The Molecular Zoo

Entry for Ethylene Ozonide

(1,2,4-trioxycyclopentane; O1COOC1)

prepared by Ben Allen (Class of 2003)

revised by Wayne E. Steinmetz

As a means of better understanding the nature of torsional strain in five-membered hydrocarbon rings, Almenningen et al. undertook the conformational analysis of ethylene ozonide by electron diffraction.¹ This study as well as later microwave work (vide infra)² showed that the material examined was not the primary ozonide but rather the cyclic product with the atom order OCOOC obtained by rearrangement via the Criegee mechanism. The primary ozonide in which the three oxygen atoms are bonded consecutively was later examined in a thorough microwave study by Gillies et al.³ Structural parameters were fit to the experimental data using two possible conformational models: the twist conformation with C₂ symmetry, and the envelope conformation with C_s symmetry (shown below). Their analysis compared the experimentally derived parameters to the theory available at the time. The experimental radial distribution curve was compared to theoretical functions for each proposed conformer; the theoretical curve for the C₂ structure gave a slightly better fit than that for the C_s structure. The RD curve provided rough estimates of interatomic distances which were analyzed by an iterative least-squares fit procedure subject to constraints. All calculations assumed symmetrically-situated methylene hydrogens and equivalent C-O bond distances. Comparisons of the structure refined under the assumption of C₂ symmetry to that of C_s symmetry were deemed inconclusive; the C₂ structure was slightly favored, through the uncertainties in the resulting structural parameters were greater. Almenningen also computed the energies of each potential structure using an early force field model, and found the C₂ structure to be favored by at least 1.5 kcal/mol. Almenningen et al. concluded in their paper that, while the assignment of the C₂ structure by electron diffraction alone was not definitive, the weight of evidence including calculations leans in that direction. They caution that, had they used a different parameterization for rotation about the O-O bond in the force field, the opposite result might have been obtained.

As part of their attempt to characterize the reaction between ozone (O₃) and ethylene gas (C₂H₂), Gillies and Kuczkowski measured microwave spectra for seven isotopomers of the reaction product, ethylene ozonide.² Data for 7 additional isotopomers and a revised structure were provided in a later paper.⁴ The molecule is a near-oblate rotor (k»0.92) belonging to the C₂ point group. They found that the spectra fit well to a rigid-rotor model and thereby ruled out pseudorotation in the ring. They also quickly ruled out the C_s conformation in favor of the C₂ by observing an alternation in intensity from rotational levels of opposite symmetry in the spectrum of the unsubstituted isotopomer; such an alternation requires the presence of a C₂ axis of symmetry. Starck effect measurements indicated that the dipole moment is entirely along the b or C₂ axis, thus confirming the determination of C₂ symmetry. C_2v conformations were ruled out by spectra for the mono-¹³C-substituted compound; an increase in P_cc was noted, indicating that the carbon atoms cannot lie in a plane of symmetry. Internal coordinates were calculated using Kraitchman's equations and a least-squares fit of the microwave data from the 14 isotopomers. Uncertainty in the O-O bond length was attributed to enhanced vibrational effects upon isotopic substitution. The caution required in the substitution method in this and similar cases is discussed by Demaison and Rudolph.⁵ The structural data indicate that the peroxide oxygens are twisted far out of the plane defined by the other three heavy atoms. Gillies and Kuczkowski speculated that, in the absence of torsional strain between adjacent methylene hydrogens (as in cyclopentane), the principle strain in ethylene ozonide is due to repulsion between oxygen lone pairs. This situation would then result in “…a C₂ conformation with a large O-O dihedral angle and a high barrier to inversion or psuedorotation.”

Tabular Results of Structural Parameters with Various Methods

All structures are in the twist conformation unless otherwise noted.

	Electron (envelope)	Electron^1 (twist)	Microwave^4	MMFF	Sybyl	AM1	HF/ 3-21g*	B3LYP/ 6-31G*	MP2/ 6-31G*
C-Ha	1.123	1.126	1.091	1.093	1.1	1.117	1.073	1.095	1.092
C-Hb	1.123	1.126	1.097	1.094	1.101	1.117	1.074	1.098	1.096
C-Oe	1.414	1.414	1.416	1.42	1.43	1.422	1.435	1.418	1.422
C-Op	1.414	1.414	1.412	1.42	1.431	1.422	1.448	1.415	1.418
O-O	1.487	1.487	1.461	1.458	1.481	1.301	1.471	1.464	1.478
COC	98.1	105.9	104.8	104.23	110.56	109.68	106.95	104.72	104.09
COO	103.27	98.57	99.3	103.16	102.16	108.58	100.6	99.56	98.43
OCO	103.0	104.81	105.0	109.33	102.43	102.78	104.5	105.9	105.87
HCH	117.2	112.3	113.3	109.41	110.15	114.32	113.26	111.75	112.35
OeCHa	108.93	109.87	110.8	109.55	113.32	110.24	110.65	111	110.77
OeCHb	108.93	109.87	109.08	109.34	108.64	109.63	110.31	110	109.99
COeCOp	51.3	-16.88	-16.2	-10.83	-14.18	-7.91	-15.06	-15.88	-16.87
COpOpC	0.00	-51.04	-49.5	-32.54	-45.71	-29.31	-47.33	-48.31	-50.69
OeCOpOp	-31.7	41.77	40.8	27.37	35.82	23	38.56	39.92	42.1
COeCHa	166.83	-134.86	-131.3	-131.13	-136.37	-122.91	-130.31	-131.3	-132.05
COeCHb	-64.24	101.11	102.8	108.99	100.86	110.4	103.57	104.51	103.15
Dipole Moment(D)			1.09	1.223				1.1028	1.1103

Observed Rotation Constants in MHz for isotopomers of ethylene ozonide.²

Uncertainties range from 0.02 to 0.30.

