The use of ethanol, which is a green fuel, in low-temperature fuel cells is attractive because of its low toxicity and high theoretical energy density (8.0 kW h kg−1) [1, 2, 3, 4, 5, 6]. However, the performance of direct ethanol fuel cells (DEFCs) is poor and need to be greatly improved. Table 1 lists the performances of state-of-the-art DEFCs with proton-exchange membranes using different anode catalysts [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]. The cell temperature cannot be greater than 110 °C because it is limited by the critical temperatures of perfluorinated proton-exchange membranes (mostly Nafion®-based membranes). It is well known that the slow kinetics of ethanol electro-oxidation in acidic media at low temperatures is a major challenge for the development of DEFCs. Much effort has been devoted to screening effective catalysts for ethanol electro-oxidation [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]. Pt is the most effective catalyst for ethanol adsorption and dissociation, but the current density on Pt is low as a result of poisoning by adsorbed CO (denoted by COads) derived from ethanol dissociation [22, 25, 26]. The use of Pt-based binary or ternary catalysts to increase the activities of ethanol electro-oxidation catalysts has been investigated.
The results listed in Table 1, which are for DEFC anode catalysts, show that PtRu- and PtSn-based catalysts achieve much greater peak power densities than Pt (12 mW/cm2) [27]. It has been reported that second elements such as Ru or Sn perform different roles during ethanol adsorption, dissociation, and oxidation. The role of Ru is to promote water dissociation at low potentials during ethanol electro-oxidation, producing oxygenated species (e.g., adsorbed OH, denoted by OHads) and giving a higher yield of CO2 than that obtained with Pt catalysts [28, 29]. When Pt is modified with Sn, the effect of COads poisoning is weakened, and the ethanol electro-oxidation pathway yields acetaldehyde, which is then oxidized to acetic acid [29, 30, 31].
We fabricated a double-layered anode catalyst using PtRu black and Pt3Sn1/C catalysts to combine the promotional roles of Ru and Sn on Pt for ethanol electro-oxidation. The performance of the DEFC with a double-layered anode catalyst was better than those of DEFCs with single-layered PtRu black and Pt3Sn1/C anode catalyst layers, and a maximum power density of 96 mW/cm2 was obtained. Another method is to use PtSnRu catalysts. Several methods have been developed for preparing PtRuSn/C catalysts, e.g., the Pechini, formic acid reduction, and liquid-crystalline templating methods. However, the DEFC performance is not high, usually less than 50 mW/cm2 [13, 14, 17, 30, 31, 32]. The preparation method has a strong influence on the particle-size and composition distribution of the catalyst and therefore affects the electrocatalytic activity. Among different approaches, the polyol process using ethylene glycol has been widely used to prepare carbon-supported Pt-based catalysts with accurate size control and a uniform distribution of metal nanoparticles on the carbon support. Polyol-synthesized Pt-based catalysts also have superior electrocatalytic activities for methanol/ethanol electro-oxidation and oxygen reduction reactions [33, 34, 35].
In this work, we prepared PtSnRu/C catalysts with different atomic ratios and metal loadings using the polyol process. The morphologies and compositions of the catalysts were determined using transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS). Single-cell tests and anode polarization measurements were conducted to evaluate the catalyst activities for ethanol electro-oxidation. Apparent activation energy analysis, in situ Fourier-transform infrared spectroscopy (FTIRS), and anode product analysis were performed to investigate the ethanol electro-oxidation mechanism on the PtSnRu/C catalyst.
H2PtCl6·6H2O, SnCl2·2H2O, and RuCl3·xH2O were used as the PtSnRu/C catalyst precursors. Three catalysts with metal loadings of 45 wt% and Vulcan XC-72R as the support material were prepared; the atomic ratios of Pt, Sn, and Ru were 3:1:0, 2.6:1:0.4, and 2:1:1. The atomic ratios of the different elements were varied by controlling the amounts of different precursors during catalyst preparation. The prepared catalysts were denoted by Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, and Pt2Sn1Ru1/C, according to the atomic ratios of the elements. A catalyst with a metal loading of 60 wt% and Ketjen Black EC 300J as the support material was prepared with Pt, Sn, and Ru atomic ratios of 2.6:1:0.4. This catalyst was denoted by Pt2.6Sn1Ru0.4/C-H. The typical preparation procedure was as follows. Calculated amounts of H2PtCl6·6H2O, SnCl2·2H2O, RuCl3·xH2O, and Vulcan XC-72R or Ketjen Black EC 300J carbon black were dissolved in ethylene glycol in a three-necked flask. The pH of the solution was increased to about 13 with 1 mol/L NaOH solution. The solution was heated to 165 °C at a rate of 10 K/min and kept at that temperature for 4 h, and then the solution was cooled to 90 °C. An HCl solution of concentration 1.5 mol/L was then added to adjust the pH to about 2, and the solution was kept at 90 °C for 24 h. Finally, the obtained catalyst was filtered, washed, and dried at 80 °C for 10 h in a vacuum oven.
