 66.86 ; H = 9.86. Condensation of Rthylic Isopropylacrylate with the Sodium Derivative of Ethylic Malonate. Formation of Ethylic a-Isopropylpropane- aa,a,-tricarboxylate, (COOC~H~)zUH*CHz*CH(C,H7)*COOC,H,.The study of the reaction which takes place when trhe mixed ethereal salts of isopropylacrylic acid a i d trimethylacrylic acid are treated with the sodium compound of ethylic malonate has been made the subject of a very large number of experiments, carried out under very varying conditions, a description of which would occupy too much space to be given here. The product formed in every case was ethylic isopropylpropanetricarboxylate produced by the condensation of ethylic isopropylacrylate with ethylic malonate, the ethylic tri- methylacrylate, as explained in the introduction, apparently taking no part in tho reaction. The yield of condensation product was found to be much influenced by the purity of the unsaturated ethereal salts ; thus, for instance, when the mixture of eth ylic isopropylacrylate and ethy lic trimethyl- acrylate had been fractionated four times with a column, and the fraction 162-175' employed, a yield of 20 per cent.of the theoretical was obtained, whereas, in one case, where the ethereal salts had been fractionated only once, and the fraction 160-177° used f o r the con- densation, the yield was only 8 per cent.OF ACRYLIC ACID, AND OF GLUTARTC ACID. 1491 Ultimately, the condensation was usually carried out in the €01 low- ing way. I n a flask connected with a reflux condenser, sodinm (46 grams) was dissolved in absolute alcohol (500 grams), and then, while still slightly warm, first ethylic malonate (320 grams) and then the mix- tiire (164 grams) of the ethereal salts of isopropylacrylic acid and trimethylacrylic acid were added ; the product was subsequently heated for one day a t GOO, and then for two days on a water bath.The opaque, somewhat dark-coloured product was mixed with twice its volume of water, acidified with bydrochloric acid, and extracted three times with etlier ; the ethereal solution was washed well with water and dilute sodium carbonate, dried over calcium chloride, evaporated, and the residual oil fractionated under reduced pressure. The yield of crude product and the weight of the various fractions obt,ained varied very much, without any apparent cause ; the follow- i n g result was obtained in one case, where the products from an experiment carried out with the quantities given above were weighed. The yield of crude product was 244 grams, and this on fractionation under reduced pressure (80 mm.) gave the following fractions : Below 160° = 98 grams.160-220 = 16 ,, 220-235 = 92 ,, Residue = 16 ,, the fraction below 160" consists principally of ethylic trimethylacryl- ate, unchanged ethylic isopropylacrylate, and some unchanged ethylic malonate.* The fraction 220-235' is nearly pure ethylic isopropyl- propanetricarboxylate. The latter, on refractionation, boiled almost constantly at 208-210 (45 mm.), and gave the following resultst on analysis. I. 0.1610 gave 0.3486 COz and 0.1247 HzO. C = 59.04 ; H = 8.60. 11. 0.1392 ,, 0.2998 ,, ,, 0.1075 ,, C = 58.74 ; H = 8.50. 111. 0.1888 ,, 0.2796 ,, ,, 0.1000 ,, C = 59.20 ; H = 8.63. CI5Hz6O6 requires C = 59.60 ; H = 8.61 per cent.Ethylic isopropylpropanetricarboxylate is a thick, perfectly colour- less oil which boils at 208-210° under a pressure of 45 mm. ; it is isomeric with the ethereal salt obtained by treating the sodium com- pound of ethylic isopropy1malonat)e with ethylic P-iodopropionate (see p. 1507), and which has the formula (C 0 OCzH5)zC (C,H,) CHz* CH2.C 0 0 C2H5 ; It A large amount of tlie etliylic malonate eniployed in the above prepayation is t These analyses were carried out with the products obtained from three hjdrolysed during the prolonged heating. &ff erent preparations. 5 H 21492 PERKIN : SOME DERIVATIVES O F PROPIONIC ACID, both ehhereal salts are very similar in properties, and boil practically at the same temperatmre, and both, on hydrolysis and subsequent elimination of carbon dioxide, yield the same isopropylglutaric acid.E - Isoprop y 1p yopane- m,al- tricarboxy Eic acid, (C 00 H) ZCH*CHa* C H ( Cs H7) *C 0 0 H. The ethylic isopropylpropanetricarboxylate, prepared as described in the last paragraph, although it gives good repults on analysis, is not quite pure, but contains traces of other ethereal salts, and on this account, €or a considerable time, i t was found impossible to obtain crystalline isopropylglutaric acid from it by hydrolysis and subsequent elimi- nation of carbon dioxide. The oily acid obtained in this way refused to crystallise after standing for a month in a cool place, and even after converting it into its anhydride, fractionating this and regenerat- ing the acid, the product did not, crystallise.Ultimately it was discovered that pure isopropylpropanetricarboxylic acid may be readily purified by the following method, and that this pure acid, on subsequent heating at 180", yields piire isopropylglutaric acid which crystallises quite readily. Ethylic