Development of Nanosized Mn₃O₄-Co₃O₄ on Multiwalled Carbon Nanotubes for Cathode Catalyst in Urea Fuel Cell

Abstract

Double-oxide Mn₃O₄-Co₃O₄ nanoparticles were synthesized and anchored on multiwalled carbon nanotubes (MWCNTs) via a single-step solvothermal method. The largest specific area (99.82 m²g⁻¹) of the catalyst was confirmed via a nitrogen adsorption isotherm. Furthermore, the uniform coating of the Mn₃O₄-Co₃O₄ nanoparticles on the surface of the MWCNTs was observed via scanning electron microscopy and transmission electron microscopy; the uniform coating provided an effective transport pathway during the electrocatalytic activities. The rotating disk electrode and rotating ring disk electrode measurements indicated that the electron transfer number was 3.96 and the evolution of H₂O₂ was 2%. In addition, the Mn₃O₄-Co₃O₄/MWCNT catalyst did not undergo urea poisoning and remained stable in an alkaline solution. Conversely, commercial Pt/C could not withstand urea poisoning for long. The performance cell achieved a power density of 0.4226 mW cm⁻² at 50 °C. Therefore, Mn₃O₄-Co₃O₄/MWCNT is an efficient and inexpensive noble-metal-free cathodic catalyst for direct urea fuel cells.

1. Introduction

In recent years, the demand for renewable energy has been gradually increasing. As a type of renewable energy carrier, urea is considered as a hydrogen carrier form because it is non-flammable, non-toxic, biodegradable, and abundant. Urea can be utilized as a power source with direct urea fuel cells (DUFCs). DUFCs represent novel technology for the efficient generation of electricity because of their numerous advantages, such as low cost, high energy-storage densities, and high levels of safety [1]. In addition, DUFCs generate electricity without releasing toxic products, thus limiting the level of environmental pollution. Hence, it constitutes an attractive power generation technology [2]. Furthermore, urea is widely available in industrial wastewater and human/animal urine. However, untreated urea could pose danger to both human health and the environment because it can hydrolyze into toxic ammonia and nitrite [3,4]. Therefore, DUFCs have two-fold benefits—they can generate electricity and produce clean water. In a DUFC, a urea electrooxidation reaction and oxygen reduction reaction (ORR) occur at the interface of the electrolyte and the electrocatalytic layers, and are expressed as follows [5]:

Anode: CO(NH2)2 + 6OH− → N2 + 5H2O+ CO2 + 6e−   E0 = −0.746 V

Cathode: 2 H2O + O2 + 4e− → 4OH−                               E0 = 0.4 V

Overall: 2CO(NH2)2 + 3O2 → 2N2 + 4H2O + 2CO2       E0 = 1.146 V

Lan et al. reported the first DUFC, which achieved a maximum power density of 1.7 mW cm⁻² using a 1 M urea solution as the fuel [6]. Although the performance of DUFCs has been significantly improved, it is still lower than that of other fuel cells in terms of the current density. A major issue is the occurrence of an ineffective electrocatalyst in the cathode side due to the strong O=O bonding and the decomposition of H₂O₂ [7]. H₂O₂ is an undesirable by-product of the ORR, where two electrons are indirectly transferred from O₂, indicating lower oxygen utilization [8]. Furthermore, H₂O₂ has strong oxidation activity, leading to the corrosion of the catalyst and chemical degradation [7].

Pt and Pt-based materials are state-of-the-art catalysts for the ORR. However, they have seriously hindered applications because of their high cost and limited availability [9]. In addition, Pt possesses poor stability, which is presented by the loss of platinum on the surface area due to (1) corrosion of the carbon support reducing the electrical contact, (2) dissolution and redeposition, and (3) agglomerations; these processes reduce the performance of Pt during fuel cell operations [10]. Among numerous efforts to replace Pt-based materials, fabricating an oxide–oxide heterostructure composite is an effective approach. The combination of two metal oxides as cocatalysts enables ORR kinetics and covalent electron transfer, owing to the interphase ligand effect between them [11]. In addition, the oxide–oxide composite favors the chemical disproportionation of the H₂O₂ intermediates to generate OH⁻ and O₂ species during the ORR [12,13,14].

