Introduction

There is a growing demand to develop clean and renewable energy sources [1]. To meet this energy growth, fuel cells have gained worldwide attention as a green, high-efficiency, and sustainable electrochemical energy conversion device [2]. Among them, the microbial fuel cell (MFC) as a bio-electrochemical apparatus has gained researchers’ interest owing to its capability of directly harvesting electrical energy during the wastewater treatment process [2, 3]. However, the MFCs’ widespread application is hindered by the high overpotentials that arise from the sluggish kinetics of the cathode oxygen reduction reaction (ORR) at neutral pH conditions [4, 5]. Although platinum (Pt) is the most efficient and widely used electrocatalyst for ORR, its commercial applicability in fuel cells is limited due to its scarcity and high cost [6]. Thus, developing inexpensive and effective ORR catalysts such as metal-free electrocatalyst [1], non-precious metal (Mn, Co, and Fe)-based catalysts [7], transition metal oxides (TMOs), and metal-nitrogen-carbon (M-N-C) [8] as alternatives that could reduce the dependence on the currently employed Pt, which has become an urgent research area.

Among these alternatives, TMOs (Fe, Ni, Mn, Cu, etc.) have gained increasing interest as ORR cathode catalysts in MFCs because of their abundance, ease of doping, structural diversity, low cost, high ORR activity, and environmental friendliness [8,9,10]. However, the TMOs suffer from synthesis uncontrollable growth, high electrical resistance, poor electrical conductivity, and low catalytic activity [11, 12]. Besides, the dissolution and sintering during the fuel cell reactions result in decreasing fuel cells’ ORR performance [13]. Combining two or more low-cost TMOs into different spinel structures (MTMOs; mixed transition metal oxides) is a powerful way to boost the ORR activity. Due to its ability to mix well into one material and variable oxidation states [11], good cycle stability [10], low cost, environmental friendliness, and higher electrical conductivity than the simple TMOs [14,15,16]. Recently, MTMOs with a spinel structure and AB2O4 general molecular formula appear as promising ORR electrocatalysts [17].

Spinel-metal ferrites (MFe2O4; M = Mn, Ni, Cu, Co, etc.) are one of the most interesting spinel-oxide structures that are used in many fields: physics, biomedical, and environmental fields [17, 18] due to their unique electrical and magnetic properties, electronic, photocatalytic ability, and outstanding conductivity [18, 19]. Their excellent electrical conductivity and electrocatalytic activity are due to the presence of surface redox-active centers (B3+‏/B2+‏ and A3+‏/A2+) for O2 adsorption and activation that arise from the electron bouncing among various valence states of metals in O-sites [15, 20].

MFe2O4’s physicochemical properties are significantly dependent on the morphology, dimension, crystallographic orientation, and composition of nanostructures [15]. From another point of view, the ORR efficiency is mainly affected by the mass transfer, so for high performance and ideal ORR catalysts, a model carbon support material is required [1]. This support should display high electrical conductivity, good mechanical properties, chemical stability [12], relative chemical inertness [13], high porosity, and high surface [21]. Conductive carbon supports such as carbon nanotubes, graphene oxide, carbon black, and activated carbon (AC) are highly accepted solutions to enhance the electrocatalytic properties of ORR in MFC applications [11, 22].

Among these carbon materials, AC is preferred because it has good characteristics such as its availability on a large scale from secondary sources [8], its low cost [23], and mesoporous structure (2–50 nm) [24] since mesopores could improve the diffusion of oxygen molecules by limiting the decrease in surface area, increasing the exposed active sites, and facilitating the charge transport to and from the catalytic site [21, 24]. This improves the electrical conductivity and the electrocatalytic activity [13]. Using AC as support for MFe2O4 spinel oxides is a promising approach to improve its ORR electrocatalytic properties so it could be used in a neutral solution [25]. The incorporation of magnetic metal oxide nanoparticles onto the AC surface is widely reported all over the literature [25], for example, CoFe2O4/AC composites [26], CuFe2O4/AC adsorbents [27], and MnFe2O4/AC composites [28].

