Immobilization of Pyrene‐Adorned N‐Heterocyclic Carbene Complexes of Rhodium(I) on Reduced Graphene Oxide and Study of their Catalytic Activity

Two pyrene‐tagged N‐heterocyclic carbene (NHC) complexes of rhodium(I) were obtained and characterized. The two complexes were supported onto reduced graphene oxide (rGO), generating two new materials in which the molecular complexes are immobilized by π–π stacking interactions onto the surface of the solid. The catalytic activity of both complexes and solid hybrid materials were studied in the 1,4‐addition of phenylboronic acid to cyclohex‐2‐one, and in the hydrosilylation of terminal alkynes. The studies showed that for both reactions, the dimetallic complex displayed better catalytic performances than the monometallic one. This accounted for both the reactions performed in homogeneous conditions and for the reactions performed with the solid. In the case of the addition of phenylboronic acid to cyclohexanone, the solid containing the dimetallic catalyst could be effectively recycled up to five times, with negligible loss of activity, whereas the monometallic catalyst rapidly became inactive. In the hydrosilylation of terminal alkynes, the selectivity towards the β‐(Z)vinylsilane was improved if the immobilized dimetallic catalyst was used, although the catalyst started to lose activity after the second run.


Introduction
Although homogeneousc atalysts usually show the advantages of high activity and high selectivity,t heir wide application in the industrial large-scale production of materials remainsl imited owing to thed ifficulties encountered when trying to separate the reaction products from the catalyst and from the reaction solvent. [1] As trategy to combinet he besta ttributes from homogeneous (high activity and selectivity) and heterogeneous (catalyst recovery and recyclability) catalysts is to use supported metal-based catalysts.T he most widely used strategy for catalysti mmobilization is the formation of ac ovalent bond between the solid and one of the ligandso ft he homogeneous catalyst. This requires that the catalysth as an additional functionalization,b ut this often raises the catalyst preparation cost and may modify its catalytic properties. Non-covalent methods for catalyst immobilization have gained interest because they offer am uch simplera pproach to catalyst immobilization, [1e] althought hey may suffer from the disadvantage that the link between the catalyst and the solid is sometimes not too strong. Carbon surfaces are interesting materials for catalyst immobilization, because they offer excellent thermal and mechanical stability,h igh surfacea rea, and ar ich surface chemistry. [2] Owing to the inherentc apability of pyrene to afford p-stacking interactions with graphitic surfaces, [3] pyrenetagged metal complexesh ave been efficiently supported onto graphitized solids, [4] and some have also been used in catalysis, proving interesting recyclabilityp roperties. [5] N-heterocyclic carbenes (NHCs) have been found to be excellent ligandsf or catalyst immobilization. [6] We recently contributed to this field by preparing as eries of pyrene-tagged NHC-based catalysts, which were supported onto reduced grapheneo xide (rGO),a nd afforded excellent recyclability in reactions such as the hydrodefluorination of aromatic fluorocarbenes, [7] hydrogenation of alkenes and alcohol oxidation, [8] and in the b-alkylation of secondary alcohols with primarya lcohols. [9] In most of thesep rocesses, we found that the presence of the pyrene tags bound to the NHC ligands allowed the effectiven on-covalenti mmobilization of the homogeneous catalyst onto the graphene derivative surface,but also induced significant modificationso ft he properties of the homogeneous catalyst, which we related to p-p stacking interactions established between the aromatic substrates and the pyrene functionalities. [9,10] Based on these precedents, we now prepared two new pyrene-tagged rhodium NHC complexes, which we used as catalysts in the addition of arylboronic acids to cyclohexanone, and in the hydrosilylation of terminal alkynes. These two catalysts were immobilized onto reducedg raphene oxide (rGO),a nd the activities and recyclability properties of the resultingh eterogenized catalysts were studied for the same two processes.
