Graphene Supported Rhodium Nanoparticles for Enhanced Electrocatalytic Hydrogen Evolution Reaction

Current research on catalysts for proton exchange membrane fuel cells (PEMFC) is based on obtaining higher catalytic activity than platinum particle catalysts on porous carbon. In search of a more sustainable catalyst other than platinum for the catalytic conversion of water to hydrogen gas, a series of nanoparticles of transition metals viz., Rh, Co, Fe, Pt and their composites with functionalized graphene such as RhNPs@f-graphene, CoNPs@f-graphene, PtNPs@f-graphene were synthesized and characterized by SEM and TEM techniques. The SEM analysis indicates that the texture of RhNPs@f-graphene resemble the dispersion of water droplets on lotus leaf. TEM analysis indicates that RhNPs of <10 nm diameter are dispersed on the surface of f-graphene. The air-stable NPs and nanocomposites were used as electrocatalyts for conversion of acidic water to hydrogen gas. The composite RhNPs@f-graphene catalyses hydrogen gas evolution from water containing p-toluene sulphonic acid (p-TsOH) at an onset reduction potential, Ep, −0.117 V which is less than that of PtNPs@f-graphene (Ep, −0.380 V) under identical experimental conditions whereas the onset potential of CoNPs@f-graphene was at Ep, −0.97 V and the FeNPs@f-graphene displayed onset potential at Ep, −1.58 V. The pure rhodium nanoparticles, RhNPs also electrocatalyse at Ep, −0.186 V compared with that of PtNPs at Ep, −0.36 V and that of CoNPs at Ep, −0.98 V. The electrocatalytic experiments also indicate that the RhNPs and RhNPs@f-graphene are stable, durable and they can be recycled in several catalytic experiments after washing with water and drying. The results indicate that RhNPs and RhNPs@f-graphene are better nanoelectrocatalysts than PtNPs and the reduction potentials were much higher in other transition metal nanoparticles. The mechanism could involve a hydridic species, Rh-H− followed by interaction with protons to form hydrogen gas.

The design of a complex that can catalyze electrochemical hydrogen activation at lower potentials than platinum electrode under ambient conditions is a challenge. The presence of the dimetallic iron and heterodimetallic iron-nickel centers at the active sites of the hydrogenases has invoked to search for a non-platinum catalyst material for the purpose [1][2][3][4][5][6][7][8] . Hydrogenases are enzymes that catalyze the interconversion of H 2 and its constituents, two protons and two electrons as shown in Eq. (1).
The three known classes of hydrogenases, [NiFe]-, [FeFe]-and FeS-cluster free hydrogenases contain iron at their active sites which are coordinated by thiolates, CO, CN − or a light sensitive cofactor 2 . The [Fe 4 (μ 3 -S 4 )] cubane type sub-cluster obtains electrons from pyruvate oxidation and transfer those electrons to the [FeFe] site which utilizes them and form hydrogen gas in an unique reaction mechanism. The mechanism involves the replacement of the labile water ligand shown in Fig. 1 by a proton followed by reductive elimination supported by a free cysteine residue (Cys 299 ) at the vicinity of the [FeFe] cluster 1 .
Although nature uses the potential redox active element "iron" in this active site, the hydrogen gas electrode is based on platinum metal. Electrocatalytic H 2 gas generation is the conversion of electricity to H 2 gas in the presence of a catalyst that can performance-wise replace the platinum electrode which functions at E p , −0.413 V at a pH value of 7.0 8 . Iron-containing electrochemical catalysts that are known to date include mononuclear iron(II) complexes, [Fe II (S 2 C 6 H 4 )(CO) 2 (PMe 3 ) 2 ], [Fe II (S 2 C 6 H 2 Cl 2 )(CO) 2 (PMe 3 ) 2 ], an FeS cluster with Fe cubane (μ-SR)Fe subsite and [Fe 2 (CN)(CO) 4 H(PMe 3 )(S 2 C 3 H 6 )] which function at high reduction potentials, Ep ≈ −1.0 V 9-16 . However, several electrochemical catalysts that contain other transition metals are known, e.g. carbon nanotube supported [Ni(PPh 2 NPh 2 ) 2 (CH 3 CN)][BF 4 ] 2 , (CpMoμ-S) 2 S 2 CH 2 , cobalt macrocyclic glyoxime/tetraimine complexes and a nickel(II) dithiolene complex which function at considerably lower reduction potentials. Several other electrochemical hydrogen evolution catalysts have also been reported