Co₃O₄ Catalysts for Complete Toluene Oxidation: Review including Meta-Analysis of Catalyst Activity Data

Abstract

Designing highly active catalysts for the oxidation of volatile organic compounds (VOCs) involves fine-tuning the catalytic surface and improving its interaction with VOCs. The present review assesses various studies conducted in the last decade on Co₃O₄ catalysts for the complete oxidation of toluene (C₇H₈) and provides information on the synthesis and physicochemical characterization of these catalysts. Taking this one step further, data from the literature were carefully chosen for an extensive meta-analysis aiming at elucidating the structure sensitivity of toluene oxidation over Co₃O₄ catalysts. Specifically, the specific reaction rate was calculated and correlated with the physicochemical properties of the catalysts, namely, the specific surface area, faceting, and Co³⁺/Co²⁺ ratio. Based on this analysis, the specific surface area does not exert a significant impact on the specific activity of Co₃O₄ catalysts for the total oxidation of toluene. Instead, the specific reaction rates are influenced by the morphology, surface concentration, exposed crystallographic planes, and oxidation state of cobalt cations. These factors determine the catalyst’s specific activity by creating defects, oxygen vacancies, or oxygen species with different reactivities. It was also found that a high surface Co³⁺/Co²⁺ ratio is necessary for achieving high catalytic performance in the oxidation of VOCs.

1. Introduction

The emission of volatile organic compounds (VOCs) has been found to have detrimental effects on both human health and the environment. The emission source determines the type and composition of the VOCs. Examples of VOCs include halogenated hydrocarbons, alcohols, aldehydes, aromatics, alkanes, ketones, olefins, ethers, esters, and sulfur-containing compounds [1,2,3].

These molecules are dispersed in the atmosphere and can travel over vast expanses [4]. There are multiple outdoor and indoor sources of VOCs [5,6]. Outdoor sources [6,7,8] include, but are not limited to, chemical industries, paper production, food processing, paint drying, transportation, petroleum refineries, automobile manufacturing, metal degreasing, textile manufacturing, electronic component plants, solvents, and cleaning products. Indoor sources [9] include household products, office supplies, printers, heat exchanger systems, insulating materials, pressed wood, wood stoves, and pipe leaks. Exposure to VOCs can cause various adverse health effects, including irritation, asthma exacerbation, allergy, respiratory effects, and cancer [10].

Aftertreatment technologies for VOC emission control can be classified into two categories based on the recovery or destruction of the pollutant. Methods for the recovery of VOCs include condensation, absorption, adsorption, and membrane separation. VOC destruction involves incineration methods, focusing on thermal [11] and catalytic oxidation [12,13], where the non-halogenated VOC is converted to carbon dioxide and water. When air is used as an oxidant in thermal oxidation, nitrogen oxides (NO_x) and dioxins are also produced as by-products. In contrast, catalytic oxidation is more effective and environmentally friendly as it takes place at lower temperatures with minor by-product formation. Operating at low temperatures is necessary to reduce energy consumption and operating costs. However, thermal incineration excels at high VOC concentrations and flow rates.

Precious metals are among the most active catalysts for the oxidation of VOCs [14,15]. Pt and Pd are the most common examples, mainly supported on carriers with high surface areas, such as alumina [16]. However, the high cost and scarcity of noble metals are limiting factors that raise sustainability concerns. Consequently, attention has been also paid to transition metal oxide (TMOs) catalysts for complete VOC oxidation [17,18,19]. Although TMOs typically have lower activity compared to noble metal catalysts, they are less prone to poisoning. Supported and unsupported oxides include single- or mixed-metal oxides containing two or more transition metal elements [20]. The present study refers to simple transition metal oxides and, more specifically, Co₃O₄. The selection of a single oxide has been made in order to exclude the influence of interactions among various phases or the support, which complicates the analysis of catalytic performance metrics. The most studied single-metal oxides for the total oxidation of VOCs are Co₃O₄ [21,22,23,24], Mn₂O₃ [25,26,27], Fe₂O₃ [28], CuO [29,30], and CeO₂ [31,32]. Co₃O₄ has been widely studied because of its high activity and stability for the complete oxidation of VOCs [32,33].

The most stable phase among CoO, Co₂O₃, and Co₃O₄ is Co₃O₄ [34]. Co₃O₄ is an oxide with a typical spinel crystal structure and is a mixed-valence oxide of CoO and Co₂O₃ [35]. Co²⁺ and Co³⁺ occupy the tetrahedral and octahedral cation sites, respectively [36]. The oxidation of VOCs over Co₃O₄ involves the adsorption of organic compounds at anionic oxygen sites in the oxide lattice, leading to the formation of an activated complex. This complex further reacts to yield oxidation products. The total oxidation of toluene proceeds via C-H bond activation, followed by the formation of benzyl species and the release of hydrogen. Active oxygen species react with the benzyl species to form benzyl alcohol, which is further oxidized to a series of products, including benzaldehyde, benzoic acid, benzene, phenol, benzoquinone, maleic anhydride, various ring-opened byproducts, and finally, carbon dioxide and water [37,38]. Lattice oxygen participates in the catalytic cycle and is regenerated by the supply of gas–phase oxygen adsorbed on a surface oxygen vacancy. Redox models can be used to describe the oxidation of VOCs on reducible metal oxides. Τhis model can be described simply as a reduction of the surface by molecules of volatile organic compounds and the reoxidation of the catalytic surface by gaseous oxygen (Equations (1) and (2)). Scheme 1 is commonly known as the Mars–van Krevelen (MvK) mechanism [39] and is given below.

Surface Reduction: VOC + {Ox} → {Red} + oxidation products

(1)

Surface Re-oxidation: {Red} + O₂ → {Ox}

(2)

where {Ox} is an oxidized surface site (e.g., lattice oxygen) and {Red} is a reduced surface site.

The corresponding MvK rate expression for the oxidation of the VOC, r, is given by Equation (3).

$$r=\frac{k_r k_{ox} p_{VOC} p_{O2}}{\gamma k_r p_{VOC}+k_{ox} p_{O2}}$$

(3)

with

• k_r, the surface reduction rate constant;
• k_ox, the surface reoxidation rate constant;
• p_VOC, the partial pressure of the VOC;
• $p_{O_2}$, the partial pressure of oxygen;
• γ, the stoichiometric coefficient for the total oxidation.

