Alternative Aqueous Phase Synthesis of a PtRu/C Electrocatalyst for Direct Methanol Fuel Cells

: Carbon-supported PtRu nanoalloy (PtRu/C) is widely used as the anode catalyst for direct methanol fuel cells (DMFC), and an aqueous phase synthesis of PtRu/C is in high demand due to for energy-saving and environmentally-benign considerations, however, it is very challenging to attain stoichiometric reduction, good dispersion and a high alloying degree. Herein, we report a facile aqueous phase approach with dimethylamine borane (DMAB) as the reducing agent to synthesize a PtRu/C(DMAB). TEM, XRD, XPS and ICP-AES characterizations indicate that the structural parameters in the PtRu/C(DMAB) are improved signiﬁcantly as compared to those obtained in a PtRu/C(NaBH 4 ) and a commercial PtRu/C, contributing to an enhanced electrocatalytic performance. It turns out that the PtRu/C(DMAB) exhibits the highest methanol electro-oxidation (MOR) performance among all of the tested samples, with the peak current up to 1.8 times as much as that of the state-of-the-art commercial PtRu/C, corroborating the highest output power density in comparative DMFC tests. In-situ attenuated total reﬂection infrared (ATR-IR) spectroscopy correlates the higher methanol electro-oxidation performance of the PtRu/C(DMAB) with its enhanced CO resistance and CO 2 generation. This simple aqueous synthetic approach may provide an alternative route for developing efﬁcient anode electrocatalysts of DMFCs.


Introduction
Direct methanol fuel cells (DMFC) are promising energy-converter to drive portable devices owing to their high efficiency, low emissions and rich fuel supply [1]. Platinum is the primary catalytic metal required for electrocatalytic methanol oxidation in acidic media [2][3][4]. However, the low abundance and poor CO tolerance of Pt calls for its alloying with a secondary metal in practice [5,6]. Among the Pt-based bimetallic catalysts, carbon supported PtRu (atomic ratio of Pt:Ru = 1:1) alloy nanoparticles (PtRu/C) serve as one type of the most practical anode catalysts for DMFC [7]. The high performance of PtRu/C in methanol electro-oxidation reactions (MOR) has been attributed mainly to the so-called bifunctional mechanism [8] together with the electronic effect [9]. For a total 6-electron oxidation reaction, the main pathway on a Pt-Ru alloy may include [7]: Pt + CH 3 OH → Pt-CO ads + 4H + + 4e − Ru + H 2 O → Ru-OH ads + H + + e − Pt-CO ads + Ru-OH ads → CO 2 + Pt + Ru + H + + e − DMAB is initially used in this work as the reductant to synthesis of a PtRu/C catalyst. It is noted that the pH plays a crucial role in the synthetic process of the PtRu/C(DMAB). On one hand, the metallic precursor solution containing Ru(III) and Pt(II) species is maintained as acidic to avoid their hydrolysis [29,34]. On the other hand, DMAB is very unstable in acidic media, leading to a significantly decreased reducing power, thus DMAB is better used in a mild alkaline solution. For a balanced consideration in this work, a reverse addition mode is introduced in the reduction process, that is, the acidic metallic precursor solution is added dropwise to the DMAB-contained carbon black slurry of a weak alkalinity. By these means, the nominal reduction of Pt(II) and Ru(III) species at the desired stoichiometric ratio of 1:1 can be obtained.
Equations (4)- (6) show that the reaction mechanism between DMAB and Pt(II)/Ru(III) is analogous to our previous reported work [31]. Firstly, DMAB reacts with OH − to form BH 3 OH − , and BH 3 OH − is to be the reducing agent. Next, it is possible for each BH 3 OH − to provide six electrons to reduce three Pt(II) or two Ru(III) ions, leading to the formation of PtRu/C(DMAB).

