Synthesis and Characterisation of Cobalt Ferrite Coatings for Oxygen Evolution Reaction

Abstract

In this paper, two novel procedures based on powder sedimentation, thermal treatment, and galvanostatic deposition were proposed for the preparation of porous cobalt ferrite (CoFe₂O₄) coatings with a metallic and organic binder for use as catalysts in the oxygen evolution reaction (OER). The electrochemical properties of the obtained electrode materials were determined as well, using both dc and ac methods. It was found that cobalt ferrite coatings show excellent electrocatalytic properties towards the oxygen evolution reaction (OER) with overpotential measured at a current density of 10 mAcm⁻² from 287 to 295 mV and a Tafel slope of 35–45 mVdec⁻¹. It was shown that the increase in the apparent activity of the CoFe₂O₄ coatings with an organic binder results mainly from a large electrochemically active area. Incorporation of the nickel binder between the CoFe₂O₄ particles causes an increase in both the conductivity and the electrochemically active area. The Tafel slopes indicate that the same rate-determining step controls the OER for all obtained coatings. Furthermore, it was shown that the CoFe₂O₄ electrodes exhibit no significant activity decrease after 28 h of oxygen evolution. The proposed coating preparation procedures open a new path to develop high-performance OER electrocatalysts.

1. Introduction

The oxygen evolution reaction (OER) plays an essential role in state-of-art energy storage and conversion devices. Nowadays, the major challenge is to obtain low-cost and high-performance catalysts for those reactions. Electrodes made of Pt, IrO₂, and RuO₂ are known OER catalysts, however they are not appropriate for large-scale applications because of their cost. Transition metals, such as Fe, Ni, Co, and Mn, show high electrocatalytic activity and stability in alkaline solutions. Thus, they are promising candidates for replacing precious metals catalysts. Sulphides, selenides, nitrides, carbides, hydroxides, oxy-hydroxides, and oxides of transition metals are extensively used as efficient electrocatalysts [1,2,3,4,5,6]. Because of their high catalytic activity, relatively low cost, and environmental friendliness, numerous studies of transition metal oxides have been performed. For example, it has been found that NiO [7], Co₃O₄ [8], MnO₂ [9], and Fe₂O₃ [10] exhibit high catalytic activity toward OER compared to RuO₂ and IrO₂. It was also found that nickel- and cobalt- based composites usually produce overpotentials in the OER reaction around 350–450 mV at a current density of 10 mAcm⁻² [4,5,6].

Generally, spinel mixed metal oxides with formula AB₂O₄ (A, B = metal) exhibit higher electrochemical activity than the single-metal oxides. This is because of the electron jump among different valence states of ions in octahedral sites and additional metallic redox-active centres [11,12,13,14,15,16]. Between various spinel oxides, Fe-based spinels, e.g., CuFe₂O₄, NiFe₂O₄, and CoFe₂O₄, have been reported as promising electrocatalysts for OER [17]. However, the CoFe₂O₄ ferrite is one of the most interesting, especially for its excellent chemical stability, efficient electrocatalytic properties, high specific capacitance, low cost, and environmental friendliness. Therefore, CoFe₂O₄ ferrite is a promising candidate as an electrode in water splitting, lithium-ion batteries, and supercapacitors [18,19,20]. Most of the mixed transition metal oxides catalysts are powders obtained by different methods, e.g., solvothermal, hydrothermal, electrodeposition, or spin coating [21,22]. Usually, to obtain an electrode, those powders are applied to a conductive substrate using a particle binder. However, polymeric binders are characterized by relatively poor electrical conductivity, which may influence the entire electrode’s electrocatalytic activity. Additionally, the low durability of polymeric binders may also cause the detachment of coating that is especially noticeable at higher current densities and vigorous O₂ evolution [23]. Another frequently used method for electrode preparation is to anchor metal oxides on a carbon substrate. In this case, to assure sufficient functional groups, the carbon substrate is treated with harsh oxidative agents, which can deteriorate the electrical conductivity of the substrate and cause environmental pollution [24]. To improve the performance of these materials, various methods of their modifications have been reported, e.g., by creating nanoporous structures, synergistic metal–metal interactions, or embedding nanoparticles of catalysts in carbon nanomaterials [24,25,26]. However, there is still a need to design new MFe₂O₄ electrocatalysts with a very porous surface, high electrical conductivity, and durability.

