Heterogeneous catalysts based on supported Rh – NHC complexes : synthesis of high molecular weight poly ( silyl ether ) s by catalytic hydrosilylation † ‡

The new rhodium(I) complexes [Rh(Cl)(COD)(R-NHC-(CH2)3Si(O Pr3)3)] (R = 2,6-diisopropylphenyl (2a); n-butyl (2b)) have been synthesised and fully characterised. The study of their application as ketone hydrosilylation catalysts showed a clear N-substituent effect, 2a being the most active catalyst precursor. Complex 2a has been immobilised in the mesoporous materials MCM-41 and KIT-6. The new hybrid materials have been fully characterised and used as catalyst precursors for the preparation of poly(silyl ether)s by catalytic hydrosilylation. The heterogeneous catalytic systems based on the materials 2a–MCM-41 and 2a–KIT-6 afford polymers with high average molecular weight (Mw) Mw = 2.61 × 10 6 g mol−1 (2a–MCM-41) and Mw = 4.43 × 10 5 g mol−1 (2a–KIT-6).


Introduction
Transition-metal catalysed hydrosilylation has been demonstrated to be a convenient route for the synthesis of new functional materials. 1,2In this context, it is remarkable that the most efficient hydrosilylation catalysts are based on platinum or rhodium complexes. 1Such metals are particularly expensive and scarce.Consequently, there is an increasing interest in the development of heterogeneous and therefore easily recyclable hydrosilylation catalytic systems.][5][6][7][8][9] During the last decades N-heterocyclic carbenes (NHCs) have been widely used as ligands in homogeneous catalysis. 10,11his could be attributed to the interesting electronic properties of NHC ligands.Indeed, they are strong σ-donor ligands with a very poor π-acceptor character. 12[15][16][17] Here we report on the synthesis and characterisation of new Rh(I)-NHC complexes functionalized with a triisopropoxysilyl tail group for their immobilisation on silica-based inorganic supports.Fine tuning of the homogeneous catalyst by modification of the N-substituent on the NHC ligand has allowed for the synthesis of active ketone hydrosilylation catalysts.Moreover, the supported catalysts have found application in the synthesis of high molecular weight poly(silyl ether)s by rhodium catalysed hydrosilylation.

General information
All manipulations were performed with rigorous exclusion of air at an argon/vacuum manifold using standard Schlenktube techniques or in a dry-box (MB-UNILAB).Solvents were dried by the usual procedures and distilled under argon prior to use or taken under argon from a Solvent Purification System (SPS).The reagents 1,1,1,3,5,5,5-heptamethyltrisiloxane, 1,1,3,3,5,5-hexamethyltrisiloxane, terephthalaldehyde and the solids MCM-41 and KIT-6 were purchased from commercial sources and used without further purification.The rhodium(I) [Rh(μ-Cl)(COD)] 2 complex 18 and I(CH 2 ) 3 ( i PrO) 3 19 were prepared according to methods reported in the literature.NMR spectra were recorded on a Varian Gemini 2000, Bruker ARX 300, Bruker Avance 300 MHz or Bruker Avance 400 MHz instrument.Chemical shifts (expressed in parts per million) are referenced to residual solvent peaks ( 1 H, 13 C{ 1 H}).Coupling constants, J, are given in hertz.C, H, and N analyses were carried out using a Perkin-Elmer 2400 CHNS/O analyzer.Mass spectrometry was performed on an Esquire 3000+ with an ion trap detector interfaced to an Agilent 1100 series HPLC system.FT-IR spectra were recorded on a Nicolet Nexus 5700 FT spectrophotometer equipped with a Nicolet Smart Collector diffuse reflectance accessory.TEM microscopy images were collected with an INCA 200 X-Sight from Oxford Instruments with an energy resolution between 136 eV and 5.9 keV.Gas chromatography analyses were performed using an Agilent 6890N with a FID detector equipped with an HP-1 column from J&W Scientific (30 m, 0.25 mm i.To an acetonitrile (30 mL) solution of 1-(2,6-diisopropylphenyl)imidazole (465 mg, 2.00 mmol), I(CH 2 ) 3 Si( i PrO) 3 (1.12g, 3.00 mmol) was added.The reaction mixture was stirred at 90 °C for 24 h.The resulting solution was filtered through Celite and the solvent was removed in vacuo.The residue thus obtained was washed with hexane (2 × 20 mL) to give a brown solid.Yield 1.07 g (89%  29 Si NMR (64.52 MHz, CDCl 3 , 293 K): δ −49.9 (CH 2 Si).Mass spectrometry (ESI + ): m/z 581.7 (M + -Cl).

