April 5, 1953
1653
CYCLOBUTANEDIONES AND THE LOWERALKYLKETENE DIMERS
[CONTRIBUTION FROM THE CHEMICAL
LABORATORIES OF MCGILL UNIVERSITY
AND
THE JOHNS HOPKINSUNIVERSITY ]
Cyclobutanediones and the Lower Alkylketene Dimers'n2 BY EVANSB. REID AND STEPHEN J. GROSZOS~ RECEIVED JUNE 20, 1952 The Schroeter synthesis of the 2,4-dimethyl and the 2,4-diethyl derivatives of 1,3-cyclobutanedione has been repeated and the products have been studied intensively. Fairly close agreement in physical properties between our methyl derivative and that of Schroeter was noted, but our ethyl derivative demonstrated a n elevated melting point, and, in comparison with the methyl analog, such great instability as to cast doubt upon its proposed cyclobutane structure. Ethereal solutions of the lower alkylketenes were prepared by the Staudinger process, and it was shown by direct comparison that both methyland ethylketenes, but not isopropylketene, dimerize to give cyclobutane products. The lactone dimers obtained by the Wedekind dehydrohalogenation of acyl halides were shown t o be absent under the conditions of our experiments, except for isopropylketene. Finally, some evidence is advanced t o show that the solid cyclobutane dimers are terminal rearrangement products of liquid polymers of as yet unknown constitution.
In the course of his work on alkylketenes Staudinger4 obtained a solid compound, presumed to be a dimer, from an ethereal solution of methylketene that was allowed to stand. Since the substance was enolic i t was assigned the cyclobutene structure I1 (Fig. 1). The only evidence substantiating the proposed structure is the abbreviated statement by Schroetefl that there was no depression on taking the mixed melting point of the ketene polymer with a compound which he had shown vzk circumstantial evidence696 to be the cyclobutanedione I. In reinvestigation~~?~ of this synthesis it has been shown conclusively that dimethylcyclobutanedione (I)7 is formed, albeit through the very unusual intermediate 3,5-dimethyl-2-hydroxy-6-ethoxyy-pyrone (IV), rather than the postulated5J cyclobutane ester V. In our work with the Schroeter synthesis6 i t was noted that the pyrone IV (R, methyl), and the 2,4dimethyl-l,3-~yclobutanedione (I) derived from it, not only melted somewhat higher than the reported values,6 but in addition I demonstrated a stability much beyond that indicated by the original author.6 On the other hand, the ethyl analogs, though melting much higher than reported, were much less stable than the methyl derivatives. After we had established, Via elaborate purifications, that this behavior was not due to impurities, a critical examination was then made of the condensation product of ethyl a,a'-diethylacetonedicarboxylate in an effort to uncover evidence at variance with either the pyrone or the cyclobutanedionecarboxylic ester structures (IV or V). This was not, a t first, successful; hydrolysis with boiling water furnished carbon dioxide and ethyl CY-butyrylb~tyrate~; alcoholic potash at room temperature was without action'; the Zeisel determination revealed one (1) For the first paper in this series see E. B. Reid, THISJOURNAL. 78,2853 (1950). (2) Most of the material in this paper was presented before the Organic Division of the Buffalo Meeting of the American Chemical Society, March 24,1952. (3) Part of this paper is taken from the doctoral dissertation of Stephen J. Groszos, The Johns Hopkins University, Baltimore, Md., 1951. (4) H. Staudinger, Bcr.,.44, 533 (1911). (5) G. Schroeter, ibid., 49, 2697 (1916); compare J. D. Roberta, R. Armstrong, R. F. Trimble, Jr., and M. Burg, THISJOURNAL, 71, 848 (1949), footnote 8. (6) G. Schroeter and C. Starsen, Bcr., 40, 1604 (1907). JOURNAL, 13, 1297 (7) R. B. Woodward and G . Small, Jr., THIS (1960). (8) These are described in the experimental part. (9) 0.Schroeter, ref. 5, had &own that hydrolyaia of the methyl analog g w e carbon dioxide and ethyl a-propioaylpmpionate.
