Efficient Preparation of (S)- and (R)-tert-Butylmethylphosphine–Borane: A Novel Entry to Important P-Stereogenic Ligands

Abstract A novel one-pot reductive methodology for the synthesis of optically pure tert-butylmethylphosphine–borane is reported. The preparation uses as the starting material tert-butylmethylphosphinous acid–borane, which is available in both enantiomeric forms from cis-1,2-aminoindanol and tert-butyldichlorophosphine. The process is based on the reduction of a mixed anhydride, the configurational stability of which has been studied in several solvents and temperatures. Tetrabutylammonium borohydride was the best reducing agent allowing for the development of a practical process. To demonstrate the utility of the new methodology, the product obtained in this manner was used in the preparation of Quinox-P*.

P-Stereogenic phosphines are a subclass of phosphine ligands that have recently grown into one of the most efficient type of ligands for asymmetric hydrogenation and other relevant industrial processes. 1 In this respect, the development of synthetic methodology allowing for the efficient synthesis of such compounds is of crucial importance. In our group, we have developed a novel strategy for the synthesis of valuable P-stereogenic synthons like amino(tert-butyl)methylphosphine-borane 1 and tert-butylmethylphosphinous acid-borane 2 (Scheme 1). 2 Compounds 1 and 2 have been employed in the synthesis of MaxPHOS, SIP, and phosphinooxazoline ligands that have proven very efficient in asymmetric hydrogenation and [2+2+2]-cycloaddition reactions. 2a,e,f Compounds 1 and 2 are valuable because they bear in common the tert-butylmethylphosphine moiety which provides a high steric bias when the phosphorus is coordinated to the metal center.

Scheme 1 Synthesis of optically pure P-stereogenic synthons
Another important P-stereogenic building-block of the same family is the tert-butylmethylphosphine-borane 3, that has been used by Imamoto for the synthesis of C 2 symmetric Quinox-P*, Benz-P*, and Pincer-P* ligands ( Figure  1). 3 These ligands have been demonstrated to be very efficient in numerous catalytic processes. 4 The synthesis reported for 3 relies on the stereoselective deprotonation of tert-butyldimethylphosphine-borane with the sparteine/s-BuLi couple, followed by oxidation of the corresponding phosphide with O 2 to yield the corresponding (hy- Pincer-P* SYNTHESIS0 0 3 9 -7 8 8 1 1 4 3 7 -2 1 0 X © Georg Thieme Verlag Stuttgart · New York 2016, 48, 2659-2663 special topic E. Salomó et al.

Special Topic Syn thesis
droxymethyl)phosphine 4 (Scheme 2). Further oxidation of 4 with RuCl 3 /K 2 S 2 O 8 leads to the phosphinecarboxylic acid which spontaneously decarboxylates to yield 3. The synthesis reported for 3 bears several shortcomings, like the use of sparteine, a natural diamine for which the unnatural enantiomer is difficult to obtain, and the need for optical enrichment of intermediate 4 by crystallization. 5 With this picture in mind, we thought that alternative preparations of optically pure 3 would be valuable for either the preparation of new or already existing P-stereogenic ligands. Here we report on the reduction of phosphinous acid 2 leading to optically pure tert-butylmethylphosphine 3 in a stereospecific fashion. To demonstrate the utility of the novel preparation, the secondary phosphineborane thus obtained was further transformed into Quinox-P*.
Optically enriched phosphinous acid-boranes are attractive synthetic intermediates; despite this, they have been scarcely used in ligand synthesis. Pietrusiewicz and Buono independently reported the preparation of optically pure tert-butylphenylphosphinous acid-borane 5 and its reduction to the corresponding secondary phosphine 7 (Scheme 3). 6 In the case of 5, a two-step procedure was necessary to accomplish the transformation. The mixed anhydride 6 was isolated, according to the authors, without loss of optical purity. Overall the reduction takes place with inversion of configuration at the P-center.

