Self-stabilized Pt–Rh bimetallic nanoclusters as durable electrocatalysts for dioxygen reduction in PEM fuel cells

Abstract

Self-stabilized Pt–Rh nanoclusters (NCs) were prepared by using a surfactant-free chemical reduction method with formic acid as the reducing agent. The elemental composition was determined by EDX analysis. The synthesized cluster was used as a supportless (SL) electrocatalyst for the reduction of oxygen (ORR) in acid medium. The composition of Pt–Rh bimetal NCs, in terms of atomic weight percentage, was optimized based on the available electrochemical surface area. Hydrodynamic linear scan voltammetric profiles show that the onset potential for oxygen reduction is 0.78 V vs. RHE at the electrode rotation rate of 2400 rpm with 17.8 μg cm⁻² loading of the SL Pt₃Rh exhibiting the limiting current density of 3.5 mA cm⁻². The durability of the electrocatalysts was investigated by performing the accelerated durability test (ADT): the electrochemical surface area (ECSA) for SL Pt₃Rh increased by nearly 9.2% while retaining nearly 85% of its initial limiting current density after 15 000 potential cycles. For comparison Vulcan-carbon-supported Pt₃Rh was synthesized under identical conditions and subjected to electrochemical investigations. Both supportless and VC-supported Pt₃Rh NC electrocatalysts were found to use a direct 4-electron transfer mechanism. In order to improve the activity, SL Pt@Pt₃Rh NC was synthesized and used as the catalyst. At 0.9 V, the mass activity (0.085 mA μg⁻¹) of the Pt@Pt₃Rh NC was found to be nearly 34 times greater than that of SL Pt₃Rh NC (0.0025 mA μg⁻¹). We conclude that the SL Pt₃Rh NC could potentially be used as an electrocatalyst for ORR in a sulfuric acid medium since it possesses good stability compared to Pt-based ORR catalysts reported in the literature.

---

1. Introduction

The state of the art for the design and production of electrocatalysts has involved the use of Pt and Pt-based nanostructures for polymer electrolyte membrane fuel cells (PEMFCs) to catalyze the cathodic reduction of molecular oxygen (ORR).¹ The inherently sluggish kinetics of ORR is driving the search for active and durable cathode electrocatalysts for PEMFCs. During the initial development of fuel cells, bulk Pt was employed as the electrocatalyst, but it suffered from low surface area and utilization. Later on, nanoelectrocatalysts garnered more attention due to their unique physical and chemical properties. The most widely employed electrocatalyst is Pt nanoparticle supported on carbon black (Pt/C).^2,3 Therefore, when Pt/C is used as a catalyst in PEMFCs, an enhanced activity is observed due to the enhancement in the electrochemical surface area (ECSA); in addition, increased mass and specific activities were obtained. This led to the development of commercial fuel cell stack systems and prototype models. In industrially used PEMFCs, employment of different nanomorphologies helped to reduce cathode catalyst (Pt) loading in membrane electrode assemblies from 0.4 mg cm⁻² to 0.1 mg cm⁻²;⁴ this reduction achieves the DOE target, which aims to reduce the exorbitant cost of fuel cell components. But the inherent disadvantage of real stack systems is the reduction in performance after several hours of operation. This reduction occurs because the carbon surface undergoes oxidation and results in Pt nanoparticle agglomeration and its leaching out from the surface, which in turn reduces the oxygen reduction kinetics and shortens the expected life span of the catalyst layer.^4–6 Although not cost effective, Pt is highly preferred due to its high catalytic activity and resistance to oxidation in acid medium compared to most of the noble and non-noble metals.^7–9 However, Pt is not the most stable of the newer catalyst materials.

In order to improve the stability and to avoid carbon corrosion issues, supportless bimetal nanostructures have been increasingly employed as durable electrocatalysts by many researchers. Sun et al. have reported the synthesis of highly durable multiarmed star-like Pt nanowires as a supportless electrocatalyst that retains ∼90% of its initial ECSA after 4000 potential cycles.⁷ Tan et al. studied the durability of Au/Pt and Au/Pt₃Ni as self-supported electrocatalysts and observed 6.8% and 9.9% ECSA loss respectively after 5000 potential cycles.⁸ Wang et al. described a Pt nanoparticle netlike-assembly, synthesized using a surfactant-assisted hydrothermal method that exhibited high durability and only 9% loss in ECSA after 20 000 potential cycles.⁹ Huang et al. reported the synthesis of highly porous Pt₃Ni nanocrystals using a PVP-assisted method and found 27.3% loss of ECSA after 6000 potential cycles.¹⁰ Li et al. prepared nanotubes consisting of mixed PtPdCu alloy nanoparticles using a Cu nanowire template-assisted method, and these nanotubes showed a nearly 20% loss of ECSA after 30 000 potential cycles with a negative shift in the half-wave potential of 5 mV.¹¹ Different successful strategies using bimetal electrocatalysts with enhanced stabilities were witnessed for ORR by a novel combination of Pt with the uncommon elements such as Rh, Re, and Os.^12–14