isopropylpropanetricarboxylate (40 grams) is hydrolysed by boiling for one hour with pure potash (50 grams) dissolved in mebhylic alcohol; the clear liquid is then mixed with water, and evaporated on a water bath, with the addition from time to time of small quantities of water, until every trace of methylic alcobol has been removed. The concentrated solution of the potassium salt, after acidifying, is extracted 10 times with pure ether (free from alcohol), the ether evaporated, the oily residue dissolved in twice its volume of water, and the solution saturated with hydrogen chloride, the temperature being kept below 40' by cooling with ice-cold water.During this operation, crystals begin to separate, and, on standing for two days, the bulk of the tribasic acid separates i n crystalline crusts ; tbese are collected on the pump, washed with concentrated hydrochloric acid, and purified by solution in water and saturation with hydrogen chloride as before. C = 49.36 ; H = 6.50. 11. 0.1243 ,, 0.2235 ,, ,, 0.0735 ,, C = 49.04; H = 6.57. C9Hl4O6 requires C = 49-54 ; H = 6.42 per cent. Isop~op~jlpropanetricart~o~~y lic acid, when moderately rapidly heated in a capillary tube, melts at 165' with rapid decomposition, due to evolution of carbon dioxide and formation of isopropy lglutaric acid, but if slowly heated, the decomposition takes place at a lower tem- perature, generally about 158".It is very readily soluble in water, I. 0.1130 gave 0,2045 CO, and 0.0661 H,O.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1493 the solnt,ion possessing a very strong acid reaction, but it is only sparingly soluble in concentrated hydrochloric acid ; it dissolves also readily in alcohol, and moderately so in ether. Salts of Isopropylpropanetricayhoxylic acid.-The siltier salt was pyepared by adding silver nitrate to a fairly concentrated warm solu- tion of the ammonium salt. It is a white, amorphous precipitlate, which, after washing well with warm water, and drying first on a porous plate over sulphuric acid in a vacuum, and then at loo', gave the following result on analysis.0.3464, on ignition, gave 0.2064 Ag. C9HllAg306 requires Ag = 60.11 per cent. A neutral so1ut;ion of the ammonium salt gives no precipitate with calcium chEoride or copper acetate, but a white, gelatinous precipitate is formed with lead acetate. On the addition of barium chhide, no precipitate is produced at first, but on standing, or more rapidly on warming, an insoluble crystalline barium salt separates. Ag = 59.58. From the experiments already described i t seemed probable, as stated in the introduction, that when a mixture of ethylic isopropyl- acrylate and ethylic trimethylacrylate is digested with the sodium derivative of ethylic malonate, that the former alone undergoes con- densation, the latter remaining unchanged ; in order to determine, therefore, whether this was really the case the foliowing experiments were made.Experiment I.--The oil boiling below 160' (80 mm.), obtained in fractionating the product of the action of the mixed ethereal salts on ethylic malonate, as described on page 1491, was collected from several operations, carefully fractionated with a column, and the fraction 160-175' digested with a large excess of the sodiuni derivative of ethylic malonate in alcoholic solution, for about three days. The product was isolated as before, and, on fractionation, yielded a small qixantity of ethylic isopropylpropanetricarboxylate, shomiiig that small quantities of ethylic isopropylacrylate were present in the ethereal salt employed. The large quantity of oil of low boiling point, which was obtained during the fractionation, was again treated with ethylic malonate as before, and yielded now only traces of ethy lic isopropylpropanetricarboxylate ; it was very carefully frac- tionated, and the fraction boiling at 160-175O hydrolysed by boiling with alcoholic potash, and the acid obtained fractionated under reduced pressure, the fraction 135-142' (90 mm.) being collected separately.Owing to the mildness of the summer weather, this oil, which should contain the recovered trimethylacrylic acid, showed no signs of crystallising ; it was, therefore, dissolved in chloroform,1494 PERKIN : SOME DERlVATXVES OF PROPICNIC ACID. and titrated with bromine until the colour remained permanent, when about 80 per cent. of the theoretical quantity of bromine was absorbed, shon-ing that this crude acid must contain a small quantity of some saturated acid as impurity.On allowing the chloroform to evaporate spontaneously, crystals gradually separated from the oily residue : these were collected, spread on a porous plate, and recrys- tallised from light petroleum ; the substance then melted a.t 185-190°, and had all the properties of dibromotrimethylprupionic acid, CBr ( C El,),* C Br ( CH,) * CO OH. C,H,,Br,O, requires Br = 58.40 per cent. 0.1391 gave 0.1906 AgBr. Br = 58-58. The formation of this substance shows that when the mixture of ethylic isopropylacrylate and ethylic trimethylacrylate is digested with the sodium derivative of ethylic malonate, sowe of the latter, at all events, remains unattacked, even when the operation is repeated three times.Ahpeiiment 11.