lS9O
FOILA. MILLERAND STANLEY D. KOCH,JR. [CONTRIBUTIOX FROM THE S O Y E S CHEMICAL
LABORATORY, I‘SIVERSITP
Vol. 70 OF ILLINOIS]
Diketene: Infrared Spectrum and Structure BY FOILA. MILLER*AND
STANLEY
D. KOCH,J R . t
strong bands are found in the double bond stretching region (1500-2000 cm. -l), whereas each of the postulated forms would have only two fundamentals there. (Form I may be an exception, with only one band in this region. The ketene CHj-C-CII=C=O CHl=C-CH, group in this molecule probably would give one I .1 0--c=o 0 band above this range and one below it.) The reI I1 maining three (or four) bands must then be exAcetylketene Vinylaceto-P-lactone plained either as combination tones, or as due to the presence of more than one form. The ultraCHj-C=CII o=c-crI, violet absorption maximum at 3130 A ( E = 2) I0-c=o I HjC--C=O I ! would seem to favor 111.9 (This too involves an 111 IV upper electronic state, however.) Electron difp-Crotonolactone Cyclobutanedione-1,3 fraction studies are said to be compatible with HOC=CH I1 or 111, incompatible with I and IV.’O UnforI I tunately none of these data provide unambiguous HzC-C=O proof for any one of the possible structures. Bev cause of this it has been suggested several times2s8 Mono-enol of IV that there may be two (or more) forms in equiThe present status of the problem has been well librium. summarized by Boese’ and by Hanford and It occurred to us that the existence of such an Sauer.2 Chemical reactions appear to have been equilibrium, Form A $ Form B, might possibly of little help in elucidating the structure, for some be demonstrated by measuring the infrared specindicate one form and some another. It is very trum as a function of temperature. The van’t difficult to evaluate such evidence critically be- Hoff equation, d In K / d T = AH/RT2, indicates cause of the considerable possibility of rearrange- that if A H is sufficiently large the equilibrium conments. Since the problem is to establish the stant will change appreciably with temperature. structure of the ground electronic state of di- The resulting alteration in the composition of the ketene, it seems undesirable to argue from evi- equilibrium mixture may then be evidenced in the dence that involves also an unknown activated infrared spectrum. Conversely, however, a specstate. One is therefore inclined to discount such trum which changes with temperature does not evidence and to rely on physical measurements indicate the existence of an equilibrium unless made only on the ground state if possible. These, other temperature effects are excluded. Two however, have also led to contradictions. The such effects come to mind immediately. (a) An non-zero dipole moment3**eliminates the possi- irreversible chemical change. This possibility bility that the structure may be symmetrical form can be eliminated if the changes in the spectrum IV alone. The Raman spectrum has been ob- are found to be reversed when the sample is tained for the l i q ~ i d , ~and ~ ~the ~ ’ infrared spec- brought back to its original temperature. (b) Altrum for both liquid and solutions.s The absence tered population of the energy levels in accordance of any 0-H stretching frequency eliminates V, with the Boltzmann factor. For a moderate rise and the many coincidences between the Raman in temperature the increased population of the and infrared spectra eliminate again the existence higher vibrational levels will increase the intensity of form IV alone. A further decision between the of difference bands, but should not appreciably remaining forms I, 11, and I11 has not yet been affect the relative intensities of fundamentals. made from the data. It is noteworthy that five Hence a very few weak bands may become * Present address: Mellon Institute of Industrial Research, markedly intensified a t higher temperatures, but Pittsburgh 13, Pa. t Present address: Department of Chemisuy, Cornell Univermty, they should be calculable as difference tones. Ithaca, N. Y. The effect of increased temperature on the ro(1) Boese, Ind. Eng. Chem., 92, 16 (1940). tational fine structure is one of broadening and (2) Hanford and Sauer, “Organic Reactions,” Vol. 3, John Wiley flattening the rotational branches. If the width and Sons,Inc., New York, N. Y , 1947, pp. 127 ff. (3) Oesper and Smyth, THISJOURNAL, 64, 768 (1942). of a vibration-rotation band of an asymmetric (4) Hurdis and Smyth, i b i d . , 66, 89 (1943). rotator is defined as the separation of points on the (5) Angus, Leckie, LeFevre, LeFCvre and Wassermann, J CLem The structure of the ketene dimer, (CHz=C= O),, has been in question ever since its discovery in 1908. Five possible forms have been seriously considered a t various times:
‘
SOL.,1751 (1935). (6) Kohlrausch and Skrabal, Proc. Indian Acad. Sci., EA, 424 (1938). (7) Taufen and Murray, T H I S JOURNAL, 67, 754 (1945). (8) Whiffen and Thompson, J . Chcm. Soc., 1006 (1946).
(9) Calvin, Magel and Hurd, THISJOURNAL, 69, 2174 (1941). (IO) Private communication from Bauer, Bregman and Wrightson to Haaford and Sauer, as reported in reference 2, p. 218, footnote. See also Abstracts of Papers, 109th meeting of American Chemical Society, April, 1946.
May, 194b
INFRARED SPECTRUM AND STRUCTURE O F
DIKETENE
1891
band envelope where the intensity has a given value, then the width of the band is proportional to the square root of the absolute temperature." Thus in going from 30 to 180' the band width will increase by a factor of about 1.2. This should not be a confusing effect, as we shall be looking for more marked changes in the spectrum. Experimental
permanent gas was present in the system. This impeded the diffusionof vapor back into the reservoir. We believe that this gas was air which degassed from the system during heating, for it gave no infrared absorption. Another experiment was performed for the purpose of locating very weak bands. In this case a dish of liquid diketene was placed within the smaller or source housing of the spectrometer. This provided a path length of 4 J cm. a t a vapor pressure of about 13 mm. Finally, the spectra of solutions of diketene in carbon tetrachloride and carbon disulfide were obtained to get a Diketene was prepared by the dimerization of ketene comparison between our samples and those used by Whifaccording to the method of Williams and K r y n i t ~ k y . ' ~ . ~fen ~ and Thompson. The spectrum was measured from 3 to 15 w with a Model Results 12-B Perkin-Elmer infrared spectrometer equipped with rocksalt optics, a General Motors breaker-type amplifier, In brief, the spectrum was found to change and a Brown recorder. A 5-cm. Pyrex cell was used for the vapor. The rocksalt windows were sealed on with markedly with temperature, and these changes glyptal resin, which was found to hold well a t tempera- were reversed on cooling. tures up t o 180' if one heats or cools rather slowly t o avoid Table I summarizes our results a t 30") and gives cracking the seal. The cell was heated in a small oven for comparison the data reported by Whiffen and consisting of a sheet metal cylinder closed with endplates of transite that were held together by four tie rods. A ThompsonBfor the pure liquid and for solutions. window was cut in each piece of transite and closed by The agreement is reasonably satisfactory. We clamping over it a polished rocksalt plate. A suitably observed the same change in the relative intensiinsulated electrical heating element was wrapped around the metal cylinder. The cell was connected by glass ties of certain bands in the 1700-1900 cm.-l region tubing to an external sample reservoir immersed in a water upon change of solvent that was noted by these bath. The connecting tube had a small U-type mercury authors. Our bands in the 3000 cm.-l region are manometer sealed t o it so that pressures in the range 0-20 accurate to only about f15 cm.-' because of poor mrn. could be read. The primary purpose of the manometer was to indicate any failure in the glyptal seals. dispersion, but the spectrum was scanned to 3600 The connecting tube and manometer were wrapped with a ern.-' to confirm the absence of hydroxyl groups spiral of resistance wire and heated electrically to a temper- in the vapor. It is noteworthy that the very inature sufficiently high so that no condensation occurred in tense band of ketene monomer a t 2153 cm.-' is not them. In this manner the reservoir was made the coolest part of the system. The pressure of the vapor was thus present in any of our spectra a t any temperature. It is necessary ta point out at this time that the determined by the temperature of the liquid reservoir, whereas the temperature of the vapor in the beam was intensities reported in this paper are only qualidetermined by that of the oven. tatively correct for the following two reasons. Spectra were measured a t vapor temperatures of apFirst, our procedure was to record a radiation proximately 30, 60, 100, 140 and 180". Experimental curve through an empty cell, and then to use this conditions for the various determinations were No.
1 2 3 4 5 6
Temp. of reservoir, OC.
28 29 34 35
< -50
r
I
8
