Nitrogen-doped carbon black supported Pd nanoparticles as an effective catalyst for formic acid electro-oxidation reaction

Abstract

Pd nanoparticles supported on nitrogen doped carbon black (Vulcan XC-72R) with two different levels of doping were prepared by the microwave-assisted ethylene glycol reduction process and used as catalyst for the formic acid electro-oxidation (FAEO). The results indicate that the different nitrogen doping contents in Pd/N-C catalysts have a significant effect on the performance of FAEO. A higher N content facilitates the uniform dispersion of Pd nanoparticles on carbon black with narrow particle size distribution. Furthermore, the electrochemical results show that the catalyst with a higher N-doping content possesses a higher catalytic activity and a long-term stability for FAEO. The peak current density of the Pd/N-C (high) catalyst is 1.27 and 2.31 times that of the Pd/N-C (low) and homemade Pd/C-H catalyst. The present paper may provide a simple method for preparation of high-performance anode catalyst for direct formic acid fuel cells (DFAFCs).

Keywords

formic acid electro-oxidation / nitrogen doped / oxidized carbon / nitrogen content

Introduction

Direct formic acid fuel cells (DFAFCs), a clean and efficient energy conversion technology, have received considerable attention due to the advantages of high electromotive force (theoretical open circuit potential 1.48 V), limited fuel crossover through Nafion membrane, and reasonable power densities at low temperatures [1–3]. The anode reaction of DFAFCs can be described as HCOOH → CO₂ + 2H⁺ + 2e⁻. In an effort to reduce indirect oxidation of FA with CO as reaction intermediate, the cost of catalyst used for FAEO, Pd catalysts has been extensively studied [4–6]. Pd-black generates much higher power densities compared with Pt-based catalysts [7]. Carbon black (e.g. Vulcan XC-72R) is widely used as catalyst support due to its good electronic conductivity, high surface area, and appropriate pore structure. Many papers have demonstrated that carbon-based support materials can enhance the catalyst activity and stability by doping nitrogen [8–14]. Because of the strong covalent bond between carbon and nitrogen and the comparable atomic size of both elements, it has been experimentally found that nitrogen atoms can replace carbon atoms at different sites and result in the formation of defects [15–16]. Bae et al. [17] have used nitrogen-containing carbon materials as a support for Pt catalyst (Pt/N-C), which exhibits increased activity and stability towards electrochemical hydrogen oxidation. Xiong [7] and his co-workers have prepared Pt nanoparticles deposited on nitrogen-doped graphene (Pt/N-G) as an effective and robust catalyst material for methanol oxidation reaction (MOR). Recently, Chang et al. [18] have used nitrogen-doped acetylene carbon black supported Pd nanocatalyst as a catalyst for formic acid electro-oxidation, and found that the heat-treatment temperature of N-doped C support affect the performance of corresponding catalyst. When the heat-treatment temperature is as high as 900°C, the corresponding catalyst, Pd/N-C-900 exhibits the largest ECSA and better CO tolerance compared with other counterparts. However, the catalysts based on Pd nanoparticles supported on N-doped carbon with controllable N content with facile method have rarely been reported.

Herein, a novel and facile strategy was developed for the synthesis of nitrogen doped carbon black supported Pd nanoparticles with superior electrocatalytic activity by hydrothermal treatment of oxidized carbon with ammonia. Compared to the palladium on carbon (Pd/C-H) catalyst (41.3 m²/g), both the Pd/N-C (high) with a higher N content and the Pd/N-C (low) with a lower N content catalysts showed a larger electrochemically active surface area (49.9 m²/g and 41.5 m²/g), which endowed the nitrogen-doped catalysts with an excellent electrocatalytic activity for formic acid electro-oxidation. The peak current density of the Pd/N-C (high) and the Pd/N-C (low) catalysts were 1294 and 1019 mA/mg_Pd, which was about 1.31 and 0.82 times higher than that of the Pd/C-H catalyst (560 mA/mg_Pd), respectively. Furthermore, the nitrogen-doped carbon black supported Pd nanoparticles showed long-term stable formic acid oxidation than the undoped sample.