-Attempts were now made to determine whether any of the ethylic trimethylacrylate had taken part i n the reaction, in which case e thylic trimethyl propanetricarboxylate must have been formed, and, on hydrolysing the product of the condensation, tri- methylpropanetricarboxjlic acid, (C 0 0 €1) ?C H*C: ( C H3) 2*C LT ( C H3) C 0 0 H, would be obtained mixed with large quantities of isopropylpropane- tricarboxylic acid. If this were the case, the former acid muRt be contained in the hydrochloric acid mother liquors, from which the latter acid had crystallised out as explained on p. 1492. These mother liquors were collected, extracted about 20 times with pure ether, the ether evapo: rated, and the residual oily acid, dissolved in a little water, was saturated with hydrogen chloride, the whole being well cooled during the operation.After standing for a fortnight, the crystals of isopropylpropane- tricarboxylic acid which had separated were removed by filtration through tt platinum cone, the filtrate again extiacted with ether, and the treatment with hydrogen cbloride repeated. Lastly, the small quantity of acid which did not crystallise was heated at 180°, and the 1-esidue etherified by treatment with alcohol and sulphuric acid, when an oily, ethereal salt was obtained which distilled, for the most part, at 170-172' (90 mm.). As the acid obtained from this on hydrolysis did not solidify, it was converted into the anhydride by boiling with acetic anhydride, and the product (3 grams) fractionated, the fraction 170-175' (30 mm.) being reconverted into the acid ; the oily acid mas thenOF ACRYLIC ACID, AND OF OLUTARIC ACID.1495 dissolved in a little water, and the solution saturated with hydrogen chloride, On standing, about 1 gram of a crystalline acid separated which melted at 90-93', and consisted evidently of isopropylglu- taric acid, since it gave an anhydride melting at 53', and an anilic acid melting at 156-158'. As far as can be judged from these experiments, it appears that ethylic trimethylacrylate is not capable of condensing with ethylic malonate, and I hope to prove this definitely by experiments with pure e t hylic trime t h ylacry late. Isopropylqhttr& acid, CH (CH,),*CH (C OOH)*CH2*C H2*COOH.In order to prepare this acid, pure isopropylpropanetricarboxylic acid is heated in an oil bath at 200' until the evolution of carbon dioxide has entirely ceased ; the residue is dissolved in boiling water, and the solution, after cooling, saturated with hydrogen chloride. After standing overnight, the colourless crystals which separate are collected on a platinum cone by the aid of the pump, washed with concentrated hydrochloric acid, and reci*ystallised by dissolving them in a little water and saturating with hydrogen chloride as before. The analysis of this acid gave the following results. I. 0.1485 gave 0.2970 CO, and 0.1073 H20. C = 54.54 ; H = 8.03. 11. 0*1188 ,, 0.2385 ,, ,, 0.0858 ,, C = 54.82 ; H = 8-02. C8H,,OI requires C = 55.18 ; H = 8.05 per cent.It is readily soluble in water, and crystallises from the hot, concentrated solution, on cooling, in magnilicent, colourless prisms ; it is sparingly soluble in concen- trated hydrochloric acid, and is, therefore, most economically purified by saturating the aqueous solution with hydrogen chloride. The finely powdered crystals dissolve readily i u alcohol or ether, and moderately i n cold benzene, but the acid is almost insoluble in light petroleum. Isopropylglutaric ucid melts at 94-95O. Salts of Isopropylglutaric acid. The silver salt, C8H,,Ag204 was prepared by adding silver nitrate t o a slightly alkaline and fairly concentrated solution of the ammo- nium salt; the white flocculent precipitate was collected, washed with water, dried at loo", and analysed. I.0.2510 gave 0.2239 C02, 0.0702 H20, and 0.1396 Ag. C = 24.55 ; H = 3.10; Ag = 55.58. Ag = 55.70. 11. 0.2487, on ignition, gave 0.1385 Ag. CeH,,Ag20t requires C = 24.74; H = 3.10 ; Ag = 55.65 per cent. A dilute neutral solution of ammonium isopropylglutarate gives no precipitate with barium nitrate or calcium chloride. Copper sdphate1496 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, gives a bright green, gelatinous precipitate, and 2ead acetate, on warm- ing, a white, flocculent precipitate. E t h y Zic Isopropy l g Zutarate, C 00 C2Hs*CH (C&) C H2* C H2*C OOC2H5. This ethereal salt was prepared by dissolving the crude acid (obtained by heating pure isopropylpropanetricarboxylic acid at 200') in 4 vols. of absolute alcohol, and adding 1.5 vols. of concen- trated sulphuric acid.After standing for two days, water was added, the oily ethereal salt extracted with ether, the ethereal solution washed with water and sodium carbonate, dried over anhydrous potassium carbonate, and the residual oil purified by distillation under reduced pressure. Ethylic isopropylglntarate is a colonrless oil, having a strong, ethereal odour ; it boils at 158-160' (45 mm.). 