Crystal structure determination of compound 2a
Crystals of 2a suitable for X-ray analysis were obtained by the slow evaporation of a solution of the complex in hexane.X-ray diffraction data were collected at 100(2) K with graphite-monochromated MoKα radiation (λ = 0.71073 Å), using narrow ω rotation (0.3°) on a Bruker APEX DUO diffractometer.Intensities were integrated with the SAINT+ program and corrected for the absorption effect with SADABS, integrated in the APEX2 package.The structures were solved by direct methods with SHELXS-97. 20Refinement, by full-matrix least-squares on F 2 , was performed with SHELXL-97. 20Hydrogen atoms were included in calculated positions and defined with displacement and positional riding parameters.A typical disorder for the oxygen atoms of the alkoxy groups was observed and modelled with two oxygens having complementary occupancy factors (0.839 and 0.161( 6)).Half a molecule of hexane, highly disordered, was also observed in the asymmetric unit of the crystal; six residual peaks were introduced in the final refinement to take account of this electron density.

Recycling of the heterogeneous catalyst
The reaction mixture was centrifuged and the corresponding heterogeneous catalyst was separated from the product by decantation, washed with CH 2 Cl 2 (3× 3.0 mL), Et 2 O (3× 3.0 mL), dried in vacuo and reused without further purification.

Synthesis and characterisation of the catalyst precursors
[Rh(X)(COD)(NHC)] (X = Cl, Br) complexes have been revealed as efficient precatalysts in homogeneously catalysed hydrosilylation processes. 10][23] The preparation of Si(OR) 3 -functionalised rhodium-NHC complexes can be accomplished using imidazolium salts as NHC ligand precursors in order to obtain silver(I)-NHC compounds, which are widely used as reagents for transmetallation reactions. 24The imidazolium iodide salts used as ligand precursors in this work were synthesized in good yield by the reaction of the corresponding 1-substituted imidazole and I(CH 2 ) 3 Si( i PrO) 3 in acetonitrile at 90 °C (Scheme 1).The imidazolium salts 1a and 1b have been fully characterised by elemental analysis, mass spectrometry (ESI + ), 1 H, 13 C{ 1 H} and 29 Si NMR spectroscopy (see Experimental).
The rhodium complexes 2a and 2b were synthesised by Lin's method of transmetallation from the intermediate silver(I) complexes prepared in situ by reaction of the corresponding imidazolium salt with Ag 2 O (Scheme 2). 24Complexes 2a and 2b were isolated, after recrystallisation, as pure yellow crystalline solids in high yield and characterised by elemental analysis, mass spectrometry (ESI + ) and 1 H and 13 C{ 1 H} NMR spectroscopy.The 1 H NMR spectra of complexes 2a and 2b provided evidence for the absence of the NCHN resonance of the parent imidazolium salts and showed the characteristic set of resonances for both the 3-triisopropoxysilylpropyl and the 2,6-diisopropylphenyl (2a) and n-butyl (2b) substituents on the NHC ligands.The 13 C{ 1 H} NMR spectra exhibit a doublet resonance at δ 183.5 ppm ( 1 J Rh-C = 52 Hz) (2a) and 183.3 ppm ( 1 J Rh-C = 51 Hz) (2b), which confirms the coordination of the carbene ligand to the rhodium centre. 23,25,26he structure of complex 2a was determined by X-ray crystallography.A view of the molecular geometry of this species is shown in Fig. 1.The rhodium atom displays a slightly distorted square planar coordination.The COD ligand bonds to the metal through its two olefinic bonds in a chelate mode, showing a typical tub conformation; the carbon atom C(9) of the NHC ligand and the chlorido ligand complete the metal coordination environment.The NHC ligand contains two different N-substituents, namely, 2,6-diisopropylphenyl at N(1) and triisopropoxysilylpropyl at N(2).The geometry of the fragment [Rh(Cl)(COD)(NHC)] compares well with those reported for the complex [Rh(Cl)(COD)(2-methoxyethyl-NHC-(CH 2 ) 3 Si(O i Pr 3 ) 3 )] (2c) 23 and related compounds. 25,26The most striking feature of the structure concerns the high structural trans effect showed by the NHC ligand compared with that of the chloride, making the two Rh-olefin bond distances clearly different (Rh-M(1) 1.989(3) Å trans to Cl, and Rh-M(2) 2.084(3) Å trans to NHC ligand; see Fig. 1