Nearly three decades ago, bulk PtRh alloy was employed as a binary electrocatalyst for oxidation of small alcohols and was found to exhibit better catalytic activity than pure Pt.^15–17 Although Rh by itself is not catalytically active, there have been several reports about the influence of Rh nanoparticles on the catalytic activity of Pt for the breaking of C–C and C–H bonds during the oxidation of small organic molecules, which would be useful for fuel cell applications.^18–21 Park et al. used a single-cell system to study the catalytic activity of PtRh and PtRuRhNi alloy nanoparticles for the methanol oxidation reaction (MOR), and concluded that the presence of Rh may influence the regeneration of an active Pt surface; note, however, that they investigated the ORR activity without much attention to stability.^22,23 Noto et al. reported the preparation of carbon-nitride-supported PtRh by pyrolysis and discussed the ORR activity in acid medium.^12,24 Friebel et al. characterized the stability of PtRh (111) nanoparticles using in situ X-ray absorption spectroscopy (XAS) for ORR.²⁵ More recently, Baraldi et al. theoretically predicted that bimetallic PtRh (111) nanoclusters could enhance the chemical reactivity of the catalyst in fuel cell reactions.²⁶ Yuge et al. analyzed the electronic structure and stability of PtRh nanoparticles using first principles calculations and predicted that PtRh alloy nanoparticles could function as a potential electrocatalyst for PEM fuel cells based on the observation of a downshift of the center of the Pt d-band in a PtRh alloy.²⁷

The objectives of the present investigation were to synthesize the PtRh nanoclusters and to evaluate the electrocatalytic activity and stability for the oxygen reduction reaction in acid medium by cyclic voltammetry in a standard three-electrode electrochemical cell. The supportless Pt–Rh nanocluster electrocatalyst was prepared by the chemical reduction method using formic acid without using any additional template, and was unambiguously characterized by X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), high-resolution transmission electron microscopy (HRTEM) and energy dispersive X-ray analysis (EDX) to determine its size, shape, distribution and composition. The Pt–Rh composition (at. wt%) was optimized based on the electrochemical activity towards oxygen reduction. For comparison, a Vulcan-carbon-(VC)-supported electrocatalyst was synthesized under identical conditions and used for electrochemical investigations. The kinetic and thermodynamic parameters for ORR were calculated for both supportless and carbon-supported Pt–Rh nanoclusters and the stabilities were evaluated under continuous potential cycling.

2. Experimental methods

2.1. Materials

Hexachloroplatinic acid (H₂PtCl₆·6H₂O) and rhodium(III) chloride monohydrate (RhCl₃·H₂O) were obtained as Pt and Rh metal precursors respectively from Sigma-Aldrich. Formic acid (98%, Sigma) was used as a reducing agent. For carbon-supported catalysts, Vulcan carbon XC-72 (VC) was used as received from Cabot (I) Ltd. as a complementary pack. Nafion perfluorinated polymer resin solution (5 wt%, Sigma-Aldrich) was applied as the catalyst binder, and sulfuric acid (Rankem), methanol (98%, Merck) and absolute ethanol (Merck) were all used as received. Millipore (18 MΩ cm) water was used to prepare all solutions.

2.2. Preparation of Pt–Rh nanoclusters

Supportless Pt–Rh nanoclusters (SL Pt–Rh NC) were synthesized by slow reduction according to the following procedure. Briefly, 20 mL of an aqueous solution containing 24 mg of H₂PtCl₆·6H₂O and 6 mg of RhCl₃·H₂O (3 : 1 Pt–Rh atomic wt%) was placed in a beaker to which 1 mL of formic acid was added. The resulting mixture was kept aside undisturbed for 72 h at room temperature. The mixture was then centrifuged at 8000 rpm and washed with water and methanol several times. Finally, the product was collected and dried in an air oven at 80 °C. Pt and Rh were used at the ratios 3 : 1, 1 : 1, and 1 : 3. The carbon-supported catalysts were prepared by adding 7.5 mg of VC at the initial stage of the above synthesis procedure before adding the reducing agent. Also, the supportless Pt@Pt₃Rh NC was synthesized by adding Pt precursor over the previously synthesized Pt–Rh (3 : 1) NC and further reducing with formic acid for another 72 h to obtain Pt skin over Pt₃Rh.

2.3. Characterization techniques

The as-prepared supportless and carbon-supported Pt–Rh NC electrocatalysts were characterized by XRD (Bruker-D8 diffractometer using Cu-K_α radiation, λ = 1.54 Å, current: 30 mA and voltage: 40 kV), FESEM, and EDX mapping (FEI instruments, Nova-Nano SEM-600, Netherlands). TEM and HRTEM images were obtained by re-dispersing the catalyst sample in absolute ethanol by sonication and drop casting on a carbon-coated copper grid (FEI Tecnai 30G2, 300 kV).

Typically, the catalyst ink was prepared by dispersing 0.5 mg of catalyst in 1 mL of absolute ethanol, and ultrasonicating for 3 min to obtain a homogeneous dispersion. Then, 7 L of catalyst ink was pipetted out onto a mirror-finished surface of the rotating-disc glassy carbon working electrode (GC-RDE, 5 mm) to obtain a metal loading of about 17.8 μg cm⁻² for both supportless and supported electrocatalysts, which was subsequently covered by 5 μL of 0.05 wt% Nafion. The indicated metal loading of each catalyst is actually the sum of Pt and Rh. Prior to coating, the GC surface was polished with a 0.05 μm alumina slurry, washed with ethanol and water and then subjected to ultrasonic agitation for 5 min in deionized water. For preliminary cyclic voltammetric investigations, a catalyst loading of 40.8 μg cm⁻² (15 μL) was maintained for all the compositions of Pt–Rh.