Besides the above factors, the ORR activity of the metal oxides is strongly dependent on synthesis strategies [15]. The traditional methods for the synthesis of ferrite materials have some drawbacks like non-uniform and large particle size, impurities, and low surface area, which seriously reduce their catalytic activity and prevent further improvement in their performance [15, 29]. To overcome these difficulties, some wet chemical processes like the solvothermal method, co-precipitation technique, sol–gel processing, microemulsion, and auto-combustion are used for the production of MFe2O4 powders with excellent properties to meet the requirements of the new applications [15, 19, 29]. Co-precipitation is a facile method that is often used to synthesize multi-metal oxides due to its simplicity, the homogeneity of the obtained material, and the application of low temperatures [18, 30]. Additionally, the synthesis of nanocomposite at a higher pH imparts its higher stability, so, a pH > 10 is selected during the co-precipitation. Also, a low-temperature hydrothermal synthesis is assumed to be beneficial in preventing nanoparticle agglomeration, enhancing the interaction of metallic moieties with the carbon support, and retaining structural integrity to favor mass and charge transport [30].

According to the published literature, only a few studies reported the kinetics carried on MFe2O4-based AC catalysts at neutral pH in MFCs. Since then, the kinetics behavior of those catalysts in neutral conditions has remained elusive, and the ORR mechanism under investigation has been complicated and is still not fully understood [4, 31]. Kodali et al. [32] discovered that comparing existing literature on MTMO-based-carbon ORR cathode catalysts in MFCs is quite complicated and becomes even more difficult to determine for the following reasons: (i) the catalysts are frequently prepared using different precursors and fabrication methods, and (ii) the performances are only compared to Pt or AC. Also, (iii) various working conditions lead to more diversity in the output, and (iv) it has not been well-established which transition metal (M = Fe, Co, Mn, Ni, etc.) -based carbon catalysts have superior ORR electrochemical performance in neutral media.

The present study introduces the synthesis of novel MFe2O4-based AC composites by a combination of different transition metals (Fe, Cu, Mn, and Ni) via a facile co-precipitation approach. The as-synthesized materials are characterized by physicochemical tools, and their ORR electrochemical performance is studied using cyclic voltammetry (CV) and linear sweep voltammetry (LSV) in neutral media (phosphate buffer solution, pH = 7).

Experimental

Materials and reagents

Carbon Vulcan XC-72R was purchased from Fuel Cell Store (USA), nitric acid, manganese (II) nitrate hydrate, ethanol, urea, and sodium hydroxide pellets were purchased from Honeywell, iron(III) nitrate nonahydrate was purchased from Alpha ChemiKa, nickel(II) nitrate hexahydrate was purchased from Alfa Aesar, copper(II) nitrate trihydrate was purchased from Fisher Scientific, and polyethylene glycol 20,000 was purchased from Merck. All materials and chemicals were used with no further purification.

Preparation of electrocatalysts

The as-received Carbon Vulcan XC-72R was chemically treated as previously described [33]; in brief, 3.0 g of carbon vulcan dispersed in 150 ml of nitric acid was sonicated in an ultrasonic bath for 1 h and then mechanically stirred at 60 °C for 3 h. The treated carbon vulcan was then washed with deionized (DI) water until reaching a neutral pH and dried in a vacuum drier at 90 °C till the treated carbon reached a constant weight. The acid-treated carbon (activated carbon; AC) was used as a support for the metal ferrite: MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC.

Typically, MnFe2O4/AC nanocomposite was synthesized according to the following modified methods: [34,35,36] and as shown in Fig. 1; 3.0 g of AC was dispersed in 250 ml of DI water and sonicated using an ultrasonic probe for 30 min. Under uniform mixing and constant magnetic stirring, 50 ml of PEG and 20 ml of urea aqueous solution were added to ferric nitrate and manganese nitrate solutions (solution 1). PEG was used as a coating agent to prevent nanoparticles from agglomeration during the precipitation process [37]. Solution 1 was added to the dispersed AC. The total mixture was heated at 80 °C for 4 h under constant mechanical stirring for gel formation and full access of the metals in the solution to the AC surface. The co-precipitation of transition metal oxides was achieved at 12.5 pH by the addition of NaOH aqueous solution. Then, the solution was filtered and washed with distilled water several times till neutral pH was attained, and dried at room temperature for 24 h under a vacuum. The obtained dried powder was sonicated in the remaining 100 ml PEG solution for 30 min, filtered, and washed with DI water several times to obtain the PEG-coated MnFe2O4/AC composite nanoparticles. The composite was dried in a vacuum oven drier up to a constant weight. All the above steps were repeated to prepare the other two combinations: NiFe2O4/AC and CuFe2O4/AC. Finally, all the prepared composites were calcined at 550 °C for 5 h.

Fig. 1
figure 1

Schematic diagram of the synthesis procedure

Full size image

Physical characterization techniques of catalysts

The crystal structure of the composites was analyzed by X-ray diffraction (XRD) (Pan Analytical Model X’Pert Pro) equipped with a Cu-Kα radiation source (λ = 0.1542 nm). Fourier transform infrared spectroscopy (FTIR) was measured by the Nicolet Is-10 FTIR spectrophotometer adopting the KBr technique; Thermo Fisher Scientific and Raman spectra were recorded at room temperature by SENTerra Bruker with a 532-nm excitation laser source. The surface morphology for all samples was investigated by field emission-scanning electron microscopy (FE-SEM, ZWESS, FEG (Field Emission Gun)) equipped with an energy dispersive X-ray spectrometer (EDX) operating at resolution (eV) 128.8, Amp time (µs) 3.84, live time (s) 100, Mag:49, and KV:5. The microanalysis was examined using high-resolution transmission electron microscopy (HR-TEM, JEOL JEM-1230). Particle size measurements of the synthesized composites were determined by dynamic light scattering (DLS) measurements using a Zetasizer Nano ZS (Malvern Instruments, UK). The nitrogen adsorption–desorption isotherms were determined by Quantachrome NOVA 3200 automated gas sorption system, USA at − 196 °C and after degassing at 120 °C and 10−5 mm Hg for 24 h to calculate the Barrett-Joyner-Halenda (BJH) surface area and pore size distributions based on the desorption branches of the isotherms. Electron spin resonance (ESR) was done using a Bruker, Germany through a standard rectangular cavity of ER 4102. Electromagnetic radiation is a coupled electric field (E) and magnetic field (H), and it oscillates at the same frequency (v) within a range of 1–100 GigaHertz (1 Hz = 1 cycle per second). The number of spins (N) is calculated based on the following equation:

$$N =K \frac{{H}_{\mathrm{e}} \times {(\Delta H)}^{2} \times (\frac{P.H}{\mathrm{wt}})}{{G}_{\mathrm{e}} \times {H}_{\mathrm{m}} \times \sqrt{P}}$$
(1)

where:

N No. of spins, H∘ magnetic resonance, ∆H peak width, Ge gain, Hm modulation amplitude, P power, and K EPR constant = 1013 spins

Electrochemical measurements

The electrocatalytic activity of the synthesized catalysts toward ORR was evaluated by CV and LSV. These measurements were carried out at room temperature in a normal three-electrode system using a Voltamaster 6 Potentiostat and a rotating disk electrode (RDE) with a rotation speed control unit. The system consists of platinum (Pt) wire (as a counter electrode), an Ag/AgCl electrode (as a reference electrode), and a glassy carbon electrode (GCE; with a geometrical surface area of 0.196 cm2) coated with the catalyst thin film as a working electrode.

Before catalyst loading on GCE, it was polished with 0.05 µm alumina powder to obtain a mirror-like surface. Then, it was washed with DI water and acetone. After that, GCE was loaded with a paste of 1 mg catalytic powder mixed with a drop of isopropanol. Then, a drop of 5 wt% Nafion solution (as a binder) was added after isopropanol drying. Finally, the second drop of isopropanol was added to the paste, and it was let to dry overnight at room temperature. In this study, all the measured potentials were relative to the potential of the Ag/AgCl electrode, and the current densities have been normalized to the electrode geometric area.

Before conducting any electrochemical test, we perform 50 cycles with a scan rate of 50 mV s−1 to clean the catalyst surface. Any electrochemical test for oxygen reduction was repeated 10 times until reaching the steady-state performance and taking the last one as a final result. CV and LSV tests were carried out in N2- and O2-saturated 0.1 M (100 mM) phosphate buffer aqueous solution (PBS; as the electrolyte for electrochemical experiments, pH = 7.2) at scan rates of 50 mV s−1 and 10 mV s−1, respectively. LSV was conducted at different rpm values (0–1400 rpm) in a voltage range (− 1000 to 1000 mV/Ag/AgCl). Before CV and LSV, the electrolyte was bubbled with N2 and O2 for 20 min and half an hour, respectively. The exact ORR kinetic parameters including kinetic current density (JK) and electron transfer number (n) were analyzed based on the Koutecky-Levich (K-L) equation [6] as follows:

$$\frac{1}{j}=\frac{1}{{j}_{L}}+\frac{1}{{j}_{k}}=\frac{1}{{B\omega }^\frac{1}{2}}+\frac{1}{{j}_{k}}$$
(2)
$$B=0.62\; n\; F\; {C}_{\circ }\;{({D}_{\circ })}^\frac{2}{3}\;{\upsilon }^{\frac{-1}{6}}$$
(3)

