Ethylene epoxidation transition states

Two extreme epoxidation modes, spiro and planar, are shown in Fig. 9 [33, 34, 53, 54, 76-85]. Baumstark and coworkers had observed that the epoxidation of cis-hexene of dimethyldioxirane was seven to nine times faster than the corresponding epoxidation of tran.y-hexene [79, 80]. The relative rates of the epoxidation of cisitrans olefins suggest that spiro transition state is favored over planar. In spiro transition states, the steric interaction for cw-olefm is smaller than the steric interaction for fran -olefm. In planar transition states, similar steric interactions would be expected for both cis- and trans-olefms. Computational studies also showed that the spiro transition state is the optimal transition state for oxygen atom transfer from dimethyldioxirane to ethylene, presumably due to the stabilizing interactions... 

The oxidation of the simplest symmetrically substituted alkene, ethylene, is noteworthy in that an asymmetric spiro transition state is observed. When constrained to Cs symmetry with eqnal forming carbon-oxygen bond lengths, the energy increases by only 0.1 kcalmol. The spiro TS has the plane of the HO—ONO (or peracid) at right angles to the axis of the C=C bond. In an idealized spiro TS this angle is exactly 90°. While the formation of snlfoxides from snlhdes by peroxynitrons acid is well-established , epoxidations have not yet been observed in solution. 

Ethylene epoxidation with unsubstituted carbonyl oxide is ca 5 kcalmol" more exothermic than with dimethylcarbonyl oxide, yielding an even earlier transition state. 

In summary, transition structures with dioxirane and dimethyldioxirane are unsymmet-rical at the MP2/6-31G level, but are symmetrical at the QCISD/6-31G and B3LYP/6-31G levels. The transition states for oxidation of ethylene by carbonyl oxides do not suffer from the same difficulties as those for dioxirane and peroxyforaiic acid. Even at the MP2/6-31G level, they are symmetrical (Figure 17). The barriers at the MP2 and MP4 levels are similar and solvent has relatively little effect. The calculated barriers agree well with experiment . In a similar fashion, the oxidation of ethylene by peroxyformic acid has been studied at the MP2/6-31G, MP4/6-31G, QCISD/6-31G and CCSD(T)/6-31G and B3LYP levels of theory. The MP2/6-31G level of theory calculations lead to an unsymmetrical transition structure for peracid epoxidation that, as noted above, is an artifact of the method. However, QCISD/6-31G and B3LYP/6-31G calculations both result in symmetrical transition structures with essentially equal C—O bonds. 

Modifying the selectivity for a particular product is a more challenging task. To understand why Ag is the most selective catalyst for ethylene epoxidation, an highly important reaction practiced industrially for decades, Linic et al. performed detailed spectroscopic and kinetic isotope experiments and DFT calculations, and they concluded that the selectivity between the partial and total oxidation of ethylene on Ag(l 11) is controlled by the relative stability of two different transition states (TS s) that are both accessible to a common oxametallacycle intermediate One results in the closure of the epoxide ring and ethylene oxide (EO), while the other leads to acetaldehyde (AC) via intra-molecular H shift and eventually combustion. The authors... 

Concise theoretical studies of Ziegler145,146 analyzed all of the possible reaction pathways including the crossover from the singlet to the triplet surface with the transition state on the singlet surface while the formed product is a triplet species. It could be shown that the epoxide precursor is formed via a [3 + 2]-addition of ethylene to two Cr=0 bonds followed by rearrangement to the epoxide product (Scheme 10). 

To test the prediction that Cu would improve the selectivity of Ag as an ethylene epoxidation catalyst, Linic, Jankowiak, and Barteau tested several Cu/ Ag catalysts on porous a-Al203 monoliths.56 The Cu concentrations in the catalysts varied from 0 to 0.8 at.%, so only a single bimetallic phase was expected. The catalysts were tested under conditions where the temperature, the feed composition of ethylene and oxygen, and the conversion of ethylene were held constant. Under these conditions, the selectivity of the catalyst increased significantly as the Cu content was increased from 0 to 0.2 at.%. This is a dramatic success the DFT calculations of transition states for competing reactions on bimetallic catalysts lead directly to the experimental identification of an improved catalyst. 

A summary of bimolecular rate constants for the acid-catalyzed hydrolysis of a series of alkyl-, vinyl- and phenyl-substituted epoxides is given in Table 1. Propylene oxide (7) is 6.6 times more reactive than ethylene oxide, and from a study of its reaction in H2018, it was shown that 70% of the glycol product results from addition of solvent to the secondary carbon and 30% from addition of solvent to the primary carbon. The reactivity per primary carbon of ethylene oxide is one-half of the observed reactivity of ethylene oxide, and thus the introduction of a methyl group on ethylene oxide results in an increase in reactivity at the primary carbon by a factor of 4 and an increase in reactivity at the secondary carbon by a factor of 9. These results are consistent with A-2 mechanisms for the acid-catalyzed hydrolyses of ethylene oxide and propylene oxide, in which some amount of positive charge generated on carbon at the transition state is stabilized by a methyl group. 

---