Introduction

The concerns related to the depletion of fossil fuels and the environmental pollution caused by their consumption call for an immediate search for alternate green and sustainable energy sources1,2. In this regard, hydrogen is a promising fuel with certain superiorities, such as high energy density, natural abundance and no harmful products created after burning3,4,5,6. Hydrogen generation through electrocatalytic water splitting is one of the most effective strategies, as water is cheap and abundant and hydrogen can be produced with high purity6,7,8. Water is thermodynamically highly stable and suitable electrocatalysts are required for water splitting. So far, noble metals-based electrocatalysts, such as IrO2/RuO2 for OER and Pt for HER, are benchmark catalysts for water splitting in terms of performance and stability, however, their high cost and scarcity are serious issues9,10. Efforts have been devoted to the exploration of suitable electrocatalysts with low overpotential, fast kinetic, and good stability for both hydrogen and oxygen evolution reactions (HER/OER) as well as overall water splitting11,12. For instance, various metal oxides8,13,14, sulfides15,16, selenides17,18, phosphides19,20, carbides and carbon nitride composites8,21, have been investigated as electrocatalysts. However, most of the non-oxide materials are unstable and convert to oxides or oxy-hydroxides during OER22,23. In this case, metal oxides can be used as stable bifunctional electrocatalysts but their activity is relatively low.

Some important pre-requisites for developing promising catalysts are that the catalyst must be composed of earth-abundant, eco-friendly and cost-effective materials. In addition, they should show high electrocatalytic activity and stability under catalytic conditions. Various reports suggest that multi-metal oxides are remarkably better than simple binary oxides24,25. The intrinsic activity can be tuned by tailoring the composition of multi-metal oxides. The activity enhancement occurs due to the creation of multiple active sites, which are critical for the rational design of superior electrocatalysts. In particular, creating atomic-scale synergistic active sites in single-phase systems is highly desirable.

Based on these principles, spinel ferrites can be earth-abundant, inexpensive, eco-compatible and highly effective materials, because of their sustainability under harsh conditions, high redox features, easy modulation in valence states, and enhanced electrical conductivity26,27,28. In particular, spinel NiFe2O4 and MgFe2O4 have gained recent attention and are being explored for possible use in energy storage and conversion technologies29,30,31,32,33,34. The two spinels crystalize in a cubic crystal symmetry and possess cations with variable oxidation states. These features enable the formation of a vast array of materials with tailored properties and promising electrochemical behavior35,36,37. To improve the electrocatalytic properties of NiFe2O4 and MgFe2O4, various strategies such as the formation of hybrid materials, doping, and interphase engineering have been employed. For instance, the Co-doped NiFe2O4 electrocatalyst with optimized composition (Co6.25Fe18.75Ni75Ox) showed excellent OER activity with a small overpotential of 186 mV and a low Tafel slope of 38.5 mV/dec at 10 mA/cm2 in 1 M KOH38. A tailored interface engineering strategy has also been reported to enhance the electrocatalytic activity of NiFe2O4/NiTe for advanced energy conversion applications39. Diverse hybrid materials of NiFe2O4 or MgFe2O4 and graphene40,41, carbon nanotubes42, and other systems have also been investigated as potential candidates for electrocatalytic applications.

Although significant efforts have been devoted to improve the electrochemical properties of NiFe2O4 and MgFe2O4, there is a lack of study on the role of their substitutional solid solutions in electrocatalytic HER and OER. The formation of a solid solution allows for engineering the materials’ properties and performance via tunable composition43. Both NiFe2O4 and MgFe2O4 crystalize in a cubic symmetry and have comparable charges (+ 2) and crystal sizes of Ni2+ (0.69 Å) and Mg2+ (0.72 Å). This similarity allows the formation of a solid solution over an entire range of compositions and varying the stoichiometric amounts of Ni2+ and Mg2+ will result in the generation of multiple different active sites, a prerequisite for highly active catalysts.