Physicochemical Characterization
The compositions of the as-synthesized PtRu/C were determined by ICP-AES analysis of metal nanoparticles dissolved in aqua regia. As listed in Table 1, the atomic percentages of B in the PtRu/C(DMAB) and PtRu/C(NaBH 4 ) are negligible (<1.5 at.%). The weight percentages of Pt in PtRu/C(DMAB) and PtRu/C(NaBH 4 ) prepared by the reverse addition are 19.0 wt.% and 19.2 wt.% while those of Ru are 10.2 wt.% and 9.9 wt.%, respectively. For the reverse addition mode, the Pt(II) and Ru(III)-contained precursor solution was acidic before reacting with the reductants, which may suppress the hydrolysis of the Ru(III) precursor. Moreover, an excessive amount of the reductant was added into the carbon black slurry with a mild alkalinity beforehand. As a result, the reverse addition aqueous phase synthesis led to the nominal reduction of Pt(II) and Ru(III) species at the desired stoichiometric ratio of 1:1. In the case of the regular addition mode, the determined weight percentages of Pt in PtRu/C(DMAB) and PtRu/C(NaBH 4 ) were 18.5 wt.% and 20.0 wt.% while those of Ru were 7.0 wt.% and 6.8 wt.%, respectively. The incomplete reduction of Ru(III) precursor may be chiefly due to the hydrolysis of the Ru(III) precursor [29,34]. In the following sections, only the PtRu/C(DMAB) and PtRu/C(NaBH 4 ) catalysts prepared by the reverse addition are investigated.  Figure 1 presents the typical TEM images of Johnson Matthey (JM) Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH 4 ) catalysts. The statistically mean sizes of metallic nanoparticles for JM Pt/C and JM PtRu/C are 2.7 ± 0.5 nm and 2.6 ± 0.4 nm, respectively. For comparison, PtRu/C(DMAB) and PtRu/C(NaBH 4 ) show mean PtRu particle sizes of 2.0 ± 0.2 nm and 2.7 ± 0.4 nm, respectively. The larger size and poorer dispersion of PtRu nanoparticles on carbon black in PtRu/C(NaBH 4 ) may arise from the stronger reducing power of NaBH 4 , showing the advantage of the milder reductant DMAB.
XRD patterns of JM Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH 4 ) catalysts are shown in Figure 2. The XRD pattern of JM Pt/C exhibits (111), (200) and (220) diffraction peaks at 2θ values of 39.8 • , 46.2 • , 67.4 • , respectively (black dashed lines, PDF#04-0802), indicating a Pt face-centered cubic (fcc) structure [35]. Separated Ru peaks are absent in all PtRu/C catalysts (red dashed lines, PDF#06-0663), which suggest that they are mainly in an alloy phase. The (111) diffraction peaks of the PtRu/C catalysts are shifted positively compared to those of JM Pt/C, in agreement with lattice shrinkage in a PtRu alloy [11]. The resulting Pt d-band downshift of the PtRu alloy is expected to contribute partly to its increased resistance to CO poisoning, in addition to the well-known bifunctional mechanism [12]. Catalysts 2021, 11, x FOR PEER REVIEW 4 of 15 XRD patterns of JM Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH4) catalysts are shown in Figure 2. The XRD pattern of JM Pt/C exhibits (111), (200) and (220) diffraction peaks at 2θ values of 39.8°, 46.2°, 67.4°, respectively (black dashed lines, PDF#04-0802), indicating a Pt face-centered cubic (fcc) structure [35]. Separated Ru peaks are absent in all PtRu/C catalysts (red dashed lines, PDF#06-0663), which suggest that they are mainly in an alloy phase. The (111) diffraction peaks of the PtRu/C catalysts are shifted positively compared to those of JM Pt/C, in agreement with lattice shrinkage in a PtRu alloy [11]. The resulting Pt d-band downshift of the PtRu alloy is expected to con- The average metallic particle sizes (d XRD ) for the above catalysts were also calculated from the (111) diffraction peak by applying the Scherrer's equation [36]: where K = 0.9 is a constant, λ is the wavelength of a Cu target X-ray (0.154 nm), B is the width of (111) a peak at half height, and θ is the diffraction angle at the maximum of (111) the peak. As shown in Table S1, the average particle sizes thus calculated are in accor-dance with those characterized by TEM, suggesting uniform size and good crystallinity of these catalysts.
Assuming that the lattice parameters on Ru content for unsupported and carbonsupported Pt-Ru alloys were of the same dependence [37], the alloying degree was semiquantitatively evaluated by Vegard's law. The lattice parameter (a) values for PtRu/C catalysts may be calculated by using (111) diffraction peaks [38]: while the alloying degree of PtRu/C (X Ru ) may be estimated from the following equation [37]: where a 0 = 0.3916 nm is the lattice parameter for pure Pt/C and k = 0.0124 nm is a constant. The values of a for JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH 4 ) are estimated to be 0.3848, 0.3848 and 0.3852 nm, respectively, yielding X Ru values of 55%, 55% and 52% for the three catalysts. The X Ru value for PtRu/C(DMAB) equals to that for JM PtRu/C but is slightly larger than that for PtRu/C(NaBH 4 ), as a result of using the milder reductant DMAB.
Catalysts 2021, 11, x FOR PEER REVIEW 5 of 15 tribute partly to its increased resistance to CO poisoning, in addition to the well-known bifunctional mechanism [12]. The average metallic particle sizes (dXRD) for the above catalysts were also calculated from the (111) diffraction peak by applying the Scherrer's equation [36]: where K = 0.9 is a constant, λ is the wavelength of a Cu target X-ray (0.154 nm), B is the width of (111) a peak at half height, and θ is the diffraction angle at the maximum of (111) the peak. As shown in Table S1, the average particle sizes thus calculated are in accordance with those characterized by TEM, suggesting uniform size and good crystallinity of these catalysts.
Assuming that the lattice parameters on Ru content for unsupported and carbon-supported Pt-Ru alloys were of the same dependence [37], the alloying degree was semi-quantitatively evaluated by Vegard's law. The lattice parameter (a) values for PtRu/C catalysts may be calculated by using (111) diffraction peaks [38]: while the alloying degree of PtRu/C (XRu) may be estimated from the following equation [37]: The chemical oxidation states and electronic properties were determined by core level XPS spectra. Figure 3a shows the Pt core level XPS spectra of JM Pt/C and PtRu/C catalysts. By deconvoluting the Pt 4f 7/2 and Pt 4f 5/2 doublets, the presence of Pt 0 , Pt 2+ and Pt 4+ components can be revealed. As listed in Table S2, the Pt 0 4f 7/2 binding energies equal to 71.1, 71.4, 72.3 and 72.0 eV for JM Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH 4 ), respectively. Specifically, the positive shift can be attributed to the partial electron transfer from Ru to Pt which lowers the d-band center of Pt sites or weakens the adsorption of CO intermediate on PtRu/C(DMAB) [12]. What's more, as seen from Figure 3b and Table S3, the relative peak intensity for Ru 0 to Ru(IV) on PtRu/C(DMAB) is larger than that on JM PtRu/C or PtRu/C(NaBH 4 ). The metallic Ru rather than the RuO 2 was actually suggested to contribute to the bifunctional mechanism [39,40], providing sufficient Ru-OH to accelerate the oxidation of Pt-CO on PtRu/C(DMAB). Table S3, the relative peak intensity for Ru 0 to Ru(IV) on PtRu/C(DMAB) is larger than that on JM PtRu/C or PtRu/C(NaBH4). The metallic Ru rather than the RuO2 was actually suggested to contribute to the bifunctional mechanism [39,40], providing sufficient Ru-OH to accelerate the oxidation of Pt-CO on PtRu/C(DMAB).

