A RhxSy / C Catalyst for the Hydrogen Oxidation and Hydrogen Evolution Reactions in HBr

Rhodium sulfide (Rh2S3) on carbon support was synthesized by refluxing rhodium chloride with ammonium thiosulfate. Thermal treatment of Rh2S3 at high temperatures (600◦C to 850◦C) in presence of argon resulted in the transformation of Rh2S3 into Rh3S4, Rh17S15 and Rh which were characterized by TGA/DTA, XRD, EDX, and deconvolved XPS analyses. The catalyst particle size distribution ranged from 3 to 12 nm. Cyclic voltammetry and rotating disk electrode measurements were used to evaluate the catalytic activity for hydrogen oxidation and evolution reactions in H2SO4 and HBr solutions. The thermally treated catalysts show high activity for the hydrogen reactions. The exchange current densities (io) of the synthesized RhxSy catalysts in H2-saturated 1M H2SO4 and 1M HBr for HER and HOR were 0.9 mA/cm2 to 1.0 mA/cm2 and 0.8 to 0.9 mA/cm2, respectively. The lower io values obtained in 1M HBr solution compared to in H2SO4 might be due to the adsorption of Br− on the active surface. Stable electrochemical active surface area (ECSA) of RhxSy catalyst was obtained for CV scan limits between 0 V and 0.65 V vs. RHE. Scans with upper voltage limit beyond 0.65 V led to decreased and unreproducible ECSA measurements. © The Author(s) 2015. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0901504jes] All rights reserved.

The intermittent availability of renewable electricity from solar and wind sources has increased significantly. 1 To match their often unpredictable power generation cycles with demand, low cost electrical energy storage systems are required.
A flow battery/regenerative fuel cell is an electrochemical storage device that stores the electrical energy as chemical energy in a fuel and converts the chemical energy of the fuel directly back to electrical energy.The potentially low cost and relatively rapid process of electrical-to-chemical or chemical-to-electrical energy, in comparison to the slower mechanical-to-electrical energy processes used with flywheel and compressed air storage, offers unique advantages.The discharge cycle of the flow battery utilizes a fuel cell which has proven to be efficient and clean devices for energy conversion.6] The core component of a H 2 /Br 2 regenerative fuel cell system is an acid-based H 2 /Br 2 fuel cell.The charge and discharge reactions occurring in the fuel cell are as follows.At the bromine electrode, during the charge cycle, bromide ions in an HBr solution are oxidized to form bromine and two electrons; The standard reduction potential for this reaction is 1.09 V.The hydronium ions migrate across proton conducting membrane to the hydrogen side of the fuel cell and are reduced to H 2 , While no precious metals are needed to catalyze the bromine reactions, the hydrogen oxidation and evolution (HOR/HER) reactions require a catalyst that is highly active, to keep the performance high and the cost low, stable and durable in the highly corrosive HBr/Br 2 environment of the cell as required by the extended life of this application.During the operation of a H 2 -Br 2 fuel cell, HBr and Br 2 could cross from the bromine side through the proton conducting membrane to the hydrogen side potentially leading to the corrosion and poisoning of the catalyst used at the hydrogen electrode.7] Consequently, alternative HER/HOR catalyst materials are needed for this system.2][13][14][15][16][17][18] In the particular case of hydrochloric acid electrolysis, Pt, known as the most effective catalyst for non-chloride systems, has been shown to have poor stability in the highly corrosive conditions of hydrochloric acid and be constantly poisoned by chloride ions.In contrast, rhodium sulfides are relatively unaffected by chloride ions.More recently, our group has found 6 that rhodium sulfides could be a promising candidate for hydrogen catalysis due to its high stability in HBr / Br 2 environment.] In this work, rhodium sulfide catalysts were synthesized on carbon supports and characterized by transmission electron microscopy (TEM) and X-ray diffraction (XRD), X-ray Photoelectron spectroscopy (XPS) and Energy-dispersive X-ray spectroscopy (EDX) techniques.Cyclic voltammetry (CV) was used for electrochemical characterization and electrochemically active surface area (ECSA) measurement.The HER/HOR activities were investigated in H 2 SO 4 and HBr solutions by multi-potential step method.

