Facile Oxygen-promoted Synthesis of Cu, N Co-doped Carbon Composites for Oxygen Reduction

LIU Qianqian, GUO Wei, PAN Mu, TU Wenmao

(1. Hubei Key Laboratory of Fuel Cell,Wuhan University of Technology, Wuhan 430070, China; 2. State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China)

Abstract: A new strategy to fabricate oxygen-promoted Cu,N co-doped carbon (OP-CuN@C) composites is reported. The strategy consists of only two simple steps: chemical polymerization and high temperature carbonization. Electrochemical measurements were conducted to investigate the catalytic activity and mechanism of ORR on the resulting samples. All the electrochemical results indicate that OP-CuN@C exhibits the best ORR catalytic activity. The ORR onset potential of OP-CuN@C is slightly lower than that of commercial Pt/C catalyst. The good performance is attributed to the large specific surface area, high content of heteroatoms (pyridinic, graphitic nitrogen, and oxygen atom) and synergistic effect between divalent copper and nitrogen dopant.

Key words: oxygen reduction reaction; oxygen promotion; OP-CuN@C; chemical polymerization approach; high temperature carbonized method

1 Introduction

Owing to their high-power efficiency, environment friendly and high reliability, metal-air batteries and fuel cells have been regarded as one of the most promising electrochemical conversion devices. However, the whole performances of these devices are restricted by the low reaction rate of oxygen reduction reaction (ORR) on their cathode side[1]. As we all know, platinum or its alloys are the best catalysts for ORR. Due to the high cost and scarcity of platinum, alternative catalysts on the basis of non-precious metals and metal-free materials have been actively investigated[2-5].

Among these promising candidates, carbon-based materials have drawn much attention due to their good conductivity, high surface area and low cost and so on[6,7]. Lots of efforts have been devoted into improving the electrochemical performance of carbon-based materials for ORR. Although some impressive progress has been made in this area, several problems (such as insufficient active sites, poor electrochemical stability, and so forth) should be overcome to achieve the goal for the commercialization of these promising electrochemical conversion devices[8,9].

It is widely accepted that Cu2+ion containing natural compounds such as cytochrome c oxidase and laccase are the good catalysts for ORR in natural biological system[10,11]. Previous literatures reported that the onset potential of laccase modified electrode is about 1.2 Vvs. RHE[12], which is roughly equal to the equilibrium potential of ORR in acidic electrolyte solution. Unfortunately, due to the absence of appropriate mediator for sequential electron transfer, it is almost impossible for using these natural complexes as the catalyst for ORR at the cathode of electrochemical conversion devices. Therefore, many studies have been implemented to utilize the ORR catalytic ability of Cu2+ions[13-17]. For example, Kanget al[14]obtained Cu-N-C materials by carbonizing metal-loaded polyaniline or Vulcan carbon and found that the mass-normalized ORR activity of Cu-PANI-NH3catalyst was about 3.90 mA·mg-1at -0.05 V. Yuet al[15]synthesized Cu, N co-doped porous carbon (Cu-N-C) by annealing Cu containing ZIF-8 and delivered that the half wave potential (E1/2) of Cu-N-C catalysts was about -0.156 V. Volosskiyet al[17]prepared Cu- and N- codoped carbon materials by carbonizing porphyrin based metal organic frameworks and pointed out that the air etching process of carbon is beneficial to exposing active ORR sites. These findings have provided reasonable strategies for utilizing the ORR catalytic ability of Cu2+complexes. However, theses Cu2+compounds were prepared via complicated methods or with poor ORR catalytic activity.

Herein, Cu,N co-doped carbon composites were fabricated by a simple two-step approach: chemical polymerization and high temperature carbonization method. Oxygen-promoted Cu,N co-doped carbon (OPCuN@C) composite shows the most positive onset potential and highest limited current density compared to those of other prepared compounds (N@C and CuN@C). Although the onset potential of OP-CuN@C is a bit lower than that of commercial Pt/C catalysts, it exhibits better electrochemical stability than Pt/C catalysts under the same testing condition. Therefore, OP-CuN@C can be considered as one of the most likely alternative ORR catalysts for electrochemical conversion devices.