According to Vannice [40], the MvK kinetic model for oxidation reactions on solid catalysts, as originally stated, is inconsistent and incorrect. Consequently, the rate expression derived from this model is inapplicable except for one particular situation that involves the adsorption of O₂ molecules on a single site. However, this model is still inconsistent because O ions (or atoms) are proposed as reactive oxygen species rather than adsorbed O₂ molecules. Consequently, the form of the rate expression in Equation (3) has no physical relevance and must be viewed only as a mathematical fitting function. Carreto et al. [41] agreed with the hypothesis proposed by Vannice [40] and stated that a conceptual shortfall of the original meaning of the “MvK mechanism” is that in a (sub) monolayer catalyst, it is hard to identify a “lattice” oxygen species, as such a catalyst has only surface atoms. Therefore, it would be most helpful to avoid the term MvK and state specifically the structural properties (oxygen abundance, oxidation state, binding energy of oxygen, oxygen species). Thus, the MvK equation is inaccurate because it does not specify the structural properties. Hence, kinetic exchange experiments using isotopes are required. However, the chemical potential that controls the kinetics of oxygen exchange has rarely been considered. Furthermore, the fact that oxides can exchange oxygen species does not imply that their catalytic functions are based on these properties.

Most surface studies have focused on the structural effects of metal oxides and surface reactivity [42]. According to Barteau and Vohs [43], the catalytic reaction rate depends on the geometric structure of the surface atoms on which it occurs, which is implicit in catalysis. These concepts are most commonly used for transition-metal-catalyzed reactions. In this case, structural sensitivity is almost always correlated with specificity [44]. Many authors have investigated the sensitivity of the structure of Pt [45,46], Rh [47], Pd [46], and Au [48], while few studies have focused on metal oxides [49] or mixed-metal oxides [50].

On the other hand, one of the first metal oxides selected to investigate structural sensitivity was MoO₃ [49,51]. The crystallite size of the catalyst can influence the specific catalytic activity, according to Boudart in 1969 [44], whose research work proposes that the catalytic behavior of the considered materials is optimized by tuning the morphology of the particles or bulk structures, and therefore, the specific surface area (S_BET) of the material. In general, the oxidizing power of a metal oxide surface can be associated with the strength of the metal–oxygen bond and the possibility of incorporating molecular oxygen into the crystal lattice, that is, the ability to re-oxidize the reduced active sites with gaseous oxygen. On the contrary, the absence of weakly bonded oxygen species may be fundamental in determining the reaction path and catalyst selectivity.

Interestingly, the influence of the acidic and basic properties of the surface should also be considered [52]. The simplest way to change the surface structure of a metal crystallite is to alter its particle size (which is related to the specific surface area). The other is by employing different synthesis methods to change the crystal shape and morphology. The structure sensitivity demonstrates how different surface structures can lead to reactions or activation within the same molecule. When such materials are designed, the activity and selectivity of the catalysts can be controlled by carefully manipulating the surface and bulk structure of the formulated particles.

Regarding the preparation of Co₃O₄ catalysts, a multitude of synthesis methods have been employed, including hydrothermal, solvothermal [53], sol–gel [54], thermal decomposition [24], chemical precipitation, and hard or soft templates [55]. Different synthesis methods can produce various morphologies and structures, such as nanowires [56], nanoboxes [57], nanoparticles [58], nanorods [59], nanoflowers [32], and nanobrushes [60]. The above studies were triggered by the fact that Co₃O₄ is a versatile oxide whose properties can be used in applications such as solid-state sensors [59,61], lithium-ion batteries [62], and medicine [63]. Although several review articles have been published on the complete catalytic oxidation of VOCs by Co₃O₄ [64] or other transition metal oxides [65,66,67], only one review article has focused on the structural activity with respect to the exposed planes of Co₃O₄ [42]. That study focused on the total oxidation of carbon monoxide.

For this purpose, the current review has been structured to shed light on the following question: “Is the specific reaction rate of VOC oxidation over Co₃O₄ dependent on its specific surface area or its crystallite size and at a second level, does it depend on its exposed crystallographic planes induced by the different morphology?” In other words, is VOC oxidation over Co₃O₄ a structure-sensitive reaction? To answer these questions, the oxidation of toluene serves as a case study. Additionally, this work presents and analyzes the main physicochemical characterization, the various synthesis methods in relation to the obtained morphologies, and the effect of the specific surface area of Co₃O₄ catalysts.

2. Evaluating the Surface Activity of Co₃O₄ Catalysts for Complete Toluene Oxidation

2.1. General Information on the Experimental Procedure for Complete Toluene Oxidation

In most of the research articles examined in this review, complete toluene oxidation was performed in fixed-bed tubular reactors. The reaction rate of VOC oxidation was calculated using Equation (4), where r_VOC is the reaction rate of VOC oxidation (mol s⁻¹ $g_{cat}^{-1}$), F is the total flow rate (mol s⁻¹), W is the mass of the catalyst (g), and ${\mathrm{C}}_{VOC}^{in}$ and $C_{VOC}^{out}$ are the inlet and outlet molar fractions of the VOC, respectively. Dividing the reaction rate by the specific surface area produces the specific reaction rate (SRR), $r_{srr}$(mol s⁻¹ $m_{cat}^{-2}$).

$$r_{VOC}=\frac{(C_{VOC}^{in}-C_{VOC}^{out})F}{W}$$

(4)

It should be mentioned that if the reaction rate was not reported in the examined research articles, it was calculated by extracting the experimental data from the light-off curves (conversion vs. reaction temperature plots) presented. The conversion should typically be less than 20%, and the data are presented as the specific reaction rate (SRR) vs. 1000/T (K⁻¹) vs. temperature (°C). In addition, in the absence of experimental data for the same temperatures for all catalysts, the points were extended along a straight line to make a relevant comparison. The toluene concentration in the reactor feed was in the range of 500–1000 ppm, while the O₂ concentration varied between 10 and 20% in all research articles.

The main intention of this review article is to examine whether the specific reaction rate of VOC oxidation over Co₃O₄ depends on the following factors:

• the specific surface area of Co₃O₄ or, equivalently, the crystallite size of Co₃O₄,
• the morphology and, therefore, the exposed crystallographic planes of Co₃O₄.
• the molar Co³⁺/Co²⁺ ratio derived mainly from XPS and H₂-TPR analyses.

This section investigates the specific activity of the catalyst related to its specific surface area vs. dominant exposed crystallographic planes obtained by various preparation methods. It is also been determined whether the specific activity can be correlated with the exposed crystallographic planes and the Co³⁺/Co²⁺ molar ratio.