Homogeneous catalytic hydrosilylation of acetophenone: influence of the N-substituent
The rhodium-catalyzed hydrosilylation of acetophenone with 1,1,1,3,5,5,5-heptamethyltrisiloxane (HepTMS) to produce PhMeCH-O-SiMe(OSiMe 3 ) 2 (Scheme 3) has been demonstrated to be an accurate activity test bench, whose results could be extrapolated to the synthetic methodology established for the preparation of poly(silyl ether)s by catalytic hydrosilylation. 23The new compounds 2a and 2b and the previously reported 2c, having a 2-methoxyethyl substituent of hemilabile character, were used as catalyst precursors for the catalytic hydrosilylation of acetophenone with HepTMS (Scheme 3).
The catalytic reactions of acetophenone with HeptMTS using the rhodium(I)-catalyst 2a, 2b or 2c were monitored by GC and 1 H NMR. As can be seen in Fig. 2, the N-substituent on the NHC ligand strongly influences the catalytic activity.Complex 2a, which contains the bulky 2,6-diisopropylphenyl substituent, exhibited the highest catalytic activity with conversion of 95% in less than 2 h and a TOF 1/2 of 73 h −1 .Complexes 2b and 2c, which contain a straight-chain flexible wingtip group of approximately the same length, were considerably less active, exhibiting TOF 1/2 values of 36 and 24 h −1 , respectively.Nevertheless, full acetophenone conversion was observed at longer reaction times with both catalyst precursors.
On the other hand, the inferior catalytic activity of 2c compared to 2b points to a negative influence of the potentially hemilabile 2-methoxyethyl substituent.This observation is in sharp contrast with the positive effect of this group on the hydrogen transfer catalytic activity of the related complex [Ir(Cl)(COD)(Me-NHC-(CH 2 ) 2 OCH 3 ], which has been substantiated by DFT calculations. 26The superior catalytic activity of 2a could be ascribed to the steric protection of the catalytic active species introduced by the bulky substituent on the NHC ligand.
1 H NMR studies of the reaction of acetophenone with HepMTS in C 6 D 6 , using 2a (10 mol%) as catalyst confirm that a pre-activation step is required as no reaction was observed below 60 °C.Initially, we observed the appearance of resonances corresponding to the hydrosilylation product, PhMeCH-O-SiMe(OSiMe 3 ) 2 , 23

Immobilisation of 2a on inorganic supports
The results described above clearly evidenced the superior catalytic performance of precursor 2a.These results prompted us to select complex 2a as precursor for the preparation of different heterogeneous catalysts.Thus, 2a-MCM-41 and 2a-KIT-6 were prepared by refluxing a mixture of 2a and the corresponding inorganic support in wet toluene for 24 h (Scheme 4).The new materials were isolated as off-white solids that were characterised by ICP-MS, FT-IR, 13 C and 29 Si CP-MAS NMR and TEM.We have determined a rhodium loading of 9.57 mg g −1 and 5.53 mg g −1 for 2a-MCM-41 and 2a-KIT-6, respectively.The FT-IR spectra of the new materials show the stretching vibration modes of the mesoporous framework (Si-O-Si) at around 1241 cm −1 , 1043 cm −1 and 800 cm −1 . 13he FT-IR spectra of the grafted materials exhibit an additional broad absorption corresponding to the ν(CN) and ν(CC) stretching modes of the ligand at around 1630 cm −1 . 21,22he resonances observed in the 13 C CP-MAS solid state NMR spectra of 2a-MCM-41 and 2a-KIT-6 compare well with those observed in the 13 31 The results of N 2 -adsorption/desorption studies of MCM-41, 2a-MCM-41, KIT-6 and 2a-KIT-6, including BET surface area, total pore volume, and Barret-Joyner-Halenda (BJH) pore size are shown in Table 1.These results showed that the surface area, pore volume and pore diameter of 2a-MCM-41 and 2a-KIT-6 decreased, which suggests the inclusion of the rhodium species 2a inside the channels of the mesoporous materials MCM-41 and KIT-6. Asa whole, the above data confirm that complex 2a has been effectively immobilised on MCM-41 and KIT-6 (Fig. 3).