The electrochemical behavior of the prepared Pt–Rh nanoclusters was investigated by CV and LSV techniques. The cyclic voltammograms were recorded in N₂-saturated 0.5 M H₂SO₄ at a sweep rate of 0.1 V s⁻¹ between −0.1 to 1.2 V vs. RHE at 25 °C. The ORR polarization curves (LSV) were obtained in an O₂-saturated 0.5 M sulfuric acid solution at a sweep rate of 0.01 V s⁻¹ and at the electrode rotation rates of 400, 800, 1200, 1600, 2000 and 2400 rpm using a rotator set up (Pine) interfaced with the potentiostat and controlled through NOVA 1.9 software (Autolab). The standard model of a double-jacketed electrochemical cell was employed for a three-electrode configuration. The electrolyte was maintained at the required temperature by circulating water in the outer jacket using a thermostat (Equibath, India). The plain Pt sheet (1 cm²) was employed as a counter electrode and a double-junction Ag/AgCl-filled (saturated KNO₃) was used as the reference electrode. All the potential values are reported to the reversible hydrogen electrode (RHE) by adding 0.21 V to the obtained potential values.²⁸ The long term stability of Pt–Rh electrocatalysts was checked by the accelerated durability test (ADT), i.e., by applying continuous potential cycles to the working electrode between 0.6 and 1.2 V in an O₂-saturated 0.5 M H₂SO₄ solution at 25 °C, at a scan rate of 0.1 V s⁻¹, and recording both CV and LSV patterns during the experiment every 1000 cycles.

3. Results and discussion

3.1. XRD, SEM and TEM results

The powder XRD patterns of both supportless and VC-supported Pt₃Rh NC are shown in Fig. 1. The XRD patterns of both catalysts show a face-centered cubic structure with the characteristic peaks at the diffraction angles (2θ): 40.4°, 47.0°, 68.7° and 82.1° corresponding to the (111), (200), (220) and (311) crystalline planes, respectively. The 2θ of the (111) peak for both SL Pt₃Rh (Fig. 1a) and Pt₃Rh/VC (Fig. 1b) is slightly shifted, by 0.6°, to a higher angle compared to pure Pt (2θ = 39.8°). The same trend was also observed for the 2θ values of the (200), (220) and (311) peaks. The shift to a higher angle confirms the formation of an alloy between Pt and Rh (i.e., incorporation of Rh atoms into the fcc lattice of Pt),²⁹ in both SL Pt₃Rh (Fig. 1a) and Pt₃Rh/VC (Fig. 1b) catalysts. In the VC-supported catalyst an additional peak observed at 2θ = 23.8° is from the hexagonal plane of the conducting carbon substrate.³⁰ The average particulate size of the nanocrystals was estimated from the XRD pattern using the Debye–Scherrer equation and found to be ∼4 nm for both catalysts.

Fig. 1 Powder XRD patterns of (a) supportless and (b) carbon-supported Pt₃Rh nanoclusters.

The FESEM images of supportless Pt₃Rh nanoclusters (ESI, Fig. S1a and b†) and VC-supported Pt₃Rh (ESI, Fig. S1c and d†) show that the supportless catalyst possesses a cluster morphology composed of tiny nanoparticles, and the VC-supported catalyst is uniformly distributed. The EDX pattern of SL Pt₃Rh, which is shown in Fig. S2a of ESI,† indicates the atomic wt% composition of Pt and Rh to be 76.9 and 23, respectively, with the expected stoichiometric ratio of nearly 3 : 1. Also, the elemental mapping (ESI, Fig. S2b–d†) clearly depicts the homogenous distribution of Pt and Rh throughout the structure of supportless Pt₃Rh. The TEM images of the supportless electrocatalyst (Fig. 2a and b) confirm the assembly of smaller nanoparticles (less than 5 nm). The fine lattice fringes of nanoclusters (Fig. 2c) confirm the crystalline nature of the Pt₃Rh bimetal nanoclusters. From the histogram (Fig. 2d) of the TEM image, the average particle size was found to be 3.3 nm with the standard deviation of around 0.74 nm, which correlates well with the calculated average particle size from the XRD pattern.

Fig. 2 (a & b) TEM images and (c) HRTEM image of supportless Pt₃Rh nanocluster with (d) corresponding histogram.

3.2. Optimization of Pt–Rh ratio

Cyclic voltammograms of SL Pt–Rh NC with different atomic ratios of Pt and Rh, such as 3 : 1, 1 : 1 and 1 : 3, were recorded in N₂-saturated 0.5 M sulfuric acid between 0.0 and +1.2 V at 25 °C and are shown in ESI, Fig. S3a.† The ECSA of the metal nanocatalyst was determined by integrating the charges measured upon desorption/adsorption of hydrogen ‘under potential deposition (H_upd) region’ between −0.1 and 0.21 V. The electrochemically active surface areas of SL Pt–Rh NCs for all the three compositions were calculated and it was found that the SL Pt–Rh 3 : 1 catalyst has the highest ECSA of 154.4 m² g⁻¹ and this value is nearly 1.3 times higher than that for the 1 : 1 (85.2 m² g⁻¹) composition and 3 times higher than that for the 1 : 3 (43.1 m² g⁻¹) composition. This relatively high magnitude of ECSA for the SL Pt–Rh NC is comparable with that (103.5 m² g⁻¹) reported for a Pt–Ru/C bimetallic ORR electrocatalyst.³¹ Also, the ECSA of 3 : 1 SL Pt–Rh was determined to be higher than those reported for nanoporous 3D-PtRh alloy and PtRh/carbon nitride catalysts.^15,27,32 Generally the observed higher surface areas of SL Pt–Rh electrocatalysts can be attributed to the fine structure of the nanocluster morphology of Pt stabilized with Rh. From the CVs, a well-defined metal-oxide reduction peak at 0.52 mV is clearly observed for 3 : 1 SL Pt–Rh, and this value is more positive compared to the other two compositions.