Experimental
Catalyst synthesis.-Thesynthesis procedure for rhodium sulfide precursor on carbon support (Rh 2 S 3 /C, Rh:C weight ratio 1:3) catalyst can be found in the patent literature. 21The carbon support used in this study was Vulcan XC-72.This electrochemically inactive precursor was thermally treated in a quartz oven at different definite temperatures in the presence of argon to make active phases (Rh 3 S 4 , Rh 17 S 15 and Rh) containing catalysts.Thermal reduction was carried out from room temperature (RT) to 600 • C, 700 • C, 800 • C and 850 • C and all catalysts were heated at final temperature for 30 minutes.The ramp rates were as follows: For 600 • C: RT to 550 -The thermogravimetric analysis (TGA) and differential thermal analysis (DTA) were performed by Perkin Elmer Diamond TG/DTA.The morphology and chemical compositions of the samples were determined by FEI Tecnai F20 XT Field Emission Transmission Electron Microscope (TEM) coupled with an energy-dispersive X-ray spectrometer (EDX) at an acceleration voltage of 200 kV.XRD data were collected on an X'Pert powder diffractometer (PANalytical, Inc.) with a Cu Kα source (corresponding to a photon wavelength of 1.54 Å).In XPS measurements, the prepared catalysts were analyzed by a Perkin−Elmer PHI 5400 electron spectrometer with an achromatic X-ray source (Al Kα (hυ = 1486.6eV)) anode.The XPS binding energies were measured with a precision of 0.1 eV.The analyzer pass energy was set to 17.9 eV, and the contact time was 50 ms.The spectrometer was calibrated by setting the binding energies of Au 4f 7/2 and Cu 2p 3/2 to 84.0 and 932.7 eV, respectively.For each measurement, all XPS spectra were corrected to the internal reference spectra of C 1s at 284.5 eV to compensate for electrostatic charging.Then, the core-level spectra were carefully deconvolved using the CasaXPS software to investigate the chemical states of elements on the electrocatalyst surface.
Electrode preparation.-Thecatalyst ink was prepared by ultrasonically dispersing 1.6 mg catalysts in 2.0 mL isopropyl alcohol (IPA) and the dispersion was then ultrasonicated for 30 min.Glassy carbon (GC) rotating disk electrodes, RDEs, with 5 mm diameter were polished with 0.05 μm alumina suspensions to a mirror finish before each experiment and served as an underlying conductive substrate of the working electrode.A quantity of 20 μL of the dispersion was pipetted out on the top of the GC.The catalyst layer was dried at room temperature.Then, an aliquot of Nafion solution (10 μl of 1 mg/mL solution in 50% IPA in water) was applied onto catalyst layer to yield an average film thickness of 0.25μm.The Nafion-coated working electrode was dried at room temperature and finally heated at 130 • C for 30 min. in air in an oven.
