AmericanChemicalSociety.com

Introduction

The goal of the funded research is to realize an asymmetric halogen-metal exchange (AHME) reaction of geminal dihalides using a recyclable planar chiral organometallic reagent derived from an iodoferrocene. The proposed AHME process (Figure 1) is a closed cycle in which a recyclable planar chiral ferrocene is used as a stereoinductive element. The complete AHME cycle incorporates three distinct phases: [1] generation of a planar chiral ferrocenylmetal (4) from an iodoferrocene precursor (1) via iodine-metal exchange using a simple achiral alkylmetal reagent (2); [2] AHME between 4 and a prochiral geminal diiodide (5) leading to the regeneration of 1 and the formation of a chiral a-haloalkylmetal (6); and [3] trapping of the scalemic carbenoid 6 in some process of interest.

Results and discussion

Geminal dihalide substrates for study were prepared by alkylation of halomethyl sodiums with either TMSCl or a 1° alkyliodide via an adaptation of known methods (Scheme 1). For initial assessment of the thermodynamic feasibility for the necessary halogen-metal interchange reactions involved in AHME, a simple chiral 2-iodo-ferrocene (8) was prepared. Following generation of the corresponding ferrocenyllithium 9 by I/Li interchange, a gem-dihalide was introduced and the quantity of iodoferrocene 8 regenerated by the reverse Li/I exchange was determined (Scheme 2). These results successfully verified that all of the requisite halogen-metal exchanges were possible. In each case, interchange efficiency was greatest for the prochiral geminal diiodides as compared to related geminal chloroiodides. Interestingly, an analogous series of bromide based dihalides failed to undergo halogen-metal exchange.

In our original design for reagent 4 we considered that YH would be important, in part because this attribute precludes a possible racemization pathway via intermolecular enantiotopic proton transfer between metallated ferrocenes. Fortunately, the fear of ferrocene racemization for carboxamide based systems with Y=H proved to be unwarranted. A selection of the above interchange cycles were repeated under otherwise identical conditions using an enantioenriched sample of iodoferrocene 8 (96% ee); in no case was a measurable drop in the ee of recovered material 8 observed.

In addition to the simple carboxamide based iodoferrocene 2, a variety of other types of planar chiral ferrocenyl iodides were also targeted and tested for efficacy as potential AHME reagents (Figure 2). In many of these compounds we tried to set YH. Aside from preventing possible racemization (vide supra), another benefit of setting YH would be to allow for fine tuning of the thermodynamic stability of the ferrocenyl metallate by a combination of steric and electronic effects. For example, a butressing effect between Y and the DoM group would result in torque and loss of chelate planarity (as shown in 10). After extensive experimentation, it was discovered that within the N,N-diisopropyl carboxamide series, substitution at the Y position precluded installation of the requisite iodine atom via DoM. Presumably, in this case the butressing effect was too great and prevented DoM altogether. Within the analogous N,N-diethyl series, however, synthesis of at least one iodide with YH (14) was possible via DoM, albeit in a poor yield (14%).

An oxazoline based series of ferrocenes (15-17), inspired by Sammakia's work and the FOP catalysts of Overman, was also studied. In this case, iodoferrocenes with YH could be prepared in high yield but the derived metallated derivatives (Li or MgCl) proved incapable of effectively reassimilating iodine from dihalides 7. Sulfoxide 18 and tertiary amine 23 were also examined, but again neither performed as effectively as carboxamide 8.

Imperative to a successful proof of concept for the proposed AHME process is validation of the third and final phase of the cycle: i.e., trapping of the product chiral carbenoid 6 in some useful reaction that could enable an indirect assay of ee. To our immense frustration, early attempts to trap carbenoid 6 with aldehydes rarely yielded an addition adduct, but instead usually led to consumption of starting dihalide and production of the alcohol arising from simple reduction of the carbonyl. After exhaustive experimentation, it was discovered that such carbenoids can be consistently added to carbonyl compounds in the desired manner in the presence of Lewis acid additives. Our recent studies have also demonstrated that low temperature quench following addition of a-chloroalkylmetals to aldehydes leads exclusively to chlorohydrins, while analogous experiments involving a-iodoalkylmetals result in facile intramolecular collapse of the initial addition adduct to provide epoxides exclusively.

Complete AHME cycles have now been successfully realized in two different scenarios (Scheme 3); these results punctuate the culmination of two years of hard work on this difficult project. Thus, enantioenriched 2-iodoferrocene 8 underwent initial I/Li exchange with n-BuLi at –78°C (THF); following subsequent addition of gem-diiodide 7 (R=SiMe₃, X=I), the resulting putative carbenoid 20 was thereafter added to benzaldehyde that had been precomplexed with Me₂AlCl. In this manner, epoxysilane 21 was obtained in 35% yield accompanied by 70% of recovered 8 that exhibited no significant loss in ee. In the second example, a similar sequence of events led to chlorohydrin 23 (²15% yield) by the use of racemic chloroiodide substrate 7 (R=Ph(CH₂)₃, X=Cl); again, 8 was recovered with excellent stereochemical fidelity. The development of a suitable HPLC based resolution technique to assay %ee for the products from these transformations is under active investigation.

Conclusion

In summary, validation of all three phases of the proposed AHME cycle has been achieved. It was further established that a simple planar chiral 2-iodoferrocenyl carboxamide reagent (8) could be subjected to a round trip I/Li/I sequence without compromising its ee, thus allaying concerns of a possible racemization pathway. Finally, a reliable method for the trapping of alpha-haloalkyl metals with aldehydes was developed. It remains to assay ee for products resulting from this process, to optimize the methodology, and to study and expand on its potential scope. This work will be conducted and concluded during the upcoming 12 month no cost extension period (9/1/10-8/31/11).

Note: Research findings have been presented by student Christopher Emerson at the 239th ACS National Meeting (3/21/10-3/25/10, San Francisco) and at the 1st Annual Graduate Research Symposium of the Organic Division of the ACS (7/15/10-7/18/10, Boston).