Synthesis of high molecular weight poly(silyl ether)s
7][38] Weber's group reported in 1998 the first example of synthesis of poly(silyl ether)s by transition metal-catalysed hydrosilylation using [Ru(CO)H 2 (PPh 3 ) 3 ] as precatalyst. 39This method allowed the synthesis of a number of poly(silyl ether)s with average M w between 10 000 and 150 000 g mol −1 . 40,41Following our interest in the chemistry and catalytic applications of Rh(I)-NHC (NHC = N-heterocyclic carbene) species, 25,26,[42][43][44] we have recently found that both homogeneous and heterogeneous catalytic systems based on the species 2c, which contain a potentially hemilabile 2-methoxyethyl substituent, are also effective for the synthesis of poly(silyl ether)s. 23However, such catalytic systems afford polymers with a relatively low average M w up to 94 000 g mol −1 , and the recycled heterogeneous catalyst showed a drastic decrease in the average M w .Thus, the design and development of effective synthetic methodologies that permit the synthesis of high molecular weight poly(silyl ether)s and recycling of heterogeneous catalysts still remain a challenge.
In this work, we have studied the reaction of terephthalaldehyde with 1,1,3,3,5,5-hexamethyltrisiloxane (HexMTS) in 1,4-dioxane at 110 °C using the rhodium chlorido species 2a, or the materials 2a-MCM-41 and 2a-KIT-6 as catalyst precursors.These catalytic reactions proceed quantitatively, affording orange oils or amber gelatinous compounds (depending on the M w ) which have been characterised  Catalysis Science & Technology Paper as the poly(silyl ether) 3 by comparison of their 1 H, 13 C{ 1 H} and 29 Si{ 1 H} NMR spectra with reported data (Scheme 5). 23t is worth noting that GPC analyses of 3 showed UV absorption at 254 nm due to the aromatic rings and all along the polymer peaks, in agreement with the copolymeric nature of the samples.It is notable that glass transition temperature (T g ) and temperature of thermic destruction (T d ) values are related to the average M w (Table 2).The T g of these polymers are in the range of −86.3 °C and −89.4 °C.The polymers with high M w (Table 2, entry 2) have higher T g values.These T g values compare well with those reported for analogous poly(silyl ether)s 23 and are lower than those reported for polymers obtained with the ruthenium catalytic system. 39,40Thermogravimetric analysis (TGA) studies showed complete destruction of the polymers in the range of 456 °C and 510 °C with a weight loss of around 96%.In this case, the average M w also influences the temperature of thermic destruction which is higher for polymers with high average M w (Table 2).Thus, an increase in the average M w produces polymers with higher thermic resistance.
The homogenous catalyst precursor 2a gave a polymer with an average M w = 34 000 g mol −1 , much higher than that obtained with 2c as catalyst precursor (M w = 5200 g mol −1 ) under the same reaction conditions (2.0 mol% of catalyst and 110 °C in 1,4-dioxane). 23The solvent has an impact on the molecular weight of the polymer (Table 3).Interestingly, the copolymerization reaction also occurs under solvent-free conditions (Table 3, entry 3).However, polymers with higher average M w were always observed when 1,4-dioxane was used as solvent (Table 3, entry 2).
On the other hand, the heterogeneous catalyst 2a-KIT-6 produced polymers of lower average molecular weight exhibiting also a bimodal distribution (Table 4, entry 4).The low molecular weight polymers represent the most abundant fraction (≈90%) with average M w lower than 4000 g mol −1 and PDI values in the range of 1.3-1.4.However, the average M w of the high molecular weight fraction obtained using the heterogeneous catalyst 2a-KIT-6 (Table 4, entry 4), M w = 134 000 g mol −1 is considerably bigger.