This work has been carried out to study the electrocatalytic activity of the CoFe₂O₄ coatings during oxygen evolution reaction. Two different methods of electrode preparation have been used. In the first method, a coating consisting of a combination of CoFe₂O₄ powder and PEG 4000 (as a binder) was deposited by sedimentation on a metallic substrate. Subsequently, such prepared coating was subjected to heat treatment at 800 °C in an inert atmosphere. The second method used was the electrolytic co-deposition of nickel and CoFe₂O₄ powder. Electrodeposition seems to be an interesting alternative in coatings production that contain mixed metal oxides powders and metallic binders. Additionally, a nickel substrate subjected to the analogous treatment like CoFe₂O₄ coatings was also investigated.

2. Results and Discussion

2.1. CoFe₂O₄ Powder and Coating Characterization

The phase composition of the as-obtained and annealed (800 °C, 1 h, air) CoFe₂O₄ powder is shown in Figure 1. In both cases, X-ray diffraction patterns show the presence of the reflexes coming from the single-phase spinel cobalt ferrite with cubic crystal structure (space group: Fd$\stackrel{-}{3}$m, lattice constant: a = 8.4101 Å, reference code: 04-009-8438 ICDD PDF4+ 2015). Note that XRD lines obtained for as-obtained CoFe₂O₄ particles are broadened due to some lattice strain [27]. The values of crystallite size calculated using the Williamson–Hall method [28] for as-obtained and annealed CoFe₂O₄ powders were 10(1) nm and 15(1) nm, respectively. Thus, one can state that CoFe₂O₄ powders are nanocrystalline phases and a 50% increase in the crystallite size is observed after annealing.

Figure 2 shows the SEM images of the examined samples, i.e., CFO_PEG, CFO_Ni, and ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$, and Ni substrate. One can see that the images obtained for CFO_PEG and CFO_Ni electrodes represent a relatively homogenous distribution of cobalt ferrite particles, while the image obtained for ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ shows some agglomerates of CoFe₂O₄ particles separated from each other. Thus, heat treatment of CoFe₂O₄ powder (800 °C, 1 h, air) influences the shape of the electrode surface. Note that CoFe₂O₄ particles, independent of fabrication procedure applied, distinctly enlarge the surface roughness of the electrode in comparison with nickel substrate. The increased surface area should increase the electrocatalytic activity of the CoFe₂O₄ coatings towards oxygen evolution.

2.2. Electrochemical Characterisation of CoFe₂O₄ Coatings

Investigations of the electrolytic oxygen evolution on the Ni and CoFe₂O₄ electrodes were carried out using dc and ac methods. In Figure 3a, quasi-stationary polarisation curves j vs. E are shown. It can be observed that the CoFe₂O₄ electrodes are more efficient than the Ni electrode. Among the CoFe₂O₄ electrodes, the lowest and the highest performance is observed for ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ and CFO_Ni electrodes, respectively. For the linear part of polarisation curves, the Tafel equation η = a + blog|j| can be used to determine characteristic parameters of the electrode-electrolyte system. The parameters provide information about the mechanism (Tafel slope b) and the rate (apparent exchange current density j₀) of OER. It can be seen from Figure 3b that all investigated electrodes show a well-defined Tafel region with a slope changing from 35 to 56 mVdec⁻¹ (see also

Table 1. The Tafel slope b, the apparent exchange current density j₀, the overpotential at a current density of 10 mAcm⁻² η₁₀ and the average roughness factor R_f(av) obtained for CoFe₂O₄ electrodes in 1 M KOH solution during OER.

	b (mVdec−1)	j0 (Acm−2)	η10 (mV)	Rf(av)	j0/Rf(av) (Acm−2)
Ni	56	3.2 × 10−10	438	12	2.7 × 10−11
CFOPEG	35	4.2 × 10−11	295	1850	2.3 × 10−14
CFONi	45	4.6 × 10−9	287	2487	1.8 × 10−12
${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$	51	1.0 × 10−9	359	93	1.1 × 10−11

).

