30 CLARKE, LAPWORTH, AND WECHSLER : CONDENSATION OFIV.--Condensa tion of lietones Containing the Group*CH2*C0 C I I : with Esters in Presence of SodiumEthoxide.By REGINALD W. L. CLARKE, ARTHUR LAPWORTH, and ELKANWECHSLER.THE acetoacetic ester condensation, in its most general form, is thatwhich occurs when a carboxylic, nitrous, or nitric ester (class E)is brought into contact with a ketone, carboxylic ester, or nitrile(class K) containing the group :CH*CO*, or :ClH.CN, in presence ofsodium ethoxide, sodamide, sodium, or similar agent. The questionof its mechanism has been the subject of much controversy and isassociated with numerous investigations. ** Geuther, Jahresber., 1863, 323 ; Zeitsch. f. Chent., 1868, 11, 662 ; Franklandand Duppa, Phil. Tram., 1866, 156, 37 ; Aanden, 1866, 138, 204, 328 ; Kolbe,Zeitsch.f. Chem., 1867, 10, 637 ; Wislicenus, Annalen, 1877, 186, 163 ; Baeyer,Ber., 1885, 18, 3640; Duisberg, Ber., 1883, 16, 133 ; Claisen and Lowman, Ber.,1887, 20, 651 ; 1888, 21, 1154 ; Bromnie and Claisen, ibid., 1888, 21, 1132 ;Claisen, ibid., 1894, 27, 114; 1905, 38, 708; also Annalen, 1893, 277, 184;1896, 291, 25 ; 1897, 297, 92 ; Michael, J. pr. Chem., 1888, [ii], 37, 507 ; Ber.,1900, 33, 3731; 1905, 31, 1922; Nef, Annalcn, 1897, 298, 319; Dieckmann,Ber., 1900, 33, 2670 ; Lapworth, Trans., 1901, 79, 1269 ; 1902, 81, 1512 ; Proc.,1903, 19, 190KETONES CONTAINING THE GROUP ‘CH;CO’CH: WITH ESTERS. 31I n spite of all evidence against it, Claisen’s theory of themechanism of the condensation (Ber., 1887, 20, 646; 1888, 21,1154) is still the one most frequently cited.The theory wasadvanced at a time when the general rule appeared to obtain that thesubstances of class K (abovo) must contain the group *CH2*CO*, orCH,*CN, and fell to the ground when Dieckmann explained theapparent absence of reactivity in compounds containing the group:CH*CO*, or :CH*CN (Ber., 1900, 33, 2670), and Perkin and Thorpediscovered a case which could not possibly be explained by Claisen’sproposition (Trans,, 1900, 79, 736 and 737).I n the work described in the latter part of the present paper, it isshown that certain simple ketones containing the group:CH*CO* CH,-,when submitted to the action of an alkyl nitrite under the conditionswhich with other esters lead, as usual, t o attack at the *CH,* group,are affected only at the :CH* group, and an explanation of this isafforded readily enough on general grounds and in the light ofDieckmann’s experiments.Claisen’s theory, however, is not appli-cable, and in this connexion his own words may be quoted : “Zuverwerfen sind natiirlich alle Vorstellungen, mit denen nur einzelne,nicht alle dieser Estercondensationen erklart werden ” (Ber., 1905,38, 715).Other views of the acetoacetic ester condensation have assumed apreliminary conversion of the compounds of class K into theirC-sodium derivatives, Na*6*CO*, or JYa*d.CN (Frankland and Duppa,Annalen, 1883, 219, 123; Baeyer, Ber., 1885, 18, 3640; comparealso Michael in numerous papers already cited), but recently thetendency has grown among chemists to suppose that the metallic deriv-atives of ketones, nitriles, kc., when they exist, are derived solely fromthe enolic forms, and such is doubtless the case with the solid com-pounds ; consequently, efforts have been made to explain the apparentlyanomalous reactions of these 0-sodium derivatives as involvingadditions at the double linking, for example :>C:d*ONa + CH,I -+ >CNe*g*ONa -+ >CMe*bO + NaII(Michael, J.pr. Chern., 1883, [ii], 27; 487; 1892, [ii], 46, 205 ; 1899,[ii], 60, 316), or, as the result of isomeric change in pre-formed0-substituted compounds :>C:’C*O*CH, --+ >CMe*&O(compare Claisen, Ber., 1905, 38, 714, where such a change isassumed as a part of the mechanism of the acetoacetic estersynthesis).32 CLARKE, LAPWORTH, AND WECHSLER : CONDENSATION OFSerious objections can be urged against such explanations, and weventure here to draw attention once more to what appears likely toprove