Platinum Particles Electrochemically Deposited on Multiwalled Carbon Nanotubes for Oxygen Reduction Reaction in Acid Media

Platinum nanoparticles supported on multiwalled carbon nanotubes (Pt/MWCNT) were prepared using controlled electrochemical deposition from Ar-saturated H2PtCl6 solution. The electrodeposition parameters were varied to observe change in the surface morphology and electrocatalytic activity of the Pt/MWCNT catalysts for oxygen reduction reaction (ORR). Surface morphology of the prepared catalysts was studied by scanning electron microscope (SEM) and scanning transmission electron microscope (STEM). For electrochemical characterization of the Pt/MWCNT electrocatalysts, CO-stripping and cyclic voltammetry (CV) experiments were performed in 0.05 M H2SO4 solution. The ORR was studied in acid medium using the rotating disk electrode (RDE) method. The results obtained were analyzed and compared to those of commercial Pt/C. The size, shape and distribution of Pt particles as well as the electrochemical behavior of Pt/MWCNT modified electrodes showed strong dependence on the deposition parameters. The catalyst materials prepared by electrochemical deposition showed higher specific activities than commercial Pt/C catalyst. The Pt/MWCNT cathode materials showed remarkable stability during repetitive potential cycling in 0.05 M H2SO4. © The Author(s) 2017. 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: 10.1149/2.0091712jes] All rights reserved.

Proton exchange membrane fuel cell (PEMFC) is identified as one of the most promising environmentally friendly sustainable energy conversion devices that efficiently convert fuel energy into electricity through electrochemical processes. [1][2][3] In comparison to the conventional internal combustion engine, low-temperature fuel cells are preferred for energy conversion in vehicles because of its green operating system, high efficiency and renewable energy resources. 4,5 One of the most critical challenges for commercializing fuel cells is the development of more efficient, highly active and durable electrocatalysts to improve the sluggish kinetics of the oxygen reduction reaction (ORR). To date, Pt-based cathode catalysts have shown significantly high electrocatalytic activity toward the ORR as compared to other metals, alloys, core-shell structures, mixed metal oxides and nonnoble metal catalysts. [6][7][8][9][10][11][12][13][14][15] Considerable efforts have been dedicated to minimize the amount of Pt loading and improve its electrocatalytic activity and durability, mainly because of its scarcity and high cost. [16][17][18][19] One of the most common and successful approach is to uniformly deposit platinum nanoparticles (PtNPs) onto an appropriate support material. Carbon-based supports like carbon black, graphene materials, carbide-derived carbon, carbon nanotubes and carbon nanofibers are most commonly used for this purpose. [20][21][22][23][24][25][26][27][28] Multiwalled carbon nanotubes (MWCNT) have some outstanding properties such as high surface area, remarkable electrical conductivity, high chemical stability, and mechanical strength, which make them a promising support material for PtNPs. [29][30][31][32][33][34] Pt/MWCNT has recently been prepared through chemical procedure using a mixture of chloroplatinic acid, polyvinylpyrrolidone and ethylene glycol. 29 Electrochemical measurements showed improved ORR activity of the Pt/MWCNT catalyst which was attributed to the dispersion of PtNPs on the pyrenecarboxylic acid functionalized MWCNTs. Rahsepar et al. synthesized Pt/MWCNT electrocatalysts for ORR by impregnation method using microwave heating. 30 Electrochemical measurements of the prepared catalysts showed high ORR activity with typical four-electron pathway. Another approach for the dispersion of PtNPs on MWCNTs is the in situ decomposition of platinum acetylacetonate in liquid polyols. 6 Our group has previously used MWCNTs for the sputter-deposition of PtNPs. 34 The prepared Pt/MWCNT electrodes showed higher ORR activity than that of bulk Pt electrode.