Experimental

Catalyst preparation

Synthesis of oxidized Vulcan XC-72R

The oxidized Vulcan XC-72R was synthesized using a modified Hummers’ method as previously reported [19]. Briefly, one gram of Vulcan XC-72R was dispersed in 23 mL of concentrated H₂SO₄ in a 250 mL flask under vigorous stirring at room temperature for 12 h. The flask was then transferred to an oil bath before 100 mg of NaNO₃ was added to the solution. After 5 minutes, one gram of KMnO₄ was added slowly to keep the reaction temperature below 40°C. Thirty minutes later, 3 mL of H₂O was added to the flask, after 5 min another 3 mL H₂O was added, followed by the addition of 40 mL of H₂O after another 5 minutes. The suspension was stirred for 15 min, and then the flask was removed from the oil bath. The mixture of 140 mL of H₂O and 10 mL of 30% H₂O₂ was added to the solution to end the reaction. After filtered and washed with 5% HCl and H₂O, the resulting material was dried at 55°C for 12 h.

Synthesis of N-doped Vulcan XC-72R (N-C)

The N-C (high) sample was prepared by the hydrothermal method of oxidized carbon and ammonia. The mixture of 50 mg of oxidized Vulcan XC-72R, 1 mL of NH₃·H₂O and 19 mL of H₂O were sonicated under ultrasonic to form a homogeneous solution. The solution was treated at 180°C for 12 h, filtered and washed with deionized water until the supernatant showed neutral PH. The sample N-C (low) was also prepared using the same method mentioned above, where the only difference was that the oxidized Vulcan XC-72R was replaced with the untreated Vulcan XC-72R.

Synthesis of N-doped Vulcan XC-72R supported Pd nanoparticles

The Pd catalyst support on nitrogen doped Vulcan XC-72R were synthesized by using the microwave-assisted ethylene glycol reduction process. The Pd content in each sample was 20% (wt). Briefly, 80 mg of N-C was dispersed into 50 mL of ethylene glycol in a beaker ultrasonic treatment and stirred for 0.5 h to form a uniform ink. Then, the H₂PdCl₄ solution (contain 20 mg Pd) was added and the pH value of the ink was adjusted to approximately 11 using the 1 mol/L NaOH aqueous solution. Subsequently, the breaker was placed in the middle of the microwave oven (LGMG-5021MW1, 2450 MHz) with 700 W with 50 s and 10 s off cycle for three times, and the solution was stirred for 3h and cooled to room temperature naturally. At last, the suspension was washed and dried in a vacuum to obtain the Pd/N-C catalyst. The palladium catalyst supported on the untreated Vulcan XC-72R (denoted as Pd/C-H) and the oxidized Vulcan XC-72R (denoted as Pd/O-C) was prepared by using the same method except for the fact that the matrix is different for comparison of the nitrogen and oxygen doping effect. The normally content of the Pd on all catalysts was 20%, and the ICP was used to detect the real content.

Electrochemical measurements

All electrochemical measurements were performed in a standard three-electrode cell using an electrochemical workstation at room temperature. The catalyst coated glassy carbon disk was used as the working electrode, a saturated calomel electrode (SCE) was used as the reference electrode, and a Pt foil as the counter electrode. All potentials in this report referred to SCE. Prior to any measurement, all electrolyte solutions were de-aerated by continuously purging with pure N₂ for at least 20 min. The electrode potential was performed with an EG&G PARC potentiostat/galvanostat (Model 273A Princeton Applied Research Co., USA) system.

The working electrode was prepared as follows. First, 5 mg of prepared catalyst powder was dispersed in the mixed solution (950mL ethanol, 50 mL of 5% Nafion solution) with ultrasonication for 30 min to form a homogeneous ink. Second, 5 μL of the suspension was pipetted onto the flat glassy carbon electrode with a diameter of 4 mm. At last, the coated electrode was dried at room temperature for about 30 min. The glassy carbon electrode was polished with alumina slurry of 0.5 and 0.03mm successively before used.