The catalytic activity of SL Pt–Rh NC towards the dioxygen reduction was investigated by observing the reduction half-cell profiles in O₂-saturated 0.5 M sulphuric acid under hydrodynamic conditions. Linear scan voltammograms were recorded for all three compositions and are compared in ESI, Fig. S3b.† Compared to the other two compositions investigated, the 3 : 1 SL Pt–Rh NC possesses a higher limiting current density of 3.92 mA cm⁻²; this composition also displayed an earlier onset potential, with a 100 to 150 mV greater magnitude. The exchange current density, transfer coefficient and the onset potential values were obtained from the corresponding Tafel plots and are given in ESI, Table S1 (ESI, Fig. S4a†). The 3 : 1 SL Pt–Rh NC displayed a 40–200 mV positive shift of the half-wave potential (E_1/2) relative to those of the compositions with 1 : 1 and 1 : 3 ratios, indicating the good catalytic activity of 3 : 1 SL Pt–Rh nanocluster for oxygen reduction.

The number of electrons transferred (n) during oxygen reduction was calculated from the corresponding Koutecky–Levich (K–L) plots and found to be 4.0 for the 3 : 1 SL Pt–Rh composition (ESI, Fig. S4b†). The other two SL Pt–Rh compositions, i.e., 1 : 1 and 1 : 3, follows two electron transfer mechanism with ‘n’ values of 2.7 and 2.6, respectively. From the calculated ‘n’ values, the 3 : 1 SL Pt–Rh NC electrocatalyst follows the most favorable electron-transfer pathway and hence this optimized composition was used for further investigations. These observations are consistent with published reports about Pt–Rh that indicate that 20–30 atomic wt% of Rh in Pt could be an optimized composition for fuel cell applications.^12,14,18

3.3. Cyclic voltammograms of Pt₃Rh nanoclusters

Fig. 3 shows the CVs of SL Pt₃Rh and VC-supported Pt₃Rh nanoclusters recorded in N₂-saturated 0.5 M H₂SO₄ between 0.0 and +1.2 V at 25 °C. A sharp H-desorption peak was seen during the anodic sweep for both supportless and VC-supported Pt₃Rh NCs in the potential region between −0.1 and 0.18 V. The oxide formation peak in the anodic sweep indicates that both Pt–O and Rh–O(OH)₃ formed between 0.7 and 1.1 V for both catalysts, and these results are similar to those reported in the literature.¹⁸ The SL Pt₃Rh NC has an ECSA of 154.4 m² g⁻¹, which is nearly 2.5 times higher than that of Pt₃Rh/VC (63.11 m² g⁻¹), nearly 45 times higher than that of the Pt-black (Hispec-1000, JM) catalyst (5.9 m² g⁻¹),³³ and comparable to those of the reported Pt–M alloy electrocatalysts.^31,34 The larger ECSA of SL Pt₃Rh NC might be due to the greater availability of exposed active Pt sites, and the smaller magnitude of VC-supported Pt₃Rh catalyst might be due to the Pt active sites being partially buried into the carbon support and also due to the surface masking by CO (formed by oxidation of the carbon substrate) over Pt active sites during the positive potential sweep.³⁵

Fig. 3 Cyclic voltammograms of supportless Pt₃Rh and Pt₃Rh/VC in N₂-saturated 0.5 M H₂SO₄ at a scan rate of 0.1 V s⁻¹ at 25 °C.

3.4. Kinetics of ORR on Pt₃Rh nanoclusters

Fig. 4a shows the LSV of SL Pt₃Rh and Pt₃Rh/VC NCs recorded in O₂-saturated 0.5 M H₂SO₄ at 2400 rpm. It can be seen that the supportless Pt₃Rh has an onset potential of 0.78 V, which is nearly 40 mV higher than that of the VC-supported Pt₃Rh (0.74 V). This SL Pt₃Rh also displays a higher half-wave potential than does SL Pt₃Rh/VC (0.59 V vs. 0.52 V), which shows the improved activity of the supportless Pt₃Rh electrocatalyst. Note, however, that there is not much variation in the magnitude of the observed limiting current density (j_L) values at 0.2 V for SL Pt₃Rh (3.50 mA cm⁻²) and Pt₃Rh/VC NC (3.34 mA cm⁻²). In order to estimate the magnitude of diffusional resistance offered by the electrolyte for the movement of oxygen molecules towards the electrodes surface, the limiting current density values were measured at different electrode rotation rates. As the rotation rate was increased from 400 to 2400 rpm the current density values also increased from 1.94 to 3.50 mA cm⁻².

Fig. 4 (a) Linear scan voltammograms and (b) K–L plots of SL Pt₃Rh and Pt₃Rh/VC nanoclusters in O₂-saturated 0.5 M H₂SO₄ at a scan rate of 0.01 V s⁻¹ at 25 °C.