As mentioned above, various chemical and physical methods have been used for the synthesis of Pt/MWCNT electrocatalysts to improve its ORR activity and long-term durability. Electrochemical deposition is one of the most efficient and versatile methods used for the direct growth of PtNPs in a controlled manner. [35][36][37][38][39][40] Chen et al. deposited porous PtNPs on a carbon nanotube (CNT) film in 0.5 M H 2 SO 4 solution using the electrodeposition technique. 41 They succeeded to prepare Pt aggregates of different shapes and morphology by tailoring the deposition parameters. The prepared electrocatalysts showed higher ORR activity and durability than commercial Pt/C in acidic medium. PtNPs have been electrodeposited on MWCNT/carbon paper using H 2 PtCl 6 (10 g L −1 ) and HCl (60 g L −1 ) bath. 42 It was found that the particle size distribution mainly depends on the electrodeposition potential, uniform distribution of PtNPs was observed at −0.6 V vs saturated calomel electrode (SCE). Kar and Sharma electrochemically deposited PtNPs on CNT-coated carbon fiber (CF) for ORR. 43 Different deposition potentials and deposition time were applied to observe change in the Pt particle size. They observed the formation of discrete Pt clusters, which may be attributed to the nucleation of particles at specific active sites on the surface of CNTs. Huang and coworkers followed rather different approach for the preparation of Pt/C electrocatalysts for ORR. 44 In order to prevent aggregation of PtNPs, they first chemically deposited Pt and Pt-Pd-Ru alloy NPs onto the carbon support as nuclei, followed by CV and pulse electrodeposition in 2.5 mM H 2 PtCl 6 solution. They observed higher ORR activity of the prepared catalysts than that of commercial Pt/C. The enhanced ORR efficiency was attributed to the improved mass transfer rate due to higher surface area of the evenly grown PtNPs. PtNPs were successfully deposited onto graphene modified glassy carbon (GC) electrode using cyclic voltammetric reduction in K 2 PtCl 4 solution. 45 Surface analysis of the prepared electrodes by SEM, X-ray photoelectron spectroscopy, and Raman spectroscopy showed that no agglomeration occurred on graphene, which can be attributed to the interaction between PtNPs and graphene sheets. The electrocatalysts showed improved ORR activity as compared to that of commercial Pt/C. Fortunato et al. electrochemically deposited Pt nanoclusters and Pt porous dendritic structures on the surface of GC electrodes modified with a wide range of carbon-based support materials namely MWCNTs, graphite, chemically converted graphene, graphene oxide (GO), graphene oxide nanoribbons (GONR) and graphene nanoribbons (GNR). 46 They reported that the electrodes prepared by electrodeposition of Pt on graphene and GNR were more active than commercial Pt/C (10 wt%) while those prepared using MWCNTs, GO and GONR failed to show improved ORR activity. To optimize the parameters for electrochemical deposition, and to control the loading, size and morphology of PtNPs, Burk and Buratto applied pulse potential deposition with varying parameters such as the pulse width, the pulse potential and the pH of the H 2 [Pt(OH) 6 ] solution. 47 The optimum potential for pulse deposition they reported was ±1.0 V vs Ag/AgCl and the optimum deposition time was 3 ms. The average Pt particle size they obtained was 2.46 nm with a narrow particle size distribution. The electrodes prepared by pulse deposition were used in PEMFCs, which showed very high gravimetric power density as compared to the commercial electrodes.
In this work we prepared Pt/MWCNT electrocatalysts for ORR by growing Pt particles on the surface of acid-treated MWCNT modified GC electrode using controlled pulse electrodeposition technique. The deposition potential and number of pulses were tailored in order to observe change in the particle size, shape and number density as well as the ORR activity of the electrocatalysts. The aim of this study is to evaluate electrodeposition conditions for the preparation of Pt/MWCNT catalysts for ORR. Preparation of MWCNT modified GC electrode.-A known amount of acid-treated MWCNTs was suspended in 2-propanol (1 mg mL −1 ). The procedure for purifying MWCNTs by acid-treatment has been reported earlier. 48 Briefly, MWCNTs were refluxed in a mixture of concentrated H 2 SO 4 and HNO 3 (1:1, v/v) at 55 • C for 2 h and then at 80 • C for 3 h. The refluxed MWCNTs were then washed thoroughly with Milli-Q water by centrifugation (3000 rpm, 10 min) followed by filtration until the filtrate became neutral. Finally, the MWCNTs were dried in vacuum for 15 h. The MWCNT dispersion was then put into the ultrasonic bath for 30 min to get a homogenous suspension. Glassy carbon (GC) electrodes of surface area of 0.126 cm 2 were cut from GC rods, fixed in Teflon holders and were polished on 1.0 and 0.3 μm alumina slurries (Buehler) to get a smooth shiny surface. After polishing the electrodes were sonicated in Milli-Q water and 2-propanol for 3 min in each. In order to get better distribution of the MWCNTs on the GC electrode surface, 2.1 μL of the suspension was transferred 6 times onto the polished GC electrode with a pipette and hence, the loading of MWCNTs on the GC was 0.1 mg cm −2 .

