Improvement of Catalytic Activity of Platinum Nanoparticles Decorated Carbon Graphene Composite on Oxygen Electroreduction for Fuel Cells

: High-performance platinum (Pt)-based catalyst development is crucially important for reducing high overpotential of sluggish oxygen reduction reaction (ORR) at Pt-based electrocatalysts, although the high cost and scarcity in nature of Pt are profoundly hampering the practical use of it in fuel cells. Thus, the enhancing activity of Pt-based electrocatalysts with minimal Pt-loading through alloy, core − shell or composite making has been implemented. This article deals with enhancing electrocatalytic activity on ORR of commercially available platinum / carbon (Pt / C) with graphene sheets through a simple composite making. The Pt / C with graphene sheets composite materials (denoted as Pt / C x :G 10 − x ) have been characterized by several instrumental measurements. It shows that the Pt nanoparticles (NPs) from the Pt / C have been transferred towards the π -conjugated systems of the graphene sheets with better monolayer dispersion. The optimized Pt / C 8 :G 2 composite has higher speciﬁc surface area and better degree of graphitization with better dispersion of NPs. As a result, it shows not only stable electrochemical surface area but also enhanced ORR catalytic activity in respect to the onset potential, mass activity and electron transfer kinetics. As shown by the ORR, the Pt / C 8 :G 2 composite is also better resistive to the alcohol crossover e ﬀ ect and more durable than the Pt / C.


Introduction
The sluggish kinetics of four electron involved electrocatalytic oxygen reduction reaction (ORR) in an acidic environment is a major obstacle for the development of cost effective fuel cells (FCs) [1][2][3][4]. The development of catalysts with high activity for the ORR is essential to proton exchange membrane fuel cells (PEMFCs), since the majority of activation losses occur at the cathode and it turns into an interesting research area [1][2][3][4][5][6][7]. The platinum (Pt)-based materials have been regarded as the most efficient electrocatalyst for ORR [8][9][10][11] because of the excellent ORR catalytic activity. Carbon supported Pt (Pt/C) is typically utilized as the state-of-the-art ORR catalyst, but is inadequate for the combined cost, performance and durability requirements [12][13][14][15]. To promote the development of highly active ORR catalysts, the United States Department of Energy sets the target as mass activity (MA, the catalytic activity per given mass of Pt, 0.44 A mg −1 , 2017-2020) [16,17]. Currently, the high cost of Pt becomes the primary limiting factor preventing the widespread adoption of FCs [18][19][20], thus, the high Pt-MA must be achieved. However, the MA is determined by the specific activity (SA, normalized by surface area) and the electrochemically active surface area (ECSA, normalized by mass), which can be optimized by tuning the exposed catalytic site on catalyst surface [21][22][23]. Therefore, several challenges, such as the lowering amount of Pt-loading in the cathode, preventing Pt nanoparticles (NPs) from severe dissolution and/or migration and corrosion of carbon substrate during

Material Synthesis
For the preparation of Pt/C x :G 10−x , 50 mg of 20 wt% Pt/C (E-TEK) was taken in a 100 mL vial, followed by the addition of 70 mL water under vigorous stirring. Under continuous stirring, 1 mg mL −1 of Graphene oxide (GO) solution (mass ratio of Pt/C:G was, x:10−x, where x = 10 to 1) was added slowly and stirred for next 30 min. Afterwards, 10 mL of 0.1 M NaBH 4 was added to the above solution and refluxed at 100 • C for 12 h. The solution was then filtered and washed with water three times. The as-obtained Pt/C x :G 10−x was then dried at 60 • C for 12 h in a vacuum oven. For comparison, G-sheets catalyst was prepared with the same process without addition of Pt/C.

Electrochemical Characterization
A 10 µL of G, Pt/C x :G 10−x or Pt/C suspension in ethanol (1 mg mL −1 with 5 µL of 5% Nafion solution) was dropped onto the prepolished glassy carbon electrode (GCE, geometric area: 0.0707 cm 2 ). All electrochemical techniques including cyclic voltammetry (CV), linear sweep voltammetry (LSV) and chronoamperometry (CA) were performed using a three-electrode potentiostat [CHI 700C electrochemical workstation (U.S.A.)]. A Pt wire was used as auxiliary electrode and Ag/AgCl electrode was used as reference electrode. The rotating ring disk electrode (RRDE) measurements were done using an EG&G Model 636 RDE system along with the CHI 700C. All electrochemical experiments were performed in high purity argon (Ar) −