Electrochemical Performance
In Figure 4a, cyclic voltammograms (CVs) of JM Pt/C and PtRu/C catalysts in 0.5 M H2SO4 present electrochemical voltametric features typical of Pt and PtRu electrodes. No distinct hydrogen adsorption/desorption peaks except larger double-layer currents can be observed on PtRu/C catalysts, compared to the corresponding counterparts on JM Pt/C. The electrochemical surface areas (ECSAs) of these catalysts may be evaluated by means of the anodic CO stripping shown in Figure 4b, according to the following equation, where Q is the integrated charge of the CO ad-layer, m is the total mass of Pt loaded on GCE, v is the scan rate and QCO is ca. 0.42 mC cm -2 for the oxidation of CO monolayer. Specifically, the previous ATR-IR studies showed that the relative CO amount on Ru sites of a PtRu alloy was very small [41,42]. As a result, the ECSAs can be normalized to the Pt mass, which are estimated to be 85.8, 94.3, 126.3 and 141.9 m 2 ·g −1 Pt for JM Pt/C, JM PtRu/C, PtRu/C(NaBH4) and PtRu/C(DMAB), respectively. Besides, as detailed in Table  S4, the onset and peak potentials of oxidative CO removal on PtRu/C(DMAB) are shifted by ca. −280 mV as compared to those on JM Pt/C, and present a slightly negative shift relative to those on JM PtRu/C and PtRu/C(NaBH4), indicating the weakest CO adsorption on PtRu/C(DMAB), which is concomitant with the XPS characterizations. The better CO resistance and larger ECSA of PtRu/C(DMAB) are expected to be consistent with its higher MOR activity and durability.

