Bifunctional Electrocatalyst of Low-Symmetry Mesoporous Titanium Dioxide Modified with Cobalt Oxide for Oxygen Evolution and Reduction Reactions

Abstract

Hybrids of low-symmetry (disordered) mesoporous titanium dioxide modified with different weight ratios of cobalt oxide nanoparticles (Co₃O₄(x)/lsm-TiO₂) are prepared using a one-pot self-assembly surfactant template. The physicochemical characterization of Co₃O₄(x)/lsm-TiO₂ hybrids by scanning and transmission electron microscopy, X-ray diffraction, N₂ adsorption–desorption isotherms, and X-ray photoelectron spectroscopy confirm the successful incorporation of cobalt oxide nanoparticles (2–3 nm in diameter) with preservation of the highly mesoporous structure of titanium dioxide substrate. Among these mesoporous hybrids, the ~3.0 wt.% Co₃O₄/lsm-TiO₂ exhibits the best performance toward both the oxygen evolution (OER) and reduction (ORR) reactions in alkaline solution. For the OER, the hybrid shows oxidation overpotential of 348 mV at 10 mA cm⁻², a turnover frequency (TOF) of 0.034 s⁻¹, a Tafel slope of 54 mV dec⁻¹, and mass activity of 42.0 A g⁻¹ at 370 mV. While for ORR, an onset potential of 0.84 V vs. RHE and OER/ORR overpotential gap (ΔE) of 0.92 V are achieved which is significantly lower than that of commercial Pt/C, hexagonal mesoporous, and bulk titanium dioxide analogous. The Co₃O₄/lsm-TiO₂ hybrid demonstrates significantly higher long-term durability than IrO₂. Apparently, such catalytic activity performance originates from the synergetic effect between Co₃O₄ and TiO₂ substrate, in addition to higher charge carrier density and the presence of disordered mesopores which provide short ions diffusion path during the electrocatalytic process.

1. Introduction

The fabrication of effective, highly stable and multi-functional electrocatalysts for oxygen evolution (OER) and reduction (ORR) reactions is in high demand for the commercialization of rechargeable metal–air and Li-ion batteries [1,2,3], fuel cells [4], and water splitting [5] applications. In particular, the development of highly efficient, stable and dual-functional electrocatalysts for the oxygen evolution (4OH⁻→O₂ + 2H₂O + 4e⁻) and reduction reactions (O₂ + 2H₂O + 4e⁻ → 4OH⁻) has attracted the interest of scientific researchers, as well as industrial R&D centers. The OER and ORR processes involve two-electron or four-electron transfer reactions that are hampered by significant overpotential, sluggish kinetics, and instability [6,7,8]. Thus, the development of highly effective, inexpensive, and bifunctional (active for both the ORR and the OER) electrocatalysts is important for the development of renewable energy technologies. It is well documented that the state-of-the-art catalysts for the OER are Ir- and Ru-based materials, while the Pt-group materials are the most active electrocatalysts for the ORR [9,10]. However, their high cost, scarcity, weak bifunctionality, and poor stability hinder the applicability of these precious metals in large-scale applications [11]. Consequently, recent research activities have been focused on developing effective bifunctional catalysts with greater performance and cost-effectiveness, and superior durability, for both the ORR and the OER, for widespread applications. Oxides and hydroxides based on transition metals, such as nickel, cobalt, iron, molybdenum, titanium, and tungsten, have emerged as a promising class of noble metal-free material catalysts [12,13,14]. These earth-abundant materials on various supports have been identified as high activity and stability electrocatalysts for the OER and the ORR under neutral or strongly alkaline conditions [15,16,17,18,19,20,21]. For example, nitrogen-doped graphene [22] and carbon nanotubes [23] modified with cobalt oxide nanoparticles hybrids have shown enhanced activity for oxygen reduction and water oxidation reactions. It is believed that the formation of interfacial Co–O–C and Co–N–C bonds and the accompanied synergistic effect are probably responsible for the ORR and OER improved activity [22,23]. Nevertheless, the ORR/OER behavior of the transition metal oxides still requires further improvement to make them highly economical and stable, in comparison to noble metal catalysts, and suitable for commercial applications.

Increasingly more mesoporous materials are being reported as OER/ORR bifunctional catalysts due to their exceptional structural characteristics of narrow pore size distribution, high specific surface area, and well-organized arrangement of mesoporous architectures, which meet the requirements of energy-related conversion, catalysis, adsorption, and sensing applications [24,25,26,27,28,29,30]. Furthermore, mesoporous electrocatalysts possess a high population of active surface sites, and the mesoporous network assists the mass transport of reactants and products, in addition to improving the catalytic performance [31,32]. Among these materials, mesoporous TiO₂ catalysts have great promise for numerous applications in electrocatalysis and energy-related processes due to good chemical and physical properties, non-toxicity, good biocompatibility, and excellent photoelectric performance [33,34,35,36]. Within this context, a considerable number of studies have focused on improving the active sites of the TiO₂ structure for the OER. Several researchers have also reported the modification of TiO₂-based materials with various cations (such as Fe, Co, Ni, Mn, Bi, etc.) to improve their electronic, chemical, optical, and magnetic properties for a better photo- and electrocatalytic performance in water splitting and energy storage applications [37,38,39,40,41,42,43]. In recent relevant work by Hoffmann et al., they reported the cobalt-doped black TiO₂ nanotube hybrid that revealed long-term durability for water decontamination and OER applications [44]. However, many of these hybrids still have a shortfall of high overpotential values in comparison with the benchmark catalysts, such as IrO₂ and Pt/C-based catalysts [36,38,40]. To this end, our group investigated the preparation of low-symmetry (disorder) mesoporous titanium dioxide (lsm-TiO₂) with high surface area and a semi-crystalline anatase wall by using a self-assembly surfactant template followed by a two-step annealing process. It was shown that the degree of porosity, order, and surface area of lsm-TiO₂ catalyst can be controlled by tuning the titanium precursor/surfactant ratio. Moreover, the addition of acetylacetone chelating agent during the self-assembly process controlled the hydrolysis and condensation of TiO₂ precursor and produced the desired phase of low-symmetry and hexagonal TiO₂ (hm-TiO₂) substrate [45,46,47]. In comparison to the highly hexagonal mesoporous structure, the lsm-TiO₂ thin films have shown a significant enhancement of the OER activity in alkaline water because of the presence of short-range order, bimodal hierarchical pores, and transverse worm-like channels, which provided fast ions diffusion and interfacial electrons transport during the electrocatalytic process [36]. Furthermore, research on TiO₂-based nanomaterials as bifunctional OER/ORR electrocatalysts has rarely been reported, and the synthesis of low-cost TiO₂ hybrid bifunctional electrocatalysts is desired. This work reports a one-pot surfactant template approach to fabricate low-symmetry (disordered) mesoporous TiO₂ substrates that modified with cobalt oxide nanoparticles to form Co₃O₄(x)/lsm-TiO₂ hybrids. The physicochemical properties of surface morphology, crystal structure mesoporosity, and specific surface area of the produced catalysts and their corresponding electrocatalytic activities for the oxygen evolution and reduction reactions were examined. The Co₃O₄(x)/lsmTiO₂ mesoporous catalyst exhibits enhanced performance as a bifunctional electrocatalyst toward the OER and the ORR in alkaline solution in comparison to the cobalt-doped hexagonal mesoporous TiO₂ (Co₃O₄/hm-TiO₂), Co₃O₄/bulk-TiO₂, and also the commercially available Pt/C and IrO₂.

2. Results and Discussion

2.1. Morphological and Physicochemical Characterization of the Cobalt Oxide-Modified Mesoporous TiO₂

Figure 1 shows the field emission scanning and transmission electron microscope (FESEM and TEM) images of the low-symmetry mesoporous TiO₂ modified with cobalt oxide (Co₃O₄(3)/lsm-TiO₂ and Co₃O₄(3)/hm-TiO₂). As shown in Figure 1a, the Co₃O₄(3)/lsm-TiO₂ catalyst exhibits a low-symmetry mesoporous structure with worm-like channels running throughout, and the pores tend to be well packed at short range. The FESEM image in Figure 1b shows the typical surface morphology of mesoporous Co₃O₄(3)/hm-TiO₂ with uniform and highly ordered two-dimensional (2D) hexagonal channels extended over a large domain. Figure 1c,d shows the representative TEM image of Co₃O₄(3)/lsm-TiO₂ and Co₃O₄(3)/hm-TiO₂ catalysts, respectively. The TEM image in Figure 1c clearly demonstrates that the low-symmetry mesoporous structure of Co₃O₄(3)/lsm-TiO₂ contains large (primary) mesopores, connected by smaller pores (hierarchical porous structure) with a disordered pore arrangement. However, it is difficult to distinguish the cobalt nanoparticles from the TiO₂ substrate. On the other hand, the TEM image in Figure 1d clearly shows the highly ordered cylindrical mesoporous channels of TiO₂, with a pore size and a wall thickness of about 2.6 and 2.1 nm, respectively. The cobalt oxide nanoparticles can be distinguished as dark spots that are uniformly distributed throughout the mesoporous hexagonal TiO₂ substrate, as shown by the red circle in Figure 1d. Figure 1f shows the high-resolution TEM image of Co₃O₄(3)/hm-TiO₂ catalyst, which demonstrates that Co₃O₄(3)/hm-TiO₂ has a crystalline nature with a lattice fringe spacing estimated at 0.345 nm index for TiO₂ (101) diffraction plane. Moreover, the crystal lattice spacing of the cobalt oxide nanoparticles is estimated at 0.243 and 0.288 nm, corresponding to the Co₃O₄ diffraction planes of (311) and (220), respectively. Upon increasing the cobalt content to 7.0 wt.%, the nanoparticles aggregated to form larger assemblies, and a disordered TiO₂ substrate was obtained, as shown in the TEM images of Co₃O₄(7)/lsm-TiO₂ catalyst in Figure S1 (Supplementary Materials).

The crystalline phases of lsm-TiO₂ and Co₃O₄(x)/lsm-TiO₂ catalysts were examined using X-ray diffraction (XRD), as shown in Figure 2. Furthermore, the XRD curve of Co₃O₄(3)/lsm-TiO₂ catalyst was compared with those of Co₃O₄(3)/bulk-TiO₂ and Co₃O₄(3)/hm-TiO₂ catalysts, as shown in Figure S2. As presented in Figure 2, the XRD patterns clearly show a series of diffraction peaks of a typical pure TiO₂ anatase phase (JCPDS card No. 01-070-6826). Clearly, no characteristic diffraction peaks of CoO or Co₃O₄ are detected when the cobalt content is lower than 7.0 wt.%. From the XRD patterns, it can be concluded that the incorporation of cobalt oxide seems not affected the anatase phase of TiO₂ substrate, which is consistent with the TEM results. However, the existence of cobalt oxide diffraction peaks are observed for Co₃O₄(7)/lsm-TiO₂ and Co₃O₄(3)/bulk-TiO₂, as shown in Figure 2 and Figure S2 (Supplementary Materials). The cobalt oxide diffraction peaks are centered at 2θ ~36.90 and 38.8, which can be assigned to the (311) and (222) diffraction planes, respectively, of the cubic Co₃O₄ phase (JCPDS Card 00-042-1467).

Figure S3 shows the effect of varying cobalt content on the average crystallite size of the lsm-TiO₂ catalysts, as assessed from the half-height width of the main diffraction peak (2θ = 25.5°) using Scherrer’s equation. The curve in Figure S3 shows that the TiO₂ crystal size increases significantly when the cobalt content exceeds 1.0 wt.% and, in general, the crystallite size of the TiO₂ substrate is in the range between 5.93 and 6.8 nm for Co₃O₄(x)/lsm-TiO₂, and 13.6 nm for Co₃O₄(3)/bulk-TiO₂. Our results demonstrate the key role of P123 surfactant, which acts as a capping agent around the nanoparticles, thereby stabilizing the porosity of the anatase phase of TiO₂ during the sol–gel process in solution and the annealing process at higher temperature [35].

The X-Ray Photoelectron Spectroscopy (XPS) analysis was carried out to define the surface bonding of mesoporous TiO₂ induced by cobalt dopant. Figure 3 displays the XPS survey spectra of pristine lsm-TiO₂ and Co₃O₄(3)/lsm-TiO₂ catalysts. Both spectra are very similar, except for the presence of the Co 2p_3/2 peak at 780.4 eV. In addition to the Ti, O, and Co 2p_3/2 peaks for TiO₂ and Co₃O₄(3)/lsm-TiO₂, carbon C 1s signals are also observed in both samples. These could originate from the burnt surfactant or from carbon contamination during sample preparation and successive handling. Figure 3b shows the high-resolution XPS spectra of Ti 2p of the pristine lsm-TiO₂ and Co₃O₄(3)/lsm-TiO₂ catalysts. The Ti 2p_3/2 and Ti 2p_1/2 spectrum peaks of pristine lsm-TiO₂ are identical to their binding energy (BE) at 459.55 and 465.29 eV, respectively, implying the Ti⁴⁺ state [48]. The spin-orbit splitting energy of those two peaks is 5.74 eV, which is identical to that reported in the literature [49].

In the presence of cobalt dopant, the Ti 2P_3/2 peaks of lsm-TiO₂ and Co₃O₄(3)/lsm-TiO₂ catalysts are almost identical to that of the pure one, except for a slight red-shift in the BE of 0.15 eV, which is consistent with previous reports [50,51,52]. Figure 3c displays the XPS spectra of the O 1s of Co₃O₄(3)/lsm-TiO₂, and three Gaussian peaks assigned as (1), (2), and (3) were used to fit the experimental data. The peak positioned at 529.83 eV belongs to the Ti–O bond in TiO₂. The peak located at 531.6 eV can be ascribed to the Ti–OH bond, which was reported to be positioned at a BE ∼1.5–1.8 eV higher than the O 1s of TiO₂ [53,54]. These results prove that the introduction of Co dopants in Co₃O₄(3)/lsm-TiO₂ creates an enriched quantity of surface O vacancies in the form of metal–OH bonds. The chemical structure and oxidation states of Co dopants in lsm-TiO₂ hybrids were well demonstrated in the XPS analysis (Figure 3d). The Co 2p spectra exhibit two main peaks with binding energies of 781.3 eV (Co 2p_3/2) and 796.9 eV (Co 2p_1/2), accompanied by two strong shakeup satellite peaks toward higher binding energies posited at nearly 6 eV from the main peaks, which suggests Co₃O₄ phase with two valence states (octahedral Co³⁺ and tetrahedral Co²⁺) coexist in Co/lsm-TiO₂ [55]. After deconvolution of Co 2p spectra, the energy difference between the Co 2p_3/2 and the Co 2p_1/2 peak had a separation of ∼15.1 eV, which further indicates that the oxidation state of cobalt in the lsm-TiO₂ lattice is mixed-valence Co₃O₄. The elemental composition analysis of XPS spectra in Figure 3a indicates a 2.92 wt.% cobalt content for the Co₃O₄(3)/lsm-TiO₂ catalyst, as shown in Table S1. In addition, Figures S4 and S5 show the Scanning electron microscope- Energy dispersive X-ray SEM–EDX elements mapping of Co₃O₄(3)/lsm-TiO₂ and Co₃O₄(3)/hm-TiO₂ catalysts, respectively. Clearly, the images of EDX elements mapping demonstrate the uniform distribution of Co₃O₄ within both the low-symmetry and hexagonal TiO₂ substrate. The EDX analysis reveals that the cobalt content is very close for both low-symmetry and hexagonal TiO₂ substrate and reaches 3.7% and 3.83%, respectively.

The mesoporosity of the obtained lsm-TiO₂, Co₃O₄(x)/lsm-TiO₂, and Co₃O₄(3)/hm-TiO₂ samples are characterized by the N₂-physisorption technique, and the results are shown in Figure 4. The N₂ sorption isotherms of the produced catalysts in Figure 4a exhibit characteristic type IV curves and H1 hysteresis loops with capillary condensation steps, which are distinctive of mesoporous materials according to the IUPAC classification [56,57]. Figure 4a demonstrates distinct capillary condensation at P/P₀ = 0.4–0.7, indicating the uniform and narrow mesopore size distribution, which is consistent with the TEM results. The pore size distribution of the lsm-TiO₂ and Co₃O₄(x)/lsm-TiO₂ samples were assessed from the adsorption data via the BJH method, and is shown in Figure 4b.

The pore size analysis shows that as the cobalt content increases, the pore size distribution becomes wider in the case of Co₃O₄(7)/lsm-TiO₂. This is probably due to the structural damage from the growth of cobalt oxide nanoparticles. The specific texture parameters of the samples are presented in

Table 1. Textural properties of the lsm-TiO₂ and Co₃O₄(x)/lsm-TiO₂ catalyst derived from nitrogen adsorption and desorption data.

Catalyst	BET Surface Area (m2 g−1) a	Pore Size (nm) b	Pore Volume (cm3 g−1) c
lsm-TiO2	200	2.7	0.290
Co3O4(0.5)/lsm-TiO2	175	2.01	0.25
Co3O4(1)/lsm-TiO2	158	2.03	0.17
Co3O4(3)/lsm-TiO2	146	2.11	0.11
Co3O4(5)/lsm-TiO2	119	1.95	0.10
Co3O4(7)/lsm-TiO2	81	1.83	0.071
Co3O4(3)/hm-TiO2	229	2.11	0.22
Co3O4(3)/bulk-TiO2	42	1.50	0.025

^a BET is evaluated by the BET equation; ^b Pore size is determined by the BJH method; ^c Total pore volume is calculated at p/p₀ = 0.950–0.995.

. We can observe that the specific surface area (SSA) and the pore volumes of low symmetry mesoporous TiO₂ substrate tend to decline as the cobalt content increases due to the deteriorating mesoporous structure, which is consistent with the above TEM and XRD results.

2.2. Electrochemical Performance of the Mesoporous TiO₂ Modified with Cobalt Oxide Nanoparticles for OER

Figure 5a shows the linear sweep voltammetry (LSV) curve of Co₃O₄(x)/lsm-TiO₂ catalyst at 10 mV s⁻¹ in 1.0 M of KOH solution in a conventional 3-electrode cell. For comparison purposes, Figure 5b shows the LSV curve of lsm-TiO₂, Co₃O₄(3)/hm-TiO₂, Co₃O₄(3)/bulk-TiO₂, and commercial IrO₂ catalysts recorded under similar conditions. Based on the LSV results, the TiO₂-base catalysts exhibit a certain degree of OER electrochemical activity in alkaline solution. The bare CP does not show any catalytic activity for the OER, while the pure lsm-TiO₂ catalyst exhibits an OER overpotential of 490 mV at a current density of 10 mA cm⁻². With the incorporation of cobalt oxide nanoparticles, the OER activity is significantly improved, as evidenced by the onset potential of 1.48 V vs. RHE and the overpotential (η) of 348 mV at 10 mA cm⁻² in the case of Co₃O₄(3)/lsm-TiO₂. This overpotential is 142 mV lower than that of the corresponding pure lsm-TiO₂. The mesoporous Co₃O₄(3)/lsm-TiO₂ catalyst shows superior performance in comparison to Co₃O₄(3)/hm-TiO₂ (452 mV) and Co₃O₄(3)/bulk-TiO₂ (470 mV) that are prepared by the same procedure, and even with respect to commercial IrO₂ catalyst (354 mV) at the same current density of 10 mA cm⁻², as shown in

Table 2. The electrochemical parameters of the mesoporous TiO₂ modified with cobalt oxide nanoparticles, as obtained from the LSV in Figure 5.

Catalysts	Onset Potential vs. RHE/V a	η (mV) at (10 mA/cm2) b	η (mV) at (20 mA/cm2) c	Current mA/cm2 at 2.0 V vs. RHE	Tafel Slope (mV/dec) d	mass Activity at 1.6 V (A g−1 ) e	TOF at η = 0.370 V (s−1 ) f
lsm-TiO2	1.620	490	530	94	98	2	0.00033
Co3O4(0.5)/lsm-TiO2	1.510	449	492	113	92	5	0.0038
Co3O4(1)/lsm-TiO2	1.490	390	432	149	65	15	0.012
Co3O4(3)/lsm-TiO2	1.485	348	370	222	54	42	0.035
Co3O4(5)/lsm-TiO2	1.490	370	410	170	84	23	0.020
Co3O4(7)/lsm-TiO2	1.530	437	481	131	87	5	0.0037
Co3O4(3)/hm-TiO2	1.58	452	490	141	121	3	0.0028
Co3O4(3)/bulk-TiO2	1.590	470	510	114	91	3	0.0020
IrO2	1.475	354	398	164	81	34	0.019

^a Onset overpotential; ^b Overpotential at j = 10, 20 mA/cm²; ^d Tafel slope estimated from the Tafel equation: η = b log (j/j0); ^e The values of mass activity; ^d Turnover frequency (TOF), see Supporting Material for the calculation method.

and Figure 5b. The OER kinetics was further investigated using a Tafel plot by fitting the LSV data to the equation: η = b log j + a, where j and b are the current density and Tafel slope, respectively [51]. As shown in Figure 5c and Table 2, the Co₃O₄(3)/lsm-TiO₂ catalyst has the smallest Tafel slope (54 mV dec⁻¹), making it the most efficient one among the investigated and reference catalysts. For further evaluation of the electrocatalytic behavior of the cobalt oxide-modified mesoporous TiO₂ hybrids for the OER, the mass activity and turnover frequency (TOF) were calculated at an overpotential of η = 0.370 V (1.6 V vs. RHE), and are reported in Table 2. The TOF and mass activity of our electrocatalysts were calculated according to the method reported by Gao et al. [58], shown in the Supplementary Materials. The mass activity of Co₃O₄(3)/lsm-TiO₂ at a potential 1.6 V vs. RHE reaches 42 A g⁻¹, and the TOF value equals 0.035 s⁻¹ at η = 370 mV. Clearly, the mass activity and TOF values of Co₃O₄(3)/lsm-TiO₂ are superior to those of the other investigated electrocatalysts and comparable to those recently reported for the state-of-the-art catalysts [59,60,61,62]. From the data in Table 2, the TOF value of Co₃O₄(3)/lsm-TiO₂ catalyst is about ~12.5, ~17.5, and ~1.85 orders of magnitude higher than that of the Co₃O₄(3)/hm-TiO₂ (0.0028), Co₃O₄(3)/bulk-TiO₂ (0.002), and IrO₂ (0.019) catalysts, respectively. The superior OER activity of Co₃O₄(3)/lsm-TiO₂ could be attributed to the synergistic effect of Co₃O₄(3)/TiO₂, as well as that the low-symmetry TiO₂ substrate provide a short diffusion path for ions mass transport, which is better than the hexagonal ordered continuous channels.

Figure 5d shows the LSV curve of the Co₃O₄(3)/lsm-TiO₂ electrode tested under various KOH concentrations (0.1, 1.0 M) and extremely alkaline conditions (such as concentrated 5.0 M of KOH solution), which are often encountered in commercial alkaline water electrolysis [63]. As shown in Figure 5d and

Table 3. Overpotential values calculated at 10 and 20 mA/cm², and current density at 1.0 V vs. SCE, of Co₃O₄(3)/lsm-TiO₂ obtained in various KOH concentrations.

Electrolyte (KOH)	η (mV) at 10 mA/cm2	η (mV) at 20 mA/cm2	I (mA/cm2) at 1.0 V	Tafel Slope mV/dec
0.1 M	445	512	61	87
1.0 M	309	337	222	54
5.0 M	243	272	318	71

, the OER performances of the Co₃O₄(3)/lsm-TiO₂ catalyst in 0.1, 1.0, and 5.0 M KOH display an onset potential of 317 , 227, and 160 mV vs. SCE, respectively, and a current density of 20 mA cm⁻² was obtained in 5.0 M KOH at an overpotential of 272 mV, which is 240 and 65 mV lower than that acquired in 0.1 and 1 M KOH, respectively.

The long-term stability of Co₃O₄(3)/lsm-TiO₂ catalyst and the ability to continuously catalyze the OER were examined using continuous cycling and chronoamperometry (CA) in 1.0 M KOH. Figure 6a displays the LSV curves of a Co₃O₄(3)/lsm-TiO₂ electrode before and after 2500 cycles in 1.0 M KOH at 10 mV s⁻¹. The catalyst clearly exhibits a similar LSV after extended cycling with negligible loss of the OER anodic current, confirming satisfactory durability in alkaline electrolytes. To further assess the superior durability of the Co₃O₄(3)/lsm-TiO₂ catalyst, Figure 6b shows the CA response of Co₃O₄(3)/lsm-TiO₂ and the benchmark IrO₂ catalysts measured for 9 h at 1.65 V vs. RHE in 1.0 M KOH. Figure 6b demonstrates that the benchmark IrO₂ electrocatalyst shows an anode current loss of 90% after continuous operation for 9 h in 1.0 M KOH solution. In contrast, the Co₃O₄(3)/lsm-TiO₂ catalyst exhibits an oxygen evolution current over 13 times higher than that of IrO₂, and offers a stable current with the negligible loss (8%) after 9 h of electrolysis at the same potential, in agreement with the results of the incessant potential CV sweeps. These results suggest that Co₃O₄(3)/lsm-TiO₂ is functioned as an effective electrocatalyst to drive water oxidation with great durability. The performance of Co₃O₄(3)/lsm-TiO₂ is not only higher than those of IrO₂, Co₃O₄(3)/hm-TiO₂, and Co₃O₄(3)/bulk-TiO₂, but also higher than those of other reported state-of-the-art cobalt-based electrocatalysts [58,59,60]. A comprehensive assessment of alkaline OER performance and catalyst parameters is delivered in Table S2.

Further the electrochemical impedance spectroscopy EIS analysis measurements were also performed to get further insight into the kinetics of electrode reactions before and after doping with cobalt oxide nanoparticles. Figure 7a presents typical Nyquist plots of the impedance data obtained for the pure lsm-TiO₂, Co₃O₄(3)/lsm-TiO₂, Co₃O₄(3)/hm-TiO₂, and Co₃O₄(3)/bulk-TiO₂ electrodes at 1.6 V vs. RHE. The low-frequency area of the Nyquist plot (Z’ vs. −Z”) and the equivalent circuit, presented in the inset, matches the charge transfer resistance (R₂) of the catalytic materials. It is clearly observed that the radii of the arc on the EIS Nyquist plots of Co₃O₄(3)/lsm-TiO₂ are lower than those of pure lsm-TiO₂, suggesting that the cobalt-doped samples possess smaller R_ct and better electrochemical OER performance. Furthermore, the result indicates that Co₃O₄(3)/lsm-TiO₂ can accelerate charge transfer kinetics and acts as a highly effective water oxidation electrocatalyst. The R₂ values for Co₃O₄(3)/lsm-TiO₂, Co₃O₄(3)/hm-TiO₂, Co₃O₄(3)/bulk-TiO₂, and lsm-TiO₂ were found to be 7.10, 184,723, and 304 Ω, respectively, as shown in

Table 4. The EIS parameters of pure lsm-TiO₂, Co₃O₄(3)/lsm-TiO₂, Co₂O₃(3)/hm-TiO₂, and Co₂O₃(3)/bulk-TiO₂ catalysts obtained through fitting EIS spectra measured at η = 500 mV to an equivalent circuit.

Catalyst	R1 (Ω)	Q2 (µF. sn−1)	R2 (Ω)	Q3 (µF. sn−1)	R3 (Ω)
lsm-TiO2	2.152	100	304	109	448
Co3O4(3)/lsm-TiO2	2.260	10,140	7.10	9760	0.480
Co3O4(3)/hm-TiO2	2.720	7170	184	3230	85
Co3O4(3)/bulk-TiO2	2.253	3870	723	190	110.4

. The Co₃O₄(3)/lsm-TiO₂ catalyst has the lowest R₂ value, implying lower electron and charge transfer resistances and, thus, faster electrode kinetics, in agreement with its highest intrinsic activity toward the OER. Figure 7b shows the Nyquist plots of the Co₃O₄(3)/lsm-TiO₂ catalyst at different overpotentials, with the corresponding impedance parameters reported in

Table 5. The EIS extracted parameters from fitting EIS plots measured at different overpotential (η) of Co₃O₄(3)/lsm-TiO₂ catalyst.

Overpotential (η)/mV	R1 (Ω)	Q2 (µF. sn−1)	R2 (Ω)	Q3 (µF. sn−1)	R3 (Ω)
220	2.26	9370	299	104,000	0.95
250	2.38	8890	53	4260	1.40
280	2.26	10,140	7.1	9760	0.48
300	2.41	12,650	3.90	--	3.4
320	2.25	75,750	2.52	9806	2.52
350	2.27	33,120	0.59	9840	1.69

. It is clear from Figure 7b that the semicircle diameter and the R₂ values are considerably decreased as the overpotential shifts from 220 to 350 mV vs. RHE, indicating that the OER process is considerably improved at higher potentials, which corresponds to earlier polarization curves.

Although the Mott−Schottky (M–S) theory applies to the flat (non-porous) electrode, the M–S plot can be used to compare the apparent charge carrier density of the mesoporous catalysts [64]. Mott−Schottky (M–S) analysis was conducted on the pure lsm-TiO₂, Co₃O₄(3)/lsm-TiO₂, and Co₃O₄(5)/lsm-TiO₂ catalysts. M–S plots were obtained in 1.0 M KOH at 500 Hz and are shown in Figure 7c. The pure lsm-TiO₂, Co₃O₄(3)/lsm-TiO₂, and Co₃O₄(5)/lsm-TiO₂ catalysts clearly exhibit a positive slope in the M–S curves, confirming n-type semiconductor character. Importantly, the Co₃O₄(3)/lsm-TiO₂ and Co₃O₄(5)/lsm-TiO₂ samples show substantially smaller slopes compared to pure lsm-TiO₂, suggesting a significant increase in charge carrier density.

The catalysts carrier densities were estimated using the Mott−Schottky equation shown below:

$$N_d=\left(\frac{2}{e_0{\mathsf{\epsilon }\mathsf{\epsilon }}_0} \right) {[ d \left(\frac{\frac{1}{C^2}}{dV}\right)]}^{-1}$$

(1)

where ε₀ is the permittivity of vacuum, e₀ is the electron charge, ε is the dielectric constant of TiO₂ (31 for anatase) [61], V is the applied electrode potential, and N_d is the donor density. The Mott−Schottky plot with the corresponding charge carrier density value for all Co₃O₄ modified lsm-TiO₂ electrodes is shown in Figure S6 (Supplementary Materials). The charge carrier densities of the pure lsm-TiO₂, Co₃O₄(3)/lsm-TiO_2, Co₃O₄(5)/lsm-TiO₂, and Co₃O₄(7)/lsm-TiO₂ were estimated to be 4.9 × 10²¹ cm⁻³, 3.9 × 10²² cm⁻³, 3.45 × 10²³ cm⁻³, and 2.24 × 10²² cm⁻³, respectively. The increment in the carrier densities of Co₃O₄(3)/lsm-TiO₂ and Co₃O₄(5)/lsm-TiO₂ could be attributed to the increased oxygen vacancy states, which acted as extra electron donors for TiO₂ substrate [65,66].

2.3. Electrochemical Performance of the Mesoporous TiO₂ Modified with Cobalt Oxide Nanoparticles for ORR

The electrocatalytic ORR activity of lsm-TiO₂ and Co₃O₄(3)/lsm-TiO₂ was studied in an N₂- and O₂-saturated 1.0 M KOH solution using glassy carbon electrode. As displayed in Figure 8a, the curves show a distinctive ORR peak for lsm-TiO₂ and Co₃O₄(3)/lsm-TiO₂ electrodes in the O₂-saturated electrolyte solution, which completely disappears in the N₂-saturated KOH solution. Moreover, the oxygen reduction onset potential is at 0.84 and 0.73 V vs. RHE, and a peak current of ~0.6 and ~0.2 mA cm⁻² was obtained for Co₃O₄(3)/lsm-and lsm-TiO₂, respectively. This proves that Co₃O₄(3)/lsm-TiO₂ significantly enhances the ORR in alkaline solution. Figure 8b displays the LSV plots obtained using a glassy carbon rotating disk electrode (RDE) loaded with lsm-TiO₂, Co₃O₄(3)/bulk-TiO₂, Co₃O₄(3)/hm-TiO₂, and Co₃O₄(3)/lsm-TiO₂ catalysts at a scan rate of 10 mV s^-1 and a rotation speed of 2000 rpm in 1.0 M KOH solution. The Co₃O₄(3)/lsm-TiO₂-based electrode achieved the best performance among the other investigated catalysts for the ORR in alkaline solution. As shown by the LSV curves of all investigated electrocatalysts (Figure 8b), the Co₃O₄(3)/lsm-TiO₂-based electrode exhibits a high current density and a characteristic ORR onset potential of about 0.84 V vs. RHE, which is respectively 182, 150, and 140 mV more positive than the overpotential of the lsm-TiO₂, Co₃O₄(3)/bulk-TiO₂, and Co₃O₄(3)/hm-TiO₂ electrodes, respectively, signifying a more facile ORR process on Co₃O₄(3)/lsm-TiO₂ hybrid.

Rotating disk electrode (RDE) analysis was also carried out to examine the ORR activity and kinetics of the Co₃O₄(3)/lsm-TiO₂ electrode in an O₂-saturated 1.0 M KOH solution. Figure S7 (Supplementary Materials) displays the LSV curves for Co₃O₄(3)/lsm-TiO₂ at rotation speeds ranging from 500 to 3000 rpm. Evidently, as the rotational speed rate increases, the current density gradually increases, due to the enhanced mass transport to the electrode surface. The number of electrons transferred (n) and participating in the ORR was assessed through the Koutecky–Levich (K–L) equation at different potentials and within the oxygen-limiting reduction current range, as shown in Figure 8c. From the line slope in Figure 8c, the average n value is about 3.6 at potentials varying from 0.25 to 0.5 V for the Co₃O₄(3)/lsm-TiO₂ catalyst. The calculated n value of Co₃O₄(3)/lsm-TiO₂ indicates a four-electron (4e⁻) transfer reaction, suggesting a direct pathway for electrocatalytic reduction of oxygen forming water, rather than from hydrogen peroxide. The results of the OER and ORR studies confirm that the Co₃O₄(3)/lsm-TiO₂ catalyst can be employed as a bifunctional oxygen electrocatalyst, and the electrocatalytic activity for both reactions can be assessed by the value of ΔE, which refers to the difference between the potential required to achieve 10 mA cm⁻² (E_J@10) for the OER and the half-wave potential (E_1/2) for the ORR [67,68,69,70,71]. Figure 8d shows the LSV of lsm-TiO₂, Co₃O₄(3)/lsm-TiO₂, Co₃O₄(3)/bulk-TiO₂, and for comparison, the commercial Pt/C catalysts loaded on carbon paper electrode in O₂-saturated 1.0 M KOH solution at a scan rate of 10 mVs⁻¹. Compared with Co₃O₄(3)/lsm-TiO₂ that uses P123 as surfactant template, the Co₃O₄(3)/bulk-TiO₂ catalyst prepared in the absence of P123 surfactant exhibits lower ORR and OER activity due to particle aggregation, which inhibits the diffusion of O₂ and induces the O₂ reduction and evolution efficiencies. The Co₃O₄(3)/lsm-TiO₂ exhibits the smallest ΔE value of 0.92 V (with E_J@10 = 1.55 V, E_1/2 = 0.65 V), and this value is noticeably lower than the ΔE obtained using commercial Pt/C catalyst (1.19 V with E_J@10 = 1.94 V, E_1/2 = 0.75 V), Co₃O₄(3)/hm-TiO₂ (0.99 with E_{J @10} = 1.69 V, E_1/2 = 0.70 V, not shown), Co₃O₄(3)/bulk-TiO₂ (1.14 with E_{J @10} = 1.69 V, E_1/2 = 0.55 V), and lsm-TiO₂ (1.18 V with E_{J @10} = 1.71 V, E_1/2 = 0.53 V), and comparable with that of the state-of-the-art cobalt-based catalysts [67,68,69,70,71]. Interestingly, the ORR current density obtained at Co₃O₄(3)/lsm-TiO₂ catalyst is higher than that at Pt/C catalyst. These results illustrate the importance of the low-symmetry mesoporous TiO₂ substrate and the role of cobalt oxide electrocatalyst dopants in boosting the ORR and OER activities for energy production and storage applications.

3. Materials and Methods

3.1. Materials

The triblock copolymer non-ionic surfactant of Pluronic^® P123 (Mw = 5800, EO20PO70EO20) and titanium(IV) butoxide (TBO, Ti(OBu)₄, 97%) were purchased from Sigma-Aldrich (Saint Louis, MO, USA). Cobalt acetate (Co(OAc)₂.4H₂O) was purchased from Alfa Aesar (Ward Hill, MA, USA). Hydrochloric acid (HCl, 37 wt.%), absolute ethanol (C₂H₆O, Analar), and acetylacetone (99%) were all purchased from Sigma-Aldrich. Carbon paper (CP, SIGRACET, GDL-24BC, SGL Technologies, Wiesbaden, Germany) was employed as the working electrode substrate. Potassium hydroxide pellets (pure) was purchased from BDH group. Deionized water with a resistivity of 18.2 MOhm cm was obtained from Ultrapure Water purification system of Milli-Q (Millipore, Inc., Darmstadt, Germany).

3.2. Synthesis of Cobalt Oxide/Mesoporous TiO₂ Materials

The crystalline mesoporous Co₃O₄(x)/lsm-TiO₂ and Co₃O₄/hm-TiO₂ catalysts (x = 0.5, 1.0, 3.0, 5.0 and 7.0 wt.% of the cobalt precursor) were prepared according to our previous reports [36] by using a chelation-assisted evaporation-induced self-assembly (EISA) approach in an ethanolic/Pluronic^® P123/HCl/acetylacetone (AcAc)/titanium tetrabutoxide (TBO) mixed solution combined with a simple template-carbonization strategy. The low-symmetry mesoporous TiO₂ modified with cobalt oxide nanoparticles was prepared by dissolving 0.50 g (0.086 mmol) of P123 surfactant in 10.0 g (217 mmol) of absolute ethanol. Then 7.5, 15.2, 46.5 79.0, and 113 mg of Co(OAc)₂ equivalent to 0.5, 1.0, 3.0, 5.0, and 7.0 wt.% was added to the above solution with continuous vigorous stirring to attain a clear homogenous solution, which is denoted as solution-A. Next, 1.5 g of TBO was added to in AcAc solution with a TBO:AcAc ratio equal to 2:3 wt.% and mixed for 20 min using magnetic stirring. Subsequently, the TBO/AcAc mixture was slowly added to P123 solution while constantly stirring for 30 min at 25 °C to obtain a homogeneous bright-yellow solution. In the meantime, concentrated HCl (1.5 g, 36 wt.%) was introduced dropwise, with continuous stirring, for a further 1.0 h. Then the obtained homogeneous solution was decanted into Petri dishes to evaporate the solvents, and then subjected to heating in an oven at 40 °C for 48 h for solidification and solvent evaporation. Later, the green transparent membrane was removed from the petri dish and subjected to pyrolysis at 350 °C for 3.0 h under N₂ atmosphere and a heating rate of 1 °C min⁻¹. Finally, the green mesoporous Co₃O₄(x)/lsm-TiO₂ product was obtained by further calcination of the grey powder in the open air at 400 °C for 4.0 h. The obtained cobalt oxide-modified low-symmetry mesoporous TiO₂ material was labeled as Co₃O₄(x)/Lsm-TiO₂, in which x represents the weight percentage of the cobalt precursor used in the synthesis process (0.5, 1.0, 3.0, 5.0, and 7.0 wt.%). In the absence of cobalt acetate addition, the sample was labeled as lsm-TiO₂. The hexagonal mesoporous TiO₂ modified by cobalt oxide (Co₃O₄(x)/hm-TiO₂) catalyst was prepared by following the same procedure of self-assembly and heat treatment, but using 2.0 g of Pluronic^® P123 (0.159 mmol) surfactant with the addition of 2.4 g of concentrated HCl (36%). The non-porous catalyst (Co₃O₄/bulk-TiO₂) was synthesized through a similar route, but without adding the Pluronic^® P123 surfactant, and used as a control catalyst.

3.3. Material Characterization

The crystal structure of the catalysts was executed using X-ray diffraction (XRD, Rigaku Mini Flex 600, Tokyo, Japan) with Cu K_α radiation (40 kV, 15 mA). The crystallite size of the materials was calculated by the Scherrer equation as D_XRD = 0.94λ/dcosθ, where λ is the X-ray wavelength, d is the diffraction peak width at half maximum, and θ is the diffraction angle. The catalysts surface morphology was acquired using scanning electron microscopy (Hitachi S4800, Japan) at 1.0 kV and 10 mA. JEOL 2100F microscope (Tokyo, Japan) operated at 200 kV was used to perform the catalysts’ fine structures. The samples were dispersed in ethanol solution and supported onto a carbon film on a Cu grid to obtain for TEM measurement. X-ray photoelectron spectroscopy (XPS) analysis was performed using an Escalab 250 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) with an X-ray source of monochromatic MgK_α. The N₂ adsorption–sorption isotherms at 77 K were measured using a NOVA 2200e surface area analyzer (Quantachrome Instruments, Florida, USA) Japan). Before the N₂ sorption analysis, the catalyst materials were degassed under vacuum at 180 °C for 6 h and the surface area was assessed by the Brunauer–Emmett–Teller (BET) method and the pore size distribution and pore volume curves were estimated by the Barrett–Joyner–Halenda (BJH) method.

3.4. Electrochemical Measurements

The catalysts’ electrochemical characterization was performed using a potentiostat (BioLogic SAS, model) in a 3-electrode assembly, with a Pt mesh (1 × 1 cm²) and saturated calomel electrode (SCE) as the counter and reference electrodes, respectively. The mesoporous catalysts were deposited on a commercial carbon paper substrate (CP, SIGRACET^®, grade GDL-24BC, geometric area 1 × 1 cm²) as the working electrode, while the potential was normalized to the reference hydrogen electrode (RHE) using the equation E_RHE = E_SCE + 0.244 V + 0.059 pH at 25 °C, where pH = 14 for 1.0 M KOH solution. The overpotential (η) was estimated as E (vs. RHE)−1.229 V. The mesoporous catalyst was deposited on the CP electrode using the electrophoretic deposition (EPD) method. In a typical procedure, 1.0 mL of iodine (40 mg in 15 mL acetone) was mixed with 15 mg of mesoporous catalyst powder and dispersed using an ultrasonic probe for 20 min to get a uniformly dispersed mixture of catalyst powder. The CP substrate (1 × 1 cm²) was placed in a small glass cell as the cathode and in parallel with the CP anode at a distance of ca. 1 cm. Then, a bias of +10 V was applied between them for 4 min using BioLogic SAS potentiostat to deposit the catalyst particles. Subsequently, the working electrode was rinsed with deionized water, dried in air, and calcined in a tube furnace under N₂ flow at 350 °C for 30 min. The average weight of the catalyst deposited on CP was 0.4 mg. The polarization curves and the OER and the ORR activity were reported without IR correction. Impedance spectroscopy analysis was executed in the frequency range of 10⁻² to 200 kHz with an AC voltage amplitude of 20 mV at a bias of 0.5 V vs. SCE in a 1.0 M KOH electrolyte. Mott–Schottky (M–S) plots were acquired in direct current potential polarization with a potential step of 10 mV at 500 Hz. The commercial Pt/C (10 wt.%) and IrO₂ catalysts were prepared and tested, under similar conditions for the sake of comparison. For ORR, linear sweep voltammetry (LSV) and cyclic voltammetry polarization measurements were performed in an O₂-saturated (if applicable) 1.0 M of KOH solution at 10 mV s⁻¹. The catalyst ink was made by dispersing the electrocatalyst (10 mg) in a mixture of water (0.5 mL), isopropanol (0.5 mL), and Nafion (10 μL of 5 wt.%) followed by sonication for 20 min. Afterwards, the catalyst ink (5 μL, equivalent to 50 μg) was casted on a glassy carbon (GC) electrode (diameter = 3 mm, area = 0.07 cm²). Rotating disk electrode (RDE) experiments were recorded using glassy carbon (3.0 mm diameter, METROHM, 628-10) at a rotation speed range of 500 to 3000 rpm at 10 mV s⁻¹. Furthermore, the total number of electrons taking part in the ORR was assessed via the Koutecky–Levich (K–L) equation shown below:

$$\frac{1}{J}= \frac{1}{J_L}+\frac{1}{J_K}=\frac{1}{B{\omega }^{1/2}}+\frac{1}{J_K}$$

(2)

$$B=0.62nF C_0 {\left(D_0\right)}^{2/3} {\nu }^{-1/6}$$

(3)

where J_K and J_L are the kinetic and limiting current densities, respectively; ω is the electrode rotating rate; B is determined from the slope of K-L curves according to the Levich formula (1); F is the Faraday constant, 96485 C mole⁻¹; n is the number of electrons transferred; C₀ is the bulk concentration of O₂, 1.2 × 10⁻⁶ mole cm⁻³; ν is the kinetic viscosity, 1.01 × 10⁻² cm² s⁻¹; and D₀ is the diffusion coefficient of O₂, 1.97 × 10⁻⁵ cm² s⁻¹.

4. Conclusions

In conclusion, we have demonstrated the synthesis of highly active bifunctional OER/ORR hybrids of cobalt oxide-doped low-symmetry mesoporous titanium oxide (Co₃O₄(x)/lsm-TiO₂) catalysts via the self-assembly surfactant template method. The effects of cobalt doping level on the morphology and electrocatalytic properties of the obtained mesoporous hybrids were investigated. The hybrids’ characterizations confirmed the formation of mesoporous titanium dioxide substrate modified with cobalt oxide nanoparticles with a diameter of 2–3 nm. The Co₃O₄(3)/lsm-TiO₂ hybrid with ~3 wt.% Co doping was found to be a highly active electrocatalyst for both the OER and the ORR in alkaline media. This catalyst exhibited excellent bifunctional OER/ORR catalytic performance, low onset potentials (1.480 V OER and 0.84 V ORR), a small OER/ORR overpotential gap (ΔE) of 0.92 V, which is significantly lower than that of reference analogue of (lsm-TiO₂ (1.18 V), Co₃O₄(3)/bulk-TiO₂ (1.14 V) hybrids, the noble-metal catalyst of Pt/C (1.19 V), and better durability in comparison with IrO₂. The enhanced activity of Co₃O₄(3)/lsm-TiO₂ catalyst is believed to originate from the synergetic effect between Co₃O₄ and TiO₂, higher charge carrier density, as well as the presence of short-range order mesopores and channels which provide short ions diffusion path during the electrocatalytic process. Our results support the new approach to fabricating novel TiO₂-based bifunctional OER and ORR catalysts modified with non-precious metal oxides for energy production, storage, and conversion technologies.