the most satisfactory view, namely, that isomeric change isinvolved, but that it is to be looked for in the ionsof the metalderivative (WislicenuF, ‘‘ Tautomerie,” Ahrens’ Voi-trage, 897 et seq ;Knorr, Anncden, 1896, 293, 39; Lapworth and Hann, Trans., 1902,81, 1512; 1904, 85, 4 8 ; Lapworth, Proc., 1903, 19, 190) or inthe sodium derivative itself (Lander, Trans., 1903, 83, 420). Itmay here be noted that Bruhl has brought forward some evidence thatmetallic derivatives of camphor can actually be isolated in both theC- and 0-forms (Ber., 1904, 37, 2170).Apart altogether from the conception of electrolytic dissociation, theexceptional lability of metals, and especially those of the alkalis andalkaline earths as exemplified by the almost universal and instanta-rieous reactivity of the metallic salts and even organo-metallic com-pounds, is sufficient to render it likely that they must be more labilethan any other types of atoms or groups.Lander’s proposition mayusefully be employed as an alternative to the earlier ionic one, beingespecially attractive, since Kahlenberg’s experiments with the oleatesin which the metals exhibit their instantaneous lability in spite ofthe absence of noticeable dissociation.From this standpoint, there is, in solutions of the sodium deriv-atives of p-ketonic esters and allied compounds, a virtual or r e dequilibrium between the 0- and C-sodium derivative ::C:CONa +-+ :CNa.CO,and this may be a t once extended to the metallic derivaiives of othercompounds containing the groups :CH*CO*, :(3H.CN, :CH*NO,, &c.,and, where amines are employed as catalytic agents in promotingcondensations with such compounds, then the ammonium radicle,NR,H*, may be supposed to functionate as a metal.For convenience,however, in the following lines, the univalent metal or ammoniumradicle f unctionating as the positive, labile, polar ” portion (compareAbegg, Bey., 1905, 38, 4112 et S e q . ) will be represented by thesymbol M.The real or virtual existence of C-metallic derivatives OF thecompounds of class K (see p.30) having been postulated, a satis-factory explanation of the acetloacetic ester condensation, as wellas of a large number of other reactions which occur with thecompounds of class K in alkaline media, follows without difficulty.The C-metallic derivatives should exhibit the characters of organo-metallic compounds proper, and a little consideration of the factsshows this to be the case, for the reactions of the metallic derivativesof ketones, and class K generally, are either those of phenols (enolsKETONES CONTAINING THE GROUP 'CH;CO'CH: WITH ESTERS. 33or of organo-metallic compounds, although, in the latter instance,developed to a less marked degree owing to their lower potentialconsequent on the more "negative" character of the organo-radicle (Michael).I n the present instance, the question under consideration is that ofthe reactions of the isomeric M-derivatives with carbonyl and cyano-groups.Organo-metallic compounds proper are characterised by theextraordinary facility with which they form additive compounds withsubstances containing the carbonyl group :>C:O+M*Alk --+ >C<iE,and analogous compounds are formed by addition to *CN, *SO,*, *N:Ogroups.The C-metallic derivatives of ketones, esters, nitriles, and nitro-paraffins behave in the same way :OM>C:O +M*d*CO ++ >C<bsc0 .the change, when it ceases here, being known as the aldol condensa-tion, and being to some extent reversible. The carbonyl compound> C O virtually selects the C-metallic derivative rather than the0-derivative, but merely because the products of its condensation withthe latter :.. J0 .OM >c:o + &foci: . . f3 >c<o.c:d . .are of a type eminently unstable and revert a t once to their generators.The aldol condensation when brought about by bases may thus itselfbe regarded as an instance of the general reaction between organo-metallic compounds and substances of class K, and, moreover, isrecognised as a necessary stage in a very large number of reactionsto which special names have been attached, such as the croton-aldehyde condensation, the Perkin synthesis, and others.The acetoacetic ester condensation is also clearly only a particularcase of the interaction of esters and organo-metallic compounds.Thus one of the latter acting on ethyl formate yields an aldehyde :whilst with t h e C-metallic derivative of a carbonyl compound preciselythe same type of change occurs :X*CO*[b*C:O] + 11-OEt, ..a P-diketone or P-ketonic ester being formed.VOL. XCIII. 