Electrochemical measurements.-Allelectrochemical measurements were made with a Gamry Instrument (G300 Potentiostat/ Galvanostat/ZRA).Electrochemical measurements were carried out in a conventional three-electrode electrochemical cell at 25 • C, with the glassy carbon disk electrode made in the above mentioned procedure as the working electrode and a saturated calomel electrode (SCE) as the reference electrode.A piece of Pt foil (1 cm 2 ) was placed in a separate chamber to serve as the counter electrode, which was connected to the working electrode chamber through a KCl salt bridge.All solutions were prepared with ultrapure water (Millipore, 18.25 M • cm).The solution of 1M H 2 SO 4 was purged with N 2 gas for nearly 30 min.before starting the cyclic voltammograms.Before the electrochemical measurements, the working electrode was soaked in de-ionized (DI) water for 2 hrs to ensure that the Nafion binder is well hydrated.Fresh electrolyte solution was used for each electrochemical measurement to ensure reproducible results.All the potentials reported herein were with respect to the reversible hydrogen electrode (RHE) potential.Cyclic voltammograms were measured in N 2 -saturated 1 M H 2 SO 4 solution from 0 to 0.65 V vs. RHE at 20 mV s −1 to avoid oxidizing the materials at high potentials, which has been found to lead to inconsistent ECSA measurements.In order to prevent H 2 bubble accumulation at electrode surface, the RDE was tilted about 5 to 10 degrees and the electrode was rotated at 10 rpm.Since ECSA measured during the anodic scan is related to the charge associated with the oxidation of adsorbed hydrogen atoms on the catalyst surface only, any small variation in the dynamics of the electrolyte from a slightly tilted electrode rotating at 10 rpm is not expected to affect this measurement.
N 2 -gas was passed through the electrolyte solution continuously during the CV scan.The "relative" electrochemical active specific surface area (ECSA) of catalyst "with respect to platinum" was calculated from the coulombic charges measured during the hydrogen oxidation step after correcting for double-layer charging current using the following equation: where ECSA is the electrochemical active surface area in cm 2 , Q H (mC) is the charge from the oxidation of adsorbed hydrogen in the hydrogen region of the CVs, 0.21 mC/cm 2 (which is also reported for other non-Pt catalysts [22][23] ) is the electrical charge associated with monolayer adsorption of hydrogen on Pt.Finally, hydrogen oxidation and hydrogen evolution (HOR/HER) was carried out at 2500 rpm in H 2 -saturated 1M H 2 SO 4 and 1M HBr solution by using multiplestep chronoamperometry method.The electrode was also tilted about 5-10 degrees in these measurements to prevent H 2 bubble accumulation at electrode surface.At these rotation speeds, the effect of the gravitational force, especially with the slight change in this force by tilting and for this small angle and electrode radius, on the fluid dynamics is minimal.Furthermore, only the data in the kinetically controlled region were used to determine the catalyst activity.IR compensation was done by measuring the resistance between the working electrode and reference electrode by electrochemical impedance spectroscopy.

Results and Discussion
Physical characterizations.-Thermalanalysis measurement of the synthesized precursor (Rh 2 S 3 /C) was carried out using thermogravimetric analysis (TGA) and differential thermal analysis (DTA).The catalyst was heated from room temperature to 850 • C with an increment of 5 • C/min in Ar.Fig. 1 shows a plot of TGA/DTA.Notably, the TGA curve can be divided into three weight-loss steps.The first step (about 5% weight loss) from room temperature to about 400 • C was due to the desorption of absorbed water molecules on the sample.The second step between 400 and 570 • C shows a ∼20% weight loss, which might be due to the release of excess sulfur because of its high volatility (boiling point at 444.6 • C), as 20% excess sulfur was used during the synthesis of the precursor.Finally, weight loss from 570 • C to 850 • C was due to the phase transformation of the Rh 2 S 3 precursor.This phase transformation is supported by XRD results which are discussed later.Within the DTA data one exothermic peak is identified at 495 • C which might be assigned to the crystallization of Rh 2 S 3 .Similar TGA/DTA results were observed by Zhang et al. 24 during the hydrothermal synthesis of rhodium sulfides.It is clear that temperature is a key factor in controlling the phase composition of rhodium sulfide in the catalyst.To obtain the active phase composition of rhodium sulfides (Rh 3 S 4 , Rh 17 S 15 and Rh), the precursor was treated from 600 • C to 850 • C.