2.2. Catalytic Evaluation of Co₃O₄ Catalysts

Bai and co-workers [68] reported on porous Co₃O₄ nanowires and nanorods prepared via the hydrothermal or microemulsion method in the presence of PEG or CTAB. The Co₃O₄-HT-PEG and Co₃O₄-HT-CTAB catalysts are porous cubic Co₃O₄ nanowires, whereas the Co₃O₄-ME-CTAB catalyst consists of porous cubic nanorods. The specific surface areas of the porous Co₃O₄ catalysts were in the range of 47–51.7 m² g⁻¹, and that of the bulk Co₃O₄ was 8 m² g⁻¹, much lower than that of Co₃O₄-HT-CTAB (51.7 m² g⁻¹). In terms of catalytic performance, the porous Co₃O₄-HT-CTAB sample demonstrated better catalytic activity than the other porous Co₃O₄ catalysts and much better activity compared to the nonporous bulk counterpart. The catalytic activity followed the order: Co₃O₄-HT-CTAB > Co₃O₄-HT-PEG > Co₃O₄-ME-CTAB > Co₃O₄-HT> Co₃O₄-Bulk. The lattice spacing (d values) shown in Figure 1 (inset HRTEM) of the Co₃O₄-HT-CTAB catalyst was found to be 0.28 and 0.47 nm, which are relatively close to those of the {220} and {111} lattice planes of the Co₃O₄ catalysts. The lattice spacing of the other porous Co₃O₄ catalysts was 0.28 nm, which was attributed to the {220} lattice plane corresponding to the surface-exposed Co³⁺. As shown in Figure 1, the Co₃O₄-HT-CTAB catalyst exhibited higher specific activity. These results could be associated with the mixed surface-exposed lattice plane {220}/{111}, which is attributed to the Co³⁺/Co²⁺ ratio. This correlation is further supported by the higher initial H₂ consumption rate of Co₃O₄-HT-CTAB during TPR (where initial H₂ consumption stands for <25% of oxygen consumption in cobalt oxide), indicating that the molar ratio of Co³⁺/Co²⁺ was higher for that sample.

Yang et al. [69] employed a modified solvothermal technique to create porous cube-aggregated monodisperse Co₃O₄ microspheres with a specific surface area of 17.4 m² g⁻¹. Meanwhile, they prepared bulk Co₃O₄ containing nano/macroparticles after the thermal decomposition of cobalt nitrate precursor at 600 °C with a specific surface area of 8.4 m² g⁻¹. Both Co₃O₄ catalysts featured solely the crystallographic {111} plane corresponding to surface-exposed Co²⁺ (Figure 2, inset). The Co₃O₄ microspheres showed better catalytic activity than the bulk Co₃O₄ catalyst, with T_100% at 300 and 340 °C, respectively. As shown in Figure 2, both catalysts show similar specific activities. This result can be attributed to the same exposed lattice plane for both samples, {111}, which corresponds to the surface-exposed Co²⁺. In addition, an XPS analysis revealed that the catalysts did not show a remarkable change in the surface Co³⁺/Co²⁺ ratio. Additionally, it should be noted that the consumption of H₂ calculated from the H₂-TPR experiments did not align with the Co³⁺/Co²⁺ molar ratio determined via the XPS analysis.

Wang et al. [70] prepared a nanoplate-aggregate microspherical Co₃O₄ catalyst via an ethylenediamine-assisted hydrothermal method. The catalyst had a cubic crystal structure and comprised nanoplate-aggregate microspheres with a specific surface area of 66 m² g⁻¹. For comparison, the authors prepared bulk Co₃O₄ after the thermal decomposition of the cobalt nitrate precursor at 600 °C. The obtained catalyst exhibited a specific surface area of 9.8 m² g⁻¹. The Co₃O₄ microspheres displayed better catalytic activity than the bulk Co₃O₄. This result is mainly due to the higher specific surface area of the Co₃O₄ microsphere catalyst. However, the Co₃O₄ bulk catalyst, as shown in Figure 2B, has a higher specific activity than the Co₃O₄ microsphere catalyst. The authors observed that the Co³⁺/Co²⁺ molar ratio of the Co₃O₄ microspheres was lower than that of the bulk Co₃O₄, indicating a higher surface concentration of Co³⁺ for the latter. From H₂-TPR, the H₂ consumption of the bulk Co₃O₄ was higher than that of the Co₃O₄ microspheres, showing that the molar ratio of Co³⁺/Co²⁺ was higher in the bulk Co₃O₄, which is in agreement with the XPS results. Hence, the increased specific activity of the Co₃O₄ bulk catalyst could be correlated with the increased molar ratio of Co³⁺/Co²⁺. It should be noted that there were no available HRTEM data to retrieve information about the exposed crystallographic planes and correlate it with the specific activity in that work.

In addition, Liu et al. [71] reported on a meso-Co₃O₄ catalyst with a specific surface area of 94 m² g⁻¹, which was prepared using a KIT-nanocasting method. For comparison, a bulk Co₃O₄ catalyst with a specific surface area of 8 m² g⁻¹ was prepared via the thermal decomposition of a nitrate salt precursor at 700 °C. Once again, the higher the specific surface area of the catalyst, the higher the catalytic activity. Specifically, meso-Co₃O₄ achieved complete conversion of toluene at 220 °C, whereas bulk Co₃O₄ achieved this at 320 °C. According to Figure 3, the calculation and extension of the specific reaction rates at comparable temperatures for the two samples led to similar specific activities, with the bulk Co₃O₄ catalyst being slightly better at low temperatures. In accordance with the previous studies, the bulk Co₃O₄ catalyst has a higher Co³⁺/Co²⁺ molar ratio and higher H₂ consumption compared to the meso-Co₃O₄ catalyst. No HRTEM data were reported for the selected catalysts.

Liu and co-workers [72] prepared a Co₃O₄ catalyst by calcinating MOFs-templated precursors. The catalyst showed a higher specific surface area (94.1 m² g⁻¹) compared to the reference sample prepared via the precipitation method (66.2 m² g⁻¹). The catalysts were evaluated in the total oxidation of toluene, where the Co₃O₄-MOF catalyst exhibited higher catalytic activity on a unit mass basis. The normalization of the reaction rate on a specific surface area basis led to similar specific reaction rates for both samples (Figure 4). According to HRTEM images for the Co₃O₄-MOF sample, the crystallographic planes observed were the {111} and the {220}, attributed to Co²⁺ and Co³⁺, respectively. No HRTEM data were given for the bulk catalyst. Additionally, neither catalyst was characterized via XPS, and thus no conclusion can be drawn for the surface composition of the catalysts.

Co₃O₄ catalysts with different morphologies and exposed crystallographic planes were successfully synthesized by Ren et al. [73] via a hydrothermal process without a surfactant. After the thermal decomposition of the cobalt precursor carried out through the hydrothermal method, 3D hierarchical cubes-stacked Co₃O₄ microspheres (catalyst C), 3D hierarchical plate-stacked Co₃O₄ flowers (catalyst P), 3D hierarchical needle-stacked Co₃O₄ spheres with an urchin-like structure (catalyst N), and 3D hierarchical sheets-stacked fan-shaped Co₃O₄ (catalyst S) were obtained. The specific surface area of the catalysts was 83.1, 58.8, 25.9, and 53.2 m² g⁻¹ for catalysts C, P, N, and S, respectively.

The HRTEM images in Figure 5B confirmed that the crystallographic planes that were primarily exposed for the catalysts C, P, N, and S were the {111}, {110}, {110}, and {112} planes, respectively. Catalyst C has a larger specific surface area than the other catalysts. The catalytic activity for the total oxidation of toluene increased as follows: C > P > N > S. On the other hand, the specific reaction rate at conversions below 20% for the Co₃O₄ catalysts followed the order C = P = N >> S. The unusually low specific activity shown in Figure 5A for the Co₃O₄-S is attributed to the lower molar ratio of Co³⁺/Co²⁺ from XPS and hydrogen consumption from the H₂-TPR analysis, or/and to the preferred exposed crystallographic plane, {112}. It would be a serious omission not to mention that the Co₃O₄-C catalyst presented a higher molar ratio of Co³⁺/Co²⁺ after the XPS and H₂-TPR analyses. Nonetheless, the specific reaction rate did not significantly differ from the other catalysts (Figure 5).

On the other hand, Liu and co-workers [74] reported on the controlled synthesis of 3D hierarchical Co₃O₄ catalysts with different morphologies. The morphology of the Co₃O₄ catalysts could be tuned by simply changing the calcination temperature. For a deeper understanding of the structures of the obtained Co₃O₄ catalysts, the authors performed HRTEM analyses of the exposed crystallographic planes of the U-80 and S-160 catalysts, as displayed in Figure 6. The specific surface areas of the U-80, US-120, and S-160 catalysts were 38.7 m² g⁻¹, 30.2 m² g⁻¹, and 26.7 m² g⁻¹, respectively. The major exposed planes for the catalyst U-80 were the {111} planes. In contrast, for S-160, the dominant exposed planes were the {110} planes. The evaluation of the catalysts in the complete oxidation of toluene and the meta-analysis of the extracted data demonstrated that the catalytic activity both on a unit mass basis and per surface area followed the order: S-160 > US-120 ≈ U-80. The increased specific activity of the S-160 catalyst, as shown in Figure 6, can be attributed to the exposed {110} planes and the notable change in the molar ratio of Co³⁺/Co²⁺ from XPS and H₂-TPR, compared with the other catalysts.

A series of Co₃O₄ catalysts, such as dumbbell-brush-type mesoporous nanosheets and nanorods, were successfully prepared by Xu et al. [75]. The specific surface areas of Co₃O₄-D, Co₃O₄-S, and Co₃O₄-R were 50, 35, and 26 m² g⁻¹, respectively. Co₃O₄-D had a porous dumbbell-brush-type shape with self-assembled nanorods, and as shown in Figure 7. A lattice spacing of 0.280 nm was observed, which can be indexed to the {220} plane of the Co₃O₄ phase. Co₃O₄-S, on the other hand, displays a hexagonal nanosheet morphology. When examining the HRTEM images in Figure 7, it is clear that the nanosheet primarily exposed the {220} facets. In addition, the dominant facets observed for the Co₃O₄-R nanorods were the {220}.

The aforementioned catalysts were evaluated for the complete oxidation of toluene. It can be observed that the difference in the specific surface area of the Co₃O₄ catalysts strongly influences their catalytic activity. More specifically, Co₃O₄-D had the highest specific surface area and was the best-performing catalyst both in terms of activity and specific activity. Taking into account that all the samples had the same exposed {220} facet, no conclusion can be drawn from the HRTEM results. Nevertheless, according to the XPS and H₂-TPR analyses, the higher specific reaction rate of Co₃O₄-D could be linked with the higher molar ratio of Co³⁺/Co²⁺ compared to the Co₃O₄-S and Co₃O₄-R catalysts.

Figure 8 depicts a summary of the specific reaction rates attained for the examined Co₃O₄ catalysts (conversions <20%) in the present review. A summary of the catalytic materials, their main physicochemical properties, and specific activity is given in

Table 1. Summary of the Co₃O₄ catalysts chosen for the meta-analysis, including the preparation methods, their main physicochemical properties (S_BET, morphology and Co³⁺/Co²⁺ ratio), and the specific reaction rates attained at 200 °C for the total oxidation of toluene.

Catalyst	Preparation Method	Morphology	SBET (m2 g−1)	Co3+/Co2+ Ratio	GHSV (ml g−1 h−1)	SRR (200 °C) (×10−10 mol s−1 m−2)	Ref.
Co3O4-HT	hydrothermal	irregular	41.9	2.13	20,000	4.00	[68]
Co3O4-HT-PEG	hydrothermal	nanowires /nanorods	50.9	2.01	20,000	5.68	[68]
Co3O4-HT-CTAB	hydrothermal	nanowires /nanorods	51.7	1.72	20,000	33.4	[68]
Co3O4-ME-CTAB	microemulsion	nanowires /nanorods	46.5	2.05	20,000	7.02	[68]
Co3O4- microsphere	solvothermal	microspheres	17.4	1.36	20,000	6.36	[69]
Co3O4-bulk	thermal decomposition	irregular	8.4	1.47	20,000	5.86	[69]
Co3O4- microsphere	hydrothermal	microspheres	66	1.44	20,000	4.88	[70]
Co3O4-bulk	thermal decomposition	polyhedra	9.8	1.65	20,000	11.2	[70]
Meso-Co3O4	KIT-nanocasting	3D ordered mesoporous	94.1	1.33	20,000	16.3	[71]
Co3O4-bulk	thermal decomposition	-	8	1.74	20,000	12.4	[71]
Co3O4-MOF	MOF precursors calcination	distorted nanocubes	94.1	-	60,000	4.53	[72]
Co3O4-bulk	precipitation	-	66.2	-	60,000	5.64	[72]
Co3O4-C	hydrothermal	3D hierarchical cubes	83.1	1.67	48,000	3.58	[73]
Co3O4-P	hydrothermal	plates	58.8	1.02	48,000	2.00	[73]
Co3O4-N	hydrothermal	3D hierarchical needles	25.9	1.02	48,000	2.36	[73]
Co3O4-S	hydrothermal	3D hierarchical sheets	53.2	1.02	48,000	0.10	[73]
U-80	hydrothermal	urchin-like	38.7	0.31	60,000	0.96	[74]
US-120	hydrothermal	urchin/shale-like	30.2	0.38	60,000	1.23	[74]
S-160	hydrothermal	shale-like	26.7	0.61	60,000	6.90	[74]
Co3O4-D	hydrothermal	dumbbell-brush-type	50.8	2.14	60,000	7.18	[75]
Co3O4-R	solvothermal	mesoporous nanorods	26	1.62	60,000	3.27	[75]
Co3O4-S	sol-gel	mesoporous nanosheets	35.9	1.95	60,000	2.02	[75]

. The catalysts were synthesized with various methods and exhibited different specific surface areas in the range of 8–100 m² g⁻¹, different facets, and different Co³⁺/Co²⁺ molar ratios. The catalytic evaluation was performed for the complete oxidation of toluene. Based on the Arrhenius plots presented in Figure 8A, the Co₃O₄ catalysts attain specific reaction rates ranging typically from 9.5 × 10⁻¹¹ to 1.6 × 10⁻⁹ mol s⁻¹ m⁻², irrespective of the synthesis method and specific surface area. The samples Co₃O₄-HT-CTAB and Co₃O₄-S constitute the obvious exceptions, with specific reaction rates of 3.3 × 10⁻⁹ and 1 × 10⁻¹¹ mol s⁻¹ m⁻², respectively, after extrapolating the Arrhenius plots to 200 °C for comparison with the rest of the samples. For a better visualization of the data, the specific reaction rates obtained at 200 °C were plotted as a function of the specific surface area of the studied catalysts (Figure 8B). If the samples Co₃O₄-HT-CTAB and Co₃O₄-S are not taken into account, it is apparent that the specific surface area does not significantly affect the specific reaction rates observed for the oxidation of toluene over Co₃O₄ catalysts. The calculated specific reaction rates are scattered within the 1 × 10⁻¹⁰–1 × 10⁻⁹ mol s⁻¹ m⁻² range with no obvious trend for values of specific surface area from 8 to 94 m² g⁻¹. Nevertheless, the variations observed in the specific rates are in the order of 3–7 times if Co₃O₄-HT-CTAB and Co₃O₄-S are excluded, and up to 330 times if these samples are included. Hence, it can be concluded that the specific surface area does not affect, to a significant extent, the specific activity of Co₃O₄ catalysts for the oxidation of toluene, at least in the S_BET region studied. However, considering the variations in the specific reaction rates, it can be deduced that factors such as the morphology and induced exposure of different facets, as well as the surface concentration and oxidation state of cobalt cations, play a key role in the specific activity of the samples.

Additionally, the apparent activation energies (E_app) of all the studied catalysts were calculated from the corresponding Arrhenius plots, and they were found to vary widely from 15 ± 3 kJ mol⁻¹ to 157 ± 2 kJ mol⁻¹.

3. Co₃O₄ Catalysts and Their Physicochemical Characterization

The physicochemical characterization of the catalytic materials is essential for a deep understanding of their catalytic performance. The correlation of the structural, textural, and chemical properties of the catalysts with the catalytic evaluations will enable the rational design and improvement of their activity. In the following section, the typical physicochemical techniques used for the detailed characterization of Co₃O₄ will be analyzed.

3.1. Powder X-ray Diffraction (PXRD)

Usually, the catalysts are scanned in the 2θ range of 10°–80° and show eleven diffraction peaks related to face-centered cubic spinel Co₃O₄ (Fd3m space group) (Figure 9A). The 2θ diffraction peaks are 19.1° (111), 31.2° (220), 36.8° (311), 38.5° (222), 44.7° (400), 55.6° (422), 58.9° (511), 64.7° (440), 73.6° (620), 76.8° (533), and 77.7° (622).

The significant broadening of the diffraction peaks reveals a nanometric particle structure. The average crystallite size is calculated using the Scherrer equation (Equation (5)) [76].

$$d_{PXRD}=\frac{K\ast \lambda }{\beta \ast cos\Theta }$$

(5)

where d_PXRD is the crystallite diameter, K is a constant with the value of 0.941, λ is the radiation wavelength, β is the peak width at half maximum in radians, and θ is the reflection angle. In addition, the lattice parameters of the catalysts can be calculated from a Rietveld refinement analysis of the PXRD patterns. From the Rietveld analysis, the microstrain related to the method and temperature of synthesis of the catalysts can be drawn. Figure 9B [42] shows a typical cubic spinel structure schematic, with Co²⁺ ions in the tetrahedral interstices and Co³⁺ ions in the octahedral interstices. In this standard spinel structure, Co²⁺ cations occupy 8a sites with tetrahedral coordination, whereas Co³⁺ cations are located at 16d sites with octahedral coordination, according to the formula [Co²⁺]_8a [Co³⁺]_16d [77,78].

The Co²⁺ cations correspond to 8a sites and form a diamond-like lattice, consisting of two displaced face-centered sublattices tetrahedrally surrounded by the four nearest Co²⁺ neighbors. The close-packed planes were {001}, {111}, and {110}, and their surface atomic configurations are shown in Figure 9b–d. Thus, the {001} and {111} planes contain only Co²⁺ cations, whereas the {110} plane primarily consists of Co³⁺ cations [42]. An increase in the crystallite size leads to a decrease in the exposed surface area (per unit volume or mass), which is supported by the example shown in

Table 2. Data from N₂ physisorption and P_XRD, adapted from reference [79].

Catalysts	SBET (m2 g−1)	dPXRD (nm)
CoDP	82	11.9
CoAP	37	21.6

.

3.2. X-ray Photoelectron Spectroscopy (XPS)

XPS is an essential technique for examining the surface chemical composition of Co₃O₄. As illustrated in Figure 10, a typical XPS spectrum shows the splitting of Co 2p_3/2 and Co 2p_1/2 at 781 and 795 eV, respectively. These show the surface species combined with the shake-up satellite located at a higher binding energy of 6.7 eV from the main peak. The estimated 2p_3/2–2p_1/2 spin–orbit splitting is 15.1 eV for Co₃O₄. Octahedral Co³⁺ is located at 779.6 eV, and tetrahedral Co²⁺ is located at 780.7 eV, with spin–orbit splitting at 794.5 eV and 796 eV, respectively. These results indicate the existence of surface Co³⁺ and Co²⁺ species [80]. After background correction and peak deconvolution, the Co³⁺/Co²⁺ molar ratio of Co₃O₄ on the surface can be expressed, which could play a crucial role in oxidation catalysis. With differences in coordination, oxidation states, and bond lengths of different Co₃O₄ catalysts, it is expected that differences exist in the chemical shifts, multiple splits, shake-up structures, and valence band structures.

The deconvolution of the O1s spectrum of Co₃O₄ indicates oxygen species with binding energies around 529.9, 531.6, and 533.5 eV, which are due to the surface lattice oxygen (O_latt), the surface hydroxyl species (O_ads), and the adsorbed water species, respectively [81]. A high relative O_ads/O_latt concentration ratio on the Co₃O₄ surface did not enhance catalytic oxidation. It should be highlighted that a material with a higher Co³⁺/Co²⁺ molar ratio shows a lower O_ads/O_latt molar ratio.

Figure 11A shows the surface Co³⁺/Co²⁺ molar ratio calculated from the XPS data as a function of the specific surface area (S_BET) for all surveyed catalysts. The Co³⁺/Co²⁺ molar ratio varies from 1.1 to 3.2 for all the catalysts with an S_BET in the range 8–94 m² g⁻¹. More specifically, twelve of the catalysts with an S_BET below 40 m² g⁻¹ show a Co³⁺/Co²⁺ molar ratio from 1.1 to 1.7, while five of the catalysts with an S_BET of 42–58 m² g⁻¹ show a Co³⁺/Co²⁺ molar ratio from 1.8 to 2.1. Moreover, eight of the catalysts with an S_BET of 8–80 m² g⁻¹ present a Co³⁺/Co²⁺ molar ratio from 2.2 to 3.2. It is worth mentioning that the theoretically expected value of the Co³⁺/Co²⁺ ratio is two. This difference presented by the research papers studied is probably due to the different synthesis methods and, therefore, the different exposed crystallographic planes.

The correlation between the calculated specific reaction rates (see Figure 8 and Table 1) and the surface Co³⁺/Co²⁺ molar ratio is illustrated in Figure 11B. It can be deduced that a higher surface concentration of Co³⁺ has a rather positive effect on the specific activity of Co₃O₄ catalysts for the total oxidation of toluene, but the experimental scattering of the results is significant. Therefore, the Co³⁺/Co²⁺ molar ratio is certainly not the single factor determining the specific reaction rate of the catalysts.

3.3. Raman Spectroscopy and Attenuated Total Reflectance Fourier Transmittance Infrared Spectroscopy (ATR-FTIR)

Co₃O₄ Raman spectra consist of five Raman bands, as reported by Dialo et al. [82] and Lei et al. [83], and they are shown in Figure 12A,B. The Raman bands at 194, 480, 520, 619, and 670 cm⁻¹ are assigned to the F_2g⁽³⁾, E_g, F_2g⁽¹⁾, F_2g⁽²⁾, and A_1g modes of the Co₃O₄ spinel structure. The sublattice vibration attributable to octahedral–tetrahedral sites is at a value of 670 cm⁻¹. Any minor variations in this vibration can be associated with a lattice distortion of the spinel system or residual stress. This phenomenon is observed in Figure 12B for Co₃O₄ prepared from different MOF precursors.

However, from the ATR, the observed 570 cm^–1 band is characteristic of the O–Co bond, denoting Co³⁺ at the octahedral site. The 668 cm^–1 band is attributable to the Co²⁺–Co³⁺–O₃ (Co²⁺ in the tetrahedral site) vibration in the spinel lattice. Any asymmetric band at 1627 cm⁻¹ is due to the OH groups from adsorbed water. It should be noted that adsorbed water can be easily desorbed at temperatures around 120–150 °C under a gaseous flow.

3.4. Hydrogen Temperature-Programmed Reduction (H₂-TPR)

Temperature-programmed reduction (TPR) is a widely used technique to investigate the reducibility of solid catalysts. TPR involves exposing the catalytic material to a stream of reducing gas, usually H₂, at various temperatures, resulting in a TPR profile, i.e., the reduction rate versus the temperature of the reactor. Qualitatively, the position of the reduction peaks provides information on the ease of reduction of these chemical species, where easier reduction is indicated by the appearance of a peak maximum at a lower temperature. Quantitatively, the amount of hydrogen consumed during the measurement can be compared to a theoretical one to determine the extent of the catalyst reduction. The reduction of the Co₃O₄ catalysts is exemplified in the H₂-TPR profiles by two reduction peaks. The peak at 200–300 °C is due to the reduction of Co₃O₄ to CoO. The peak within the 300–450 °C range is due to the reduction of CoO to metallic Co⁰. The theoretical H₂ consumption for the complete reduction of Co₃O₄ to Co⁰ is 16.6 mmol g⁻¹. The overall reduction is described by Equations (6) and (7) as follows:

Co₃O₄ + H₂ → 3CoO + H₂O

(6)

3CoO + 3H₂ → 3Co + 3H₂O

(7)

It is worth mentioning that the peak below 200 °C is attributed to the reduction of oxygenated surface species formed during oxidation pre-treatment [84]. The range of the reduction temperature varies according to the surface area, preparation method, and morphology of the catalysts. Typical H₂^-TPR profiles are presented in Figure 13, taken from the work of Zhang et al. [79], Ma et al. [60], and Ren et al. [73].

As depicted in Figure 13A, a slight modification in the precipitation method from Zhang et al. [79] resulted in CoAP having a 2.2 times lower specific surface area compared with CoDP, and thus revealing different crystallite sizes. Furthermore, the difference in material composition results in a difference in the Co³⁺/Co²⁺ ratio and, finally, in the temperature reduction profile. Figure 13B compares the H₂-TPR profiles of the Co₃O₄ rods (Co₃O₄-R), Co₃O₄ hexagonal plates (Co₃O₄-H), and Co₃O₄ double-sided helianthus discs (Co₃O₄-D). These materials were also synthesized using the hydrothermal method by Ma et al. [60] by adjusting the water/ethanol ratio. In this case, Co₃O₄ with different morphologies did not show remarkably different H₂-TPR profiles or reduction temperatures. This is possibly because the catalysts have only minor differences in their specific surface area. The main difference was the highest Co³⁺/Co²⁺ ratio, indicating that more Co³⁺ species were exposed on the surface of the Co₃O₄-B catalyst. This result is probably due to the unique double-sided nanobrush structure of the Co₃O₄-B catalyst.

Figure 13C shows the H₂-TPR profiles of the hydrothermally synthesized Co₃O₄ catalysts present in the form of cube-stacked microspheres (C catalyst), plate-stacked flowers (P catalyst), needle-stacked double-spheres with an urchin-like structure (N catalyst), and sheet-stacked fan-shaped Co₃O₄ (S catalyst) [73]. According to these H₂-TPR profiles, the catalyst morphology and, therefore, the specific surface area played a significant role compared with the Co³⁺/Co²⁺ molar ratio. In this case, it is also necessary to note that the C catalyst has a higher S_BET of 83 m² g⁻¹, those of the S and P catalysts are close to 55 m² g⁻¹, and that of the N catalyst has the lowest value of 30 m² g⁻¹. In contrast, the mean crystallite size of the C, P, N, and S catalysts were 16, 18, 30, and 17 nm, respectively. Thus, it was demonstrated that the cube-stacked microspheres (catalyst C) with the highest S_BET had slightly lower reduction temperatures.

To summarize, three different H₂-TPR profiles are shown in Figure 13. In all cases, the morphology and S_BET played an essential role. It is observed that the main factor in the profile and the temperature of reduction in Co₃O₄ is the Co³⁺/Co²⁺ molar ratio. It should be noted that higher Co³⁺/Co²⁺ ratio values resulted in an alteration of the reduction profile. Moreover, it is evident that the reduction of CoO to Co is highly dependent on the Co₃O₄ morphology.

3.5. Oxygen Temperature-Programmed Desorption (O₂-TPD)

The presence of oxygen, either adsorbed, lattice-bound, or in gaseous form, is a critical factor in the oxidation reactions that take place over transition metal oxide catalysts. In this context, the O₂ Temperature Programmed Desorption (O₂-TPD) technique is a valuable tool for assessing oxygen storage and mobility, and it is capable of discerning between various oxygen species on the catalytic surface. Typical TPD profiles (O₂ desorption rate vs. temperature) comprise one or more peaks attributed to different adsorption strengths of oxygen species. The representative O₂-TPD profiles of the Co₃O₄ catalysts are shown in Figure 14. Figure 14A refers to the Co₃O₄/KIT-6 ordered mesoporous catalysts prepared using the impregnation method. These experiments were performed on Co₃O₄ catalysts to explore the relationship between adsorbed oxygen species and the Co₃O₄ surface structure. The O₂-TPD profiles of the transition metal oxides showed that the oxygen species were desorbed within a wide temperature range. The O₂-TPD profiles were divided into three main categories: physically adsorbed oxygen (O₂, ads), chemically adsorbed oxygen (O₂⁻/O⁻), and lattice oxygen (O²⁻). It is also worth noting that physically adsorbed oxygen is easily desorbed and can be removed before the start of the O₂-TPD experiment using a pure He flow.

Chemisorbed oxygen molecules can be desorbed above 100 °C from metal oxides, whereas lattice oxygen is challenging to desorb, and the temperature required is generally close to 500 °C. In Figure 14A, Xie et al. [85] describe the α and β oxygen desorption regions of Co₃O₄/KIT-6 catalysts. The oxygen desorption peaks in areas α and β belong to the surface desorption of weakly chemisorbed oxygen species and strongly chemisorbed oxygen (O₂⁻/O⁻) species from oxygen vacancies, respectively.

On the other hand, Ren et al. [73] (Figure 14B) define four kinds of oxygen species that are observed in the temperature ranges of 80–120 °C, 120–450 °C, 450–700 °C, and above 700 °C. The above temperatures could be assigned to molecular oxygen species adsorbed on oxygen vacancies (O_v), surface-adsorbed oxygen ions (O_ads), surface lattice oxygen (O_latt,_s), and bulk-phase lattice oxygen (O_latt,_b), respectively. The desorption peak at approximately 100 °C for the C, P, and N catalysts was attributed to the desorption of molecular oxygen. Thus, the O₂-TPD profiles show that the chemisorbed oxygen species can desorb at relatively low temperatures. The desorption profiles varied depending on the preparation method and the catalyst morphology.

3.6. Scanning Electron Microscopy (SEM) and High-Resolution Transmission Electron Microscopy (HRTEM)

Figure 15 presents the SEM and HRTEM images of the Co₃O₄ catalysts prepared using different synthesis methods. In addition, SEM and high-resolution TEM images were used to examine various catalyst morphologies and exposed crystallographic planes, respectively. The HRTEM technique is employed to further study the size distribution of crystallites and the active facets for catalysis. It is worth noting that in most cases, the HRTEM size distribution and the crystallite size calculated from XRD must be comparable.

Figure 15A shows the HRTEM images of M-Co₃O₄-350 (Figure 15(Ae)), M-Co₃O₄-450 (Figure 15(Ah)), and M-X-250 (Figure 15(Ab)). The lattice spacing in Figure 15Ac for M-X-250 was 0.24 nm, corresponding to the {311} plane. For the M-Co₃O₄-350, the three lattice fringes are shown in Figure 15(Ae,f), which are 0.24 nm, 0.28 nm, and 0.467 nm, and these are derived from the {311}, {220}, and {110} crystallographic planes, respectively. Nevertheless, the {110} plane is the dominant exposed surface for this material. However, M-Co₃O₄-450 (h, i) demonstrates that the {111} plane is the predominant one. From the above data, it is clear that the exposed planes of Co₃O₄ can be manipulated by controlling the Co-MOF calcination temperature. Figure 15A presents the Co-MOF (ZSA-1) materials synthesized through a hydrothermal method. The as-received Co-MOF was used as a precursor to obtain the final catalysts. The precursor was calcined in air at different temperatures (250, 350, and 450) for 1 h. The SEM images in Figure 15(Aa,d,g) present the catalysts obtained after calcination. These Co₃O₄ catalysts exhibited octahedral morphologies with slightly different surfaces depending on the calcination temperature.

The SEM images in Figure 15(Ba,d,g,j) show the double-sided nanobrush Co₃O₄ (Co₃O₄-B), Co₃O₄ rods (Co₃O₄-R), Co₃O₄ hexagonal plates (Co₃O₄-H), and Co₃O₄ double-sided helianthus discs (Co₃O₄-D), which were prepared by tuning the water/ethanol ratio using a hydrothermal method.

Figure 15(Bc,f,i,l) presents the HRTEM images of the aforementioned materials. The Co₃O₄-B lattice spacing in Figure 15(Bc) is 0.234 nm, corresponding to the {222} plane of Co₃O₄-B. In addition, the HRTEM image of the Co₃O₄-R catalyst showed a lattice spacing of 0.235 nm (Figure 15(Bf)), which corresponds to the {222} plane of Co₃O₄. A {220} crystallographic plane with a d-spacing of 0.275 nm is observed in Figure 15Bi for the Co₃O₄-D catalyst. The HRTEM image in Figure 15Bl shows that the Co₃O₄–H catalysts possessed a {220} plane with a lattice spacing of 0.282 nm. Based on these results, it can be concluded that the exposed facets of the catalysts can be modified by changing their morphology.

4. Designing Co₃O₄ Catalysts

In this section, the preparation methods and the precursors employed to obtain Co₃O₄ catalysts with different textural and structural characteristics are summarized. Several cobalt compounds have been used as precursors to prepare Co₃O₄ for the complete oxidation of VOCs. These include cobalt nitrate hexahydrate [Co(NO₃)₂·6H₂O], cobalt chloride hexahydrate [CoCl₂·6H₂O], cobalt sulfate hexahydrate [Co(SO₄)₂·6H₂O)], cobalt hydroxide [Co(OH)₂], cobalt basic carbonate [2CoCO₃·3Co(OH)₂·H₂O], cobalt acetate [Co(CH₃COO)₂], and cobalt acetylacetonate (Co(C₅H₇O₂)₂). Co-MOFs as precursors have also been a new trend in recent years. From Figure 16, it can be concluded that the most used precursor is cobalt nitrate, employed in 72% of the reported studies, followed by cobalt chloride and cobalt acetate (11% each).

The most representative Co₃O₄ synthesis methods are shown in Scheme 2. These include the hydrothermal/solvothermal method [53], the sol–gel process [54], thermal decomposition [24], chemical precipitation, and hard or soft templates [55]. The simple thermal decomposition of a cobalt precursor was used by Rivas et al. [33], Wang et al. [70], Liu et al. [58], Bai et al. [68], and Yan et al. [32]. The thermal decomposition method results in catalysts with a specific surface area in the range 3–82 m² g⁻¹, which depends on the calcination temperature, typically 300–700 °C.