0.1251 gave 0.2860 GOz and 0.1079 H20. C = 62.35 ; H ClzHz2O4requires C = 62-61 ; H = 9.56 per cent. = 9.59. yH2* CH2* CO CH(C~H,)*C o>O- Isopropylglutaric Anhydride, This is readily prepared from isopropylpropanetricarboxylic acid by heating it at 200' until the evolation of carbon dioxide has ceased, dissolving the residue in freshly distilled acetic anhydride, heating the solution to boiling for 10 minutes, and then placing it over solid potash in a vacuum desiccator.After some days, the crystals of the anhydride are collected, left in contact with porous porcelain until colourless, and then recrystallised from boiling light petroleum (b. p. 50-60'). The following results were obtained on analysis. 0.1448 gave 0.3280 CO, and 0.1007 H20. C = 61.17; H = 7.72. 0.1490 ,, 0.3336 ,, ,, 0.1017 ,, C = 61.06; H = 7-56, C8H1103 requires C = 61.54 ; H = 7*68 per cent. IsopropylgZutaric anhydride melts at about 53", but not quite sharply, and the melted substance, on cooling, sets to a transparent, jelly-like mass, which, on rubbing, soon becomes cryatalline and opaque. It is only sparingly soluble, even in boiling light petroleum (b.p. 50-SO'), and is almost completely deposited on cooling in very voluminous, feathery masses, consisting o€ slender needles, somewhat resembling cotton wool in appearance. It melts in boiling water, and dissolves only very slowly ; on concentrating the solution and allowing it to stand, isopropylglutaric acid separates in magnificent, colourless prisms.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1497 Isopropy Eghtaranilic acid, C 0 OH*C Hz*CH2*CH ( C3H,) *CO*NH*CSH5, or COOH*CH(CsH7)*CHz*CH,*CO*NH*C~Ha (3). In order to prepare this substance, 0.9 gram of pure isopropyl- glutaric anhydride was dissolved in 5 C.C. of pure dry benzene, and 0.7 gram of freshly distilled aniline added ; this occasioned a con- siderable rise of temperature, and i n about a minute the whole crys- tallised.The crystals were collected, dyained on porous porcelain, and recrystallised from dilate methylic alcohol, when beautiful, colourless, glistening crystals were obtained, which gave the following results on analysis. 0.1384 gave 6.8 C.C. moist nitrogen at 18" and 755 mm. N = 5.64. C,H,,NO, requires N = 5.62 per cent. ~sTsopro~ylglutaranilic acid softens at 150°, and melts at 158-1 59'. It is readily soluble in alcohol, but only sparingly in benzene, light petroleum, or water. It does not appear to give off wster, even when heated at 250°, but when heated to boiling for some time in a test-tube, decomposition takes place, probably with formation of the anil; as, howevey, the product waB not readily purified, and I had very little substance at my disposal, the matter was not further in- vestigated.Oxidation of Isoproyylylutnric acid. Isopropylglutaric acid is only very slowly attacked by potassium dichromate and sulphuric acid even on boiling; 10 grams of the acid heated to boiling with 15 grams of potassium dichromate, 40 grams of concentrated sulphuric acid, and 300 C.C. of water for 10 days in a reflux apparatus, still contained unreduced chromic acid. The oxidation takes place, however, much more rapidly when the acid (10 grams) is digestled with chromic acid (30 gramsj and sulphuric acid (50 grams, diluted with 200 C.C. of water), the reduction of the chromic acid being complete after three days' boiling, The product from both experiments was distilled with steam until the condensed water ceased to have an acid reaction, the distillate was boiled with excess of barium carbonate, filtered, the filtrate evapo- rated to a small bulk, and precipitat,ed with silver nitrate.The white crystalline silver salt was collected with the aid of the pump, washed with water, dissolved in boiling. water, the hot solution filtered from traces OE reduced silver, and allowed to stand; the glistening crystals of pure silver acetate which separated were collected, dried at IOO", and analysed. 0.1940, on ignition, gave 0.1252 Ag. Ag = 64.54. The residue from the steam distillation was boiled with excess of CH,-COOAg requires Ag = 64.67 per cent.1498 PERKIN : SOXE DERIVATIVES OF PROPIONIC ACID, caustic ~ o ~ I I , , filtered from the chromium hydroxide, and the latter repeatedly washed with water; the combined liquors were then evaporated to a small bulk, acidified, and extracted three times with pure ether (A), and then again SO times with pure ether (13). The ethereal solution A, on evaporation, deposited a greeuisli crystalline acid, which, on recrystallisation from hydrochloric acid, at once yielded an acid, melting at 92-95', which proved to be uiiattacked isopropylglutaric acid. 0.1489 gave 0.2991 GO2 and 0.1075 H20.C,H,,02 requires C = 55.18 ; H = 8.05 per cent. The ethereal solution B, on evaporahion, yielded a greenish crys- talline mass, which contained only slight traces of isopropylglutayic acid. On repeatedly fractionally crystallising this substance from hydi*ochloi*ic acid , beautiful colourlesr;l crystals were ultimately obtained, which melted at 183-185', and consisted of pure succinic acid, as