Pt Nanoparticles Loaded on W18O49 Nanocables–rGO Nanocomposite as a Highly Active and Durable Catalyst for Methanol Electro-Oxidation

Highly active and durable electrocatalysts are vital for commercialization of direct methanol fuel cells. In this work, a three-dimensional nanocomposite consisting of platinum nanoparticles, W18O49 nanocables, and reduced graphene oxide composite (Pt/W18O49 NCs–rGO) has been prepared as an electrocatalyst for methanol oxidation reaction (MOR). The catalyst is prepared through a two-step method. The W18O49 nanocables and the reduced graphene oxide composite are prepared by a solvothermal process. Then, Pt nanoparticles are loaded on the W18O49 nanocables and the reduced graphene oxide composite by a hydrogen reduction at ambient condition. The obtained catalyst has a special three-dimensional architecture consisting of two-dimensional nanosheets, assembled one-dimensional nanocables, and the loaded nanoparticles on their surface. The Pt/W18O49 NCs–rGO catalyst shows 1.56 time mass activities than the Pt/C, with the current density of the forward anodic peak reaching 1624 mA/mgPt at 0.854 V versus reversible hydrogen electrode potential in 0.1 M HClO4 and 0.5 M CH3OH mixed electrolyte. It also shows a strong antipoisoning property toward CO. For the durability testing, the current density of Pt/W18O49 NCs–rGO shows a 37% decay, whereas the current of Pt/C catalyst shows a 41% degradation from 600 to 3600 s at 0.7 V. The high activity toward MOR, good antipoisoning for intermediate products, and excellent stability are ascribed to strong metal–support interaction effects between the Pt nanoparticles and the W18O49 NCs.


INTRODUCTION
Platinum-based catalysts were used in both electrodes of direct methanol fuel cells (DMFCs). Preparing platinum nanoscale catalysts with special morphologies, such as platinum nanoparticles, 1−4 platinum nanonetwork, 5 and other designed Pt nanostructures, 6−10 enable to improve the activity of platinum. But, it will lose active sites during working, which results from poisoning of intermediate species during anode reaction process in DMFCs. 11−14 To overcome this problem, various metal elements, such as Pd, Ni, and others were used for incorporation into platinum to prepare alloy catalysts. 15−17 It not only reduces the amount of platinum but also improves the ability of CO tolerance via a bifunctional mechanism. 18−20 Unfortunately, many alloy catalysts are not stable at the inclement circumstance in DMFCs. Great efforts have been made to enhance the stability; some compositions are easy to dissolve from the reaction interface of the alloy catalyst. 21−25 Transition-metal oxides as a support composite with noblemetal catalysts have attracted much attention due to their high activity, stability, special microstructure and morphology, and other properties. 26−29 In previous works, our group found that platinum and transition-metal oxide composites, such as Pt/ Mo 4 O 11 30 and Pt/W 18 O 49 , 31 can improve both anti-COpoisoning ability and stability in harsh circumstance. It was ascribed to the contribution of a strong metal−support interaction (SMSI). A reversible reaction between WO x and tungsten bronze compound (H x WO y ) is an important medium in the methanol electro-oxidation reaction process. 31 It is helpful for the small molecules adsorption on the Pt/W 18 O 49 catalyst surface. Meanwhile, the oxygen defect of W 18 O 49 is beneficial for oxygen transport from catalyst to reaction interface. But, the low conductivity between the reaction interfaces limits the rate of electrocatalytic reaction. 31 Although we can find some transition-metal oxides with a better conductivity, such as Mo 4 31,33 sometime, it is still not good enough for the electrocatalytic reactions in DMFCs. The conductivity of the catalysts still needs to be improved.
In this work, platinum nanoparticles and W 18 O 49 nanocables loaded on reduced graphene oxide nanosheet composite (Pt/ W 18 O 49 NCs−rGO) as a new three-dimensional (3D) nanostructured catalyst for methanol electro-oxidation was designed and prepared successfully via a self-assembly method. The methanol oxidation reaction (MOR) and CO stripping tests were used to compare the electrocatalytic activities of the as-synthesized catalyst and commercial Pt/C catalyst. The Pt/ W 18 O 49 NCs−rGO shows better activity, durability, and anti-CO-poisoning ability than the commercial Pt/C catalyst, which can be attributed to the highly dispersed platinum nanoparticles and the SMSI between Pt, W 18 O 49 , and rGO. Various characterizations provide evidence of the mechanism of enhanced catalytic properties. The present work not only provides a promise catalyst for methanol electrocatalytic oxidation but also presents an approach for designing various multifunctional materials.  Figure 1a shows that the as-sythesized product has a special 3D architecture consisting of two-dimensional nanosheets rGO, assembled one-dimensional nanocables (W 18 O 49 NCs), and highly dispersed Pt nanoparticles loaded on their surface. Figure 1b