where j, jL, and jK are the measured, limiting diffusion, and kinetic current density (mA cm−2), respectively, and ω represents the electrode rotation speed. B is the slope of the K-L plot, n is the number of electrons transferred per O2 molecule in ORR, and F is Faraday’s constant (F = 96,487 C mol−1). In 0.1 M PBS, C∘ (1.117 × 10−6 mol mL−1) is the O2 bulk concentration, D∘ (1.9 × 10−5 cm2 s−1) is the O2 diffusion coefficient, and υ (0.01073 cm2 s−1) is the electrolyte kinetic viscosity [31].

Results and discussion

Electrocatalysts physical characterization

XRD analysis

Figure 2a shows the XRD pattern of activated carbon, with a distinct broad peak at 2θ ≈25° and a weak peak at 2θ = 43° corresponding to (002) and (100) crystallographic planes, respectively. The recorded peaks were indexed to the hexagonal crystalline activated carbon [21, 38].

Fig. 2
figure 2

XRD of (a) AC and (b) CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC

Full size image

Figure 2b shows the XRD patterns of MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC. CuFe2O4/AC exhibited spinel diffraction peaks at 2θ values of 18.39°, 30.06°, 34.69°, 36.15°, 44.11°, and 62.19° which correspond to crystallographic planes (101), (112), (103), (211), (220), and (224) respectively, in agreement with the literature [27] (ICDD: 04–006-4007). Furthermore, CuFe2O4/AC spectra showed the presence of a weak peak at 2θ = 38.91° for CuO (111) [39] (ICDD: 04–006-4186). Spinel diffraction peaks were observed in MnFe2O4/AC spectra at 2θ = 24.25°, 33.25°, 35.65°, 41.03°, 49.59°, 54.15°, 62.51°, and 64.21° congruent to (220), (311), (222), (400), (422), (333), (440), and (521), which agrees with the literature [40, 41]. The MnFe2O4/AC main diffraction peak at 2θ = 33.25° is considered a measurable degree of its crystallinity [42], and its sharpness indicates that the MnFe2O4/AC possesses a higher crystallinity [29]. Finally, NiFe2O4/AC has diffraction peaks at 2θ = 18.52° (111), 30.38° (220), 35.80° (311), 37.43° (222), 43.48° (400), 57.59° (511), and 63.11° (440) [43] (ICDD: 04–006-6580). Upon the deposition of different TMOs on the AC surface, the intensity of carbon peaks decreased due to the decrease in graphitization degree, and this proves the combination of metal oxides with AC [5].

Raman spectroscopy

Figure 3 shows the Raman spectrum of NiFe2O4/AC and AC. The Raman spectra of AC exhibited two obvious peaks around 1332 cm−1 (D band) and 1609 cm−1 (G band) of carbon-based materials [44]. Generally, the D peak is concerned with a sequence of lattice defects and carbon atoms’ disorder degree, including bond angle and bond length disorder, as well as hybridization. The G peak is evolved from the in-plane stretching vibration of sp2 hybridized carbon atoms, and it reveals the carbon graphitization degree [45]. The peak intensity ratio of ID/IG could be used to evaluate the degree of defects and disordered structures for each material [46]. The ID/IG ratio of the AC sample is approximately equal to 1.06 implying the good defect degree of this sample. In general, the appropriate defect structure and higher disorder degree in carbon-based catalysts would benefit in increasing the affinity of oxygen adsorption by supplying more active sites, thus enhancing the ORR catalytic activity [22, 47]. The deposition of transition metals (Fe, Cu, Mn, and Ni) on the AC surface affected the D and G bands, and these Raman bands disappeared after the deposition, thereby changing the ID/IG value. This change is ascribed to the change in geometry, purity, crystallinity, structure, and surface composition of the AC. However, Fig. 3 showed that NiFe2O4/AC exhibited peaks around 486 and 691 cm−1 that were pointed to NiO, while no peaks were detected for CuFe2O4/AC and MnFe2O4/AC, since these composites are Raman silent [48]. These results proved the deposition of Fe, Ni, Cu, and Mn on the AC surface.