Besides tailoring the composition, other ways of bolstering the performance are by engineering the morphology and avoiding the use of capping agents. The capping agents with long alkyl chains block the active sites and act as insulating layers, resulting in poor charge transfer properties44. Generally, the development of 2-dimensional morphology improves catalytic performance by providing enhanced surface area, which in turn results in more active sites. Likewise, the use of surfactants can be avoided by adopting solid-state pyrolysis of the precursors route. Therefore, this study reports a facile solvent-free approach to prepare nanosheets of inexpensive and eco-friendly Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions via pyrolysis of metal acetylacetonates. This straightforward and green synthesis procedure has afforded the formation of monophasic nanoferrite solid solutions that crystallize in a cubic spinel structure. The electrochemical performance of the nanoferrite solid solutions was investigated by tuning the local chemical environment of Ni in NiFe2O4 via Mg substitution.

Experimental

Chemicals

Nickel (II) acetylacetonate (98%, Merck-Schuchardt), magnesium (II) acetylacetonate (98%, Merck-Schuchardt), and iron (III) acetylacetonate (97%, Sigma-Aldrich). These metal complexes were used as received.

Synthesis of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions

The Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions of different stoichiometric compositions were prepared by solventless thermolysis of metal acetylacetonates. Briefly, for the synthesis of NiFe2O4 nanoparticles, 0.10 g (0.39 mmol) of nickel acetylacetonate and 0.27 g (0.78 mmol) of iron acetylacetonate were mixed to form a solid mixture. The solid mixture was ground using a pestle and mortar for ≈ 15 min to homogenize the mixture. The precursor mixture was then placed into a ceramic boat, which was placed in a reactor tube. The reactor tube was then placed inside the carbolite tube furnace, followed by thermal treatment at 450 °C, at a heating rate of 20 °C per minute for 1 h. After 1 h of annealing, the heating was switched off, and the furnace was allowed to cool to room temperature. The reactor tube was taken out of the furnace upon cooling, and the product was collected for analysis without any post-treatment. Similarly, the synthesis of MgFe2O4 nanoparticles was achieved by employing similar procedures except that magnesium acetylacetonate was used instead of nickel acetylacetonate and the amount of magnesium and iron complexes were maintained in the same mole ratio of 1:2.

For the synthesis of Ni1−xMgxFe2O4 (x = 0.2, 0.4, 0.6, 0.8) solid solutions, a known quantity of nickel acetylacetonate was partially substituted by appropriate amounts of magnesium acetylacetonate by adjusting the mole ratios of Mg and Ni in the intervals of 0.2, 0.4, 0.6, and 0.8, while keeping the amount of iron acetylacetonate unchanged in the reaction mixture. The reaction procedures for the entire series of solid solutions were kept similar to those employed to synthesize the ternary nickel and magnesium ferrites.

Characterization of the Ni1−xMgxFe2O4 nanocatalysts

Structural analysis of the Ni1−xMgxFe2O4 nanoparticles was ascertained by powder X-ray diffraction (p-XRD) analysis employing a Bruker AXS D8 Advance X-ray diffractometer. The instrument uses nickel-filtered Cu Kα radiation (λ = 1.5418 Å) at 40 kV, 40 mA. SEM imaging was carried out on a ZEISS-Auriga Cobra SEM Field Emission Scanning Electron Microscope (FE SEM) while EDX elemental analysis was performed on a JEOL JSM-7500F Field Emission Scanning Electron Microscope (FE-SEM) equipped with Energy Dispersive X-ray spectroscopy (EDX). The SAED, TEM and HRTEM analyses were performed on a JEOL 2100 HRTEM at accelerating voltages of 200 kV.

Electrochemical characterization

The electrocatalytic property of the Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) was examined via a three-electrode system using a Versastat 4-500 electrochemical workstation (Princeton Applied Research, Oak Ridge, TN, USA). For the preparation of the working electrode, the electrode paste was synthesized using Ni1−xMgxFe2O4 material (80%), PVDF (10%), carbon black (10%) with N-methyl pyrrolidinone (NMP) solvent as active materials, binder, and conducting agent, respectively. The paste was dipped into the clean Ni foam and dried for 48 h. While Pt wire was used as a counter electrode, Hg/HgO was used as the reference electrode. To examine the performance of the electrocatalyst for HER and OER, linear sweep voltammetry (LSV) was carried out at a scan rate of 2 mV/s. Also, electrochemical impedance spectroscopy (EIS) was performed at the potential of 0.6 V (V, SCE) in the frequency range of 0.05 Hz–10 kHz at an applied AC amplitude of 10 mV. For the stability of electrocatalysts, chronoamperometry techniques were utilized at the potential of 0.57 V (V, SCE). All measurements for electrocatalysts were conducted using 1 M KOH electrolyte.

Results and discussion

Structural analysis of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions

A set of Ni1−xMgxFe2O4 nanocatalysts were synthesized by a direct solid-state thermolysis process. Figure 1a shows the typical powder X-ray diffraction (p-XRD) patterns of the as-prepared Ni1−xMgxFe2O4. The diffraction peaks found in the pure ternary systems prepared at x = 0 and x = 1 are exclusively indexed with the cubic spinel crystal system having the space group Fd3̅m. The pristine ferrites are consistent with the cubic phases of pure trevorite, NiFe2O4 (ICDD # 01-086-2267) and magnesioferrite, MgFe2O4 (ICDD # 01-089-3084) for x = 0 and x = 1, respectively. The p-XRD data for the nanoferrites with x = 0.2 to x = 0.8 indicate the formation of solid solution phases with variable stoichiometric composition of Ni2+ and Mg2+ in the spinel matrix. Notably, their diffraction peaks lie in between those of pure NiFe2O4 and MgFe2O4, and these solid solutions maintained the same cubic symmetry throughout the composition regulation.

Figure 1
figure 1

(a) Powder-XRD patterns (b) Variation of lattice constant (left y-axis) and cell volume (right y-axis) with Mg2+ content for Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions.

Full size image

The lattice constants (a = b = c) of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) nanospinels were ascertained from the p-XRD data by employing the formula shown by Eq. (1) and the results are shown in Table 1.

$$\frac{1}{{d^{2} }} = \frac{{\left( {h^{2} + k^{2} + l^{2} } \right)}}{{a^{2} }}$$
(1)
Table 1 Lattice parameter (a), crystallite size (d), unit cell volume (V), and EDX composition of nanospinel Ni1−xMgxFe2O4 solid solutions at various magnesium contents (x).
Full size table

The lattice constants of NiFe2O4 were found to be 8.313 Å, conforming to those reported in the standard data (8.337 Å, ICDD #: 01-086-2267). After the incorporation of Mg2+, a slight increase in the values of the lattice parameters is observed, which is also ascribed to the slightly larger size of Mg2+ (0.72 Å) relative to the Ni2+ (0.69 Å)45. The lattice parameters computed for pure MgFe2O4 (8.344 Å) are also comparable with the values reported in standard data (8.369 Å, ICDD #: 01-089-3084). The values of lattice parameters were then plotted as a function of Mg2+ content (x) as shown in Fig. 1b. It is obvious that the lattice constant increases in a linear fashion with Mg2+ inclusion from 8.313 Å for NiFe2O4 to 8.344 Å for MgFe2O4. This linear relationship between the lattice parameters and Mg2+ content is in agreement with Vegard’s law46. The values of the lattice constants obtained in this study are consistent with previously reported values for magnesium-substituted nickel ferrite nanoparticles prepared via a co-precipitation route47. Likewise, the data in Table 1 and Fig. 1b demonstrate that the cell volume increases monotonically with increasing magnesium content. All these findings confirm the successful inclusion of Mg2+ into the crystal structure of NiFe2O4. The Debye–Scherrer formula (Eq. 2), Ref.48 was employed to compute the average crystallite size of Ni1−xMgxFe2O4 samples.

$$L = \frac{0.89\lambda }{{\beta cos\theta }}$$
(2)

In the formula, L = average crystallite size, λ = X-ray wavelength, β = full width at half maximum, and θ = Bragg’s angle of the (311) plane. The average crystallite sizes of the as-prepared Ni1−xMgxFe2O4 nanoparticles vary between 10 and 20 nm (Table 1). The average crystallite size obtained for the pristine nickel ferrite was larger compared to those exhibited by magnesium-substituted samples.

