Graphene Oxide-supported Oxime Palladacycles as Efficient Catalysts for the Suzuki–miyaura Cross-coupling Reaction of Aryl Bromides at Room Temperature under Aqueous Conditions

Palladacycles are highly efficient precatalysts in cross-coupling reactions whose immobilization on carbonaceous materials has been hardly studied. Herein, we report a detailed study on the synthesis and characterization of new oxime palladacycle-graphene oxide non-covalent materials along with their catalytic activity in the Suzuki–Miyaura reaction. Catalyst 1-GO, which has been fully characterized by ICP, XPS, TGA, and UV-Vis analyses has been demonstrated to be an efficient catalyst for the Suzuki–Miyaura coupling between aryl bromides and arylboronic acids using very low catalyst loadings (0.002 mol % of Pd) at room temperature under aqueous conditions.

Very recently, we have shown that palladium nanoparticles (Pd NPs), supported on graphene and reduced graphene oxide, efficiently catalyze the Suzuki-Miyaura coupling between aryl bromides and potassium aryltrifluoroborates under aqueous and low loading conditions (0.1 mol % Pd) employing conventional or microwave heating [65].Based on the experience and knowledge of our research group with oxime palladacycles as highly efficient precatalysts in cross-coupling reactions [66][67][68], herein we report a detailed study on the synthesis and characterization of new oxime palladacycle-graphene oxide non-covalent materials along with their catalytic activity in the Suzuki-Miyaura reaction.

Results and Discussion
We initially chose GO as support for oxime palladacycles 1-3 (Figure 1) since this carbon material shows interesting hydrophilicity and stability properties in liquid phase reactions due to the abundant oxygen-containing functional groups on its surface.A non-covalent immobilization strategy, through hydrogen-bond and π-stacking interactions between the pre-formed palladacycles and the functionalized carbon support was selected to prepare the new catalytic materials.This approach should integrally preserve the palladacycle structure, avoiding an uncontrolled generation of other Pd(II) or Pd(0) species, as demonstrated by other authors when using a post-palladation strategy [17].Furthermore, we expect the carbonaceous support to act as a reservoir and stabilizing medium for the in situ generated Pd NPs, which could extend the catalyst lifetime.
Very recently, we have shown that palladium nanoparticles (Pd NPs), supported on graphene and reduced graphene oxide, efficiently catalyze the Suzuki-Miyaura coupling between aryl bromides and potassium aryltrifluoroborates under aqueous and low loading conditions (0.1 mol % Pd) employing conventional or microwave heating [64].Based on the experience and knowledge of our research group with oxime palladacycles as highly efficient precatalysts in cross-coupling reactions [65][66][67], herein we report a detailed study on the synthesis and characterization of new oxime palladacycle-graphene oxide non-covalent materials along with their catalytic activity in the Suzuki-Miyaura reaction.

Results and Discussion
We initially chose GO as support for oxime palladacycles 1-3 (Figure 1) since this carbon material shows interesting hydrophilicity and stability properties in liquid phase reactions due to the abundant oxygen-containing functional groups on its surface.A non-covalent immobilization strategy, through hydrogen-bond and π-stacking interactions between the pre-formed palladacycles and the functionalized carbon support was selected to prepare the new catalytic materials.This approach should integrally preserve the palladacycle structure, avoiding an uncontrolled generation of other Pd(II) or Pd(0) species, as demonstrated by other authors when using a post-palladation strategy [17].Furthermore, we expect the carbonaceous support to act as a reservoir and stabilizing medium for the in situ generated Pd NPs, which could extend the catalyst lifetime.Preliminary studies on the immobilization of the palladacycles over GO were carried out with complex 1.The preparation procedures tested to obtain the 1-GO catalyst are indicated in Scheme 1. Basically, they consisted of the synthesis of the corresponding oxime palladacycle 1 from 4,4′dichlorobenzophenone [68] and a subsequent immobilization of this complex over GO [69] in THF or water as solvents (Scheme 1).As depicted, sonication of the palladacycle in the presence of GO for 1 h (methods B and C) significantly improved the content of the immobilized complex on the carbon surface as determined by Inductively Coupled Plasma Optical Emission Spectrometry ICP-OES, being the highest (2 wt % Pd) when using THF as solvent (method B).Finally, following B conditions but starting from 4-hydroxyacetophenone and 1-[4-(2-hydroxyethoxy)phenyl]ethan-1-one, the supported catalysts 2-GO (2.72 wt % Pd by ICP-OES) and 3-GO (2.98 wt % Pd by ICP-OES) were also prepared (Scheme 1).Preliminary studies on the immobilization of the palladacycles over GO were carried out with complex 1.The preparation procedures tested to obtain the 1-GO catalyst are indicated in Scheme 1. Basically, they consisted of the synthesis of the corresponding oxime palladacycle 1 from 4,4 -dichlorobenzophenone [69] and a subsequent immobilization of this complex over GO [70] in THF or water as solvents (Scheme 1).As depicted, sonication of the palladacycle in the presence of GO for 1 h (methods B and C) significantly improved the content of the immobilized complex on the carbon surface as determined by Inductively Coupled Plasma Optical Emission Spectrometry ICP-OES, being the highest (2 wt % Pd) when using THF as solvent (method B).Finally, following B conditions but starting from 4-hydroxyacetophenone and 1-[4-(2-hydroxyethoxy)phenyl]ethan-1-one, the supported catalysts 2-GO (2.72 wt % Pd by ICP-OES) and 3-GO (2.98 wt % Pd by ICP-OES) were also prepared (Scheme 1).