The catalyst morphologies were investigated by TEM using a TECNAI F30 or G2 microscope (FEI Corp.), and more than 200 particles were measured to obtain the average particle size and particle-size distribution of the catalysts. XPS measurements were carried out using a JEOL JPS-9010MC spectrometer with an Mg Kα radiation source. The Pt 4f, Ru 3p, and Sn 3d signals were collected. The position of the C 1s peak, i.e., 284.2 eV, was used to correct the catalyst binding energies.
Anode and cathode diffusion layers were prepared by brushing a slurry consisting of carbon black, polytetrafluoroethylene (PTFE), and ethanol on hydrophobic carbon paper (0.19 mm thick, SGL Carbon Group, Short Hills, NJ, USA) and subsequent drying as previously described [36]. Anode catalyst layers containing Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, or Pt2Sn1Ru1/C were prepared as follows. A water/alcohol solution of the catalyst was mixed with a Nafion® solution (Dupont), and the resulting mixture was ultrasonically agitated with vigorous mechanical stirring to form a homogeneous ink. The ink was then brushed onto the anode diffusion layer held on a heating table at 60 °C. The resulting catalyst loading was 4.5 mg/cm2 and the Nafion® content of the anode catalyst layer was 10 wt%. The anode catalyst layer with Pt2.6Sn1Ru0.4/C-H had a catalyst loading of 5.8 mg/cm2, and a Nafion® content of 10 wt%. In all cases, identical cathode catalyst layers were prepared by brushing the cathode ink onto the cathode diffusion layer held on a heating table at 60 °C. The resulting Pt loading was 3.7 mg/cm2, and the Nafion content was 12 wt%. The obtained anode and cathode were placed on both sides of a Nafion® 1135 membrane and hot pressed at 135 °C and 2000 pounds for 2 min to form an MEA.
Single-cell tests were carried out using an Arbin® instrument at 90 °C. The anode was fed with 1.5 mol/L ethanol solution at a flow rate of 1 mL/min. The cathode was fed with oxygen at a flow rate of 200 mL/min and gas pressure of 0.2 MPa. Linear sweeping voltammetry of the anode catalyst layers was performed at 90 °C using an EG&G PAR 273 potentiostat at a scan rate of 1 mV/s. For these measurements, the anode was fed with 1.5 mol/L ethanol solution at a flow rate of 1 mL/min as the working electrode, and the cathode was fed with hydrogen gas as the counter and pseudo-reference electrodes. The hydrogen pressure and flow rate were 0.1 MPa and 50 mL/min, respectively.
In situ FTIRS was performed in a PTFE cell with a CaF2 optical window using a JASCO FT/IR-6100 spectrometer equipped with a triglycine sulfate detector. A gold disk (10 mm in diameter) was used as the electrode substrate. The catalyst layers were deposited on a gold electrode surface using a previously described method [37]. The electrode was pressed against the CaF2 window to create a thin ethanol layer in 0.1 mol/L HClO4 solution. IR spectra were recorded at a scan rate of 0.25 mV/s, and 25 interferograms were co-added to each spectrum. All potentials were referenced to the reversible hydrogen electrode. The measurements were performed at room temperature. The trap for the outlet products has been described previously [12]. The DEFC was operated at a certain cell voltage for 3 h under the same operating conditions as in the single-cell tests. The concentrations of ethanol, acetaldehyde, and ethyl acetate were determined using a VARIAN CP-3800 gas chromatograph, and the concentration of acetic acid was determined by titration with standardized NaOH solution. The content was calculated from the weight of BaCO3.
Figure 1 shows TEM images and corresponding size distribution diagrams of the catalysts. It can be seen that the particles are uniformly distributed on the carbon support with mean diameters of 2.6, 2.1, and 1.8 nm for the Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, and Pt2Sn1Ru1/C catalysts, respectively. The particles in the Pt2.6Sn1Ru0.4/C and Pt2Sn1Ru1/C catalysts are distributed more evenly with a narrower range on the support than those in the Pt3Sn1/C catalyst, which might be because the addition of Ru inhibits nanoparticle aggregation.
