Revisiting the iridacycle-catalyzed hydrosilylation of enolizable imines

Jorge Pèrez-Miqueo a, Virginia San Nacianceno a, F. Borja Urquiola a and Zoraida Freixa *ab
aDepartment of Applied Chemistry, Faculty of Chemistry, University of the Basque Country (UPV-EHU), San Sebastián, 20018, Spain. E-mail: Zoraida_freixa@ehu.eus
bIKERBASQUE, Basque Foundation for Science, Bilbao, 48013, Spain

Received 15th June 2018 , Accepted 28th August 2018

First published on 28th August 2018


The hydrosilylation of enolizable imines using iridacycle catalysts has been investigated using a family of half-sandwich iridium(III) complexes. Several pitfalls had been detected when using standard procedures, and the use of in situ NMR spectroscopy appeared decisive to assess the actual reaction pathway. The results obtained clearly demonstrate that iridacycle-catalyzed hydrosilylation of enolizable imines with monohydrosilanes should be described as a domino reaction sequence, analogous to the one described by Oestreich et al. for boron-based Lewis acid catalysts. Equimolecular quantities of amine (b) and N-silylated enamine (c) were formed in the first stage, consuming 0.5 eq. of hydrosilane, which were subsequently converted into N-silylated amine (d) through a consecutive catalytic cycle using an additional 0.5 eq. of hydrosilane. However, when polyhydrosilanes were used as hydrosilylating agents, N-silylated amine (d) was directly obtained as the reaction product, but after a fast initial conversion, the reaction suffered from product inhibition. Additionally, iridacycles were also found to be active in the dehydrocoupling of Et3SiH and amines and in the hydrolysis of hydrosilanes, which can be competitive processes. These findings lead to an intricate mechanistic picture containing interrelated processes, which need to be taken into account for the development of more efficient, selective or enantioselective systems.


1. Introduction

In the past years, hydrosilylation reactions experienced a great progress, prompted by the development of more efficient processes based on both transition metal and main group Lewis acid catalysts.1–5 These new catalysts broadened the scope of substrates, achieving high selectivities mainly due to the variety of catalysts and reaction mechanisms. The pioneering platinum- and rhodium-based systems developed for alkene hydrosilylation were found to operate through the so-called Chalk–Harrod or modified Chalk–Harrod mechanism.6–8 A similar mechanism was postulated by Ojima for ketone hydrosilylation using Wilkinson's catalyst.9,10 These systems relied on two-electron oxidative addition and reductive elimination elementary steps, which narrowed considerably the scope of the catalysts (Scheme 1). The discovery of main group Lewis acid (LA) catalysts for hydrosilylation reactions constituted a milestone in the area.11,12 The LA catalysts follow an ionic mechanism, in which the LA enhanced the electrophilic character of the silicon center through activation of the Si–H bond of the hydrosilane (Scheme 1). Subsequent intermolecular nucleophilic attack of the unsaturated substrate on the silicon center forms the ionic intermediate [image file: c8cy01236a-t1.tifC[double bond, length as m-dash]XSiR3]+[HLA] (X = NR′′, O), considered the key intermediate.13–15 This type of hydrosilane activation, also coined “frustrated Lewis pair” bond activation,16,17 has also been proposed for several transition metal-based hydrosilylations,2,18–26 hydrosilane hydrolysis or alcoholysis,27–31 and CO2[thin space (1/6-em)]32,33 and ether reductions,2,34,35 among other reactions.36,37 The key adduct LA-hydrosilane was isolated for the first time by Brookhart in 2008 for an iridacycle catalyst and described as an η1-H(Si) iridium(III) complex (Scheme 1).38 Recently, the molecular structure of a half-sandwich iridacycle hydrosilane adduct has been obtained by X-ray diffraction showing that hydrosilylation processes with these catalysts proceed through η2-(H–Si) metal-induced activation of the hydrosilane, which is best described as a Lewis donor–acceptor pair of an iridium hydride and a silylium cation (Scheme 1).18,30
image file: c8cy01236a-s1.tif
Scheme 1 Hydrosilane activation in (A) Chalk–Harrod and modified Chalk–Harrod and (B) Lewis acid hydrosilylation mechanisms.

While transition metal-catalyzed reduction of carbonyl groups with hydrosilanes has been widely studied, the related hydrosilylation of imine functionalities remains less explored. Recently, important contributions in this area have been made by the groups of Michon, Nikonov, and Oestreich. After seminal work by Djukic et al. showing the catalytic activity of half-sandwich iridium(III) metallacycles in tandem hydroamination–hydrosilylation/protodesilylation of alkynes (involving hydrosilylation of in situ generated imines),39 Michon et al. published a detailed study on the latter process in 2015.40 These Ir catalysts were able to reduce a wide variety of aldimines and ketimines to the corresponding amines (after protodesilylation). Although the experimental evidence did not permit one to conclude whether the system operated through the classical Chalk–Harrod or through a LA mechanism, the involvement of an iridium-hydrido intermediate species was assessed.1,40 Recent evidence of LA activation of hydrosilanes by half-sandwich iridacycles18,30 reinforces the idea that imine hydrosilylations with these catalysts could proceed through a mechanism analogous to the one described originally by Piers for boron-based LA catalysts.15

In 2013, Oestreich et al. observed that borane-promoted hydrosilylation of enolizable imines proceeds through a sequential reaction pathway rather than through the classical Piers mechanism, and they identified the corresponding amine and N-silylated enamine as equimolecular reaction intermediates.41 This finding constitutes a milestone for the development of enantioselective variants of the process. In spite of the relatively large number of organometallic catalysts described for imine hydrosilylations and the similitude established between the boron-based and the organometallic-based LA hydrosilane activation mechanism,42 to the best of our knowledge, this sequential pathway has never been observed for an organometallic catalyst.

In 2014, Oestreich also showed that N-silylated enamines could be obtained as main reaction products from enolizable imines under hydrosilylation conditions when using a tethered ruthenium catalyst (Ohki–Tatsumi complex).43,44 This reactivity was attributed to the capability of the bifunctional catalyst to cleave the Si–H bond of hydrosilanes, forming a ruthenium hydride that contains an electrophilic silicon atom coordinated to the thiolate residue of the ligand (Scheme 2). After the transfer of the electrophilic silylium cation to the imine N atom, the ruthenium hydride acts as a Brønsted base that deprotonates the silyliminium ion rather than as a hydride donor; this way the dehydrogenative pathway is favoured over the reductive coupling. Although several half-sandwich iridium(III) complexes containing the hydroxypyridine motif showed the involvement of the bifunctional ligand in several multifunctional catalytic processes,45–47 their effect on hydrosilylation reactions has not yet been studied.


image file: c8cy01236a-s2.tif
Scheme 2 Hydrosilane activation by a tethered ruthenium catalyst.44,48

In the present study we aimed to establish the actual reaction mechanism operative in the hydrosilylation of enolizable imines using half-sandwich iridium(III) complexes and how it depends on the nature of the hydrosilane. Its resemblance with the one proposed by Oestreich41 for boron-based LA catalysts together with the importance of coexistent competitive processes will be discussed. Eventually, the catalytic behaviour of related multifunctional hydroxypyridine catalysts will be presented.