ACS Omega
Article rGO composite. The surface of the Pt nanoparticle is composed of alternating the {100}, {111} facets and kinks at the corners, which normally are the active sites for catalytic reactions. Figure 1f shows the particle size distribution statistics of Pt nanoparticles, with a mean size of 2.71 ± 0.39 nm. It was reported that platinum nanoparticles with 2−4 nm size exhibit the highest activity for electrocatalysis. 36  It is noted that there are defects in W 18 O 49 NCs as pointed by the two arrows ( Figure S4a). From the enlarged part of the defect region ( Figure S4b), the shear plane marked by arrows is the main crystal structure of the Magneĺi phase. Figure S4c presents another part of the same crystal, showing relatively perfect crystal planes of (010) and (103). Figure S4d shows the computer-simulated atom sites, further indicating that the W 18 O 49 NCs grow along the 010 direction.
Raman spectroscopy and thermogravimetry (TG) analysis were introduced for further characterization on the W 18  For the spectrum of graphene oxide, two broad peaks at about 1334 and 1596 cm −1 can be assigned to the D and G peaks of graphene nanosheets. The D band is stronger than the G band in the spectrum of graphene oxide, which can be ascribed to a disordered structure of the surface oxygenous groups. 39,40 But the intensity ratio of I G /I D does not show an obvious change after the reduction. It could be ascribed to the disordered structure of rGO, which is induced by interactions between W 18 Figure S6, a mass loss is observed before 200°C for both samples, which may be due to dehydration. Another mass loss at 561°C for W 18 O 49 NCs−rGO may be due to the carbon oxidation reaction. The latter mass loss is actually a combination process of graphene oxidation and the oxidation of W 18 31 The catalytic activity of Pt is normally determined by zerovalance Pt 0 and its oxides (PtO and/or PtO 2 ). The XPS spectra of Pt 4f in Pt/W 18 O 49 NCs−rGO and Pt/C were deconvoluted to reveal the oxidation species ( Figure S7b). The binding energy at 71.3, 72.2, and 75.3 eV for the Pt 4f 7/2 peak, together with three corresponding subpeaks at 74.6, 75.5, and 78.6 eV for Pt 4f 5/2 peak in Pt/W 18 O 49 NCs−rGO corresponded to Pt 0 , PtO, and PtO 2 , respectively. 43,44 The peaks positions of Pt in Pt/W 18 O 49 NCs−rGO are lower than those of the Pt/C. For example, the peak position of zerovalence Pt 0 in the Pt/W 18 O 49 NCs−rGO is left-shifted about 0.2 eV compared to that of Pt 0 of the Pt/C due to a strong metal−support interaction effect. 31,45,46 The actual Pt species loading was obtained by integrating the peak emission line (Table S2). The percent ratio of Pt with the higher valence states in the Pt/W 18 O 49 NCs−rGO (57.76%) is equivalent to those in the Pt/C (57.34%). Normally, the high valence states of Pt show high catalytic activity toward oxidation.   49 NCs act as a bridge to connect Pt nanoparticles and graphene nanosheets. This connection improves the conductivity of the catalyst and reduces the ohmic drop of the electrode. In Figure S6b and  O 49 NCs causes the formation of the tungsten bronze (H x WO y ) with a good proton conductivity. Hydrogen spillover effect would promote the dehydrogenation of CH 3 OH. Second, the number and efficiency of the active sites are increased due to the synergistic effect because W 18 O 49 NCs act as a oxygen reservoir, which may provide more O ads species on the Pt surface. 31 In previous works, it has been reported that the anodic peak in the reverse scan is related to the removal of the intermediate species absorbed on the catalyst surface, e.g., CO and HCOO − . 31,45,47−51 These species occupy the active site of Pt, leading to decreasing catalytic activity and energy transformation efficiency. Consequently, the ratio of the forward anodic peak current density (I f ) to the backward anodic peak current density (I b ) can be also used to indicate the tolerance of the catalyst toward intermediate products species. Usually, a catalyst with a high ratio of I f /I b shows a good tolerance toward intermediate products. It can be clearly found that the ratios of I f /I b for Pt/W 18 O 49 NCs−rGO and Pt/ C are 1.16 and 0.75, respectively, as shown in Figure 3b. The I f /I b ratio of Pt/W 18 O 49 NCs−rGO is 1.55 times than that of the Pt/C catalyst.
To further examine the anti-methanol poisoning property, CO stripping voltammetry was tested as a function of CO poisoning time in 0.1 M HClO 4 at ambient condition at a sweep rate of 50 mV/s. A CO oxidation peak is found at 0.822

