The oxygen reduction reaction (ORR) is very important because it is the cathodic reaction in fuel cells and metal-air batteries. Pt-based materials are extensively used as ORR catalysts under acidic conditions, but under alkaline conditions, a wide range of non-noble metals and their oxides are stable and active for practical applications [1, 2, 3]. An understanding of the ORR pathway on a catalyst surface is critical in both fundamental and practical terms. Although the ORR pathway on a smooth catalyst surface can be easily studied using the rotating ring disk electrode (RRDE) technique, for a carbon-based practical porous electrode, the ORR pathway is complex because the considerable amount of intermediate HO2− generated at the carbon surface can be readsorbed for further reactions. Specifically, hydroperoxide species have three possible subsequent pathways, as shown in Scheme 1, i.e., (1) diffusion directly into the electrolyte as a product, (2) further electrochemical reduction to form OH−, and (3) chemical decomposition to produce O2 and OH−. Unlike case (1), cases (2) and (3) are apparent four-electron processes, although only case (2) is a real four-electron process. Our previous study of carbon-supported cobalt oxides [4, 5] clearly showed that although the apparent electron transfer number was close to four in a large potential window, different electrochemical reactions occurred depending on the electrode potential, i.e., a two-electron transfer process followed by chemical decomposition of hydroperoxide at low overpotentials, but two-electron transfer followed by further electrochemical reduction of hydroperoxide, namely a serial four-electron transfer process, at high overpotentials.
Other promising transition-metal oxides, e.g., manganese oxides, are also worth investigating to help us to understand catalysis of the ORR by transition-metal oxides. Previously, it was concluded that a two-electron transfer process with subsequent chemical decomposition of hydroperoxide occurred on a manganese oxide surface [6, 7]. However, there is still a lack of experimental evidence to support this deduction. In this paper, continuing our previous study on the effects of manganese valences in manganese oxides on ORR activity [8], we aim to elucidate the ORR pathway on carbon-supported manganese oxides. The manganese oxide samples used in this study are the same as those used in our previous study [8]; therefore, the preparation procedures will not be described again. First, we calculated the electron transfer number of the ORR according to the Koutecky-Levich equation by measuring the ORR polarization curves via the rotating disk electrode (RDE) technique. We then distinguished between direct and indirect four-electron transfer processes by detecting the HO2- yields at manganese oxides of different catalyst thicknesses via the RRDE technique. Finally, we determined the main reactions in peroxide chemical decomposition (HODR) and electrochemical reduction (HORR) by measuring the turnover frequencies (TOFs) of the two reactions.
A traditional three-electrode system was used for RDE and RRDE measurements. The RDE (Φ 5 mm, glassy carbon (GC)] measurements were performed using a CHI 760D electrochemical workstation. The RRDE measurements were performed using a Bi-potentiostat (Pine Instruments). GC covered by a porous catalyst film was used as the working electrode, and Pt wire and a Hg/HgO electrode (MMO, in 1 mol/L NaOH, 0.93 V vs reversible hydrogen electrode after calibration) served as the counter and reference electrodes, respectively. The potential of the Pt ring electrode was kept at 0.2 V vs the MMO during the RRDE tests.
The preparation of the working electrode has been described in detail in the literature [9]. Briefly, catalyst powder (3 mg) was dispersed in ethanol (2 mL). Carbon powder (2 mg; Vulcan XC-72, Cabot Corp.) was added to increase the conductivity, and 5 wt% Nafion solution (50 μL; DuPont) was added as a binder. The mixture was ultrasonicated to form a well- dispersed ink. A certain amount of the ink was pipetted onto the GC electrode and then the solvent was evaporated at room temperature to form a catalyst thin film. All the MnOx catalyst samples were fresh, without any electrochemical pretreatment.
The ORR and HORR polarization curves were recorded in O2-saturated 1 mol/L NaOH solution and N2-saturated 1 mol/L NaOH containing 0.85 mmol/L H2O2 solution, respectively, at a scanning rate of 10 mV/s.
GC covered with a Pt/C catalyst was used as a probe to detect the changes with time in the H2O2 concentration in the electrolyte after adding MnOx as the HODR catalyst. The measurements were also carried out in a three-electrode system [10]. The preparation procedure for the thin-film electrode with Pt/C as the electrocatalyst was the same as that described above for the MnOx electrode. After electrochemically cleaning the electrode surface at a scanning rate of 100 mV/s in 1 mol/L NaOH, the ORR limiting current (ilim,ORR) was recorded by holding the electrode potential at -0.5 V for 10 s, with the electrode rotating at a speed of 1600 rpm. Then H2O2 (30%) was added to the electrolyte to ensure a concentration of 0.85 mmol/L (the same concentration as that of O2 in the O2-saturated 1 mol/L NaOH solution) [11]. Then MnOx (300 μg) dispersed in water (1 mL) was quickly added to the H2O2-containing electrolyte under magnetic stirring to ensure good contact of the MnOx catalyst with H2O2. The Pt electrode was held at -0.5 V for 10 s to collect the current (ilim) at intervals. The concentration of residual HO2- in the electrolyte could be calculated according to the current difference on the Pt electrode (ilim,HORR = ilim− ilim,ORR). The measurements continued for only 10 s at each interval and then the Pt electrode was rapidly removed from the electrolyte after the tests, therefore it was assumed that no extra HO2-decomposed at the Pt surface.
The ORR polarization curves of the carbon-supported manganese oxide electrocatalysts were measured in O2-saturated 1 mol/L NaOH electrolyte; the results are shown in Fig. 1. The ORR onset potentials of the four carbon-supported MnOx catalysts with different manganese valences, i.e., MnOOH/Mn(OH)4, MnO2, Mn2O3, and Mn3O4/Mn2O3 [8], are very similar, and the ORR limiting currents increase with increasing electrode rotating rate. For comparison, the ORR limiting current with Vulcan® XC-72 carbon as the electrocatalyst is lower than that with MnOx. According to the Koutecky-Levich equation [12]:
(where n is the ORR electron transfer number, F is the Faraday constant, A is the geometric area of the electrode, D is the diffusion coefficient of O2 in the electrolyte, ω is the rotation speed of the electrode in radians, ν is the viscosity of the electrolyte, and Co* is the concentration of O2 at the electrode surface), the limiting current is proportional to the square root of the electrode rotating rate, therefore the relationship between 1/ilim and ω-0.5 (@E = -0.5 V) is linear, as shown in Fig. 2. After fitting, the electron transfer numbers for MnOOH/Mn(OH)4, MnO2, Mn2O3, and Mn3O4/Mn2O3 were calculated to be 3.5, 3.5, 3.6, and 3.5, respectively, suggesting that a four-electron transfer is the dominant process for carbon-supported MnOx. For comparison, the electron transfer number for Vulcan® XC-72 carbon is only 2.1, suggesting that a two-electron process is dominant with HO2- as the main product.
To elucidate the apparent four-electron transfer process, we investigated the changes in the HO2- yield with varying catalyst thickness: if HO2- is the intermediate in the ORR, the HO2- yield should decrease with increasing catalyst thickness, because the HO2- species has a greater probability of chemically decomposing to form O2 [10, 13]; otherwise, the ORR should be a direct four-electron transfer process with negligible HO2- production [14, 15].
To avoid any changes in the surface manganese valence, fresh MnOx samples were used for the RRDE tests. The RRDE measurements were performed in an O2-saturated 0.1 mol/L NaOH electrolyte. The measured disk currents are shown in Fig. 3(a). The HO2- yield can be calculated from Eq. (2):
where iringand idisk are the ring current and disk current, respectively, and N is the collecting efficiency of the RRDE (N = 0.38 after calibration). The calculated HO2- yield (XHO2-) is shown in Fig. 3(b). To enable a clear comparison, the HO2− yields (@-0.4 V vs MMO) for different electrocatalysts are listed in Table 1. It can be seen that for all catalysts, as the catalyst loading increases from 3.75 to 30 μg, the quasi-limiting current increases, but the ORR onset potential remains almost constant. For the MnOOH/Mn(OH)4 sample, with a catalyst loading of 3.75 μg, the maximum XHO2- is 87.4%, indicating a dominant two-electron process. The HO2- yield sharply decreases with increasing catalyst loading. As the catalyst loading increases to 30 μg, the HO2- yield decreases to around 21%. For the MnO2 sample, XHO2- decreases from 26% to 6.6% with increasing catalyst loading from 3.75 to 30 μg, suggesting a four-electron process at high catalyst loading. These results clearly show that O2 is first reduced to HO2-, which is then reduced or chemically decomposes in a thick catalyst layer, therefore the HO2- yield detected by the Pt ring electrode decreases with increasing catalyst layer thickness. Bonakdarpour et al. [13] investigated the ORR at Fe/N/C electrodes and made similar observations, i.e., the HO2- yield decreased from 60%-80% to 2%-15% as the catalyst layer increased from 40 μg/cm2 to 800 μg/cm2.
To further clarify whether chemical decomposition or electrochemical reduction of the generated HO2-intermediate is dominant, we calculated and compared the TOFs of the HORR and HODR. The HODR TOF was calculated as described in the Experimental section. The difference between the ORR limiting current (ilim,ORR), measured in O2-saturated 1 mol/L NaOH, and the time-dependent limiting current of ilim, measured in O2-saturated 1 mol/L NaOH + 0.85 mmol/L H2O2, gives the limiting current for HO2- electrochemical reduction (ilim,HORR) as ilim − ilim,ORR. The changes in ilim,HORR with time can be used to calculate the TOF of the HODR at the MnOx surface, as described below [10]. It should be noted that the NaOH solution was saturated with O2 in measuring the H2O2 reduction current, to eliminate the influence of O2 on HO2- chemical decomposition. It should also be mentioned that the calculation of TOFHODR was based on the hypothesis that (1) TOFHODR is independent of the HO2- concentration and (2) all the MnOx takes part in the reaction. Theoretically, the HORR limiting current can be expressed as
where D2 is the diffusion coefficient of HO2- in the electrolyte, c2 is the bulk concentration of HO2-, and d is the diffusion layer thickness of the electrode at a rotating speed of 1600 rpm.
The HODR is approximately first order at low HO2- concentrations, therefore the reaction rate can be expressed as
where kHODRis the apparent reaction rate constant of the HODR, which is related to the amount of catalyst in the electrolyte. Combining Eqs. (3) and (4), the relationship between ilim,HORR and kHODRis expressed as
The changes in log(ilim,HORR) with time are shown in Fig. 4, which is fitted linearly, indicating that the HODR is a first-order reaction, as supposed. If the catalyst amount added to the electrolyte is the same for different samples, the slopes of the lines reflect the HODR rates for different MnOx catalysts, i.e., Mn3O4/Mn2O3 > Mn2O3 > MnO2 > MnOOH/Mn(OH)4.
Furthermore, kHODR is related to the TOF by Eq. (6) [10]:
where V is the volume of the electrolyte, m is the mass of MnOx, c2 is the bulk concentration of HO2-, W is the mass percentage of manganese in MnOx, and MMn is the atomic weight of manganese (54.95 g/mol). The calculated results are listed in Table 2.
The HORR polarization curves for the MnOx catalysts were measured in a N2-saturated 0.85 mmol/L H2O2 + 1 mol/L NaOH electrolyte and are shown in Fig. 5(a). To eliminate the influence of mass transport, the kinetic HORR currents were derived from Fig. 5(a) according to [12]:
The obtained Tafel plots are shown in Fig. 5(b). It can be seen that reduction currents are detected for the four MnOx catalysts and the onset potentials are around −0.05 V. Mn2O3 and MnO2 have the most positive and the most negative half-wave potential, respectively. The limiting currents are in the range -0.06 to -0.10 mA. For comparison, the theoretical limiting current of the HORR was calculated according to the Koutecky-Levich equation; at around 1.6 mA/cm2, it is ca. 0.31 mA for an electrode of f 5 mm, with a diffusion coefficient of HO2- in 1 mol/L NaOH of 1.3 × 10-9 m2/s, electrolyte viscosity of 1 × 10-4 m2/ s, and electrode rotating speed of 1600 rpm. The much lower experimental limiting current compared with the calculated value shows that only some HO2- is electrochemically reduced, and some is chemically decomposed to form O2, which is either electrochemically reduced or released to the atmosphere. In the following calculation, it is assumed that the limiting current results entirely from reduction of HO2-, to obtain the maximum TOF of the HORR. The TOF of the HORR is calculated as [10]
where ik,HORRis the kinetic current of the HO2- electrochemical reduction, n is the electron transfer number, and F, W,and MMn have the same meanings as in Eqs. (3) and (6). The values at -0.12 V are listed in Table 2 for comparison.
From Table 2, it is clearly seen that for the four MnOx catalysts, the TOF for the HODR is two orders of magnitude higher than that for the HORR, indicating that the intermediate HO2- mainly undergoes subsequent chemical decomposition. These MnOx samples have similar morphologies; one reason for the better ORR activities of Mn2O3 and Mn3O4/Mn2O3 is probably faster chemical decomposition of the HO2- intermediate on these catalysts.
A comparison of the TOFs of the HODR and HORR showed that although apparent four-electron transfer processes were observed by measuring the HO2− yield via the RRDE technique, the real ORR pathway involves a two-electron transfer process to generate HO2−, with subsequent chemical decomposition. These results are expected to help us to understand the intrinsic catalytic behavior of carbon-supported transition-metal oxides for the ORR in alkaline electrolytes.
2015, Vol. 36 Issue (2): 175-180
PDF (6525 KB)
2015, Vol. 36 Issue (2): 175-180
PDF (6525 KB)