XPS measurements were carried out to elucidate the surface compositions and oxidation states of Pt, Sn, and Ru in the Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, and Pt2Sn1Ru1/C catalysts. The spectra of Pt 4f, Ru 3p, and Sn 3d in these catalysts are shown in Fig. 2. The Pt 4f spectra contain two peaks corresponding to the Pt 4f7/2 and Pt 4f5/2 states from spin orbital splitting. The Pt 4f7/2 spectra were deconvoluted with three different Pt oxidation states, i.e., Pt(0), Pt(II), and Pt(IV), corresponding to the three peaks at 71.0, 72.1, and 73.9 eV, respectively, for the Pt3Sn1/C catalyst. The standard XPS Pt(0) peak is at 70.9 eV according to previous reports [38]. The small deviation from the standard value could be attributed to interactions between the metals and the carbon support and/or a small-particle effect [39]. The data in Table 2 show that the atomic ratio of Pt(0) in the Pt2.6Sn1Ru0.4/C catalyst is 35.3%, which is lower than that in Pt3Sn1/C (49.4%), indicating that the content of alloyed Pt decreases on addition of a small amount of Ru to the Pt2.6Sn1Ru0.4/C catalyst. However, the content of alloyed Pt increases to 42.5% with increasing Ru content in the Pt2Sn1Ru1/C catalyst, indicating that a considerable amount of PtRu alloy was formed in the catalyst.
The Ru 3p3/2 signals can be deconvoluted into two peaks with binding energies corresponding to Ru(0) and Ru(IV) oxide as shown in Fig. 2(b). The Ru(IV) contents are 41.8% and 41.7% for the Pt2.6Sn1Ru0.4/C and Pt2Sn1Ru1/C catalysts, respectively. This indicates that there are considerable amounts of ruthenium oxide in the catalysts. During catalyst preparation using the polyol method, some Ru3+ is reduced to Ru(0) by ethylene glycol to form a PtRu alloy, and the rest of the Ru is oxidized to amorphous RuO2 in the air [40]. As the data in Table 2 shows, the atomic oxidation states of Ru in the Pt2.6Sn1Ru0.4/C and Pt2Sn1Ru1/C catalysts are similar because the same preparation method was used. From the spectra in Fig. 2(c), it can be seen that the main Sn 3d5/2 peak is at 486.5 eV with a half-peak width of about 1.8 eV. This is assigned to one surface state of SnOx (e.g., SnO, SnO2, or Sn hydroxides); SnO and SnO2 cannot be distinguished because the difference between Sn(II) and Sn(IV) binding energies is only 0.2 eV. Sn atoms on the catalyst surfaces could have been oxidized when the catalysts were exposed to air before the XPS measurements because of the low redox potentials for Sn/Sn2+ and Sn2+/Sn4+ [41, 42].
Figure 3 shows the single-cell performance of DEFCs with Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, Pt2Sn1Ru1/C, and Pt2.6Sn1Ru0.4/C-H as the anode catalysts. The open-circuit voltage (OCV) for Pt2.6Sn1Ru0.4/C catalyst is 0.86 V, which is slightly higher than those of the other catalysts (0.84 and 0.81 V for Pt3Sn1/C and Pt2Sn1Ru1/C, respectively). The OCV of a DEFC is mainly determined by the catalytic activity for ethanol electro-oxidation on the anode catalyst and ethanol crossover from the anode to the cathode. It is known that ethanol crossover is affected by the cell temperature, ethanol concentration, and cathode gas back-pressure. The single-cell-test operating conditions were the same for all the catalysts, and therefore the higher OCV reflects a greater catalytic activity for ethanol electro-oxidation on the Pt2.6Sn1Ru0.4/C catalyst. The cell voltage decreases in the order Pt2.6Sn1Ru0.4/C > Pt3Sn1/C > Pt2Sn1Ru1/C with increasing current density. As the current density continues to increase (>100 mA/cm2), the anode is strongly polarized, and the cell voltage declines more rapidly for the Pt3Sn1/C catalyst than for the Pt2Sn1Ru1/C catalyst, leading to a crossover point of 0.33 V at 200 mA/cm2. This can be explained by the presence of some PtRu alloy in the Pt2Sn1Ru1/C catalyst because the PtRu catalyst is more active than the Pt3Sn1/C catalyst at high current densities [12]. There is little difference between the maximum power densities of the Pt3Sn1/C and Pt2Sn1Ru1/C catalysts: 70 mW/cm2 for Pt3Sn1/C and 66 mW/cm2 for Pt2Sn1Ru1/C. The Pt2.6Sn1Ru0.4/C catalyst gives higher DEFC performance across the entire current density region than Pt3Sn1/C and Pt2Sn1- Ru1/C catalysts, with a maximum power density of 91 mW/cm2 at 175 mA/cm2. Figure 3 shows that when the metal loading on the Pt2.6Sn1Ru0.4/C-H catalyst is further increased to 60 wt%, the maximum power density reaches 121 mW/cm2 at 213 mA/cm2; this is a better performance than those previously reported for similar operating conditions [7, 8, 11, 12, 15, 16, 17].