Electrochemical Performance
In Figure 4a, cyclic voltammograms (CVs) of JM Pt/C and PtRu/C catalysts in 0.5 M H 2 SO 4 present electrochemical voltametric features typical of Pt and PtRu electrodes. No distinct hydrogen adsorption/desorption peaks except larger double-layer currents can be observed on PtRu/C catalysts, compared to the corresponding counterparts on JM Pt/C. The electrochemical surface areas (ECSAs) of these catalysts may be evaluated by means of the anodic CO stripping shown in Figure 4b, according to the following equation, where Q is the integrated charge of the CO ad-layer, m is the total mass of Pt loaded on GCE, v is the scan rate and Q CO is ca. 0.42 mC cm -2 for the oxidation of CO monolayer. Specifically, the previous ATR-IR studies showed that the relative CO amount on Ru sites of a PtRu alloy was very small [41,42]. As a result, the ECSAs can be normalized to the Pt mass, which are estimated to be 85.8, 94.3, 126.3 and 141.9 m 2 ·g −1 Pt for JM Pt/C, JM PtRu/C, PtRu/C(NaBH 4 ) and PtRu/C(DMAB), respectively. Besides, as detailed in Table  S4, the onset and peak potentials of oxidative CO removal on PtRu/C(DMAB) are shifted by ca. −280 mV as compared to those on JM Pt/C, and present a slightly negative shift relative to those on JM PtRu/C and PtRu/C(NaBH 4 ), indicating the weakest CO adsorption on PtRu/C(DMAB), which is concomitant with the XPS characterizations. The better CO resistance and larger ECSA of PtRu/C(DMAB) are expected to be consistent with its higher MOR activity and durability.
The methanol electro-oxidation activities of JM Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH 4 ) catalysts were also compared by CVs. Figure 5a depicts Pt-mass normalized CVs for Pt-based catalysts in 0.5 M H 2 SO 4 containing 1 M CH 3 OH solution at 50 mV s −1 . It can be seen that PtRu/C(DMAB) exhibits a higher peak current density of 0.53 A mg −1 Pt , which is about 1.5 and 1.8 times of those obtained on JM Pt/C (0.36 A mg −1 Pt ) and JM PtRu/C (0.30 A mg −1 Pt ), respectively. Besides, as shown in Table S5, the ratio between the forward (i f ) and backward peak current density (i b ) on PtRu/C(DMAB) is comparable to that on JM PtRu/C but is higher than the counterparts on JM Pt/C and PtRu/C(NaBH 4 ). A higher i f / i b ratio was previously used as a criterion for CO tolerance [43][44][45][46], and more recently as an indicator of a more oxyphilic surface or enhanced OH formation [47,48], in favor of MOR to CO 2 conversion at lower potentials in the preceding forward scan. The MOR parameters of the PtRu/C catalysts from this work and recent publications are listed in Table S6 for comparison. Notably, the peak potential of CO oxidation on PtRu/C(DMAB) is 0.252 V (vs. SCE), which is more negative than those reported in the literature, in favor of the MOR to CO 2 conversion at lower potentials. The better CO tolerance and higher mass activity of the PtRu/C(DMAB) together with the aqueous synthesis benefits the scaled-up application of this catalyst. The methanol electro-oxidation activities of JM Pt/C, JM PtRu/C, PtRu/C(DMAB) and PtRu/C(NaBH4) catalysts were also compared by CVs. Figure 5a depicts Pt-mass normalized CVs for Pt-based catalysts in 0.5 M H2SO4 containing 1 M CH3OH solution at 50 mV s −1 . It can be seen that PtRu/C(DMAB) exhibits a higher peak current density of 0.53 A mg −1 Pt, which is about 1.5 and 1.8 times of those obtained on JM Pt/C (0.36 A mg −1 Pt) and JM PtRu/C (0.30 A mg −1 Pt), respectively. Besides, as shown in Table S5, the ratio between the forward (if) and backward peak current density (ib) on PtRu/C(DMAB) is comparable to that on JM PtRu/C but is higher than the counterparts on JM Pt/C and PtRu/C(NaBH4). A higher if / ib ratio was previously used as a criterion for CO tolerance [43][44][45][46], and more recently as an indicator of a more oxyphilic surface or enhanced OH formation [47,48], in favor of MOR to CO2 conversion at lower potentials in the preceding forward scan. The MOR parameters of the PtRu/C catalysts from this work and recent publications are listed in Table S6 for comparison. Notably, the peak potential of CO oxidation on PtRu/C(DMAB) is 0.252 V (vs. SCE), which is more negative than those reported in the literature, in favor of the MOR to CO2 conversion at lower potentials. The better CO tolerance and higher mass activity of the PtRu/C(DMAB) together with the aqueous synthesis benefits the scaled-up application of this catalyst. Figure 5b shows the CVs in 0.5 M H2SO4 containing 1 M CH3OH solution at 1 mV s −1 to mimic a quasi-steady-state measurement. The corresponding Tafel plots in Figure 5c shows the apparent Tafel slopes for the PtRu/C(DMAB) and PtRu/C (NaBH4) are both ca. 183 mV dec −1 , close to that obtained for the commercial PtRu/C (200 mV dec −1 ) and appreciably lower than that for the commercial Pt/C (226 mV dec −1 ). The smaller apparent Tafel slope on the PtRu/C(DMAB) together with the larger oxidation peak current suggests that the DMAB-derived PtRu/C catalyst at least demonstrates a higher activity in a half cell measurement condition, thus deserving of a further test as the real anode catalyst in a single direct methanol fuel cell. In order to further evaluate the long-time stability, chronoamperometric curves were recorded at a practical potential of 0.25 V. As shown in Figure 5d