including metalloporphyrins, low-valent transition metal complexes forming hydrides upon reaction with acids, mononuclear iron(II) complexes, cobalt-dithiolene complexes and a nickel complex [Ni(P Ph 2 N Ph ) 2 ][BF 4 ] 2 (P Ph 2 N Ph = 1,3,6-triphenyl-1aza-3,6-diphosphacycloheptane) which electrocatalyze H 2 production with high turnover frequencies but at significantly high reduction potentials E p > −1.1 V 17-24 . Various kinds of heterogeneous non-precious metal based electrocatalysts for hydrogen evolution reaction (HER) and oxygen reduction reaction (ORR) including metal sulfides, selenides, carbides, nitrides, phosphides, three dimensional porous carbon nanostructures and heteroatom doped nanocarbons have been reviewed [25][26][27] . Multilayer thin films of metal nanocrystals and graphene quantum dots have been prepared by layer-by-layer assembly approach and these composites demonstrate efficient and versatile electrocatalytic performances such as reduction of aromatic nitro compounds, methanol oxidation and water splitting 28 . Plasmonic TiO 2 NRs@Ag@GQDs ternary heterostructures which have been prepared by a layer-by-layer assembly strategy combined with an in situ light irradiation display an enhanced photoelectrochemical water splitting performance 29 . Pd-CdS nanowire heterostructures display remarkable photocatalytic reduction of nitroarenes and photocatalytic hydrogen production under visible light irradiation 30 . Our research focused towards the electrochemical hydrogen evolution resulted in a mononuclear iron(III) dithiolene of severely distorted square pyramidal geometry and a nickel(II)-sulfur based radical ligand complex that catalyze electrochemical hydrogen gas evolution at lower potentials in CH 3 CN. In these cases, along with metals, the dithiolene ligands are also potentially redox active forming sulfur based radicals [31][32][33][34][35][36] .
It has been demonstrated that Pt 3 Ni(111) surface is 90-fold more active for ORR than the current platinum catalyst on porous carbon used in PEMFC 37 . Core-shelled TiC@TiO 2 has been shown to be a promising catalyst support for proton exchange membrane fuel cells (PEMFCs). TiC is thermally stable with low solubility in sulfuric acid and high electronic conductivity. Both these materials are used as supports for platinum and platinumpalladium alloy catalysts (Pt/TiC, Pt 3 Pd/TiC and Pt 3 Pd/TiC@TiO 2 ) and their catalytic activity toward ORR are much higher than those for Pt/TiC 38 . Earth abundant transition metal nanocatalysts that include Mn, Fe, Co, Ni and Cu and early transition metals such as Ti, V, Cr, Zr, Nb and their nanocomposites for reduction of various aromatic compounds have been reported 39 . Synthesis of different types of graphene-based composite photocatalysts and their applications in reduction of CO 2 , nitroarenes, methanol oxidation, elimination of pollutants and photochemical water splitting have been reviewed 40 . Au-Pd nanoalloys supported on graphene (Au-Pd/GR) have been reported which display higher photocatalytic performance than the monometallic, GR supported nanoparticles towards degradation of dye, rhodamine B (RhB) 41 . Basic principles of photocatalytic water splitting, engineering strategies for photocatalysts optimization, and promising photocatalytic materials for water splitting have been recently reviewed 42 . Various metal oxides, metal non-oxides and non-metal catalysts for oxygen evolution reaction (OER) have also been recently reviewed 43 . There are several reports on the catalytic activities of rhodium nanoparticles (RhNPs) and rhodium nanoparticles on polymer, graphene and carbon nanotube matrices for the reduction of aromatic compounds, amino boranes as a means of H 2 storage. But there are no reports on the application of RhNPs and graphene supported RhNPs (RhNP@f-graphene) as electrocatalysts for the generation of hydrogen gas from water. Here in we report preparation, characterisation and electrocatalytic properties of rhodium nanoparticles (RhNPs) and graphene supported rhodium nanoparticles (RhNPs@f-graphene) which display better electrocatalytic performance than the platinum nanoparticles (PtNPs) and graphene supported platinum nanoparticles (PtNPs@f-graphene) under similar experimental conditions.