Fig. 3
figure 3

Raman spectra of NiFe2O4/AC and AC

Full size image

FT-IR spectroscopy

The structures of AC, CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC were examined by FT-IR spectroscopic analysis as shown in Fig. 4. The FT-IR spectra of the prepared materials showed strong bands in the lower mid-infrared 400–700 cm−1 range that originate from the stretching vibrations of the metal–oxygen bond (M–O; M = Mn, Ni, Cu, and Fe) [13]. The strong peaks observed at 578 and 619 cm−1 were attributed to (Mn–O) and (Ni–O) stretching vibration bands of MnFe2O4/AC and NiFe2O4/AC, respectively [42, 48]. In the case of CuFe2O4/AC, the peak at 430 cm−1 was assigned to Fe–O, while the peak at 603 cm−1 was confirmed to Cu–O bond [49, 50]. The peaks around 1140 cm−1 were devoted to (C–N) stretching vibrations for MnFe2O4/AC and NiFe2O4/AC, whereas the peak at 1543 cm−1 was assigned to (N–H) bending of amide II for CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC [51, 52]. All the fabricated materials show a peak around 1650 cm−1 assigned to (C═O stretching) [48, 50]. All the catalysts (AC, CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC) show a strong and broad peak around 3450 cm−1 due to the presence of (OH-free asymmetric stretching), which is attributed to the physisorbed water molecules that were absorbed from the surrounding atmosphere [13, 21, 48]. The absorption peak around 3650 cm−1 (one peak) is assigned to (N–H) stretching frequency of amide II.

Fig. 4
figure 4

FT-IR of CuFe2O4/AC, MnFe2O4/AC, NiFe2O4/AC, and AC

Full size image

Surface analysis

The surface properties and the pore structure of the prepared composites (MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC) were studied by N2 adsorption–desorption isotherms. It is commonly regarded that carbon materials with a multi-level porous structure provide a large number of catalytic active sites with fine dispersion and suitable channels to facilitate reactant diffusion and electron transfer [53]. Figure 5a showed that the catalysts exhibited classical type IV isotherms and a well-defined hysteresis loop close to high relative pressure in the P/P∘ range (0.7–1) for all samples, which is characteristic for mesoporous materials. This is greatly important to ORR since mesopore is mainly required to enable mass transport of sufficient oxygen in ORR [10, 54]. Furthermore, according to the desorption branch, the BJH surface area and total pore volume of MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC catalysts have partially decreased than that of AC after deposition, which could be due to pore-filling and pore-blocking of channels during the precipitation process [1], as confirmed by the disappearance of AC micropores in the prepared composites.

Fig. 5
figure 5

(a) N2 adsorption–desorption isotherms and (b) pore size distribution for MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC

Full size image

As listed in Table 1, MnFe2O4/AC had a low BJH surface area of 53.236 m2 g−1 with a total pore volume of 0.09 cm3 g−1, all of which originated from mesopores compared to that of AC that possessed a high BJH surface area of 180.149 m2 g−1 and total pore volume of 0.44 cm3 g−1. The decrease in BJH area resulted from a decrement in mesopores that was likely owed to the occupancy of partial mesopores for AC by the deposition of Mn and Fe nanoparticles [13]. The BJH surface area and total pore volume of mesopores were slightly increased in CuFe2O4/AC compared to MnFe2O4/AC, to be 56.481 m2 g−1 and 0.1 cm3 g−1, respectively. On the other hand, the BJH surface area and total pore volume of mesopores in NiFe2O4/AC were remarkably increased to be 76.259 m2 g−1 and 0.27cm3 g−1, respectively.

Table 1 Structural and textural parameters for MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC composites
Full size table

Pore size distribution curves are shown in Fig. 5b, pointing out that all the samples have mesopore sizes in the range of 10–50 nm. This indicates that mesopores contributed the majority of the volume, which was thought to be beneficial for ORR and the transport of electrolyte ions (proton) into the interior catalyst layers, while their pore sizes are 10.476, 11.831, and 50.501 nm for MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC, respectively. In particular, the mesoporous structure is profitable for the rapid transmission of O2, fuel, and electrolytes, which can accelerate the redox reaction rate and improve electrochemical performance [10]. The electrochemical performances are highly linked to the catalysts’ surface-active sites and mobility into the electrolyte [46].