Twop yrene-tagged N-heterocyclic carbene (NHC) complexes of rhodium(I) wereo btained and characterized. Thet wo complexes were supported onto reduced graphene oxide (rGO), generating two new materials in which the molecular complexes are immobilized by p-p stacking interactions onto the surfaceo ft he solid. The catalytic activity of both complexes and solid hybrid materials were studied in the 1,4-addition of phenylboronic acid to cyclohex-2-one, and in the hydrosilylation of terminal alkynes. The studies showed that for both reactions,t he dimetallic complex displayed better catalytic performances than the monometallic one. This accounted for both the reactions performed in homogeneous conditions and for the reactions performed with the solid. In the case of the addition of phenylboronic acid to cyclohexanone,t he solid containing the dimetallic catalystc ould be effectively recycled up to five times, with negligible loss of activity,w hereas the monometallic catalyst rapidlyb ecame inactive. In the hydrosilylation of terminal alkynes, the selectivity towards the b-(Z)vinylsilane was improved if the immobilized dimetallic catalyst was used, although the catalyst startedt ol ose activity after the second run.

Results and Discussion
The imidazolium and bisazolium salts A and B were prepared accordingt ot he literature procedures. [11] The coordination to rhodiumw as performed accordingt ot he methodd epicted in Scheme 1. The reaction of the imidazolium or bisazolium salts A or B with [RhCl(COD)] 2 (COD = cyclooctadiene)i namixture of THF/DMF at 80 8Ci nt he presence of K 2 CO 3 gave complexes 1 and 2.K Br was added to the reaction mixture to facilitate the formation of the bromide containingc omplexesa nd avoid the presence of mixtures of halides in the reactionp roducts. The resulting products, 1 and 2,w ere obtained in moderate to high yields (56-65 %) after workup.
Both complexes were characterized by means of NMR spectroscopyand mass spectrometry,and gave satisfactory elemental analysis. 13 CNMR spectroscopy of complexes 1 and 2 shows the distinctive doubletso wing to the metalated carbene carbons at 183.6 ( 1 J Rh-C = 50 Hz) and 183.4 ( 1 J Rh-C = 50 Hz) ppm. Both 1 Ha nd 13 CNMR spectra of complex 2 are in agreement with its twofold symmetry.T he ESI-mass spectrometry of the complexes shows representative peaks at m/z values of 583.2 (assigned to [1ÀBr] + )a nd complexes at 1169.2 (assigned to [2ÀBr] + ).
Complexes 1 and 2 were grafted ontor educed graphene oxide (rGO) by mixing complexes 1 or 2 and rGO in dichloromethanei na nu ltrasound bath for 20 min. Then, the suspension wass tirred for 12 h( Scheme2). The first visual evidence of the anchoringo ft he catalysts onto the solid is the disappearance of the yellow color of the solution. The resulting black solids were filtered and washed with methylene chloride. The 1 HNMR spectra of the filtrate confirmed the absence of signals resulting from 1 or 2 in the solution, thus constituting evidencef or the effective immobilization of the molecular complexes on the solid. The exact rhodiumc ontent in solids rGO-1 and rGO-2 was determined by digestion of the samples in hot HCl/HNO 3 followed by inductively coupledp lasma mass spectrometry (ICP-MS) analysis,a nd accounted for 0.9 and 1.11wt% of rhodium in rGO-1 and rGO-2,r espectively.T he elemental mapping by energy-dispersive X-ray spectroscopic analysis(EDS) performed by means of high-resolution transmission electron microscopy (HRTEM) of rGO-2 confirmed the homogeneous distribution of rhodiumi nt he hybrid material ( Figure 1s hows the HRTEM and EDS elemental mapping images obtained for rGO-2).
The catalytic activity of complexes 1 and 2 was tested in two reactions typically catalyzed by rhodium(I) complexes, namely the 1,4-addition of arylboronic acids to cyclohexen-2one, and in the hydrosilylation of terminal alkynes. The two ligands in 1 and 2 induce quasi-identical stereoelectronic properties, and therefore the comparison of the activities of these two complexes offer an excellento pportunity to compare the activity of the dimetallic complex with respect to its monometallic analog. The activity of the heterogenized materials rGO-1 and rGO-2 will also be compared with the one shown by their homogeneous counterparts.