2. Experimental

2.1 General procedures

All solvents were dried and purified by known procedures and freshly distilled under nitrogen from appropriate drying agents prior to use. CDCl3 and CD2Cl2 were freeze–pump–thaw degassed and stored under N2 over a 4 Å molecular sieve. All manipulations and reactions involving air- and/or moisture-sensitive organometallic compounds were performed under an atmosphere of dry nitrogen using standard Schlenk techniques. NMR spectra were referenced against an internal capillary of TMS for catalytic experiments, and added TMS (0.1% v/v) or residual signals of solvent for characterization spectra.

Hydrosilanes, DSiEt3 (>97% D incorporation) and acetophenone-D3 (>99% D incorporation) were acquired from Sigma-Aldrich and used without further purification. Complexes 1,49,502,45551N-(1-phenylethyl)aniline, N-(1-phenylethylidene)aniline, N-(1-phenylethylidene)aniline-D3,52,53 and NaBArF24[thin space (1/6-em)]54 were synthesized using standard procedures. Imines were stored under nitrogen at 10 °C.

2.1.1 General procedure for the hydrosilylation reaction. N-(1-Phenylethylidene)aniline (0.1946 mmol, 1 eq.), the selected iridium(III) catalyst (× mol%), NaBArF24 (2× mol%) and an internal capillary containing a solution of TMS (10% v/v in CDCl3) were introduced in an NMR tube which was then purged 3 times with vacuum/nitrogen cycles. Under nitrogen, CDCl3 (0.5 mL) was added, the tube was sealed with a septum, and an initial spectrum was registered. Subsequent addition of the corresponding hydrosilane (0.2335 mmol, 1.2 eq.) through the septum was considered the beginning of the reaction. The reaction progress was monitored by 1H NMR spectroscopy.

When specified, aliquots (0.1 mL) were taken at selected time intervals, filtered through a Celite™ pad (0.5 mm in a Pasteur pipette) and washed with 3 mL of CH2Cl2. The solvent was evaporated and the sample was analyzed by 1H NMR spectroscopy in CDCl3.

2.1.2 Amine (N-(1-phenylethyl)aniline) hydrosilylation in a closed system. Catalytic reactions were carried out in a 10 mL glass reactor using the kinetic kit (MOTM) series X102 from manonthemoontech,55 which permits electronic monitoring of the variation of pressure and temperature of the gas phase inside a closed glass reactor. N-(1-Phenylethyl)aniline (0.3892 mmol, 1 eq.), 0.002 mmol (1.0 mg) of precatalyst 1 (0.5 mol%) and NaBArF24 (0.004 mmol, 3.4 mg) were stirred in 1.0 mL of CH2Cl2. Addition of 74.6 μL (0.4670 mmol) of Et3SiH to this mixture was considered the starting time of the reaction.
2.1.3 Hydrolytic dehydrogenation of Et3SiH in a closed system. Catalytic reactions were carried out in a 10 mL glass reactor using the kinetic kit (MOTM) series X102 from manonthemoontech.55 0.002 mmol (1.0 mg) of precatalyst 1 (1 mol%) and NaBArF24 (0.004 mmol, 3.4 mg) were stirred in 1 mL of CDCl3 and 7 μL of distilled water. Addition of 74.6 μL (0.4670 mmol) of Et3SiH to this mixture was considered the starting time of the reaction.

3. Results and discussion

3.1 Optimization of the experimental procedure

Preliminary experiments on the hydrosilylation/protodesilylation reaction of N-(1-phenylethylidene)aniline, using [IrCp*(ppy)Cl]Cl (1) as a precatalyst, were conducted following the experimental procedure and conditions described by Michon et al. for this specific reaction (precatalyst 0.1 mol%, NaBArF24 0.2 mol%, Et3SiH, 1.2 eq. [imine] = 0.34 M, 2 mL CH2Cl2).40 Although the results obtained confirmed the efficiency of the catalytic system, inconsistent conversions were obtained over several runs (0.5 h, conv. 33–43%; 2 h, conv. 47–88%; 24 h, 91–100%). To make sure that the protodesilylation procedure commonly used for this reaction (filtration of 0.1 mL aliquots of reaction through a Celite™ pad and washing with CH2Cl2) was not altering the measured conversions, the progress of the hydrosilylation reaction was also monitored by in situ1H NMR spectroscopy in CD2Cl2 (Fig. 1).
image file: c8cy01236a-f1.tif
Fig. 1 Time-resolved 1H NMR study of the hydrosilylation of N-(1-phenylethylidene)aniline using precatalyst 1 (300 MHz). Reaction conditions: precatalyst 1 (0.13 mg, 0.00025 mmol), NaBArF24 (0.46 mg, 0.0005 mmol), N-(1-phenylethylidene)aniline (50.9 mg, 0.2591 mmol), Et3SiH (50.0 μL, 0.3130 mmol), 0.7 mL CD2Cl2, internal capillary TMS (10% v/v in CDCl3).

The spectra obtained showed that under these catalytic conditions, all of the starting imine (a in Fig. 1) was consumed in only 400 minutes, but the N-silylated amine (d in Fig. 1) was not the only reaction product. Consumption of imine (a) produced new signals assigned to the corresponding amine (b in Fig. 1) and N-silylated enamine (c in Fig. 1). Signals attributed to N-silylated amine (d) started to appear only at a later stage (see Fig. 1). A reaction profile was constructed based on the integration of selected signals of imine- and silane-derived products (Fig. 2). On the one hand, the sigmoid shape of the curve that represents N-silylated amine (d) formation together with the simultaneous and equimolecular formation and consumption of amine (b) and N-silylated enamine (c) suggests that the latter were reaction intermediates towards the formation of the N-silylated amine (d), which leads to a simplified reaction scheme (2ab + c → 2d), as found before by Oestreich for borane-based LA catalysts.41 On the other hand, the hydrosilane consumption profile showed a fast decay when imine (a) was still present and a much slower consumption when this substrate was consumed, which suggests that the second reaction was a slower process (see B in Fig. 2).


image file: c8cy01236a-f2.tif
Fig. 2 Reaction profiles showing (A) imine-derived and (B) silane-containing product distribution according to in situ1H NMR spectroscopy. Red arrows indicate the points at which aliquots were taken. C) Bar diagram showing imine-derived product distribution at different reaction times by in situ1H NMR spectroscopy (A, 25 min; B, 130 min; C, 791 min; D, 24 h) and after acidic workup (A′, B′, C′, D′). Reaction conditions: precatalyst 1 (0.13 mg, 0.00025 mmol), NaBArF24 (0.46 mg, 0.0005 mmol), N-(1-phenylethylidene)aniline (50.9 mg, 0.2591 mmol), Et3SiH (50.0 μL, 0.2310 mmol), 0.7 mL CD2Cl2, internal capillary TMS (10% v/v/CDCl3).