the following analysis shows : C = 54.78 ; H = 8.02.0*1100 gave 0.1639 CO, and 0.0523 H,O. C4H,04 requires C = 40.68; H = 5.10 per cent. No other acids could be isolated from the products of the oxida- tion, and it therefore follows that when boiled with chromic acid and sulphuric acid, isopropylglutaric acid is oxiilised with formation of acetic and succinic acids. C = 40.66 ; H = 5.28. Action of E'thplic a-Bronzotrimethylpropionate on the Soditmh Desizatit-e of Ethylic Jfalonate in Alcoholic Solutiorc. I n studying this decomposition, sodium (6 grams) was dissolved in alcohol (75 grams) in a flask connected with a reflux condenser, ethylic malonate (40 grams) and ethylic a-bromotrimethylpropionate (56 grams) were added, and the whole allowed to stand for some time.Very little action seemed to take place in the cold, but, on gently warming on a water bath, sodium bromide separated rapidly, the separation, after half ail hour, being apparently so complete that in the earlier experiments the action was stopped a t the end of this time ; subsequently, however, it was found that a very much larger yield of condensation product could be obtained if the liquid was kept in vigorous ebullition for 2-3 days. At the eiid OE this time, the alcohol was distilled off from a salt bath, the residue mixed with watw: the oil which WRS precipitated extracted with ether, the ethereal solution washed, well dried over calcium chloride, the ether distilled off, and the light brown, oily residue fractionated under reduced pressure (100 mm.).OF-ACRYLIC ACID, AND OF GLUTARIC ACIJ). 1499 About half of' this product distilled between 125' and 145O, the thermometer then rose rapidly, the greater bulk (22 grams) passing over between 222' and 230' (100 mm.) ; the latter, on refractiona- tion, distilled for the most part constantly at 220' (100 mm.) as a colourless oil, which gave the Eollowing numbers on analysis : 0.1501 gave 03255 GO, and 0.1168 H,O.C = 59.14 ; H = 8.65. The careful examination of this substance showed that it was not C,sH2606 requires C = 59.60 ; H = 8.61 per cent. ethylic methylisopropylethanetricarboxylate, but that it consisted of very nearly pure ethylic isopropylpropanetri- carboxylate (which it was expected would be formed during this reaction), identical with the product obtained by condensing ethylic isopropylacrylate with the sodium derivative of ethylic malonate (p.1491) ; this was proved in the following. way. The ethereal salt was hydrolysed with alcoholic potash, and the tricarboxylic acid isolated and purified exactly as described in the case of isopropylpropanetricarboxy~ic acid (p. 1492). The crystalline product, like the acid just named, melted at 163-165' with decom- position, and gave, on analysis, the following numbers : 0.1695 gave 0.3064 GOz and 0.1008 H,O. C = 49.30 ; H = 6-56, C9H,,06 requires C = 49.54 ; H = 6.42 per cent. 011 heating this acid at 200°, carbon dioxide was evolved, and the residue, after being twice recrystallised from hydrochloric acid, melted at 94-95', and consisted of isopropylglutark acid, as was ahown by converting it into its anhydride (m.p. 33'). and into isopropylglu- taranilk acid (m. p. 157-159'), and also from the results obtained on analysis. (C OOCaHb) C (CH,) (C3H,)*CH (COO CZH5)2, 0.1472 gave 0.2956 GO, and 0.1065 H,O. C,H,404 requires C = 55.18 ; H = 8.05 per cent. There can be no doubt that the product of the action of ethylic bromotrimethylpropionate on the sodium derivative of ethylic malotiate in alcoholic solution, is ethylic isopropylpropanetricarb- oxylate. C = 54-77; H = 8.10. Action of h'thylic Bromot,-imethylp'rprop~onate on the Xodiuin Derivative of Ethylic Maloiaate in Xylene Solutioiz. As stated in the introduction to this paper, it has often been found that in cases where a glutaric acid derivative is formed by t h e action of an a-bromo-ethereal salt on the sodium derivative of ethylic malonate in alcoholic sohition, a widely different result mar be1500 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, obtained by substituting xylene for alcohol as the solvent, the product consisting then for the most part of a derivative of succinic acid.In order t o determine whether this would be the case in the present instance, the following experiment was made. Sodium (3 grams) was melted under boiling xylene and the whole violently agitated so as to bring the sodium into a state of the finest possible division, ethylic malonate (20 grams) was then added, and the whole heated to boiling in a reflux apparatus until the sodium had completely dis- appearzd ; ethglic a-bromotrimethylpropionate (28 grams) was then added, and tbe mixture heated to boiling for two days. The brownish product was shaken with water and a quantity of ether, the ethereal solution was beparated, washed well with water and dilute hydro- chloric acid, dried over calcium chloride, and the ether and xylene distilled off, at first under ordinary, but subsequently under reduced pressure.On fractionating the oily residue under 100 mm. pressure, the greater portion distilled below 160°, only about 6 grams passing over between 210' and 230'. The latter portion, on refractionation, passed over between 218' and 222' as a colourless oil, which, on aualpsis, gave the followii~g results. 