Electrodeposition of PtNPs on MWCNT modified GC electrode.-The electrodeposition experiments were carried out in
Ar-saturated 1 mM H 2 PtCl 6 solution in 0.05 M H 2 SO 4 using conventional 3-electrode cell configuration. An Autolab PGSTAT30 potentiostat/galvanostat (Metrohm Autolab, The Netherlands) was used for electrochemical measurements. The deposition potential values were applied with respect to SCE, while Pt wire separated by a glass frit from the working solution was used as counter electrode. A series of experiments were performed using two different deposition potentials namely at 0 and −0.35 V, which were denoted as procedure A and B, respectively. In each procedure, the number of pulses was changed in the range of 100-1500 to observe change in the surface morphology and ORR electrocatalytic activity of the catalysts. Pulse time for deposition was kept constant at 250 ms during the whole range of experiments, while the interval at 1.0 V between two consecutive pulses was 3 s. For adequate mass transfer, the electrode was rotated at 1000 rpm in all experiments. The prepared electrodes were then designated as Pt/MWCNT-X y where X is the deposition procedure (A or B) and y represents the number of pulses.
Instrumentation and measurements.-The surface morphology and particle size and number density of the prepared Pt/MWCNT catalysts was studied by a Helios NanoLab 600 (FEI) SEM and STEM (Titan 200, FEI), supplied with energy dispersive X-ray spectrometer (EDS) Super-XTM system. To perform electrochemical measurements, a five-necked glass cell with a standard three-electrode configuration was used. For surface decontamination and characterization, CO-stripping was performed in 0.05 M H 2 SO 4 solution at 20 mV s −1 . Cyclic voltammograms were recorded at a scan rate (ν) of 50 mV s −1 in Ar-saturated 0.05 M H 2 SO 4 . The ORR measurements were carried out at 10 mV s −1 in O 2 -saturated 0.05 M H 2 SO 4 solution at different electrode rotation rates (ω), which ranged from 360 to 4600 rpm, using rotating disk electrode (RDE) system consisting of an EDI101 rotator and a CTV101 speed control unit (Radiometer). All the potential values are applied with respect to the reversible hydrogen electrode (RHE).

Results and Discussion
Surface morphology of Pt/MWCNT samples.-SEM images show that Pt particles of different shapes, sizes and number density are formed on the surface of MWCNTs (Figures 1a-1f). It was observed that the shape of the particles was strictly dependent on the deposition potential, while the particle size and number density was attributed to the number of pulses. The particles deposited by procedure A acquired spherical, cauliflower-like shape (Figures 1a-1c) while those deposited by procedure B were cube-shaped (Figures 1d-1f). It was also revealed that the Pt nanoparticles formed by procedure A are larger in size than those formed by procedure B.  (Figures 1d-1f). Moreover, it can be seen in Figures 1d-1f that at more negative potential of deposition the Pt particles are formed not only on the outer MWCNT surface, but also on the areas underneath while at 0 V the Pt particles were detected at the outer surface of MWCNTs only (Figures 1a-1c).
It was revealed from SEM measurements that the formation of the Pt particles follows a regular pattern and these particles are formed at certain nucleation sites. The number of such Pt particles increased with increasing number of pulses and decreasing the deposition potential. In order to figure out if there is any deposition of smaller PtNPs dispersed on the surface of MWCNTs, the STEM measurements were carried out as shown in Figure 2. It was confirmed that PtNPs deposited at certain specific nucleation sites only and there were no smaller Pt nanoparticles dispersed on MWCNTs. The formation of similar clusters at certain active sites on the surface of CNTs has been reported earlier. 43,44,49 Sharma and Kar suggested that the corroded rough surface of the CNTs after acid-treatment may provide active sites for nucleation. 43 Huang et al. suggested that the number of Pt nuclei formed on the carbon surface depends on the hydrophilicity and uniformity of the support material. 44 Earlier reports confirmed that formation of various functional groups in acid-treated CNTs act as nucleation sites. 49 Figure 3a. The current densities were obtained by normalizing the measured current to geometric surface area of GC electrode. Complete blockage of the Pt surface due to the pre-adsorbed CO can be seen during the first potential cycle in the potential range from 0.05 to 0.4 V, which is removed by oxidative stripping during the second potential cycle in  the potential range from 0.05 to 1.0 V. It has been reported earlier that the pre-peak in the CO-stripping voltammograms of Pt nanoparticles can be attributed to a change in surface morphology such as particle size distribution and aggregation of the nanoparticles. 53 Studies with stepped surfaces have shown that the presence of different steps catalyze the oxidation of CO. [54][55][56] During the third potential cycle, in the same potential range, it was confirmed that the pre-adsorbed CO is now completely oxidized. 31 A comparison of the CO-oxidation profile for Pt/MWCNT-A and Pt/MWCNT-B is shown in Figures 3b  and 3c, respectively. It was found that all the electrodes prepared by both electrodeposition procedures (A and B) followed almost the same CO-oxidation profile. The only exception was Pt/MWCNT-A 100 where no Pt was detected and the electrode behaved more like a GC electrode modified with pristine MWCNTs and hence, the electrochemical results for Pt/MWCNT-A 100 are not presented hereafter. Since there were no desorption peaks in the hydrogen underpotential deposition region, it was revealed that the electrode surface is completely blocked by the pre-adsorbed CO. A broad CO oxidation peak at ≈0.72 V and a shoulder at ≈0.75 V was detected in the second potential cycle. 57 The shoulder at ≈0.75 V can be easily detected in Pt/MWCNT-B catalysts but not in the case of Pt/MWCNT-A, especially with increased number of pulses. This can be attributed to the difference in the nature of crystal facets at the surface of the PtNPs formed at different deposition parameters. 57 The CO-oxidation peak height increases with increasing the number of pulses, which is attributed to the metal loading. After decontamination of the surface by  (Figure 4). Three consecutive potential cycles were measured for each electrode in the range of 0.05-1.45 V in order to get a stable CV curve while the potential scan rate was 50 mV s −1 . Figure 4a presents a comparison of the CV curves obtained from Pt/MWCNT-A at different number of pulses while Figure 4b shows CV curves obtained from Pt/MWCNT-B. Like CO oxidation behavior, Pt/MWCNT electrodes prepared by both procedures showed almost the same profile for cyclic voltammetry. Clear hydrogen adsorption/desorption peaks are detectable for almost all the electrodes in the potential region between 0.05 and 0.4 V, which corresponds to the different crystal facets of the PtNPs. The hydrogen desorption peak at ∼0.14 V is attributed to Pt(110) while that at ∼0.30 V is attributed to Pt(100) crystal facets. 57,58 Assuming that a charge of ∼210 μC cm −2 is required to desorb a monolayer of hydrogen from the Pt catalyst surface, the real surface area (A r ) of the electrocatalyst was calculated by integrating the charge under the hydrogen desorption peaks. 59 The A r values calculated for all the electrodes studied are given in Table I. It can be seen that Pt/MWCNT-B series catalysts possess higher A r values than their Pt/MWCNT-A counterparts. This can be attributed to the particular dendritic-type Pt clusters formed by procedure B as shown in Figure 2b.
Oxygen reduction reaction studies.-The electrochemical reduction of oxygen was investigated in O 2 -saturated 0.05 M H 2 SO 4 solution using the RDE method. The ORR polarization curves were recorded at different electrode rotation rates in the potential range from 0.05 to 1.1 V at a scan rate of 10 mV s −1 . The background current was measured using the same procedure in Ar-saturated solution, which was then subtracted from the RDE data. The ORR polarization curves of Pt/MWCNT-B 400 measured at various rotation rates are presented in Figure 5.
The number of electrons transferred per O 2 molecule (n) during the ORR was calculated from the RDE data using the Kouteckey-Levich (K-L) equation as given below: where j is the measured current density, j k and j d are the kinetic and diffusion-limited current densities, respectively, F is the Faraday constant (96,485 C mol −1 ), k is the rate constant for ORR, is the diffusion coefficient of O 2 in 0.05 M H 2 SO 4 (1.93 × 10 −5 cm 2 s −1 ), 60 v is the kinematic viscosity of the solution (0.01 cm 2 s −1 ), 61 and ω is the electrode rotation rate (rad s −1 ).
The RDE data presented in Figure 5 was used to construct the K-L plots as shown in the inset of Figure 5. It was confirmed form the RDE data analyzed by Eq. 1 that all the Pt/MWCNT electrocatalysts followed the typical four-electron pathway for the electrochemical reduction of oxygen. The four-electron pathway for Pt-based catalysts prepared by electrodeposition method has been reported earlier. 37 Table I. It was observed that overall the E 1/2 increased with increasing number   of pulses. The lowest E 1/2 value of 0.75 V was shown by Pt/MWCNT-A 200 . In the case of Pt/MWCNT-A, all the electrodes prepared with the number of pulses ≥800 showed higher E 1/2 value than that of the commercial Pt/C while in the case of Pt/MWCNT-B, the electrodes prepared with the number of pulses ≥400 showed higher E 1/2 value than that of the commercial Pt/C. The highest E 1/2 value of 0.88 V was shown by Pt/MWCNT-B 1500 which is attributed to the increased A r value. It can be observed in Figure 6 that the values of diffusion limited current density slightly deviate from the theoretical value of 6.2 mA cm −2 at 1900 rpm. This can be attributed to the rough surface of the MWCNT coated GC electrodes. Further analysis of the ORR data was carried out by constructing mass-transfer corrected Tafel plots for Pt/MWCNT electrodes as shown in Figure 7. The Tafel analysis provides an insight into the ORR mechanism by measuring kinetic currents in the absence of diffusion limitations. Two characteristic Tafel slope regions were observed from which the slope values were determined (Table I). The Tafel slope values for all Pt/MWCNT-A ( Figure 7a) and Pt/MWCNT-B (Figure 7b) electrodes were close to −60 and −120 mV dec −1 at low current densities and higher current densities, respectively, which indicated that all the electrodes followed the same mechanistic pathway and that the rate of the ORR is determined by the transfer of the first electron to the adsorbed O 2 molecule. 32,34,62 The specific activity (SA) values of the Pt/MWCNT electrodes for ORR were calculated from the equation given below: where I k is the kinetic current and A r is the real electroactive surface area of the Pt catalyst. The specific activities calculated at 0.9 V  are presented in Table I. All the Pt/MWCNT modified electrodes showed higher SA values than that of commercial Pt/C. The electrodes prepared by procedure A showed higher specific activities than those prepared by procedure B. It was observed that the SA decreased with increasing the number of pulses. In the case of Pt/MWCNT-A, the highest SA of 0.45 mA cm −2 was shown by Pt/MWCNT-A 200 , which is more than 4-times higher than that of the commercial Pt/C. Moreover, a decrease in the SA value was observed with increasing number of pulses up to 800 after which a constant value of 0. was observed in the case of Pt/MWCNT-B 1500 , which was still higher than that of commercial Pt/C catalyst. The specific activities of the electrodes prepared with lower number of deposition pulses are higher than those reported earlier for Pt/C catalysts. In our previous work heat-treated Pt/MWCNT catalysts were studied where maximum SA of 0.32 mA cm −2 at 0.9 V was observed after heat-treatment at 300 • C. 32 Gasteiger and co-workers obtained SA values of 0.20 mA cm −2 at 0.9 V in acid electrolyte for state-of-the-art Pt/C catalysts. 63 Li et al. studied composite materials prepared by mixing Pt-coated reduced graphene oxide with carbon black. The SA value obtained from this catalyst was 0.21 mA cm −2 . 64 Similarly, SA value of 0.22 mA cm −2 at 0.9 V was reported for Pt deposited on gDNA-graphene oxide composite. 65 Since the material has dendritic structure and the surface is rather porous, SA is increased with the decrease of surface area. There might be mass-transfer restrictions due to the porous nature of the catalyst material, which could lead to lower SA values, although a detailed characterization of the catalyst structure has not been made in this work. While dealing with the K-L analysis we cautiously considered important methodological aspects of this analysis method for studying the kinetics of the ORR. 66 In addition, long-term durability tests were performed in order to investigate the stability properties of the prepared catalyst materials in acid media. Figure 8a shows CV profiles of Pt/MWCNT-B 200 during ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 193.40.12.10 Downloaded on 2017-08-10 to IP 5000 potential cycles. As can be seen from the inset of Figure 8a the true area of Pt degraded by almost 48%. Nevertheless, despite the harsh durability testing conditions used the values of half-wave potential shifted only 15 mV. Figure 8b presents RDE data before and after 5000 repetitive potential cycles. Recently, many groups have contributed their efforts to figure out the optimum conditions for the electrochemical deposition of PtNPs onto the support material surface. [37][38][39][40][41][42][43][44][45][46]67,68 Porous Pt aggregates of different shape and surface morphology have been electrochemically deposited on carbon nanotube layer. 41 Basically, two different deposition techniques were applied namely, pulse electrodeposition in which the working electrode was kept switching between two constant potentials and CV at a constant scan rate. It was revealed that Pt aggregates of different size, shape and morphology could be achieved by changing the deposition parameters. The catalysts showed high ORR activity and stability as compared to that of commercial Pt/C, even after 5000 repetitive potential cycles in the potential range from 0.6 to 1.24 V vs RHE. 41 Another approach is to prepare hierarchical structured catalyst layer for ORR by electrochemical deposition of PtNPs onto CNTs coated on CF, which was used as a substrate. 43 Briefly, CNTs were chemically grown on the surface of CF followed by electrodeposition of PtNPs in H 2 PtCl 6 solution. SEM images showed thick forest of Pt-NPs coated CNTs finely grown on the surface of CF. The cluster size was ranging from ∼80 to ∼170 nm with an average particle size of ∼10 nm. Furthermore, the Pt/CNT/CF electrode showed an enhanced electrocatalytic activity toward the ORR, which is attributed to the PtNPs clusters formed at the active sites of the CNTs. Unwin and coworkers electrochemically deposited PtNPs on singlewalled carbon nanotubes (SWCNT) grown on Si/SiO 2 . 37 The electrodeposition was performed using cyclic voltammetry for a certain number of cycles, in K 2 PtCl 6 solution where SWCNT network was used as a working electrode. They reported the formation of Pt particles with their size ranged from 175 to 375 nm at certain nucleation sites. They introduced microcapillary electrochemical method as a versatile technique that can be further optimized to prepare highly active Pt/SWCNT electrocatalysts for ORR. PtNPs have been successfully deposited onto reduced graphene oxide (RGO) modified GC electrode by electroreduction of Pt salt. 69 SEM measurements showed homogenous distribution of the high density PtNPs on the modified GC electrode surface. The Pt/RGO modified electrodes showed considerably higher ORR activity than that of RGO modified GC electrode. The durability of various nanocarbon-supported Pt catalysts has been investigated recently. 70 The work reported herein will contribute to the preparation of Pt/MWCNT catalysts that may find application as cathode materials of PEMFCs.

Conclusions
PtNPs were electrochemically deposited onto MWCNTs in a controlled manner, not only on the outer surface but also on the available surface underneath. Pt particles with increased surface area and excellent electrocatalytic activity were formed. It was revealed that the surface morphology of the PtNPs depends on the deposition parameters. Briefly, the shape of the Pt particles depends on the deposition potential while the particle size and number density mainly depends on the number of pulses. High surface area dendritic Pt particles are formed at more negative deposition potential. Moreover, all the prepared Pt/MWCNT electrodes showed higher SA values than that of commercial Pt/C. The specific activities decreased with increasing number of pulses independent of the deposition potential. Using the K-L equation, it was confirmed that all the electrodes follow the typical four-electron pathway while the Tafel slope values revealed that slow transfer of the first electron to O 2 molecule is the rate-determining step for ORR on Pt/MWCNT prepared by electrochemical deposition.