Intermediates b and c were never observed in the earlier investigations of this reaction studied by protodesilylation of reaction aliquots.40 To study the fate of these intermediates after the standard workup procedure, several aliquots (0.1 mL) were taken at 25 min, 130 min, 791 min and 24 h, submitted to the protodesilylation treatment, and analyzed by 1H NMR spectroscopy. It is worth mentioning that N-silylated enamine (c) was not observed in any of the samples after acidic workup, which hampered other authors to suspect about its intermediacy in the process. Only the last sample, in which most of the starting imine (a) had already been transformed into N-silylated amine (d) (100% conversion, 85% selectivity, according to in situ1H NMR analysis) showed a comparable conversion after acidic treatment. 1H NMR analysis of the rest of the aliquots after workup showed conversions considerably different from the ones observed by in situ NMR spectroscopy (Fig. S3 in the ESI). At short reaction times the inferred conversions were much larger than the actual ones (A vs. A′ in Fig. 2C). This could be attributed to the fact that the reaction could continue (even at a higher reaction rate!) during the time needed for the workup (approx. 10 minutes). Surprisingly, when the starting imine (a) was fully consumed according to in situ1H NMR spectroscopy (C in Fig. 2C), the sample after acidic workup showed amine (b) and imine (a) in an 85/15 ratio, respectively. This result could be explained by taking into account that not only does the reaction continue during the workup process, but the formed N-silylated enamine (c) evolves towards the highly unstable enamine (a′), which tautomerizes to the starting imine (a) (Scheme 3).44 Both facts have counterproductive effects on the presumed conversion measured after protodesilylation, and the final effect on the calculated conversion is difficult to predict. The difference in the measured conversions is smaller at intermediate reaction times (B vs. B′ in Fig. 2C). It should be noted that the total amount of amine (b) observed after acidic workup at this stage does not correspond to that of the N-silylated amine (d) formed during the catalytic process (only ∼5%), and thus it must be a direct reaction product.


image file: c8cy01236a-s3.tif
Scheme 3 Evolution of N-silylated enamine (c) and N-silylated amine (d) during the protodesilylation workup.

These results plead for the use of in situ1H NMR spectroscopy as a standard methodology to study the reaction progress. In order to reduce the analysis cost, the reaction was studied using CDCl3 as solvent instead of CD2Cl2, which gave analogous results (see Fig. S4 and S5 in the ESI). Acidic treatment of the final reaction mixtures rendered amine b in quantitative yields after column chromatography.

Blank experiments conducted in the absence of iridacycle 1 confirmed the involvement of the organometallic catalyst in the process. When the reaction was carried out at larger catalyst loadings (0.5, 1 and 2 mol%) the same reaction profile was observed, although the reaction proceeded at much higher rates (see Fig. S6–S11 in the ESI). Using 2 mol% of catalyst, the starting imine (a) was consumed in only 20 min, and 80 min sufficed to attain the maximum conversion to N-silylated amine (d) (∼97%). Analysis of the TOF at 20% and 80% conversion (based on imine (a) consumption and N-silylated amine (d) formation, respectively) suggested a first-order dependence of the reaction rate on the iridacycle 1 concentration for both consecutive processes (Fig. S12 in the ESI).

When the reactions were not run under a rigorously dry nitrogen atmosphere, much faster (but irreproducible) reaction rates were observed (Fig. S13 in the ESI).§ In a controlled experiment, a clear rate enhancement was observed upon fast exposure of the sample to wet air (see Fig. S14 and S15 in ESI). This acceleration was attributed to diffusion of moisture in the sample which could stabilize polar transition states and reaction intermediates.56,57 Oxygen was not considered responsible for the rate enhancement as it should have a detrimental effect, if any, on the activity. Additionally, the slight excess of amine (b) (compared to N-silylated enamine (c)) observed in these experiments was also consistent with the presence of adventitious H2O in the samples. To confirm this hypothesis several experiments were run under strictly inert conditions in the presence of controlled quantities of H2O (0.5 mol% catalyst loading). When the reaction was run with 0.1 equivalent of H2O per substrate, all the imine (a) was fully consumed in 13 minutes and converted to N-silylated amine (d) (84%) and amine (b) (16%) in only 40 minutes (Fig. S16 in the ESI). If the H2O content was increased to 1 equivalent per imine (a), in the first spectra acquired (4 minutes after hydrosilane addition) 87% conversion was already observed, with amine (b) being the main reaction product (Fig. S17 in the ESI). It is worth mentioning that in this high-water-content reaction an evident gas evolution was observed upon addition of the hydrosilane, and a signal indicative of the presence of H2 was observed in the 1H NMR spectra at 4.56 ppm (Fig. S18 in the ESI). For this reason, it could not be concluded if the excess of amine (b) observed was formed through hydrolysis of the formed N-silylated amine (d) or by a consecutive metal-catalyzed hydrosilane hydrolysis and imine (a) hydrogenation (routes A and B in Scheme 4, respectively).


image file: c8cy01236a-s4.tif
Scheme 4 Possible routes for the formation of amine (b) in the presence of H2O under catalytic conditions.

To determine the process that leads to the observed excess of amine when H2O is present, an independent experiment was run under identical conditions but notably in the absence of imine (a). The immediate formation of triethyldisiloxane (98%) and silanol (2%) with concomitant gas evolution was observed in less than 4 minutes by 1H NMR spectroscopy (Fig. S19 in the ESI). This reaction was also performed in a closed reactor equipped with a pressure transducer;55 for this experiment 95% conversion was calculated based on the final pressure. These results confirmed the capacity of iridacycle 1 to act as a very efficient Et3SiH hydrolysis catalyst. This activity is not surprising, since the dehydrosilylation of alcohols with Et3SiH using iridacycle 1 has been assessed recently.30 Most probably, the mechanism responsible for this process is the metal-catalyzed hydrosilane activation followed by a nucleophilic attack of water on the activated silane (as described for other iridium-based systems) (see eqn (1)–(5)).27,58 No hydrosilane hydrolysis occurred in the absence of catalyst even after 24 h of reaction, as observed already by other authors.28

 
image file: c8cy01236a-t2.tif(1)
 
image file: c8cy01236a-t3.tif(2)
 
image file: c8cy01236a-t4.tif(3)
 
image file: c8cy01236a-t5.tif(4)
 
image file: c8cy01236a-t6.tif(5)

To assess the capability of compound 1 to catalyze also the subsequent hydrogenation of the imine (a), a screw-capped Young NMR tube was charged with N-(1-phenylethylidene)aniline (38.00 mg, 0.1946 mmol), 0.5 mol% iridacycle 1 (0.50 mg, 0.001 mmol), NaBArF24 (1.70 mg, 0.0019 mmol), an internal capillary TMS (10% v/v/CDCl3) and 0.5 mL of distilled CDCl3. The tube was pressurized with 2.8 bar of H2, to simulate the maximum quantity of hydrogen liberated by hydrolysis of the hydrosilane. The evolution of the reaction was monitored by 1H NMR spectroscopy at regular time intervals. No reaction was observed after 24 h, which discards the capability of iridacycle 1 to act as an imine (a) hydrogenation catalyst at such low hydrogen pressures and confirms that even if iridacycle 1 were an active catalyst for hydrosilane hydrolysis, the excess of amine (b) observed in the high-water content reaction stems from hydrolysis of the formed N-silylated amine (d).