Surface Analysis of Pt/C x :G 10−x
The Pt/C x :G 10−x including Pt/C and G-sheets were first characterized by transmission electron microscopy (TEM) analysis and the images of as-prepared composites are displaying in Figure 1. The TEM image of the G-sheets showed the typical thin film and nanosheets-like structure (Figure 1a). The G-sheets with a laminar characteristic assembled by a wrinkled silk veil were obtained after removal of oxygen functional groups form the surface due to reduction. The partially crinkled nature of G-sheets probably originated from the defective nature, which was generated from the reduction process of GO [27]. Pt/C ( Figure 1b) and Pt/C 8 :G 2 (Figure 1c) consist of clear crystal Pt NPs with spherical size and without any serious aggregation or agglomeration. The images of both samples are showing that the Pt NPs were mostly in uniform size within the range of 2 to 4 nm. The most striking feature is that the Pt NPs were migrated onto the surfaces of G-sheets (Figure 1d). For a better understanding, the high resolution TEM (HRTEM) images were also recorded, which show the migration of Pt NPs from Pt/C to G-sheets with excellent mono-dispersion that can greatly boost the electrocatalytic activity. 1a). The G-sheets with a laminar characteristic assembled by a wrinkled silk veil were obtained after removal of oxygen functional groups form the surface due to reduction. The partially crinkled nature of G-sheets probably originated from the defective nature, which was generated from the reduction process of GO [27]. Pt/C ( Figure 1b) and Pt/C8:G2 (Figure 1c) consist of clear crystal Pt NPs with spherical size and without any serious aggregation or agglomeration. The images of both samples are showing that the Pt NPs were mostly in uniform size within the range of 2 to 4 nm. The most striking feature is that the Pt NPs were migrated onto the surfaces of G-sheets (Figure 1d). For a better understanding, the high resolution TEM (HRTEM) images were also recorded, which show the migration of Pt NPs from Pt/C to G-sheets with excellent mono-dispersion that can greatly boost the electrocatalytic activity.
To understand the influence and optimization of G-sheets addition with Pt/C, the morphology has also been investigated as in Figure 2. Figure 2, shows the TEM images of Pt/C (a), Pt/C8:G2 (b), Pt/C6:G4 (c) and Pt/C1:G9 (d). All samples contained Pt NPs on carbon and G-sheets and the Pt NPs were transferred onto G-sheets upon simple addition. The TEM image of Pt/C showed that the higher density of NPs with such aggregation on carbon matrix (Figure 2a). Upon addition of 20% (w/w) of G-sheets, the Pt NPs were migrated onto G-sheets with aggregation-free distribution (Figure 2b). At 40% (w/w) addition, the Pt NPs were also migrated with lower amount in Figure 2c. For comparison, it is also observed that the Pt NPs were migrated with lower amount at the surface of Pt/C1:G9 (Figure 2d). Therefore, as observed in Figures 1 and 2, the addition of 20% of G-sheets was optimal for the preparation of good quality of Pt/Cx:G10−x with better Pt NPs dispersion morphology.  To understand the influence and optimization of G-sheets addition with Pt/C, the morphology has also been investigated as in Figure 2. Figure 2, shows the TEM images of Pt/C (a), Pt/C 8 :G 2 (b), Pt/C 6 :G 4 (c) and Pt/C 1 :G 9 (d). All samples contained Pt NPs on carbon and G-sheets and the Pt NPs were transferred onto G-sheets upon simple addition. The TEM image of Pt/C showed that the higher density of NPs with such aggregation on carbon matrix (Figure 2a). Upon addition of 20% (w/w) of G-sheets, the Pt NPs were migrated onto G-sheets with aggregation-free distribution (Figure 2b). At 40% (w/w) addition, the Pt NPs were also migrated with lower amount in Figure 2c. For comparison, it is also observed that the Pt NPs were migrated with lower amount at the surface of Pt/C 1 :G 9 (Figure 2d). Therefore, as observed in Figures 1 and 2, the addition of 20% of G-sheets was optimal for the preparation of good quality of Pt/C x :G 10−x with better Pt NPs dispersion morphology.

Raman Analysis of Pt/Cx:G10−x
To investigate the internal structure of Pt/Cx:G10−x composite, Raman spectroscopy was used, which is known as an important tool to investigate the graphitic, crystalline and electronic properties of carbon-based materials [43][44][45]. Characterization of carbon-based materials can be done by comparing D and G bands while the D band represents the disordered graphitic structure with a significant level of defects and the G band is related to the E2g phonon of the ordered sp 2 -graphitic structure. Therefore, the intensity ratio of D band and G band (ID/IG) is considered to reflect the relative disorder stemming from structural imperfection of graphitic structure. In Figure  3a, the narrow D band at 1351 cm −1 and strong G band at 1599 cm −1 were observed at G-sheets spectrum and the ID/IG was calculated as 0.88 which indicating the distortion in G-sheets [46]. Additionally, ID/IG was 0.99 for pristine Pt/C, indicating sp 3 -C rich carbon black in Pt/C sample. The value was similar to the previous report [47]. After analyzing Raman data, the ID/IG was found to be increased from 0.88 (or 0.99) to 1.03 and 1.11 for Pt/C6:G4 and Pt/C8:G2, respectively, which indicated a better graphitization. Unlikely the G band of Pt/C (1597.5 cm −1 ), G bands of Pt/C8:G2 (1598 cm −1 ) and Pt/C6:G4 (1598.3 cm −1 ) were up-shifted, which indicated a better composite formation in between C and G-sheets. Moreover, the intensity of 2D band from Raman spectroscopy indicates the multilayer stacking of G-sheets. However, the intensity of the 2D band was higher at Pt/C6:G4 than that at Pt/C8:G2 in Figure 3a that indicating better dispersion of Pt/C onto few layer G-sheets in Pt/C8:G2 sample compared to multilayer assembly of G-sheets in Pt/C6:G4 sample [48]. This feature could be facilitate to the SSA and subsequent electrocatalytic activity [48,49].