Results and Discussion preparation of nanoparticles of rhodium (Rhnps), platinum (ptnps) and cobalt (conps) (1-3). The detailed synthetic procedures for the preparation of the three nanoparticles RhNPs, PtNPs and
CoNPs, (1-3) are given in methods. The metal halides were reduced by NaBH 4 in water at laboratory temperature (30 °C) under high dilution conditions. The black colored particles which precipitated out immediately were sonicated, centrifuged and isolated after 12 h. The transition metal nanoparticles (TMNP, 1-3) were insoluble in water, they were sonicated in water and deposited on to polymer coated carbon grids and aluminum stubs for transmission electron microscopic (TEM) and scanning electron microscopic (SEM) analysis respectively. The TEM analytical results of the rhodium nanoparticles are shown in Fig. 2A and a higher magnification image of RhNP is diaplayed in Fig. 2B. The RhNPs of <10 nm diameter are found to agglomerate to form nanostructures of 60 nm diameter as shown in Fig. 2B. The SEM images of RhNPs on aluminium stubs are displayed in Fig. 3A which further confirms the agglomeration of RhNPs to nanostructures of 60 nm diameter. The nanoparticles were confirmed to be pure rhodium nanoparticles by EDX analysis as shown in Fig. 3D. Powder XRD spectrum of RhNPs was measured in the range of 2θ from 5° to 90°. As shown in Fig. 3E, a sharp peak was observed at 2θ, 41.2° and broad, weak peaks were displayed at 2θ, 47.4°, 69.5° and 83.8°. The TEM images of the PtNPs and CoNPs are shown in Fig. 4. The PtNPs were polydispersed and are of 50 nm and 30 nm diameters as shown Fig. 4A and the CoNPs were found to group into spherical structures of 89 nm diameter as shown in Fig. 4B. The SEM image of the CoNPs also indicate the agglomeration of the particles as shown in Supplementary Information. preparation of composites, tMnps@f-graphene. The functionalized graphene was obtained from graphitic powder by following a previously reported procedure 31 . Graphitic powder was sonicated in tetrahydrofuran (thf) for 1 h followed by centrifugation. The supernatant thf layer was decanted and the residue was washed with acetone and dried thoroughly which was heated under reflux in a mixture of H 2 SO 4 and HNO 3 . The composites of transition metal nanoparticles and functionalized graphene were obtained by sonication of both in acetonitrile. The solvent was evaporated and the residue was washed with acetone and then finally dried. The composites were characterized by TEM and SEM methods. The TEM images of the RhNP@f-graphene are given in Fig. 2C,D which clearly indicate the presence of graphene sheets and RhNPs. The TEM images further indicate that RhNPs of <10 nm diameter are dispersed on graphene sheets. The aggregated RhNPs are found to segregate www.nature.com/scientificreports www.nature.com/scientificreports/ upon formation of composite with f-graphene as shown in Fig. 2C. The SEM images of the RhNP@f-graphene composite shown in Fig. 3B,C indicate that the dispersion of RhNPs on graphene sheets resemble dispersion of water droplets on lotus leaves. The SEM images of the RhNP@f-graphene further confirm that the size of the rhodium nanoparticles is about 11-13 nm. The EDX analysis of the composite show that the percentage of rhodium is 66.2% and carbon content is 8.3% as shown in Fig. 3F. The powder XRD spectrum of RhNPs@-graphene in the range of 2θ from 5° to 90° is displayed in Fig. 3G which shows an additional weak and broad peak at 2θ, 24.   www.nature.com/scientificreports www.nature.com/scientificreports/ electrocatalytic studies. The transition metal nanoparticles and their f-graphene composites were tested for electrocatalytic efficiency and hydrogen gas generation by cyclic voltammetry. The electrocatalysis was conducted in acidic water containing p-TsOH. Glassy carbon working, platinum wire auxillary and Ag/AgCl reference electrodes were used. In a typical experiment, nanoparticles (0.02 g) were taken in distilled water (8 ml) in an electrochemical cell, with potassium nitrate (0.1 mM) as a supporting electrolyte. Aqueous solutions of p-TsOH were added to the electrochemical cell and the cyclic voltammograms were recorded as a function of increasing concentrations of p-TsOH. The results of electrocatalytic experiments using RhNPs and RhNP@f-graphene as the catalysts are displayed in Figs 5 and 6 respectively. Pure RhNPs did not show any peak (Fig. 5, black). Upon adding aqueous solutions of