Preliminary DMFC Test
The electrical characteristics of the DMFCs incorporating the above anode catalysts   Figure 5c shows the apparent Tafel slopes for the PtRu/C(DMAB) and PtRu/C (NaBH 4 ) are both ca. 183 mV dec −1 , close to that obtained for the commercial PtRu/C (200 mV dec −1 ) and appreciably lower than that for the commercial Pt/C (226 mV dec −1 ). The smaller apparent Tafel slope on the PtRu/C(DMAB) together with the larger oxidation peak current suggests that the DMAB-derived PtRu/C catalyst at least demonstrates a higher activity in a half cell measurement condition, thus deserving of a further test as the real anode catalyst in a single direct methanol fuel cell.
In order to further evaluate the long-time stability, chronoamperometric curves were recorded at a practical potential of 0.25 V. As shown in Figure 5d, PtRu/C(DMAB) exhibits a current density of 35.4 mA mg −1 Pt for 1 h electrooxidation, which is 44, 1.52 and 1.11 times as high as that obtained on JM Pt/C (0.8 mA mg −1 Pt ), JM PtRu/C (23.3 mA mg −1 Pt ) and PtRu/C(NaBH 4 ) (31.9 mA mg −1 Pt ), respectively. Again, the enhanced MOR performance on PtRu/C(DMAB) is consistent with the structural characterization results.