Cobalt and manganese oxides have good ORR catalytic ability; hence, they are widely studied for ORR catalysis [15,16,17]. Possessing a spinel structure, the Co²⁺/Co³⁺ redox couple found in tetrahedral and octahedral voids facilitates the formation of defects, resulting in electron transfer [18]. Like cobalt oxide, manganese oxides (Mn_xO_y) are suitable metal oxides for the catalysis of ORRs owing to their various oxidation states, chemical stability, and low cost. Various other metal oxide-based ORR electrocatalysts, such as the oxides of Cu, Fe, Zr, and Sn, have also been investigated, but they showed lower activity than Co or Mn-based catalysts. However, cobalt and manganese oxides have intrinsically low conductivity and particle aggregation, which contribute to hinder their chemical and thermal stability [9]. Therefore, dispersing cobalt and manganese oxides homogeneously on multiwalled carbon nanotubes (MWCNTs) as the conductive substrate is an effective way to overcome the aforementioned problems [19].

Here, we report a nanosized composite that consists of Mn₃O₄ and Co₃O₄ nanoparticles, which were uniformly anchored to MWCNTs via a one-step simple solvothermal method without adding any intermediate substances. The impact of the homogeneous oxide–oxide structure of the composite on catalytic activity was thoroughly investigated. Upon integrating into DUFCs, the Mn₃O₄-Co₃O₄/MWCNT composite exhibited highly electrocatalytic properties compared to the corresponding individual components. The achieved power density was relatively low compared to other currently developed devices, but the catalyst’s components are inexpensive and abundant. Moreover, there is a promising approach of a simple and efficient synthesis of the catalyst in the present work. Therefore, the DUFC could be applied for simultaneous wastewater treatment as an efficient process of the proper treatment for urea-containing wastewater from domestic and industrial activities at low cost.

2. Materials and Methods

2.1. Preparation of Mn₃O₄-Co₃O₄/MWCNT Composite

For the fabrication of Mn₃O₄-Co₃O₄/MWCNT, 2 mmol Co(CH₃COO)₂·4H₂O (Sigma Aldrich) was dissolved in 100 mL of ethanol, and the mixture was stirred at room temperature for 2 h. Subsequently, the mixture was transferred into a Teflon-lined autoclave for the solvothermal reaction at 150 °C for 5 h. The obtained mixture was cooled to room temperature. Next, 80 mg of ox-MWCNTs (acid-treated MWCNT) and 2 mmol Mn(CH₃COOH)₂·4H₂O (Sigma Aldrich) were added, and the mixture was stirred at 85 °C for 3 h. The dark brown precipitates were collected via centrifugation and washed with ethanol and ultrapure water several times. Furthermore, Co₃O₄/MWCNT and Mn₃O₄/MWCNT were synthesized using the same process.

2.2. Material Characterization

The morphology and particle sizes of the composite were observed via field emission scanning electron microscopy (SEM) (FE-SEM, Hitachi S-4200 system, Tokyo, Japan) and transmission electron microscopy (TEM) (HR-TEM, CRYO-ARM200F, JEOL, Massachusetts, USA). A Brunauer–Emmett–Teller (BET) measurement was performed to investigate the specific surface areas of the obtained samples (Autosorb iQ Station 2). The X-ray diffraction (XRD) spectra were recorded using the Panoalytical Empyrean XRD apparatus with Cu 1.8 KW (Cu k-α 1.54 Å, max 60 kV, and 55 mA) and PICXEL 3D with a prefix interface Xenon proportional detector (0D point detector).

2.3. Electrochemical Measurements

Rotating ring disk electrode (RRDE) measurements were performed with an RRDE-3A instrument in the SP-240 Bio-Logic SAS mode to evaluate the activities of Co₃O₄/MWCNT, Mn₃O₄/MWCNT, and Mn₃O₄-Co₃O₄/MWCNT in ORR electrocatalysis. For this measurement, a three-electrode cell was used, which consisted of an RRDE, Ag-AgCl/KCl (0.973 V vs. RHE (reversible hydrogen electrode)), and a Pt wire employed as the working, reference, and counter electrodes, respectively. Typically, 10 mg of a prepared catalyst powder was ultrasonically suspended in 1 mL isopropanol and 100 µL of Nafion for 30 min. Subsequently, 10 µL of the resulting slurry was coated on the surface of the RRDE. All measurements, including cyclic voltammetry (CV), chronoamperometry, and linear sweep voltammetry (LSV) were carried out in 1 M KOH at room temperature. The measuring system was purged with O₂ for at least 30 min before measurements. We also analyzed the scan rate at 20 mV s⁻¹ between 0.4 and 1.2 V vs. RHE for CV and 0.5 to 1.2 V vs. RHE at rotating speeds ranging from 400 to 2000 rpm for LSV. The commercial Pt/C (Pt 20%, from Premetek) was subjected to the same method as a reference standard.

