Stereospecific SN 2 @ P reactions : novel access to bulky P-stereogenic ligands

Chiral phosphines are the cornerstone of asymmetric metal catalysis. Among this class of ligands, bulky P-stereogenic phosphines have proved to be highly efficient in asymmetric hydrogenation and other relevant processes. However, their synthesis in the optically pure form is often not straightforward. The stereoselective synthesis of tert-butyl P-stereogenic phosphines has relied mostly on the selective deprotonation of tert-butyldimethylphosphine borane complexes and on the reaction of borane lithium phosphides with electrophiles. Nucleophilic substitution reactions at the ‘‘bulky’’ P-center (SN2@P) are generally avoided since it is overly unreactive. Furthermore, when forced to react, it usually provides non-stereospecific substitution processes. A relevant exception to this behavior is the reaction of halo-tertbutylmethylphosphine-boranes with alkynyllithium reagents. However, halophosphines are configurationally unstable and have to be generated and reacted in situ at low temperature. Jugé and others showed that P-stereogenic aryl and alkyl aminophosphine boranes undergo stereospecific acid-promoted methanolysis to yield the resulting methyl phosphinites with inversion at the P-center. The acidolysis can also be carried out with HCl/toluene, in this case yielding the optically enriched chlorophosphine boranes. However, these reactions directly fail or provide reduced optical purity when a bulky group (e.g., tertbutyl) is attached to phosphorus. We recently reported the stereospecific synthesis of aminophosphines 1a and 1b which have been used by us in the preparation of MaxPHOS and SIP type ligands. We considered that these compounds were ideally suited to evaluate whether the acidolysis of bulky aminophosphines could be performed in a stereospecific manner, and whether the resulting products could be further transformed into valuable ligand structures (Scheme 1). Herein we report on the stereospecific hydrolysis of bulky aminophosphine boranes and how the resulting phosphinous acids can be transformed into P*-ligands through SN2@P reactions. The initial acidolysis studies on 1a and 1b were conducted in H2SO4/MeOH (Table 1). Methanolysis of 1a at 50 1C for 16 h did not afford the expected methyl phosphinite borane 2a but the corresponding phosphinous acid borane 3a and the secondary phosphine oxide 4a that results from borane deprotection and P(III)/P(V) tautomerization of 3a (Table 1, entry 1). We attributed the formation of 3a to the residual content of water in the solvent used. To confirm this hypothesis, we next ran the acidolysis reaction of 1a in a MeOH/H2O (20 : 1) mixture. This solvent mixture and heating to 40 1C for 16 h afforded exclusively phosphine oxide 4a in low enantiomeric excess (Table 1, entry 2). By simply reducing the reaction time to 2 h, we were able to isolate the corresponding phosphinous acid 3a with an excellent yield and optical purity (499% ee, Table 2, entry 3).