Preliminary DMFC Test
The electrical characteristics of the DMFCs incorporating the above anode catalysts were measured to evaluate their potential application. Figure 6 compares the plots of the output voltage and power density versus current density of the DMFCs fed with 1 M methanol at 80 • C at a given Pt mass loading. The highest current density of the DMFC with PtRu/C(DMAB) is 464 mA cm −2 , which is higher than that of JM Pt/C (200 mA cm −2 ), PtRu/C(NaBH 4 ) (400 mA cm −2 ) or JM PtRu/C (439 mA cm −2 ). In terms of peak power density, the DMFC with JM Pt/C yields the lowest value (41.1 mW cm −2 ), and the one with PtRu/C(DMAB) gives the highest one (96.3 mW cm −2 ). The latter is ca. 18% and 9% higher than that with PtRu/C(NaBH 4 ) (81.8 mW cm −2 ) and the state-of-the-art JM PtRu/C (88.7 mW cm −2 ), respectively, in harmony with the electrochemical measurements. What's more, we understand that the fuel cell performance relies heavily on the membrane electrode assembly (MEA) preparation technology besides the catalysts [28,49,50], optimizing the microstructures of MEA in the near future could lead to an upgraded output performance of the DMFC with the PtRu/C(DMAB). By the way, the JM PtRu/C is speculated as a sosoloid by considering its poor solubility in a hot aqua regia, for which structure special post heat-treatment is often required following the organic phase synthesis of a catalyst. In this regard, our facile aqueous phase synthesis of the PtRu/C(DMAB) provides an alternate solution to the anode catalyst of DMFCs.
Catalysts 2021, 11, x FOR PEER REVIEW 9 thesis of a catalyst. In this regard, our facile aqueous phase synthesis of PtRu/C(DMAB) provides an alternate solution to the anode catalyst of DMFCs.

Interfacial Species Evolved during MOR
In-situ ATR-IR spectroscopy was employed to trace the evolution of adsorbed termediates and dissolved products of methanol dissociation and oxidation while potential was scanned positively. The spectral results are shown in Figure 7 and the tailed band assignments are summarized in Table S7. Briefly, in Figure 7a-c, the str band at 2019~2053 cm −1 can be assigned to the linearly bonded CO on Pt (Pt-COL), and