FE-SEM and EDX microscopy

Figure 6 shows the FE-SEM images, mapping, and EDX analysis obtained for MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC composites. The morphology of all the prepared samples shows to some extent a spherical-like nanoparticle aggregate and uniformly grown on the AC surface [13]. Generally, using AC as a support improves the dispersion of the transition metal oxides and reduces their agglomeration [50]. As shown in Fig. 6, CuFe2O4/AC shows spherical nanoparticles sintered together to form a connected structure with a smaller primary size 44.7–71.6 nm. The micrographs of NiFe2O4/AC show clusters of fine particles sticking jointly. The sample surface contains number of fine pores or voids (defects) that could be attributed to the large quantity of gases released during the calcination process. The grain size estimated from the micrograph is in the range of 19.97–43.98 nm. MnFe2O4/AC reveals semi-quasispherical particles with unqualified sharp crystals in a cluster shape, homogeneously distributed in the AC frame. The grain size ranged from 35.78 to 107.6 nm. The EDX spectrum analysis confirms the characteristic peaks of the Cu, Fe, Mn, Ni, C, N, and O, while no other peaks of foreign elements were observed. The involved table in Figure 5 depicts the wt.% of Cu, Fe, Mn, Ni, C, N, and O elements, and it proves the presence of these elements in the prepared composites. The chemical analysis showed that the Wt% Fe/wt% Cu=1.83 and the at% of Fe/at%Cu= 2.08 while the Wt% of Fe/wt% of Ni=2.09 and the at% of Fe/Ni=2.1 and Wt% of Fe/Mn= 2.17 and the at% of Fe/Ni=2.13. This proves the formation of spinels. The consistent elemental mappings in selected areas additionally reveal a homogeneous distribution of O, Fe, C, N, Mn, Cu, and Ni. Remarkably, carbon had the lowest percentage of all the prepared composites, revealing that the metal-oxide layers covered most of the catalysts’ surface [48]. These results are consistent with the FE-SEM findings.

Fig. 6
figure 6

SEM, Mapping, and EDX Images for CuFe2O4/AC, NiFe2O4/AC, and MnFe2O4/AC

Full size image

HR-TEM and DLS analysis

HR-TEM of carbon vulcan (Fig. 7a) shows aggregated particles with spherical shape and irregular size. The reduction in size and ordered uniform structure (Fig. 7b) is noticed and attributed to the shrinkage and dispersion of particles during the acidic treatment of carbon vulcan. TEM images of all composites show low agglomeration, high dispersion, and homogeneous distribution of the composite nanoparticles on the AC surface [10, 50]. CuFe2O4/AC (Fig. 7c) displays a uniform and wide distribution cuboid structure, and the DLS analysis shows that the average particle size is 712 nm. Figure 7d of NiFe2O4 displayed a semi-spherical particles with a random range of particle size (295–995 nm) as confirmed in the histogram of DLS analysis. HR-TEM image of MnFe2O4 (Fig. 7e) reveals the shell formation of the hollow nanoparticles, which are made up of some nanocrystals. There are fundamentals for the formation of a hollow structure as follow [55]: (1) a deep permeation of metal ions within the carbon nanoparticles, and (2) the two rates of the precursor shell formation and that of carbon nanosphere decomposition must match with each other. When the heating temperature is low, the breakdown rate of the carbon species is slow, allowing the metal atoms to assemble within the carbon templates and interconnect to form shell layers. The diameter distribution from the DLS curve shows that MnFe2O4 has a wide range (220–1106 nm) of particle size.

Fig. 7
figure 7

HR-TEM images for (a) carbon vulcan, (b) AC, (c) CuFe2O4/AC, (d) NiFe2O4/AC, and (e) MnFe2O4/AC

Full size image

ESR spectroscopy

Electron spin resonance (ESR) measurements were carried out to verify the presence of oxygen vacancies and demonstrate the assimilation of the oxygen defects, qualitatively, in MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC. Figure 8 points out that all the samples display a single ESR signal at a g-value of 2.29 (g//) and 2.7 (g┴) for MnFe2O4/AC. A CuFe2O4/AC sample has a g// of 2.49 and a g┴ of 2.91, whereas NiFe2O4/AC has a g// of 2.42 and a g┴ of 2.93, which is related to unpaired electrons trapped in oxygen vacancies (OV) [56].