X-ray photoelectron spectroscopy (XPS) was used to analyze the elemental composition of the surface of the different catalysts.Interestingly, XPS analysis of the palladacycles 1-3 and all the supported catalyst batches showed all the binding energies associated with the oxime palladacycle structure (see SI).This point confirmed the stability of the cyclopalladated complexes during the immobilization process to GO.In fact, in the case of the preparation of 1-GO, no structural changes were observed by 1 H-NMR analysis in the eluted palladacycle after the immobilization process (see SI). Generally, none of the prepared catalysts contained palladium oxides or palladium(0) species on their surface according to the XPS analyses.Indeed, in isolated cases we could identify small amounts of the precursor Li 2 PdCl 4 by XPS analysis, easily removed by simple washings with water.On the other hand, in the case of using water as solvent (method C) for the preparation of 1-GO, we also observed small amounts (~8%) of a non-identified oxidized palladium specie (336.0 and 341.1 eV by XPS, see SI).Scheme 1. Synthesis of supported oxime palladacycles.
X-ray photoelectron spectroscopy (XPS) was used to analyze the elemental composition of the surface of the different catalysts.Interestingly, XPS analysis of the palladacycles 1-3 and all the supported catalyst batches showed all the binding energies associated with the oxime palladacycle structure (see SI).This point confirmed the stability of the cyclopalladated complexes during the immobilization process to GO.In fact, in the case of the preparation of 1-GO, no structural changes were observed by 1 HNMR analysis in the eluted palladacycle after the immobilization process (see SI). Generally, none of the prepared catalysts contained palladium oxides or palladium(0) species on their surface according to the XPS analyses.Indeed, in isolated cases we could identify small amounts of the precursor Li2PdCl4 by XPS analysis, easily removed by simple washings with water.On the other hand, in the case of using water as solvent (method C) for the preparation of 1-GO, we also observed small amounts (~8%) of a non-identified oxidized palladium specie (336.0 and 341.1 eV by XPS, see SI).
Figures 2 and 3 show the XPS spectra of 1 and 1-GO where the spectroscopic signature of palladium, carbon, oxygen, and nitrogen atoms has been determined [70][71][72] Regarding palladium (Figure 3a), the Pd3d XPS spectrum for 1 and 1-GO showed the corresponding binding energies at 337.70 (3d5/2) and 343.02 (3d3/2) eV and 337.62 (3d5/2) and 342.83 (3d3/2) eV, respectively.These energies were assigned to the Pd(II) from the oxime palladacycle.On the other hand, O1s XPS analysis for catalyst 1 showed three different components related to the C=N-OH bond at 531.76 eV, the O-H bond at 533.5 eV and the C-O bond at 532.20 eV (Figure 3b).The higher content of the O-H and C-O groups in 1-GO (532.20 and 533.32 eV, respectively) can be associated with the corresponding groups on the graphene oxide surface.The carbonaceous material was also responsible for the higher content of the C=C/C-H aromatic ring binding energies (284.60 and 286.71 eV) in the C1s XPS spectrum of 1-GO when compared with 1 (Figure 3c).The N1s XPS spectra of 1 and 1-GO showed the N=C bonding energies at 400.36 and 399.93 eV, respectively.Furthermore, the peaks corresponding to the N-Pd moiety appeared at 402.86 eV for 1 and 401.93 eV for 1-GO (Figure 3d).Finally, the Cl2p XPS spectra of 1 was analyzed showing the corresponding binding energies at 198.6 eV (2p), 199.9 (2p3/2) and 200.3 eV (2p1/2).These energies were assigned to the Cl-Pd bonding energy from the oxime palladacycle 1.The band at 200.1 8 eV corresponded to the 2p3/2 aromatic-Cl bonding energy (Figure 3e) [62].Thus, performing the palladation reaction using THF as solvent before immobilization onto GO led to the integral cyclopalladated species being supported.In fact, Pd(0) has been previously detected when anchoring oxime palladacycles to different supports such as, 3-hydroxypropyltriethoxysilylfunctionalized MCM-41 [17] and 3-(aminomethyl)pyridine-functionalized graphene oxide [63].