Elemental compositional analysis

The composition and elemental distributions of Ni, Mg, Fe, and O were analyzed by energy-dispersive X-ray (EDX). The EDX results (Supplementary Information, Fig. S1) indicate the presence of Ni, Fe and O for x = 0, and Mg, Fe and O for x = 1 in the desired ratio. For the solid solution nanocrystals with compositions x = 0.2–0.8, the presence of Ni, Mg, Fe, and O was confirmed. A summary of the atomistic composition of each of the components in the alloyed nanoferrites is provided in Table S1. The stoichiometry of the elements obtained is consistent with the expected values within the substitution limits, suggesting that there is no side reaction or significant loss of the starting materials. In Fig. 2, the relationship between the amount of Ni2+ and Mg2+ detected from EDX with respect to the mole fraction of [Mg]/[Mg + Ni] in precursor feed indicates a decrease in nickel content with a linear increase in magnesium content. In addition, the EDX mapping of the as-prepared Ni1−xMgxFe2O4 solid solutions is given in Fig. 3, indicating that the distribution of the respective elements in the spinel structure is nearly uniform, ruling out the possibility of de-alloying or phase segregation. This also confirms the formation of the solid solution between NiFe2O4 and MgFe2O4 in the single-crystalline alloyed nanospinel.

Figure 2
figure 2

Change in Mg and Ni content as a function of mole fraction of [Mg]/[Mg + Ni] in precursor feed.

Full size image
Figure 3
figure 3

Elemental mapping of Ni1−xMgxFe2O4 (0.2 ≤ x ≤ 0.8) nanosheets.

Full size image

Microstructure and morphological studies

The morphology of nanoparticulate Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions was examined by using SEM and TEM analyses. Figure S2 shows SEM micrographs of Ni1−xMgxFe2O4 nanoparticles. It can be seen that for all compositions, nanoclusters were formed with a narrow size distribution, and a closer look indicates that the clusters are composed of smaller particles grouped together. To have clear information about the particle morphology, size and microstructure of Ni1−xMgxFe2O4 samples, TEM analysis was carried out. The TEM images displayed in Fig. 4 show sheet-like structures with truncated edges. The nanosheets were of different sizes in the range of 10–100 nm, where smaller particles coalesced together to form relatively bigger sheets. The TEM analysis also shows that the size of nanosheets was relatively small for NiFe2O4 and the size increases with increasing Mg concentration in Ni1−xMgxFe2O4. The nanosheets formed were of low thickness and some sheets were horizontally stacked on top of other sheets. Additionally, the lattice fringes with interplaner spacings of d = 4.81, 2.94, 2.51, 2.40, and 2.08 Å were observed, corresponding to the (111), (220), (311), (222), and (400) planes of cubic spinel Ni1−xMgxFe2O4 nanosheets. These results are consistent with the characteristic d-spacing and Miller indices observed from p-XRD data. Further, the SAED patterns (displayed as insets) reveal well-defined spots, which suggest the crystalline nature of Ni1−xMgxFe2O4 samples.

Figure 4
figure 4

TEM and SAED (insets) images of Ni1−xMgxFe2O4 solid solutions over the entire range, by solid-state decomposition of metal acetylacetonate complexes.