Raman Analysis of Pt/C x :G 10−x
To investigate the internal structure of Pt/C x :G 10−x composite, Raman spectroscopy was used, which is known as an important tool to investigate the graphitic, crystalline and electronic properties of carbon-based materials [43][44][45]. Characterization of carbon-based materials can be done by comparing D and G bands while the D band represents the disordered graphitic structure with a significant level of defects and the G band is related to the E 2g phonon of the ordered sp 2 -graphitic structure. Therefore, the intensity ratio of D band and G band (I D /I G ) is considered to reflect the relative disorder stemming from structural imperfection of graphitic structure. In Figure 3a, the narrow D band at 1351 cm −1 and strong G band at 1599 cm −1 were observed at G-sheets spectrum and the I D /I G was calculated as 0.88 which indicating the distortion in G-sheets [46]. Additionally, I D /I G was 0.99 for pristine Pt/C, indicating sp 3 -C rich carbon black in Pt/C sample. The value was similar to the previous report [47]. After analyzing Raman data, the I D /I G was found to be increased from 0.88 (or 0.99) to 1.03 and 1.11 for Pt/C 6 :G 4 and Pt/C 8 :G 2 , respectively, which indicated a better graphitization. Unlikely the G band of Pt/C (1597.5 cm −1 ), G bands of Pt/C 8 :G 2 (1598 cm −1 ) and Pt/C 6 :G 4 (1598.3 cm −1 ) were up-shifted, which indicated a better composite formation in between C and G-sheets. Moreover, the intensity of 2D band from Raman spectroscopy indicates the multilayer stacking of G-sheets. However, the intensity of the 2D band was higher at Pt/C 6 :G 4 than that at Pt/C 8 :G 2 in Figure 3a that indicating better dispersion of Pt/C onto few layer G-sheets in Pt/C 8 :G 2 sample compared to multilayer assembly of G-sheets in Pt/C 6 :G 4 sample [48]. This feature could be facilitate to the SSA and subsequent electrocatalytic activity [48,49].

SSA Measurement of Pt/Cx:G10−x
The porosity is an important characteristic that involves with the SSA and electrocatalytic properties of carbon-based materials [50,51]. The SSA and pore distribution of as prepared composite samples has been figured out by N2 adsorption/desorption curves as can be seen in Figure  3b. The type-IV isothermal curves are confirmed the presence of mesopores in all composite samples [52]. Therefore, the SSA (trough Brunaue-Emmett-Teller method) of Pt/C8:G2 was measured as 471 m 2 g −1 , which was significantly greater than the Pt/C6:G4 (423 m 2 g −1 ) and the Pt/C (318 m 2 g −1 ). This is probably due to the few layer G-sheets assembly and the less aggregation of carbon black compared to Pt/C6:G4 and Pt/C, respectively. The pore size distribution (using the Barrett-Joyner-Halenda method, BJH) curves derived from the N2 desorption which confirming the presence of the abundant pores in Pt/C8:G2 sample with average pore diameter of 10.3 nm and a pore volume of 1.8 cm 3 g −1 (Figure 3b inset). Both values were much lower for Pt/C6:G4 and Pt/C samples. Therefore, larger SSA, higher pore volume and wider range of pore size distribution are indicating that of facile electrocatalysis at Pt/C8:G2 than those of Pt/C6:G4 and Pt/C samples.

