Asymmetric epoxidation enantiofacial selectivity

As pointed out by Hosoya et al.92 the enantiofacial selection of ra-olefins is mainly controlled by the asymmetric centers at the C-8(8 ) carbons, while that of trans-olefins is preferentially controlled by the asymmetric centers at the C-9(9 ) carbons in 119 or 120. Optically active Mn(III)-salen complexes have catalyzed the epoxidation of m-olelins with higher ee (>90%), especially when they are conjugated with an acetylene or phenyl group. However, the epoxidation of trans-olefins with these salen complexes shows rather poor enantio-selectivity (Table 4-18). 

The presence of the stereogenic centre at C(l) introduces an additional factor in the asymmetric epoxidation now, besides the enantiofacial selectivity, the diastereoselectivity must also be considered, and it is helpful to examine epoxidation of each enantiomer of the allylic alcohol separately. As shown in Fig. 10.2, epoxidation of an enantiomer proceeds normally (fast) and produces an erythro epoxy alcohol. Epoxidation of the other enantiomer proceeds at a reduced rate (slow) because the steric effects between the C(l) substituent and the catalyst. The rates of epoxidation are sufficiently significative to achieve the kinetic resolution and either the epoxy alcohol or the recovered allylic alcohol can be obtained with high enantiomeric purity [9]. 

The nonconventional tartrate esters 1-3 have been used to probe the mechanism of the asymmetric epoxidation process [20a]. These chain-linked bistartrates when complexed with 2 equiv. of Ti(0-f-Bu)4 catalyze asymmetric epoxidation with good enantiofacial selectivity. 

The scope of allylic alcohol stmctures that are subject to asymmetric epoxidation was foreshadowed in the first report of this reaction. Examples of nearly all the possible substitution patterns were shown to be epoxidized in good yield and with high enantiofacial selectivity [2], The numerous results that have appeared since the initial report have confirmed and extended the scope of the stmctures that have been epoxidized. This section of the chapter illustrates the structural scope without being exhaustive in coverage of the literature. Examples were chosen... 

In the epoxidation of the allylic alcohol shown in Eq. 6A.3b, the epoxy alcohol is obtained in 96% yield and with a 14 1 ratio of enantiofacial selectivity [40b], An interesting alternate route to the epoxide of entry 12 (Table 6A.3) has been described in which 2-r-butylpropene is first converted to an allylic hydroperoxide via photooxygenation and then, in the presence of Ti-tartrate catalyst, undergoes asymmetric epoxidation (79% yield, 72% ee) [38b]. The intermediate hydroperoxide serves as the source of oxygen for the epoxidation step. 

Allylic alcohols with a cis-3-substituent (45) are the slowest to be epoxidized, and they give the most variable enantiofacial selectivity. Both these characteristics suggest that allylic alcohols of this structure have the poorest fit to the requirements of the active epoxidation catalyst. Nevertheless, asymmetric epoxidation of these substrates is still effective and in most cases gives an enantiomeric purity of at least 80% ee and often as high as 95% ee. Patience with the slower reaction rate usually is rewarded with chemical yields of epoxy alcohols comparable with those obtained with other allylic alcohols. A number of representative examples are collected in Table 6A.5 [2,4,38,59,62a,71-78],... 

Extensive use in synthesis has been made of the asymmetric epoxidation of (2,3 )-disubstituted allylic alcohols. With few exceptions enantiofacial selectivity is excellent (90-95% ee). The results for a number of epoxidations of allylic alcohols with smaller substituents are collected in Table 6A.6 [2,4,41,61b,79-84], while a variety of other compounds with larger groups are illustrated by structures 47-60. 

The presence of a stereogenic center at Cj of an allylic alcohol introduces an additional factor into the asymmetric epoxidation process in that now both enantiofacial selectivity and dias-tereoselectivity must be considered. It is helpful in these cases to examine epoxidation of each enantiomer of the allylic alcohol separately. Epoxidation of one enantiomer proceeds normally and produces an erythro epoxy alcohol in accord with the rules shown in Figure 6A.1. 

The hallmark of Ti-tartrate catalyzed asymmetric epoxidation is the high degree of enantiofacial selectivity seen for a wide range of allylic alcohols. It is natural to inquire into what the mechanism of this reaction might be and what structural features of the catalyst produce these desirable results. These questions have been studied extensively, and the results have been the subject of considerable discussion [6,135,136]. For the purpose of this chapter, we review the aspects of the mechanistic-structural studies that may be helpful in devising synthetic applications of this reaction. 