34 CLARKE, LAPWORTH, AND WECHSLER : CONDENSATION OFConfusion has probably arisen simply because the product inthe latter instance is, from its very nature, prone to further change,but the substances isolated are exactly those which the compoundX*CO*c'*C:O would yield under the experimental conditions. Withthe latter point, Dieckmann has already dealt (Zoc. cit.), and referencemay be made to his paper, but the following forms a very briefsummary of the possibilities.* *Representing the above Condensation product aswx*co*c*co*z,-irthen1.( a ) If X, W, Y, and Z are all alkyl groups, the compoundis unstable under the experimental conditions, and falls into itsgenerators :WX*CO*OEt + H*b*CO*Z,-iTby reversal, or (6) might break up into two new compounds :'CVX*CO*bH + OEt*CO*Z.ri-2. If either W or Y is a hydrogen atom, then the substance isconverted by the metallic alkyl oxide into a stable metallic derivativeof the enolic form, reversal thus being obviated.3. If Z (or X?) has the structure -CHR*CO*R' or *CHR.CN, tben,even if W and Y are both alkyl groups, a stable metal enolicderivative may be formed and reversal inhibited.Case 1 ( a ) has been dealt with by Dieckmann (Zoc.cit.), but case1 ( b ) has not yet been observed. Case 2 is the ordinary acetoaceticester type of synthesis so fully elaborated by Claisen. Case 3 is thatnoticed by Perkin and Thorpe (Trans., 1900, 79, 736 and 737).The condensations dealt with in the present paper were made withketones containing the groups *CH2*CO*6H and CH2*CO*C':C! respect-tively, the ester with which they were made to react being a nitrousand not a carboxylic ester. It has usually been supposed that, as withother esters, these condense only with ketones which contain thegroup *CH,*CO-, but such is, in reality, not the case. In theinstances we have examined, attack appears to be directed almostexclusively a t the :CH*CO* group.The behaviour of menthone with alkyl formates, on the one handKETONES CONTAINING THE GROUP 'CH;CO'CH: WITH ESTERS, 35and alkyl nitrites, on the other, is very instructive.instance, the initial products may be eitherIn the formerCH,-C( CHMe,)*CHO CH,4CH(CHMe,)dlH, 60 or 6H2 60eHMe*bH, kHMe*kH*CHOOf these, the former is perhaps formed more rapidly than the othera t first, but belongs to a highly unstable type and readily falls into itsgenerators, yielding, by absorption of alcohol, alkyl formate andmenthone, The latter, however, is rendered stable by conversion intothe sodium derivative of the enolic form, which the former cannotyield.With alkyl nitrite, the corresponding forms are :CH,-C( CHMe,)*NO CH,-CH(CHMe2)dlH, 60 and &€, 60~HM&H, dHRle*dH*NOand here the former type does not revert to its generators, owingno doubt to the greater stability of the linking *C*NO comparedwith *C*CO*.As a result, it is at the grouping *CaCO* that re-absorp-tion of the elements of alcohol occurs, and the next products in succes-sion are :. .bCH,--CH( CHMe,) *NOdlHMe*bH, dHMe*bH,CH,-C( CHMe,): No0 H6H2 C0,Et --+ dH, C0,EtThe compound obtained as the product of the reaction, on removingalcohol, neutralising, and extracting, has the properties of an ester, andthis, if boiled with alkalis, is converted into the same hydroximino-acidas is obtained by treatment of menthone with amyl nitrite and hydro-chloric acid, but the yield in the latter instance is comparatively verypoor.The first observation of this apparently anomalous behaviour ofnitrous esters when used in conjunction with sodium ethoxide wasmade by Hantzsch (Ber., 1887, 20, 579; compare also Dieckmann,Be?