Scheme 3 Reduction of tert-butylphenylphosphinous acid according to Pietrusiewicz and Buono
We have recently reported that methanesulfonyl (mesyl) anhydrides derived from 2 and 5 can undergo nucleophilic substitution reaction at phosphorus (S N 2@P) with amine nucleophiles. 2a In this work, we noticed that the mesyl anhydride derived from 2 was not configurationally stable and could not be isolated as the phenyl analogue 6.
In order to determine the optimal solvent and temperature conditions in which intermediate 8 would preserve the initial optical purity, we studied the transformation of 2 into the amine derivative 1 (Table 1). Optically pure phosphinous acid 2, was treated with mesyl anhydride and Et 3 N to yield the mixed anhydride 8 which was left some time in solution before bubbling an excess of NH 3(g) into the reaction mixture. This was a convenient method since the optical purity of 1 could be readily determined by chiral GC. Initially, using different solvents, the formation of 8 was carried out at 0 °C and the solution was left for 1 h at the same temperature before bubbling ammonia (Table 1, entries 1-6). The use of toluene, THF, Et 2 O, and DME afforded the final substitution product 1 with a high degree of racemization (22-46% ee). On the other hand, acetonitrile and dichloromethane produced less racemization affording the final product in 80 and 91% ee, respectively. At this stage we studied the effect of the temperature. Using the best solvent in the series and lowering the temperature to -10 °C the enantiomeric excess increased to 98% ee (Table 1, entry 7). Running the reaction at -20 °C in CH 2 Cl 2 the racemization was completely suppressed and the product was isolated in 99% ee (Table 1, entry 8). To show the importance of temperature, an almost complete racemization took place when a CH 2 Cl 2 solution of 8 was stirred for 3 h at room temperature (Table 1, entry 9). From the previous study, we concluded that intermediate 8 was stable to racemization at -20 °C in CH 2 Cl 2 . Thus, the reduction of phosphinous acid 2 had to be ideally performed under the same conditions of solvent and temperature. With this constraint in mind we began to search for the reagent that could fulfil such requirements ( Table 2). The use of NaBH 4 , which was used successfully in the reduc-

Special Topic Syn thesis
tion of 6, did not produce any reduction product (Table 2, entry 1). We attributed this lack of reactivity to the poor solubility of NaBH 4 in CH 2 Cl 2 . Reduction with BH 3 ·SMe 2 or NaBH(OAc) 3 was also unproductive ( Table 2, entries 2 and  3). Diisobutylaluminum hydride at -20 °C produced a low yield (13%) of the desired secondary phosphine (Table 2, entry 4). Increasing the reaction temperature to 0 °C and shortening the reaction time to 2 h improved the yield to 34% but with a concomitant loss of optical purity (Table 2, entry 5). We reasoned that the sluggish reactivity observed for the DIBAL-H reagent was due to the steric hindrance created by the isobutyl groups of the reagent. Hence, we next tried the use of the smaller alane (AlH 3 ) generated from LiAlH 4 and AlCl 3 . Addition of alane over the mixed anhydride 8 in CH 2 Cl 2 at -20 °C afforded this time the secondary phosphine 3 in 80% yield and 97% ee (Table 2, entry 6). Finally, in the search for a commercial reducing agent that could be easily handled and stored, we turned our attention to tetrabutylammonium borohydride ([Bu 4 N][BH 4 ]). The high solubility of this reagent in CH 2 Cl 2 permits reductions to be carried out in the absence of protic solvents. 7 The use of [Bu 4 N][BH 4 ] provided an efficient reduction of the intermediate 8 producing the secondary phosphine 3 with inversion of configuration in 86% yield and 99% ee (Table 2, entry 7). Using the opposite enantiomer of the phosphinous acid, the enantiomer of 3 was obtained in 99% ee (Table 2, entry 8), thus demonstrating that the reduction process is completely stereospecific.
To demonstrate the utility of this novel reduction methodology, optically pure tert-butylmethylphosphine-borane prepared by us was employed in the preparation of Quinox-P* ligand following Imamoto's procedure (Scheme 4). 3b Deprotonation of 3 with n-BuLi at -78 °C provided the corresponding lithium phosphide which was reacted in situ at low temperature with dichloroquinoxaline. Removal of the borane protecting groups provided, in a single-pot process, Quinox-P* in 70% yield and 99% optical purity as determined by optical rotation. 8