0*1680 gave 0.3632 CO, and 0.1296 H?O.CI5Hz6O6 requires C= 59-60 ; H = 8-61 per cent. That this substance consisted, in this case also, at all events for the most part, of ethglic isopropylpropanetricarboxylate was proved in the following way. The ethereal salt, on hydrolysis, yielded a tricarboxylic acid which melted at 163', and at slightly higher temperatures decomposed, yielding an acid which, after recrystallisation, melted at 96-95' ; this acid gave an anhydride melting at 53", and an anilic acid melting at 156-158' ; it was, therefore, evidently isopropjlglutaric acid. It is, of course, quite possible that, in this experiment, some cthylic methylisopropylethanetricarboxylate may have been formed, but the quantity present in the product of the reaction can only have been small.C = 58.96 ; H = 8.58. Action of Yheiaoxyethylic Bromide, CsHoO*CH2*CH2Br, on the Sodium Dsrivutive of Ethy lic Dirneth ylpropanetricarboxylate, (COOC,H,),C l3.C (CH,),*CH,*C OOC2Hb. The reason for investigating this reaction haB been mentioned in the introduction to this paper; the experiment was carried out as follows. Sodium (5 grams) was dissolved in absolute alcohol (70 grams) in a Bask connected with a reflux apparatus ; the solution, when nearly cold, was mixed with ethylic dimethylpropnnetricni*boxylak (62OF ACRYLIC ACID, AND OF OLUTARIC ACID. 1501 grams), then phenoxyethylic bromide (45 grams) was added, and the whole gradually heated to boiling on a water bath. No action appeared to take place until the temperature had risen to about 50', then sodium bromide began to separate rapidly, and after heating to boiling for four hours the liquid had a neutral reaction.Water was now added, the oily product extracted four times with ether, the ethereal solution well washed with water, dried over calcium chloride, and evaporated, when 88 grams of a very slightly coloured oil was obtained ; t h i s was not analysed but directly converted into the corre- sponding acid by hydrolysis. %'or this purpose, the oil was heated for four hours in a reflux apparatus wiih potash (80 grams), dissolved in pure methylic alcohol, the bulk of the alcohol was then distilled off, water added, and the neutral oil (phenoxyethyl ethyl ether, C6H50*CH2*CH2*OC2H5, see p. 1503) which separated was extracted with ether ; the aqueous solution was then evaporated on a water bath until quite free from alcohol, and the cold, moderately concentrated solution acidified with hydro- chloric acid.This caused the precipitation of a viscous oil, which, however, on adding ice and shaking well, solidified in the course of half an hour, to a hard, ochre-coloured mass. In order to purify it, the crude substance was ground up with water, well washed by means of the pump, dissolved i n dilute sodium carbonate, and the brown solution heated on a water bath with purified animal charcoal for an hour ; after filtration, the liquid was very much lighter coloured, and, on the addition of hydrochloric acid, gave an almost colourless, oily pre- cipitate, which, however, very rapidly solidified. The solid acid was collected, washed with water, left in contact with porous porcelain for a few days, and then purified by recrystallisation from dilute acetic acid ; it was thus obtained iii the form of colourless prisms, which, on analysis, gave the following numbers.I. 0.1106 gave 0,2676 GO2 and 0.0626 HzO. C = 65.99 ; H = 6.33. IT. 0.1392 ,, 0.3381 ,, ,, 0.0750 ,, C = 66.23 ; H = 5.98. (C6H5*O*CH2*CH2)zC(COOH)2 requires C = 66.28; H = 5.81 p.c. This acid melted at about 150-152' with decomposition, and was found, on examination, to be diphenoxyethylmalonic acid, identical with the acid which was subsequently prepared by the action of phenoxyethylic bromide on the sodium derivative of ethylic malonate, and is described i n a previous paper (Bentley, Haworth, and Perkin, this vol., 169).Sults of Di~he?aoxyethylnialonic m i d . These have not been previously described. The siZver saM, ( C6H50*CH,*CH,)C(COOAg)2, was obiained as a curdy, white precipitate on adding an excess of silver nitrate to1502 PERKIN : SOME DERIVATIVES OF PROPIONIC ACID, the slightly alkaline solution of the ammonium salt of the acid. It was collected, well washed with water, dried on a porous plate over sulphuric acid and t8hen for a short time in the water oven. C = 40.09; C,,H,,Ag20s requires C = 40.86; H = 3.22; Ag = 38.71 per cent. Thedilute solution of the ammoilium salt of this acid shows the 0.1981 gave 0.2907 C02, 0.0621 H,O, and 0.0759 Ag. H = 3.45; Ag = 38.31. following behaviour wit11 reagents. Calcium Chloride.-A heavy, white, amorphous precipitate.H n ~ i z ~ n ~ Chloride.-A white, amorphous precipitate, which, on boiling, apparently becomes crystalline. Copper XuZphate.-A light blue gelatinous precipitate, insoluble in water. Lead Acetate.