Full size image

Electrocatalytic HER and OER

For electrocatalytic analysis, Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) will be referred to as NMF with a specified “x” value. The HER activities of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) catalysts were investigated under alkaline conditions (1 M KOH) in a usual three-electrode arrangement. Figure 5a displays the LSV polarization curves of Ni1−xMgxFe2O4 with varying mole ratios of Ni and Mg at a scan rate of 5 mV/s. Pure NiFe2O4 displays low catalytic performance with an overpotential of 159 mV which was needed to produce a current density of 10 mA/cm2. Upon Mg incorporation, the electrocatalytic activity was enormously improved as manifested by the reduction of overpotential from 159 mV of bare NiFe2O4 to Ni0.8Mg0.2Fe2O4 (135 mV), Ni0.6Mg0.4Fe2O4 (130 mV), Ni0.4Mg0.6Fe2O4 (121 mV), Ni0.2Mg0.8Fe2O4 (134 mV), and MgFe2O4 (153 mV). Remarkably, at the current density of 10 mA/cm2, the Ni1−xMgxFe2O4 (x = 0.6) electrode exhibited the best electrocatalytic activity for HER with an overpotential of 121 mV which is smaller compared to its counterparts. This reduction in overpotentials demonstrates that the incorporation of the proper content of Mg in the crystal lattice of NiFe2O4 can effectively improve its catalytic activity for HER. These results further demonstrated that the value of overpotential of Ni0.4Mg0.6Fe2O4 at the geometric current density of 10 mA/cm2 is superior to many binary and ternary metal oxide catalysts such as NiFe2O4 (290 mV)49, Ni/Co3O4 (145 mV)50, Fe2O3/NCs (350 mV)51, δ-MnO2 (196 mV)52, and MgFe2O4 (402 mV)53. It is noteworthy that the Ni0.4Mg0.6Fe2O4 also shows great superiority to other previously reported HER electrocatalysts summarized in Table S2. In Fig. 5b, the Tafel slopes of all electrodes were measured from the LSV measurements. Remarkably, the Tafel slope of Ni0.4Mg0.6Fe2O4 was found to be 125 mV/dec, which is lower than that of NiFe2O4 (136 mV/dec), Ni0.8Mg0.2Fe2O4 (146 mV/dec), Ni0.6Mg0.4Fe2O4 (130 mV/dec), Ni0.8Mg0.2Fe2O4 (184 mV/dec), and MgFe2O4 (143 mV/dec). The reduction of Tafel slope from 136 mV/dec (NiFe2O4) to 125 mV/dec (Ni0.4Mg0.6Fe2O4) may be ascribed to probable modification effect of the surface electronic state due to incorporation of Mg element, which in turn enhances the inherent conductivity of Ni0.4Mg0.6Fe2O454. These changes indicate that among other factors, the electrochemical kinetics depend on the ratio of Mg dopants. Generally, the lower Tafel slope of the electrode indicates better process kinetics, even when significant H2 generation is needed at elevated voltage or current densities.

Figure 5
figure 5

(a) HER polarization curves, (b) HER Tafel slopes, (c) OER polarization curves, (d) OER Tafel slopes, (e) Nyquist plots at 0.5 V, and (f) Chronoamperometric measurements at 0.55 V, for Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) electrodes (NMF stands for Ni1−xMgxFe2O4 (0 ≤ x ≤ 1)).

Full size image

The OER activities of a series of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) catalysts in 1 M KOH solution were also investigated. The polarization curves of Ni1−xMgxFe2O4 catalysts with different mole ratios of Ni to Mg show a significant decrease in the overpotential upon Mg incorporation in NiFe2O4 (Fig. 5c). While the pristine NiFe2O4 needed an overpotential of 407 mV to deliver a current density of 10 mA/cm2, the Mg-doped NiFe2O4 samples exhibited lower overpotentials of 391, 389, 382, and 284 mV for x = 0.2, 0.4, 0.6 and 0.8, respectively. Also, the pristine MgFe2O4 displayed a lower overpotential of 326 mV compared to NiFe2O4. Among the studied series of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) catalysts, Ni0.2Mg0.8Fe2O4 demonstrated the best OER performance with a lower overpotential of 284 mV, within the window of potential examined. The electrocatalytic activity demonstrated by Ni0.2Mg0.8Fe2O4 surpasses many other previously reported metal oxide-based catalysts. For example, MnFe2O4 was reportedly synthesized by Li et al. and showed an overpotential of 470 mV at a current density of 10 mA/cm2 in alkaline media55. In a similar study, CoFe2O4 exhibited 370 mV under similar electrolytic conditions. Hirai et al. reported that Mn3O4 needed an overpotential of 600 mV to produce a current density of 10 mA/cm2 in 1 M KOH solution. They further reported the synthesis of Mn2.4Co0.6O4 which exhibited a high overpotential of 510 mV56. Also, Co3O4 nanocubes fabricated by Chen et al. were reported to display an overpotential of 580 mV (at 10 mA/cm2) in alkaline electrolytes57. Table S3 shows the comparison of the values of overpotentials Ni0.2Mg0.8Fe2O4 with other non-precious metal catalysts. The values of the Tafel slope indicated in Fig. 5d were obtained in the range of 54–112 mV/dec. The low overpotential and small Tafel slope make Ni0.2Mg0.8Fe2O4 a more promising OER catalyst.