The OER mechanism consists of the following steps [29]:

$$\mathrm{M}+{\mathrm{OH}}^{-}\leftrightarrow \mathrm{MOH}+{\mathrm{e}}^{-},$$

(1)

$$\mathrm{MOH}+{\mathrm{OH}}^{-}\leftrightarrow {\mathrm{MO}}^{-}+{\mathrm{H}}_2\mathrm{O},$$

(2)

$${\mathrm{MO}}^{-}\leftrightarrow \mathrm{MO}+{\mathrm{e}}^{-},$$

(3)

$$2\mathrm{MO}\leftrightarrow 2\mathrm{M}+{\mathrm{O}}_2.$$

(4)

Either (1) or (3) electrochemical reaction may be the rate-determining step (rds). When the surface coverage by OH is much smaller than one and the reaction (1) is rds, then the Tafel slope b = 2.3RT/βF. However, if the reaction (3) is rds, then b = 2.3RT/(1 + β)F [29]. Assuming transfer coefficient β = 0.5 and temperature 25 °C, the theoretical Tafel slope for the reaction (1) is equal to 120 mVdec⁻¹ and for the reaction (3) 40 mVdec⁻¹. The Tafel slopes b obtained for all investigated CoFe₂O₄ coatings indicate that the OER proceeds via the same mechanism, with the third reaction step as the rate-determining step. The experimental b values higher than 40 mVdec⁻¹ indicate that the transfer coefficient is lower than 0.5. According to literature reports, it can be caused by anion adsorption and the nonuniform distribution of the surface electric field on the rough electrodes [29].

The parameter j₀ obtained for CFO_Ni and ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ is ca. one order of magnitude higher in comparison with nickel electrode and ca. two orders of magnitude higher than for CFO_PEG. Such behaviour could be explained by the higher electrical conductivity of nickel (σ ~ 10⁶ Ω⁻¹cm⁻¹) in comparison with cobalt ferrite particles (σ ~ 10⁻⁷ Ω⁻¹cm⁻¹) [30]. Therefore, for CFO_Ni and ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ electrodes, the electrolytically deposited nickel used as a particle binder causes an increase in the apparent exchange current density of the whole electrode.

The apparent activity of different electrodes could be compared using overpotential η determined at selected current density (in this work, j = 10 mAcm⁻²). For CoFe₂O₄ electrodes, the parameter η₁₀ is from 80 mV to 150 mV lower in comparison with the nickel substrate (see Table 1). Thus, one can state that CoFe₂O₄ electrodes characterise higher apparent activity towards OER with respect to Ni electrode. It was also found that the apparent activity of CFO_Ni electrode modified by nickel binder is the highest among all investigated electrodes. For comparison, different kinds of CoFe₂O₄ electrodes reported in the literature for OER in alkaline media (1 M KOH or 1 M NaOH) are gathered in

Table 2. Comparison of cobalt ferrite catalysts reported in the literature.

Catalyst	b (mVdec−1)	η10 (mV)	Reference
CoFe2O4 (powders)	35	295	this work
CoFe2O4 (powders)	45	287	this work
CoFe2O4 (powders)	69	360	[12]
CoFe2O4 (powders)	126	435	[12]
CoFe2O4 (hollow nanofibers)	95	414	[11]
CoFe2O4 (thin films)	54	490	[13]
CoFe2O4 (nanoparticles)	73	378	[14]
CoFe2O4 (nanoplates)	61	360	[15]
CoFe2O4 (nanofibers)	107	340	[16]

. All these data confirm that the CFO_PEG, CFO_Ni, and ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ electrodes reveal high activity towards OER.