The XRD of the synthesized catalysts at different temperatures shown in Fig. 2 2), it is difficult to index them.These results indicate that the Rh 2 S 3 material was transformed to other phases by loss of sulfur during thermal reduction.Similar thermal treatment phenomenon for rhodium sulfide catalyst was also reported. 16,24he average catalyst particle size was calculated from the XRD diffraction peak width using the Scherer equation: 25 where L is the particle size, λ is the X-ray wavelength in nanometer (0.15418 nm), β is the peak width of the diffraction peak profile at half maximum height in radians and K is a constant, normally taken as 0.9.The value of β in 2θ axis of diffraction profile must be in radians.We were unable to calculate the average particle size for the poorly crystallized Rh 2 S 3 precursor phase.The particle sizes for both the 600 • C and 700 • C catalysts are 8.7 nm, 9.9 nm for the 800 • C catalyst and 19.9 nm for the 850 • C catalyst, respectively.The increase in particle size with temperature might be attributed to the migration and agglomeration of catalyst at higher temperature.To avoid the formation of elemental Rh, we limited the catalyst synthesis temperature to 800 • C for the thermal reduction in the rest of study.
The morphology and distribution of the catalysts were evaluated by TEM.Fig. 3 shows TEM images of the precursor (Rh 2 S 3 /C) and thermally reduced catalysts.The carbon particles are the 40 nm objects, with the Rh x S y as the darker particles.Well-defined crystallites and particle size distribution were not observed for untreated Rh 2 S 3 /C precursor.After thermal treatment of the precursor, consistent particle size distribution was observed (at 600 • C, 700 • C and 800 • C).The particle size was measured directly from the TEM images at different randomly selected regions of each sample.Histograms of the particle size distributions of Rh x S y catalysts are included as insets in Fig 3 .Similar particle size distribution on the order of 3-12 nm (mean of 6 nm) of all catalysts was found on carbon support, some agglomeration of individual crystals was noted.These catalyst particle sizes are in good agreement with the XRD results.
EDX was used to measure the elemental compositions and the rhodium to sulfur atomic ratios of these catalysts, and the results are shown in Figure 4.The Rh:S atomic ratio for the precursor was about 0.57 which is slightly lower than the theoretical atomic ratio of Rh 2 S 3 (0.67) indicating the presence of some elemental sulfur in the catalyst.The Rh:S ratio was found to increase upon thermal treatment.It is noteworthy to point out that Rh:S ratios greater than 1.13, such as that synthesized at 800 • C (1.21), are due to the formation of elemental Rh.The electronic states of Rh in Rh x S y /C catalysts were studied by XPS, and the results are shown in Fig. 5. Figure 5a shows that the binding energies of 3d levels of Rh in different catalysts were shifted to higher values (i.e., 309.2 [3d 5/2 ] and 314.2 [3d 3/2 ] eV for the precursor; 308.9 [3d 5/2 ] and 313.9 [3d 3/2 ] eV for 600 • C; 308.5 [3d 5/2 ] and 313.0 [3d 3/2 ] eV for 700 • C and 307.8 [3d 5/2 ] and 312.9 [3d 3/2 ] eV for catalyst 800 • C) in comparison with those (307 [3d 5/2 ] and 312 [3d 3/2 ] eV) of elemental Rh.This indicates that Rh existed in a cationic state rather than in an elemental state in these catalysts.Figure 5b shows the XPS spectra of O1s.The peak at 532 eV is due to the presence of oxygen-containing groups on the carbon surface. 26Oxidic rhodium (at ca.530 eV 27 ) is not detected in any of these catalysts, thus ruling out the formation of rhodium oxide.As shown in Fig. 5c, the contributions at ca. 162.1 eV at the lower binding energy portion of the S 2p peak originate from S δ− present in sulfide [28][29] and are clearly visible in all samples.In the precursor, the shoulder at higher binding energy at around 164 eV implies the existence of elemental sulfur, in accord with the EDX results.This is an indication that excess S higher than the stoichiometric ratio of 1.50 in the composite exists as elemental sulfur.