Interfacial Species Evolved during MOR
In-situ ATR-IR spectroscopy was employed to trace the evolution of adsorbed intermediates and dissolved products of methanol dissociation and oxidation while the  Figure 7 and the detailed band assignments are summarized in Table S7. Briefly, in Figure 7a-c, the strong band at 2019~2053 cm −1 can be assigned to the linearly bonded CO on Pt (Pt-CO L ), and the weak bands at 1935~1972 cm −1 belong to the Ru-CO L due to methanol dehydrogenation, together with the band at~1610 cm −1 due to the δ(HOH) of interfacial water. The downward IR bands at~1723 and~1323 cm −1 can be assigned to dissolved formic acid (HCOOH) and bridge-bonded fomite (HCOO B ), respectively, given that the reference spectrum is taken at 1.0 V. These observed results are consistent with the so-called dual reaction pathways for MOR, i.e., the CO-pathway and the non-CO (via fomite-formic acid) pathway [51][52][53]. The normalized CO L band intensities as a function of potential are plotted in Figure 8. The CO L bands on JM Pt/C, PtRu/C and PtRu/C(DMAB) decrease by ca. 35%, 50%, and 60%, respectively, as the potential increases from 0.1 V to 0.6 V. The more rapid drop of the CO L band with increasing potential, or the higher CO-resistance arises from the bifunctional mechanism on PtRu alloys [54]. The CO 2 produced during MOR was detected with the asymmetric stretch band at~2340 cm −1 , as shown in Figure 7d. The normalized CO 2 band intensities versus potential are compared in Figure 8 for the three corresponding catalysts. It can be seen that the CO 2 band on the PtRu/C(DMAB) shows up at the most negative potential, as its band intensity reaches 100%, indicating that it is the most promoted 6-electron MOR on this home-synthesized catalyst. Thus, the ATR-IR results further prove the enhanced electrocatalytic and DMFC performance of the as-synthesized PtRu/C(DMAB) at a molecular level.

Preparation of Catalysts
The PtRu/C(DMAB) was prepared by a so-called reverse addition of aqueous phase synthesis. Briefly, in a 100 mL three-necked bottle, 56 mg of refluxed Vulcan XC-72 carbon black was added to 20 mL of Milli-Q water. The slurry was sonicated for 45 min, and stirred at 750 rpm for 1 h in an ice-water bath with a N 2 -bubbled atmosphere. Then, 250 mg of Na 3 C 6 H 5 O 7 ·2H 2 O and 294.6 mg of DMAB were added into the slurry, which was further stirred to ensure the complete dissolution of the two chemicals. Next, 10 mL of aqueous solution (to suppress the hydrolysis of the metallic precursors, additional 1 vol.% HCl was added) containing 34 mg of K 2 PtCl 4 and 17 mg of RuCl 3 was added dropwise to the suspension via a peristaltic pump at 1 mL min −1 . The suspension was kept stirring at 900 rpm in an ice-water bath and a N 2 -bubbled atmosphere for another 4 h. After that, the slurry was stirred at 30 • C overnight to ensure the complete decomposition of the remaining DMAB. The powder was thereafter filtered out of the suspension, and rinsed with 5 mL of 0.1 mol L −1 CH 3 COOH and copious amount of ultrapure water, respectively. The as-prepared PtRu/C(DMAB) was dried in a vacuum oven at 80 • C for 8 h. The PtRu/C(NaBH 4 ) catalyst was prepared according to the otherwise same procedures except that 189.2 mg of NaBH 4 was used as the reducing agent instead of DMAB.
For comparison, the PtRu/C(DMAB) and PtRu/C(NaBH 4 ) catalysts were also prepared by a regular addition of aqueous phase synthesis, namely, one of the reductants was added dropwise into the aqueous solution containing carbon black and Pt(II)-Ru(III) precursors with otherwise the same conditions as described above.

Material Characterization
The compositions of the catalysts were determined by inductively coupled plasmaatomic emission spectroscopy (ICP-AES, Thermo Fisher iCAP 7400, Thermo Fisher China, Shanghai, China). The crystalline phase structures of the nanoparticles were examined by X-ray diffraction (XRD, Bruker D8 Discovery A25, Bruker AXS, Karlruher, BW, Germany) with the Cu Kα radiation at the 2θ angle from 30 • to 80 • . The morphology and size distributions of the catalysts were characterized by transmission electron microscope (TEM, FEI Tecnai G2 F20 S-Twin, FEI, Hillsboro, OR, USA). The metallic electronic structures were analyzed by X-ray photoelectron spectroscopy (XPS, PHI 5000C&PHI 5300, PHI, Lafayette, LA, USA) with Mg Kα radiation and the C 1s peak at 284.6 eV was used for calibrating the binding energies.