The results described above confirm that exposure of the reaction mixture to moisture not only accelerates the reaction progress, most probably through a stabilization of polar transition states, but it also triggers a competitive metal-catalyzed hydrolysis of the hydrosilane and the eventual hydrolysis of the formed N-silylated amine (d). These observations are also consistent with the “acceleration” of the reaction progress observed when aliquots were submitted to acidic workup in an open-air atmosphere (vide supra).

Due to the lack of reproducibility obtained when the reactions were not conducted under dry N2, all reactions described in this manuscript were performed under inert conditions. The equimolarity among the N-silylated enamine (c) and the amine (b) formed and consumed during the reaction progress was taken as an indirect proof of the inert conditions of the experiment.

3.2 Mechanistic studies

As mentioned above, the formation of equimolecular quantities of N-silylated enamine (c) and amine (b) as intermediates in the hydrosilylation of enolizable imines was already observed by Oestreich using electron-deficient boron LA catalysts.41 Their existence was attributed to borohydride reduction of an iminium ion (II in Scheme 5) generated in the N-silylated enamine (c) formation step (Scheme 5).
image file: c8cy01236a-s5.tif
Scheme 5 Amine (b) and N-silylated enamine (c) formation steps in LA-catalyzed imine hydrosilylation, as proposed by Oestreich.41

To give support to their mechanistic proposal, they performed a series of well-designed experiments: initially, they confirmed the product-forming ability of these intermediates by reacting, under catalytic conditions, equimolecular quantities of amine (b), N-silylated enamine (c), and hydrosilane in the presence of 3 mol% electron-deficient boron catalyst. They also studied independently the activity of the LA catalyst in the dehydrogenative hydrosilylation of amine (b) demonstrating the capability of the borane-catalyst to activate the hydrosilane and to silylate the amine nitrogen atom.41 In our case, we did not have access to isolated samples of N-silylated enamine (c), which hampered us from performing the first experiment described by Oestreich. However, we were able to study separately the capacity of iridacycle 1 to act as a catalyst in dehydrogenative silylation of amine (b) by in situ1H NMR spectroscopy (see Fig. S20 and S21 in the ESI). The data obtained were in agreement with a first-order decay of the signal due to the N-(1-phenylethyl)aniline (b) and hydrosilane, and exponential formation of N-silylated amine (d) with a calculated pseudo-first order constant kobs of 0.0048 min−1.

A singlet at 4.6 ppm, which could be assigned to evolved hydrogen, was detected at the early stage of the reaction, while at larger conversions it was masked by the overlapping methine signal of the formed N-silylated amine (d). These results evidenced the capacity of iridacycle 1 to catalyze the dehydrocoupling of Et3SiH and N-(1-phenylethyl)aniline (b) to form N-silylated amine (d), most probably through the intermediacy of a silylammonium ion (III) and an iridium(III)-hydride species, according to the reaction sequence 1 + 6 + 7 + 4, in analogy to Oestreich's proposal.41

 
image file: c8cy01236a-u1.tif(6)
 
image file: c8cy01236a-u2.tif(7)

Next we looked at the kinetics of a standard hydrosilylation reaction, viz. the initial decay of imine a. The curve was analyzed by a pseudo first-order reaction, which strictly speaking may not be correct, as a is involved in two reactions, 8 and sequential 10. The rate of decay was found at kobs (0.024 min−1), much larger than for the dehydrogenative amine hydrosilylation pathway (1 + 6 + 7 + 4).

Taking into account that the concentration of amine (b) is never larger than 0.13 M during N-(1-phenylethylidene)aniline (a) hydrosilylation, reactions 6 and 7 are most probably not competitive under imine (a) hydrosilylation conditions, although a contribution to the final production of N-silylated amine (d) cannot be completely ruled out.

Following Oestreich's methodology,41 deuterium-labeling experiments were also conducted using precatalyst 1. When Et3SiD was used as the “hydro”silane, in situ1H and 2H NMR spectroscopy were used to study the fate of the hydride/deuteride of the hydrosilane during the whole catalytic process (Fig. S22 and S23 in the ESI). The spectra obtained confirmed its selective incorporation on the methine group of both the amine (bD) and the N-silylated amine (dD) (Scheme 6). Additionally, although virtually identical profiles for the consumption of imine and formation of the intermediates were registered (within the experimental error), compared to the ones obtained when HSiEt3 was used as the hydrosilane (see Fig. S24 in the ESI), a slow consumption of the intermediates b and c and formation of N-silylated amine was observed. Due to the complexity of the reaction sequence, it was not possible to calculate a KIE, but this observation indicates that, in fact, the N-silylated amine (d) is formed through two consecutive catalytic processes. In the first one, consumption of half of the hydrosilane to form equimolecular quantities of b and c, most probably the cleavage of the Si–H/D or Ir–H/D bond is not involved in the rate-determining step (rds). The second is a much slower process and consumes an additional 0.5 eq. of hydrosilane, converting these intermediates (b and c) into two equivalents of the N-silylated amine (d). In this second process, the slightly slower reaction rate observed when deuterium labelled silane was used suggested that either Si–H/D or Ir–H/D bond cleavage could be involved in the rds, although it could not be established unambiguously.


image file: c8cy01236a-s6.tif
Scheme 6 Fate of the hydride of the hydrosilane according to in situ1H and 2H NMR spectroscopy.

To confirm that the origin of the amine (b) was intermolecular proton transfer from the methyl group of the silyliminium cation (I in Scheme 5) to the unreacted imine (a), an experiment was run using N-(1-phenylethylidene)aniline deuterium-labeled at the α-carbon of the imine (aCD3). When this substrate was submitted to a standard hydrosilylation procedure (using 0.5 mol% 1 as precatalyst) the in situ2H NMR spectra obtained showed a broad singlet around 3.95 ppm corresponding to the N–D of the amine (bCD3-ND) in addition to the expected signals due to deuterated methyl and methylene groups of all the imine-derived compounds (see Fig. S25 in the ESI and Scheme 7).


image file: c8cy01236a-s7.tif
Scheme 7 Deuterium distribution in the reaction products when aCD3 was used as the substrate according to in situ2H NMR spectroscopy.

All these experiments are consistent with a two-step domino reaction mechanism analogous to the one proposed by Oestreich for electron-deficient borane catalysts. Both cycles are closely interrelated and are based on the same catalyst and hydrosilane activation mechanism. The entire process starts with the iridium(III)-catalyzed hydrosilane activation (eqn (1)), in agreement with recent evidence for related reactions and catalysts.18,30 This step is common for the competing hydrosilane hydrolysis and amine hydrosilylation (vide supra). The different reaction rates observed for the three processes under identical catalytic conditions reflect the nucleophilic character of H2O, imine (a) and amine (b), respectively. This tendency points to a common mechanism based on a subsequent nucleophilic attack of either H2O, amine (b) or imine (a) on the metal-coordination activated hydrosilane. The different activities depending on the nucleophile together with the lack of Si–D KIE observed on the imine (a) consumption under hydrosilylation conditions points to hydrosilane activation not being the rds of the process.