p-TsOH (0.2 ml of 50 mM) a reduction signal was observed at E p , −0.67 V (Fig. 5) whereas the RhNP@f-graphene displayed the reduction signal at E p , −0.601 V (Fig. 6). The reduction onset potentials were observed at E p , −0.186 V and −0.117 V for RhNPs and RhNP@f-graphene respectively. Compared with the pure RhNPs, the reduction potential of the composite RhNP@f-graphene was shifted by 60 mV towards E p , 0.00 V. Additions of p-TsOH to the cyclic voltammetric cell resulted in increasing currents as shown in Figs 5 and 6. The RhNPs and RhNPs@f-graphene consumed 0.8 g of p-TsOH and the current exceeded 10 mA. Gas bubbles sticking to the glassy carbon working electrode were observed during electrolysis. The gas bubbles were analyzed and confirmed to be hydrogen gas by head space analysis 32 . A mixture of RhNPs@graphene (0.02 g) and p-TsOH (0.8 g, ~4 mmol) in water with KNO 3 (0.2 M) as supporting electrolyte was purged with nitrogen gas for 15 min. A 2 ml syringe was inserted into the electrochemical  cell and the argon gas supply inlet was closed. The reaction mixture was subjected to controlled potential electrolysis at −0.6 V. The gas bubbles formed at the glassy carbon working electrode surface were tapped off to the surface and the gas over the solution was taken into a syringe which was analysed by gas chromatography and confirmed to be H 2 gas. Cyclic voltammograms of functionalized graphene without electrocatalysts were recorded as a function of addition of aqueous solutions of p-TsOH and the results are displayed in Fig. 7. As a function of increasing concentrations of added p-TsOH (black to blue) under similar experimental conditions, only graphene electrocatalyzes proton reduction at a potential of E p , −1.7 V but the onset potentials were observed at E p , −1.1 V. The reduction of pure p-TsOH in water in the absence of RhNPs or RhNPs@f-graphene was observed at a higher negative potential, E p , −1.4 V vs. Ag/Ag + .
Mechanism of proton activation. The SEM analytical results of the composite, RhNPs@f-graphene ( Fig. 3C) indicate that the RhNPs are dispersed over f-graphene layers similar to the dispersion of water droplets on a lotus leaf. The interaction between the RhNPs and functionalized graphene could be π-bonding interaction or dihapto type (η 2 -type) of boding and protons could also bind to rhodium leading to the formation of Rh-H + species. This could be reduced upon application of electrode potential to a hydridic species, Rh-H − as shown in Fig. 8. Structures of rhodium-hydride complexes coordinated to fullerenes (C 60 ) in an η 2 -mode viz., (η 2 -C 60 ) Rh(CO)H(PPh 3 ) 2 have been reviewed 46 . This Rh-H − species could interact with further protons and could form hydrogen gas. In the process, rhodium nanoparticles remain unaffected and provide site for fixing protons which in other words, an adsorption medium. The rhodium metal nanoparticles could play an important role in bringing down the reduction potential, E p , because the reduction of protons and evolution of hydrogen gas takes place at higher reduction potentials when CoNPs (E p , −0.98 V) and FeNPs (E p , −1.59 V) were used as electrocatalysts. This might be due to the ease of formation and optimum stability of Rh-H − species.

Recycling and durability of Rhnps and Rhnps@f-graphene. After the cyclic voltammetric experiments,
the RhNPs and the RhNPs@f-graphene were filtered from the reaction mixture and were analysed by SEM and TEM techniques. The results obtained in the case of RhNPs are shown in Supplementary Information which indicate the presence of mixtures of large crystals along with the RhNPs. The large crystals could be KNO 3 supporting electrolyte and p-toluene sulfonate used in the cyclic voltammetric experiment. The electrocatalysts, RhNPs and RhNPs@f-graphene were recycled as follows: After the cyclic voltammetric experiment, the RhNPs and RhNPs@f-graphene were filtered and were sonicated in water (50 mL) for 30 min. Then they were centrifuged and the water layer was decanted. The process was repeated for 6 times in order to get rid of KNO 3 and p-toluene sulfonate and the recycled and purified RhNPs and RhNPs@f-graphene were reused as the electrocatalysts. The recycled RhNPs were found to electrocatalyze further generation of hydrogen gas by addition of p-TsOH (0.8 g) and the current again exceeded 10 mA. This proves that the RhNPs and RhNPs@f-graphene have excellent durability and electrocatalytic efficiency.