All electrochemical measurements were conducted in a 0.5 mol/L H₂SO₄ solution with or without 0.5 mol/L HCOOH, and the solutions were deaerated by pure nitrogen for at least 20 min prior to any measurement. For the electro-oxidation of formic acid, the potential ranged from ‒0.2 to 1.0 V. For further exploration, the COad stripping voltammograms were measured in a 0.5 mol/L H₂SO₄ solution and the working electrode was kept at 0.2 V. In order to enable the CO to be absorbed onto the catalyst completely, the CO was purged into the 0.5 mol/L H₂SO₄ solution for at least 15 min, and the excess in the electrolyte was purged out with N₂ for 15 min. The amount of the COad was evaluated by integration of the COad stripping peak. All the measurements were made at room temperature and the stable results were reported.

Physical characterization

The particle morphology and the size of the catalyst were characterized by using a transmission electron microscopy (TEM) operating at 200 kV (Philips TECNAI G2). X-ray diffraction (XRD) measurements were performed with a PW-1700 diffraction using a Cu Ka (l=1.5405 À) radiation source (Philips Co). X-ray photoelectron spectroscopy (XPS) measurements were carried out on a Mg Ka radiation source (Kratos XSAM-800 spectrometer). The bulk compositions were evaluated using an inductively coupled plasma-optical emission spectrometer (ICP-OES, X Series 2, Thermo Scientific USA).

Results and discussion

Figure 1 shows the structure characterization of all catalysts. As shown in Fig. 1(a), the XRD patterns of the three catalysts exhibit a broad peak at 2q=24.8º, which can be assigned to the (002) reflection peak of the hexagonal structure of graphite. The 2q peaks at 40.3º, 46.7º, and 68.3º are corresponding to the (111), (200), and (220) crystal planes of the face centered cubic crystalline structure of the Pd, respectively. Figure 1(b) shows the X-ray photoelectron spectroscopy (XPS), where the total amounts of nitrogen in the Pd/N-C(high) and Pd/N-C(low) catalysts are 1.95% (at) and 0.62% (at), respectively. The reason for the higher level of doping in the Pd/N-C (high) catalysts can be explained by the fact that the oxidation process for the Vulcan XC-72R through the Hummer method leads to the formation of oxygen-containing functional groups and more defects, thereby creating more sites for the anchoring of nitrogen. As illustrated in Fig. 1(c), 4 N-containing groups are present in the Pd/N-C(high) catalyst, ascribing to pyridinic N (N1, 398.1 eV), amino N (N2, 399.6 eV), pyrolic N (N3, 400.5 eV) and graphitic N or quaternary N (N4, 403 eV) respectively. Further, the XPS spectra of N in the samples of Pd/N-C (high) is deconvoluted, and it is found out that the pyridinic N and pyrolic N are the main nitrogen containing component of Pd/N-C(high) catalyst, which serves as the active sites for anchoring the metal/alloy NPs [20]. Figure 1(d) depicts the Pd 3d XPS spectra. For the Pd/N-C(high) catalyst, the peaks at 334.9 eV and 340.1 eV are assigned to metallic Pd, while the peaks at 336.3 eV and 341.6 eV correspond to Pd(II) species. For the Pd/N-C (low) catalyst, the peaks at 335.5 and 340.8 eV correspond to metallic Pd, and the peaks at 336.45 and 341.75 eV correspond to Pd(II) species. For the counterpart Pd/C-H catalyst, both the metallic Pd peaks (335.8 and 341.1 eV) and the Pd(II) peaks (336.4 and 342.8 eV) are found at higher values. The negative shift of Pd binding energies for both Pd/N-C catalysts can be attributed to the N doping effect, in which N serves as an electron donor. Therefore, the partial electron transfer from N to Pd decreases the electron affinity of Pd, which may alter the catalytic behaviour towards formaic acid oxidation [20].

The TEM characterization was used to estimate the influence of the nitrogen on the morphology, as well as the size and the dispersion of the Pd nanoparticle. Figure 2 demonstrates the TEM images of Pd/N-C (high), Pd/N-C (low), Pd/C-H and Pd/O-C catalysts. It can be seen clearly from the TEM images that both the Pd/C-H and Pd/O-C show obvious aggregation, while the Pd particles on the N-C support are much smaller and more uniformly dispersed than those on the carbon support without nitrogen. This suggests that the addition of nitrogen into carbon support can effectively inhibit the aggregation of particles, which leads to much a smaller particle size and a much higher particle density [18]. The average particle sizes of Pd/N-C (high), Pd/N-C (low) and Pd/C-H and Pd/O-C catalysts are 3.6 nm, 4.09 nm, 4.22 nm and 4.34 nm, respectively, clearly demonstrating the positive effect of N doping on reducing the particle size. The ICP results indicate that the mass content of Pd in Pd/N-C (high), Pd/N-C (low), Pd/O-C and Pd/C-H is 19.86%, 19.79%, 19.95% and 18.83% respectively, which is consistent with the normal content (20%), thus the influence of mass loading of Pd can be excluded.