Figures 2 and 3 show the XPS spectra of 1 and 1-GO where the spectroscopic signature of palladium, carbon, oxygen, and nitrogen atoms has been determined [71][72][73] Regarding palladium (Figure 3a), the Pd3d XPS spectrum for 1 and 1-GO showed the corresponding binding energies at 337.70 (3d5/2) and 343.02 (3d3/2) eV and 337.62 (3d5/2) and 342.83 (3d3/2) eV, respectively.These energies were assigned to the Pd(II) from the oxime palladacycle.On the other hand, O1s XPS analysis for catalyst 1 showed three different components related to the C=N-OH bond at 531.76 eV, the O-H bond at 533.5 eV and the C-O bond at 532.20 eV (Figure 3b).The higher content of the O-H and C-O groups in 1-GO (532.20 and 533.32 eV, respectively) can be associated with the corresponding groups on the graphene oxide surface.The carbonaceous material was also responsible for the higher content of the C=C/C-H aromatic ring binding energies (284.60 and 286.71 eV) in the C1s XPS spectrum of 1-GO when compared with 1 (Figure 3c).The N1s XPS spectra of 1 and 1-GO showed the N=C bonding energies at 400.36 and 399.93 eV, respectively.Furthermore, the peaks corresponding to the N-Pd moiety appeared at 402.86 eV for 1 and 401.93 eV for 1-GO (Figure 3d).Finally, the Cl2p XPS spectra of 1 was analyzed showing the corresponding binding energies at 198.6 eV (2p), 199.9 (2p3/2) and 200.3 eV (2p1/2).These energies were assigned to the Cl-Pd bonding energy from the oxime palladacycle 1.The band at 200.1 8 eV corresponded to the 2p3/2 aromatic-Cl bonding energy (Figure 3e) [63].Thermogravimetric analysis (TGA) of 1 (line blue), GO (black), and 1-GO (line red) was next performed (Figure 4).As shown, palladacycle 1 presents a 39% mass loss below 300 °C which corresponds with the degradation of the 4,4′-dichlorobenzophenone oxime.On the other hand, regarding the support, GO exhibited around a 13% mass loss below 100 °C and almost a 27% loss at 210 °C resulting from the removal of the labile oxygen-containing functional groups and H2O.Finally, 210 °C resulting from the removal of the labile oxygen-containing functional groups and H2O.Finally, 1-GO exhibited a 28% mass loss at 200 °C due to the loss of the carbonaceous oxygenated species from GO and the loss of the organic ligand of the palladacycle.A final and significant common drop in mass was observed for the three materials around at 925 °C connected with the pyrolysis of the carbon skeletons (Figure 4).According to the TGA, the Pd content for catalyst 1-GO (measured over different material batches) was between 1.35% to 2% which slightly differs from the content obtained from the more accurate ICP-OES technique (1.10% Pd).Thermogravimetric analysis (TGA) of 1 (line blue), GO (black), and 1-GO (line red) was next performed (Figure 4).As shown, palladacycle 1 presents a 39% mass loss below 300 • C which corresponds with the degradation of the 4,4 -dichlorobenzophenone oxime.On the other hand, regarding the support, GO exhibited around a 13% mass loss below 100 • C and almost a 27% loss at 210 • C resulting from the removal of the labile oxygen-containing functional groups and H 2 O. Finally, 1-GO exhibited a 28% mass loss at 200 • C due to the loss of the carbonaceous oxygenated species from GO and the loss of the organic ligand of the palladacycle.A final and significant common drop in mass was observed for the three materials around at 925 • C connected with the pyrolysis of the carbon skeletons (Figure 4).According to the TGA, the Pd content for catalyst 1-GO (measured over different material batches) was between 1.35% to 2% which slightly differs from the content obtained from the more accurate ICP-OES technique (1.10% Pd).
Catalysts 2017, 7, 94 5 of 16 Catalyst 1-GO was also characterized by solid UV-vis spectroscopy.As depicted in Figure 5, 4,4′-dichlorobenzophenone oxime (EtOH solution) showed the characteristic absorption C band at 265 nm.This absorption band was slightly blue shifted to 261 nm for palladacycle 1, which also exhibited the specific metal-to-ligand charge transfer band at about 357 nm.As shown in the UV-vis absorption spectra, graphene oxide exhibited two absorption peaks, a maximum at 230 nm Catalyst 1-GO was also characterized by solid UV-vis spectroscopy.As depicted in Figure 5, 4,4 -dichlorobenzophenone oxime (EtOH solution) showed the characteristic absorption C band at 265 nm.This absorption band was slightly blue shifted to 261 nm for palladacycle 1, which also exhibited the specific metal-to-ligand charge transfer band at about 357 nm.As shown in the UV-vis absorption spectra, graphene oxide exhibited two absorption peaks, a maximum at 230 nm corresponding to π/π* transitions of aromatic C-C bonds, and a shoulder at 293 nm attributed to n/π* transitions of C=O bonds [74].Finally, in the UV-vis spectrum of 1-GO, the absorption C band from the palladacycle at 262 nm could be detected (Figure 5).Catalyst 1-GO was also characterized by solid UV-vis spectroscopy.As depicted in Figure 5, 4,4′-dichlorobenzophenone oxime (EtOH solution) showed the characteristic absorption C band at 265 nm.This absorption band was slightly blue shifted to 261 nm for palladacycle 1, which also exhibited the specific metal-to-ligand charge transfer band at about 357 nm.As shown in the UV-vis absorption spectra, graphene oxide exhibited two absorption peaks, a maximum at 230 nm corresponding to π/π* transitions of aromatic C-C bonds, and a shoulder at 293 nm attributed to n/π* transitions of C=O bonds [73].Finally, in the UV-vis spectrum of 1-GO, the absorption C band from the palladacycle at 262 nm could be detected (Figure 5).Palladacycles 1 and 2 have been demonstrated by our group to be very active precatalysts in the Suzuki-Miyaura reaction under organic [75] and aqueous conditions [76].On the other hand, the amphiphilic character of GO (hydrophilic edges and hydrophobic basal plane) [77] should be beneficial in aqueous processes, converting this material in a potential phase-transfer catalyst.Thus, the synthesized materials 1-GO, 2-GO, and 3-GO were tested as precatalysts in the aqueous Suzuki-Miyaura cross-coupling between 4-bromoanisole and phenylboronic acid at room temperature under low loading conditions (0.02 mol % Pd) (Table 1).Initially, we demonstrated that GO (2 wt %) was not active in the cross-coupling process (Table 1, entry 1).As depicted in entries 2-4, the non-supported palladacycles 1-3 afforded 4-methoxybiphenyl (4) with high yields after 20 h (83, 80, and 99%, respectively).Similar results were obtained for the supported catalysts 1-GO (85%) and 3-GO (82%) in the cross-coupling reaction, while catalyst 2-GO afforded 4 in a lower 69% conversion (Table 1, entries 5-7).In order to compare catalytic activities under the studied reaction conditions, Herrmann's catalyst (5) (Herrmann's catalyst was purchased from Aldrich (Madrid, Spain)) was also supported in GO following immobilization method B (Scheme 1) to afford catalyst 5-GO with a 3.22% of Pd content according to the ICP-OES analysis (see SI for the synthesis and characterization data).As depicted in entries 8 and 9, both 5 and 5-GO afforded 4-methoxybiphenyl with good conversion, showing similar catalytic activities than the rest of the tested palladacyclic catalysts.1, entries 5-7).In order to compare catalytic activities under the studied reaction conditions, Herrmann's catalyst (5) (Herrmann's catalyst was purchased from Aldrich (Madrid, Spain)) was also supported in GO following immobilization method B (Scheme 1) to afford catalyst 5-GO with a 3.22% of Pd content according to the ICP-OES analysis (see SI for the synthesis and characterization data).As depicted in entries 8 and 9, both 5 and 5-GO afforded 4-methoxybiphenyl with good conversion, showing similar catalytic activities than the rest of the tested palladacyclic catalysts.The activity of immobilized palladium nanoparticles on graphene nanoplatelets (PdNPs-G) and reduced graphene oxide (PdNPs-rGO) as a catalyst was also evaluated in the model Suzuki-Miyaura coupling.As previously described, these supported catalysts, obtained from NanoInnova Technologies S.L., have shown good activity in the Suzuki cross-coupling of potassium aryltrifluoroborates with aryl halides [64].However, under the studied conditions, these materials afforded 4-methoxybiphenyl (4) in a 30 and 40% conversion, respectively (Table 1, entries 10 and 11).
In view of the obtained results, it became evident that a reduction in the catalyst loading was mandatory to determine the most active catalyst in the model Suzuki cross-coupling.Thus, supported catalysts 1-GO, 2-GO, 3-GO, and 5-GO were tested reducing the catalyst loading down to 0.002 mol % of Pd in order to test the limits of the new materials.As depicted in Table 1 (entries 12-15), only oximepalladacycle-derived catalyst 1-GO was active enough to give 4-methoxybiphenyl in a good 73% isolated yield at room temperature.

Entry
Pd Catalyst Mol % Pd 4a (Conv., %)  The activity of immobilized palladium nanoparticles on graphene nanoplatelets (PdNPs-G) and reduced graphene oxide (PdNPs-rGO) as a catalyst was also evaluated in the model Suzuki-Miyaura coupling.As previously described, these supported catalysts, obtained from NanoInnova Technologies S.L., have shown good activity in the Suzuki cross-coupling of potassium aryltrifluoroborates with aryl halides [65].However, under the studied conditions, these materials afforded 4-methoxybiphenyl (4) in a 30 and 40% conversion, respectively (Table 1, entries 10 and 11).
In view of the obtained results, it became evident that a reduction in the catalyst loading was mandatory to determine the most active catalyst in the model Suzuki cross-coupling.Thus, supported catalysts 1-GO, 2-GO, 3-GO, and 5-GO were tested reducing the catalyst loading down to 0.002 mol % of Pd in order to test the limits of the new materials.As depicted in Table 1 (entries 12-15), only oxime-palladacycle-derived catalyst 1-GO was active enough to give 4-methoxybiphenyl in a good 73% isolated yield at room temperature.