The addition of arylboronica cids to a,b-unsaturated ketones [12] is ap rocess for which some Rh I -NHC complexes have afforded excellent activities and chemoselectivities. [12a, 13] As a model reaction, we studied the addition of phenyl boronic acid to cyclohexen-2-one. The reactions were performed by using equimoleculara mountso fp henylboronic acid and cyclohexanonef or 6h in toluene at 100 8Ci nt he presence of KOH. The catalyst loading was 0.2 mol %f or 1 and 0.1 mol %f or 2, to studyt he reactions by using the same amount of rhodium loading. Under these reaction conditions, catalyst 1 yielded 85 %o ft he final product, whereas the dimetallic catalyst 2 afforded almost quantitative yields (98 %). This result contrasts with our previousfinding that atrimetallic NHC-based rhodium catalysta fforded quasi-identical activities as its relatedm onometallica nalog for this same catalytic reaction. [13d] In this case, we believe that the presence of the pyrene functionalities in the rhodium catalysts may have some influencei nt he activity of the catalysts owing to possible p-stacking interactions with the aromatic substrates, andt his may induce substantial differences with the activities shownb yr elatedr hodium catalysts not containing pyrene tags. [9] Next, we studied the activitieso f the hybrid materials rGO-1 and rGO-2 in the same reaction.
Each individual run was performedb yu sing 0.2 mol %b ased on the rhodium loading (the amounto fs olid was calculated based on the amount of rhodium determined by ICP-MS). After completion of each run (6 h), the solidw as filtered, washedw ith CH 2 Cl 2 ,and reused in the following run.
As can be observed from the resultss hown in Figure 2, both catalysts rGO-1 and rGO-2 showeds imilarc atalytic activities as their homogeneousc ounterparts in the first cycle. Catalyst rGO-1 showed gradually diminishing activity during five cycles (from 90 %t o7 3% yield), and then the activity abruptly ceased. The activity of rGO-2 was constantly higher than that shown by rGO-1 in all the recyclability experiments. During the first four cycles, rGO-2 produced quantitative yields of the final Scheme2.Immobilization of catalysts 1 and 2. Figure 1. STEM image (left) and EDS (HRTEM) elemental mapping imageo frGO-2 (right;green dots indicaterhodium atoms). ChemCatChem 2018ChemCatChem , 10,1874ChemCatChem -1881 www.chemcatchem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim product, and then the activity gradually decreased until complete deactivation of the catalyst after the ninth run. Considering the accumulated turnover numbers (TONs) of these recyclability experiments,t he results shown here constitute ac lear enhancement of the catalytic activity of the hybrid materials compared with the resultss hown in homogeneous conditions. In particular,c atalyst rGO-2 accumulated3 500 turnovers. The fact that the dimetallic complex with the two pyrene tags affords higher activity and better recyclability properties than the catalystw ith only one pyrene tag is in accord with previous studies, whichs uggestst hat leaching is significantly reduced if the metal complex is anchored to the surface of the solid by more than one pyrene tag. [4b, 9] The solids resulting from the recycling experimentsw ere analyzed by meanso fH RTEM spectroscopy and ICP-MS. The ICP-MS analysis of the solids revealed that the rhodium content for rGO-1 and rGO-2 was 0.73 %a nd 1.11 %, respectively,t herefore there was a3 0% loss of rhodium after five cycles in the case of rGO-1,w hereas the amount of metal was maintained over the nine cycles in the case of rGO-2.A sp reviously mentioned, this is in agreement with our previous findings regarding the more effective immobilization of catalysts containing two pyrene tags compared with those containing only one pyrene tag. [9] The HRTEM images (see the Supporting Information for details) did not show any traces of formation of metal nanoparticles, and the EDS analysisr evealed that the homogeneous distribution of rhodium along the solid surface wasm aintained. The morphology of graphene does not showsignificant changes compared with the originalo ne, apart from an increase in the number of wrinkles. Taking all these resultst ogether,w eb elieve that the loss of activity of the catalysts may be owing to partial loss of bulk hybrid catalystd uring the filtration process after each run, combinedw ith the partial poisoning of the surfaceo ft he catalyst by the addition of KOH for each of the cycles of the experiment.
Next, we studied the activity of the rhodiumc omplexes 1 and 2 and their hybrid heterogenized analogs, rGO-1 and rGO-2,i nt he hydrosilylation of terminal alkynes. This reaction constitutes ag ood challenge for study,b ecause the search for an effective catalystr equires that the catalystn ot only affords ah igh activity,b ut also high selectivity toward one of the three possible isomers that can be obtained. Severalr hodium complexes have provided good activities and selectivities in this reaction with important practical applications in synthetic organic chemistry. [14] We first studied the homogeneously catalyzed reactions between 1-hexynea nd phenylacetylene with HSiMe 2 Ph. The reactions were performed in CDCl 3 at room temperature by using catalystl oadings of 2mol %( based on rhodiumc oncentration). Under these reaction conditions, both catalysts showedg ood activity in the hydrosilylation of 1-hexyne, withaclear preference for the formation of the anti-Markovnikovanti-addition b-(Z)-vinylsilane product, which reaches 69 %y ield if the dimetallic catalyst 2 was used (Table 1, entry 2). This observation is very interesting, because the b-(Z)p roduct is the less frequent product in this reaction, and thus av ery elusivet arget. [14c,d, 15] Catalyst 2 is more active than catalyst 1 in the hydrosilylation of phenylacetylene with HSiMe 2 Ph, and also more selective in the production of the b-(Z)-vinylsilane (56 %y ield, entry 4). For this reaction, we observed that the conversion was slightly higher than the sum of the product yields if catalyst 2 was used. This is because, together with the formation of the three vinylsilanes, we observed the formationo fs tyrene, ac lear indication that the hydrosilylation of the terminal alkyne was accompanied by the dehydrogenative silylation, with styrene formed as ah ydrogen-trappinga gent. The product resulting from the dehydrogenative silylation (in our case, PhCCSi-Me 2 Ph) is easy to miss if the reaction is followed by 1 HNMR spectroscopy,a nd the formation of styrene is the clearest indication that this process is occurring. [14g, 16] Because we wanted to haveacompletep icture of the product formation along the courseo ft he reaction, we studied the time-dependent reactionp rofiles of the hydrosilylation reactions. As can be seen from the graphics shown in Figure 3, the ratio betweent he three isomersformed in the process is maintained all along the reaction course. This discards the option that any of the reaction products are formed through the metal-assisted isomerizationo fa ny of the other vinylsilanes. [14e, 17] Interestingly,t he reactionp rofiles arising from the   [b] 1 phenylacetylene1 70 .8 14 46 4 [b] 2 phenylacetylene5 62 .3 15 86 Reaction conditions:0 .077 mml of alkynea nd 0.085 mmol of HSiMe 2 Ph in 0.5 mL of CDCl 3 ,a tr oomt emperature for 6h,w ith 2mol %o fc atalyst loading, based on the rhodium content.
[a] Yields andc onversions determined by 1 HNMR spectroscopy,b yu sing anisole as internals tandard. [b] As mall amount of styrenew as observed togetherw ith all other reaction products. ChemCatChem 2018ChemCatChem , 10,1874ChemCatChem -1881 www.chemcatchem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim evolution of the reactions of 1-hexyne with HSiMe 2 Ph show the curved profiles typical for ap rocess that follows af irstorder reactionw ith respect to the concentration of the alkyne. Ad ifferent situation arises from the analysiso ft he plots referred to the evolution of the reactiono fp henylacetylene with the silane. For these processes, the reactionc ourse is represented by as traight line, indicating that the reaction is pseudo-zeroth order with respectt ot he concentration of the substrate, and therefore the reactionr ate does not dependo n the concentrationo fp henylacetylene. This type of situation normally arises if only as mall portion of molecules of the substrate are in al ocation where they are able to react, and this fraction is continually replenished from the larger pool. We previously found that this situation is produced if using pyrene-tagged catalysts and aromatic substrates. We attributed the effect to the non-covalent interaction between the aromatic substrate and the pyrene tag of the catalyst, whichs aturates the catalyst, and renders ar eactionr ate that is non-dependent on the concentration of the substrate. [9] We next studied the activity of the hybrid materials rGO-1 and rGO-2 in the hydrosilylationo f1 -hexynea nd phenylacetylene. The study of these reactions under heterogeneous conditions allows us to compare the activity and selectivity of the solid materials with the ones provided by the homogeneous catalysts, and to determine the recyclability properties of the heterogenized catalysts. The studies of the heterogenization and reutilization of alkyne hydrosilylation catalysts are very rare, althoughM esserlea nd co-workersr ecently described an interesting example in whichaRh I -NHC based complex was covalently immobilizedo nto graphene, and afforded excellent activities and high reusability in the hydrosilylation of diphenylacetylene. [18] We performed the reactions in CDCl 3 by using 2mol %c atalystl oading. Althoughw eo bserved that the reactions at room temperature did not show any significant product formation after 8h,t he reactions at 60 8Cp roduced conversions above 80 %a fter 6h.F igure 4s hows the results for the reactions performed between phenylacetylenea nd HSi-Me 2 Ph using catalysts rGO-1 and rGO-2.T he reaction performed in thep resence of rGO-1 afforded 85 %c onversion, with the b-(Z)-vinylsilane being the major product formed (61 %y ield). Both b-(E)a nd a isomersw ere obtainedw ith yields below 20 %. In terms of selectivity,this result is in perfect accordw ith the results provided by the homogeneous catalyst 2.A fter the first run finished, the solid was filtered, washed with CH 2 Cl 2 ,a nd reused forasecond cycle. In the second run, the activity of rGO-1 dropped down to only 17 %c onversion, with the same relative proportion of vinylsilanes formed. The activity of catalyst rGO-2 (87 %c onversion, 61 %y ield of b-(Z)) was very similart ot hat shown by rGO-1 during the first run. However,t his catalystc ould be reused as econd time without loss of activity,s howingt he same relative formation of isomers as in the first cycle. The activityo ft he catalyst was gradually reduced during the third (58 %c onversion)a nd fourth cycles (32 %c onversion), and was negligible in the fifth run.
Given the highera ctivity and reusability properties of catalyst rGO-2,w ed ecided to study the activity of this catalyst in Figure 3. Time-dependent reaction profiles of the reactionsof1 -hexyne and phenylacetylenewith HSiMe 2 Ph by using catalysts 1 and 2.Reactionconditions: 0.077 mmol of alkyne and 0.085mmol of HSiMe 2 Ph in 0.5 mL of CDCl 3 ,atroomtemperaturew ith 2mol %c atalystloading (based on rhodium content). a) 1-Hexyne + HSiMe 2 Ph by using catalyst 1,b )1-hexyne + HSiMe 2 Ph by using catalyst 2,c )phenylacetylene + HSiMe 2 Ph by using catalyst 1,a nd d) phenylacetylene + HSiMe 2 Ph by using catalyst 2. ChemCatChem 2018ChemCatChem , 10,1874ChemCatChem -1881 www.chemcatchem.org 2018 Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim the hydrosilylation of 1-hexyne. The results are shown in Figure 5. The analysis of the results allowed two interesting conclusions to be reached. First, the catalyst showedg ood activity during the first cycle (71 %c onversion), and then the activity was gradually reduced until the sixth run (6 %c onversion). Second, the selectivity of the catalyst toward the production of the b-(Z)-vinylsilane was clearlye nhanced compared with the homogeneous catalyst 2,a sc an be observed from the very small differences shown between the conversion valuesa nd the yields of production of this isomer. We did not detect the formation of any other vinylsilanes during any of the cycles performed in this series of experiments. Given that the b-(Z)-vinylsilane is av ery elusive and pursued target in the hydrosilylation of terminala lkynes, we believe that this result is of major importance.
The solids resulting from the recycling experiments were analyzed by means of HRTEM and ICP-MS. The ICP-MS analysis of the rhodium content in rGO-2 after the recycling experiments gave 1.08 and 1.10 % for the hydrosilylations of phenylacetylenea nd 1hexyne, respectively.T hese resultsi ndicate that the dimetallic catalysti se ffectively supported onto the solid, and that for rGO-2 leaching is negligible after all recycling experiments.O nt he contrary,f or the monometallic catalyst, with only one pyrene functionality,the final rhodiumcontent was 0.8 %, thus in-dicatinga15 %l oss of rhodiuma fter only two recycling experiments. This result is in agreement with the fact that the immobilizationo ft he monometallic catalysti sm uch less efficient than for the catalyst containing two pyrene functionalities.
The HRTEM images (in the Supporting Information) did not show traces of formation of nanoparticles. The EDS analysisr evealed that the homogeneous distributiono fr hodiuma longt he solid surface was maintained. The morphology of graphene does not show significant changes compared with the original one. Ta king all these resultst ogether,w eb elieve that the loss of activity of the catalysts may be owing the partial poisoning of the surface of the catalystb yt he reagents.

Conclusions
We prepared two rhodium(I) catalysts with N-heterocyclic carbene ligandsd ecorated with pyrene functionalities. The two catalysts were immobilizedo n the surface of reduced graphene oxide (rGO), and the catalytic activity of the homogeneous catalysts and the solids were compared for two organict ransformations: the 1,4-addition of phenylboronic acid to cyclohexan-4-one, and in the hydrosilylation of terminal alkynes. In both catalytic reactions, the dimetallic complex showed better catalytic performances than the monometallic one. For the reactions performed with the heterogenized catalysts, we observedt hat the dimetallic catalystw ith the two pyrene functionalities displayedh igher catalytic activity thant he material with the catalyst containing only one pyrene tag. Thes olid containing the dimetallic catalystc ould be effectively recycled for five times without measurable loss of activity in the addition of phenylboronic acid to cyclohexanone. In the case of the hydrosilylation of terminal alkynes, the solid containing the dimetallic catalyst afforded ac lear improvement of the selectivity of the process, with the reactionb eing stereoselective in the produc- . Reactionconditions: alkyne (0.077 mmol), HSiMe 2 Ph (0.085 mmol), and rGO-1 or rGO-2 (2 mol %b ased on rhodium)i nCDCl 3 (0.5 mL) at 60 8Cf or 8h.C onversionsand yieldsdeterminedb y 1 HNMR spectroscopy by usinga nisole as standard. Figure 5. Recycling experiments of the reaction of 1-hexyne with HSiM 2 Ph by using hybrid catalyst rGO-2.R eactionc onditions: alkyne( 0.077mmol), HSiMe 2 Ph (0.085 mmol), and rGO-2 (1 mol %) in CDCl 3 (0.5 mL) at 60 8Cf or 8h.C onversions and yieldsdetermined by 1 HNMR spectroscopy by using anisole as standard. ChemCatChem 2018ChemCatChem , 10,1874ChemCatChem -1881 www.chemcatchem.org tion of the b-(Z)-vinylsilane for the case of the hydrosilylation of 1-hexyne. This result is very interesting, because it indicates that the immobilization of the catalyst onto rGO not only can afford catalysts that can be effectively reused, but also that the selectivity may be improved.
The analysiso ft he solids after the recyclinge xperiments indicatedt hat leaching was negligible for the catalysts with two pyrene fragments, whereas there was as ignificant loss of rhodium content for the materialc ontaining the catalyst with only one pyrene tag. These results illustrate that the effectiveness of the immobilization is clearly dependento nt he number of anchoring pyrene sites on the catalyst.

Experimental Section
General comments All manipulations were performed under nitrogen by using standard Schlenk techniques and high vacuum. Anhydrous solvents were either distilled from appropriate drying agents (SPS) and degassed prior to use by purging with dry nitrogen and kept over molecular sieves. The azolium salts 2, 3,a nd 4 were obtained according to the procedures reported in the literature. [11] All other reagents were used as received from commercial suppliers. NMR spectra were recorded with Varian spectrometers. Electrospray mass spectra (ESI-MS) were recorded with MicromassQuatro LC instrument, and nitrogen was employed as the drying and nebulizing gas. Ag as chromatograph GC-2010 (Shimadzu) equipped with aF ID and Te chnokroma (TRB-5MS, 30 m 0.25 mm 0.25 mm) column and ag as chromatograph/mass spectrometer GCMS-QP2010 (Shimadzu) equipped with aT echnokroma (TRB-5MS, 30 m 0.25 mm 0.25 mm) column were used. Elemental analyses were performed with an EA1108 CHNS-O Carlo Erbaanalyzer.E lectrospray mass spectra (ESI-MS) were recorded with aM icromass-Quatro LC instrument, and nitrogen was employed as the drying and nebulizing gas. High-resolution transmission electron microscopy images (HRTEM) and high-angle annular dark-field (HAADF)-STEM images of the samples were obtained by using aJ em-2100 LaB6 (JEOL) transmission electron microscope coupled with an INCA Energy TEM 200 (Oxford) energy-dispersive X-ray spectrometer (EDX) operating at 200 kV.T he determination of the metal loading was done by ICP-MS Agilent 7500 CX. Synthesis of 2:Amixture of B (80.0 mg, 0.10 mmol), [RhCl(COD)] 2 (49.3 mg, 0.10 mmol), K 2 CO 3 (82.9 mg, 0.6 mmol), and KBr (70.0 mg) in THF/DMF (10:3 mL) was stirred at 75 8Cf or 8h.T he reaction was performed under N 2 .T he resulting suspension was cooled to room temperature and the solvent was removed under vacuum. The crude product was purified by column chromatography.T he pure compound 2 was eluted with dichloromethane/ethyl acetate (8:2) and precipitated from am ixture of dichloromethane/n-hexane to give ay ellow solid. Yield:6 5% (90. Preparation of rGO-1:r GO (150 mg) and CH 2 Cl 2 (10 mL) were introduced into ar ound-bottom flask. The suspension was sonicated for 30 min. Then, compound 1 (40 mg) was added. The suspension was stirred at room temperature for 12 hu ntil the solution become clear.T he black solid was filtrated and washed with CH 2 Cl 2 (2 25 mL), affording the resulting product as ab lack solid. The filtrates were combined and evaporated to dryness under reduced pressure. Unsupported compound 1 was analyzed by 1 HNMR spectroscopy by using anisole as internal standard. Integration of the characteristic signal of anisole (-OMe) versus (CH 2 -pyrene) reveals the amount complex that has been deposited on the rGO. The exact amount of complex supported was determined by ICP-MS analysis (0.9 %wtRh).

Synthesis and characterization of rhodium complexes
Preparation of rGO-2:r GO (150 mg) and CH 2 Cl 2 (10 mL) were introduced into ar ound-bottom flask. The suspension was sonicated for 30 min. Then, compound 2 (20 mg) was added. The suspension was stirred at room temperature for 12 hu ntil the solution become clear.T he black solid was filtered and washed with CH 2 Cl 2 (2 25 mL), affording the resulting product as ab lack solid. The filtrates were combined and evaporated to dryness under reduced pressure. Unsupported compound 2 was analyzed by 1 HNMR spectroscopy by using anisole as internal standard. Integration of the characteristic signal of anisole (-OMe) versus (CH 2 -pyrene) reveals the amount complex that has been deposited on the rGO. The exact amount of complex supported was determined by ICP-MS analysis (1.11% wt Rh).
Recycling experiments:I naround-bottom flask, amixture of 2-cyclohexen-1-one (0.5 mmol), arylboronic acid (0.6 mmol), KOH (0.09 mmol), and rGO-1 or rGO-2 (0.2 mol %b ased on the metal) was heated at reflux in toluene (1.5 mL) for 6h.T he monitoring of the reaction, yields, and conversions were determined by GC analyses by using anisole as an internal standard. After completion of each run (6 h), the reaction mixture was allowed to reach room temperature and was filtered. The remaining solid was washed thoroughly with CH 2 Cl 2 ,d ried under reduced pressure, and reused in the following run.
General procedure for the hydrosilylation of terminal alkynes: In aN MR tube, am ixture of the alkyne (0.077 mmol), HSiMe 2 Ph (0.085 mmol), and ac atalytic amount of 1 or 2 (2 mol %b ased on the metal) were dissolved in CDCl 3 (0.5 mL). The mixture was stirred at room temperature. The progress of the reaction was monitored by 1 HNMR spectroscopy.T he identity of the products formed was assessed from the literature.
Recycling experiments:I na2mL vial, am ixture of the alkyne (0.077 mmol), HSiMe 2 Ph (0.085 mmol), and the rGO-1 or rGO-2 (2 mol %b ased on the metal) were dissolved in CDCl 3 (0.5 mL). The mixture was stirred at 60 8Cf or the required time. The monitoring of the reaction, yields, and conversions were determined by GC analyses by using anisole as an internal standard. After the completion of each run (6 h), the reaction mixture was allowed to reach room temperature and was filtered. The remaining solid was washed thoroughly with CH 2 Cl 2 ,d ried under reduced pressure, and reused for the following run.