Fig. 8
figure 8

ESR spectra of MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC composites

Full size image

The magnetic properties can be illustrated by peak-to-peak line width (∆Hpp), resonant magnetic field (Hr), and g-factor [57]. Table 2 shows that ∆Hpp values increase from 671.9 to 972.3 and 1019.8 Oe for CuFe2O4/AC, NiFe2O4/AC, and MnFe2O4/AC, respectively. Strong dipole interactions resulted in a large broad peak-to-peak line width (ΔHPP) and higher crystallite size. Our results show that the broad signal narrows down with MnFe2O4/AC and it gives the highest Hr (Oe), while small ΔHPP gives rise to strong super-exchange interactions as in NiFe2O4/AC. However, the large ΔHPP indicates the large ferrimagnetic particles [58]. ΔHPP was found to decrease with the decrease in crystallite size of the particle due to strong super-exchange interactions. The intensity of the ESR peak is much higher for NiFe2O4/AC compared with MnFe2O4/AC or CuFe2O4/AC. This indicates that NiFe2O4/AC is opulent in oxygen defects.

Table 2 ESR data for MnFe2O4/AC, CuFe2O4/AC, and NiFe2O4/AC composites
Full size table

It has been recognized that the decrease in the metal oxidation number happens when the crystal consumes an oxygen atom in vacancy formation. Also, theoretically, it was noticed that the reduction in particle size increases the metal stability in its several applications [59].

The g-factor values increase from 2.12 to 2.16 with the increase in the particle size and calcination temperature, hence increasing the microscopic magnetic interactions. In ferrites, the difference between ∆Hpp and g-factor values can explain the super-exchange and dipole–dipole interactions [60]. Table 2 illustrates that the value of the resonant magnetic field reduces from 2989 to 2976 Oe due to the incorporated metal.

Electrochemical performance of the electrocatalysts toward ORR

The ORR activity of CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC composites was investigated using RDE measurements (CV and LSV) in N2- and O2-saturated 100 mM PBS neutral solution at scan rates of 50 mV s−1 and 10 mV s−1, respectively. Figure 9 shows the CVs of CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC composites, which reveals that no redox peaks were observed for the samples in both the absence (black line) and presence (red line) of oxygen. In the presence of oxygen, all of the composites demonstrated capacitive behavior and efficient catalytic activity toward ORR [14]. MnFe2O4/AC has a higher ORR activity than NiFe2O4/AC and CuFe2O4/AC because at potential − 24 mV vs. Ag/AgCl, MnFe2O4/AC exhibits a larger current density (− 138.2 µA cm−2) response than NiFe2O4/AC (− 104.75 µA cm−2) and CuFe2O4/AC (− 83.09 µA cm−2).

Fig. 9
figure 9

CV curves in N2-saturated (black line), O2-saturated (red line), and background current subtract (green line) of NiFe2O4/AC, MnFe2O4/AC, and CuFe2O4/AC composites at 50 mV s−1

Full size image

For more studying the electrochemical properties of the prepared composites, the background subtraction of cyclic voltammetry (BS-CV) technique is applied. The concept of background current subtraction denotes to subtraction of current density in the presence of O2 to the current density without O2 to exclude the reaction of the background solution [61]. After using BS-CV technique (Fig. 9), no redox peaks were observed for the three samples. The BS-CV (green line) reveals that the electrocatalysts still keep on their capacitive behavior, and this confirms its electrocatalytic activity.

To further investigate the ORR performance of CuFe2O4/AC, MnFe2O4/AC, and NiFe2O4/AC composites, LSV tests were performed in an O2-saturated 100 mM PBS neutral medium at different rpm by employing a voltage that varies linearly across the working electrode and recording the corresponding current change. Figure 10a shows that MnFe2O4/AC had onset potential (Eonset) (− 0.223 V vs. Ag/AgCl) and kinetic current density (jK) (− 5 mA cm−2) higher than CuFe2O4/AC (− 0.280 V vs. Ag/AgCl, − 3.05 mA cm−2), and NiFe2O4/AC (− 0.270 V vs. Ag/AgCl, − 2.67 mA cm−2). This implies that MnFe2O4/AC has enhanced catalytical activity and is kinetically more facile for O2 reduction, so it is the highest ORR active electrocatalyst among the prepared composites in this work.

Fig. 10
figure 10

(a) LSV curves comparison of NiFe2O4/AC, CuFe2O4/AC, and MnFe2O4/AC composites at 800 rpm and 10 mV s−1 scan rate, and the corresponding (b) Tafel fitting lines

Full size image

Additionally, the kinetic performance of the prepared composites was further evaluated by Tafel plots which were calculated from the LSV curve at 800 rpm (Fig. 10b). The Tafel curve is one of the main indicators for ORR catalytic activity. The electrocatalyst Tafel slopes are in the order of MnFe2O4/AC (− 330 mV dec−1) < NiFe2O4/AC (− 414 mV dec−1) < CuFe2O4/AC (− 577 mV dec−1), which agrees with the ORR results. In addition, the Tafel slope is closely related to electron transfer rate, and the fact that the MnFe2O4/AC composite presents the lowest Tafel slope suggests that it can easily adsorb oxygen molecules onto its surface and activate it. Also, this means a quicker electron transfer rate, thereby exhibiting the best and fastest ORR kinetics and performance under neutral conditions [21, 53, 62].