Figure 3 is an estimation of electrochemical surface area for all catalysts. Figure 3(a) shows the cyclic voltammograms (CVs) of the Pd/N-C (high), Pd/N-C (low), Pd/O-C and Pd/C-H catalysts in the 0.5 mol/L H₂SO₄ solution at a scan rate of 50 mV/s between –0.2 V and 1.0 V. As exhibited in Fig. 3(a), the four catalysts exhibit hydrogen under the potential deposition behaviour (–0.2‒0.05 V) and the Pd redox behaviour (>0.25 V), with current response increases in the sequence of Pd/C-H<Pd/O-C<Pd/N-C (low)<Pd/N-C(high), indicating the highest electrochemical surface area (ECSA) of the Pd/N-C(high) catalyst. Due to the effect of hydrogen spillover and H₂ absorption in Pd crystal lattice, the ECSA calculated from the hydrogen using the potential deposition method (Fig. 3(a)) only provides rough information for the surface area. The more accurate ECSA data were obtained from CO-stripping experiments, as shown in Fig. 4. The ECSA values of electrocatalysts were calculated by integrating the charges (Q) associated with the peak from CO-stripping, assuming 0.420 mC/cm was needed for the removal of the CO monolayer [21]: ECSAco=Qco/( 0.420×[Pd]), where [Pd] indicates the Pd loading on the electrode. By calculation, the ECSA of the catalysts is 49.9, 41.8, 41.3 and 34.6 m²/g for Pd/N-C (high), Pd/N-C (low), Pd/C-H and Pd/O-C, respectively. The ECSA of Pd/N-C (high) is approximately 1.2 times that of the Pd/C-H catalyst (41.3 m²/g). This result is consistent with the results obtained by the hydrogen using the potential deposition method, further suggesting the positive role of N content in reducing the particle size of Pd. It is worth noting that the much higher current density of Pd/O-C may come from the much bigger electrical double-layer capacitor (Fig. 3(a)), considering the fact that the ECSA of the catalysts is 49.9 m²/g, 41.8 m²/g, 41.3m²/g and 34.6 m²/g for Pd/N-C (high), Pd/N-C (low), Pd/C-H and Pd/O-C, respectively. It can be observed that the much bigger ECSA does not come from oxygen-containing groups but from N-doped. Hence, the oxygen-containing functional groups may enhance the double-layer capacitor, but cannot lead to the uniform distribution of Pd nanoparticles (Fig. 2(d)) or higher Pd loading. Thus, the oxidized Vulcan XC-72R was used as precursor for easily preparing N-doped Vulcan XC-72R with a higher N content.

Figure 5 displays the CVs of different catalysts in the 0.5 mol/L H₂SO₄ + 0.5 mol/L HCOOH solution at a scan rate of 50 mV/s. For all Pd-based catalysts, an anodic peak appears at 0.3 V in both forward and reverse scans. Clearly, the peak current density of the catalysts shows an order of Pd/C-H<Pd/O-C<Pd/N-C(low)<Pd/N-C (high). Particularly, the peak current density of Pd/N-C (high) and Pd/N-C (low) catalysts are 1294 and 1019 mA/mg_Pd, which are about 1.31 and 0.82 times higher than the peak current density of the Pd/C-H catalyst (560 mA/mg_Pd), respectively. While the peak current density of Pd/O-C is 908 mA/mg_Pd, which is higher than that of Pd/C-H but lower than those of Pd/N-C(low) and Pd/N-C (high). Thus, the catalysts doped with nitrogen have a better performance than the catalysts without doping and oxygen doping. Moreover, a higher content of N doping leads to a higher formic acid oxidation activity. This corresponds well to the increase in Pd utilization, as evidenced in the ECSA measurements. Furthermore, as proved by the XPS, with the increase of nitrogen content, a stronger interaction between Pd and the support is observed, which may indicate a higher stability of the catalysts.

