The extensive use of fossil fuel has caused high levels environmental pollution, leading to climate change, and an energy crisis. Therefore, there is an urgent need for the development of new power sources. As a high-efficiency energy conversion device, fuel cells have a critical role in the energy utilization sector [1]. Proton exchange membrane fuel cells use hydrogen and small organic molecules, such as CH3OH and HCOOH, as the anodic fuels and have attracted much attention [2, 3]. Compared with hydrogen, small organic molecules have a higher energy density and are easily transported and stored. Their application is becoming more promising over the next decade [4, 5, 6]. Direct formic acid fuel cells (DFAFCs) with the electrooxidation of formic acid as the anodic reaction have advantages such as a higher theoretical open circuit potential, lower rate of fuel crossover, non-toxic, and non-flammable. They are expected to become commercialized fuel cells [7].
However, there are problems for the development of DFAFCs. For example, palladium and platinum catalysts are easily poisoned by intermediates, and the catalytic properties including activity and stability need further improvement. For these reasons, efficient promoters, such as second metals [8, 9, 10], nonmetals [11], and metal oxides [12, 13], have been proposed to enhance the catalytic properties of Pt or Pd catalysts via synergetic effect (bifunctional mechanism) or electronic effect (change in the Pt/Pd electronic structure). Conductive polymers have long been considered as an attractive support material and promoter of catalysts because of their high surface area, low resistance, and high stability. A number of studies on the synthesis, characterization, and properties of conducting polymer-supported catalysts have been reported [14]. The list of conducting polymers applied to Pt or Pd catalysts includes polyaniline (PANI) [15, 16], polypyrrole [17, 18], polyfuran [19], polythiophene [20], and polycarbazole [21]. Polycarbazole can weaken the adsorption strength of CO on catalysts and thus allows CO oxidation under lower potentials [21]. PANI has an important role in changing the electronic structure of Pt and increases the activity and stability of Pt toward oxygen reduction reaction [15].
In the present study, two types of PANI-promoted Pd samples are prepared as catalysts for the formic acid electrooxidation reaction (FAOR). nPANI/Pd catalysts are prepared by the electropolymerization of aniline on the glassy carbon electrode and the electrodeposition of Pd. The Pd/C/nPANI catalysts are prepared through the direct electropolymerization of aniline on a commercial Pd/C catalyst. The catalytic properties of Pd for FAOR with and without PANI are compared to evaluate the efficiency of PANI as a promoter for Pd catalyzing FAOR.
A glassy carbon electrode (d = 5 mm) embedded in a Teflon holder was used as the working electrode. Prior to each use, the electrode was polished with 0.5 and 0.05 μm alumina suspensions followed by washing ultrasonically with HNO3 (1:1), ethanol, acetone, and deionized water, sequentially. The solution for the electropolymerization of PANI was H2SO4 (0.5 mol/L) and aniline (0.05 mol/L). PANI was obtained after different cycles of cyclic voltammetry (CV) in a potential range from −0.2 to 0.9 V at a scan rate of 50 mV/s. The subsequent deposition of Pd nanoparticles (NPs) was obtained after 20 cycles of CV in H2SO4 (0.5 mol/L) and K2PdCl4 (1 mmol/L), in a potential range from −0.2 to 0.6 V at a scan rate of 50 mV/s. The catalysts obtained through the above processes were marked as nPANI/Pd, and n refers to the number of CV cycles for the electropolymerization of PANI.
Pd/C catalyst (20 wt%) was purchased from Johnson Matthey Company. The Pd/C catalyst ink was prepared by sonicating a suspension of the carbon-supported catalyst (5.0 mg) in isopropanol (1.0 mL). A 10 μL sample of the suspension was transferred onto the disk electrode and air-dried. The subsequent electropolymerization of PANI was obtained via the same method as that for nPANI/Pd. PANI was obtained after different cycles of CV in the potential range from −0.2 to 0.9 V at the scan rate of 50 mV/s. The catalysts were marked as Pd/C/nPANI.
Aniline was purified by vacuum distillation and stored under nitrogen gas. All other reagents were of analytical grade and used as received.