Chiral phosphines are the cornerstone of asymmetric metal catalysis. 1Among this class of ligands, bulky P-stereogenic phosphines have proved to be highly efficient in asymmetric hydrogenation and other relevant processes. 2,3However, their synthesis in the optically pure form is often not straightforward.The stereoselective synthesis of tert-butyl P-stereogenic phosphines has relied mostly on the selective deprotonation of tert-butyldimethylphosphine borane complexes 4 and on the reaction of borane lithium phosphides with electrophiles. 5 Nucleophilic substitution reactions at the ''bulky'' P-center (S N 2@P) are generally avoided since it is overly unreactive.Furthermore, when forced to react, it usually provides non-stereospecific substitution processes. 6,7 relevant exception to this behavior is the reaction of halo-tertbutylmethylphosphine-boranes with alkynyllithium reagents. 8owever, halophosphines are configurationally unstable and have to be generated and reacted in situ at low temperature.
Juge ´and others showed that P-stereogenic aryl and alkyl aminophosphine boranes undergo stereospecific acid-promoted methanolysis to yield the resulting methyl phosphinites with inversion at the P-center. 9The acidolysis can also be carried out with HCl/toluene, in this case yielding the optically enriched chlorophosphine boranes. 10However, these reactions directly fail or provide reduced optical purity when a bulky group (e.g., tertbutyl) is attached to phosphorus.
We recently reported the stereospecific synthesis of aminophosphines 1a and 1b which have been used by us in the preparation of MaxPHOS and SIP type ligands. 11We considered that these compounds were ideally suited to evaluate whether the acidolysis of bulky aminophosphines could be performed in a stereospecific manner, and whether the resulting products could be further transformed into valuable ligand structures (Scheme 1).Herein we report on the stereospecific hydrolysis of bulky aminophosphine boranes and how the resulting phosphinous acids can be transformed into P*-ligands through S N 2@P reactions.
The initial acidolysis studies on 1a and 1b were conducted in H 2 SO 4 /MeOH (Table 1).Methanolysis of 1a at 50 1C for 16 h did not afford the expected methyl phosphinite borane 2a but the corresponding phosphinous acid borane 3a and the secondary phosphine oxide 4a that results from borane deprotection and P(III)/P(V) tautomerization of 3a (Table 1, entry 1).We attributed the formation of 3a to the residual content of water in the solvent used.To confirm this hypothesis, we next ran the acidolysis reaction of 1a in a MeOH/H 2 O (20 : 1) mixture.This solvent mixture and heating to 40 1C for 16 h afforded exclusively phosphine oxide 4a in low enantiomeric excess (Table 1, entry 2).By simply reducing the reaction time to 2 h, we were able to isolate the corresponding phosphinous acid 3a with an excellent yield and optical purity (499% ee, Table 2, entry 3).
The hydrolysis of 1b was also carried out stereospecifically to yield the corresponding phosphinous acid 3b in 499% ee and 63% yield (Table 1, entry 4).Hydrolysis of 1a and 1b took place with inversion of the configuration at the P-center as confirmed by X-ray crystallography of the corresponding benzoyl derivative (see ESI ‡). 12 The opposite enantiomer of 3a and and 3b were obtained when starting from (R P )-1a and (S P )-1b thus confirming that the process is completly stereospecific.These results indicate that the acidolysis of bulky aminophosphines is extremely sensitive to the nature of the incoming nucleophile.We believe that the minimal steric differences between methanol and water allow the latter to act as an efficient nucleophile, while making the former unreactive.
To further explore the scope of this transformation we submitted diastereomerically pure aminophosphines 5a and 5b (the synthetic precursors of aminophosphines 1a and 1b) to acidic hydrolysis in MeOH/H 2 O mixtures (Scheme 2).The reaction was slower than that achieved using the parent compounds 1a and 1b; 13 however, again, an increase in the reaction temperature and a longer reaction time produced the enantiomerically pure phosphinous acids 3a and 3b in 84% and 66% yields respectively.The hydrolysis of 5a and 5b is a practical approach to prepare the corresponding phosphinous acids; furthermore, it allows the recovery of the chiral auxiliary.
