6337
ergy-transfer quenching than the low-lying rubrene triplet. A second marked difference in this system’s behavior, as compared to that of rubrene in benzonitrile, is that none of the luminescence parameters shows a dependence on the ion generation sequence. The dependence seen in the pievious case was thought to result from ineffectual quenching of the rubrene triplet by its ion radicals, and it was further supposed that control of 7’s magnitude was partially defaulted to anodically generated foreign substances. In the present case, the anodic potential excursions were less extreme than in the rubrene work; hence the lack of an effect of generation order may merely reflect the inconsequential production of foreign substances. In an alternative view, it may be that ‘T is totally governed by well-controlled substances, such as the ion radicals. Triplet quenching by the ions may also bear importantly on the third important contrast these results show with the rubrene work: the concentration dependence of ?. The o, us. (t,/tr)’” plots were, in all cases, highly linear; hence the existence of a timeindependent f(t)/T2ratio is implied. At the higher concentrations, ? decreases slowly with increasing concentration, but since the parameter’s definition includes an inverse concentration term, this behavior seems entirely reasonable. At the lower concentrations, however, 0 quickly assumes very large values. At present it is not possible to thoroughly rationalize this phenomenon, but it should be noted that at low-substrate concentrations it is likely that the “walls” radical ions, which bound the reaction zone, will be less formidable than at higher concentrations. If so, as the concentrations of the aromatics are decreased, the control of 7 may
well shift from highly reproducible effects such as those associated with radical ions, to more obscure and variable ones, such as quenching by the solvent or trace impurities. In addition to the foregoing contrasts, the results given here offer some noteworthy parallels with previous reports. Specifically, an extrapolation of a to higher concentrations indicates apparent triplet yields in accord with those measured earlier by sensitized isomerization of stilbene. Freed and Faulkner recorded a value of 7 i 2 X for the yield at [lo-MP] = 2 X 10-2 M , compared to the value of 4 X obtained from the current data by extrapolation. Considering the imprecision attendant to the extraction of Qt from a and to the extrapolation, one can regard this agreement as encouraging support for the interpretation presented here. It is further interesting to note that the triplet yields of fluoranthene-10-MP systems are comparable to those reported for the rubrene anion-cation reaction. Moreover, it remains mildly surprising that the fractional participation of triplets in annihilation is so high, even though it is comparatively lower in this case. Thus one can reiterate the earlier conclusion that the low values of Qecl result primarily from inefficiencies in the annihilation process itself, rather than from low excitation yields. Acknowledgment. The support of this work by a departmental grant from the E. I. du Pont de Nemours Co., Inc., and by Grant GP-28375 from the National Science Foundation is gratefully acknowledged. We wish to thank the National Science Foundation further for its award of a Graduate Fellowship to R. B.
Mechanism of Ozonolysis. Microwave Spectrum, Structure, and Dipole Moment of Ethylene Ozonide C. W. Gillies and R. L. Kuczkowski”
Contribution f r o m the Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48104. Received February 17, 1972
Abstract: The structure of ethylene ozonide, H2COOCHZ0,has been determined from the microwave spectra of seven isotopic species. The half-chair conformation with CZsymmetry has been established and there is no evidence for free or hindered pseudorotation. Th? parameters determined for tpe ring atoms are as follows: r(C-0,) = 1.436 i 0.006 A, r(C-0,) = 1.395 0.006 A, r(Op-O,) = 1.470 i 0.015 A, LCOC = 102.83 + 0.44’, LCOO = 99.23 i 0.38”, LOCO = 106.25 i 0.64”, and the dihedral angles ‘T(C~O,C,O,> = -16.60 + 0.40’, T(C,O,O,C,) = -50.24 1.26”, and ‘T(O,CO,O,) = 41.27 + 0.96’. The dipole moment is 1.09 i .01 D. Three excited vibrational states of the normal isotopic species arising from the asymmetric ring-bending vibration have been assigned. The ozonolysis of HDC=CH1 yielded six isotopic species of ethylene ozonide including two singly deuterated ozonides, three doubly deuterated ozonides, and the undeuterated ozonide. This result was analyzed in view of several proposals for the mechanism of ozonolysis.
*
*
T
he mechanism of the ozonolysis of alkenes has recently been reviewed by Murray.’ The most extensive mechanistic scheme previous to 1960 was proposed by Criegee.2 According t o his proposal (1) R. W. Murray, Accounts Chem. Res., 1 , 3 1 3 (1968). (2) R. Criegee, Rec. Chem. Progr., 18, 11 1 (1957).
(Scheme I), the initial ozone-alkene adduct I fragments to give a zwitterion I1 and a carbonyl compound 111. The normal ozonide is formed by the recombination of 11 and 111. Experimental data have accumulated which cannot be rationalized on the basis of the Criegee mechanism. In particular, the discovery that cis-
Gillies, Kuczkowski J Spectrum, Structure, and Dipole Moment of Ethylene Ozonide
6338 Scheme I
00-0
/ \
/
rr
111
IV
trans ratios of cross ozonides prepared by the ozonolysis of cis and trans unsymmetrical alkenes depend on alkene geometry led Murray, et u I . , ~ to suggest the possibility of a competing mechanism. The additional pathway suggested by them involves the attack of I? the molozonide, by a foreign or Criegee-produced aldehyde. Bailey, et u I . , suggested ~ later that the accumulated data can be rationalized by refining the original Criegee mechanism. To account for the steric effects, it was proposed that the original Criegee zwitterion can exist as syn and anti isomers. Subsequent oxygen-18 labeling studies of some aliphatic alkenesj-8 supported the Murray aldehyde interchange mechanism. The results of other oxygen-18 labeling investigations of phenylethylene~~ were rationalized in terms of the original Criegee mechanism. This microwave study of 1,2,4-trioxacyclopentane (ethylene ozonide) was undertaken in order to elucidate the mechanism of the ozonolysis of ethylene through isotopic labeling. Microwave spectroscopy is potentially a powerful tool in such investigations since the spectra of different isotopic species can be uniquely predicted once the structure is known. The resolution of a microwave spectrometer readily permits unambiguous detection of the diffcrent isotopic spectra even in a multicomponent mixture. The technique also does not require destruction of the ozonide in contrast to isotopic mass spectral studies or chemical degradation to simpler species with subsequent mass spectral analysis. Hence, assumptions about fragmentation patterns or degradation mechanisms are not necessary. In order to pursue mechanistic studies, the detailed structure of ethylene ozonide had to be determined. The structure of this ozonide was interesting in its own right since little information is available in the literature on such compounds. In fact, previous to this work, the conformation had not been established unambiguously for any ozonide; moreover, the possibility of a nonrigid ring showing the effects of free or hindered pseudorotation had not been eliminated. An electron diffraction study of ethylene ozonide10found that molecular models with C2symmetry (half-chair conformation) (3) R. W. Murray, R. D. Youssefyeh, and P. R. Story, J . Amer. Chew?. Soc., 89, 2429 (1967).
(4) N. L . Bauld, J. A. Thompson, C. E. Hudson, and P. S . Bailey, ibid.,90, 1822 (1968). ( 5 ) C. E. Bishop and P.R. Story, ibid., 90, 1905 (1968). (6) P. R. Story, C. E. Bishop, J. R. Burgess, R. W. Murray, and R. D. Youssefyeh, ibid.,90, 1907 (1968). (7) R. W. Murray and R. Hagen,J. O r g . Chem., 36,1103 (1971). ( 8 ) P. R. Story, J. A . Alford, J. R . Burgess, and W. C. Ray, J . Amer. Chem. SOC.,93,3042 (1971). (9) S. Fliszar and J . Carles, ibid.,91,2637 (1969). (10) A. Almenningen, P. Kolsaker, H . M. Seip, and T. Willadsen, Acta Chem. Scatzd., 23,3398 (1969).
Journal of' the American Chemical Society
and C, symmetry (envelope conformation) both gave equally good fits to the experimental intensity and radial distribution curves. Models corresponding to free or restricted pseudorotation were also not excluded. Conformational energy calculations, however, favored the model with C, symmetry by 0.81-1.29 kcal/mol. Recently, the structure of 3-carbomethoxy-5-anisyl1,2,4-trioxacyclopentane was investigated by X-ray diffraction. l 2 The conformation of the ozonide ring could not be established because of a structural disorder associated with the five-membered ring. In this investigation, ethylene ozonide has been found to have the C, conformation in the ground vibrational state and a detailed structure is reported. The compound shows no evidence of free or hindered pseudorotation. Several low-frequency vibrational states arising from excitations of the ring-bending vibration were assigned and the dipole moment was determined. Information on the mechanism of ozonolysis was obtained by analyzing the products of the ozonolysis of HDC=CH2. The presence of five partially deuterated species as well as the normal isotopic species of ethylene ozonide in the products of this ozonolysis is discussed in the final section. Experimental Section Ethylene ozonide was prepared by ozonizing 0.1 M solutions of ethylene (CP grade, Matheson Co.) in methyl chloride (high purity grade, Matheson Co.) at -78". All ozonolyses were continued to 100% of the theoretically required amount of ozone. Ozonide yields ranged from 60-70x. Due to the unstable nature of ethylene ozonide, 1-mmol quantities were prepared. Flow rates of approximately 2 mg/min of ozone were used for the reaction. The sample outlet of a Welsbach Model T-408 ozonator was used to obtain the low ozone flow rates which were necessary to ozonize 1mmol quantities of ethylene. The ozonide was separated from the solvent by low-temperature vacuum distillation. A series of traps cooled to -45.2, -95, and - 196' were used; the ozonide was collected in the -95' trap, high boiling impurities were stopped by the -45.2" trap, and the solvent was trapped at -196". An identical procedure was used for the ozonolysis of HDC=CH? (98Z enriched from Merck and Co., Inc.). All samples of the ozonide were stored at liquid nitrogen temperature because samples were found to decompose overnight at room temperature. The sample cell of the microwave spectrometer was maintained at Dry Ice temperature for all spectroscopic work. Even at this temperature, frequent fresh doses of the sample were necessary due to ozonide decomposition. All rotational transitions were measured with a conventional 80kHz Stark-modulated spectrometer to an accuracy of i 0 . l MHz.13 A precision dc power supply (Fluke, Model 412B) was employed for Stark shift measurement^'^ and the dipole moment of OCS15 was used to determine the average spacing of the Stark septum in the cell.
Spectra The rotational spectrum of ethylene ozonide, a near0.92, consisted of a number of oblate rotor with K strong b-type R-branch transitions. Each transition was accompanied by several regularly spaced vi brational satellites on the low-frequency side. The rotational transitions assigned to the ground state and three vibrational satellites are given in Tables I and TI. The frequencies of all transitions from each state fit well to a rigid-rotor model. indicative of a ring with
-
(11) (12) (13) (14) (I 5)
1 94:18 1 September 6 , 1972
H. M. Seip, ibid.,23,2741 (1969). P. Groth, ibid., 24,2137 (1970). R. L. I