2.4. Fabrication and Examination of Membrane Electrode Assembly

A single fuel cell was fabricated to evaluate the practical application of the Mn₃O₄-Co₃O₄/MWCNT composite using the as-prepared nanostructure as a cathode catalyst and the commercial Ni/C (Ni 20%, from Premetek) as an anode catalyst with Fumasep FAA-3-50 (Fumatech) anion exchange membranes. For this, 50 mg of the as-prepared catalyst (in powder form) and the commercial Ni/C were dispersed in a mixture of 5% Nafion (200 µL) and isopropanol (800 µL), and sonicated for 30 min to form a uniform ink. The catalyst inks (2 mg cm⁻²) were then loaded onto a 5.0 cm² carbon paper. Both electrodes were attached to the membrane electrolyte by hot-pressing for 2 min at 100 °C. The membrane electrode assemblies were examined in a fuel cell station (Scitech Korea Inc., Seoul, Korea). The experiments were performed in a 0.33 M urea/1 M KOH solution fed as an anolyte and wet O₂ as the catholyte at room temperature.

3. Results and Discussion

The two-step synthesis route of the Mn₃O₄-Co₃O₄/MWCNT catalyst is shown in Figure 1. The MWCNT was peroxided with HNO₃ to enrich the epoxy and hydroxyl functional surface groups. Specifically, Co₃O₄/MWCNT was prepared through the hydrolysis reaction of the cobalt-salt precursor and ox-MWCNT in an ethanol solution without any reducing agents under solvothermal treatment at 150 °C. In the ox-MWCNT template, the Co species were coated homogeneously owing to the attractive electrostatic forces and the good solubility of the reduced MWCNT [20]. In the next step, Mn(CH₃COOH)₂ was added to the intermediate product solution and stirred at 80 °C for 3 h to form Mn₃O₄ nanoparticles over Co₃O₄/MWCNT. Thus, the Mn₃O₄ nuclei were homogenously anchored to the surface of the MWCNTs.

The successful synthesis of manganese–cobalt oxide nanoparticles on the MWCNT composite was confirmed via TEM and SEM, as shown in Figure 2. SEM and TEM images of the obtained products showed that the Mn₃O₄ and Co₃O₄ nanoparticles were homogeneously coated in a high quantity on the surface of the MWCNT template without any aggregation. According to the TEM images shown in Figure 2, the average size of both oxides was small, ranging from 8 to 10 nm, which could be due to the low hydrolysis rate of the precursor [21].

The structures and crystallographic phase of the synthesized catalyst were studied through XRD, as shown in Figure 3. For the Mn₃O₄/MWCNT and Co₃O₄/MWCNT catalysts, the diffraction peaks were indexed according to JCPDS 24-0734 and JCPDS 43-1003, respectively. The XRD patterns of the prepared Mn₃O₄-Co₃O₄/MWCNT nanoparticles exhibited well-defined peaks at 18.96° (111), 31.09° (220), 36.86° (311), 44.78° (400), 59.11° (511), and 65.28° (440), which were indexed to the Co₃O₄ spinel structure (JCPDS 43-1003). No residual peaks were observed. The individual Co₃O₄ phase had a crystalline structure. In addition, the metal–oxygen–carbon (M–O–C) composite presented a peak at 55° (303), which corresponded to the formed Mn₃O₄, in reference to JCPDS 27-0734. This observation suggests the formation of Mn₃O₄ in addition to the Co₃O₄ phase in the composite, indicating the presence of a multiphase structure.