To characterize the morphology of the nPANI/Pd samples, an ITO electrode was used as the working electrode for the deposition of PANI and Pd NPs. The morphology of the nPANI/Pd samples was investigated by field-emission scanning electron microscopy (SEM; JSM 7401F, JEOL, Tokyo, Japan) at an operating voltage of 3.0 kV. The morphology was also investigated using a JEM-2100 transmission electron microscope (TEM; JEOL) operating at 120 kV. X-ray photoelectron spectroscopy (XPS) measurements were taken on a Thermo ESCALAB 250 instrument (Rochester, NY, USA) with Al Kα radiation (hν = 1486.6 eV). X-ray diffraction (XRD) patterns were collected on a MiniFlex 600 X-ray diffractometer (Rigaku Co. Ltd, Tokyo, Japan) at a scan rate of 4°/min (15° < 2θ < 80°) and the wavelength of the incident radiation was 0.15406 nm (Cu Kα). To characterize the crystal structure of the Pd NPs, the electrodeposition of Pd was carried out on carbon paper electrode (TGP-090, Toray, Tokyo, Japan) instead of the glassy carbon electrode as previously reported [22]. The mass of Pd electrodeposited on each nPANI/Pd electrode was determined by inductively coupled plasma atomic emission spectrometry (PerkinElmer Optima-4300DV Spectrometer, Norwalk, CT, USA).
Electrochemical measurements were carried out in a conventional three-electrode cell at 25 °C using a CHI 760E electrochemical workstation (Shanghai Chenhua Apparatus, Shanghai, China). A Pt foil (1.0 cm × 1.0 cm) and a saturated calomel electrode (SCE) were used as the counter electrode and reference electrode, respectively. All potentials in this work are referred to the SCE. CV and chronoamperometry tests of FAOR were performed in H2SO4 (0.5 mol/L) and HCOOH (2.0 mol/L) solution. The chronoamperometry curves were tested under a constant potential of 0 V vs SCE.
Fig. 1 shows the SEM and TEM images of the nPANI/Pd and Pd/C/nPANI samples. The deposited Pd NPs were uniformly dispersed on the surface of the electrode (Fig. 1(a)), and the interwoven fibrous network structure of PANI was formed after electropolymerization (Fig. 1(b)). The fiber network is not densely packed, making the polymer film highly porous. The morphology of the electrode surface changed significantly after the deposition of Pd NPs on the PANI electrode (Fig. 1(c)-(f)). In all of the nPANI/Pd samples, the Pd NPs were immobilized on the dendrites of PANI. The dendrites of PANI were decorated with highly dispersed Pd NPs, which created a close proximity between the NPs and PANI. Such an affinity is desirable for the composite nPANI/Pd materials to show a fast electron transfer rate and high conductivity. Additionally, increasing the electropolymerization cycles (n) of PANI gradually increased the thickness of the fiber network. Fig. 1(g) and (h) shows representative TEM images of the Pd/C/20PANI sample. After the deposition of PANI, the Pd/C catalyst particles were covered with a PANI layer, and the boundary zones of the particles were unclear. Similarly, with the increase in electropolymerization cycles of PANI, the thickness of the layer gradually increased (results of other Pd/C/nPANI samples are not shown here). The thickness of the PANI layer may be crucial for the catalytic activities of Pd catalysts for FAOR.
To examine the crystal structure of the Pd-based catalysts, XRD measurements were carried out using a carbon paper electrode to replace the glassy carbon electrode. Because the intensity of the signals arising from the carbon paper is much higher than that from the Pd catalysts, we show a close-up of the XRD patterns in the range 2θ = 35°-45°. Fig. 2(a) and (b) shows the XRD patterns of the nPANI/Pd and Pd/C/nPANI samples, respectively. The signals of the uncoated carbon paper (Fig. 2(a)(1)), the deposited Pd (Fig. 2(a)(2)) and Pd/C (Fig. 2(b)(1)) samples without PANI are also shown for comparison. The peak at 40.3° was observed in all of the nPANI/Pd and Pd/C/nPANI samples, and corresponds to the diffraction at the (111) planes of the face-centered cubic structure of metallic Pd (PDF #46-1043) [23]. Other broad peaks associated with an overlap of peaks at 2θ = 42.3° and 44.4° correspond to the (100) and (101) planes of the graphite carbon, respectively (PDF #41-1487) [24], and these peaks showed the same positions as those of the uncoated carbon paper. There were no other diffraction peaks from the nPANI/Pd and Pd/C/nPANI samples, suggesting that the PANI electropolymerized on the electrode was in an amorphous state.