XPS Analysis of Pt/Cx:G10−x
The X-ray photoelectron spectroscopy (XPS) analysis has been used to reach further insight into the composite formation as well as the elemental analysis. In Figure S1, mainly three peaks appeared in XPS survey spectra of Pt-containing samples at ~70 eV, ~284.1 eV and ~532.1 eV, which signify the presence of Pt, C and O elements, respectively [8,53], whereas the G-sheets XPS survey spectrum showed only C and O peaks. However, the C and O ratio (C/O, at%) was calculated as 17 for Pt/C, which was decreased to 7.6 and 6.3 for Pt/C8:G2 and Pt/C6:G4, respectively. Indicating significant amount of oxygen was incorporated in all composite samples due to addition of G-sheets [12] although, C/O of G-sheets was 6.5. Additionally, the amount of Pt was confirmed as 20, 16.5 and 14.7 wt% in Pt/C, Pt/C8:G2 and Pt/C6:G4 samples, respectively.
For a detail analysis of distribution of carbonaceous groups in all samples, the core level of C 1s The porosity is an important characteristic that involves with the SSA and electrocatalytic properties of carbon-based materials [50,51]. The SSA and pore distribution of as prepared composite samples has been figured out by N 2 adsorption/desorption curves as can be seen in Figure 3b. The type-IV isothermal curves are confirmed the presence of mesopores in all composite samples [52]. Therefore, the SSA (trough Brunaue-Emmett-Teller method) of Pt/C 8 :G 2 was measured as 471 m 2 g −1 , which was significantly greater than the Pt/C 6 :G 4 (423 m 2 g −1 ) and the Pt/C (318 m 2 g −1 ). This is probably due to the few layer G-sheets assembly and the less aggregation of carbon black compared to Pt/C 6 :G 4 and Pt/C, respectively. The pore size distribution (using the Barrett-Joyner-Halenda method, BJH) curves derived from the N 2 desorption which confirming the presence of the abundant pores in Pt/C 8 :G 2 sample with average pore diameter of 10.3 nm and a pore volume of 1.8 cm 3 g −1 (Figure 3b inset). Both values were much lower for Pt/C 6 :G 4 and Pt/C samples. Therefore, larger SSA, higher pore volume and wider range of pore size distribution are indicating that of facile electrocatalysis at Pt/C 8 :G 2 than those of Pt/C 6 :G 4 and Pt/C samples.

XPS Analysis of Pt/C x :G 10−x
The X-ray photoelectron spectroscopy (XPS) analysis has been used to reach further insight into the composite formation as well as the elemental analysis. In Figure S1, mainly three peaks appeared in XPS survey spectra of Pt-containing samples at~70 eV,~284.1 eV and~532.1 eV, which signify the presence of Pt, C and O elements, respectively [8,53], whereas the G-sheets XPS survey spectrum showed only C and O peaks. However, the C and O ratio (C/O, at%) was calculated as 17 for Pt/C, which was decreased to 7.6 and 6.3 for Pt/C 8 :G 2 and Pt/C 6 :G 4 , respectively. Indicating significant amount of oxygen was incorporated in all composite samples due to addition of G-sheets [12] although, Processes 2019, 7, 586 6 of 14 C/O of G-sheets was 6.5. Additionally, the amount of Pt was confirmed as 20, 16.5 and 14.7 wt% in Pt/C, Pt/C 8 :G 2 and Pt/C 6 :G 4 samples, respectively.
For a detail analysis of distribution of carbonaceous groups in all samples, the core level of C 1s XPS spectra for Pt/C, Pt/C 8 :G 2 , Pt/C 6 :G 4 and G-sheets are presented in Figures 2c and S2. The C 1s was deconvoluted by five different components ( Figure S2) assigned to the oxygen-free sp 2 -C (C=C) at 285.2 eV and oxygen containing sp 3 -C, such as C-O at 285.7 eV, C=O at 287.4 eV and O-C=O at 289.2 eV with a π-π* at 291.2 eV [15,43,54]. However, the C 1s spectrum of Pt/C shows only a sharp peak for sp 2 -C, which signifies that a smaller amount of oxygen-containing species were presence in this sample. Compared to Pt/C, the high intensity peak appeared at other C 1s XPS spectra for sp 3 -C species, which indicates that oxygen-species were gradually increased due to the gradual addition of G-sheets. Although the sp 2 -C peak intensity showed a similar amount of sp 2 -C in all samples, the detected sp 2 -C species was higher in Pt/C 8 :G 2 sample, as shown in the pie chart in Figure 2c inset. This pie chart describes the percentage of sp 2 -C species was higher in Pt/C 8 :G 2 sample among all samples possibly due to better reduction with limited addition of G-sheets. Finally, the comparative Pt 4f XPS spectra for Pt/C, Pt/C 8 :G 2 and Pt/C 6 :G 4 are shown in Figures 3d and S3. In Figure 3d, all Pt 4f spectra present a doublet consisting with two bands at 72.3 and 75.5 eV, which were assigned to the Pt 4f 7/2 and Pt 4f 5/2 , respectively. The distribution of Pt metallic states in those samples was displayed in Figure  S3, which confirms to that all Pt in those samples were mostly in Pt 0 state. However, both Pt 4f 7/2 and Pt 4f 5/2 peak positions of those samples were in the same position even upon addition of G-sheet, indicating that all PtNPs in those samples were in the same electronic environment and no significant change was observed due to G-sheet addition, except reducing total Pt content.