Comparison of electrocatalytic efficiency of RhNPs@f-graphene. In order to compare the catalytic efficiency of RhNPs and RhNP@f-graphene with that of Platinum electrodes under identical experimental conditions, electrocatalytic experiments were conducted using PtNPs and PtNPs@f-graphene. Electrocatalytic experiments were also conducted using CoNPs and CoNPs@f-graphene as electrocatalysts. The cyclic voltammetric results using PtNPs, PtNP@f-graphene, CoNPs, CoNP@f-graphene as electrocatalysts and as a function of increasing concentrations of three additions of p-TSOH are shown in Fig. 9. In these experiments the onset potential which is the lowest potential at which the reduction current starts to flow in the electrochemical system was monitored. The PtNPs displayed the onset potential at E p , −0.367 V (Fig. 9B) where as CoNPs displayed the onset www.nature.com/scientificreports www.nature.com/scientificreports/ potential at Ep , −0.98 V (Fig. 9C). The composite PtNP@f-graphene displayed the onset potential at E p , −0.38 V (Fig. 9E) and the composite, CoNP@f-graphene displayed the onset potential at E p , −0.968 V (Fig. 9F). But the onset potentials for RhNPs and RhNPs@f-graphene are very low at E p , −0.186 V (Fig. 9A) and E p , −0.117 V (Fig. 9D) respectively. The onset potentials for the hydrogen evolution from acidic water by RhNPs@f-graphene, PtNPs@f-graphene and pure RhNPs are compared in Fig. 9G. The cyclic voltammetric profile obtained on using RhNPs@f-graphene as electrocatalyst is displayed in black, that for PtNPs@f-graphene in red and that for the pure RhNPs in blue. These onset potentials clearly indicate the higher catalytic performance of RhNPs and its composite, RhNP@f-graphene than even PtNPs and PtNP@f-graphene under identical experimental conditions. The RhNPs and RhNP@f-graphene were also found to be durable and extremely stable as explained above. Hence it can be concluded that the RhNPs immobilized on the functionalized graphene support is an ideal and better candidate that can substitute the platinum electrode (Platinum on porous carbon support) in fuel cells.

conclusion
In summary, a series of transition metal nanoparticles (RhNPs, PtNPs, CoNPs, FeNPs) and their composites with f-graphene (RhNPs@f-graphene, PtNPs@f-graphene, CoNPs@f-graphene) have been prepared and characterized by SEM, TEM and p-XRD analytical methods. The nanoparticles were mostly aggregated and segregate upon composite formation with f-graphene. The RhNPs immobilized on f-graphene resemble the dispersion of water droplets on a lotus leaf. The RhNPs and the composite RhNPs@f-graphene electrocatalyze hydrogen evolution from acidic water at much lower potentials than that of PtNPs and PtNPs@f-graphene under identical experimental conditions. The reduction onset potentials for RhNPs and RhNPs@f-graphene were found to be E p , −0.117 V and −0.186 V respectively whereas for the PtNPs and PtNPs@f-graphene were E p , −0.367 V and −0.38 V respectively. The CoNPs, FeNPs and CoNPs@f-graphene electrocatalyze hydrogen evolution at higher reduction potentials E p , −0.97 V, −1.58 V and −0.98 V respectively. The lower electropotentials and high current values observed in the case of RhNPs and RhNPs@-graphene reveal the crucial role played by rhodium in electrocatalytic hydrogen evolution. The plausible mechanism could be the formation of hydridic Rh-H − species under the influence of electrode potential which interact with further proton to form hydrogen gas. Methods preparation of Rhnps, ptnps, conps and fenps. The metal halides, RhCl 3 .3H 3 O or anhydrous PtCl 2 or CoCl 2 .6H 2 O or anhydrous FeCl 3 (0.05 g) was stirred in distilled deionized water (30 mL) at 30 °C for 5 min. NaBH 4 (0.02 g) was added to the reaction mixture in portions carefully. A bright red colored solution was obtained in the case of RhCl 3 .6H 2 O which became colorless after the precipitation of black particles. After stirring for 8 h, the black precipitate was collected by filtration. The precipitate was washed with water (10 mL), sonicated with acetone (25 mL), centrifuged and the acetone layer was decanted. The black precipitate thus obtained was dried and weighed (0.02 g; yield, >95%).
preparation of functionalized grapheme. Graphite powder (0.5 g) was taken in THF (80 mL) and water (20 mL), stirred at 38 °C for 2 h and ultra-sonicated for 10 h. The solvents were decanted after centrifugation and the residue was dried thoroughly. This residue was treated carefully and dropwise with concentrated H 2 SO 4 (30 mL) and fuming nitric acid (10 mL) at 0 °C. The reaction mixture was heated under reflux for 12 h and allowed to stand at 38 °C for 10 h. The supernatant acid layer was decanted and the residue was washed thoroughly with water by centrifugation and dried. FT-IR (v, cm −1 ): 3400 (br), 1714 (br.), 1600 (w).