Highly Enantioselective Hiyama Cross-Coupling via Rh-Catalyzed Allylic Arylation of Racemic Allyl Chlorides

Highly enantioselective Hiyama cross-coupling reactions have been achieved through rhodium(I)-catalyzed dynamic kinetic asymmetric transformations between aryl siloxanes and cyclic racemic allyl halides. This process affords valuable enantiomerically enriched aryl-substituted cyclic allyl products and is compatible with heterocyclic allyl chloride electrophiles.


■ RESULTS AND DISCUSSION
We first evaluated racemic 3-chlorocyclohex-1-ene 1a in combination with silicon-based coupling partners that have previously been used in Hiyama couplings (Table 1). In these initial studies we used the cationic rhodium(I) complex [Rh(COD)(MeCN) 2 ][BF 4 ] (5 mol %) and (S)-BINAP (6 mol %) as the ligand in THF at 60°C. Under these conditions, the Hiyama−Denmark protocol 13 involving a silanol and a bulky base did not give rise to desired product (entry 1). The use of trisiloxane was also unsuccessful (entries 2 and 3). Fortunately, we found that phenyl triethoxysilane in combina-tion with TBAF afforded the desired product in low yield but good enantioselectivity (entry 4, 25% yield, 92% ee). Both the yield and the enantioselectivity improved when phenyl trimethoxysilane was employed (entry 5), and carrying out the reaction at reflux gave higher yield (55% yield, 99% ee, entry 6). Dimethoxydiphenylsilane showed the desired reactivity and excellent enantioselectivity but the yield was quite low (21% yield, 99% ee, entry 7). Interestingly, the Rhcatalyzed asymmetric 1,4-addition of organosilanes 8a uses fluoride-free conditions, and when similar conditions were tried here we did not observe any desired product (entry 8). Finally, the use of alternative fluoride sources such as CsF or AgF did not give rise to 3a (entries 9 and 10).
We then proceeded to optimize further the conditions with the aim of improving the reaction yield (Table 2). As above, heating the reaction mixture to reflux in THF in the presence of [Rh(COD)(MeCN) 2 ][BF 4 ] (5 mol %), (S)-BINAP (6 mol %), and TBAF (2 equiv) gave 3a in 55% yield with 99% ee (entry 1). Using SEGPHOS as ligand afforded similar results (entry 2), whereas the employment of other bisphosphine ligands led to diminished yields, although maintaining complete enantioselectivity (entries 3 and 4). When dienetype ligand L3 was tested, no trace of 3a was detected (entry 5). 14 We were able to slightly improve the yield by increasing the metal/ligand ratio up to 1:1.5. Thus, 7.5 mol % (S)-BINAP gave 3a in 65% yield (entry 6). Carrying out the reaction in 1,4-dioxane at 90°C resulted in 11% yield (entry 7). Finally, other Rh(I) complexes were tested. However, none of them proved superior to the tetrafluoroborate salt (entries 8−11). 15 Using the reaction conditions described in entry 6,  4 ] (5 mol %), S-BINAP (7.5 mol %), TBAF (2 equiv), 2 equiv. siloxane in THF heated to reflux} we explored the scope of this transformation. Pleasingly, we found that the enantioselective Rh(I)-catalyzed Hiyama cross-coupling with racemic allyl chlorides could be accomplished with several aryl siloxanes (Scheme 2). Different aromatic motifs such as naphthyl or biphenyl are also compatible, giving rise to the corresponding products 3b,c in moderate yields and excellent enantioselectivity. The para-and meta-alkyl substituted arenes were tolerated, providing 3d−f with similar results in terms of both yield and enantioselectivity. In contrast, the use of an ortho-methyl substituted substrate only afforded traces of the desired product (not shown, <5% yield). A para-methoxy group slightly diminished the yield, although maintaining great enantioselectivity (98%) (3g). A similar effect was observed for the thioanisole derivative, leading to 34% yield and 95% ee (3h). However, with a methoxy group in the meta position, the desired product can be obtained in 51% yield and 96% ee (3i). Furthermore, a 1,3-dioxole moiety can be used to prepare 3j with high enantioselectivity (99% ee). In addition, halogen-substituted aryl siloxanes were suitable coupling partners, affording the products 3k−n with near complete enantioselectivity (97 → 99% ee).
Regarding heteroaryl siloxanes, commercially available triethoxy(thiophen-2-yl)silane was tested under these reaction conditions, but no product was detected. Additionally, some vinyl siloxanes were also tried, and only traces of product were observed in some cases.
Despite the promise of organosilicon coupling partners in synthesis, a limitation of this method is that there are currently few general robust procedures for the synthesis of aryltrimethoxysilanes, preventing the use of more elaborately functionalized silane coupling partners. However, we anticipate that if such aryl species were available then many of them would be tolerated in this reaction. We hope researchers in the field will develop methods to prepare such silanes.
Next, we examined if this protocol is applicable to heterocyclic allyl chloride coupling partners (Scheme 3). Using 3-chloro-3,6-dihydro-2H-pyran 1b with the conditions described above allowed us to prepare compound 3m with  high enantioselectivity (98%), albeit with only 34% yield. The synthesis of piperidine derivatives holds a significant importance within chemistry as it is present in numerous natural alkaloids, pharmaceuticals, and various synthetic substances with important properties. 15 Therefore, we decided to apply this asymmetric Hiyama cross-coupling in the synthesis of enantioenriched dihydropiperidine derivatives using racemic N-Boc-protected allyl chloride 1c in combination with different arylsiloxanes. We were pleased to find that this process gave rise to compounds 3n−q with high levels of enantiomeric excess. During this study, variable amounts of the homocoupling and proto-demetalation products of the silane were detected.
Additionally, β-elimination on the allyl chloride could sometimes be observed. These competing processes probably account for the moderate yields of the coupling products.
Regarding the reaction mechanism, we tentatively hypothesize the catalytic cycle operates as outlined in Scheme 4. In contrast to our work with boronic acid nucleophiles, 11c the use of a cationic Rh source with a BF 4 − counterion gave the best results. We suggest that initial activation of silicon by fluoride sets the stage for Si to Rh transmetalation. Then, oxidative addition to the allyl chloride likely takes place to provide a Rh(III) species. This intermediate could equilibrate between diastereomeric Rh-allyl species through suprafacial 1,3-isomerization. 16 If reductive elimination takes place preferentially in

■ CONCLUSION
In summary, we have reported highly enantioselective Rhcatalyzed cross-couplings between arylsiloxanes and racemic cyclic allyl chlorides. This process represents a rare example of asymmetric Hiyama coupling, The method enables the preparation of valuable allyl arenes with uniformly high enantioselectivity (92 → 99% ee). Important heterocyclic scaffolds are compatible with this transformation, leading to highly enantioenriched dihydropyran and piperidine derivatives. At this stage, it is proposed that diastereoselective 1,3isomerization between two competing Rh-σ-allyl species accounts for enantioselection.