A number of derivatives of the tartaric acid structure have been examined as substitutes for the tartrate ester in the asymmetric epoxidation catalyst. These derivatives have included a variety of tartramides, some of which are effective in catalyzing asymmetric epoxidation (although none display the broad consistency of results typical of the esters). One notable example is the dibenzyltartramide, which in a 1 1 ratio (in reality, a 2 2 complex as shown by an X-ray crystallographic structure determination [138]) with Ti(0-i-Pr)4 catalyzes the epoxidation of allylic alcohols with the same enantiofacial selectivity as does the Ti-tartrate ester complex [18], It is remarkable that, when the ratio of dibenzyltartramide to Ti is changed to 1 2, epoxidation is catalyzed with reversed enantiofacial selectivity. These results are illustrated for the epoxidation of a-phenylcinnamyl alcohol (Eq. 6A.12a). 

The essence of the asymmetric epoxidation process, including correlation of enantiofacial selectivity with tartrate ester stereochemistry, is outlined in Figure 6A. 1. No exceptions to the face-selectivity rules shown in Figure 6A.1 have been reported to date. Consequently, one can use this scheme with considerable confidence to predict and assign absolute configuration to the epoxides obtained from prochiral allylic alcohols. When allylic alcohols with chiral substituents at C-l, C-2, and/or C-3 are used in the reaction, the assignment of stereochemistry to the newly introduced epoxide group must be done with considerably more care. 

The nonconventional tartrate esters (1) to (3) have been used to probe tite mechanism of the asymmetric epoxidation process. These chain-linked bis(tartrate) molecides when complexed with 2 equiv. of Ti(OBu )4 catalyze asymmetric epoxidation with good enantiofacial selectivity. A number of tartrate-like ligands have been studied as potential chiral auxiliaries in the asymmetric epoxidation and kinetic resolution processes. Although on occasion a ligand has been found that has the capability to induce high enantioselectivity into selected substrates (see Section 3.2.T.3), none has exhibited the broad scope of effectiveness seen with the tartrate esters. 

In contrast to allylic alcdiols, the asymmetric epoxidation of homoallylic alcohols shows the following three general characteristics (i) the rates of epoxidation are slower (ii) enantiofacial selectivity is reversed, i.e. oxygen is delivered to the opposite face of the alkene when the same tartrate ester is used and (iii) the of oiantiofacial selectivity is lower with enantiomeric excesses of the epoxy alcohols... 

Several other allylic alcohols with primary C-2 substituents have been epoxidized with good results (Table 3, entries 7-10 and 14). Epoxy alcohols have been obtained with 93-96% ee and when the catalytic version of the reaction is used, as in Table 3, entry 10, the yield is excellent. When the C-2 substituent is more highly branched, as in entries 11-13, there may be some interference to high enantiofacial selectivity by the bulky group, since the ee in two cases (entries 11 and 12) is 86%. Another example which supports this possibility of steric interference to selective epoxidation is summarized in equation (3). In this case, the optically active allylic alcohol (12) was subjected to epoxidation with bo antipodes of the titanium tartrate catalyst. With (+)-DIPT enantiofacial selectivity was 96 4 ( matched pair ), but with (-)-DIPT selectivity fell to only 1 3 ( mismatched pair ), a further indication that a secondary C-2 substituent can perturb the fit of the substrate to the active catalyst species. In the epoxidation of the allylic alcohol shown in elation (4), the epoxy alcohol is obtained in 96% yield and with a 14 1 ratio of enantiofacial selectivity. An interesting alternate route to the epoxide of entry 12 (Table 3) has been described, in which 2-r-butylpropene is first converted to an allylic hydroperoxide via photooxygenation and then, in the presence of the titanium tartrate catalyst, undergoes asymmetric epoxidation (79%... 

There is a rough correlation between the enantiomeric excess observed for these qx)xy alcohols and the steric complexity at the a-carbon of the C-3 substituent. When the C-3 substituent is a primary group (Table 5, entries 1,2,4,6-12 and 19-21), enantiofacial selectivity is highest and enantiomeric excesses of 80-95% are observed for these compounds. When the substituent is secondary (entries 3 and 15-18) or tertiary (entry 5), enantiofacial selectivity is much more variable. When the substituent is asymmetric, enantiofacial selectivity depends on the ab lute configuration, as is evident in comparison of entry 15 with 16 and of 17 with 18 in Table 5. Epoxidation of these chiral allylic alcohols with one antipode of catalyst yields moderate to good diastereoselectivity, while with the odier antipode diastereoselectivity is virtually lacking. 

In the case of terminal olefins, asymmetric epoxidation typically results in relatively low enantiomeric excess. For (salen)Mn(III) catalysis, it is not clear whether the low degree of asymmetric induction is due to poor enantiofacial selectivity during... 

---