Fig. 3 shows representative FTIR spectra of the 15PANI and 15PANI/Pd samples. Characteristic infrared peaks corresponding to the benzenoid and quinoid forms of the aromatic phenyl ring system at the fingerprint region between 800 and 1700 cm−1 were observed. Absorption bands at around 802 cm−1 were attributed to the bending of C-H on the benzene ring out-of-plane [25]. The peaks at 1298, 1483, and 1566 cm−1 were for the stretching of aromatic-N, stretching of N-benzene ring-N, and stretching of N=quinoid ring=N [26, 27]. Nearly all of the above mentioned functional groups on PANI were observed on nPANI/Pd, but with slight changes, which may be because of the charge-transfer between the Pd NPs and PANI.
The CV method was used to electropolymerize PANI because it is controllable. The thickness of PANI can be tuned through different numbers of potential cycling curves. Fig. 4(a) represents the successive CV curves during the course of the electropolymerization of aniline in H2SO4 (0.5 mol/L) and aniline (0.05 mol/L) (30 cycles). The potential range was from −0.2 to 0.9 V, and the scan rate was 50 mV/s. The initial cycles did not show any redox signals in the whole potential range, and three couples of defined redox peaks were gradually observed with increasing potential cycling curves. The first peak at around 0.21 V was associated with the leucoemeraldine-emeraldine transition, the second at around 0.50 V corresponds to the oxidation of a head-to-tail dimer, and the third peak at the highest potential of 0.82 V was attributed to the conversion of emeraldine to pernigraniline. These CV curve profiles are similar to those reported in the literature [28, 29]. Upon sequential cycles, the redox peaks gradually increased, suggesting the formation of an electroactive and conductive layer on the surface of the electrode.
After the PANI films were formed, Pd NPs were electrodeposited by CV from −0.20 to 0.60 V in H2SO4 (0.5 mol/L) and K2PdCl4 (1 mmol/L) solution for 20 cycles at a scan rate of 50 mV/s. Fig. 4(b), (c), and (d) shows successive CV curves (20 cycles) during the electrosynthesis of Pd NPs on PANI formed after 10, 20, and 30 potential cycles, respectively. Compared with the redox peaks of PANI without the decoration of Pd NPs, those of PANI in nPANI/Pd samples showed a lower intensity. For example, the current density of the 30PANI/Pd oxidation peak at 0.23 V was 0.58 mA/cm2 (Fig. 4(d)), which was much lower than that of the 30PANI sample (0.73 mA/cm2, Fig. 4(a)). For all of the nPANI/Pd samples, the peaks related to the hydrogen adsorption/desorption of Pd on the low potential side of CV were more pronounced with increasing cycle numbers, showing the increases in the Pd amount on the electrode and the high surface area of Pd NPs. In addition, the intensity of hydrogen adsorption/desorption peaks of the 20th CV curve in Fig. 4(b), (c), and (d) were different, even though the cycle numbers of Pd electrodeposition were the same (20 cycles). The intensity of the hydrogen adsorption/desorption peaks is strongly dependent on the thickness of the PANI substrate. With increasing the cycle numbers of PANI from 10 to 30, the intensity of hydrogen adsorption/desorption peaks gradually decreased. These results suggest that a thick PANI film leads to the Pd NPs embedding into the underlying PANI network and thus cannot be used for the catalysis. For the optimal efficiency of Pd NPs for catalytic reactions, the appropriate thickness of Pd film is necessary and crucial. The proximity between PANI and Pd NPs is important to enhance Pd utilization and may also be important for the catalytic activity of Pd.