Following this reasoning, after hydrosilane activation (eqn (1)), nucleophilic attack of the imine (a) on the activated silane would cleave the Si–H bond, rendering a silyliminium cation I and an iridium hydride (eqn (8)). Most probably, in the case of non-enolizable imines, hydride transfer from the iridium hydride to silyliminium cation I would form the N-silylated amine (d) and regenerate the catalyst (eqn (9)), following the “classical” mechanism described by Piers for electron-deficient LA catalysts (cycle A in Scheme 9). The contribution of this process to the total reaction outcome (N-silylated amine (d) formation) would depend on the relative Brønsted acidity of the protonated LA catalyst and the basicity and concentration of the imine (k9vs. k10) (vide infra).

 
image file: c8cy01236a-u3.tif(8)
 
image file: c8cy01236a-u4.tif(9)

Not surprisingly, in the case of substrates containing α-protons, at low conversions, hydride transfer from the iridium hydride to silyliminium cation (I) is a slow process compared to the abstraction of one of the α hydrogens by the unreacted imine (a), forming N-silylated enamine (c) and an iminium cation (II) (eqn (10)). It is worth noting that this competitive reaction, described only for boron-based LA catalysts, should be operative in all the cases in which enolizable imines are used as substrates as it is a catalyst-independent process. The iminium cation II could also suffer the attack of an imine (a) as a base (especially at low conversions), but this is an unproductive reaction as II and the enamine a′ (which tautomerizes to the most stable imine a) would be formed (Scheme 8).


image file: c8cy01236a-s8.tif
Scheme 8 Unproductive reactivity between iminium cation II and imine (a).

Eventually, hydride transfer from iridium hydride to cation II would occur, forming the amine b and regenerating the active catalyst (eqn (11)). This catalytic sequence (B in Scheme 9) would be mostly operative at low conversions and would transform imine (a) into equimolecular quantities of amine (b) and N-silylated enamine (c), using 0.5 eq. of hydrosilane, through the equation sequence 1 + 8 + 10 + 11.

 
image file: c8cy01236a-u5.tif(8)
 
image file: c8cy01236a-u6.tif(10)
 
image file: c8cy01236a-u7.tif(11)


image file: c8cy01236a-s9.tif
Scheme 9 Coexisting catalytic cycles under imine hydrosilylation conditions using iridacycle 1 as catalyst.

When amine (a) and N-silylated enamine (c) concentrations start building up, a second catalytic cycle (C in Scheme 9) becomes operative. After hydrosilane activation by the iridacycle (eqn (1)), nucleophilic attack of the amine (b) on the activated silane would render N-silylated ammonium cation (III) and an iridium hydride (eqn (6)). The feasibility of this elementary step has been experimentally confirmed for the dehydrogenative Si–H coupling of Et3SiH and N-(1-phenylethylidene)aniline (b) (vide supra). Deprotonation of III by the N-silylated enamine (c) would render one equivalent of N-silylated amine (d) and iminium cation (I) (eqn (12)). The latter, could either re-enter into cycle B or at low imine (a) concentrations would eventually evolve to N-silylated amine (d), following eqn (9).

 
image file: c8cy01236a-u8.tif(6)
 
image file: c8cy01236a-u9.tif(12)
 
image file: c8cy01236a-u10.tif(9)

To confirm the synergism of both catalytic cycles (B and C), an experiment was conducted using only ∼0.5 equivalent of hydrosilane. When the reaction was run using a low catalyst loading (0.1 mol%), it apparently stopped when all the hydrosilane was consumed (∼400 min) (Fig. 3). At this point there was a 75% conversion, with amine (b), N-silylated enamine (c) and N-silylated amine (d) being the reaction products in an approximate ratio of 1/1/0.5 (see Fig. 3, top). The sigmoid shape of the curve representing N-silylated amine (d) formation suggested that it started forming after a certain induction period, most probably needed for building up concentrations of the intermediates b and c. This observation is consistent with the fact that cycle A (Scheme 9) is not operative at low conversions, and N-silylated amine (d) arises from deprotonation of silylammonium cation (III) (eqn (12)). Although apparently the reaction did not evolve further once the hydrosilane was consumed, at extended reaction times (24 h) the product distribution expected for a system following Pier's mechanism (A in Scheme 9) was observed: “unreacted” imine (a) (48%), N-silylated amine (d) (42%) and amine (b) (10%) were the only reaction products. When the reaction was conducted at a larger catalyst loading (0.5 mol%, Fig. 3) the same conversion and similar initial product distribution was observed in only 200 min, but the evolution of this species towards the final distribution could already be intuited (which is coherent with a positive order dependence of the reaction rate on catalyst concentration for both cycles B and C) (Scheme 9). To accelerate this evolution, in accord with our former observations, the reaction was performed under a non nitrogen-protected atmosphere (Fig. 3, bottom). A clear evolution of the initially formed products towards imine (a) and N-silylated amine (d) upon full consumption of Et3SiH was observed. These results are consistent with catalytic cycle B being composed of reversible elementary steps driven towards the product due to the consumption of both amine (b) and N-silylated enamine (c) through cycle C. When all the hydrosilane was consumed, the formed amine (b) and N-silylated enamine (c) were trapped into these reversible steps and eventually evolved back towards imine (a) and N-silylated amine (d) triggered by the irreversible liberation of the latter through cycle A. As a result of these equilibria, a single measurement of the reaction progress at extended reaction times showed the “expected” ∼50% conversion towards the N-silylated amine (d) if only 0.5 eq. of hydrosilane was used.


image file: c8cy01236a-f3.tif
Fig. 3 Reaction profile showing imine-derived product distribution according to in situ1H NMR spectroscopy. Imine (a), blue circles; amine (b), purple circles; N-silylated enamine (c) orange circles; N-silylated amine (d), green circles. Reaction conditions: precatalyst 1 (top: 0.1 mg, 0.0002 mmol; bottom: 0.5 mg, 0.0010 mmol), NaBArF24 (top: 0.34 mg, 0.0004 mmol; bottom: 1.7 mg, 0.0019 mmol), N-(1-phenylethylidene)aniline (38.0 mg, 0.1946 mmol), Et3SiH (15.6 μL, 0.0973 mmol), 0.5 mL CDCl3, internal capillary TMS (10% v/v/CDCl3).

All these experiments offer an intricate picture of interrelated reactions and equilibria operative under catalytic conditions. The eventual reaction outcome depends on the actual relative concentration of reagents, intermediates and adventitious H2O (see Scheme 9).