Figure 4 shows the anode polarization curves of the DEFCs with Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, and Pt2Sn1Ru1/C as the anode catalysts. The concentration and flow rate of ethanol were set at 1.5 mol/L and 1.0 mL/min, respectively, to ensure a fuel supply. The maximum polarization potential was set at 0.6 V to prevent Sn dissolution; the maximum current was set at 1.0 A and was limited by the potentiostat. The faradic current was normalized based on the geometric electrode area. It can be seen from the profiles that the current density in the studied potential region is greatest for Pt2.6Sn1Ru0.4/C. Ethanol oxidation takes place at 0.05 V on the Pt2.6Sn1Ru0.4/C, which is a negative shift of about 20 and 90 mV compared with the Pt3Sn1/C and Pt2Sn1Ru1/C, respectively, suggesting a positive effect on the catalytic activity for ethanol electro-oxidation. The decisive factor is the activation energy, which is related to the reaction barrier of ethanol electro-oxidation, and will be discussed in Section 3.4. In the potential region 0.0-0.49 V, the ethanol electro-oxidation current density is greatest for the Pt2.6Sn1Ru0.4/C catalyst, corresponding to the highest cell voltage in the single-cell tests. The order of the catalyst activity for ethanol oxidation is Pt2.6Sn1Ru0.4/C > Pt3Sn1/C > Pt2Sn1Ru1/C.
The apparent activation energy (Ea) values are calculated by plotting the current density at a specific potential as a function of temperature according to the Arrhenius equation. The slope of a linear fit is used to calculate Ea:
where j is the current density at a specific potential, j0 is the exchange current density, R is the gas constant, T is the temperature, and Ea is the apparent activation energy at the potential. Previous work [3, 21, 28, 43, 44, 45] shows that ethanol electro-oxidation in an acidic solution can be described by two parallel reaction pathways. One pathway is C-C bond cleavage to form CO2, and the other is a direct oxidation pathway to form acetaldehyde and then acetic acid. Ethanol is first dissociated on the active catalytic sites by either pathway. The dissociated species (COads and CH3COads) are adsorbed on these active sites and then oxidized by OHads species generated from water. Figure 5 shows the apparent activation energies at different potentials on the Pt3Sn1/C, Pt2.6Sn1Ru0.4/C, and Pt2Sn1Ru1/C catalysts. It can be seen that the Ea value on the Pt2Sn1Ru1/C catalyst is minimal (21.43 kJ/mol) when the electrode potential is 0.1 V; this means that ethanol adsorption and dissociation are easiest on Pt2Sn1Ru1/C. However, the C1 or C2 species produced by ethanol dissociation cannot be oxidized at these potentials because no current density generated from ethanol electro-oxidation on the Pt2Sn1Ru1/C catalyst is observed in the linear sweep voltammetry curves in Fig. 4. The Ea value on Pt2Sn1Ru1/C increases with increasing electrode potential and becomes greater than those on Pt2.6Sn1Ru0.4/C and Pt3Sn1/C. The atomic states and compositions of Pt and Sn in the Pt2Sn1Ru1/C catalyst are very different from those in the Pt3Sn1/C catalyst as a result of the formation of PtRu alloy, resulting in increasing Ea with increasing electrode potential.
The Ea value on the Pt2.6Sn1Ru0.4/C catalyst is 37.12 kJ/mol at 0.1 V, which is greater than that on Pt3Sn1/C (34.61 kJ/mol). This indicates that adsorption and dissociation of ethanol on the active catalytic sites become more difficult. However, the Ea value decreases sharply with increasing electrode potential, resulting in the lowest value among the catalysts when the electrode potential is greater than 0.2 V. It is therefore proposed that the adsorbed species can be oxidized more quickly on Pt2.6Sn1Ru0.4/C than on Pt3Sn1/C and Pt2Sn1Ru1/C. The small amount of Ru in the Pt2.6Sn1Ru0.4/C catalyst compared with the Pt3Sn1/C catalyst offers more oxygenated species, facilitating electro-oxidation of adsorbed species derived from ethanol dissociation. The synergetic effects of Pt, Sn, and Ru in the Pt2.6Sn1Ru0.4/C catalyst therefore improve the ethanol electro-oxidation efficiency, and thus the maximum power density of the DEFC.