In practical terms, an important criterion is the long-term stability of the catalysts. The stability of the cobalt ferrite electrodes in 1 M KOH solution was tested using the chronopotentiometry method at the constant current density of 50 mAcm⁻² (see Figure 3c). Percentage change of the measured potential determined for time period from 1 to 28 h is 5% for Ni, 3% for CFO_PEG, 1% for CFO_Ni and 7% for ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$. Thus, it can be stated that all investigated electrodes are relatively stable in an alkaline environment. Figure 4 shows the SEM images and corresponding maps of Ni, Fe, Co elements distribution for the CFO_PEG and CFO_Ni electrodes after 28 h of oxygen evolution. Comparison with images obtained before OER (Figure 2) indicates that the surface of CFO_PEG and CFO_Ni electrodes remained almost unchanged. Thus, SEM images confirm the conclusion obtained from electrochemical measurements. The distribution maps of elements indicate that the electrode surface consists of the porous coating made of CoFe₂O₄ particles (or their agglomerates) and nickel substrate in the case of CFO_PEG or nickel binder in the case of CFO_Ni.

Impedance spectra measured for selected potentials are shown in Figure 5. For all investigated electrodes, two semicircles on Nyquist plots are observed. There is a small potential-independent semicircle in the high-frequency region, and a potential-dependent semicircle in the low-frequency region. Such a shape of spectra is typical for porous electrodes. The potential-independent semicircle can be related to the geometry of the electrode surface, whereas the potential-dependent semicircle corresponds to the Faradaic reaction [31,32,33]. Thus, a model with two-time constants can adequately describe the response of porous CoFeO₄ electrodes during OER.

An electrical equivalent circuit used to describe the investigated process is shown in Figure 6. It contains solution resistance (R_s), coating capacitance (CPE_l.), coating resistance (R_l), charge transfer resistance (R_ct), and double layer capacitance (CPE_dl). Note that the deviation of solid electrodes from purely capacitive behaviour (caused by physical nonuniformity or uneven distribution of active sites) has been taken into account using CPE elements instead of capacitors. The following equation describes the impedance of the CPE element: Z_CPE = 1/T(jω)^ϕ, where T is the capacity parameter and ϕ is the dispersion parameter related to the depression angle [31,32,33].

For all obtained electrodes, the value of charge transfer resistance decreases with increasing overpotential, which corresponds to the decreasing diameter of the low-frequency semicircle (see Figure 5). Furthermore, comparing R_ct with the R_l, it was found that R_ct >> R_l. Thus, R_ct governs the electrode kinetic. The relation η vs. log(R_ct⁻¹) for all investigated electrodes is shown in Figure 7. It was found that R_ct values obtained for CoFe₂O₄ at lower overpotentials can be over 100 times lower than that obtained for Ni electrode and in both cases gradually decreases with increasing overpotential. Additionally, η vs. log(R_ct⁻¹) curves show well-defined linear regions with slopes from 40 to 48 mVdec⁻¹. Note that the obtained slopes correlate with Tafel coefficients obtained from η vs. log(j) curves.

The double layer capacitance C_dl can be determined using the equation: C_dl = [T/((R_s + R_l)⁻¹ + R_ct⁻¹)^(1-ϕ)]^1/ϕ [32]. The relation between C_dl and η for all obtained coatings is shown in Figure 8. It was found that the smallest value of C_dl was obtained for the nickel electrode, as should be expected. CFO_Ni and CFO_PEG coatings exhibit, respectively, ca. 200 times and ca. 150 times higher C_dl (average value) in comparison with nickel electrode.

Note that double layer capacitance is directly proportional to the electrochemically accessible surface area. Assuming that the double layer capacitance of the smooth oxide surface is 60 µFcm⁻² [29,33], the roughness factor R_f can be calculated as R_f = C_dl/60 µFcm⁻². The intrinsic exchange current density (calculated as j₀/R_f) indicates that the superior catalytic properties of CFO_PEG electrode are mainly the result of the large electrochemically accessible surface area. In turn, the excellent CFO_Ni electrode properties result from both high intrinsic catalytic activity and large electrochemically accessible surface area. Note that the most useful case for the electrocatalysis is when the whole area of an electrode is accessible to reactants. As can be seen in Figure 8, the double layer capacitance obtained for all electrodes decreases with increasing overpotential. This may result from the fact that oxygen bubbles partially block the electrode surface. However, it should be emphasized that even at high overpotentials, CFO_PEG and CFO_Ni electrodes are characterized by much higher values of C_dl compared to the nickel substrate.

3. Materials and Methods

3.1. Cobalt Ferrite Powder Synthesis

The CoFe₂O₄ powder was synthesised using the coprecipitation method [27] and reagents of analytical purity (POCh, Gliwice, Poland). Hence, 0.6 moldm⁻³ of FeCl₃·6H₂O and 0.3 moldm⁻³ of CoCl₂·6H₂O solutions were mixed in ultrapure water with a resistivity of 18.2 MΩcm. Precipitating agent, i.e., NaOH (2 moldm⁻³), was slowly added to the stirred chlorides solution. Coprecipitation was conducted in a Teflon vessel at a temperature of 50 °C for two hours. Subsequently, the synthesised powder was rinsed in distilled water, centrifuged for five minutes at 300 rpm, and dried at 50 °C for 48 h. The powder was also ground in an agate mortar to break agglomerates into smaller particles. In further study, the as-received and post-thermal treatment (800 °C, 1 h, air) CoFe₂O₄ powder was used. The phase composition of CoFe₂O₄ powder was determined by X-ray diffraction technique (XRD) using a Philips X’Pert PW 3040/60 diffractometer (PANalytical, Almelo, Netherlands) equipped with Cu_Kα radiation. The ICDD cards were used for phase identification.

3.2. Preparation of Cobalt Ferrite Coatings

Cobalt ferrite coatings were deposited on a nickel plate (Nickel 201, Ni ≥ 99.0%). Before the deposition process, the nickel plate with a working area of 0.5 cm² was mechanically polished using abrasive papers (P320, P600, P1000), and subsequently rinsed in acetone in an ultrasonic bath for 10 min. The studied coatings were produced by applying two different procedures. In the first case, a mixture of the thermally treated CoFe₂O₄ powder (20 mg per 10 mL of solution) and polyethylene glycol 4000 (PEG 4000) (20 mg per 10 mL of solution) in acetone was prepared. Additionally, 5 mg of sodium dodecyl sulfate (SDS) was added as a surfactant to the mixture. To obtain a homogeneous colloid, the prepared mixture was placed in an ultrasonic bath for 30 min. Subsequently, the nickel substrate was immersed in the mixture. The coating was deposited by sedimentation on a metallic substrate until all the solvent had evaporated. The obtained coating was subjected to heat treatment at 800 °C for 1 h in an argon atmosphere. In this paper, coating obtained in that way will be referred to as CFO_PEG. In the second case, to obtain CoFe₂O₄ coating with a metallic binder, electrochemical deposition was used. Nickel was deposited from the bath with the following composition (concentrations in gdm⁻³): NiSO₄·7H₂O—84, NiCl₂·6H₂O—10, H₃BO₃—8, to which 20 mg of cobalt ferrite powder and 5 mg of SDS was added. Note that nickel electrodeposition started when CoFe₂O₄ powder settled by sedimentation on the surface of the metallic substrate. Nickel binder was deposited under galvanostatic conditions at current density j = 10 mAcm⁻² and time 20 min. The deposition process was carried out at the temperature 20 °C. The Faradaic yield of nickel deposition was ca. 70%. In this paper, coatings obtained in that way will be referred to as CFO_Ni (in the case of thermally treated CoFe₂O₄ powder) and ${\mathrm{CFO}}_{\mathrm{Ni}}^{\mathrm{a}\text{-}\mathrm{r}}$ (in the case of as-received CoFe₂O₄ powder). Using a JEOL JSM—6480 (JEOL Ltd., Tokyo, Japan) scanning electron microscope (SEM) with energy dispersive spectroscopy (EDS) attachment, the surface morphology and chemical composition of the obtained CoFe₂O₄ coatings were determined.