3.3 Screening of hydrosilanes

The reaction mechanism of the hydrosilylation of N-(1-phenylethylidene)aniline using other hydrosilanes (some of which were already described in the original publication by Michon) was also studied.40 All the reactions were analyzed by in situ1H NMR spectroscopy and the speciation diagram constructed based on the integration of signals selected individually to minimize overlapping with other signals. The results obtained are summarized in Fig. 4. Time-resolved NMR spectra are collected in Fig. S26–S39 in the ESI. Visual inspection of the different reaction profiles obtained at 0.5 mol% catalyst loading clearly indicated that amine (b) and N-silylated enamine (c-type) were observed as equimolecular reaction intermediates not only with Et3SiH, but also when Ph3SiH, MePh2SiH or Me2PhSiH was used as the hydrosilane. In the latter, the extremely fast reaction observed hampered the detection of any reaction intermediate under these catalytic conditions, but they were clearly observed at lower catalyst loadings (0.1 mol%). When PhSiH3 or Ph2SiH2 was used as a hydrosilylating agent a completely different reaction profile was observed. Not only was there no N-silylated enamine (c-type) detected along the reaction course, but after a fast and initial conversion of part of the imine (a) to N-silylated amine (d-type), the reaction evolved towards complete conversion at much lower reaction rates. The observed amine (b) (5–10%), should be considered a side product rather than a reaction intermediate in these cases.
image file: c8cy01236a-f4.tif
Fig. 4 Hydrosilane screening. Reaction profiles showing imine-derived products distribution, according to in situ1H NMR spectroscopy. Reaction conditions: precatalyst 1 (0.1 or 0.5 mol%), NaBArF24 (0.2 or 1.0 mol%), N-(1-phenylethylidene)aniline (38.0 mg, 0.1946 mmol), hydrosilane (37.3 μL, 0.2335 mmol), 0.5 mL CDCl3, internal capillary TMS (10% v/v/CDCl3).

The observed profile would be compatible with a fast initial reaction suffering from product inhibition. This is not surprising since the formed Si–H containing N-silylated amine could compete with the unreacted hydrosilane for the active site of the catalyst.

The fast initial reaction and the absence of amine and N-silylated enamine as intermediates imply a very fast hydride transfer from the iridium hydride to silyliminium cation I compared to the abstraction of a methyl hydrogen by the unreacted imine (k9k10).

A rate acceleration (and different selectivity) when using di- and trihydrosilanes compared to monohydrosilanes was already reported for rhodium-catalyzed ketone hydrosilylation.59–61 With these rhodium-based systems, the hydrosilane activation occurs through two-electron oxidative addition of the hydrosilane to the metal center, according to the so-called Ojima's mechanism. The different activity and selectivity when polyhydrosilanes were used was initially explained by Zheng and Chan.59 In contraposition with the traditional Ojima mechanism, based on the insertion of a coordinated ketone into the Si–Rh bond rendering a Rh(III) alkyl species, they proposed an alternative mechanism in which, after oxidative addition of the hydrosilane, the ketone “coordinates” to the silicon atom and then inserts into the Si–H bond, rendering a Rh(III) silyl species. More recently, Gade has revisited this reaction, proposing a silylene-based mechanism which could be operative in the case of polyhydrosilanes (Scheme 10).61


image file: c8cy01236a-s10.tif
Scheme 10 Initial steps of the different mechanisms proposed for ketone hydrosilylation with rhodium catalysts.

In the case of iridacycle 1, oxidative additions of hydrosilane are not very likely to occur as they would require the partial ring slippage of Cp*or decoordination of phenylpyridyl ligands. Additionally, as mentioned before, the key adduct LAH-silane has already been isolated by Djukic for the cation of 1 and established as an active species in the nitrile-to-amine hydrosilylation processes.18 Assuming this activation procedure, and inspired by the mechanism of Gade and Zheng and Chan, we considered that the fast initial conversion observed in the case of polyhydrosilanes could proceed through a direct attack of the imine on the activated silane and insertion into the Si–H bond of the hydrosilane, rendering a very reactive iminium ion (I′ in Scheme 11) (large k9). Alternatively, formation of I′ can be also explained by the nucleophilic attack of the imine on the activated silane, releasing silyliminium cation, I and the tautomeric equilibrium of the latter. The large k9 observed, compared to the one observed for R3SiH derivatives, could be due to the less steric hindrance of intermediates I and I′ (Scheme 11). Unfortunately, we could not go further into more detailed mechanistic studies, so this proposal is purely speculative.


image file: c8cy01236a-s11.tif
Scheme 11 Postulated mechanism for imine hydrosilylation using polyhydrosilanes and iridacycle 1 as catalysts.

3.4 Screening of catalysts

With the aim to study the effect of a potentially bifunctional 2-hydroxy-6-phenylpyridyl ligand on the complex, two complexes containing such functionality (2 and 6) were synthesized. An analogous methoxy-substituted (3), the N–O chelate (4) and the cationic complex (5) were also synthesized for comparative purposes (Chart 1) (see the ESI for synthetic details).
image file: c8cy01236a-c1.tif
Chart 1 Iridium(III) complexes studied as precatalysts.

The catalytic behaviour of complexes 2–6 under standard conditions (0.1 mol% cat. loading in CDCl3 using Et3SiH as the hydrosilane). The results obtained are presented in Fig. S40–S50 in the ESI. In all cases the same reaction profile was observed, i.e. amine (b) and N-silylated enamine (c) were formed and consumed at equal rates (within the experimental error). N-Silylated amine was the main reaction product at extended reaction times. The similar profiles and reaction rates obtained with catalysts 1 and 2 discarded the involvement of the hydroxyl group of phenylpyridine in the catalytic process, indicating that this system did not resemble the bifunctional catalyst designed by Oestreich.44 The replacement of the coordinated chloride by an aquo ligand slightly accelerated the catalytic process, as inferred by comparison of reaction profiles obtained with catalysts 5 and 1. This acceleration could be due to an easier creation of the vacant site needed for the activation of the hydrosilane or to the accelerating character of the released aquo ligand. Interestingly, when catalyst 5 was studied without NaBArF24 no catalytic activity was observed. This was attributed to the insolubility of compound 5 in the reaction medium. In this case the role of NaBArF24 was that of a solubilizing anion rather than a chloride abstractor. Independent 1H NMR experiments were performed to confirm this hypothesis.

In the case of compound 6, a rather low catalytic activity was observed. In this case it was explained based on an additional stabilization of the coordinated acetate through hydrogen bonding with the hydroxyl group of the phenylpyridine ligand. The molecular structure of this compound was confirmed by X-ray diffraction of a crystalline sample, obtained by slow diffusion of ether on a CDCl3 solution of the complex (see Fig. S73 in the ESI). The molecular structure of 6 confirmed this hydrogen-bonded arrangement in the solid state.

According to this preliminary screening, compounds 3 and 5 represent the most active catalyst of the series, consuming all imine in less than 100 min with only 0.1 mol% of catalyst loading.