Electrochemical Analysis of Pt/C x :G 10−x
The Pt/C x :G 10−x catalysts were placed onto clean GCE with different Pt-mass loading, the electrochemical measurements were then carried out by CV technique in Ar-saturated 0.1 M HClO 4 solution as displayed in Figure 4a. The Figure 4a exhibitions CV curves which recorded on Pt/C, Pt/C 8 :G 2 and Pt/C 6 :G 4 , respectively, in the potential range of 1.2 to 0 V (versus RHE) at a fixed scan rate of 50 mV s −1 . The PtO reduction peak located around 0.7 V and the H 2 adsorption/desorption peaks at the range of 0.4 to 0 V in the backward CV scan was clearly appeared in all curves. The electrochemical surface area (ECSA) of all Pt-containing catalysts was derived from the Coulombic charge (Q) that are associated with H 2 adsorption-desorption on Pt-site using Equation (1) [55,56]. The ECSA of Pt-containing catalysts was calculated as 63, 77 and 65 m 2 g −1 Pt , while the Pt-loading was 28.3, 23.3 and 20.8 µg Pt cm −2 for the Pt/C, Pt/C 8 :G 2 and Pt/C 6 :G 4 , respectively. The obtained ECSA for Pt/C 8 :G 2 was significantly higher than those of Pt/C and Pt/C 6 :G 4 , which indicates that the Pt-site in Pt/C 8 :G 2 was much exposed and electroactive. It is well known that the ECSA value is an important indication of the precious metals utilization with minimal metal loading. Therefore, this result signifies a better utilization of active Pt-sites in Pt/C 8 :G 2 catalyst probably due to the higher SSA, better dispersion of Pt NPs with less dissolution nature.
where Q o is the charge associated with the H 2 adsorption/desorption region, Q r is the charge required for the monolayer adsorption of H 2 on Pt surface (assumed to be 210 µC cm −2 ) [55,56] and m is the Pt-mass loading (in µg cm −2 ).

Electrochemical ORR on Pt/Cx:G10−x
For exploring the catalytic activity of the Pt/Cx:G10−x composite catalysts for electrochemical ORR, the LSV curves on RRDE were then recorded for Pt/C, Pt/C8:G2, Pt/C6:G4 and G-sheets electrodes in O2-saturated 0.1 M HClO4 solution at a scan rate of 5 mV s -1 and at a constant rotation speed (1600 rpm), which are shown in Figure 4c. As can be seen in Figure 4c, the onset potential (Eonset) of ORR for the Pt/C electrode was at 0.95 V (versus RHE) and the limiting diffusion current density (jL) was 4.47 mA cm -2 at 0.4 V (versus RHE). Upon the addition of G-sheets, the ORR Eonset shifted positively to 1.01 V, which was accompanied by a concomitant increase in the ORR jL up to 4.8 mA cm -2 at 0.4 V (versus RHE) on Pt/C8:G2 electrode, indicating an improved ORR catalysis on Pt/C8:G2 electrode in terms of both parameters. However, the excess amount of G-sheets addition the ORR activity decreased in respect to the Eonset and jL on Pt/C6:G4 electrode (0.95 V and 4.24 mA cm -2 ) and both parameters were tremendously decreased on G-sheets electrode (0.9 V and 3 mA cm -2 ) (Eonset confirmed from Figure S4). These results clearly demonstrated a significant enhancement in the ORR electrocatalytic activity for the Pt/C8:G2 than not only the Pt/C6:G4 and G-sheets electrodes but also the pure benchmark Pt/C electrode.
To compare the electrocatalytic activity towards ORR against Pt-mass, the Pt-mass normalized kinetic current density (jk) was calculated at 0.9 V (versus RHE) in Figure 4c inset. The Pt/C8:G2 catalyst provided the highest MA (252 mA mg −1 ) towards ORR which was 3.3 and 1.9 times greater than those of Pt/C (76.5 mA g −1 ) and Pt/C6:G4 (135.6 mA g −1 ), respectively. This is consistent with the significantly higher specific activity (SA, ECSA normalized jk) at 0.9 V (versus RHE) for ORR at Pt/C8:G2 (0.27 mA −2 ) than those of Pt/C (0.12 mA −2 ) and Pt/C6:G4 (0.15 mA −2 ) electrodes as shown in Figure S5. In addition, LSV curves of ring electrode on RRDE for Pt/C, Pt/C8:G2, Pt/C6:G4 and G-sheets electrodes have been compared in Figure 4d. The ring current density was produced from the generation of H2O2 during ORR [57,58]. As can be seen, all electrodes The electrochemical stability of catalysts under practical operating condition is a key factor for commercialization and extensive use of FCs. The CV experiment was performed for evaluating longterm ECSA stability and the variation of ECSA has shown in Figure 4b. In Figure 4b, for Pt/C 8 :G 2 , the initial value was 77 m 2 g −1 Pt , which increased to 81 and 79 m 2 g −1 Pt in the first 200 and 500 cycles, respectively, and finally decreased to 76 m 2 g −1 Pt after 1000 cycles. The ECSA of Pt/C dropped at 60 m 2 g −1 Pt from its initial ECSA (63 m 2 g −1 Pt ) after 200 cycles, and the final ECSA after 1000 cycles dropped to 52 m 2 g −1 Pt . Similarly, the ECSA of Pt/C 6 :G 4 dropped at 54 m 2 g −1 Pt after 1000 cycles from 65 m 2 g −1 Pt as this was its initial ECSA. Thus, the Pt/C 8 :G 2 demonstrated a good ECSA stability up to the 1000 CV cycles which was significantly higher than the benchmark carbon supported Pt/C.