*., 1900, 33, 579), but referred to a-monosubstituted /3-ketonicesters, and he does not attribute any special significance to his results,probably because such compounds are resolved in so many differentways at the point between the a- and @carbon atoms.The compoundsdealt with in the present paper are simple ketones, and the investiga-tion was undertaken with the object of finding an explanation of theresults obtained when an attempt was made to prepare isonitroso-cyanodihydrocarvone (Lapworth and W echsler, Trans., 1907, 9 1,978and 1919). Here we mere forced to the conclusion that attack tookD 36 CLARKE, LAPWORTH, AND WECHSLER : CONDENSATION OFplace, not at the *CH,*CO.group, but exclusively at the :CH*CO*complex.From pulegone (I), by similar treatment, we obtained an esterhaving the structure (11) :NOH C02Etso that a migration of a hydrogen atom from the side-chain to thea-carbon atom must have occurred at an intermediate stage; in otherwords, a partial conversion of pulegone into isopulegone seems tohave taken place.The constitution of the product was determined by oxidising theacid with permanganate, when it was found to be converted intoformic, acetic, nitrous, and P-methyladipic acids :Other compounds examined in respect to their behaviour with nitrousCHBresters and sodium ethoxide were a-bromocamphor, C,H,,<bo ,which reacted readily, but yielded ordinary isonitrosocamphor, andfenchone which was unchanged, although a variety of conditions wasimposed, and thus further support is offered to the contention ofSemmler (Ber., 1906, 39, 2581) that fenchone contains the groupingc\ c- c G O * c-c./"C/ \cE x P E R I M E N T A L.Action of Amy? Nitvile on Jlenthone in Pszesence of SodiumE'thoxide.I n the first experiments on this reaction, it was found that an oilysubstance only was obtained if the product was isolated merely bydilution and extraction with solvents. This oily material, it masnociced, evolved a considerable quantity of amyl alcohol when treatedwith alkalis or acids, and this led to the surmise that esters werepresent, so that the following process was adoptedKETONES CONTAINING THE GROUP 'CH,'CO'CH: WITH ESTERS. 37Menthone was added to an ice-cold solution of slightly more thanone atomic proportion of sodium in absolute alcohol, and subsequentlya molecular proportion of amyl nitrite was introduced a t such a ratethat the temperature did not rise more than one or two degreesabove zero.After standing in the cold for some hours, the wholewas distilled with the aid of a current of steam until the distillatewas odourless, the resulting liquid being then agitated with a littleanimal charcoal and filtered. Hydrochloric acid was next added tothe cooled solution until no further precipitate formed, the oily sub-stance which separated being allowed to solidify, when i t was removedand crystallisod from methyl alcohol.I n this way, 75 grams wereobtained in a nearly pure state from 85 grams of menthone :0.3102 gave 0.6738 CO, and 0.2644 H,O.0,2153 ,, 15.45 C.C. moist nitrogen a t 14' and 746 mm. N = 7.1.C,,H,,O,N requires C = 59.7 ; H = 9.5 ; N = 7.0 per cent.0.961 required 4'7.1 C.C. N/10 NaOH for neutralisation, whence theequivalent = 204 (calculated = 201).As the properties of the acid agreed closely with those of theoxime of P[-dimethyloctan-E-onoic acid, obtained by the action ofamp1 nitrite on menthone in presence of acid as catalyst, some of thatoxime was prepared by the latter process. No difference between thetwo substances could be detected, and their melting points wereunaltered on admixture.The following new derivatives of the Rcidwere prepared.The p-nitrophenylhydraxone formed a bright yell0 w, crystallinepowder melting at 130" :0.1059 gave 11.9 C.C. moist nitrogen a t 16" and 758 mm. N = 13.1.The semicar6axone was obtained in small, white crystals :0*1208 gave 18.1 