-A white, gelntinous precipitate, which, on boiling, cakes together to a caseous mass quite insoluble in water. Dip henox yet h y lace tic acid, ( C6HD* O*CH2*C H,) H* C 0 0 H. When diphenoxyethylmalonic acid is heated at 180', it is rapidly decomposed wit,h evolution of carbon dioxide and formation of diphenoxyet,hylacetic acid, the reaction taking place quantitatively, as is shown by the following experiment. 9.0224 grams of the pure dibasic acid, heated in a small flask in an oil bath, lost 1.1574 gram of CO, = 12.83 per cent., whereas accord- ing to the equation { C,H,.O*CH2*CH,),C (COOH),= (C6H6*O*CH,*CH2),CH0COO€€ + Cot, the calculated loss is 12.78 per cent.The residual monobasic acid was dissolved in dilute sodium carbonate, and the solution digested with purified animal charcoal and filtered; the filtrate was then acidified with hydrochloric acid, and the precipitated acid, which at first was oily but which soon solidified t o a hard cake, was washed with water, and purified by I ecrystallisation from dilute alcohol. 0.1264 gave 0.3334 CO, and 0.0751 H20. C = 71.93 ; H = 6.60. (C,H6*O*cH2*CH2)2CH*COOH requires C = 72.00 ; H = 6.67 p. c. It melts at 87-88' and is identical wit,h the acid obtained from the sodium compound of ethylic malonate by the action of phenoxyethylic bromide (Bentley, Haworth, and Perkin, this vol., 169).The follon-- iug salts of this acid, which have not been described, were prepared when the acid was first obtained. The silver salt, (CsH,*O*CH2*CH2),CH*COOAg, obtained on adding silver nit'rate to a slightly alkaline solution of the ammoilium salt of the acid, is a caseous precipitate, which after repeated washingOF ACRYLIC ACID, AND OF GLUTARIC ACID. 1503 with water, becomes quite brittle. It was ground up, washed well with water and then with methylic alcohol, and dried over sulphuric acid. C = On analysis the following results were obtained. I. 0.2480 gave 0.4830 GO2, 0.1068 H,O! and 0.0646 Ag. 53.11 ; H = 4.78 ; Ag = 26.05. IT. 0.1240 gave 0.0326 Ag. The dilute neutral solution of the ammonium salt shows the follow- Ag = 26.45. C,,H,,AgO, requires C = 53.07; H = 4.67; Ag = 26.53 per cent.ing behaviour with reagents. Calcium Chloride.-A white gelatinous Frecipitate, almost insoluble Barium Nitrate.--No precipitate even on boiling. Copper XuZphate.-A light blue, caseous precipitate, which, on Lend acetate.-A white, caseous precipitate, which, on warming in water. boiling, becomes granular. with water, becomes pasty and apparently decomposes. Phenoxyethy l Ethyl Ether, Ce&'O*C H2*CH2*O*C2H5. During the hydrolysis of the product of the action of phexoxy- ethylic bromide on the sodium compound of ethylic dimethylpropane- tr-icarboxylate it was found, as stated on p. 1501, that a considerable quantity of a neutral substance mas formed. The product from several experiments was collected, dissolved in ether, washed with water, dried over calcium chloride, and the ethereal solution eca- poi-ated ; on submitting the oily residue to fractional distillation, the greater portion boiled at 227-229*, and gave the following results on analysis. 0.1.502 gave 0.3974 CO, and 0.1150 H20.C,oH,102 requires C = 72.28 ; H = 8-43 per cent. There can be no doubt that this substance is phenoxyethyl ethyl ether, produced by the action of sodium ethoxide on the phenoxy- ethylic bromide used in the experiments. C6H5*0-CH ,*CH,Br + NaOC,H, = C6E1[5.0.CH,.CH?*OoC?~5 + NaBr. It is a colourless oil possessing a penetrat'ing odour somewhat similar to that of benzyl ethyl ether, C6&,-CB2*O*C2H5. C = 78-16 ; H = 8.50.1504 PERKIN : SOME DERJYATIVES OF PROPIONIC ACID, Action of Pheizoxyethylic Bromide, C6Hs*O*CH2GH,Br, on the Sadiun~ Foimation of Deyivative of Ethylic Isopropyl@-opanetricarboxylate.Pheizoq et hy lisoprop y lpropanetricarbox y lic acid, (COOH),? ---CHa*$JH*COOH CH2*CH2*O*Cp,H5 CH (CH,), In carrying out this experiment, 1.2 gram of sodium was dissolved in 15 grams of absolnte alcohol, and, after cooling well, 15 grams of ethylic isopropylpropanetricarboxylate was added ; the yellow solu- tion thus produced was mixed with 11 grams of phenoxyethylic bromide, and the whole heated in a reffnx apparatus on the water bath; at first very little action appeared to take place, but after a time the liquid became suddenly cloudy and quantitiee of sodium bromide separated rapidly. After heating for 10 hours to boiling, the product, which was still alkaline, was mixed with water, and extracted five times with ether; the ethereal solution was washed well, dried over calciuni chloride, and evaporated, when 22 grams of a yellow oil were ob- tained, which did not solidify after standing over sulphnric acid in a vacuum for two days.This oil, on examination, proved to be a mixture ; it was therefore not analysed but converted into the corresponding acids, by boiling for two hours with methyl alcoholic potash (20 grams). The alkaline product was evaporated until free from alcohol, extracted with ether to remove any neutral oil (phenoxyethyl ethyl ether), again evapo- rated, the residue dissolved-in water, and, after cooling