4. Conclusions

In agreement with recent publications on related processes,18,30 the role of half-sandwich iridacycles as LA catalysts in the activation of hydrosilane in the hydrosilylation of imines was clearly established. This activation procedure was shown to be operative not only in the imine hydrosilylation but also in the related hydrolysis of hydrosilanes and hydrosilylation of amines (which were revealed as competing reactions under catalytic conditions). It was found that iridacycle-catalyzed hydrosilylation of enolizable imines operates through a reaction mechanism more intricate than anticipated. N-(1-Phenylethylidene)aniline (a) transformed into the expected N-silylated amine (d), not following the “direct” Piers mechanism, but through a sequential process, via amine (b) and N-silylated enamine (c) formed in equimolecular amounts as reaction intermediates. This cascade mechanism, described by Oestreich for electron-deficient boranes,41 has never been observed before for organometallic-based catalysts. A detailed experimental study of the mechanism revealed that it consists of an intricate net of equilibria, which makes the product outcome dependent on the reagents, H2O and catalyst relative concentrations, and reaction time.

All the catalysts assayed followed this sequential mechanism, catalysts 3 and 5 being the most active of the series.

The sequential mechanism was operative for all the monohydrosilanes assayed, but when polyhydrosilanes were used as hydrosilylating reagents completely different reaction profiles were observed, demonstrating product inhibition after an initial fast, direct hydrosilylation reaction.

Additionally, it was evidenced that the protodesilylation workup procedure, commonly used to determine the reaction progress, not only rendered unreliable results but also hampered the observation of the actual reaction intermediates. In situ NMR spectroscopy, permitting a direct observation of the most stable reaction intermediates, reactants and products, should be considered as the standard procedure to monitor this reaction.

We consider that these findings constitute a solid pillar for further catalyst improvement. They should necessarily be taken into account for the development of asymmetric versions of the reaction using either chiral organometallic catalysts or chiral hydrosilylating agents.

Abbreviations

LALewis acid
LBLewis base
NaBArF24Sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate
TMSTetramethylsilane
rdsRate-determining step
KIEKinetic isotope effect

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Funding from Ikerbasque, the Spanish MINECO/FEDER (CTQ2015-65268-C2-1-P) and the UPV-EHU (GIU16/25) is acknowledged. The UPV-EHU SGIker is acknowledged for technical assistance in the NMR and HR-MS analyses.