R
.>C-c=o
O=c-c(,
)c-c H I I I o=c-c
I
I1
H I I H
o=c R
/OH
R '
R
c=c I11
R )c-cRO
)c=c(
H
I
I/R
O=C-C(
OH
v
IV
COOCnH6
Fig. 1.
ethoxy group; and although the Kurt-Meyer enol titration was negative, methylmagnesium iodide indicated the presence of one active hydrogen with additional Grignard carbonyl reaction to the extent of 80%. We repeated the carbethoxylation6 of the condensation product with ethyl chloroformate, obtaining presumably the 2-carbethoxy pyrone derived from IV, a substance of considerably lower boiling point than that reported by the German author. We do not at this time regard this difference as being of great significance1l0since our product duplicated the behavior of Schroeter's in regenerating the original condensation product under the catalytic influence of sodium ethoxide. Chemical proof was finally secured for the pyrone structure (IV) of the acidic condensation product through treatment with diazoethane. The basic pyrone I11 (R, ethyl) was obtained; I11 was also isolated as the acid-soluble, base-insoluble fraction from the sulfuric acid condensation of ethyl CY, a'-diethylacetonedicarboxylate when carried out in the known manner.7e1.6 No doubt attaches to the structure of 111,11and its identity with the ethylation product would seem to eliminate definitely the cyclobutanedionecarboxylic ester from consideration (compare 7 and 1). Physical evidence substantiating the chemical proof was obtained from the infrared absorption spectral2 of the two related pyrones, viz., 3,5-dimethyl-2hydroxy-6-ethoxy- y-pyrone and 3,5-diethyl-6ethoxy-2,4-pyronone. While the curves are similar, certain important differences appear. In the (IO) The possibility exists, and is under investigation, that the acidic sulfuric acid condensation products obtained by Schroeter, et ul.,*p' were in fact the cyclobutanecarboxylic esters (V) as originally formulated, and not the isomeric pyrones (IV) that have been obtained by Woodward and Small,' and by Reid,l and in the present work. (11) G. Schroeter, Bcr., 59, 973 (1926). (12) Our thanks are due Dr. Lester P. Kuhn of the Ballistic Laboratory, Aberdeen Proving Grounds, and Mrs. Marjorie H. Melville of The Johns Hopkins University, for their assistance in these determinations.
EVANSB. REIDAND
1R,jR
STEPHEN
J. GROSZOS
Irol. '73
case of the ethyl pyrone no OH absorption is ob- molecular weight determinations showed the subserved a t about 3 p (3400 an.-'), in contrast with stance to be a dimer of methylketene, and direct the wide band shown in this region by the acidic comparisons Via physical and chemical tests proved methyl pyrone. Of interest also is the strong that the substance possessed structure I. Not carbonyl absorption for both pyrones, occurring only were the infrared absorption spectral2 of both at 5.8-6.0 p (1700 em. -l), excluding the possibilities specimens virtually superposable, but the typical of either conjugated chelated structures or ex- bands ascribed to the OH frequency13 of either The ethylpyrone associated or non-associated hydroxyl groups were ternally hydrogen-bonded OH. is thus the non-enolic analog of the methyl deriva- absent. The substance thus appears to be diketonic in the solid state and this was further inditive described earlier. Turning now to the question of the diethyl- cated by the position and sharpness of the carbonyl cyclobutanedione, I, we found that our ethyl- band a t 5.8 p. The spectra also show a maximum pyrone underwent saponification and decarboxyla- at 10.9 p, which is of interest in connection with the tion with hot baryta to yield a crystalline strongly cyclobutane structure of the substance since it has acidic enol that was empirically identical with been reported recently" that cyclobutane hydroSchroeter's product, although its melting point was carbons exhibit a characteristic absorption a t much higher. Again, various tests were carried 10.9-11.0 p, ascribed to the methylene rocking out to establish either the identity or non-identity frequency. Likewise the infrared spectra of the of our substance with his. These did not, however, dioximes of the two specimens were identical, yield decisive information. Thus, hydrolysis read- and