Fig. 5(a) compares the cyclic voltammograms of the nPANI/Pd samples in N2-saturated H2SO4 (0.5 mol/L) solution. Increasing the cycle numbers of PANI electropolymerization gradually increased the current density of the redox peaks associated with PANI. In addition, significant changes in the hydrogen adsorption/desorption peaks of the nPANI/Pd samples were observed that were not seen with those of the Pd electrode. The hydrogen adsorption/desorption peak was very broad for the Pd electrode, which was similar to those reported for bulk Pd or Pd wire electrodes [16, 30], whereas the profiles of the peaks for the nPANI/Pd samples were well-defined. Because the exposed PANI surface would be inactive for hydrogen adsorption/desorption, the sharply increased peaks may be a consequence of the highly dispersed Pd NPs. The hydrogen atoms are adsorbed on the surface of Pd and intermixed with the inner layer of Pd NPs (absorption) [31, 32]. The hydrogen region can then be divided into two parts: the most intensive peaks (−0.13 V) on the low potential side of the CV curve originating from the oxidation of adsorbed hydrogen; and the flat peaks at relatively high potentials (−0.10 to 0.05 V) related to the oxidation of absorbed hydrogen on the surface of Pd [16].
Fig. 5(b) shows the CV curves of the nPANI/Pd catalysts for FAOR in a H2SO4 (0.5 mol/L) and HCOOH (2.0 mol/L) solution. The data of the reference Pd catalyst were also included for comparison. Besides the hydrogen region on the low potential side and the redox peaks corresponding to PANI, anodic peaks from the oxidation of formic acid on Pd catalysts at around 0.05 V were observed for all the nPANI/Pd catalysts. The inset of Fig. 5(b) shows a close-up of the positive-going CV curves of formic acid oxidation. The current density of FAOR on the nPANI/Pd catalysts showed a significant increase compared with that of the reference Pd catalyst. When the 10PANI film was used as the substrate, the current density of FAOR (at 0.05 V) on 10PANI/Pd was 1.41 mA/cm2, which was ~5.2 times that of the reference Pd catalyst (0.27 mA/cm2). Increasing the cycle number to 15 (15PANI/Pd), the current density of FAOR increased to 1.70 mA/cm2, which was ~6.3 times that of the Pd catalyst. When the cycle number of PANI was further increased to 20 and again to 30, the current density of FAOR did not increase but decreased. There is a volcano-shaped dependence of the catalytic activity of Pd for FAOR on the cycle numbers of PANI electropolymerization. Because the mass of Pd on each nPANI/Pd electrode was slightly different, the mass-specific activity (MSA) was appropriate for comparing catalytic activities. The data are listed in Table 1. The highest catalytic activity of Pd in the nPANI/Pd catalysts was that of 15PANI/Pd, the MSA was 0.209 A/mg, which is 7.5 times that of the Pd catalyst (0.028 A/mg). This result may be because of the proximity between the deposited Pd NPs and PANI. Because PANI is inactive for the oxidation of formic acid, the enhanced catalytic properties may arise from a promoting effect of PANI [16].
To further evaluate the promoting effects of PANI on Pd, a commercial Pd/C catalyst was used as the substrate for PANI coating (denoted as Pd/C/nPANI). Fig. 6(a) shows 20 cycles of electrochemical polymerization of aniline on the Pd/C catalyst in a solution of H2SO4 (0.5 mol/L) and aniline (0.05 mol/L). Similar to the profiles of the CV curves for the polymerization of aniline on a glassy carbon electrode, three couples of peaks corresponding to the redox process of PANI were observed. In addition, pronounced peaks associated with hydrogen adsorption/desorption on the Pd catalyst were also detected in the low potential side of the CV curves. Fig. 6(b) shows the comparison of the CV curves of Pd/C/nPANI catalysts in H2SO4 (0.5 mol/L) (n = 0-40). Increasing the cycle numbers of PANI polymerization showed a slight decrease in the intensity of hydrogen adsorption/desorption peaks, which may be because of the partial overlapping of the PANI networks on the surface of Pd NPs. Because the decrease in the intensity of hydrogen adsorption/desorption peaks is not pronounced, most of the PANI network may be embedded with Pd NPs, which leads to a close proximity between PANI fiber and Pd NPs and enables a promoting effect of PANI on Pd catalysts.