In order to investigate the different phases within the samples, Rh3d spectra are carefully deconvolved and presented in Fig. 6.The binding energy observed for Rh3d in the precursor (Fig. 6a) shows a doublet at 309.2 eV and 314.2 eV, corresponding to 3d 5/2 and 3d 3/2 states, respectively.Normally, the peak at 309.2 eV arises from covalently bound oxidized Rh (sulfide and/or oxide, but we can exclude the possibility of rhodium oxide as no O 1s peak is found at 530 eV).Combined with S 2P analysis, these peaks are attributed to covalently bonded, sulfurized Rh (Rh 2 S 3 ) which is also evident from XRD result.However, it is noted the Rh 2 S 3 can be transformed to Rh 3 S 4 , Rh 17 S 15 and Rh by loss of sulfur during thermal reduction in presence of Ar (from XRD results).Deconvolution spectrum of Rh3d 5/2 in Fig. 6b for 600 • C shows the presence of a small amount of Rh 17 S 15 phase (ca.308.2 eV) whereas the main contributor of the spectrum is Rh 2 S 3 (ca.309.2 eV).Similar deconvolved spectra is observed for the catalyst at 700 • C (Fig. 6c) where the main peaks are due to Rh 2 S 3 (ca.309.2 eV) with Rh 17 S 15 (ca.308.2 eV) and Rh 3 S 4 (ca.308.7 eV).As demonstrated in Fig. 6d for the 800 • C sample, the spectrum of the lowest binding energy of Rh 3d 5/2 is 1.2 eV lower than that of Rh 2 S 3 .This shift confirms that at relatively high temperature, Rh 2 S 3 is transformed to other phases, mainly Rh 17 S 15 (ca.308.2 eV) with a small amount of Rh 3 S 4 (ca.308.7 eV) and Rh (ca.307.3 eV).However, a small peak at ca. 309.2 eV, which corresponds to Rh 2 S 3 , is also observed indicating that some Rh 2 S 3 remained.The different phase compositions of the catalysts are calculated from the area of deconvolved peaks and summarized in Table I.Cyclic voltammograms.-Previousstudies 16,[30][31] had reported cyclic voltammograms (CV) in N 2 -saturated H 2 SO 4 solution of Rh x S y /C from 0 to 1.2 V (vs.RHE) and that electrochemical active surface area (ECSA) measurements by CV had been problematic.
Similarly, we have found inconsistent relative ECSA, with respect to polycrystalline Pt surface, when CV was scanned at this potential range.To resolve this issue, we have measured the CVs in N 2saturated 1M H 2 SO 4 solution of Rh x S y /C at different anodic potential limit to get consistent and reproducible ECSA.A family of cyclic voltammograms of Rh x S y /C in 1M H 2 SO 4 is shown in Fig. 7a.The anodic limit was increased progressively on each sweep after the electrode was cleaned from 0 to 1.0 V (vs.RHE) for 2 cycles at a scan rate of 100 mV s −1 at the beginning of each sweep.After the final anodic limit, an additional CV scan was conducted from 0 to 0.6 V (dotted line) to check the reproducibility of results.Electrochemical active surface area (relative to Pt) of each sweep was calculated and summarized in the inset of Fig. 7b.Interestingly, the ECSA was found to decrease as the potential window was extended in the positive direction.This phenomenon might be due to the formation  and adsorption of OH ions on Rh sites of the catalyst and formation of rhodium hydrous oxide species which could not be reduced during the cathodic scan.The formation of these species can suppress the hydrogen adsorption/desorption on the electrode surface.Furthermore, the double-layer charging current was found to increase with increasing anodic potential limit which could indicate the incomplete reduction of adsorbed oxides species at the Rh x S y surface.Thus, in this study, we have performed the CV scan from 0 to 0.65 V (vs.RHE) to avoid the interference of oxide/hydroxide formation and reduction that lead to inconsistent ECSA.ception of the precursor, the catalysts showed a similar behavior with well-defined hydrogen adsorption/desorption peaks.No peak corresponding to hydrogen adsorption/desorption was observed for the precursor.Relative ECSA with respect to Pt surface was calculated and the highest ECSA was obtained for 800 • C, in agreement with the XRD/XPS results which showed that the 800 • C contained the highest composition of active phases (Rh 3 S 4 , Rh 17 S 15 , Rh) active phase composition.ECSAs of these catalysts are presented in Table II.
HOR/HER activity.-Theelectrode current for the HOR and HER reactions can approximately be represented by a Butler-Volmer relation with no mass transport resistance as follows: where i represents the measured current density in A per cm 2 of geometric area, a is the catalyst area per geometric area, i o is the exchange current density in A per cm 2 of catalyst area, α a and α c are the anodic and cathodic transfer coefficients, F is Faraday's constant, R is the universal gas constant, T is temperature, and η = E-E eq is the overpotential.The exchange current density (i o ) is an important kinetic parameter representing the catalytic activity of a catalyst.The Butler-Volmer equation above can be simplified when the value of η is sufficiently small, (|η| < 25 mV): From the slope of I versus η curve, i o can be calculated as shown in equation 4.
The evaluation of the experimental current I vs. η for the hydrogen oxidation and hydrogen evolution reaction on Rh x S y /C electrode was carried out in 1M H 2 SO 4 as well as 1M HBr solution under hydrogen atmosphere at rotation rate of 2500 rpm.After the open circuit voltage (OCV) reached the equilibrium value (0.0 ± 0.0002 V vs. RHE), multi potential steps (each step was for 10 s) were applied in increments of 0.002 V from OCV to ±0.016 V. Chronoamperometric response of the electrode in 1M H 2 SO 4 is shown in Fig. 9.It can be observed that the current quickly reached steady state, as expected at these low current densities.Readings of the current value were made every 1 s and the mean value of the current data measured in the last 5 s  of each potential step was used to create the current density versus overpotential plots shown in Figures 10 and 11 for 1M H 2 SO 4 and 1M HBr, respectively.The exchange current densities of the catalysts were obtained from the slopes of the I vs. η lines and summarized in Table III.Note that conversion of the current to the current density is not necessary because i and a in Equations 3 and 4 are both referenced to the same geometric area.
Exchange current density (i o ) per active area is an intrinsic property of the catalyst.In this case where multiple active phases are present, the measured exchange current densities represent the average values of all the active phases.The fact that the exchange current densities are almost the same says that the active phase is probably the same for all three samples.It might not be certain that it is Rh 17 S 15 , but the fact that the ECSA is increasing with Rh 17 S 15 indicates that Rh 17 S 15 is likely to be contributing to activity.Catalyst synthesized at 800 • C gave the highest mass specific exchange current density as it was composed of the maximum fraction of active phases.The i o in 1M HBr for all catalysts were found to be lower those obtained in 1M H 2 SO 4 solution.For Pt, 7,32 this is known to be due to the adsorption of Br − on the active surface, so it is possible a similar effect is seen here.The slightly higher exchange current density for the • could be the formation of the higher activity Rh phase, which exists at 15% in this sample.
It has been reported 19 that the bifunctional HOR/HER activities of supported Rh x S y were due to undefined synergistic effect of the conductive Rh 3 S 4 and Rh 17 S 15 phases in the catalyst (Rh 2 S 3 is a semiconductor).Similarly, it is evident from XPS results of the cata-  and Rh) increases with increasing pretreatment temperature, from 600 • C to 800 • C, the bifunctional HOR/HER activity was also found to increase accordingly.