Electrochemical ORR on Pt/C x :G 10−x
For exploring the catalytic activity of the Pt/C x :G 10−x composite catalysts for electrochemical ORR, the LSV curves on RRDE were then recorded for Pt/C, Pt/C 8 :G 2 , Pt/C 6 :G 4 and G-sheets electrodes in O 2 -saturated 0.1 M HClO 4 solution at a scan rate of 5 mV s −1 and at a constant rotation speed (1600 rpm), which are shown in Figure 4c. As can be seen in Figure 4c, the onset potential (E onset ) of ORR for the Pt/C electrode was at 0.95 V (versus RHE) and the limiting diffusion current density (j L ) was 4.47 mA cm −2 at 0.4 V (versus RHE). Upon the addition of G-sheets, the ORR E onset shifted positively to 1.01 V, which was accompanied by a concomitant increase in the ORR j L up to 4.8 mA cm −2 at 0.4 V (versus RHE) on Pt/C 8 :G 2 electrode, indicating an improved ORR catalysis on Pt/C 8 :G 2 electrode in terms of both parameters. However, the excess amount of G-sheets addition the ORR activity decreased in respect to the E onset and j L on Pt/C 6 :G 4 electrode (0.95 V and 4.24 mA cm −2 ) and both parameters were tremendously decreased on G-sheets electrode (0.9 V and 3 mA cm −2 ) (E onset  Figure S4). These results clearly demonstrated a significant enhancement in the ORR electrocatalytic activity for the Pt/C 8 :G 2 than not only the Pt/C 6 :G 4 and G-sheets electrodes but also the pure benchmark Pt/C electrode.
To compare the electrocatalytic activity towards ORR against Pt-mass, the Pt-mass normalized kinetic current density (j k ) was calculated at 0.9 V (versus RHE) in Figure 4c inset. The Pt/C 8 :G 2 catalyst provided the highest MA (252 mA mg −1 Pt ) towards ORR which was 3.3 and 1.9 times greater than those of Pt/C (76.5 mA µg −1 Pt ) and Pt/C 6 :G 4 (135.6 mA µg −1 Pt ), respectively. This is consistent with the significantly higher specific activity (SA, ECSA normalized j k ) at 0.9 V (versus RHE) for ORR at Pt/C 8 :G 2 (0.27 mA cm −2 ECSA ) than those of Pt/C (0.12 mA cm −2 ECSA ) and Pt/C 6 :G 4 (0.15 mA cm −2 ECSA ) electrodes as shown in Figure S5. In addition, LSV curves of ring electrode on RRDE for Pt/C, Pt/C 8 :G 2 , Pt/C 6 :G 4 and G-sheets electrodes have been compared in Figure 4d. The ring current density was produced from the generation of H 2 O 2 during ORR [57,58]. As can be seen, all electrodes started to generate the ring current at the E onset of ORR and the intensity of steady state current region indicated the degree of H 2 O 2 generation. However, on the Pt/C 8 :G 2 electrode, the H 2 O 2 generation was significantly less than that on other electrodes which is favorable for ORR process, indicating that Pt/C 8 :G 2 was comparatively better electrocatalyst towards ORR. Therefore, the electrochemical results show that the Pt/C 8 :G 2 was efficient electrocatalyst than those of other tested electrodes in respect to the E onset , MA, and SA towards ORR because of the higher porosity, larger SSA and improved ECSA.

Kinetics of ORR on Pt/C x :G 10−x
It is known that the ORR in aqueous solution occurs mainly by two pathways: the direct four-electron reduction pathway from O 2 to H 2 O and the two-electron reduction pathway from O 2 to H 2 O 2 [59]. However, to evaluate the ORR pathway on Pt/C, Pt/C 8 :G 2 , Pt/C 6 :G 4 and G-sheets electrodes, the RRDE data was then used for the calculation of transferred electron numbers per O 2 (n) and the corresponding H 2 O 2 formation on aforementioned electrodes during the ORR process from Equations (1) and (2) [60,61] in Figure 5a. As shown in Figure 5a, the n value was found to be dependent on the potential for all electrodes. Initially, the n value rapidly increased afterwards slowly increased with the subsequent potential range. However, the n for ORR at the Pt/C 8 :G 2 electrode (3.95 to 4.02) was always higher than those on commercial Pt/C (3.85 to 3.92), Pt/C 6 :G 4 (3.83 to 3.89) and G-sheets (3 to 3.6) at the potential range of 0.8 V to 0.2 V. Therefore, the ORR proceeded via a four-electron pathway at Pt/C 8 :G 2 electrode which was even better than those of other tested electrodes in this study [62]. This is consistent with the relatively lower H 2 O 2 production during ORR at the Pt/C 8 :G 2 electrode with compared to those electrodes ( Figure 5b). Therefore, it has been confirmed that the Pt/C 8 :G 2 composite electrode was comparatively better ORR electrocatalyst than Pt/C and other tested electrodes in respect to the higher number of electron transfer per O 2 as well as a lower production of H 2 O 2 .
where i d and i r are the disk and ring electrode currents, respectively, and N is the collection efficiency of the RRDE (0.37), as calculated from Equation (3).
all catalysts in Figure 5d. As shown in Figure 5d, the Tafel slope for Pt/C8:G2 electrode was 63 mV dec -1 at a low overpotential which was very close to the theoretical value and Pt/C (62 mV dec -1 ), and obviously better than the Pt/C6:G4 (72 mV dec -1 ) and G-sheets (91 mV dec -1 ). The result also indicated a better ORR catalysis on Pt/C8:G2 catalyst with improved ORR mechanism and faster electron transfer.