ACS Omega
Article V (vs reversible hydrogen electrode (RHE)) for the commercial Pt/C after purging with pure CO for 0.1−5 s, and the equilibrium CO coverage reaches within 0.1 s ( Figure  4a). For Pt/W 18 O 49 NCs−rGO, the CO oxidation peak is observed at about 0.743 V (vs RHE) after purging with pure CO for 1−30 s, and even after 30 s, the CO coverage does not reach an equilibrium (Figure 4b). Pt/W 18 O 49 NCs−rGO exhibits a stronger CO tolerance ability compared with the Pt/ C. It is attributed to the SMSI between Pt and composite support. As a poisoning molecular, CO is easily adsorbed and covered at the active Pt sites and blocks the catalytic reaction. The interactions between Pt and the W 18  The morphology of the catalysts also has a great influence on the activity of the reaction. Nanoparticles tend to agglomerate spontaneously by Ostwald ripening. And, it is a thermodynamically favorable process. To gain insight into the degradation mechanism of the Pt/C and Pt/W 18 O 49 NCs− rGO composite activity, Pt/W 18 O 49 NCs−rGO and Pt/C samples were prepared and observed by TEM before and after chronoamperometry test, respectively. As shown in the TEM image ( Figure S8a,c), the Pt NPs are uniformly dispersed on the substrates in both samples initially. Mean sizes of Pt NPs are 3.25 ± 0.32 and 2.18 ± 0.29 nm in Pt/W 18 O 49 NCs−rGO and Pt/C, respectively. After the test, it is observed that some platinum nanoparticles dropped from carbon support due to the electrochemical corrosion ( Figure S8b). Meanwhile, some Pt NPs become agglomerated particles with a mean size of 7.25 ± 0.89 nm. The agglomeration of Pt due to Ostwald ripening reduces the specific area of Pt NPs. It leads to a decrease in the current density and the activity loss of the Pt/C catalyst. No significant difference was observed for the Pt/W 18 O 49 NCs− rGO catalyst before and after the test. Although some larger nanoparticles with a diameter around 6−9 nm were observed after the test, the size of most Pt NPs in Pt/W 18 O 49 NCs−rGO after the test is 3.30 ± 1.20 nm. So, it maintains a high current density during the chronoamperometry test and exhibits a high stability. The high catalytic activity and excellent stability of the Pt/W 18 O 49 NCs−rGO are ascribed to the following two possible reasons. First, there are strong SMSI effects between platinum and W 18 O 49 NCs. If no hydrogen gas is introduced into the suspension system consisting of W 18 O 49 NCs−rGO and K 2 PtCl 4 during synthesis, we can still observe the Pt nanoparticles marked by blue circles deposited on the surface of W 18 O 49 NCs, the (010) plane was marked 0.378 nm, as shown in Figure S9. Pt NPs are difficult to oxidize due to the redox property of W 18 O 49 NCs. Second, Pt NPs are well dispersed and separated by the nanocables in Pt/W 18 O 49 NCs−rGO. This increases the barrier of the movement of the Pt NPs and reduces the possibility of the collision of the particles and agglomeration. Ostwald ripening is a thermodynamically spontaneous process, the barrier is favorable for preventing them from agglomeration. Moreover, the 3D structure is also very stable because there is still strong connection between W 18 Figure S10 shows the electrochemical impedance spectroscopy (EIS) spectra of commercial Pt/C and Pt/ W 18 O 49 NCs−rGO. It can be seen that the Pt/W 18 O 49 NCs− rGO shows a smaller charge transfer resistance compared to that of the commercial Pt/C catalyst. This combination of Pt NPs, W 18 O 49 NCs, and graphene provides more reactive centers for MOR and enhances the antipoisoning property toward intermediate products. The unique structure of the composite consisting of two-dimensional nanosheet and onedimensional nanocable is favorable for enhancing the stability of the materials. The well-dispersed morphology is attributed to the interaction between tungsten oxide and graphene oxide.