C.C. moist nitrogen a t 15' and 773 mm. N= 17.8.CllH2103N3 requires N = 1'7.3 per cent.These compounds were formed on warming the oxime with aqueoussolution, p-n i t rophen yl hydrazine acetate, and semicarbazide acetaterespectively, the hydroxylamine being eliminated with great ease.C = 59.3 ; H = 9.5.C16H2505N3 requires N = 13.1 per cent.Actiolz of Amyl Nitrite on Pulegone in Presence of Xodium Ethoxide.This reaction was carried out in a manner similar to that describedin the case of menthoce, but, as the esters formed appeared to be morestable, the product, some hours before the steam distillation, was mixedwith an excess of strong aqueous potassium hydroxide.The materialwhich separated on acidifying the aqueous residue at the end of thesteam distillation was very gummy, and it was found necessary t38 CLARKE, LAPWORTH, AND WECHSLER : CONDENSATION OFpurify it by dissolving it in ether and extracting the acidic matterfrom this by shaking it with sodium carbonate solution. Afterdissolved ether was removed from the .alkaline liquid, hydrochloric acidprecipitated a viscid mass, which slowly became semicrystalline aftertrituration with acetic acid. The solid portion was the oxime of a newacid, which we propose to term “isopulegonic acid,” in order to indi-cate its near relationship to isopulegone.I t was finally purified bycrystallisation from dilute methyl alcohol :0.2070 gave 0.4604 GO, and 0.1646 H,O. C= 60.7; H= 8.8.0.1538 ,, 9.9 C.C. moist nitrogen at 18’ and 738 mm. N= 7.2.CI,ET170,N requires C = 60.3 ; H = 8.5 ; N = 7.0 per cent,0,4115 required 20.5 C.C. N/10 NaOH for neutralisation, whence theequivalent = 201 (calculated = 199).The compound is readily soluble in methyl or ethyl alcohoI, ether,benzene, chloroform, ethyl acetate, or carbon disulphide, but dissolvesonly sparingly in light petroleum or hot water. It melts at 85’.When heated above its melting point, the compound decomposed,ammonia and an unpleasant smelling vapour being evolved.It reducesa hot ammoniacal solution of silver nitrate, but has no effect onFehling’s solution unless it has been previously heated with a mineralacid, when the product reduces this solution in the cold, a behaviourwhich indioates that the substance is the oxime of a ketonic acid.The compound gave no crystallisable compound on acetylation.Attempts to obtain a specimen of the pure ketonic acid in a statesuitable for analysis were unsuccessful. When it was warmed withhydrochloric acid, the product cooled, and extracted with ether, an oilwas removed in small quantity in which the presence of a ketonic acidwas proved by warming it with aqueous semicarbazide acetate. Thesemicarbaxone separated as a white solid, and, after repeated crystal-lisatian from alcohol, melted at 100’ :0.1060 gave 15.2 C.C.moist nitrogen at 9 O and 757 mm.C,oH160,N*NH*CO*NH, requires N = 17.4 per cent.Unlike the allied oxime of Pc-dimethyloctan-r-onoic acid, thehydroximino-acid from pulegone i s too stable to yield the semimb-azone or p-nitrophenylhydrazone when merely warmed with theacetates of the corresponding bases.N = 17.2.Oxidation of the Oxime of isoPulegonic Acid.As it was found impracticable to isolate pure pulegonic acid, theexperiment of oxidising the pure oxime itself was undertaken. Twelvegrams of that compound were dissolved in a solution of sodiumcarbonate, and to the ice-cold solution was added very gradually a2 per cent. solution of potassium permanganate. The colour of thKETONES CONTAINING THE GROUP 'CH,'CO'CH: WITH ESTERS. 39latter at first disappeared instantaneously, and the addition of theoxidising agent was continued until an appreciable interval elapsedbetween its introduction and the disappearance of the pink colour,which was the case when about 18 litres had been used.The liquidwas then freed from manganese dioxide in the usual manner, evaporatedto a bulk of about 100 c.c., mixed with carbamide, acidified withsulphuric acid, and distilled in a current of steam,(a). Volatile Acid Products.