To study the chemical bonding of the synthesized catalysts, we performed X-ray photoelectron spectroscopy (XPS). From the survey spectrum, we confirmed the existence of cobalt, manganese, oxygen, and carbon. The XPS profiles of Co2p^3/2 and Co2p^1/2 are presented in Figure 4b. The fitting peaks were observed at 780.2 and 795.3 eV for the binding energy at a spin energy gap of 15.1 eV, which further verified the presence of Co₃O₄ in the composite [22]. The presence of Mn₃O₄ was verified by two peaks at 641.8 and 653.4 eV with a separating energy of 11.6 eV, indicating the Mn2p^3/2 and Mn2p^1/2 spin-orbit states, respectively [23,24]. The C1s deconvolution spectrum contained three distinguished peaks at 284.5, 286.04, and 282.1 eV, which corresponded to carbon atoms bonded to carbon (C‒C), oxygen (C‒O), and oxygen in carboxyl groups (O‒C=O), respectively [25]. The O1s spectrum was divided into four distinct peaks, which indicated the formation of oxygen bonds at the interface between the oxides and carbon.

In addition, the O1s XPS profiles of the individual components Mn₃O₄/MWCNT and Co₃O₄/MWCNT were recorded for reference and are shown in Figure 4e,f. The two peaks at 531.84 and 533.04 eV corresponded to O‒C=O and C‒O bonding, respectively, which corroborated previously reported results [26,27]. The third peak observed at 529.7 eV was consistent with an M‒O bond (M = Co, Mn) [28,29], which is slightly positively shifted. Furthermore, the intensity of this peak decreased compared to that of the individual oxides alone on MWCNTs, demonstrating the coexistence of Mn‒O and Co‒O in the composite catalysts. The formation of the M‒O‒C bonding at the interface of the oxide and MWCNTs was also confirmed by comparing the binding energy and intensity in the O1s of the synthesized catalysts. The characteristic peak at 530.14 eV in the spectrum of the Mn₃O₄-Co₃O₄/MWCNT composite was similar to that of Mn₃O₄/MWCNT and Co₃O₄/MWCNT. Nonetheless, this peak showed an increased intensity and a slight change in the binding energy, indicating the formation of M‒O‒C bonding, including Mn‒O‒C and Co‒O‒C [30]. In addition, the bonding energy of the well-defined peaks shifted positively from 1 to 3 eV in contrast with that in the M‒O bond, which further proved the formation of M‒O‒C in Mn₃O₄-Co₃O₄/MWCNT [20].

The XPS results demonstrated that the Co₃O₄ and Mn₃O₄ nanoparticles were decorated over the surface of the MWCNTs, and this was highly related to the enhancement of the electrochemical performance. The high concentration of oxygen vacancies could facilitate oxygen adsorption and covalent metal oxide–nanocarbon bonding and enable improved electron transfer across the interface. The specific areas of Mn₃O₄-Co₃O₄/MWCNT were 99.82 m²/g as obtained from the BET analysis shown as Figure 5. The large specific surface areas of the synthesized composite could facilitate the adsorption and transportation of O₂ and H₂O during the ORR.

To assess the catalytic activity of Mn₃O₄-Co₃O₄/MWCNT, we evaluated the individual components using commercial Pt/C 20% as a comparative sample via CV on an RRDE in O₂-saturated 1.0 M KOH. The observed ORR peaks for the obtained sample are depicted in Figure 6. The CV curves of the Mn₃O₄-Co₃O₄/MWCNT composite suggest much higher activity than those of Co₃O₄ and Mn₃O₄. Additionally, the curves were compared with those of Pt/C. The reduction peaks of all synthesized samples had the same position peak (0.7 V vs. RHE), but the diffusion limiting current density followed this sequence: Mn₃O₄-Co₃O₄/MWCNT > Co₃O₄/MWCNT > Mn₃O₄/MWCNT. The reduction peak reached approximately 3.55 mA cm⁻² for the Mn₃O₄-Co₃O₄/MWCNT composite and was approximately 1.3 times higher than that of Co₃O₄ (2.54 mA cm⁻²) and Mn₃O₄ (2.49 mA cm⁻²), indicating that the combination of metal oxides can significantly improve the ORR activity. In addition, the current density of the Mn₃O₄-Co₃O₄/MWCNT composite was close to that of Pt/C; however, the more positive the peak potential was (0.7 V), the more positively shifted the onset potential (0.85 V) was, as shown in Figure 6b. Thus, the Mn₃O₄-Co₃O₄/MWCNT composite was an active electrocatalyst for the ORR.