gave a doublet a t 10.6 and 10.9 p , probably due ily cleaved the molecule to yield carbon dioxide and to ring stretching and methylene rocking respecdi-n-propyl ketone (compare the dimethylcyclo- tively. l8 The X-ray diffraction patterns of both butanedione, I).6 Although the properties of specimens were obtained by the powder method, derivatives seemed to indicate complete similarity and since the lines may be superposed one upon the between the two, our compound failed to form other, it follows that the two samples are composed either a mono- or a dioxime, and in addition proved of the same substance and are not dimorphic modiinert to hydrogen in the presence of Adams cat- fication~,'~ despite our inability to obtain them in alyst. Further, attempted ozonization merely identical crystalline form. caused rapid decomposition of the compound, Although I appears to be diketonic in the solid apparently to the same oily product formed on state, in solution i t is substantially ionic. This keeping the solid under ordinary conditions for was made apparent from a study of the effects of about ten days. Iri the Grignard machine a speci- concentration and solvent upon the ultraviolet men behaved abnormally, showing 1.3 active absorption spectra of the compound.20 In ethanol, hydrogens with 1.7 additions to carbonyl groups.'* using concentrations of 4, 10, 20 and 200 X Since it appeared impossible either to duplicate M , a set of intersecting curves was obtained, giving Schroeter's synthesis of 2,4-diethyl-1,3-cyclobu- the isosbestic point21a t X 244 mp, log E 4.01. The tanedione or to prove by degradations the cyclo- non-applicability of Beer's law, shown by the shift butane structure of our product, our attention of maxima to longer wave lengths with decreasing turned to the preparation of the dimers of methyl- concentration, and the isosbestic point clearly and ethylketene, for such would make possible show the presence of an equilibrium between two direct comparisons of (a) the dimer of methylketene species, each of which obeys Beer's law. Changing with the known7 2 &dimethyl- 1,3-cyclobutane- from ethanolic to aqueous solution also caused a dione, and (b) the dimer of ethylketene with our shift of maxima toward longer wave lengths, high-melting derivative. An apparatus was de- but the further change from water to dilute sodium signed (see Experimental part) that successfully hydroxide occasioned very little shift of the maxima. overcame experimental diffi~ulties,'~ and furnished This, together, with the fact that only a slight ethereal solutions of methyl-, ethyl- and isopropyl- displacement toward shorter wave length occurred ketenesL6 From methyl- and ethylketene solu- on going from ethanol to dilute hydrochloric acid tions by appropriate treatment were obtained indicates that the enol is mainly ionic in hydroxylic both solid and liquid polymers; isopropylketene solvents. solutions failed to furnish solid material. Solid Ethylketene Dimer.-As was found with Solid Methylketene Dimer.-The solid product the solid methylketene dimer, analysis, mixed was obtained from solutions that had been allowed melting points, infrared and ultraviolet absorption to stand. It was obtained in the form of a micro- spectra, and X-ray diffraction patterns clearly crystalline powder, and attempts to crystallize it showed the solid ethylketene product to be idenin the form of long needles identical with those tical with 2,4-diethyl-l,3-~yclobutanedione as preformed by I were without success. Nevertheless (17) J. M. Derfer, E. E. Pickett and C. E. Boord, THISJOURNAL, (13) Compare R. S. Rasmussen, D. D. Tunnicliff and R. R. Brattain, T ~ I JOURNAL, S 71,1068 (1949); also E.B. Reid and W. R . Ruby, rbid., 7 5 , 1054 (1951). (14) Dimedone also behaves abnormally. See ref. 7. (15) C. D. Hurd, F. W. Cashion and P. Perletz, J . Ow. Chqm., 8, 367 (1943), report unsuccessful attempts to obtain ketene or methylketene by the Staudinger method (16) Isopropylketene has not heretofore been prepared H. Staudinger and H. W Klever, R i r , 41, 906 (InOS), report the preparation of an ethereal solution o f r t h s l k e t e n e hut n n polymeric prodiicts rchtld
Ilr:
