OBSERVATIONS ON THE CLAI!3EN REACTION. 2891 CCCC1.-Observations on the Cluisen Reaction. By GILBERT T. Momm and EUSEBIUS Horns. UNTIL recently the condensation of the higher methyl monoketones with esters under the influence of sodium had only been inve&igated up to the homologues containing the radicals n-hexyl aad n-nonyl. The present investigation wsts undertaken to fmd out where and if possible why the reaction stopped when applied to methyl ketones containing also one of the higher alkyl radicals. The nomenclature throughout this paper is based on the following condemation with ethyl acetate : It is shown that under the conditions stated the reaction proceeds quite normally from 12. = 7 to n = 19 a satisfactory yield of 1 3-diketone being obtained in every case.The general method of procedure has been to convert the cor-reaponding fatty acid into its barium salt and to distil this with three molecular proportions of anhydrous barium acelate in a specially constructed flat vacuum pan (Morgan and Holm- J . Soc. C h . I d . 1925 44 108~). The crude product wa8 p d e d and subjected to the Claisen condensation with sodium and ethyl acetate the sodium salt of the enolic form of the @-diketone being treated with acetic acid and the liberated diketone precipitate 2893 MORGAN ANT HOLMES : and purified aa its co-ordinated copper derivative. The acids C&H,+,*CO$€ where n = 10,12,14 18 were obtained by oxidising the methyl ketone from the next higher naturally occurring acid, to give acetic acid and the required acid from which mixture the lrttter wa8 isolated as its relatively insoluble barium salt.The copper mlts of the @-diketones are blue as are the lower members of the series but the colour becomes less pronounced as the homologous series is wended. The diketones from n = 7 to 19 have been obtained as solids a t the ordinary temperature, whilst the monoketones from ?a = 10 upwards are colourless solids with progressively higher melting points. Since this work originally arose out of the researches on cyclic kllurium compounds of bactericidal potency it is of interest to note that Mr. C. J. A. Taylor (this vol. p. 2615) has found it possible to prepare cydotelluropentanedione dichlorides reducible to bacteri-cidal cydotelluropentanediones in the case of the following diketones, all of which are described in the present paper n-octoylacetone, n-nonoylacetone and n-duodecoylacetone (lauroylacetone).These are the only ones so far condensed with tellurium tetrachloride, but it appears probable that the whole of the series woiild give similar results. E X P E R I M E N T -4 L. In the case of the lower members of the series n = 9 and below, the procedure was as described below in the case of n-octoylacetone. For the higher members three to nine mob. of the ester were employed no preliminary coohg being necessary. After heafing under reflux for from 4 to 5 hours the mixture was poured on to ice as before. The mixtures were then neutralised with acetic acid and satur-ated cupric acetate was run in. In most cases the copper salts of the diketones were precipitated a t once but for the higher ones 72 = 15 and upwards it was found necessary to add alcohol to bring the reagents together.The free diketones were obtained by treating the copper salts with dilute sulphuric acid in the presence of ether which was subsequently removed. The lower diketones were purified by vacuum distillation the higher ones by crystallisation from alcohol. n-Octoylacetone.-n-Heptyl methyl ketone (35 g.) obtained by the dry distillation of barium n-octoate with barium acetate (3 mols.) was added to ethyl acetate (2.5 mols.) the mixture cooled to O" and sodium (1 atom.) added in the form of thin slices. After 12 hours the mixture was heated under reflux for 3 hours allowed to cool and poured on to ice.The liquid was then made ver OBSERVATIONS ON THE CLAISEN REACTTON. 5893 slightly acid with dilute acetic acid and the oily layer of liberated diketone precipitated with a saturated solution of cupric acetate. After being stirred a t intervals for 4 hour the precipitate was collected and washed with a little water and cold alcohol. A yield of 26 g. represented 4743%. T w o subsequent batches gave yields of 41.0 and aS*6% respectively. Crystam successively from alcohol and benzene the ~olvper salt w&s obtained in pale blue needles and melted a t 118". The salt (0-5 g.) was shaken with a few C.C. of dilute sulphuric acid in the presence of a little ether. The acid layer was separ-ated and a water washing of the ether layer added. The copper was precipitated by means of caustic potash and weighed as the oxide (Found Cu 14.6; C 61.6; H 8.85.C,,H,804Cu requires Cu 14.8; C 61.45; H 8.8%). The n-octoylacetone recovered from the copper salt had the characteristic d o u r of the p-diketones of this series and boiled at 248"/755 mm. and 118"/5 mm. n-NonoylQcetone.-A quantity of the corresponding ketone from a reputed pure specimen of barium pelargonate (23 g.) was added to 37 g. of pure dry ethyl acetate and 3.8 g. of sodium were added to the cooled mixture. The copper salt of a diketone was obtained by the general method and weighed 26.3 g. a yield of 68.8%. Recrystdised from benzene it melted at 107" and was of the expected pale blue colour (Found Cu 13-9. C,E,,O,Cu requires Cu 13.9%). Since the above melting point was not in accord with those of the other members of the series and furthermore other dis-crepancies arose when the diketone was condensed with tellurium tetrachloride another specimen of pelargonic acid was obtained and the reactions were repeated.Again the copper salt of a diketone waa obtained purified and analysed (Found Cu 13-75; C, 63.05; H 9-4. Theory requires Cu 13-9; C 62-9; H 9.2%). "his specimen melted at 11,5.5" a value bringing it in line with its homologues. It appears probable that the earlier specimen of acid contained an acid isomeric with pelargonic acid but having a branched chain remote from the carboxyl group. The n-nonoylacetone recovered from the copper salt boiled a t 150"/15 mm. n-Decoylacetone.-This diketone has been described (Morgan and Holmes J.1924 125 760; J . P k m . Soc. June 14 1924). n- UndecoyEarcetone.-n-Decyl methyl ketone obtained from the corresponding undecoic acid which was itself prepared by the ketone and acidified dichromate degradation of lauric acid, was condensed with sodium and ethyl acetate (6 mob.). Th 2894 MORGAN AND HOLMES: yield of copper salt was 3.9 g. (70%). After crystallisation from benzene the salt melted at 112" (Found Cu 12.4; C 65-6; H, 10.0. C,,Hm04Cu requires Cu 12-4; C 65.4; H 9.7%). The free diketone was obtained as a colourless solid melting at 28". n-Duodeuylucetone (Lauroyhcetone).-Three separate batches of the monoketone obtained from lauric acid were condensed with 3, 6 and 9 molecular proportions of ethyl acetate respectively the period of refluxing being extended to 4 hours.The yields of copper diketone from 5 g . of ketone were 2.6 g. 3.45 g. (50.2%), and 3-40 g. respectively. It appears that the optimum yield of rather more than 50% was reached by using 6-8 molecular pro-portions of ester in the condensation. The pure salt melted at 112.5" (Found Cu 11.7; C 66.6; H 10.3. CsH,O,Cu requires Cu 11.7; C 66.5; H 10.0%). The free diketone wm a colourless solid of melting point 31-32'. n-Tridecoyhetone.-Myristic acid was converted into the barium salt and -distilled with barium acetate. The resulting methyl ketone was then oxidised with acid sodium dichromate to give the next lower acid which on subsequent &illation of its barium salt gave the duodecyl methyl ketone. This (5 g.) was condensed with ester fo give 3.7 g.of the copper dikdone a yield of 54%. The pure salt melted at 111" (Found Cu 11.2. C3,H5,04Cu requires Cu 11.2%). The pure diketone obtained from the copper derivative was a colourless solid melting at 35". It had scarcely any d o u r and gave the red ferric coloration with alcoholic ferric chloride only on warming (Found C 75.4; H 11.7. C1,HSO2 requires C 75-6; H 11.8%). n - T d r w l a c e t o n e (Nyristqyklcetone).-Five g. of the ketone obtained from myristic acid were subjected to the Claisen reaction, giving 3-42 g. of copper salt (yield 52%) which after crystallisation from alcohol and from benzene melted at 112" (Found Cu 10.5. CMHs2O4Cu requires Cu 10.6%). The free diketone melted at 39' (Found C 76.1; H 12.3.C1,H3202 requires C 76.1; H 11.9%). n- Pentcsdecoy laceton e .