Chronoamperometric curves were used to compare the electrochemical stabilities of different catalysts for formic acid electro-oxidation. Figure 6(a) shows the chronoamperometric curves of Pd/N-C(high), Pd/N-C(low) and Pd/C-H catalysts in 0.5 mol/L H₂SO₄ containing 0.5 mol/L HCOOH solution at 0.2 V. The Pd/N-C (high) exhibits a higher current density and a slower current decay than the Pd/N-C(low) and Pd/C-H, which indicates a better tolerance to the intermediate poisoning species during the process of formic acid electro-oxidation [22]. It can be seen that the current density of Pd/N-C(high), Pd/N-C(low) and Pd/C-H catalysts is 284, 133 and 74 mA/mg_Pd at 1000 s. All the curves reach a steady state after 3500 s and the current density of Pd/N-C(high), Pd/N-C(low) and Pd/C-H catalysts is 118, 54 and 27 mA/mg_Pd. Even at 6000 s, the current density of Pd/N-C (high) is still 75 mA/mg_Pd, which is 0.9 times higher than that of Pd/N-C(low) (39 mA/mg_Pd) and 5.8 times higher than Pd/C-H (11 mA/mg_Pd). Thus, Pd/N-C(high) remains a higher current density than the other two catalysts, which indicates that the Pd/N-C(high) catalyst has a much better activity and anti-poisoning feature than Pd/N-C(low) and homemade Pd/C.

The EIS technique has been used to probe the interfacial processes and kinetics of electrode reactions in electrochemical system. The formic acid electro-oxidation on different catalysts at different potentials shows different impedance patterns. The Nyquist plots of the three catalysts are illustrated in Fig. 6(b). The diameter of the EIS semicircle correlates with the charge transfer resistance. As the diameter decreases, the resistance decreases, and the formic acid oxidation reaction rate increases. It can be seen from Fig. 6(b) that the impedance for the Pd/N-C (high) catalyst has the smallest diameter. Therefore, the Pd/N-C (high) catalyst has the lowest resistance and the highest activity for formic acid oxidation compared to the other two duplicates.

Conclusions

In summary, the Pd catalysts supported on nitrogen-doped carbon black has been prepared with different nitrogen contents. With the increase of nitrogen contents in the carbon support, the catalysts show a larger electrochemical active surface area and a better catalytic activity and stability. The enhanced performance is mainly attributed to the higher metal utilization ratio and the electronic structure modification by the nitrogen functional groups. The amount of N in the carbon matrix plays a critical role in adjusting the diameter and dispersion of Pd NPs, as well as modifying the binding energy of Pd 3d core levels, thus improving the electrocatalytic activity and stability of Pd NPs. The present work provides a simple method to prepare Pd/N-C catalyst with different N contents, which will be a candidate anode catalyst for future DFAFCs.

References

[1] Rice C, Haa  S, Masela R I ,  Wieckowski A . Catalysts for direct formic acid fuel cells. Journal of Power Sources, 2003, 115(2): 229–235 

[2] Wang J, Liu  Y, Okada T . Novel platinum-macrocycle composite catalysts for direct formic acid fuel cells. Journal of Applied Electrochemistry, 2016, 46(8): 901–905 

[3] Yu X, Pickup  P G. Recent advances in direct formic acid fuel cells (DFAFC). Journal of Power Sources, 2008, 182(1): 124–132 

[4] Antolini E. Palladium in fuel cell catalysis. Energy & Environmental Science, 2009, 2(9): 915–931 

[5] El-Nagar G A, Darweesh  A F, Sadiek  I. A novel nano-palladium complex anode for formic acid electro-oxidation. Electrochimica Acta, 2016, 215: 334–338 

[6] Feng L, Yao  S, Zhao X ,  Yan L, Liu  C, Xing W . Electrocatalytic properties of Pd/C catalyst for formic acid electrooxidation promoted by europium oxide. Journal of Power Sources, 2012, 197: 38–43 