The number of electrons transferred (n) per oxygen molecule during the reduction of dioxygen was calculated from the corresponding K–L plots (j_L⁻¹ and ω^−1/2) (Fig. 4b). The slopes of these plots were found to be 11.3 for SL Pt₃Rh and 12.46 for Pt₃Rh/VC and the corresponding ‘n’ values were found to be 4.0 and 3.5 respectively. These values clearly show that both the supportless and VC-supported Pt₃Rh catalysts predominantly follow the direct 4-electron transfer mechanism during ORR. The kinetic current density for SL Pt₃Rh was calculated to be 0.08 mA cm⁻², which is nearly 7 times higher than that of Pt₃Rh/VC (0.01 mA cm⁻²). The mass activity (MA) at 0.9 V was calculated by dividing the current density by catalyst loading (17.8 μg cm⁻²) at 2400 rpm and was found to be 0.0025 mA μg⁻¹ for SL Pt₃Rh. This value is higher than that of the VC-supported Pt₃Rh, and comparable to those of some of the reported ORR-catalyzing Pt/C¹¹ and Pt nanostructures.³⁶

The mass-transfer-corrected Tafel plots were constructed for SL Pt₃Rh and Pt₃Rh/VC catalysts (ESI, Fig. S5†) and the corresponding Tafel slopes were found to be 115 and 109 mV per dec respectively. These values are similar to the 120 mV per dec value corresponding to the clean Pt surface as reported for unsupported Pt black,³⁷ suggesting that the surface properties are quite similar to that of the clean Pt surface. This result clearly indicates that the first electron transfer is the rate-limiting step during the reduction of dioxygen.^1,35 ORR kinetic parameters including half-wave potential (E_1/2), kinetic current density (j_k), kinetic rate constant (k), Tafel slope (b), electron transfer co-efficient (α), exchange current density (i₀) and mass activity were calculated at 0.9 V for both supportless and VC-supported Pt₃Rh NCs and are summarized in Table 1. It can be clearly seen that the SL Pt₃Rh possesses better oxygen reduction activity than Pt₃Rh/VC in acid medium.

Table 1 Summary of ORR kinetic parameters for supportless and VC-supported Pt₃Rh nanoclusters at 25 °C^a

Catalyst	jd (mA cm^−2)	On set potential (V)	E1/2 (V)	jk (mA cm^−2)	10^3k (cm s^−1)	b (mV per dec)	α	10^6i0 (A cm^−2)	n	MA (mA μg^−1)
a jd-limiting current density; E1/2-half-wave potential; jk-kinetic current density; b-Tafel slope; α-electron transfer coefficient; i0-exchange current density; n-number of electron transfer; MA-mass activity @ 0.9 V.
SL Pt3Rh	3.50	0.78	0.59	0.08	3.50	115	0.52	0.96	4.0	0.0025
Pt3Rh/VC	3.34	0.74	0.52	0.01	3.25	109	0.55	0.32	3.5	0.0001

---

3.5. Accelerated durability tests of Pt₃Rh nanoclusters

The electrochemical stability of supportless and VC-supported Pt₃Rh nanocluster electrocatalysts was investigated by ADT potential cycling. The catalyst-coated GC-RDE working electrode was subjected to continuous potential cycling between 0.6 and 1.4 V at 25 °C in O₂-saturated 0.5 M H₂SO₄. During the potential cycling, the CVs were recorded with N₂ purging and LSVs were recorded with O₂ purging every 1000 potential cycles and the performances of the electrocatalysts were then compared with regards to normalized ECSA and ORR limiting currents. The ADT-CVs for SL Pt₃Rh (Fig. 5a) and Pt₃Rh/VC (ESI, Fig. S6a†) electrocatalysts were recorded at different intervals up to 15 000 cycles. As the potential cycling was continued, the following changes were observed: (i) the hydrogen adsorption/desorption regions became more distinct and visibly developed; (ii) the shape of the anodic Pt-oxide region attained the pattern characteristic of pure Pt; (iii) the cathodic reduction peak shifted positively by 85 mV; and (iv) the capacitance current was reduced nearly 2.5-fold from the initial value. During the ADT test, the microstructural changes of supportless Pt₃Rh nanoclusters caused by Rh leaching was clearly observed from the continuously changing CV pattern during potential cycling, which finally evolved to being similar to the CV pattern of polycrystalline Pt. The dissolution of Rh from the SL Pt–Rh with a cluster morphology could increase the availability of Pt active sites, and hence greater values of ECSA occurred after long-term potential cycling. At the end of 15 000 cycles, the ECSA of SL Pt₃Rh was found to be increased from 154.4 to 168.2 m² g⁻¹ while for Pt₃Rh/VC the ECSA decreased from 63.1 to 6.2 m² g⁻¹, i.e., a nearly 9.2% increase in ECSA was seen for SL Pt₃Rh whereas a nearly 91.5% loss of ECSA was observed for VC-supported Pt₃Rh.

Fig. 5 (a) ADT CVs and (b) LSVs of SL Pt₃Rh at 2400 rpm. (c) Normalized ORR current density of supportless and VC-supported Pt₃Rh nanoclusters recorded at different potential cycling during ADT.

Fig. 5b shows the ORR polarization curves recorded for SL Pt₃Rh and ADT potential cycling at various intervals. The ADT-LSV curves reveal that the SL Pt₃Rh NC (Fig. 5b) retained nearly 85% of its initial limiting current density with 14 mV positive shifts in half-wave potential with the same onset potential. In contrast, Pt₃Rh/VC (ESI, Fig. S6b†) lost nearly 61% of its initial limiting current after 15 000 potential cycles, and the half-wave potential (80 mV) shifted negatively, hence increasing the overvoltage for dioxygen reduction. These results could be due to oxidation of the carbon support, and the subsequent effects such as catalyst dissolution and surface poisoning by adsorption of CO (that was produced from the carbon substrate oxidation) as reported in the literature.^5,6 In the mixed kinetic-diffusion control region (0.9 to 0.65 V), the polarization curves of the SL Pt₃Rh NC show a 14.9% current drop after the ADT, while there is a significant 60.6% reduction for the Pt₃Rh/VC. The percentage of normalized ECSA and the loss of ORR current densities were obtained from ADT-CVs (ESI,† Fig. 6c) and ADT-LSVs (Fig. 5c) at different intervals of potential cycling to assess the performance of the electrocatalysts. Table 2 shows the comparison of loss of ECSA and loss of ORR limiting current density for SL Pt₃Rh and Pt₃Rh/VC with the reported Pt-based nanostructures either in supportless or carbon-supported (commercial Pt/C) catalysts from various published reports.^34–45 It could be observed and concluded that the presence of Rh not only helps to stabilize the Pt nanoclusters but also affects the oxygen binding and desorption kinetics due to the improved electronic effects of the Pt–Rh nanoalloy. The synergistic effect of bimetallic catalysts possessing an enhanced catalytic activity could be due to the following influencing factors: bi-functional nature, ‘electronic’ ligand effects, change in inter-atomic distance, shift in d-band position and surface segregation.^46–48

Fig. 6 TEM images and corresponding histograms of supportless Pt₃Rh nanoclusters (a & c) before and (b & d) after ADT potential cycling (15 000 potential cycles).

Table 2 Comparison of the durability of Pt₃Rh nanocluster electrocatalysts with Pt-based nanostructures from various published reports^a

Reference	Catalyst	Loading (μg cm^−2)	ECSA (m^2 g^−1)	Loss of ECSA after ADT (%)/PC	ORR current density loss (%)/PC
a NA-nanoassembly; ND-nanodentrites; NW-nanowires; NC-nanoclusters; HP-hollow porous; PC-potential cycles.
37	Pt/C	11	71.0	74/10 800	-NA-
38	Self supported Pt NAs	NA	40.8	27.5/10 000	-NA-
39	Pt65Ir11Co24/C	16	70	85 90 70/20 000	-NA-
41	Pd–Pt ND	15.3	48.5	50/10 000	-NA-
42	Pt–Pd NW/C HP Pt C^−1	NA	0.45 cm^2	37& 50/15 000	∼3% with 15 mV degradation/15 000
43	Pt36Ni15Co49/C	6.6	56.6	28/10 000	-NA-
44	Pt/Vulcan	51.9	∼43.0	80.3/10 000	-NA-
45	Pt/Vulcan XC 72R	23.8	47.0	19.1/10 000	-NA-
Present study	SL Pt@Pt3Rh	17.8	214.3	7.8% increase/15 000	10.2/15 000
	SL Pt3Rh NC		154.4	9.1% increase/15 000	14.9/15 000
	Pt3Rh NC/VC		63.1	91.5/15 000	60.6/1500

---

Fig. 6a and b show the TEM images of SL Pt₃Rh NC recorded before and after ADT potential cycling (15 000 cycles), respectively. From these images, it can be clearly observed that there was no appreciable change in the cluster morphology of the bimetal catalyst. From the corresponding histograms (Fig. 6c and d), the average particle size was found increased nearly by ∼1.5 nm after 15 000 continuous potential cycles. It can be concluded that the SL Pt₃Rh NC electrocatalyst possesses greater stability than do VC-supported Pt₃Rh and the PtRh/C reported by Sethuraman et al. (∼77% loss of ECSA after 20 000 potential cycles).³⁶ These experimental findings demonstrate that the nanocluster morphology of SL Pt₃Rh with the optimized composition of 3 : 1 possesses more accessible active sites due to the presence of Rh and Rh-oxides as predicted by theoretical calculations reported by Friebel et al.,²⁵ and Baraldi et al.²⁶

3.6. Enhanced activity of Pt@Pt₃Rh nanoclusters

Although the stability of SL Pt₃Rh was substantially better than that of the standard Pt/C catalyst, the onset potential and activity still need to be improved. To address these issues, a thin layer of Pt was produced over the surface of previously formed SL Pt₃Rh, by reducing Pt precursor with formic acid under the same conditions as described above. The obtained catalyst is labeled as SL Pt@Pt₃Rh. The increase in Pt content was confirmed by EDX analysis (ESI, Fig. S7b†) and the SEM and TEM images (ESI, Fig. S7a†) revealed the retention of a cluster morphology without significant change. The CV profile of SL Pt@Pt₃Rh in N₂-saturated 0.5 M H₂SO₄ at a scan rate of 0.1 V s⁻¹ is shown in ESI, Fig. S7c.† It is seen that the peak potential of oxygen reduction is shifted in the positive direction by 150 mV compared to that of SL Pt₃Rh. But the LSV profiles (ESI, Fig. S7d†) at different rotation rates show an increased resistance as inferred from the sloppy behavior instead of a steep increase in cathodic current as expected for an active catalyst material, and the current density continues to increase slowly without reaching the limiting plateau. When ADT was performed over SL Pt@Pt₃Rh (Fig. 7a and b), nearly the same trend in stability up to 15 000 potential cycles was observed as that of SL Pt₃Rh.

Fig. 7 (a) CVs and (b) LSVs of SL Pt@Pt₃Rh nanoclusters recorded at different potential cycling during ADT.

In summary, further enriching the surface with a layer of Pt resulted in a 140 mV improved onset potential (ESI, Fig. S8†) compared with SL Pt₃Rh. Limitations of the activity, however, were seen in the form of sloppy LSV behavior. Such behavior might have been caused by the smooth surface of the Pt skin covering the rough surface structure and by kinks in the structure of SL Pt₃Rh that are ideal sites for oxygen binding.

4. Conclusions

In the present work we investigated the synthesis of Pt–Rh nanoclusters using formic acid as the reducing agent. TEM images showed finely distributed nanoparticles with an average size of ∼4 nm, and the shift in 2θ value substantiates alloy formation between Pt and Rh. The synthesized electrocatalyst showed good activity towards oxygen reduction, and the atomic composition with regards to Pt and Rh was optimized based on the electrochemical performance in sulfuric acid medium. The composition that showed the highest activity was found to be 76.9% Pt and 23 % Rh (atomic wt%). The number of electrons transferred per mole of oxygen was found to be 4.0 for SL Pt₃Rh, indicating that the reduction follows a 4-electron transfer mechanism. Accelerated durability tests showed that SL Pt₃Rh exhibits the highest stability for 15 000 potential cycles: it showed only a 9.2% increase in ECSA and retained 85% of the ORR limiting current. An attempt was made to further increase the activity of Pt–Rh by forming a Pt skin over the optimized composition of Pt₃Rh and this in fact increased the mass activity nearly 34 times compared to SL Pt₃Rh NC while possessing nearly the same stability up to 15 000 cycles. To conclude, the presence of Rh in the Pt₃Rh bimetal nanocluster structure enhances the catalytic activity, provides considerable stability to Pt, and seems to be the promising electrocatalyst for oxygen reduction in acid medium for PEM fuel cells.

Acknowledgements

This work was financially supported by the Board of Research in Nuclear Sciences (BRNS), Department of Atomic Energy (DAE), Govt. of India, under Basic Sciences program through sanction no. 2009/37/29/BRNS. S.B. profoundly thanks the management of SCSVMV University for having funded and created the necessary laboratory facilities to carry out this research project. B.N. thanks the CSIR, New Delhi for the award of Senior Research Fellowship (SRF) through the Sanction no. 09/1104(0001)/2013-EMR-I. K.K.R. Datta acknowledges financial support by the Operational Program Research and Development for Innovations-European Regional Development Fund (project CZ.1.05/2.1.00/03.0058 of the Ministry of Education, Youth and Sports of the Czech Republic) and the Operational Program Education for Competitiveness-European Social Fund (project CZ.1.07/2.3.00/30.0004 of the Ministry of Education, Youth and Sports of the Czech Republic).

References

1. N. M. Markovic, T. J. Schmidt, V. Stamenkovic and P. N. Ross, Fuel Cells, 2001, 1, 105.
2. H. Yano, E. Higuchi, H. Uchida and M. Watanabe, J. Phys. Chem. B, 2006, 110, 16544.
3. Y. H. Shih, G. V. Sagar and S. D. Lin, J. Phys. Chem. C, 2008, 112, 123.
4. Y. S. Horn, W. C. Sheng, S. Chen, P. J. Ferreira, E. F. Holby and D. Morgan, Top. Catal., 2007, 46, 285.
5. K. H. Kangasniemi, D. A. Condit and T. D. Jarvi, J. Electrochem. Soc., 2004, 151, E125.
6. J. Wang, G. Yin, Y. Shao, S. Zhang, Z. Wang and Y. Gao, J. Power Sources, 2007, 171, 331.
7. S. Sun, G. Zhang, D. Geng, Y. Chen, R. Li, M. Cai and X. Sun, Angew. Chem., Int. Ed., 2011, 50, 422.
8. Y. Tan, J. Fan, G. Chen, N. Zheng and Q. Xie, Chem. Commun., 2011, 47, 11624.
9. Z. Xu, H. Zhang, H. Zhong, Q. Lu, Y. Wang and D. Su, Appl. Catal., B, 2012, 111–112, 264.
10. X. Huang, E. Zhu, Y. Chen, Y. Li, C.-Yi. Chiu, Y. Xu, Z. Lin, X. Duan and Y. Huang, Adv. Mater., 2013, 25, 2974.
11. H. H. Li, C. H. Cui, S. Zhao, H. B. Yao, M. R. Gao, F. J. Fan and S. H. Yu, Adv. Energy Mater., 2012, 2, 1182.
12. V. D. Noto, E. Negro, K. Vezzu, L. Toniolo and G. Pace, Electrochim. Acta, 2011, 57, 257.
13. D. Yoo, C. Pak, S.-A. Jin and K. H. Lee, US Pat., 13 710 173, 2013.
14. M. T. Reetz, M. Lopez, W. Grunert, W. Vogel and F. Mahlendorf, J. Phys. Chem. B, 2003, 107, 7414.
15. D. F. A. Koch, D. A. J. Rand and R. Woods, J. Electroanal. Chem., 1976, 70, 73.
16. P. N. Ross, K. Kinoshita, A. J. Scarpellino and P. Stonehart, J. Electroanal. Chem. Interfacial Electrochem., 1975, 59, 177.
17. N. R. De Tacconi, J. M. Leger, B. Beden and C. Lamy, J. Electroanal. Chem., 1982, 134, 117.
18. F. H. B. Lima, D. Profeti, M. Chatenet, D. Riello, E. A. Ticianelli and E. R. Gonzalez, Electrocatalysis, 2010, 1, 72.
19. T. Kawaguchi, Y. Rachi, W. Sugimoto, Y. Murakami and Y. Takasu, J. Appl. Electrochem., 2006, 36, 1117.
20. Q. Yuan, Z. Zhou, J. Zhuang and X. Wang, Chem. Mater., 2010, 22, 2395.
21. R. T. S. Oliveira, M. C. Santos, P. A. P. Nascent, L. O. S. Bulhoes and E. C. Pereira, Int. J. Electrochem. Sci., 2008, 3, 970.
22. K. W. Park, J. H. Choi, S. A. Lee, C. Pak, H. Chang and Y. E. Sung, J. Catal., 2001, 224, 236.
23. K. W. Park, D. S. Han and Y. E. Sung, J. Power Sources, 2006, 163, 82.
24. V. D. Noto and E. Negro, J. Power Sources, 2010, 195, 638.
25. D. Friebel, D. J. Miller, D. Nordlund, H. Ogasawara and A. Nilsson, Angew. Chem., Int. Ed., 2011, 50, 10190.
26. A. Baraldi, L. Bianchettin, S. de Gironcoli, E. Vesselli, S. Lizzit, L. Petaccia, G. Comelli and R. Rosei, J. Phys. Chem. C, 2011, 115, 3378.
27. K. Yuge, T. Ichikawa and J. Kawai, Mater. Trans., 2010, 51, 321.
28. P. D. Tran, S. Y. Chiam, P. P. Boix, Y. Ren, S. S. Pramana, J. Fize, V. Artero and J. Barber, Energy Environ. Sci., 2013, 6, 2452.
29. Y. Zhang, M. Janyasupab, C. W. Liu, X. Li, J. Xu and C. Liu, Adv. Funct. Mater., 2012, 22, 3570.
30. K. Mohanraju and L. Cindrella, RSC Adv., 2014, 4, 11939.
31. Q. He, X. Yang, W. Chen, S. Mukerjee, B. Koel and S. Chen, Phys. Chem. Chem. Phys., 2010, 12, 12544.
32. W. Sheng, S. Chen, E. Vescovo and Y. S. Horn, J. Electrochem. Soc., 2012, 159, B96.
33. H. A. Gasteiger, S. S. Kocha, B. Sompalli and F. T. Wagner, Appl. Catal., B, 2005, 56, 9.
34. W. Sheng, H. A. Gasteiger and Y. S. Horn, J. Electrochem. Soc., 2010, 157, B1529.
35. X. Q. Gong, Z. P. Liu, R. Raval and P. Hu, J. Am. Chem. Soc., 2004, 126, 8–9.
36. V. A. Sethuraman, J. W. Weidner, A. T. Haug, M. Pemberton and L. V. Protsailo, Electrochim. Acta, 2009, 54, 5571.
37. N. M. Markovic, H. A. Gasteiger and P. N. Ross, J. Phys. Chem., 1995, 99, 3411.
38. J. C. Meier, C. Galeano, I. Katsounaros, A. A. Topalov, A. Kostka, F. Schuth and K. J. J. Mayrhofer, ACS Catal., 2012, 2, 832.
39. B. Y. Xia, W. T. Ng, H. B. Wu, X. Wang and X. W. Lou, Angew. Chem., Int. Ed., 2012, 51, 1.
40. R. Loukrakpam, B. N. Wanjala, J. Yin, B. Fang, J. Luo, M. Shao, L. Protsailo, T. Kawamura, Y. Chen, V. Petkov and C. J. Zhong, ACS Catal., 2011, 1, 562.
41. B. Lim, M. Jiang, P. H. C. Camargo, E. C. Cho, J. Tao, X. Lu, Y. Zhu and Y. Xia, Science, 2009, 324, 1302.
42. C. Koenigsmann, A. C. Santulli, K. Gong, M. B. Vukmirovic, W.-ping. Zhou, E. Sutter, S. S. Wong and R. R. Adzic, J. Am. Chem. Soc., 2011, 133, 9783.
43. B. N. Wanjala, R. Loukrakpam, J. Luo, P. N. Njoki, D. Mott, C. J. Zhong, M. Shao, L. Protsailo and T. Kawamura, J. Phys. Chem. C, 2010, 114, 17580.
44. F. Takasaki, S. Matsuie, Y. Takabatake, Z. Noda, A. Hayashi, Y. Shiratori, K. Ito and K. Sasaki, J. Electrochem. Soc., 2011, 158, B1270.
45. F. Hasche, M. Oezaslan and P. Strasser, J. Electrochem. Soc., 2012, 159, B24.
46. S. Alayoglu and B. Eichhorn, J. Am. Chem. Soc., 2008, 130, 17479.
47. J. A. Rodriguez and D. Wayne Goodman, J. Phys. Chem., 1991, 95, 4196.
48. H. L. Jiang and Q. Xu, J. Mater. Chem., 2011, 21, 13705.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra08490j

---