Stability and Selectivity Towards ORR on Pt/Cx:G10−x
The Pt/C8:G2, Pt/C6:G4 and Pt/C electrodes were further subjected to investigate the stability and selectivity toward ORR which is regarded as the key factor for the widespread commercialization of FCs. The stability of those electrodes was evaluated during ORR through CA technique at 0.8 V (RHE) in an O2-saturated 0.1 M HClO4 at a rotation rate of 1600 rpm in Figure 6a. As depicted in Figure 6a, the relative current from all electrodes decreased with time and the Pt/C8:G2 electrode showed a slower decrease at ~25% after continuous reaction for 30,000 s, whereas the Pt/C6:G4 and Pt/C clearly decreased to ~48% and ~45%, respectively, at the same time, indicating that the Pt/C8:G2 electrocatalyst was approximately two times higher stable than the commercial Pt/C and Pt/C6:G4 electrodes. Additionally, the possible crossover effect was also investigated in presence of anodic fuel molecules (e.g., ethanol, methanol) during ORR at the Pt/C, Pt/C6:G4 and Pt/C8:G2 electrodes as in Figure 6b. As described in Figure 6b, a sharp response and huge decease in relative current was observed in CA curve for the Pt/C electrode and considerably lower decease in relative current was observed in CA curve for the Pt/C6:G4 electrode upon addition of 2 M ethanol and methanol, respectively. However, the CA curve for the Pt/C8:G2 electrode was slightly changed in relative current upon addition of ethanol and methanol, indicating that the Pt/C8:G2 electrocatalyst was better selective toward ORR and resistive to the crossover effect than the commercial Pt/C and Pt/C6:G4 electrocatalysts. To gain further insight into the ORR electrocatalysis on Pt/C x :G 10−x and G-sheets electrodes, the reaction kinetics were investigated by RDE at various rotation speeds in same electrolyte ( Figure S6). In Figure S6, the LSV curves showed that the limiting current density (j) of ORR was increased by the increasing rotating speed in all cases. At a constant rotation rate, upon controlled addition of G-sheets with Pt/C, the j of ORR at the Pt/C 8 :G 2 electrode was always higher than those of other tested electrodes, indicating an improved ORR process at Pt/C 8 :G 2 electrode. However, the Koutecky-Levich (K-L) plots resultant from the corresponding LSV curves showed the linear relationships between j −1 L and angular velocity (ω −1/2 , ω = 2π* rpm) for all electrodes at various potentials (0.6 V to 0.2 V versus RHE). Particularly, the K-L plots for Pt/C 8 :G 2 electrode ( Figure S6) maintained nearly same slopes over the potential range was studied in respect to the K-L plots for other tested electrodes, indicating a consistent number of electron transferred per O 2 molecule during ORR and comparatively better first-order reaction kinetic ORR possess on Pt/C 8 :G 2 electrode than Pt/C, Pt/C 6 :G 4 and G-sheets electrodes [63,64]. However, for a better understanding, the comparison of K-L plots and n-values (calculated from K-L slopes) at 0.4 V (versus RHE) are also presented for all catalysts in Figure 5c. As can be seen in Figure 5c, all K-L plots were parallel and near to the dotted line that represented to the theoretical value of four-electron involved ORR. Particularly, the K-L plot of Pt/C 8 :G 2 electrode was much betted parallel to the four-electron line, indicating a higher n-value towards ORR at any rotation speed. The n-value for ORR at the Pt/C 8 :G 2 electrode (4.02) was higher than those on Pt/C (3.91), Pt/C 6 :G 4 (3.88) and G-sheets (3.5) at 0.4 V (versus RHE) (Figure 5c inset). These values were concomitant to the previous calculation.
To analyze the Tafel behavior, the Tafel plots were derived for further kinetic investigation for all catalysts in Figure 5d. As shown in Figure 5d, the Tafel slope for Pt/C 8 :G 2 electrode was 63 mV dec −1 at a low overpotential which was very close to the theoretical value and Pt/C (62 mV dec −1 ), and obviously better than the Pt/C 6 :G 4 (72 mV dec −1 ) and G-sheets (91 mV dec −1 ). The result also indicated a better ORR catalysis on Pt/C 8 :G 2 catalyst with improved ORR mechanism and faster electron transfer.