Fig. 7 shows the positive-going FAOR CV curves of Pd/C/nPANI catalysts. The Pd/C/nPANI peaks at around 0.10 V were attributed to the oxidation of formic acid, which were more negative than that of the Pd/C catalyst (0.18 V). In addition, most of the oxidation current density of the Pd/C/nPANI catalysts show an increase compared with that of the Pd/C catalyst, showing a promoting effect of PANI to Pd for FAOR. The Pd/C/nPANI catalysts retained a volcano-shape dependence between the MSA of Pd and the cycle numbers of PANI polymerization as shown in Table 1. Increasing the cycle numbers from 5 to 10 and then to 20 increased the MSA of Pd (at 0.05 V) from 0.085 to 0.112 and then to 0.135 A/mg, respectively, and further increases to 30 and 40 cycles decreased the current density of FAOR. For the Pd/C/40PANI catalyst, the signal corresponding to the FAOR was difficult to detect, and at the same time, the peak at around 0.22 V corresponding to the redox process of PANI became pronounced, indicating that the Pd NPs were almost completely covered by PANI. Among the Pd/C/nPANI catalysts, Pd/C/20PANI showed the highest MSA for FAOR at 0.135 A/mg, which was 2.3 times that of the commercial Pd/C catalyst at 0.059 A/mg. These data indicated that PANI shows a significant promoting effect of Pd for the FAOR, which is similar to the results of the nPANI/Pd catalysts described above. This enhanced performance of Pd may be closely related to the change in the electronic structure of Pd in the presence of PANI.
To identify the change in electronic properties of Pd catalysts, XPS was employed. Fig. 8 shows the XPS spectra of the 15PANI/Pd and Pd/C/20PANI catalysts. Results of the corresponding reference catalysts included the deposited Pd without PANI and commercial Pd/C after potential cycling from −0.2 to 0.9 V. The vertical dashed lines refer to the binding energy (BE) positions for the reference Pd catalysts without PANI. For the 15PANI/Pd catalyst, the Pd 3d signals can be deconvoluted into two doublets. The intense doublet at BE = 335.5 and 340.8 eV was characteristic of Pd(0), and another doublet (BE = 336.4 and 341.7 eV) indicated Pd(II). Fig. 8 shows that the BE positions of Pd(0) for the 15PANI/Pd catalyst shifted positively by ~0.3 eV compared with those of the Pd catalyst (BE = 335.2 and 340.5 eV), suggesting that there is electron transfer from Pd to PANI, and Pd is in an electron deficiency state. Similar results were also observed for the Pd/C/20PANI and Pd/C catalysts. The electron delocalization between the Pd NPs and PANI alters the electronic structure of Pd, resulting in an improved catalytic activity of Pd for FAOR [33].