-The odour of fatty acids being percep-tible, three-fourths of the distillate was neutralised with normal sodiumhydroxide, mixed with tho remaining fourth, and the whole evaporated.The residue had the appearance of sodium acetate, and gave an esterhaving the odour of ethyl acetate on treatment with sulphuric acid andalcohol; i t was therefore redissolved in water, mixed with silvernitrate solution, and the silver salt which separated was washed, dried,and analysed :0.3066 gave 0.1981 Ag, whence the equivalent of the volatile acidThe solution from which the silver salt had been removed blackenedconsiderably on standing, a fact indicating the presence of a smallquantity of formic acid.(b).Non-uoZcctiZe P.rocZzccts.-The liquid from which the volatile acidshad been removed was acidified, extracted repeatedly with ether,which was afterwards washed with a little water, dried, andevaporated, the residue being freed from alcohol by frequent evapora-tions with water.A semi-solid mass was 6nttlly left which was foundto yield a considerable quantity of a sparingly soluble copper salt, andthe whole was therefore dissolved in water, neutralised with ammonia,and mixed with copper acetate, the precipitated copper salt beingremoved, washed with water, and decomposed with hydrogen sulphidein the usual way. In this manner, a semicrystalline material wasobtained which was freed from adherent oil and crystallised repeatedlyfrom ethyl chloride and light petroleum :was 60.1, the equivalent of acetic acid being 62.0.1511 gave 0.2920 CO, and 0.1047 H,O.0.1232 required 15.7 C.C. B/10 NaOH for neutralisation, whence theequivalent = 78.4, whilst a dibasic acid, C7H,,04, has theequivalent 80.0.The substance was readily soluble in most of the organic solventswith the exception of light petroleum, and separated in slender needlesmelting at 84-85'.It did not yield an anhydride when heated a t200', and had all the characters of P-methyladipic acid.To confirm the production of formic acid during the above oxidationa specimen of the pure oxime of isopulegonic acid was boiled with diluteC = 52.7 ; H = 7.7.C7H1204 requires C = 52.5 ; H = 7.5 per cent40 CONDENSATION OF KETONES WITH ESTERSsulphuric acid and potassium dichromate, and the distillate, which hada faint odour of formaldehyde, collected. The presence of form-aldehyde was confirmed by adding hydrochloric acid and phloroglucinolto a portion of the liquid, when a pink colour developed. Theremainder of the distillate was neutralised with sodium carbonate,evaporated, and tested for formate with ammoniacal silver nitrate andwith mercuric chloride, and in both cases a positive result was obtained.No appreciable quantities of products other than nitrous, formic,acetic and P-methyladipic acids could be detected as oxidationproducts.Action of Amy1 Nityite on a-Bromocanaphor in Presence of SodiumEthoxide.When a-bromocamphor was subjected to treatment in the mannerdescribed in the case of menthone, rapid action occurred, and potassiumbromide separated in considerable quantities. The product was dis-tilled in a current of steam, when much unchanged bromocamphorpassed over, and the residue, which was yellow, gave a solid pre-cipitate on neutralisation. This was collected and crystallised fromalcohol, when it was found to melt at 152-154'; it was insoluble insolutions of alkali carbonates, although freely soluble in sodiumhydroxide, and was therefore not a carboxylic acid. It containednitrogen, gave camphorquinona when boiled with formaldehyde andhydrochloric acid, and had all the other characteristics of ordinaryisoni trosocamp hor.Fenchone did not react with alkyl nitrites under the conditionsadopted in any of the experiments above described.Much of the cost of the investigation was defrayed by a grantawarded by the Research Fund Committee of the Chemical Society,for which we wish to express our indebtedness.CHEMICAL DEPARTMEKT,GOLDSMITHS' COLLEGE,NEW CROSS, S.E