Isotopomer labels: C₁-O_e-C₂-O_p2-O_p1-H_a1-H_b1-H_a2-H_b2

Isotopomer	A0(MHz)	B0(MHz)	C0(MHz)
12-16-12-16-16-1-1-1-1	8243.84	8093.8	4584.74
13-16-12-16-16-1-1-1-1	8233.28	7928.25	4530.18
12-16-12-16-16-2-1-1-1	8203.59	7576.8	4418.74
12-16-12-16-16-1-2-1-1	8050.3	7668.24	4500.84
Cis-12-16-12-16-16-2-1-2-1	7971.7	7225.88	4340.32
trans-12-16-12-16-16-2-1-2-1	8084.21	7175.87	4260.34
trans-12-16-12-16-16-1-2-1-2	7861.84	7288.19	4413.44

Comparison of structures generated by various means:

• The microwave structure and the electron diffraction structure produced with the twist conformation assumed. These structures are quite similar, differing by at most 0.032 Å for bond lengths, 3.1^o for angles, and less than 1^o for heavy-atom dihedrals. The errors in internal coordinates produced by both methods had similar errors; the main difference in errors was for the positions of the hydrogens. Not surprisingly, the errors on these values were much greater in the ED structure than in the microwave structure.
• Molecular mechanics structure produced with the Merck Molecular Force Field. This method gave fairly good agreement with the microwave structure: bond lengths were nearly exact, while heavy-atom angles were off by around 3^o each. Heavy-atom dihedrals deviated significantly from the microwave structure; the differences for the CO_pO_pC and O_eCO_pO_p dihedrals were particularly striking.
• The Tripos SYBYL force field also gave decent results, with near perfect bond lengths. Whereas the dihedrals from the sybyl structure were in better agreement with the microwave structure than those produced by MMFF, the angles were markedly worse; the CO_eC angle was particularly off.
• The structure generated by the semi-empirical AM1 method was in worse agreement with experiment than either of the two mechanics methods: greater deviations were found from the microwave and electron diffraction structures in the bond lengths, angles and dihedrals.
• Results from ab initio quantum mechanics at the Hartree-Fock level with a 3-21G* basis set were generally better than any of the methods discussed so far, with angle and dihedral deviations from experiment about 3^o or much less. The CO_eC angle and CO_p bond length were the worst offenders.
• The quantum mechanical result using B3LYP hybrid density functional theory and a 6-31G* basis set was better still; this was the closest result to experiment seen so far, with all angles and dihedrals less than 2^o off. This calculation took considerably longer, however
• Finally, and electron correlation method at the MP2 level of theory was tried with a 6-31G* basis set. This method gave results slightly better still than the DFT method and took slightly longer as well.

Structural Files in pdb and SYBYL mol2 format generated by experimental and computational methods:

microwave structure pdb microwave structure mo2

electron diffraction twist pdb electron diffraction twist mo2

electron diffraction envelope pdb electron diffraction envelope mo2

MMFF pdb MMFF mo2

sybyl pdb sybyl mo2

am1 pdb am1 mo2

hf/3-21g* pdb hf/3-21g* mo2

b3lyp/6-31g* pdb b3lyp/6-31g* mo2

mp2/6-31g* pdb mp2/6-31g* mo2

References

1. A. Almenningen, P. Kolsaker, H. M. Seip, and T. Willadsen. “ Studies on Molecules with Five-membered Rings: III An Electron Diffraction Investigation of Gaseous 1,2,4-Trioxacyclopentane (Ethylene Ozonide)”. Acta Chem. Scand, 23, 3398-3406 (1969).

2. C. W. Gillies and R. L. Kuczkowski. “Mechanism of Ozonolysis. Microwave Spectrum, Structure, and Dipole Moment of Ethylene Ozonide”. J. Am. Chem. Soc., 94, 6337-6343 (1972).

3. J. Z. Gillies, C. W. Gillies, R. D. Suenram, and F. J. Lovas, "The Ozonolysis of Ethylene: Microwave Spectra, Molecular Structure, and Dipole Moment of Ethylen Primary Ozonide (1,2,3-Trioxolane)", J. Am. Chem. Soc., 110, 7991-7999 (1988).
4. R. L. Kuczkowski, C. W. Gillies, and K. L. Gallagher, "Microwave Spectrum and Structure of Ethylene Ozonide: Effects of Large Axis Rotations in Structure Calculations", J. Mol. Spectrosc., 60, 361-72 (1976).

5. J. Demaison and H. D. Rudolph, "When Is the Substitution Structure Not Reliable", J. Mol. Spectrosc., 215, 78-84 (2002).