For comparison, the polarization curves for the ORR were also recorded in a 1 M KOH solution at 1600 rpm, as shown in Figure 6c. The electrocatalytic activity of the composite catalyst was higher than the activities of its individual components and approximately equal to that of the commercial Pt/C owing to their higher current density and positive half-wave potential. The results illustrated that the onset potential of the composite (0.87 V) was higher than that of Mn₃O₄ and Co₃O₄ (0.84 V) and lower than that of the Pt/C catalyst (0.97 V). The current density on the composite was 4.97 mA cm⁻² at 0.6 V, which was higher than that of Co₃O₄ (4.70 mA cm⁻²) and Mn₃O₄ (4.53 mA cm^-2) and similar to that of Pt/C. To gain insight into the kinetics of the ORR, the LSV of Mn₃O₄-Co₃O₄/MWCNT was recorded in O₂-saturated 1.0 M KOH at different rotating rates. As shown in Figure 7a, the polarization curves suggested that the measured current intensity increased with the high-speed rotation rates because of the enhanced diffusion. Based on the diffusion in the kinetically limited regions, we used the Koutecky–Levich (K–L) plot to determine the electron transfer number. The K–L equation is [31]:

$$\frac{1}{\mathrm{J}}=\frac{1}{{\mathrm{J}}_{\mathrm{L}}}+\frac{1}{{\mathrm{J}}_{\mathrm{k}}}=\frac{1}{{\mathrm{B}\mathsf{\omega }}^{1/2}}+\frac{1}{{\mathrm{J}}_{\mathrm{k}}}$$

(1)

B = 0.62nFC_oD_o^2/3 ν^−1/6

(2)

where J is the measured current density; J_k is the kinetic current density; B is the Levich constant or proportionality coefficient, which could be determined from the slope of the K–L plot; ω is the electrode rotating rate, F = 96485 C mol⁻¹; ν is the kinematic viscosity of the electrolyte (ν = 0.01 cm² s⁻¹); C_o = 7.8 × 10⁻⁷ mol cm⁻¹ is the concentration of O₂ in the electrolyte; and D_o = 1.8 × 10⁻⁵ m² s⁻¹ is the diffusion coefficient of O₂ in 1.0 M KOH. The electron transfer number was calculated using Equations (1) and (2) reaching up to 3.71 electrons, indicating a four-electron oxygen reduction pathway.

To gain further insight into the activities of the catalysts in the ORR, we investigated the reaction kinetics at a rotation rate of 1600 rpm. During the experiment, the disk potential was scanned at 0.55–0.9 V vs. RHE, while the ring potential was maintained at 0.627 V. As shown in Figure 7c, both ring and disk currents increased rapidly because the oxygen supply was sufficient on the surface of the working electrode to enhance the acceleration rate of the ORR. Co₃O₄/MWCNT and Mn₃O₄/MWCNT had very similar disk currents, whereas the composite catalyst exhibited a higher ring current. The onset potential and current density peaked at 0.855 V and 0.465 mA, which were higher than those of the other catalysts. The half-wave potential and diffusion-limited current increased in the following order: Co₃O₄/MWCNT ≤ Mn₃O₄/MWCNT < Mn₃O₄-Co₃O₄/MWCNT, illustrating the increase in the ORR activity.

In addition, the electron transfer number and peroxide percentages were obtained as a function of both disk and ring currents using the following equation from the RRDE technique [31].

$$\mathrm{n}=\frac{4{\mathrm{I}}_{\mathrm{D}}}{{\mathrm{I}}_{\mathrm{D}}+{\mathrm{I}}_{\mathrm{R}}·\mathrm{N}}$$

(3)

$${\%\mathrm{H}}_2{\mathrm{O}}_2=\frac{100{\mathrm{nI}}_{\mathrm{R}}}{2{\mathrm{I}}_{\mathrm{D}}\mathrm{N}}$$

(4)

where I_D and I_R are the currents at the disk and ring electrodes, respectively, and N = 0.317 (I_D/I_R) is the RRDE collection efficiency. For the as-prepared catalysts Mn₃O₄-Co₃O₄/MWCNT, Co₃O₄/MWCNT, and Mn₃O₄/MWCNT, the χ H₂O₂ were 2%, 4.46%, and 2.5%, and the n values were 3.96, 3.91, and 3.94, respectively. These values were similar to those obtained from the K–L plot, but a small difference remained due to the calculation process. However, the results presented a preference for a direct four-electron reduction pathway. Mn₃O₄-Co₃O₄/MWCNT showed the highest n value (3.96) and the lowest χ H₂O₂ (2%), which proved its superior electrochemical activity compared to Co₃O₄/MWCNT and Mn₃O₄/MWCNT.