References

  1. M. C. Lipke, A. L. Liberman-Martin and T. D. Tilley, Angew. Chem., Int. Ed., 2017, 56, 2260–2294 CrossRef PubMed.
  2. T. Robert and M. Oestreich, Angew. Chem., Int. Ed., 2013, 52, 5216–5218 CrossRef PubMed.
  3. M. Iglesias, F. J. Fernández-Alvarez and L. A. Oro, ChemCatChem, 2014, 6, 2486–2489 CrossRef.
  4. V. K. Dioumaev and R. M. Bullock, Nature, 2003, 424, 530–532 CrossRef PubMed.
  5. F. J. Fernández-Alvarez, A. M. Aitani and L. A. Oro, Catal. Sci. Technol., 2014, 4, 611–624 RSC.
  6. A. J. Chalk and J. F. Harrod, J. Am. Chem. Soc., 1965, 87, 16–21 CrossRef.
  7. C. L. Randolph and M. S. Wrighton, J. Am. Chem. Soc., 1986, 108, 3366–3374 CrossRef.
  8. S. B. Duckett and R. N. Perutz, Organometallics, 1992, 11, 90–98 CrossRef.
  9. I. Ojima, M. Nihonyanagi, T. Kogure, M. Kumagal, S. Horiuchi and K. Nakatsugawa, J. Organomet. Chem., 1975, 94, 449–461 CrossRef.
  10. I. Ojima, T. Kogure, M. Kumagai, S. Horiuchi and T. Sato, J. Organomet. Chem., 1976, 122, 83–97 CrossRef.
  11. D. J. Parks and W. E. Piers, J. Am. Chem. Soc., 1996, 118, 9440–9441 CrossRef.
  12. M. Pérez, Z.-W. Qu, C. B. Caputo, V. Podgorny, L. J. Hounjet, A. Hansen, R. Dobrovetsky, S. Grimme and D. W. Stephan, Chem. – Eur. J., 2015, 2015, 6491–6500 CrossRef PubMed.
  13. M. Mewald and M. Oestreich, Chem. – Eur. J., 2012, 18, 14079–14084 CrossRef PubMed.
  14. A. Y. Houghton, J. Hurmalainen, A. Mansikkamäki, W. E. Piers and H. M. Tuononen, Nat. Chem., 2014, 6, 983–988 CrossRef PubMed.
  15. J. M. Blackwell, E. R. Sonmor, T. Scoccitti and W. E. Piers, Org. Lett., 2000, 2, 3921–3923 CrossRef PubMed.
  16. D. W. Stephan, Org. Biomol. Chem., 2008, 6, 1535–1539 RSC.
  17. D. W. Stephan and G. Erker, Angew. Chem., Int. Ed., 2010, 49, 46–76 CrossRef PubMed.
  18. M. Hamdaoui, C. Desrousseaux, H. Habbita and J.-P. Djukic, Organometallics, 2017, 36, 4864–4882 CrossRef.
  19. Y. Corre, X. Trivelli, F. Capet, J.-P. Djukic, F. Agbossou-Niedercorn and C. Michon, ChemCatChem, 2017, 9, 2009–2017 CrossRef.
  20. Y. Corre, V. Rysak, F. Capet, J.-P. Djukic, F. Agbossou-Niedercorn and C. Michon, Chem. – Eur. J., 2016, 22, 14036–14041 CrossRef PubMed.
  21. J. Fuchs, H. F. T. Klare and M. Oestreich, ACS Catal., 2017, 7, 8338–8342 CrossRef.
  22. D. V. Gutsulyak, A. van der Est and G. I. Nikonov, Angew. Chem., Int. Ed., 2011, 50, 1384–1387 CrossRef PubMed.
  23. T. T. Metsänen, P. Hrobárik, H. F. T. Klare, M. Kaupp and M. Oestreich, J. Am. Chem. Soc., 2014, 136, 6912–6915 CrossRef PubMed.
  24. S. Park and M. Brookhart, Organometallics, 2010, 29, 6057–6064 CrossRef PubMed.
  25. D. V. Gutsulyak, S. F. Vyboishchikov and G. I. Nikonov, J. Am. Chem. Soc., 2010, 132, 5950–5951 CrossRef PubMed.
  26. M. Iglesias, P. J. Sanz Miguel, V. Polo, F. J. Fernández-Alvarez, J. J. Pérez-Torrente and L. A. Oro, Chem. – Eur. J., 2013, 17559–17566 CrossRef PubMed.
  27. X.-L. Luo and R. H. Crabtree, J. Am. Chem. Soc., 1989, 111, 2527–2535 CrossRef.
  28. K. Garcés, F. J. Fernández-Alvarez, V. Polo, R. Lalrempuia, J. J. Pérez-Torrente and L. A. Oro, ChemCatChem, 2014, 6, 1691–1697 CrossRef.
  29. D. Ventura-Espinosa, S. Sabater, A. Carretero-Cerdan, M. Baya and J. A. Mata, ACS Catal., 2018, 8, 2558–2566 CrossRef.
  30. M. Hamdaoui, M. Ney, V. Sarda, L. Karmazin, C. Bailly, N. Sieffert, S. Dohm, A. Hansen, S. Grimme and J.-P. Djukic, Organometallics, 2016, 35, 2207–2223 CrossRef.
  31. M. Aliaga-Lavrijsen, M. Iglesias, A. Cebollada, K. Garcés, N. García, P. J. Sanz Miguel, F. J. Fernández-Alvarez, J. J. Pérez-Torrente and L. A. Oro, Organometallics, 2015, 34, 2378–2385 CrossRef.
  32. S. Park, D. Bézier and M. Brookhart, J. Am. Chem. Soc., 2012, 134, 11404–11407 CrossRef PubMed.
  33. F. J. Fernández-Alvarez, J. J. Pérez-Torrente and L. A. Oro, Angew. Chem., Int. Ed., 2012, 51, 12824–12827 CrossRef PubMed.
  34. J. Yang, P. S. White and M. Brookhart, J. Am. Chem. Soc., 2008, 130, 17509–17518 CrossRef PubMed.
  35. M. P. McLaughlin, L. L. Adduci, J. J. Becker and M. R. Gagné, J. Am. Chem. Soc., 2013, 135, 1225–1227 CrossRef PubMed.
  36. J. Yang and M. Brookhart, J. Am. Chem. Soc., 2007, 129, 12656–12657 CrossRef PubMed.
  37. S. Park and M. Brookhart, J. Am. Chem. Soc., 2012, 134, 640–653 CrossRef PubMed.
  38. J. Yang, P. S. White, C. K. Schauer and M. Brookhart, Angew. Chem., Int. Ed., 2008, 47, 4141–4143 CrossRef PubMed.
  39. W. Iali, F. La Paglia, F. Le Goff, D. Sredojevic, M. Pfeffer and J.-P. Djukic, Chem. Commun., 2012, 48, 10310–10312 RSC.
  40. Y. Corre, W. Iali, M. Hamdaoui, X. Trivelli, J.-P. Djukic, F. Agbossou-Niedercorn and C. Michon, Catal. Sci. Technol., 2015, 5, 1452–1458 RSC.
  41. J. Hermeke, M. Mewald and M. Oestreich, J. Am. Chem. Soc., 2013, 135, 17537–17546 CrossRef PubMed.
  42. T. Robert and M. Oestreich, Angew. Chem., Int. Ed., 2015, 52, 5216–5218 CrossRef PubMed.
  43. Y. Ohki, Y. Takikawa, H. Sadohara, C. Kesenheimer, B. Engendahl, E. Kapatina and K. Tatsumi, Chem. – Asian J., 2008, 3, 1625–1635 CrossRef PubMed.
  44. J. Hermeke, H. F. T. Klare and M. Oestreich, Chem. – Eur. J., 2014, 20, 9250–9254 CrossRef PubMed.
  45. K.-I. Fujita, T. Yoshida, Y. Imori and R. Yamaguchi, Org. Lett., 2011, 13, 2278–2281 CrossRef PubMed.
  46. R. Kawahara, K. Fujita and R. Yamaguchi, Angew. Chem., Int. Ed., 2012, 51, 12790–12794 CrossRef PubMed.
  47. W.-H. Wang, J. T. Muckerman, E. Fujita and Y. Himeda, New J. Chem., 2013, 37, 1860–1866 RSC.
  48. C. D. F. Königs, H. F. T. Klare, Y. Ohki, K. Tatsumi and M. Oestreich, Org. Lett., 2012, 14, 2842–2845 CrossRef PubMed.
  49. C. Scheeren, F. Maasarani, A. Hijazi, J.-P. Djukic and M. Pfeffer, Organometallics, 2007, 26, 3336–3345 CrossRef.
  50. L. Li, W. W. Brennessel and W. D. Jones, J. Am. Chem. Soc., 2008, 130, 12414–12419 CrossRef PubMed.
  51. Y.-K. Sau, X.-Y. Yi, K.-W. Chan, C.-S. Lai, I. D. Williams and W.-H. Leung, J. Organomet. Chem., 2010, 695, 1399–1404 CrossRef.
  52. P. Renzi, J. Hioe and R. M. Gschwind, J. Am. Chem. Soc., 2017, 139, 6752–6760 CrossRef PubMed.
  53. J. S. M. Samec and J.-E. Bäckvall, Chem. – Eur. J., 2002, 8, 2955–2961 CrossRef.
  54. M. Brookhart, B. Grant and A. F. Volpe Jr, Organometallics, 1992, 11, 3920–3922 CrossRef.
  55. http://www.manonthemoontech.com/ .
  56. Z.-F. Xu, R. Affleck, P. Wangikar, V. Suzawa, J. S. Dordick and D. S. Clark, Biotechnol. Bioeng., 1994, 43, 515–520 CrossRef PubMed.
  57. R. Affleck, Z.-F. Xu, V. Suzawa, K. Focht, D. S. Clark and J. S. Dordick, Proc. Natl. Acad. Sci. U. S. A., 1992, 89, 1100–1104 CrossRef.
  58. M. Aliaga-Lavrijsen, M. Iglesias, A. Cebollada, K. Garcés, N. García, P. J. Sanz Miguel, F. J. Fernández-Alvarez, J. J. Pérez-Torrente and L. A. Oro, Organometallics, 2014, 34, 2378–2385 CrossRef.
  59. G. Z. Zheng and T. H. Chan, Organometallics, 1995, 14, 70–79 CrossRef.
  60. P. Gigler, B. Bechlars, W. A. Herrmann and F. E. Kühn, J. Am. Chem. Soc., 2010, 133, 1589–1596 CrossRef PubMed.
  61. N. Schneider, M. Finger, C. Haferkemper, S. Bellemin-Laponnaz, P. Hofmann and L. H. Gade, Angew. Chem., Int. Ed., 2009, 48, 1609–1613 CrossRef PubMed.
  62. S. Wübbolt and M. Oestreich, Synlett, 2017, 28, 2411–2414 CrossRef.
  63. S. Bähr and M. Oestreich, Organometallics, 2017, 36, 935–943 CrossRef.
  64. S. Wübbolt, M. S. Maji, E. Irran and M. Oestreich, Chem. – Eur. J., 2017, 23, 6213–6219 CrossRef PubMed.
  65. C. Lorentz and U. Schubert, Chem. Ber., 1995, 128, 1267–1269 CrossRef.

Footnotes

Electronic supplementary information (ESI) available: Synthetic procedures, reaction profiles and representative in situ NMR spectra during catalysis, X-ray crystallographic data. CCDC 1849635. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8cy01236a
The same reactivity was described for carbonyl substrates48 and recently extended to the reduction of nitriles and pyridines by catalytic hydrosilylation,62 and to the enantioselective hydrosilylation of imines and ketones using chiral versions of the catalyst.63,64
§ The same rate enhancement upon exposure to air was already noticed for the related hydrosilane alcoholysis process using a copper(I) catalyst.65
Previous reports on this reaction describe strictly inert conditions as necessary to avoid catalyst decomposition.40

This journal is © The Royal Society of Chemistry 2018