Interestingly, upon reducing the catalyst loading and reaction time (0.2 mol% catalyst and 1 day of reaction at 110 °C) and using the heterogeneous catalysts 2a-MCM-41 and 2a-KIT-6, we observed an extraordinary enhancement of the average M w of the obtained polymers.This is remarkable in the case of the catalytic systems based on the material 2a-MCM-41 which allow for the preparation of polymers with an average M w = 2.61 × 10 6 g mol −1 (Table 4, entry 2).Additionally, we have proved that increasing the reaction time from 1 day to 2 days did not produce a noticeable change (Table 4, entries 3 and 6).This is an exciting result and represents the first synthetic route to poly(silyl ether)s with high average molecular weight.
The results described above provide evidence that the inorganic support strongly influences the average M w of the polymers.Thus, polymers with higher average M w were obtained using 2a-MCM-41 and 2a-KIT-6 as catalysts (Table 4).The best catalytic performance, in terms of average M w , was achieved with 2a-MCM-41.The polymers obtained using the hybrid heterogeneous catalysts have a bimodal   distribution This fact could be attributed to a confinement effect exerted by the support.GC monitoring of the copolymerisation reactions of terephthalaldehyde with 1,1,3,3,5,5-hexamethyltrisiloxane (HexMTS) revealed the total consumption of the co-monomers after 5 min, using both the homogeneous catalyst precursor 2a and the heterogeneous catalyst 2a-MCM-41.Additionally, studies on the variation of M w with reaction time provide evidence that in both homo-and heterogeneously catalysed processes, polymerisation proceeds in a stepwise manner with the M w of the polymer continuously increasing with time.These results are in agreement with step-growth polymerisation. 45ecycling studies using 2a-MCM-41 and 2a-KIT-6 as catalysts showed that in both cases and independent of the catalyst loading, the average molecular weight of the obtained polymers decreased after three uses, being dramatic in the case of 2a-MCM-41.We have studied the possibility of leaching of active species from the heterogeneous material; however, the solutions obtained after decantation of the heterogeneous catalysts showed no activity in the hydrosilylation of acetophenone.Furthermore, in order to discard the leaching of active species during reaction, we heated a suspension of 2a-MCM-41 in 1,4-dioxane for 24 hours at 110 °C.The 1,4-dioxane solution was filtered, and acetophenone and HepTMS were added.The mixture was heated for 16 hours and no trace of the hydrosilylation product was observed.An important detail observed is the slight enhancement of the weight of the recycled heterogeneous catalyst.Although other reasons cannot be excluded, this fact is compatible with the presence of polymer chains blocking the channels of the heterogeneous catalyst, which could be an explanation for diminishing M w after each cycle.
together with a doublet resonance at δ −20.62 ppm ( 1 J Rh-H = 40 Hz) due to a Rh-H intermediate species.After 30 minutes at 60 °C, the 1 H NMR spectra showed an increase of the resonances due to PhMeCH-O-SiMe(OSiMe 3 ) 2 and the formation of a mixture of unidentified rhodium hydride complexes.On the other hand, solutions of 2a in C 6 D 6 and wet C 6 D 6 (5% H 2 O) were stable for 24 h at 80 °C and no traces of any decomposition product nor i PrOH were observed.These observations suggest a hydride-mediated mechanism resulting from the silane activation, in full agreement with the established rhodiumcatalysed ketones hydrosilylation mechanisms.[27][28][29][30]
C{ 1 H} NMR spectra of the parent complex 2a.The most prominent resonances in the 13 C CP-MAS solid state NMR spectra of the solids are those due to the CH imd [δ 124.8−123.1 ppm], CH 2 N [δ 53.9−52.6 ppm] and CH 2 Si [δ 8.5−8.3 ppm] carbon atoms.The 29 Si CP-MAS solid state NMR spectra of these new materials exhibit resonances of great intensity corresponding to the silicon atoms of the silica (at around −92.0, −100 and −110 ppm) and less intense broad resonances in between δ −51.4 and −70.2 ppm (2a-MCM-41) and δ −52.6 and −66.3 ppm (2a-KIT-6) assigned to T 1 , T 2 and T 3 environments of the CH 2 Si silicon atoms.

Scheme 5 Table 2
Glass transition temperature (T g ) and temperature of thermic destruction (T d ) a Bimodal polymer.b (weight loss %).

Table 3
Solvent effect on the homogeneously catalysed synthesis of 3 a 4 days at 110 °C using 2.0 mol% of 2a as catalyst precursor, 100% of conversion. a