The improved performance (higher electron transfer and lower ring current) of the Mn₃O₄-Co₃O₄/MWCNT catalyst can be attributed to the following reasons (Figure 8). First, the uniformity of the Co₃O₄ and Mn₃O₄ metal oxides decorated on the surface of the MWCNTs enabled C‒O‒M covalent bonding at the interface [32,33,34,35]. The fine decoration of the Mn₃O₄-Co₃O₄ nanoparticles contributed to this activation [36]:

	Mn3+ + H2O2 → Mn2+ + HOO• + H+
	HOO• + Mn3+ → Mn2+ + H+ + O2
	Mn2+ + H2O2 → Mn3+ + HO• + OH−
	2OH• → H2O + ½ O2
Overall:	H2O2 → H2O + ½ O2

In this process, a trace amount of the H₂O₂ intermediate undergoes chemical disproportionation to generate O₂ species, which can be used as reactants for a further ORR [37]. Similarly, OH⁻ is produced by catalyzing the decomposition of H₂O₂ with Co₃O₄ nanoparticles [34]. In addition, the Co³⁺ species of Co₃O₄ provide a surface electronic state, which carries electrons from bulk oxides to produce an excited cationic state, facilitating the ORR [37]. Moreover, the uniform metal oxides decorated on MWCNTs confine agglomerations, which contribute to the active surface areas. Additionally, this process can enhance the electronic conductivity due to the efficient generation of ions in the alkaline media [38,39].

Fuel tolerance and stability are vital characteristics of a high-performing ORR catalyst. Urea crossover tests were performed in a 1.0 M KOH solution containing 5 mL of 0.33 M urea, as shown in Figure 9a,b. The current density immediately decreased by 43% at 0.7 V for Pt/C after injecting the urea solution, whereas that of the Mn₃O₄-Co₃O₄/MWCNT catalyst showed a minor decrease and retained 92% of its initial value. This behavior demonstrated that Pt/C could not withstand the urea crossover. With regard to the long-time durability of the catalysts, the voltammograms showed high stability of Mn₃O₄-Co₃O₄/MWCNT with no distinct current change; the initial limiting current density was 0.672 V (vs. RHE). Figure 9c,d presents the chronoamperometric responses of the commercial Pt/C and the as-prepared Mn₃O₄-Co₃O₄/MWCNT catalyst at a rotation rate of 1000 rpm in a 1M KOH solution for 5000 s. The commercial Pt/C dropped approximately 10% of its initial current due to the formation of Pt hydroxide on its surface. It is clear that Mn₃O₄-Co₃O₄/MWCNT underwent an inconsiderable decrease and remained stable at 97% of its initial current after 5000 s. The obtained results demonstrate that Mn₃O₄-Co₃O₄/MWCNT has a superior long-term stability in comparison with commercial Pt/C. [39]. All chronoamperometric responses demonstrate that the Mn–Co oxide composite possessed high durability and favorable kinetics. Therefore, the composite can be used effectively as a cathode catalyst in an alkaline fuel cell.

Furthermore, we evaluated the cell performance of the prepared sample used as a catholyte in DUFC. Figure 10 shows the polarization and power density curves of Mn₃O₄-Co₃O₄/MWCNT at room temperature. A current density of 2.13 mA cm⁻² and a maximum power density of 0.4226 mW cm⁻² at 50 °C were obtained.

4. Conclusions

We prepared a Mn₃O₄-Co₃O₄/MWCNT nanocomposite with high catalytic performance in the ORR by growing Co₃O₄ and Mn₃O₄ nanoparticles on the surface of MWCNTs via a solvothermal method without adding any reducing agents. The electrochemical evaluation showed that the composite presented both improved activity compared to the individual components and stability in an alkaline solution comparable to that of commercial Pt/C. Moreover, the Mn₃O₄-Co₃O₄/MWCNT catalyst could tolerate urea poisoning and did not degrade in the long term. Thus, Mn₃O₄-Co₃O₄/MWCNT can be considered as a low-cost, environmentally friendly, and efficient catalyst for DUFCs.