Figure 6 shows in situ spectra recorded during ethanol electro-oxidation on the Pt2.6Sn1Ru0.4/C catalyst from potentials of 0.05 to 0.70 V. A positive peak is located at 2343 cm−1, which is attributed to the asymmetric stretching vibrations of CO2 in water. This appears at 0.25 V and the intensity increases with increasing potential, indicating that CO2 is one of the products, and its yield increases with DEFC discharging. Another dominant feature of the spectra is the peak at 2042 cm−1. The bipolar shape is the result of subtraction of two Stark-shifted single-beam spectra at the sample and reference potentials [46]. The stretching vibrations of the carbonyl groups (C=O) in acetic acid, acetaldehyde, or ethyl acetate cause the peak at 1717 cm−1. However, this peak is affected by the strong negative peak located at 1650 cm−1, which indicates consumption of interfacial water during the reaction. The positive absorption band at 1395 cm−1 is associated with C-O stretching in acetic acid and/or ethyl acetate. The peak at 1368 cm−1 is attributed to CH3 symmetric deformation in acetaldehyde. The positive peak at 1274 cm−1 is attributed to O−H deformation in acetic acid. The peak at about 1100 cm−1 is assigned to adsorption of ClO4− on the electrode surface from the electrolyte solution. The broad peak located at 2620 cm−1 is probably the characteristic vibration of aldehydic (O=)C-H. The results show that CO2, acetic acid, acetaldehyde, and ethyl acetate are produced during ethanol electro-oxidation.
Table 3 lists the anode product distributions and current efficiencies of DEFCs with Pt3Sn1/C and Pt2.6Sn1Ru0.4/C as the anode catalysts. It can be seen that acetaldehyde, acetic acid, ethyl acetate, and CO2 are detected in the DEFC anode effluent, in agreement with the in situ FTIRS results. Acetic acid is the major anode product, with over 90% current efficiency. The current density contributions of acetaldehyde and ethyl acetate are less than 5% and 2%, respectively. The yield is lowest, lower than 1%. Ethanol therefore cannot be completely oxidized to CO2 under the present operating conditions. The C2 products are the key factors in the DEFC performance, and the product distributions and current efficiencies on the Pt2.6Sn1Ru0.4/C and Pt3Sn1/C catalysts are similar, indicating that the addition of Ru to PtSn enhances the electrochemical performance [14].
The ethanol electro-oxidation efficiencies were calculated using Eq. (2):
where η is the ethanol electro-oxidation efficiency, MCO2, MEA, MAAL, and MAA are the concentrations of CO2, ethyl acetate, acetaldehyde, and acetic acid, respectively, Methanol is the initial concentration of ethanol, and Vin and Vout are the inlet and outlet solution volumes, respectively, for the DEFC anode. The results are shown in Fig. 7. It can be observed that the ethanol electro-oxidation efficiency increases as the cell voltage decreases because more ethanol can be oxidized on further discharging. The electro-oxidation efficiencies at the studied cell voltages were higher for Pt2.6Sn1Ru0.4/C than for Pt3Sn1/C, which resulted in a higher current density and therefore a better DEFC performance.
PtSnRu/C catalysts with uniform and narrow particle-size distributions were prepared using the polyol process. Optimization of Pt, Sn, and Ru ratios and increasing the catalyst metal loading resulted in a maximum power density of 121 mW/cm2 for the DEFC with Pt2.6Sn1Ru0.4/C-H as the anode catalyst under the operating conditions 90 °C, 1.5 mol/L ethanol solution, and 0.2 MPa oxygen. The product distribution with the Pt2.6Sn1Ru0.4/C catalyst was similar to that with Pt3Sn1/C catalyst: acetic acid was the major anode product with over 90% current efficiency, and the yield of CO2 was lowest with less than 1%. The ethanol oxidation efficiency on the Pt2.6Sn1Ru0.4/C catalyst was greater than that on the Pt3Sn1/C catalyst. These results suggest that the synergetic effects of Pt, Sn, and Ru in the Pt2.6Sn1Ru0.4/C catalyst decreased the activation energy for ethanol electro-oxidation, thus improving the ethanol electro-oxidation efficiency and maximum power density of the DEFC.
2014, Vol. 35 Issue (8): 1394-1401
PDF (853KB)
2014, Vol. 35 Issue (8): 1394-1401
PDF (853KB)