Article
This work may provide a new approach to prepare 3D nanostructured composite catalysts.

CONCLUSIONS
In summary, we prepared a 3D nanostructured composite Pt/ W 18 O 49 NCs−rGO as an efficient and durable catalyst for methanol oxidation reaction. Graphene oxide nanosheet was reduced to graphene through a solvothermal process. As a conductive support material, W 18 O 49 NCs−rGO nanocomposite were characterized by X-ray powder diffraction (XRD) (D8 Advance, Bruker, Deutschland). Raman spectra were measured with a micro-Raman spectrometer (HR 800, Jobin Yvon, France) at 457.9 nm wavelength. X-ray photoelectron spectroscopy (XPS) was performed with Al Kα radiation (Kratos AXIS Ultra DLD, Britain). The thermal analysis was measured by thermogravimeter (TG/differential thermal analysis (DTA), 6300, SEIKO, Tokyo, Japan). Scanning electron microscopy (SEM) was performed on a scanning electron microscope (Hitachi S-4300, Japan). Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) were performed on a microscope (JEOL JEM2100F, Japan) operated at 200 keV. The compositions of the catalysts were analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES, Thermo Fisher).
4.6. Electrochemical Measurements. Glassy carbon electrodes were polished with alumina powder, ultrasonically washed, and blow dried before being dropped into the catalyst ink. The electrode preparation and electrochemical measurements are similar to that reported in the literature previously. 31 The same procedures were used to prepare Pt/C electrode (Hispec3000, Johnson Matthey).
MOR tests were performed in a 0.1 M HClO 4 and 0.5 M CH 3 OH solution at a scan rate of 100 mV/s. Chronoamperometry measurements were carried out in a 0.1 M HClO 4 and 0.5 M CH 3 OH solution after a 50 cycle CV activity in Arpurged 0.1 M HClO 4 solution at a scan rate of 100 mV/s. For details, please refer to the literatures. 31 For CO stripping voltammetry, the procedures are similar to those reported in the literature previously. 31 For comparison, the potentials measured were calculated to the values (vs RHE) automatically by the testing software. Electrochemical impedance spectroscopy (EIS) measurements were performed on an Autolab electrochemical workstation (PGSTAT302N) in the frequency range from 100 kHz to 100 mHz with a bias voltage of 10 mV at 0.6 V (vs RHE). The solution for the EIS experiment is 0.1 M HClO 4 with 0.5 M methanol, which was purged by argon gas.

ACS Omega
Article Author Contributions ∇ Y.W. and S.W. contributed equally to this work.

Notes
The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS
Authors thank the support of the National Natural Science Foundation of China (NSFC 51472009, 51172007, 51372271, and 51672029) and acknowledge the support of Beijing municipal high-level innovative team building program (IDHT20170502). The large-scale instrument and equipment sharing forum of Beijing University of Technology (BJUT) is also duly acknowledged for providing excellent analysis conditions and support. H.Z. acknowledges the support from the U.S. Department of Commerce, National Institute of Standards and Technology under the financial assistance award 70NANB17H249.