3.3. Electrochemical Measurements

Electrochemical tests were performed in 1 M KOH solution using a three-electrode cell with saturated calomel electrode (SCE) as a reference electrode and platinum mesh as a counter electrode. Working electrodes were CoFe₂O₄ coatings with a geometric surface area of 0.5 cm². The ohmic drop between the reference and working electrode was reduced using a Luggin capillary. All electrochemical experiments were carried out at a temperature of 20 °C. For the data registration, a PARSTAT 2273 system and the PowerSuite 2.58 software (Princeton Applied Research, Oak Ridge, TN, USA) were used. Values of measured potentials were converted from SCE to reversible hydrogen electrode (RHE) according to the following formula: E_RHE = E_SCE + 0.059pH + 0.241. The overpotential (η) for the oxygen evolution reaction was calculated using the equation: η = E_RHE − 1.23.

Quasi-stationary polarisation curves j vs. E were recorded using the linear sweep voltammetry (LSV) technique within the potential range from 1.22 V to 1.85 V vs. RHE and sweep rate v = 10 mVmin⁻¹. Before measurements, CoFe₂O₄ electrodes were conditioned at an anodic potential of 1.85 V for 5 h. The ohmic drop compensation was conducted during measurements using the current interrupt technique [4].

The spectra were registered using the electrochemical impedance spectroscopy (EIS) method. The impedance spectra were recorded potentiostatically at selected dc potentials from the interval where the OER takes place. Before the recording of each spectrum, electrodes were held at the appropriate potential for 5 min. Spectra were registered in the frequency range from 20 kHz to 10 mHz with a density of 10 points per decade. The amplitude of the ac signal was 10 mV_rms. For the quantitative analysis of obtained data, the ZSimpWin 3.21 software was used.

To evaluate long-term stability of the CoFe₂O₄ electrodes, the chronopotentiometry technique was used. The test was carried out at the current density j = 50 mAcm⁻² for 28 h.

4. Concluding Remarks

This paper reports on two new procedures for the preparation of porous cobalt ferrite coatings with a metallic and organic binder for use as catalysts in the oxygen evolution reaction (OER). The parameter η₁₀ indicates that all investigated CoFe₂O₄ coatings exhibit significantly higher apparent activity towards OER than nickel substrate in 1 M KOH. In particular, CoFe₂O₄ coating with PEG applied as a binder shows overpotential η₁₀ = 295 mV at a current density of 10 mAcm⁻² and Tafel slope b = 35 mVdec⁻¹. CoFe₂O₄ coating with Ni binder (η₁₀ = 287 mV and b = 45 mVdec⁻¹) shows comparable catalytic activity towards OER. It was stated that the main reason for the superior catalytic activity of CFO_PEG coating is the large electrochemically active surface area. In contrast, the excellent catalytic activity of CoFe₂O₄ coating containing nickel binder between the ferrite particles is caused by increasing both the conductivity and the electrochemically active surface area. It was also shown that the obtained CoFe₂O₄ electrodes maintain their catalytic activity for at least 28 h at a current density of 50 mAcm⁻². It can be stated that this work offers a new path for the design of high-performance OER electrocatalysts using both sintering and electrodeposition techniques.