-n - Tetr adec yl met h y 1 ketone was prepared from palmitic acid by the general degradation process. Condensed with ethyl acetate it gave 2.25 g. of the copper diketone a yield of 35%. The purified salt melted at 111" (Found Cu 10-5. C,,H6,0,Cu requires Cu 10.2%). The colourless diketone melted at 42" (Found C 76.4; H, 12.0. C18Ha02 requires C 76.6; H 12.1%). n-HexadeciyZmetone (Pdmitoykrcetone) .-n-Pentadecyl methy OBSERVATIONS ON !lTiE CLBISEN REACTION. 2895 ketone (5 g.) obtained from p h i t i c acid wa.~ condensed with ethyl acetate (7 mols.) by the general method the time of refluXing being raised to 5 hours. The copper salt (yield 2.2 g.; 34%) after purification melted at 112" (Found Cu 9.6.C38Hm04h requires Cu 9.7%). The decompoaifion of the copper salt by dilute sulphuric mid in presence of ether took an excessive time. Accordingly in this cam and in the case of all the higher members of the aeries the copper derivative was suspended in water an equal volume of concentrated acid added and the mixture allowed to cool before being extracted with ether. The free diketone cryshllised from alcohol in colourless plates, m. p. 49" (Found C 76.8; H 12.3. C19H3602 requires C 77.0; H 12.2%). n-Heptadeuqhetone (Margaroyhetone).-The condensation of the ketone (5 g.) obtained from margaric acid was carried out with 3 6 and 9 mob. of ester respectively the time of refluxing being 5 hours. The yields of copper salt were 0.35 g.1.7 g. (27%) and 1.4 g. and of recovered ketone 4.0 g. 2.5 g. and 2-9 g. respectively. In each case a small amount of copper margarate was produced. The copper salt of the diketone was washed with cold ether. The unchanged ketone was then separated from the copper margarate in the waahings by acidification treatment with barium acetate, and filtration of the insoluble barium salt. The pure copper salt melted at 112" (Found Cu 9.3. C,H,,O,CU requires Cu 9*3y0) and the free diketone at 51" (Found C 77.3 ; H 12.3. n-C&oyluc&tone (Stmroyhetone) .-In a repetition of previous work 10 g. of n-heptadecyl methyl ketone were condensed with ethyl acetate (35 mols.) and sodium and 0.75 g. of copper diketone waa obtained on ad- cupric acetate and alcohol to the reaction mixture (yield 6%).With 8 mols. of ester 6.9 g. of copper salt were obtained (yield 55%). No diketone copper salt was ever obtained until alcohol was added to the reaction mixture this accounting for some of the earlier negative results. The purified salt melted at 113" (Found Cu 9.0. C,,H,,O,Cu requires Cu, The free dikdone crystallised from alcohol in colourless plates, m. p. 52-5" and gave a red coloration with alcoholic ferric chloride on warming (Pound C 77.5; H 12.4. C,1Ha02 requires C, 77.8; H 12.35%). n-~70nabdecoylctcetone.-)~-Octadecyl methyl ketone wa8 obtained from the C acid by the general degradation process. Condensing with ethyl acetate gave a 320/ yield of a copper dikdone some C,H,,O requires C 77.4; H 12.3%). 9*OY0) 2896 HINSHELWOOD AND BURK : unchanged ketone being recovered.The pure salt orysfallised from benzene melted a t 112.5" (Found Cu 8.6. CMHWO4Cu requires Cu 8.6%). The recovered dikane consisted of thin plates m. p. 55" (Found : C 77.9; H 12.5. n-Eicosan~~etone.-Pot~ium erucate wits fused with caustic potash giving acetic acid and the normal C acid eicosanic acid. After purification this was converted into its barium salt which C22H4i02 requires C 78-1 ; H 12.4%). F I G . 1. 360 220 t 180 h 140 loot was then distilled with -barium acetate. The wnonadecyl methyl ketone obtained was condensed with ethyl acetate in the general manner to give the corresponding copper derivative 5 g. yielding 1-75 g. (28.4%). Crystallbed from benzene the salt melted at 114" (Found Cu 8-15. C4,H,,04Cu requires Cu 8.3%). The free diketone melted a t 57" (Found C 78.4; H 12.6. C&HMO2 requires C 7 8 4 ; H, In the diagram curve I shows a noteworthy change in the fusi-1 2 -5 yo ) * bility of copper p-diketones with lengthening of the carbon chain. Curves I1 and I11 indicate the boiling points of the lower diketones and the melt4ing points of the higher diketones respectively. The authors desire to thank the Advisory Council for Scientific and Industrial Research and the Birmingham University Research Committee for grants which have helped to defray the expense of this investigation. 0 4 S 12 16 20 n in CnH,+,.CO.CH,.CO.CH, UN~VERSITY OF BIRMINGHAM, EDGB-4STON. [Received il-ovember 4th 1925.