Scheme 4 Preparation of Quinox-P* starting from tert-butylmethylphosphine-borane prepared by the new reduction methodology
In summary, we have devised a novel reductive methodology for the synthesis of optically pure tert-butylmethylphosphine-borane, which is a strategic P-stereogenic intermediate for the synthesis of important chiral phosphine ligands. The novel preparation uses as the starting material tert-butylmethylphosphinous acid-borane which is available in both enantiomeric forms. The process is based on the reduction of the mixed mesyl anhydride derivative 8 which was found to be configurationally stable in CH 2 Cl 2 at -20 °C. Tetrabutylammonium borohydride was the reducing agent of choice, allowing for the development of a practical one-pot process. The usefulness of the new reductive methodology was demonstrated by the preparation of Quinox-P* following the original Imamoto's procedure. We think that the novel preparation will improve the availability of 3 and thus foster its incorporation into novel ligand structures.
All reactions were carried out under a N 2 atmosphere with dried solvents; THF, Et 2 O, and CH 2 Cl 2 were dried in a PureSolv purification system from Innovative Technology, Inc. NMM was dried with molecular sieves and kept under N 2 . Other commercially available reagents and solvents were used with no further purification. TLC was carried out using TLC-aluminum sheets with silica gel (Merk 60 F 254 ). Silica gel chromatography was performed by using 35-70 mm silica or an automated chromatography system (Combiflash®, Teledyne Isco). NMR spectra were recorded at 23 °C on a Varian Mercury 400 or Varian 500. 1 H and 13 C NMR spectra were referenced either to relative internal TMS or to residual solvent peaks. 31 P NMR spectra were referenced to H 3 PO 4 . Optical rotations were measured at r.t. (25 °C) using a Jasco P-2000 iRM-800 polarimeter. Concentration is expressed in g/100 mL. The cell sized 10-cm long and had 1 mL of capacity, measuring λ was 589 nm, which corresponds to a sodium lamp. Synthesis of (S)-2 and (R)-2 was performed as previously described. 2a

Reduction with Alane
A solution (S)-tert-butylmethylphosphinous acid-borane [(+)-2, 100 mg, 0.75 mmol) and Ms 2 O (195 mg, 1.12 mmol) in CH 2 Cl 2 (4 mL) was cooled to -20 °C. To this solution, anhyd Et 3 N (0.26 mL, 1.87 mmol) was slowly added and the mixture was stirred for 1 h at -20 °C. A solution of AlH 3 [1.5 M in Et 2 O, made in situ by mixing LiAlH 4 (4 equiv) and AlCl 3 (1 equiv) in Et 2 O] (2 mL, 3.0 mmol) was added dropwise and the mixture was stirred for 4 h at -20 °C. Consumption of the starting material was observed by TLC. The solution was warmed to 0 °C and 1 M aq HCl was slowly added. The resulting suspension was filtered through a plug of Celite. The organic layer was separated and the aqueous phase was extracted with CH 2 Cl 2 (3 ×). The combined extracts were washed with brine, dried (MgSO 4 ), and concentrated on a rotary evaporator under reduced pressure. Purification by column chromatography (silica gel, isocratic CH 2 Cl 2 ) gave (+)-3 as a colorless semisolid; yield: 90 mg (80%); 97% ee. Spectroscopic data was in agreement with the literature. 3a

Enantiomeric Excess Determination for 3
Optical purity for 3 was determined by derivatization to the corresponding benzylphosphine which was analyzed by chiral HPLC.