Optically pure P-stereogenic phosphinous acid boranes are attractive synthetic intermediates; however, they have barely been used in ligand synthesis. 14While optically enriched 3b has been prepared independently by Pietrusiewicz and Buono via resolution or H-menthylphosphinate technology, this is the first time that phosphinous acid 3a has been reported.It is known that mesylactivated phosphinous acids undergo effective nucleophilic reductions in the presence of NaBH 4 with inversion at the phosphorus center, 14 and we speculated whether this process could be extended to nucleophiles larger than hydride.Hence, phosphinous acid (S P )-3a was treated with Ms 2 O in the presence of triethylamine, and the reactions of the resulting mixed anhydride with several nucleophiles were studied (Table 2). 15nitial experiments using ammonia as the nucleophile indicated that, in solution, the phosphinyl-mesyl anhydride underwent slow racemization.Fortunately, a judicious choice of solvent and lowering of the reaction temperature to À20 1C permitted the nucleophilic substitution with ammonia in an almost completely stereospecific fashion.This produced (R P )-1a in 99% yield and 96% ee with the inverted configuration at the P-center (Table 2, entry 1).Importantly, primary amines also acted as efficient nucleophiles in this process, producing the corresponding aminophosphines in satisfactory yield and excellent enantiomeric excess (Table 2, entries 2-5).Chiral primary a-branched amines like (S) and (R)-1-phenylethylamine and phenylglycinamide also yielded the substitution products (R P )-10, (R P )-11 and (R P )-12 in 95-98% diastereomeric excess (Table 2, entries 6-8).A cyclic secondary amine like pyrrolidine and an aromatic amine like p-anisidine also efficiently produced the substitution products in 98 and 91% ee (Table 2, entries 9 and 10).In contrast, dibenzylamine was not good enough as a nucleophile and did not afford the expected aminophosphine (Table 2, entry 11).Also, oxygen nucleophiles failed to provide substitution products (Table 2, entry 12).In contrast, benzylthiol and thiophenol did provide the corresponding sulfides 17 and 18, albeit in low yield (22 and 36% respectively) (Table 2, entries 13 and 14).We attributed these low yields of isolation to the instability of these molecules.Finally, we tested the tertbutylphenylphosphinous acid borane 3b as the electrophile.Reaction with ammonia and (R)-phenylethylamine provided the corresponding aminophosphines 1b and 19 with excellent optical and diastereomeric purity (Table 2, entries 15 and 16).
The hydrolysis of aminophosphine boranes combined with the substitution on the resulting phosphinous acid represents a novel means of obtaining P*-ligands.To highlight the beneficial impact that this methodology could have in asymmetric catalysis, we followed the procedure shown in Table 2 to prepare a novel P-stereogenic N-phosphinooxazoline ligand and its corresponding cationic iridium complex (Scheme 3).Activation of (S P )-3a with mesyl anhydride and triethylamine and reaction with aminooxazoline 20 provided the borane-protected ligand 21 in 43% yield and in 498% dr, as determined by 1 H NMR. Borane was removed in neat pyrrolidine at 90 1C. 16We found that the approach of using neat pyrrolidine was superior to that of existing borane-deprotection methods like DABCO or neat diethylamine.Thus, from 21 and using a one pot-three reaction sequence, the cationic iridium complex 22 was obtained as an orange solid in 89% yield.With the iridium complex 22 in hand, we tested its performance in the asymmetric hydrogenation of a,b-unsaturated esters (Table 3). 17,18Reduction of ethyl trans-b-methylcinnamate under standard non-optimized conditions (50 bar of hydrogen in dichloromethane with 1 mol% of 22 as the catalyst) produced the (R) hydrogenated product in 95% ee (Table 3, entry 1).Under the same reaction conditions, hydrogenation of the isopropyl and cyclohexyl b-substituted cinnamates afforded the reduced products in 97% ee (Table 3, entries 2 and 3).Finally, paramethyl-substituted methylcinnamate afforded the reduced compound with complete selectivity (499% ee, Table 3, entry 4).
In summary, we have shown that the acid hydrolysis of bulky primary and secondary aminophosphine boranes occurs in a completely stereospecific manner with inversion of the configuration at the P-center to yield the corresponding optically pure phosphinous acid boranes.Also, we have demonstrated that, upon activation, phosphinous acid boranes undergo stereospecific nucleophilic substitution reactions at the P-center with amine nucleophiles.The potential of this process has been demonstrated with the synthesis of a P-stereogenic phosphinooxazoline ligand which has been applied to the asymmetric Ir-catalyzed hydrogenation of trans-b-alkylcinnamates, achieving selectivities of up to 99% ee.
We thank financial support from the Spanish Ministerio de Economia y Competividad (CTQ2014-56361-P and CTQ2013-40568P), the IRB Barcelona, the Generalitat de Catalunya (2014SGR670), the ICREA Foundation (M.Die ´guez and O. Pa `mies ICREA Academia awards).S.O.thanks the Generalitat de Catalunya for a FI fellowship. A. F.-G. thanks the CONACYT for a postdoctoral fellowship.

aScheme 2 Scheme 3
Scheme 2 Acidolysis of N-secondary aminophosphines and recovery of the chiral auxiliary.

Table 2
Substitution reactions with several nucleophiles

Table 3
Asymmetric hydrogenation of olefins using complex 22 as the catalyst