2 Experimental

2.1 Samples preparation

The sample precursors were synthesized by chemical polymerization method under different conditions. More precisely, aniline monomer (3 mL) was dispersed in deionized water (100 mL), and then 0.75 mol·mL-1(NH4)2S2O8solution as an initiator was added slowly into aniline solution. Next, the resulting solution was kept up for 24 h under continuous stirring in ice bath to obtain sample precursors. Afterwards, the obtained precursors were filtered, cleaned with deionized water and dried under vacuum at 80 ℃ for 8 h, and were marked as sample I. For the preparation of sample II, 5 mL 0.1 M CuSO4solution as a polymerization catalyst and dopant was added into aniline solution before the addition of (NH4)2S2O8solution, and all the other steps were same as the synthesis procedure of sample I. The preparation processes of sample III were same as the preparation of sample II, besides the whole polymerization process of sample III was conducted under oxygen atmosphere. It needed to be pointed that the oxygen purging process was before adding CuSO4solution. Finally, the three samples were pyrolyzed by means of gradual heating up to 900 ℃ for 2 h at a heating rate of 5 ℃·min-1under nitrogen atmosphere. The obtained products derived from sample I, II, and III were denoted as N@C, CuN@C, and OP-CuN@C, respectively.

2.2 Characterization

The D/Max-RB X-ray diffractometer with Cu Ka radiation was employed to analyze the phase composition of the materials, with which the accelerating voltage and applied current were 40 kV and 80 mA, respectively. Raman microscope (INVIA) was conducted to study the composites. The surface composition of the composites were examined by X-Ray photoelectron spectroscopy (ESCALAB 250Xi). Structure and morphology of the materials were observed by using scanning electron microscope. Specific surface area of the samples was evaluated by Brunauer-Emmet-Teller (BET) method with a Micromeritics ASAP 2020 gas adsorption apparatus.

2.3 Electrochemical measurements

The electrode was made as our previous work[4]. More specifically, 5 mg of sample was added into a mixture of 20 µL of 5% aqueous Nafion solution, 100 µL of ultrapure water and 900 µL of isopropanol. The resulting mixture was treated by ultrasonic wave in ice bath for 30 min in order to gain homogeneous catalyst ink. Then, 20 μL aliquot of the well dispersed ink was added onto the polished glassy carbon electrode to prepare working electrode. The prepared electrode was air-dried under ambient condition for 24 hours and then was immersed into electrolyte (0.1 M KOH) solution. Homemade reversible hydrogen electrode (RHE) was used as a reference electrode and a Pt foil was employed as the counter electrode. Before testing, the working electrode was electrochemically cleaned. Cyclic voltammetry (CV) tests were performed at a voltage range from 0 to 1.2 V with a scan rate of 50 mV·s-1in N2-saturated and O2-saturated electrolyte, respectively. Linear sweep voltammetry (LSV) tests were conducted at the potential range from 0.2 to 1.0 V with a scan rate of 5 mV·s-1in O2-saturated electrolyte at different rotating speeds. Electrochemical Impedance Spectra (EIS) was also recorded with the AC amplitude of 5 mV by sweeping the frequency from 0.1 Hz-1 000 kHz. The electrochemical stability of OP-CuN@C and commercial Pt/C catalysts were carried out at 0.6 V in O2-saturated electrolyte for 40 000 s via chronoamperometric measurements. All these electrochemical measurements were implemented by using CHI660E electrochemical workstation (Shanghai Chenhua Co. Ltd., China) at ambient temperature.

3 Results and discussion

3.1 Physicochemical characterization

Figs.1(a) and 1(b) show the X-ray diffraction (XRD) patterns and Raman spectra of N@C, CuN@C, OP-CuN@C，respectively. As shown in Fig.1(a), two XRD peaks are observed at about 24 ° and 43 °, which are indexed as the (002) and (100) planes of graphite, respectively[18,19]. The broaden and widen peaks indicate that all the samples exhibits an disordered and amorphous structure[20]. More information can be seen from Fig.1(b), two obvious peaks at 1 580 cm-1and 1 330 cm-1are derived from the G-band and D-band of carbon materials, respectively. As we all know, D-band is associated to the disorder arising from the anomalies and defects, while G-band is related to the phonon mode withE2gsymmetry of graphite[21]. It is also reported that the intensity ratio of D-band to G-band (ID/IG) can be used to evaluate the presence of disorder or defects in carbon materials[22]. TheID/IGvalues of N@C, CuN@C, and OP-CuN@C are calculated to be 1.02, 1.02, and 1.04, respectively. The results indicate that all the samples are the common amorphous carbon materials, agreeing well with XRD results, and also suggest that the introduction of copper ions and oxygen atmosphere would not affect the amorphous structure of all the samples.

Fig.1 XRD patterns (a) and Raman spectra (b) of N@C, CuN@C, and OP-CuN@C

To further illustrate their microstructures, SEM is employed to observe the morphology of the composites. The low- and high-resolution SEM micrographs of N@C, CuN@C, and OP-CuN@C are respectively shown in Figs.2(a-f). As can be seen in Figs.2(a-b), N@C exhibits a grid structure, which is composed of staggered carbon-based fibers. It can be seen from Figs.2(c-d) that a handful of grid structure is converted into tubular structure after the addition of Cu ions. The reason for this phenomenon is that the polyaniline film coated on the surface of carbon-based fibers, which is producedviathe oxidative polymerization of aniline monomers catalyzed by Cu ions in an aqueous solution[23], can wrap them to form a tubular structure on account of surface tension. In addition, on the basis of the Ref.[23], the mass production of polyaniline film can be realized by using a small amount of metal salts as a catalyst under oxygen atmosphere. Therefore, it can be speculated that a good deal of carbon-based materials with grid structure can be transformed into tubular products under our experiment condition, which is confirmed by the SEM results in Figs.2(f-e). As we all know, the microstructure change of materials may have impact on their surface area[24]. For this reason, nitrogen adsorption-desorption is performed to obtain the surface area of the prepared samples.

Fig.2 Low and high-resolution SEM images of N@C (a, b), CuN@C (c, d), and OP-CuN@C (e, f)

Fig.3 shows the nitrogen adsorption-desorption isotherms of N@C, CuN@C, and OP-CuN@C. In line with the hysteresis loop features of the nitrogen adsorption-desorption isotherms in Fig.3, the type of all the isotherms is categorized as type IV and the type of hysteresis loop is classified as type H2, which contains inkbottle-shaped pores and a small amount of parallel-plate pores or cylindrical pores[25]. It is calculated that the surface area of OP-CuN@C is 639.72 m2·g-1, higher than N@C (381.67 m2·g-1) and CuN@C (438.87 m2·g-1). The enhanced surface area of OP-CuN@C may be ascribed to the formation of tubular structure, which is beneficial to exposing more active sites and improving the ORR activity.

Fig.3 Nitrogen adsorption-desorption isotherms of N@C, CuN@C, and OP-CuN@C

On the other hand, the surface elemental composition of the samples is characterized by using XPS. Figs.4(a-d) show the XPS spectra of composites, and the high-resolution spectra of carbon, copper, and nitrogen, respectively. From Fig.4(a), N@C is composed of carbon, nitrogen and oxygen species. Besides including the above-mentioned three species, CuN@C and OPCuN@C also contain copper species. No sulfur species is detected, meaning that it is completely eliminated during the heating process. As listed in Table 1, the content of oxygen species in OP-CuN@C (5.36%) is higher than that in N@C (4.73%) and CuN@C (4.31%), at the same time, its copper species content (0.21%) is slightly lower than CuN@C (0.28%). This result shows that the chemical polymerization of aniline monomers occurred under oxygen atmosphere can promote oxygen atom to incorporate into carbon matrix.

Fig.4 XPS spectra of N@C, CuN@C, and OP-CuN@C(a); The high-resolution spectra of carbon(b), copper(c), and nitrogen(d)

Table 1 Percentages of various carbon and nitrogen states for all the samples obtained from the deconvolution of XPS C 1s and N 1s peaks, respectively

The high resolution XPS spectra of carbon is displayed in Fig.4(b). Peaks of C sp2 at 284.6 eV and C sp3 at 285.3 eV are ascribed to the graphitic carbon and amorphous carbon, respectively. The other peaks at 286.2 and 287.2 eV are characteristic peaks of carbon atoms bonded to an oxygen atom through a single bond and a double bond, respectively. As expected, it is obtained from Table 1 that OP-CuN@C has the highest total content (30.9%) of C-O and C=O among all the samples. More information can be obtained from the high resolution XPS spectra of copper (Fig.4(c)). The two peaks for CuN@C at about 932.8 and 952.7 eV are ascribed to the 2p1/2and 2p3/2peaks of Cu+ions, respectively[26], whereas the two peaks for OP-CuN@C at about 933.9 and 954.1 eV are attributed to the 2p1/2and 2p3/2peaks of Cu2+ions, respectively[27]. When Cu2+ions are used as catalysts for the production of polyaniline, although the Cu2+ions is reduced to Cu+ions during the polymerization reaction of aniline monomers, the produced Cu+ions can be oxidized to return back to Cu2+ions under oxygen atmosphere[23]. Therefore, for OPCuN@C, a large amount of Cu2+could participate the polymerization during the whole process and end in the form of Cu2+. Furthermore, satellite peaks corresponding to electronic shake-up are also observed, which proves the existence of Cu2+ions in CuN@C and OPCuN@C[28]. It can also be seen from the spectra of N 1s (Fig.4(d)) that for each prepared sample four peaks are observed at about 398.4, 397.7, 401.1, and 402.8 eV, which are attributed to pyridinic, pyrrolic, graphitic nitrogen and nitrogen-oxide, respectively[29].

Table 1 shows the various carbon and nitrogen states for all the samples obtained from the deconvolution of XPS C1s and N1s peaks. As listed in Table 1, pyridinic and graphitic nitrogen, which are considered as active sites for oxygen reduction reaction[30,31], are easily produced by using Cu ions as catalysts under oxygen atmosphere.

3.2 Electrochemical analysis

To examine the ORR activity of all composites, Fig.5 shows the CV curves of N@C, CuN@C, and OPCuN@C. No obvious current peaks are detected when the electrolyte is saturated with nitrogen. When oxygen is added into electrolyte, well-defined current peaks are observed in all CV curves, which are attributed to the ORR on the samples[32]. As compared with N@C and CuN@C, OP-CuN@C exhibits good catalytic activity with a peak voltage of 0.72 V and a peak current of 1.09 mA. For explaining their intrinsic catalytic activity, EIS was carried out to assess the charge transfer resistance at electrode surface and describe the catalytic kinetics of the samples, as shown in Fig.6. The charge transfer resistance of OP-CuN@C is about 42 Ω, which is smaller than that of N@C (67 Ω) and CuN@C (48 Ω). The EIS curves of N@C, CuN@C and OP-CuN@C also have been fitted, and according to the fitting curves, the charge transfer resistence are 19.6 Ω, 6.3 Ω, and 1.2 Ω, respectively, which are consistent with the trend observed without fitting, indicating that the charge transfer of OP-CuN@C is an efficient procedure.

Fig.5 CV curves for N@C, CuN@C, and OP-CuN@C in the N2-saturated and O2-saturated 0.1 M KOH electrolyte

Fig.6 AC impedance spectra of N@C, CuN@C, and OP-CuN@C performed at 5 mV in the frequency range of 1 000 kHz to 50 MHz

To gain further insight about the ORR catalytic activity of the three samples, LSVs were recorded via RDE at 1600 rpm in O2-saturated 0.1 M KOH electrolyte. For comparison, LSV curve for commercial 20wt% Pt/C catalyst was also conducted under the same testing condition and is shown in Fig.7(a). The ORR onset potential of OP-CuN@C (0.91 V) is more positive than that of N@C (0.81 V) and CuN@C (0.88 V) in Fig.7(a). Moreover, the limited current density of OP-CuN@C (6.04 mA·cm-2) is explicitly higher than that of other samples (3.07 mA·cm-2for N@C and 4.38 mA·cm-2for CuN@C). On the other hand, the onset potential of OP-CuN@C is slightly smaller (about 50 mV) than that of Pt/C catalyst (0.96 V). To our best knowledge, the unexpected high limited current density of OP-CuN@C is comparable or even superior to the analogous catalysts recently reported[13-17]. In addition, according to the data from Fig.7(a) and Tafel equation[33], it can be calculated that the Tafel slope of OPCuN@C is 51.8 mV·dec-1at high potential area, which is superior to that of Pt/C catalyst (57.6 mV·dec-1), indicating that the rate-determined step for ORR on OP-CuN@C is faster than Pt/C catalyst[30]. The excellent ORR catalytic performance of OP-CuN@C may be attributed to the following reasons: (i) the large specific surface area contribute to the increased contact area between the OH-radicals and OP-CuN@C[4], (ii) the synergistic effect between doped nonmetal atoms (such as oxygen and nitrogen) can adjust the electronic structure of carbon matrix and optimize the adsorption strength of ORR intermediates on OP-CuN@C[34,35], (iii) the formation of Cu2+-N structure is beneficial for tuning the energy level of Cu2+of the d-electrons as well as facilitate the adsorption of ORR intermediates on OP-CuN@C[13], (IV) the doping of Cu2+ions is also conductive to improve the conductivity of carbon matrix and accelerating the charge transfer between OPCuN@C and electrolyte.

To shed light on catalytic mechanism of ORR, Fig.7(b) shows the LSV curves for OP-CuN@C at different rotation rates. For comparison, the LSV curves for N@C, CuN@C catalysts were also tested and shown in Figs.8(a-b). From Fig.7(b) and Figs.8(a-b), three different controlled areas (i e,dynamic-controlled area at high potential, dynamic-diffusion-controlled area at middle and diffusion-controlled area at low potential[36]) are observed in all the LSV curves. Based on the data from Fig.7(b) and Koutecky-Levich theory (Eq.(1)[37]), Fig.7(c) and Figs.8(c-d) display the Koutecky-Levich (K-L) plots of OP-CuN@C, C@N, and CuN@C catalysts, respectively.

Fig.7 (a) LSV curves for N@C, CuN@C, OP-CuN@C and commercial Pt/C catalysts in the O2-saturated 0.1 M KOH electrolyte at a scan rate of 5 mV·s-1 with a rotation rate of 1600 rpm; (b) LSV curves for OP-CuN@C in the O2-saturated 0.1 M KOH electrolyte solution at a scan rate of 5 mV·s-1 with different rotation rates; (c) Koutecky-Levich plots for OP-CuN@C at different potentials derived from the RDE measurements; (d) The electron transfer number of N@C, CuN@C, OP-CuN@C and commercial Pt/C catalyst at 0.6 V

Fig.8 LSV curves for (a) N@C and (b)CuN@C in the O2-saturated 0.1 M KOH electrolyte solution at a scan rate of 5 mV·s-1 with different rotation rates. Koutecky-Levich plots for (c)N@C and (d))CuN@C at different potentials derived from the RDE measurements

where,JandJkis the measured and kinetic current density, respectively;ωis the rotating rate (rpm);COis the O2saturation concentration in aqueous solution;DOis the O2diffusion coefficient in 1 M KOH electrolyte solution; andυis the dynamic viscosity of aqueous solution. According to Eq.(1), the electron-transfer numbers (n) at 0.6 V were calculated and presented in Fig.7(d) on basis of the slopes of the linear fitted K-L plots in Fig.7(c) and Figs.8(c-d). In Fig.7(d), thenvalue of OP-CuN@C (3.608) is higher than that of N@C (2.623) and CuN@C (3.189). Thenvalue indicates that the ORR catalyzed by OP-CuN@C is a close 4e-reduction process leading to the formation of H2O.

To further analyze the electrochemical properties of OP-CuN@C, its electrochemical stability was investigated by using chronoamperometric tests. The electochemical properties of OP-CuN@C and commercial Pt/C are shown in Fig.9. From Fig.9, we can see that after 40000 s, the catalytic activity retention of OP-CuN@C and Pt/C catalyst for ORR is 82.7% and 62.7%, respectively, indicating that the electrochemical stability of OP-CuN@C for ORR is much better than that of Pt/C catalyst. According to the previous literatures[38], it is speculated that the good electrochemical stability of OP-CuN@C is attributed to the strong covalent bonds between C, N and O atoms as well as the protection of Cu2+ions in carbon matrix. All the electrochemical behaviors show that OP-CuN@C is one of the most likely candidate materials for the development of high performance ORR catalysts.

Fig.9 Chronoamperometric responses of OP-CuN@C and commercial Pt/C catalyst at 0.6 V in the O2-saturated 0.1 M KOH electrolyte for 40 000 s

4 Conclusions

In summary, OP-CuN@C catalysts with specific surface area of 639.72 m2·g-1are prepared via a facile two-step approach. The electrochemical results indicate OP-CuN@C is a good ORR catalyst. First of all, the limited current density of OP-CuN@C is much higher than that of commercial Pt/C catalyst. Secondly, the electron transfer number of OP-CuN@C is 3.608 at 0.6 V, suggesting that the ORR process of OP-CuN@C is approximated a four-electron-transfer path. Thirdly, the OP-CuN@C shows better electrochemical stability than Pt/C catalyst. It is believed that OP-CuN@C is one of the most likely alternative ORR catalysts. And the easy prepared method for the composites also can be used wildly.