Next, we performed a substrate scope using catalyst 1-GO under the optimized reaction conditions (Table 2).Initially, we checked the reactivity of different boron-derived nucleophiles with 4-bromoanisole (Table 2, entries 1-4), none of them being as active as the phenylboronic acid.Then, we confirmed the activity of the catalytic system using deactivated aryl bromides as electrophiles, since not only 4-bromoanisole but also 4-bromophenol afforded, after reaction with phenylboronic acid, the corresponding [1,1 -biphenyl]-4-ol in a 61% yield (Table 2, entry 5).Both the activated 4-bromoacetophenone and the neutral bromobenzene reacted with phenylboronic acid and 4-tolylboronic acid to afford biphenyls 4c and 4d in 66 and 94% yields, respectively.On the other hand, as depicted in entries 8 and 9, 1-GO seemed to be very sensitive towards the steric hindrance of the reactants since a 42% yield was obtained for the reaction between 2-bromotoluene and phenylboronic acid (entry 9), while no reaction was observed when 4-bromoanisole reacted with (2,6-dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (  dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).acid (entry 9), while no reaction was observed when 4-bromoanisole reacted with (2,6dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).acid (entry 9), while no reaction was observed when 4-bromoanisole reacted with (2,6dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).acid (entry 9), while no reaction was observed when 4-bromoanisole reacted with (2,6dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).reactants since a 42% yield was obtained for the reaction between 2-bromotoluene and phenylboronic acid (entry 9), while no reaction was observed when 4-bromoanisole reacted with (2,6dimethylphenyl)boronic acid (Table 2, entry 9).Regarding the electronic nature of the nucleophile, the electron-rich 4-tolylboronic acid afforded 4g in a 68% yield after reaction with 4-bromoanisole (Table 2, entry 10).On the contrary, the electron-poor 4-(trifluoromethyl)phenylboronic acid showed, as expected, a very low reactivity with the same deactivated electrophile giving biphenyl 4h in a 12% yield (Table 2, entry 11).Finally, we could also perform a Suzuki alkenylation of 4-bromoanisole with styrylboronic acid, which afforded compound 4i in a 35% yield (Table 2, entry 12).With respect to the active species involved in the Suzuki reaction employing 1-GO as a precatalyst, it is well-known that oxime palladacycles act as stable precursors to highly active and truly Pd(0) catalyst, usually generated after transmetallation by the activated boron nucleophile followed by a reductive elimination reaction of the aryl-ligated palladacycle [77].The presence of water in the reaction medium, as demonstrated by Blackmond [78], would also play a key role in accelerating the initial formation of the monomeric active catalytic species.As commented above, palladium nanoparticles on graphene nanoplatelets (PdNPs-G, average nanoparticle size: 4.50 nm) and reduced graphene oxide (PdNPs-rGO, average nanoparticle size: 10.9 nm) have shown lower activity in the model Suzuki reaction under the optimized reaction conditions, probably due to the agglomeration of the nanoparticles under the reaction conditions.The slow release of the active Pd species from the oxime palladacycle in 1-GO, would account for the better activity of this material in the Suzuki reaction.
A simple recovery and reuse of transition-metal immobilized catalytic systems is highly desirable from both economic and environmental points of view.Therefore, a study of 1-GO recyclability was carried out on the model reaction, i.e., the coupling between 4-bromoanisole and phenylboronic acid using 0.1 mol % Pd as catalyst loading.After each cycle, the catalyst was easily separated from the reaction mixture by washing the crude reaction with a EtOAc/H2O: 1/1 solvent mixture and subsequent centrifugation (see SI, for details).As depicted in Scheme 2, catalyst 1-GO showed good catalytic activity for the two first reactions runs, decreasing the conversion of the reaction from the third cycle.With respect to the active species involved in the Suzuki reaction employing 1-GO as a precatalyst, it is well-known that oxime palladacycles act as stable precursors to highly active and truly Pd(0) catalyst, usually generated after transmetallation by the activated boron nucleophile followed by a reductive elimination reaction of the aryl-ligated palladacycle [77].The presence of water in the reaction medium, as demonstrated by Blackmond [78], would also play a key role in accelerating the initial formation of the monomeric active catalytic species.As commented above, palladium nanoparticles on graphene nanoplatelets (PdNPs-G, average nanoparticle size: 4.50 nm) and reduced graphene oxide (PdNPs-rGO, average nanoparticle size: 10.9 nm) have shown lower activity in the model Suzuki reaction under the optimized reaction conditions, probably due to the agglomeration of the nanoparticles under the reaction conditions.The slow release of the active Pd species from the oxime palladacycle in 1-GO, would account for the better activity of this material in the Suzuki reaction.
A simple recovery and reuse of transition-metal immobilized catalytic systems is highly desirable from both economic and environmental points of view.Therefore, a study of 1-GO recyclability was carried out on the model reaction, i.e., the coupling between 4-bromoanisole and phenylboronic acid using 0.1 mol % Pd as catalyst loading.After each cycle, the catalyst was easily separated from the reaction mixture by washing the crude reaction with a EtOAc/H2O: 1/1 solvent mixture and subsequent centrifugation (see SI, for details).As depicted in Scheme 2, catalyst 1-GO showed good catalytic activity for the two first reactions runs, decreasing the conversion of the reaction from the third cycle.With respect to the active species involved in the Suzuki reaction employing 1-GO as a pre-catalyst, it is well-known that oxime palladacycles act as stable precursors to highly active and truly Pd(0) catalyst, usually generated after transmetallation by the activated boron nucleophile followed by a reductive elimination reaction of the aryl-ligated palladacycle [78].The presence of water in the reaction medium, as demonstrated by Blackmond [79], would also play a key role in accelerating the initial formation of the monomeric active catalytic species.As commented above, palladium nanoparticles on graphene nanoplatelets (PdNPs-G, average nanoparticle size: 4.50 nm) and reduced graphene oxide (PdNPs-rGO, average nanoparticle size: 10.9 nm) have shown lower activity in the model Suzuki reaction under the optimized reaction conditions, probably due to the agglomeration of the nanoparticles under the reaction conditions.The slow release of the active Pd species from the oxime palladacycle in 1-GO, would account for the better activity of this material in the Suzuki reaction.
A simple recovery and reuse of transition-metal immobilized catalytic systems is highly desirable from both economic and environmental points of view.Therefore, a study of 1-GO recyclability was carried out on the model reaction, i.e., the coupling between 4-bromoanisole and phenylboronic acid using 0.1 mol % Pd as catalyst loading.After each cycle, the catalyst was easily separated from the reaction mixture by washing the crude reaction with a EtOAc/H 2 O: 1/1 solvent mixture and subsequent centrifugation (see SI, for details).As depicted in Scheme 2, catalyst 1-GO showed good catalytic activity for the two first reactions runs, decreasing the conversion of the reaction from the third cycle.Figure 6 shows the reasons why 1-GO undergoes a loss of catalytic activity along the successive reaction cycles.The Pd3d X-ray photoelectron spectroscopy (XPS) spectrum (Figure 6a) indicated the presence of two different Pd(II) entities.The component at the higher binding energy was assigned to oxime palladacycle 1, while the component at the lower binding energy [336.2(Pd3d5/2) and 341.8 (Pd3d3/2) eV] was assigned to PdO due to the surface oxidation of the Pd NPs generated from the palladacycle during the catalytic cycles [79].C1s XPS analysis also showed a marked decrease of the oxygen content in 1-GO after the fifth run: Fresh 1-GO (O/C: 0.714); 1-GO after the fifth cycle (O/C: Scheme 2. 1-GO recyclability and yields in the Suzuki-Miyaura coupling in different reaction cycles. Figure 6 shows the reasons why 1-GO undergoes a loss of catalytic activity along the successive reaction cycles.The Pd3d X-ray photoelectron spectroscopy (XPS) spectrum (Figure 6a) indicated the presence of two different Pd(II) entities.The component at the higher binding energy was assigned to oxime palladacycle 1, while the component at the lower binding energy [336.2(Pd3d5/2) and 341.8 (Pd3d3/2) eV] was assigned to PdO due to the surface oxidation of the Pd NPs generated from the palladacycle during the catalytic cycles [80].C1s XPS analysis also showed a marked decrease of the oxygen content in 1-GO after the fifth run: Fresh 1-GO (O/C: 0.714); 1-GO after the fifth cycle (O/C: 0.371).The low catalytic activity of palladium oxide, especially being agglomerated as demonstrated by the TEM analysis (Figure 6b), would account for the observed decrease of the catalytic activity of the 1-GO system.Also, inductively coupled plasma mass spectrometry (ICP-OES) analysis of the washings after the fifth run showed 67 ppb of Pd leaching.Figure 6 shows the reasons why 1-GO undergoes a loss of catalytic activity along the successive reaction cycles.The Pd3d X-ray photoelectron spectroscopy (XPS) spectrum (Figure 6a) indicated the presence of two different Pd(II) entities.The component at the higher binding energy was assigned to oxime palladacycle 1, while the component at the lower binding energy [336.2(Pd3d5/2) and 341.8 (Pd3d3/2) eV] was assigned to PdO due to the surface oxidation of the Pd NPs generated from the palladacycle during the catalytic cycles [79].C1s XPS analysis also showed a marked decrease of the oxygen content in 1-GO after the fifth run: Fresh 1-GO (O/C: 0.714); 1-GO after the fifth cycle (O/C: 0.371).The low catalytic activity of palladium oxide, especially being agglomerated as demonstrated by the TEM analysis (Figure 6b), would account for the observed decrease of the catalytic activity of the 1-GO system.Also, inductively coupled plasma mass spectrometry (ICP-OES) analysis of the washings after the fifth run showed 67 ppb of Pd leaching.Oxime palladacycles are considered reservoirs of highly active palladium nanoparticles [65][66][67].Usually, PdNps supported on graphene materials are prepared by reduction of a Pd salt or Pd catalyst in the presence of the corresponding carbonaceous support [44][45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60][61].Then, we next turned our attention to the reduction of 1 with NaBH4 and the catalytic activity of the supported PdNPs was thus obtained.When unsupported palladacycle 1 was reduced with NaBH4 in the presence of GO (Figure 7, Eq. a) [80], a carbonaceous material [Pd(0)-rGO] with an 8.40% Pd content (by ICP-OES analysis) was obtained.XPS analysis of this material showed a 75.8% content of Pd(0), which unfortunately suffered agglomeration during the reduction process as demonstrated by TEM analysis (Scheme 3 Eq.a).On the other hand, the reduction of 1-GO with NaBH4 (Scheme 3, Eq.b) afforded a material [Pd(II)-rGO] which, according to ICP-OES analysis, contained a 5.97% Pd.XPS analysis of [Pd(II)-rGO] demonstrated the stability of palladacycle under these reduction conditions since 76.8% of 1 was not reduced.Both materials suffered a strong decrease of oxygen content as a consequence of the reduction of GO (O/C: 0.655) to rGO (O/C: 0.237-0.369).Oxime palladacycles are considered reservoirs of highly active palladium nanoparticles [66][67][68].Usually, PdNps supported on graphene materials are prepared by reduction of a Pd salt or Pd catalyst in the presence of the corresponding carbonaceous support [45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60][61][62].Then, we next turned our attention to the reduction of 1 with NaBH 4 and the catalytic activity of the supported PdNPs was thus obtained.When unsupported palladacycle 1 was reduced with NaBH 4 in the presence of GO (Scheme 3, Eq.a) [81], a carbonaceous material [Pd(0)-rGO] with an 8.40% Pd content (by ICP-OES analysis) was obtained.XPS analysis of this material showed a 75.8% content of Pd(0), which unfortunately suffered agglomeration during the reduction process as demonstrated by TEM analysis (Scheme 3 Eq.a).On the other hand, the reduction of 1-GO with NaBH 4 (Scheme 3, Eq.b) afforded a material [Pd(II)-rGO] which, according to ICP-OES analysis, contained a 5.97% Pd.XPS analysis of [Pd(II)-rGO] demonstrated the stability of palladacycle under these reduction conditions since 76.8% of 1 was not reduced.Both materials suffered a strong decrease of oxygen content as a consequence of the reduction of GO (O/C: 0.655) to rGO (O/C: 0.237-0.369).
We also studied the activity of Pd(0)-rGO and Pd(II)-rGO in the model Suzuki reaction between 4-bromoanisole and phenylboronic acid under the optimized reaction conditions (0.002 mol % Pd) (Scheme 4).As expected, Pd(0)-rGO was catalytically inactive in the process.On the other hand, Pd(II)-rGO afforded 4a in a 40% yield (Scheme 4).This result ratified the important amphiphilic effect of GO on the Suzuki coupling carried out under aqueous conditions.Palladacycle 1 was also immobilized over rGO following Method B (see SI), but the material thus obtained (1-rGO) only contained a 0.035% of Pd, according to the ICP-OES analysis.We also studied the activity of Pd(0)-rGO and Pd(II)-rGO in the model Suzuki reaction between 4-bromoanisole and phenylboronic acid under the optimized reaction conditions (0.002 mol % Pd) (Scheme 4).As expected, Pd(0)-rGO was catalytically inactive in the process.On the other hand, Pd(II)-rGO afforded 4a in a 40% yield (Scheme 4).This result ratified the important amphiphilic effect of GO on the Suzuki coupling carried out under aqueous conditions.Palladacycle 1 was also immobilized over rGO following Method B (see SI), but the material thus obtained (1-rGO) only contained a 0.035% of Pd, according to the ICP-OES analysis.

General
Unless otherwise noted, all commercial reagents and solvents were used without further purification. 1H-NMR (300 MHz) and 13 C-NMR (75 MHz) spectra were obtained on a Bruker AC-300, using CDCl3 as solvent and TMS (0.003%) as reference, unless otherwise stated.Low-resolution mass spectra (MS) were recorded in the electron impact mode (EI, 70 eV, He as carrier phase) using an Agilent 5973 Network Mass Selective Detector spectrometer, being the samples introduced through a GC chromatograph Agilent 6890N equipped with a HP-5MS column [(5%-phenyl)methylpolysiloxane; length 30 m; ID 0.25 mm; film 0.25 mm].Analytical TLC was performed on Merck aluminum sheets with silica gel 60 F254.Silica gel 60 (0.04-0.06 mm) was employed for flash chromatography.The conversion of the reactions was determined by GC analysis on an Agilent 6890N Network GC system.Centrifugations were carried out in a Hettich centrifuge (Universal 320, 6000 rpm, 15 minutes).ICP-MS analyses were performed on an Agilent 7700x equipped with HMI We also studied the activity of Pd(0)-rGO and Pd(II)-rGO in the model Suzuki reaction between 4-bromoanisole and phenylboronic acid under the optimized reaction conditions (0.002 mol % Pd) (Scheme 4).As expected, Pd(0)-rGO was catalytically inactive in the process.On the other hand, Pd(II)-rGO afforded 4a in a 40% yield (Scheme 4).This result ratified the important amphiphilic effect of GO on the Suzuki coupling carried out under aqueous conditions.Palladacycle 1 was also immobilized over rGO following Method B (see SI), but the material thus obtained (1-rGO) only contained a 0.035% of Pd, according to the ICP-OES analysis.Scheme 4. Suzuki coupling using rGO as a catalyst support.

General
Unless otherwise noted, all commercial reagents and solvents were used without further purification. 1H-NMR (300 MHz) and 13 C-NMR (75 MHz) spectra were obtained on a Bruker AC-300, using CDCl3 as solvent and TMS (0.003%) as reference, unless otherwise stated.Low-resolution mass spectra (MS) were recorded in the electron impact mode (EI, 70 eV, He as carrier phase) using an Agilent 5973 Network Mass Selective Detector spectrometer, being the samples introduced through a GC chromatograph Agilent 6890N equipped with a HP-5MS column [(5%-phenyl)methylpolysiloxane; length 30 m; ID 0.25 mm; film 0.25 mm].Analytical TLC was performed on Merck aluminum sheets with silica gel 60 F254.Silica gel 60 (0.04-0.06 mm) was employed for flash chromatography.The conversion of the reactions was determined by GC analysis on an Agilent 6890N Network GC system.Centrifugations were carried out in a Hettich centrifuge (Universal 320, 6000 rpm, 15 minutes).ICP-MS analyses were performed on an Agilent 7700x equipped with HMI Scheme 4. Suzuki coupling using rGO as a catalyst support.