well, acidified with dilute hydrochloric acid. The yellowish scmi-solid mass which was thus precipitated soon became quite solid, especially after adding ice and shaking well; it was collected, ground up with water, and washed with the aid of the pump, the mother liquors being preserved for subsequent investigation (see below). In order to purify the crude acid thus obtained, it was dissolved in dilute sodium carbonate, digested with purified animal charcoal, filtered, and the filtrate, after cooling with ice, acidified with hydrochloric acid, the whole being stirred during the operation. I n this way an almost colourless precipitate of nearly pme phenoxyethylisopropyl- propanetricarboxylic acid was obtained, which, after washing and drying, resembled starch powder in appearance ; in this condition, i t melted at 178-180° with vigorous evolution of gas. This product was now dissoIved in much pure ether, the ethereal solution evaporated to a small bulk, and allowed to stand in a closed flask, when the pure acid separated in hard, glistening, crystalline crusts ; these crystals were washed with ether, dried at loo", and nnalysed with the following results.OF ACRYLIC ACID, AND OF GLUTARIC ACID. 1505 I. 0.1080 gave 0.2368 CO, and 0*0640 H20. C = 59.80 ; H = 6.58. IT. 0.1032 ,, 0.2286 ,, ,, 0.0630 ,, c = 60.41 ; H = 6-78, C17H120, requires C = 60.36; H = 6.51 per cent. Phenoxyethylisopropylpropunetricurboxylic acid, when heated in a capillary tube, softens at 178" and melts at 179-180° with I-apid decomposition into carbon dioxide and the corresponding dibasic acid. It is readily soluble in alcohol, but only sparingly in benzene, light petroleum, chloroform, and cold water; the finely divided sub- stance dissolves with difficulty in boiling water, and crystallises out again on cooling almost completely in curious nodular masses. The yield of the phenoxy-acid obtained by the above method is only 4-5 grams from 15 grams of ethylic isopropylpropanetricarboxylate, whereas, assuming this ethereal salt to be pure, over 17 grams should have been obtained. One reason for this is the formation of consider- able quantities of phenoxyethyl ethyl ether during the reaction, and thus a portion of the e thylic isopropylpropanetricarboxylate employed takes no part in the decomposition, and remains unchanged in the crude product of the reaction. On hydrolysing this crude pro- ductl, as described above, and precipitating the phenoxy-acid with hydrochloric acid, isopropylpropanetricarboxylic acid remains in the filtrate and in the mother liquors from the purification of this acid, and is easily recovered by neutralising with sodium carbonate, evapor- ating to a small bulk, acidifying, and extracting 10 times with pure ether. After drying over calcium chloride and distilling off the ether, the oily residue was etherified by treatment with alcohol and sulph- uric acid, and the ethereal salt, after -purification by distillation under reduced pressure, was employed in subsequent operations. Ultimately a good yield of phenoxyethylisopropylpropanetricarb- oxylic acid may be obtained in this way ; but the preparation of this acid is, nevertheless, a very tedious one, many weeks being required in preparing it-even in small quantities. Phertoxyethylisopropylglutaric acid, COOH*~H-CH2-FH*COOH CH,*CH,*O*C,H, CH(CH,), - When phenoxyethylisopropylpropanetricarboxylic acid is heated at 180-19U0, it rapidly loses carbon dioxide, and, apparently also, water vapour, and is converted into a colourless oil which, on cooling, does not solidify, even on long standing, and may possibly consist of the anhydride of the dibasic acid. It was dissolved in boiling dilute sodium carbonate, the solution filtered, acidified, and extracted with pure ether, the ethereal solution dried over calcium chloride, evaporated at a low temperature, and the residual oil allowed to stand for two VOL. LXIX. 5 11506 HEINKE AND PERKIN : ETHYLIC &IODOPROPIONATE days, when it had almost completely solidified. Considerable diffi- culty was experienced in endeavouring to recrystallise the acid, but ultimately this was accomplished by dissolving the crude substance i n a large quantity of light petroleum (b. p. 100-120°), boiling with animal charcoal, filtering, and allowing the solution to stand for some days exposed to the a i r ; nodular crystals then separated, which, after washing with light petroleum (b. p. 40-45'), gave the following numbers on analysis. 0.1317 gave 0.3161 CO, and 0.0899 H,O. C = 65.43; H = 7.50. 0.1431 ,, 0.3415 ,) ,, 0.0969 ,, C = 65.10; H = 7.51. C16H2205 requires C = 65.30; H = 7.49 per cent. Phenoxyethylisopropylglutaric acid melts at about 90-93O ; it is readily soluble in alcohol and ether, but almost insoluble in water. When heated for some time in a water bath, it is converted into an oil which, on cooling, sets to a transparent jelly; this does not ciyitallise, even on mbbing with a crystal of the pure acid, and, therefore, may possibly be the anhydride of the acid, especially as i t dissolves in cold sodium carbonate with difficulty.