Conclusions
We have successfully synthesized an active Rh x S y catalyst on carbon support with particle sizes between 3 and 12 nm.Thermal reduction of the inactive Rh 2 S 3 phase created active phases of Rh 3 S 4 , Rh 17 S 15 and Rh.Among the catalysts prepared, the catalyst synthesized at 800 • C exhibited the highest ECSA as it contained the highest fraction of the active phases as confirmed by XRD and XPS analyses.The ECSA of these rhodium sulfide catalysts was found to decrease as the upper potential scan limit was extended beyond 0.65 V.This phenomenon might be attributed to the formation of rhodium hydrous oxide species which could not be reduced during the cathodic scan.The HOR/HER activity of the catalysts in H 2 SO 4 and HBr solution were quantified as the exchange current density (i 0 ) per active area.The specific exchange current densities (i 0 ) in 1M H 2 SO 4 (0.89 to 1.04 mA/cm 2 ) of Rh x S y /C catalysts are found to be comparable with Pt catalyst (1.0 to 1.2 mA / cm 2 ).The HER/HOR exchange current density in 1M HBr solution for the catalysts synthesized in the study are 0.81 ± 0.01 mA/cm 2 , 0.81 ± 0.02 mA/cm 2 , and 0.92 ± 0.03 mA/cm 2 for 600 • C, 700 • C and 800 • C catalysts, respectively.The HER/HOR exchange current density of these catalysts may be lower in HBr than H 2 SO 4 due to the adsorption of Br − on the active surface.The highest mass specific activity i o * was obtained with the catalyst synthesized at 800 • C, due to the high active phase composition.

Figure 1 . 15 Figure 2 .
Figure 1.TGA-DTA curves of precursor (Rh 2 S 3 /C).Scan rate of 5 o /min. ) unless CC License in place (see abstract).ecsdl.org/site/terms_useaddress.Redistribution subject to ECS terms of use (see 129.237.45.148Downloaded on 2016-04-01 to IP confirmed the presence of different phases of rhodium sulfides, such as Rh 2 S 3 (ICSD No.15344), Rh 3 S 4 (ICSD No. 410813), Rh 17 S 15 (ICSD No. 410838) and Rh (ICSD No.426969).The result shows that the precursor was poorly crystallized and potentially composed of Rh 2 S 3 phase.However, the catalysts were crystallized after thermal treatment, and the crystallized phases of the products varied with the heat-treatment temperature.For the catalysts synthesized at 600 • C and 700 • C, mixtures of Rh 2 S 3 and Rh 17 S 15 phases were obtained.The catalyst synthesized at 700 • C contained more of the Rh 17 S 15 phase than was observed in the 600 • C synthesized catalyst.Synthesis at 800 • C resulted in samples composed of mostly the Rh 17 S 15 phase with a small amount of Rh 2 S 3 and Rh phases.It is noted that the diffraction spectra between 31 • to 42 • clearly mark the fingerprint of Rh 17 S 15 phase.The XRD pattern of the sample at 850 • C was dominated by elemental Rh phase.As the XRD spectrum of Rh 3 S 4 is almost superimposed with the spectrum of Rh 17 S 15 (see the reference spectrum in Fig.

Figure 3 .Figure 4 .
Figure 3. TEM images of catalyst precursor, Rh 2 S 3 /C and after thermal treatment at 600 • C, 700 • C, and 800 • C. The corresponding histograms of particle size distributions were plotted in the insets.

Figure 7 .
Figure 7. (a) CVs of Rh x S y /C (800 • C) electrode in N 2 -saturated 1M H 2 SO 4 solution with different anodic potential limits.Potential scan rate: 0.02 V s −1 .(b) Enlarged CVs of Fig. 7a (dotted rectangular) and inset is the plot of ECSAs calculated at different anodic limits.

Figure 10 .
Figure 10.Linear plots of current density versus applied potentials of synthesized catalysts in H 2 -saturated 1M H 2 SO 4 solution.

Figure 11 .
Figure 11.Linear plots of current density versus applied potentials of synthesized catalysts in H 2 -saturated 1M HBr solution.