Figure 11a shows that the kinetic and diffusion current density increase with the rotation speed, implying the decrease of diffusion resistance among the catalyst surface and saturated oxygen molecules [46]. Furthermore, the ORR pathway at the MnFe2O4/AC surface was kinetically and quantitatively studied by using the LSV plots collected from RDE experiments performed at different rotation speeds, and the slope of the curve was obtained at applied potentials between − 0.5 and − 0.7 V vs. Ag/AgCl to calculate the number of electron transfer from the K-L equation.

Fig. 11
figure 11

(a) LSV curves of MnFe2O4/AC composite at different rpm and 10 mV s−1 scan rate, and the corresponding (b) K-L plots at different potentials

Full size image

The fitted lines of K-L plots (Fig. 11b) show good linearity at different potentials, and this indicates first-order ORR reaction kinetics toward the concentration of dissolved oxygen in neutral electrolytes [46, 53]. The average electron transfer number was 4.3 demonstrating that ORR catalyzed by MnFe2O4/AC mainly proceeds via a 4-electron mechanism. Among the prepared materials, MnFe2O4/AC showed the highest ORR catalytic activity owing to a good combination of crystal structure and morphology. This improved ORR activity of MnFe2O4/AC can be ascribed to the strong interaction between Fe, Mn, and AC [63]. Besides, the transition valence state of Mn plays an essential role in encouraging the adsorption and reduction of oxygen [12, 64].

Compared with Pt/C electrocatalyst, MnFe2O4/AC displayed current density equal to that of Pt/C (− 5.1 mA cm−2) [65]. Also as shown in Table 3, MnFe2O4/AC showed a higher current density than electrocatalysts in the literature. The high electrocatalytic performance of MnFe2O4/AC may be attributed to both the defective structure and mesoporosity of AC and the redox activity of metal centers (Fe and Mn) [30]. Aside from the high dispersion of Fe and Mn nanoparticles [24], and the predominant presence of manganese in a + 2 oxidation state, Mn(II) entities are highly accepted as efficient centers for oxygen adsorption and electron transfer via the Mn(II) to Mn(III) transformation [30, 64]. These factors helped in the contact between the active sites and electrolyte solution, hence enhancing the ORR process [12].

Table 3 ORR performance comparison of electrocatalysts reported in literature
Full size table

Conclusion

MFe2O4/AC composites were successfully prepared and characterized through different routes. The high dispersion of the ferrites on the surface of the carbon is considered to be responsible for the high electrochemical activity. Additionally, our results collected from LSV carried at 800 rpm revealed that MnFe2O4/AC had higher Eonset (− 0.223 V vs. Ag/AgCl), kinetic current density (jK) (− 5 mA cm−2), and lower Tafel slope (330 mV dec−1) than NiFe2O4/AC (− 0.270 V vs. Ag/AgCl, − 2.67 mA cm−2, and 414 mV dec−1), and CuFe2O4/AC (− 0.280 V vs. Ag/AgCl, − 3.05 mA cm−2, and 577 mV dec−1), respectively. Meanwhile, the K-L plot calculations for ORR catalyzed by MnFe2O4/AC in neutral phosphate buffer solution (pH = 7) and at different rpm confirmed that it is a 4e− pathway mechanism. MnFe2O4 displayed the highest ORR performance, and this can be pointed to (i) mesoporosity that provides more accessibility to active sites, (ii) highly dispersed Mn and Fe species that work as the main active sites, and (iii) the large ferrimagnetic particles. Thus, MnFe2O4/AC is suggested to be a model alternative for the commercial Pt/C catalyst. Finally, this research work has contributed to a simple, efficient, and low-cost synthesis technique for the development of naturally abundant and efficient ORR composite electrocatalysts with improved electrocatalytic performance. MnFe2O4/AC is considered a low-cost Pt-alternative ORR cathode catalyst for MFC applications at pH = 7. We suggest taking into consideration our research results, since it sheds light on the design and synthesis of spinel-structure ferrite oxides and expands their applications in energy conversion technologies, including fuel cells, particularly, MFCs operated in neutral conditions at pH = 7.