General
Unless otherwise noted, all commercial reagents and solvents were used without further purification. 1H-NMR (300 MHz) and 13 C-NMR (75 MHz) spectra were obtained on a Bruker AC-300, using CDCl 3 as solvent and TMS (0.003%) as reference, unless otherwise stated.Low-resolution mass spectra (MS) were recorded in the electron impact mode (EI, 70 eV, He as carrier phase) using an Agilent 5973 Network Mass Selective Detector spectrometer, being the samples introduced through a GC chromatograph Agilent 6890N equipped with a HP-5MS column [(5%-phenyl)-methylpolysiloxane; length 30 m; ID 0.25 mm; film 0.25 mm].Analytical TLC was performed on Merck aluminum sheets with silica gel 60 F254.Silica gel 60 (0.04-0.06 mm) was employed for flash chromatography.The conversion of the reactions was determined by GC analysis on an Agilent 6890 N Network GC system.Centrifugations were carried out in a Hettich centrifuge (Universal 320, 6000 rpm, 15 min).ICP-MS analyses were performed on an Agilent 7700x equipped with HMI (high matrix introduction) and He mode ORS as standard.Elemental analyses were determined with a CHNS elemental micro analyzer with Micro detection system TruSpec LECO.X-ray Powder Diffraction (XRD) was performed in a Bruker D8-Advance with mirror Goebel (non-planar samples) with high temperature Chamber (up to 900 • C), with a generator of x-ray KRISTALLOFLEX K 760-80F (power: 3000 W, voltage: 20-60 KV and current: 5-80 mA) with a tube of RX with copper anode.X-ray photoelectron spectroscopy (XPS) was performed in a VG-Microtech Mutilab 3000 equipment equipped with a hemispherical electron analyzer with 9 channeltrons (with energy of passage of (2-200 eV) and an X-ray radiation source with Mg and Al anodes.Transmission electron microscopy (TEM) was performed in JEOL Model JEM-2010.This microscope features an OXFORD X-ray detector model INCA Energy TEM 100 for microanalysis (EDS).The image acquisition camera is of the brand GATAN model ORIUS SC600.It is mounted on the axis with the microscope at the bottom and is integrated into the GATAN Digital Micrograph 1.80.70 image acquisition and processing program for GMS 1.8.0.The supported catalyst were sonicated in a ultrasons P-Selecta (360 W).Solid state UV-Vis spectroscopy was performed in a JASCO V-670 dual-beam UV-Vis/NIR spectrometer covering the wavelength range from 190 to 2700 nm.The equipment has a single monochromator with double netting, one for the UV-Vis region (1200 grooves/mm) and one for the NIR region (300 grooves/mm).The detectors are a photomultiplier tube for the UV-Vis region and a PbS detector for the NIR region.The switching of both the detectors and the networks is automatically effected at a wavelength set by the user between 750 and 900 nm.The sources used are a deuterium lamp (190 to 350 nm) and a halogen lamp (330 to 2700 nm).UV-Vis spectroscopy analysis of oxime precursor of catalyst 1 was performed in ethanol solution (1.04 mg of oxime in 20 mL of ethanol aproxymately) in a SHIMAZU UV-1603 spectrophotometer covering the wavelength range from 190 to 2700 nm.

Synthesis of Oxime Palladacycle 1
A MeOH (0.8 mL) solution of the oxime derived from 4,4 -dichlorobenzophenone (1 eq), sodium acetate (61.70 mg, 0.752 mmol, 1 eq) and Li 2 PdCl 4 (1.5 mL of a 0.5 M solution in MeOH, 1eq) was stirred at room temperature under argon atmosphere for 4 days.Then, water was added to precipitate the corresponding palladium complex and subsequently, the palladacycle was filtered and washed with water and hexane.Finally, the catalyst was dried under reduced pressure overnight giving oxime palladacycle 1 as a yellow solid (47% yield).

Synthesis of 1-GO
A 10 mL glass vessel was charged with GO (200 mg) and 1 (45.89mg).Then, anhydrous THF (4 mL) was added and the reaction mixture was sonicated for 1 h.Afterwards, the reaction was stirred at room temperature for 48 h.After this time, the catalyst was submitted to four washing (THF, 10 mL)-centrifugation (6000 rpm, 20 min) cycles, with the solvent being eliminated after each cycle with a syringe equipped with a 4 mm/0.2µm PTFE syringe filter.The residual solvent was completely removed under reduced pressure affording 1-GO (2% Pd by ICP-OES).

Typical Procedure for the Recovery of the Catalyst in the Suzuki-Miyaura Reaction
Once the reaction was finished, the mixture was suspended and stirred for 15 min in a 10 mL mixture of EtOAc/H 2 O: 1/1.Then, this mixture was centrifuged (6000 rpm, 20 min) and the solvent was eliminated using a syringe equipped with 4 mm/0.2µm PTFE syringe filter.The washing/centrifugation sequence was repeated four additional times until no product was detected in the liquid phase by thin layer chromatography.The residual solvent was completely removed under reduced pressure affording the supported palladacycle catalyst which was directly used in the same tube with fresh reagents for the next run.This procedure was repeated for every cycle and the conversion of the reaction was determined by GC using decane as internal standard.

Scheme 3 .
Scheme 3. Palladacycle reduction studies in the presence of GO.

Scheme 4 .
Scheme 4. Suzuki coupling using rGO as a catalyst support.

Scheme 3 .Scheme 3 .
Scheme 3. Palladacycle reduction studies in the presence of GO.
1Reaction conversion towards 4 determined by GC analysis.Isolated yield in brackets.