CO stripping voltammograms were used to evaluate the CO tolerance and electrochemically active surface areas (EAS) of Pd. Because the electrodeposited quantity of Pd on the electrode was low (1.60-1.80 μg) for the nPANI/Pd catalysts, the CO oxidation peaks were overlapped with oxidation peaks of PANI and the CO stripping voltammograms of nPANI/Pd are not shown in this study. Fig. 9 shows representative CO stripping voltammograms of the Pd/C/nPANI catalysts coated on the electrodes. The voltammetry of the pre-adsorbed CO was carried out in H2SO4 (0.5 mol/L) for 15 min. Results for the corresponding reference Pd catalysts are also shown for comparison. The anodic peaks at around 0.23 and 0.49 V were associated with the oxidation of PANI, and the pronounced peaks at around 0.78 V were from the oxidative removal of CO monolayer on the electrode. Profiles of CO oxidation on the Pd/C/nPANI catalysts were similar to those on the Pd/C catalyst, whereas the onset and peak potentials showed a slight difference. Pd/C/20PANI and Pd/C/30PANI catalysts showed the most CO oxidation negative peak potential (0.76 V), which was ~20 mV more negative than those of the other Pd/C/nPANI and Pd/C catalysts. These observations suggest that the presence of PANI improves the CO tolerance of Pd, and interaction between Pd and PANI was crucial for this improvement. In addition, the EAS of Pd can be calculated by integrating the CO stripping peaks, followed by subtracting the area arising from the double layer charging and oxidation formation, which was measured from the second cycle after the oxidative removal of CO monolayer (Table 1). For the Pd/C/nPANI catalysts, all of the EAS values of Pd were lower than that of the Pd/C catalyst (56.0 m2/g) because of the PANI coating. Increasing electropolymerization cycles of PANI gradually decreased the EAS of Pd in the Pd/C/nPANI catalysts from 43.0 m2/g at n = 5 to 39.2 m2/g and 36.5 m2/g at n = 10 and 30, respectively. When the electropolymerization cycles of PANI increased to 40, the EAS data decreased to 33.3 m2/g. The intrinsic activity (IA) of Pd in the Pd/C/nPANI catalysts was obtained by normalizing the current density of FAOR to the EAS of Pd. Except for Pd/C/40PANI, all of the Pd/C/nPANI catalysts showed improved IA compared with that of the commercial Pd/C catalyst, indicating that PANI can act as a promoter to Pd for FAOR. Among the Pd/C/nPANI catalysts, Pd/C/20PANI showed the highest IA, which was 3.3 times that of the Pd/C catalyst. The enhanced catalytic activity may be associated with the proximity of Pd and PANI and the change in the electronic structure of Pd.
Chronoamperometric analyses were carried out to evaluate the activity and stability of 15PANI/Pd, Pd/C/20PANI, and the reference Pd and Pd/C catalysts. Fig. 10 shows the chronoamperograms (at 0 V) of the catalysts in H2SO4 (0.5 mol/L) and HCOOH (2.0 mol/L) solution. Because the loadings of Pd on the electrode for the Pd/C/20PANI and Pd/C catalysts were higher than those for the 15PANI/Pd and Pd catalysts, the initial oxidation current density was also higher. For the Pd/C/20PANI catalyst, both the initial and the steady-state oxidation current density were higher than those for the Pd/C catalyst, the current density on Pd/C/20PANI catalyst at 2000 s (1.49 mA/cm2) was around 50 times that on the Pd/C catalyst (0.03 mA/cm2). Similar results were also found for the 15PANI/Pd and Pd catalysts, where the current densities at 2000 s were 0.22 and 0.02 mA/cm2 (inset of Fig. 10), respectively. The MSA of 15PANI/Pd (0.027 A/mg) was 12.9 times that of the Pd catalyst (0.002 A/mg). These data suggest that the presence of PANI in Pd catalysts enhances the catalytic activity and stability for FAOR. These results are consistent with the above voltammetric results. The improved electronic conductivity of the composite catalysts because of the presence of PANI may be one of the reasons for the enhanced catalytic properties of Pd [25], and comprehensive investigations need to be carried out in future studies.
This study reports two PANI-promoted Pd catalysts for FAOR. PANI alone has no catalytic property for FAOR. However, PANI shows significant promoting effects on Pd catalysts. These promoting effects are strongly dependent on the numbers of potential cycles for the polymerization of PANI. For both the nPANI/Pd and Pd/C/nPANI catalysts, the current density of FAOR shows distinct enhancements compared with those of the Pd reference catalysts. The highest current density of FAOR on the nPANI/Pd and Pd/C/nPANI catalysts are for 15PANI/Pd and Pd/C/20PANI, respectively. These findings show important application for the design and preparation of electrocatalysts for the FAOR and the development of DFAFCs and other catalytic technology.
2015, Vol. 36 Issue (7): 943-951
PDF (1542 KB)
2015, Vol. 36 Issue (7): 943-951
PDF (1542 KB)