59th Annual Report on Research 2014

Significant work was carried out by two students based on preliminary results in which commercially available, unprotected cinchonidine catalyzed the reduction in 25-30% ee (Scheme 1). As expected, opposite enantiomers are obtained when quinine and hydroquinidine are employed. This is a remarkable first result considering all of the potential binding sites for the borenium ion on the chiral Lewis base and the possibility of actual reaction with the free OH group.       

We began by attempting to optimize conditions using quinine itself.  A small and reproducible improvement was observed when the reaction temperature was lowered.  We then prepared several derivatives with OH group protected, and others with significantly different structures.  As shown in Scheme 1, unfortunately all modifications led to considerably lower enantioselectivity. In addition, we also attempted the selective reduction of simple ketimines, but they also reacted with low enantioselectivities (Scheme 2).

Scheme 1. Attempted optimization of enantioselective borenium-catalyzed reductions

Scheme 2.  Effect of imine structure.

The interesting conclusion then was that the presence of the OH group on the benzylic position was required.  Considering the difficulty of obtaining dramatically different structures that would possess this sub-unit, we decided to focus on a different class of borenium catalyst that might be expected to have higher reactivity, and where tunability would be considerably easier. 

Thus we focused on borenium ions based on triazolylidenes (Scheme 3).  These are the mesoionic cousins of N-heterocyclic carbenes (NHC), that they are assembled by a Click reaction, which greatly aides the preparation of derivatives.  Additionally, they can be prepared with significantly less steric bulk than their NHC counterparts. As we have shown, the triazolylidenes are better able to stabilize the borenium and the resulting borohydride has greater reactivity.  Finally, unlike typical FLP catalysts such as B(C₆F₅)₃, borenium-based catalysts do not require rigorously dried hydrogen or high temperatures and pressures to catalyze the hydrogenation of imines and heterocycles. 

Scheme 3.  Synthesis and characterization of mesoionic boranes

In a paper recently accepted at Angewandte Chemie, we described the use of borenium catalysts for hydrogenation chemistry and showed that in isosteric systems, the borenium systems outperformed the corresponding NHCs (compare 3a/b in Scheme 4).  Additionally, the ability to access systems in which one of the flanking groups is a C–H, which is unique to triazolylidenes, results in the most active catalysts observed to date (3b).  These complexes catalyze the room temperature and atmospheric pressure hydrogenation of ketimines (4b) under conditions where B(C₆F₅)₃ is completely inactive (Scheme 4, Table).  In addition, heterocycles can also be reduced (Scheme 5).

Scheme 4.  Borenium catalyzed hydrogenations

Scheme 5.  Borenium catalyzed hydrogenations of heterocycles

As shown in Scheme 6, the first and obvious choice to prepare chiral triazolylium species employs alpha-chiral amines.  However in all cases, racemization occurred during deprotonation and was confirmed by deuteration studies.  We attempted to use a silver oxide route to transfer the chiral carbene to boron, and while this worked, racemization was again observed. 

Scheme 6. Attempted synthesis of chiral triazolylidene-based borohydrides as precursors to borenium ions. Racemization observed in all cases.

            We then moved to bicyclic substructures including the synthesis of chiral NHC-derived borohydride 11, which was converted into 9-BBN derivatives 12a/b (Scheme 7). The presence of two isomers is likely due to the high steric shielding of the boron center, and will likely be corrected by treating the product at elevated temperature. 

            Using the same strategy, 11 was converted into alkylated derivatives 13-17, which provide different steric environments and the possibility of chirality adjacent to boron (14-17), with diastereoselctivity of the BH addition across COD possibly influenced by the bisoxazoline ligand.

Scheme 7. Preparation of chiral NHC-based borenium ions.

            In addition, we are also examining axially chiral borenium species. Scheme 8 summarizes the results to date towards the synthesis of planar chiral paracyclophane 1,2,3-triazol carbene borane. Paracyclophane (18) can be brominated (19) or formylated (20) quantitatively. Initially a Sonogashira-route to convert rac-19 into the corresponding TMS-protected rac-22 was attempted, however, in all cases only starting materials and trace amounts of paracyclophane were isolated. Thus this route was abandoned in favour of a Corey-Fuchs reaction sequence (20 into 23 then 22), which has the advantage of allowing resolution via rac-PCP-CHO. This has been accomplished on small scale giving the (R)-isomer in low yield (14%) and modest 68% ee, which we hope to improve on larger scale.

Scheme 8. Preparation of chiral paracyclophane-based borenium ions.

            In the racemic series, alkyne 22 could be converted into triazole 24 by a Cu-catalyzed Click reaction and then alkylation took place in 75% yield.  The next steps, currently in progress, are deprotonation and trapping by BH₃, followed by conversion to an alkylated derivative. Alternatively direct reaction with 9-BBN would give the 9-BBN derivative of 27.

            Work is also underway to prepare axially chiral borenium ion 39. Thus far we have completely up to compound 35 on large scale, and on small scale have completed the synthesis of the azide derivative (Scheme 9).

Scheme 9. Preparation of chiral binaphthyl-based borenium ions.

            In conclusion, we have made significant advances in this first year towards the development of mild, enantioselective borenium-based metal-free catalysts for the hydrogenation of imines.  Work is underway on the use of this technique for the formation of C–C bonds and will be the focus of work in the next year of the grant. 

report narrative = 935 words, 9 schemes