[7] Xiong B, Zhou  Y, Zhao Y ,  Wang J, Chen  X, O’Hayre R ,  Shao Z. The use of nitrogen-doped graphene supporting Pt nanoparticles as a catalyst for methanol electrocatalytic oxidation. Carbon, 2013, 52: 181–192 

[8] Liang J, Hassan  M, Zhu D ,  Guo L, Bo  X. Cobalt nanoparticles/nitrogen-doped graphene with high nitrogen doping efficiency as noble metal-free electrocatalysts for oxygen reduction reaction. Journal of Colloid and Interface Science, 2017, 490: 576–586 

[9] Liu D, Li  L, You T . Superior catalytic performances of platinum nanoparticles loaded nitrogen-doped graphene toward methanol oxidation and hydrogen evolution reaction. Journal of Colloid and Interface Science, 2017, 487: 330–335 

[10] Ramakrishna S U B ,  Srinivasulu Reddy D ,  Shiva Kumar S ,  Himabindu V . Nitrogen doped CNTs supported Palladium electrocatalyst for hydrogen evolution reaction in PEM water electrolyser. International Journal of Hydrogen Energy, 2016, 41(45): 20447–20454 

[11] Xiong Q, Chi  H, Zhang J ,  Tu J. Nitrogen-doped carbon shell on metal oxides core arrays as enhanced anode for lithium ion batteries. Journal of Alloys and Compounds, 2016, 688, Part B: 729–735

[12] Xu G, Dou  H, Geng X ,  Han J, Chen  L, Zhu H . Free standing three-dimensional nitrogen-doped carbon nanowire array for high-performance supercapacitors. Chemical Engineering Journal, 2017, 308: 222–228 

[13] Xu J, Ju  Z, Cao J ,  Wang W, Wang  C, Chen Z . Microwave synthesis of nitrogen-doped mesoporous carbon/nickel-cobalt hydroxide microspheres for high-performance supercapacitors. Journal of Alloys and Compounds, 2016, 689: 489–499 

[14] Zhang Y, Chen  L, Meng Y ,  Xie J, Guo  Y, Xiao D . Lithium and sodium storage in highly ordered mesoporous nitrogen-doped carbons derived from honey. Journal of Power Sources, 2016, 335: 20–30 

[15] Wu G, Zelenay  P. Nanostructured nonprecious metal catalysts for oxygen reduction reaction. Accounts of Chemical Research, 2013, 46(8): 1878–1889 

[16] Matter P H, Zhang  L, Ozkan U S . The role of nanostructure in nitrogen-containing carbon catalysts for the oxygen reduction reaction. Journal of Catalysis, 2006, 239(1): 83–96 

[17] Bae G, Youn  D H, Han  S, Lee J S . The role of nitrogen in a carbon support on the increased activity and stability of a Pt catalyst in electrochemical hydrogen oxidation. Carbon, 2013, 51: 274–281 

[18] Chang J, Sun  X, Feng L ,  Xing W, Qin  X, Shao G . Effect of nitrogen-doped acetylene carbon black supported Pd nanocatalyst on formic acid electrooxidation. Journal of Power Sources, 2013, 239: 94–102 

[19] Xiao M, Zhu  J, Ge J ,  Liu C, Xing  W. The enhanced electrocatalytic activity and stability of supported Pt nanopartciles for methanol electro-oxidation through the optimized oxidation degree of carbon nanotubes. Journal of Power Sources, 2015, 281: 34–43 

[20] Ham D J, Pak  C, Bae G H ,  Han S, Kwon  K, Jin S A ,  Chang H ,  Choi S H ,  Lee J S . Palladium-nickel alloys loaded on tungsten carbide as platinum-free anode electrocatalysts for polymer electrolyte membrane fuel cells. Chemical Communications, 2011, 47(20): 5792–5794 

[21] Kawaguchi T, Sugimoto  W, Murakami Y ,  Takasu Y . Temperature dependence of the oxidation of carbon monoxide on carbon supported Pt, Ru, and PtRu. Electrochemistry Communications, 2004, 6(5): 480–483 

[22] Zhang J F, Xu  Y, Zhang B . Facile synthesis of 3D Pd-P nanoparticle networks with enhanced electrocatalytic performance towards formic acid electrooxidation. Chemical Communications, 2014, 50(88): 13451–13453 

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg