Fe–Mn bimetallic oxides-catalyzed oxygen reduction reaction in alkaline direct methanol fuel cells

Two Fe–Mn bimetallic oxides were synthesized through a facile solvothermal method without using any templates. Fe2O3/Mn2O3 is made up of Fe2O3 and Mn2O3 as confirmed via XRD. TEM and HRTEM observations show Fe2O3 nanoparticles uniformly dispersed on the Mn2O3 substrate and a distinct heterojunction boundary between Fe2O3 nanoparticles and Mn2O3 substrate. MnFe2O4 as a pure phase sample was also prepared and investigated in this study. The current densities in CV tests were normalized to their corresponding surface area to exclude the effect of their specific surface area. Direct methanol fuel cells (DMFCs) were equipped with bimetallic oxides as cathode catalyst, PtRu/C as the anode catalyst and PFM as the electrolyte film. CV and DMFC tests show that Fe2O3/Mn2O3(3 : 1) exhibits higher oxygen reduction reaction (ORR) activity than Fe2O3/Mn2O3(1 : 1), Fe2O3/Mn2O3(1 : 3), Fe2O3/Mn2O3(5 : 1) and MnFe2O4. The much superior catalytic performance is due to its larger surface area, the existence of numerous heterojunction interfaces and the synergistic effect between Fe2O3 and Mn2O3, which can provide numerous catalytic active sites, accelerate mass transfer, and increase ORR efficiency.


Introduction
The rapid depletion of fossil fuel and the increase in environmental pollution have driven us to search for sustainable and clean energy resources. Fuel cells have been considered promising power sources owing to their advantage of transforming chemical energy directly into electrical energy. 1 At present, direct methanol fuel cells (DMFCs) are obtaining great attention in virtue of their high energy density, environment friendliness and comparatively lower operating temperature. [2][3][4][5] Furthermore, methanol is convenient and safe for transport and storage, swi to refuel and available at a low price. 6 At present, DMFCs have great potential application as a portable power supply or electric vehicle power supply. Nevertheless, the inertial oxygen reduction reaction (ORR) and methanol oxidation reaction (MOR) dynamics and the high cost of noble-based catalysts and proton exchange membrane (PEM) hinder the commercial application of DMFCs. 4,7,8 Recently, polymer ber membranes (PFMs) have been demonstrated to be an excellent alternative to PEMs for higher performance liquid fuel cells at a reduced cost in our previous study. 9,10 The bers in PFMs are neutral and possess pores and gaps, which allow molecules, ions, and liquid fuel to transport or move through the PFM freely. Consequently, the cathode catalysts should have both outstanding tolerance for methanol poisoning and excellent stability. The widely used cathode catalysts are Pt or Pt-based metal alloy catalysts, such as Pt-Co, 11 Pt-Pd, 4 Pt-Ni, 12 and Pt-Fe. 13 However, these catalysts have both ORR and MOR catalytic activity, leading to a mixed potential at the cathode and poisoning by methanol. In terms of lower cost, a variety of non-Pt catalysts, such as Ru-Se, 14 Pd-Ni, 15 Pd-Fe, 16 Co-Se, 17 Fe-N-C, 18 Cu-Fe-S, 19 and Co-O, 20 which display ORR catalytic activity and better methanol tolerance than Pt-based catalysts, also have been researched.
Among them, transition metal (Fe, Co, Ni, Mn, etc.) oxides have gained increasing interest as ORR catalysts in virtue of their high activity, low cost and environmental friendliness. 21

Materials characterization
The structures and compositions of the as-prepared Fe 2 O 3 / Mn 2 O 3 and MnFe 2 O 4 were characterized via X-ray diffraction (XRD, D/Max 2200PC, Japan) and high-resolution TEM (HRTEM). The morphological properties were characterized by eld emission scanning electron microscopy (FESEM, Hitachi S-4800, Japan) and transmission electron microscopy (TEM, FEI company Tecnai G2 F20) equipped with energy-dispersive spectrometer (EDS). The Brunauer-Emmett-Teller (BET) method was carried out to determine the pore volumes, pore size and the specic surface area distribution of the samples using a surface area and porosimetry system (ASAP 2460, Micromeritics Instrument Corporation, USA). X-ray photoelectron spectroscopy (XPS) measurements (VG Thermo ESCALAB 250 spectrometer) were used to quantitatively analyze the chemical compositions of samples.

Electrochemical measurements
Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were measured using an electrochemical workstation (CHI 660E, Chenhua Instruments, Shanghai, China). A standard three-electrode system consisted of the catalyst-modied glassy carbon electrode as the working electrode, Hg/HgO electrode as the reference electrode and the Pt network as the counter electrode. The glassy carbon electrode was modied as follows: 4 mg catalyst, 1 mg CNTs, 0.2 mL distilled water, 0.5 mL absolute ethyl alcohol and 50 mL Naon solution (5 wt%) were ultrasonically dispersed into a homogeneous suspension for about 1 h; then, the suspension was poured on the glassy carbon electrode surface and dried at room temperature.

Electrode preparation and DMFC measurements
The cathode electrode was a sandwich structure, including catalyst layer, current accumulating matrix and gas diffusion layer. The gas diffusion layer was obtained by mixing 60 wt% acetylene black and 40 wt% polytetrauoroethylene (PTFE, 30 wt% solution) with ethanol under ultrasonication and pressing the slurry into a thin layer of 0.3-0.5 mm and then treating at 350 C for 1 h. The catalyst layer was obtained rst through mixing 24 mg catalyst, 6 mg CNTs and 6.7 mg 30 wt% PTFE solution into slurry with addition of a certain amount of absolute ethanol; the slurry was pasted on nickel foam (porosity > 95%) and then dried at 80 C for 2 h. Finally, the cathode was obtained by pressing the catalyst layer on nickel foam and the gas diffusion layer under 2 MPa.
The anode was obtained via mixing PtRu/C (60 wt%) and Naon solution (5 wt%) at a mass ratio of 1 : 1. The anode preparation process is consistent with that of the cathode without the gas diffusion layer. The loading of PtRu/C was 5 mg cm À2 .
The cathode, PFM and anode were assembled into a fuel cell. At the cathode, the oxygen ow rate was 20 cubic centimeters per minute; the anode aqueous solution was 4 M KOH and 5 M methanol. The structure of PFM-DMFCs was introduced and described in our previous study. 9 A battery testing system (Neware Technology Co., Ltd., Shenzhen, China) was used to measure the performance. nanosheets-self-assembled globular structure ( Fig. 2(b)). The microspheres of Fe 2 O 3 /Mn 2 O 3 (3 : 1) are 3-4 mm in diameter and the pore size is about 30 nm. From the XRD analysis results shown in Fig. 1(a), it can be inferred that the formation of nanopores is due to the release of CO 2 , which comes from MnCO 3 decomposition during the calcination process. In particular, mesoporous structure is protable for the rapid transmission of O 2 , fuel and electrolyte, which can accelerate the redox reaction rate and improve electrochemical performance. 28 Further, the EDS elemental mappings of Fe 2 O 3 / Mn 2 O 3 (3 : 1) (Fig. 2(c)-(f)) were recorded to obtain elemental distribution of Fe, Mn and O in the structure and it could be observed that the three elements are distributed homogeneously. As shown in Fig. 3 Fig. 4(b), a distinct heterojunction boundary between Fe 2 O 3 nanoparticles and Mn 2 O 3 substrate could be detected as shown by the red line. Fig. 4(c) shows that MnFe 2 O 4 exists as nanospheres with diameters of 300-500 nm. The lattice fringe with d-spacing is 0.25 nm, which can be well indexed to the (311) plane of MnFe 2 O 4 phase ( Fig. 4(d)).

Structural and morphological characterization
XPS was used to measure the surface chemical composition and conrm the Fe/Mn ratio of the as-prepared Fe 2 O 3 /Mn 2 O 3 samples. As shown in Fig. 5(a), the common peaks of Fe 2p, Mn 2p and O 1s are present. The element contents are calculated and summarized in Table 1, illustrating that the results of Fe/ Mn ratios are approximately equal to the corresponding experimental values. The N 2 adsorption-desorption technique at 77 K was used to investigate specic surface areas and pore structures of the as-prepared samples. The nitrogen adsorptiondesorption curves (Fig. 5(b)) manifest a type IV isothermal line with a delay loop-line in the P/P 0 range of 0.

ORR activity and DMFC performance
CV tests were performed to describe ORR catalytic activity. The current densities were normalized to their corresponding surface area. Capacitance correction was acquired by subtracting the measured current densities under N 2 from those measured under O 2 under the same condition. Fig. 6(a) shows the CV curves of    slightly more negative than that of MnFe 2 O 4 (À0.237 V), the oxygen-reduction peak current density is much greater than that of MnFe 2 O 4 (À26. 26 Fig. 6(b) Fig. 6(d), at a constant current of 10 mA cm À2 at room temperature, the Fe 2 O 3 /Mn 2 O 3 -based DMFC has much higher cell voltage than MnFe 2 O 4 -based DMFC for 75 000 s. In about 50 000 seconds, the voltage of MnFe 2 O 4based DMFC decreases sharply. For the Fe 2 O 3 /Mn 2 O 3 -based DMFC, no distinct attenuation phenomenon is found, indicating that this cell is quite stable.

ORR mechanism of Fe 2 O 3 /Mn 2 O 3
In conclusion, Fe 2 O 3 /Mn 2 O 3 (3 : 1) exhibits higher ORR activity and superior DMFC performance than MnFe 2 O 4 . The rst reason is that Fe 2 O 3 /Mn 2 O 3 (3 : 1) has a much larger specic surface area (21.73 m 2 g À1 ) than MnFe 2 O 4 (3.05 m 2 g À1 ), which plays a key role in enhancing ORR activity, providing numerous active sites and accelerating mass-transfer. It is worth noting that although current densities are normalized to their corresponding surface area in the CV tests, Fe 2 O 3 /Mn 2 O 3 still demonstrates higher ORR activity than MnFe 2 O 4 .
The second reason is due to the existence of the numerous heterojunctions between Fe 2 O 3 and Mn 2 O 3 , which provides an intensive internal electric eld at the interface of the two oxides and increases the catalytic active sites, electron transfer and ORR efficiency. 29,30 EIS was applied to describe the internal resistance of Fe 2 O 3 /Mn 2 O 3 (3 : 1) and MnFe 2 O 4 . As shown in Fig. 7(a)     Paper microspheres are 3-4 mm in diameter and the pore size is about 30 nm. The formation of nanopores is due to the release of CO 2 , which comes from MnCO 3 decomposition during the calcination process. The TEM and HRTEM images show Fe 2 O 3 nanoparticles uniformly dispersed on the Mn 2 O 3 substrate and a distinct heterojunction boundary between Fe 2 O 3 nanoparticles and Mn 2 O 3 substrate. MnFe 2 O 4 has a hierarchical structure, in which the nanoparticles are 20-30 nm in diameter and the self-assembled globular shapes have diameters of 300-500 nm.

Conflicts of interest
There are no conicts to declare.