Stability and Selectivity Towards ORR on Pt/C x :G 10−x
The Pt/C 8 :G 2 , Pt/C 6 :G 4 and Pt/C electrodes were further subjected to investigate the stability and selectivity toward ORR which is regarded as the key factor for the widespread commercialization of FCs. The stability of those electrodes was evaluated during ORR through CA technique at 0.8 V (RHE) in an O 2 -saturated 0.1 M HClO 4 at a rotation rate of 1600 rpm in Figure 6a. As depicted in Figure 6a, the relative current from all electrodes decreased with time and the Pt/C 8 :G 2 electrode showed a slower decrease at~25% after continuous reaction for 30,000 s, whereas the Pt/C 6 :G 4 and Pt/C clearly decreased to~48% and~45%, respectively, at the same time, indicating that the Pt/C 8 :G 2 electrocatalyst was approximately two times higher stable than the commercial Pt/C and Pt/C 6 :G 4 electrodes. Additionally, the possible crossover effect was also investigated in presence of anodic fuel molecules (e.g., ethanol, methanol) during ORR at the Pt/C, Pt/C 6 :G 4 and Pt/C 8 :G 2 electrodes as in Figure 6b. As described in Figure 6b, a sharp response and huge decease in relative current was observed in CA curve for the Pt/C electrode and considerably lower decease in relative current was observed in CA curve for the Pt/C 6 :G 4 electrode upon addition of 2 M ethanol and methanol, respectively. However, the CA curve for the Pt/C 8 :G 2 electrode was slightly changed in relative current upon addition of ethanol and methanol, indicating that the Pt/C 8 :G 2 electrocatalyst was better selective toward ORR and resistive to the crossover effect than the commercial Pt/C and Pt/C 6 :G 4 electrocatalysts. Finally, the morphology of Pt/C (Figure 6c) and Pt/C8:G2 (Figure 6d) was evaluated by TEM analysis after a stability test via CA. As can be seen, Pt NPs were found to be detached from the carbon support with higher agglomeration in Pt/C catalyst (Figure 6c) as a result of severe dissolution and corrosion of carbon support [65]. On the other hand, the Pt/C8:G2 composite catalyst achieved a more uniform Pt NPs distribution after the stability test and Pt NPs were not detached from the graphene and carbon support (Figures 6d and S7). Therefore, all described results confirmed that the Pt/C8:G2 composite catalyst is much efficient ORR catalyst in terms of better electrocatalysis, stability and fuel selectivity among Pt/Cx:G10−x including Pt/C.

Conclusions
Finally, the Pt/C with G-sheets composite could act as an efficient ORR electrocatalyst, which was prepared through a simple composite making. The optimum ration of carbon and graphene composite exhibited stable ECSA with minimal Pt-loading which resultant from higher SSA and aggregation-free dispersion of Pt NPs. As a result, Pt/C8:G2 catalyst showed superior activity toward ORR in respect to the not only high Eonset, limiting current, MA and SA but also high stability and resistance to the crossover effect. Notably, the Pt/C8:G2 composite catalyst showed significantly better ORR kinetics with lowering H2O2 formation and higher long-term operational stability than that of benchmark Pt/C and other composite electrodes. Finally, the morphology of Pt/C (Figure 6c) and Pt/C 8 :G 2 (Figure 6d) was evaluated by TEM analysis after a stability test via CA. As can be seen, Pt NPs were found to be detached from the carbon support with higher agglomeration in Pt/C catalyst (Figure 6c) as a result of severe dissolution and corrosion of carbon support [65]. On the other hand, the Pt/C 8 :G 2 composite catalyst achieved a more uniform Pt NPs distribution after the stability test and Pt NPs were not detached from the graphene and carbon support (Figures 6d and S7). Therefore, all described results confirmed that the Pt/C 8 :G 2 composite catalyst is much efficient ORR catalyst in terms of better electrocatalysis, stability and fuel selectivity among Pt/C x :G 10−x including Pt/C.

Conclusions
Finally, the Pt/C with G-sheets composite could act as an efficient ORR electrocatalyst, which was prepared through a simple composite making. The optimum ration of carbon and graphene composite exhibited stable ECSA with minimal Pt-loading which resultant from higher SSA and aggregation-free dispersion of Pt NPs. As a result, Pt/C 8 :G 2 catalyst showed superior activity toward ORR in respect to the not only high E onset , limiting current, MA and SA but also high stability and resistance to the crossover effect. Notably, the Pt/C 8 :G 2 composite catalyst showed significantly better ORR kinetics with lowering H 2 O 2 formation and higher long-term operational stability than that of benchmark Pt/C and other composite electrodes.

Conflicts of Interest:
The authors declare no conflict of interest.