Electrochemical Measurements
The working electrode was a glassy carbon electrode (GCE, 3 mm in diameter) coated with different catalyst layers. A saturated calomel electrode (SCE) and a Pt gauze were used as the reference and counter electrodes, respectively, in a three-compartment cell. An electrochemical workstation (Model 605B, CH Instruments, Shanghai, China) was used to control potential and measure current with an 80% iR drop compensation. All electrochemical measurements were run at 25 ± 1 • C.
First, 2 mg of the catalysts, 800 µL of Milli-Q water, 200 µL of isopropanol and 20 µL of 5 wt.% Nafion were mixed and sonicated to form a uniform catalyst ink, then 5.1 µL of the ink was coated onto a polished GCE via a micropipette and dried with the aid of infrared radiation. The Pt loading was set to be 28 µg cm −2 on each working electrode. Cyclic voltammograms of MOR on different catalysts were measured in 1 M CH 3 OH + 0.5 M H 2 SO 4 at 50 mV s −1 , while chronoamperometric curves were recorded in the same solution at 0.25 V for 3600 s. Prior to the above measurements, the catalysts were subjected to multiple cycles of potential scan in 0.5 M H 2 SO 4 to attain clean surfaces. Notably, the upper potential limit was set to be 0.6 V to avoid the oxidative leaching of the Ru component [55,56]. The ECSAs were evaluated by the anodic CO stripping voltammetry at 10 mV s −1 in which the tested catalysts were predosed in CO-saturated solution at 0.05 V followed by N 2 -purging.

Fuel Cell Performance Test
In the single cell test, the MEA was a catalyst coated membrane electrode prepared by ultrasonic spraying. Further, 20 wt.% Pt/C or 30 wt.% PtRu/C was employed as the anode catalyst, with a Pt loading of 0.8 mg cm −2 , and 60 wt.% JM Pt/C was used as the cathode catalyst, with a Pt loading of 2 mg cm −2 . The cathode and anode catalysts were sprayed onto one and the other sides of a Nafion membrane, respectively. The MEA output performance was assessed at 80 • C by feeding 1 M methanol at a flow rate of 35 mL min −1 and pure oxygen at a flow rate of 0.3 slpm (standard liters per minute).

In-Situ ATR-IR Measurement
The in-situ ATR-IR measurement was run on a Nicolet iS50 IR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) with a built-in MCT-A detector at an incidence angle of ca. 60 • in conjunction with the electrochemical workstation. The working electrode was a catalyst layer coated on an Au electroless under-film pre-deposited on the reflecting plane of a hemicylindrical Si prism, as detailed in our previous reports [51,57,58]. A graphite rod and an SCE served the counter and reference electrodes, respectively. The catalyst ink was prepared by sonicating a mixture of 5 mg of the catalysts, 1 mL of isopropanol and 50 µL of 5 wt.% Nafion, then 105 µL of the ink was pipetted on the IR window and dried to attain a Pt loading of 127 µg cm −2 . The resulting ATR-IR spectra were calculated according to A = -lg(R s / R r ), where R s and R r represented the reflected single-beam intensities obtained at the sample and reference potentials, respectively.

Conclusions
In summary, we have initially reported a simple, cost-effective and facile DMAB-based aqueous phase approach to synthesize the PtRu/C(DMAB) with a desired 1:1 stoichiometric ratio. PtRu nanoparticles are well-dispersed on carbon support in the PtRu/C(DMAB) with mean particle sizes being smaller than those in the PtRu/C(NaBH 4 ) or the commercial PtRu/C, concomitant with the most enhanced CO tolerance and MOR activity. What's more, in-situ ATR-IR spectroscopy further corroborates the electrocatalytic performance at a molecular level. The higher power density of a DMFC with the PtRu/C(DMAB) anode catalyst together with the facile aqueous synthesis of this catalyst paves the way for its potential scaled-up application.

Data Availability Statement:
The data presented in this study are available in the manuscript ( Table 1, Figures 1-8) and supplementary materials (Tables S1-S7).