The electrical conductivity of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) nanocatalyts was elucidated by electrochemical impedance spectroscopy. The Nyquist plot displayed in Fig. 5e shows that the pristine NiFe2O4 nanoparticles possess a large semicircle, demonstrating poor electron transfer capability, compared to Ni1−xMgxFe2O4 (0.2 ≤ x ≤ 0.8) solid solutions. The lowest charge resistance values displayed by the solid solutions imply intimate contact between the current collector and Ni1−xMgxFe2O4 and is an indication of more swift charge transfer kinetics. The results confirm further that the incorporation of Mg in spinel NiFe2O4 lattices contributed to the improvement of electrical conductivity via reduction of the charge transfer resistance, and consequently boosting the electrocatalytic properties of Ni1−xMgxFe2O4 electrodes. Of all electrode configurations investigated, Ni0.2Mg0.8Fe2O4 shows the smallest semicircle, indicating superior conductivity, and hence high electrocatalytic activity towards water splitting.

The diameter of the semicircle obtained at a lower frequency provides information on the ionic resistance of the electrolyte, indicating series resistance (Rs) with charge transfer resistance Rct. The Rs for all the samples studied were the same, while Rct depends on the composition. The simplest equivalent circuit for these samples is where one resistance Rs in series with Cdl and Rct parallel to the Cdl. The composition of Ni1−xMgxFe2O4 with X = 0 showed the highest charge-transfer resistance while with X = 0.8 displayed the lowest charge-transfer resistance.

Evaluation of the catalyst’s electrochemical stability is important for practical water splitting applications. To explore the electrochemical stability of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) electrocatalysts, chronoamperometry measurements were performed at 0.55 V. Remarkably, no significant changes in the current density were observed during 17 h tests, signifying excellent electrochemical stability of all Ni1−xMgxFe2O4 systems in the alkaline electrolyte (Fig. 5f). The stability of Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) electrocatalysts was further examined by continuous LSV scans in 1 M KOH. The results indicate the absence of significant change in the polarization curve after 1000 cycles, signifying superior stability of the nanocatalysts for both HER (Fig. 6a,b, and Fig. S3) and OER (Fig. 6c,d, and Fig. S4) in alkaline solution.

Figure 6
figure 6

Comparison of HER (a, b) and OER (c, d) polarization curve between LSV 1 curve and LSV 1 k curve for Ni1−xMgxFe2O4 (x = 0 and 0.6) electrodes.

Full size image

Conclusion

In conclusion, this study reports a composition-controlled fabrication of homogeneous Ni1−xMgxFe2O4 (0 ≤ x ≤ 1) solid solutions by solventless pyrolysis method. Experimental investigation demonstrates that by regulating the molar composition of Mg and Ni in the preparation process, the physicochemical and electrochemical performance of the material were modified. The as-synthesized Ni0.4Mg0.6Fe2O4 nanoparticles exhibited the best electrocatalytic activity for HER with an overpotential of only 121 mV which is much smaller compared to its analogues, at a current density of 10 mA/cm2 and the electrode exhibits good stability during long-term electrolysis. Meanwhile, Ni0.2Mg0.8Fe2O4 showed the best OER activity, requiring an overpotential of 284 mV to deliver the same current density within the window of potential examined. The outstanding electrocatalytic performance of these solid solutions is largely ascribed to the enhanced conductivity due to surfactant free surfaces, nanoparticulate nature and synergic effect of different metals (Mg, Ni and Fe) which either directly or indirectly promoted the catalytic activity. The results described in this work pave the way for the design of mixed spinel oxides with high electrocatalytic activity for applications in sustainable energy systems.