Other research groups such as Li et al. [22], Ren et al. [87], Ma et al. [60], Zhang et al. [79], Liu et al. [74], Liotta et al. [88], Ma et al. [89], Fan et al. [90], and Anke et al. [91] employed a precipitation method to prepare Co₃O₄ as a catalyst for complete VOC oxidation. The specific surface area of the produced catalysts varied from 5 to 151 m² g⁻¹ depending on the calcination temperature. The employed precipitation agents were, in most cases, urea (CH₄N₂O), hydroxycarbonate (CH₂O₅), sodium carbonate (Na₂CO₃), ammonium hydroxide (NH₃·H₂O), sodium hydroxide (NaOH), potassium hydroxide (KOH), ammonium bicarbonate (NH₄HCO₃), potassium bicarbonate (KHCO₃), potassium carbonate (K₂CO₃), and oleylamine (C₂H₂NH₂).

Fei et al. [92], Xu et al. [75], Hu et al. [93], Xu et al. [75], Fei et al. [92], Zhao et al. [94], Wang et al. [95], Zhou et al. [96], Wang et al. [70], Yang et al. [69], Bai et al. [68], Ma et al. [60], Ren et al. [73,97], Liu et al. [74], Liu et al. [98], and Li et al. [99] used a solvothermal–hydrothermal method at a reaction temperature of 80–180 °C to obtain Co₃O₄ catalysts with different morphologies. Typical patterns resulting from this method include cubes, nanorods, spheres, nanowires, nanotubes, microspheres, double-sided nanobrushes, 2D-nanoplates, 3D-nanoflowers, 1D-nanoneedles, 3D hierarchical cube-stacked microspheres, 3D hierarchical stacked needles, two-spheres with an urchin-like structure, 3D hierarchical fan-shaped stacked sheets, and dumbbell-brush-type. The specific surface area of these materials varied between 8 and 83 m² g⁻¹. Additionally, Dissanayake et al. [21], Xie et al. [85], Mei et al. [100], Ren et al. [73], Liu et al. [58,71], Deng et al. [101], and Ma et al. [102] [89] synthesized three-dimensional (3D) or two-dimensional (2D) mesoporous Co₃O₄ materials using the template method with SBA-15, KIT-6, P123, and PMMA as templates. Except for PMMA, which was removed after calcination, the other templates were removed by washing with a highly concentrated solution of NaOH. The specific surface areas of the received Co₃O₄ catalysts were in the range of 21 to 553 m² g⁻¹.

A sol–gel method was employed by Ercolino et al. [103], Puértolas et al. [104], Rivas et al. [33], Kirchnenova et al. [105], and Liu et al. [106] to synthesize Co₃O₄ catalysts. The final catalysts are calcined at a temperature of 350–600 °C, and the typical values of the specific surface area vary between 1 and 27 m² g⁻¹. Several research groups, including Rivas et al. [33], Liu et al. [106], and Solsona et al. [107], have prepared Co₃O₄ catalysts through the application of a grinding technique that involves metal precursors and appropriate concentrations of C₆H₈O₇ or NH₄HCO₃. The resulting specific surface area ranged between 30 and 160 m² g⁻¹ and was observed to be significantly dependent on the grinding time and the calcination temperature.

Recently, there has been a growing trend in synthesizing mesoporous catalysts through the thermal decomposition of MOFs (Liu et al. [72], Zhang et al. [108], Lie et al. [86], and Zhao et al. [109]). The shape of the final catalyst varies depending on the shape of the organic ligand but is often in the form of void nanocages, octahedrons, and macroscopic hexagonal crystals. This synthesis method yields a specific surface area ranging from 2.5 to 118 m² g⁻¹.

Figure 17 presents the percentage of Co₃O₄ catalysts prepared through various methods in relation to their specific surface area (S_BET). The majority of Co₃O₄ catalysts exhibited a specific surface area within the range 30–60 m² g⁻¹, constituting 35% of the samples. Catalysts with a S_BET of 15–30 m² g⁻¹ accounted for 25% of the samples. Finally, catalysts with specific surface areas in the low (<15 m² g⁻¹) or high (>60 m² g⁻¹) ranges represented around 20% of the samples, respectively.

5. Overall Conclusions, Challenges, and Perspectives

Catalytic oxidation is an effective after-treatment technology for the abatement of VOC emissions from industrial sources. The preparation of noble metal-free catalysts with high activity and stability remains challenging. This review focuses on Co₃O₄ catalysts, aiming at identifying the structure sensitivity of toluene oxidation as a case study. Particular attention has been paid to the specific reaction rate of the catalysts and its correlation with the textural and structural characteristics of Co₃O₄ samples. Extensive reference has been made to synthetic methods and characterization techniques to infer whether the catalytic performance is related to the specific surface area, morphology, facets, and surface oxidation state of the catalysts. From the aforementioned findings, it can be concluded that the specific surface area does not have a significant impact on the specific reaction rate of Co₃O₄ catalysts for the complete oxidation of toluene. However, the specific reaction rates differ due to factors such as the morphology, different exposed crystallographic planes, surface concentration, and oxidation state of cobalt cations. These factors are crucial in determining the specific reaction rate of Co₃O₄ by causing the formation of defects (steps, kinks, etc.) and oxygen vacancies or the development of oxygen species with different reactivities. More specifically, the preparation of Co₃O₄ with {110}, {220}, or {220}-{111} exposed crystallographic planes results in enhanced specific activity that could be attributed to the higher Co³⁺/Co²⁺ surface ratio induced by the preferential exposure of Co³⁺ cations on the catalytic surface. On the other hand, when the only exposed plane is the {111} plane, which is populated by Co²⁺, the catalytic surface is inactive for the oxidation of toluene. A high Co³⁺/Co²⁺ ratio is a prerequisite for a high catalytic performance in the oxidation of VOCs.

In a broader context, the topic of the structure sensitivity of reactions taking place on catalytic surfaces has been raised mostly for supported noble metal catalysts. In the case of metal oxide catalysts, structure sensitivity might theoretically manifest itself through two pathways as a function of the primary crystallite size of the oxide. The first pathway refers to the number density of surface defects and how this is influenced by the crystallite size and the specific surface area of the catalyst. It is typically expected that surface defects would lead to more active surface sites due to the presence of cations with a lower coordination number. Therefore, an increase in crystallite size (hence a decrease in the specific surface area) would lead to fewer defective crystals and, as a result, to a decrease in specific catalytic activity. However, such a behavior was not observed according to the analysis of this work since the specific activity of the catalysts did not show a definitive trend with variations in the specific surface area. The second pathway, which refers to the expression of different crystallographic planes (faceting) with a variation in crystallite size and the employed synthesis method, appears to be more relevant. Based on our analysis, the focus on developing Co₃O₄ catalysts with high specific surface areas does not suffice to prepare a highly active oxidation catalyst since the real challenge is to tailor the synthetic route to favor the development of surface traits that also contribute to the enhancement of the intrinsic activity of Co₃O₄ catalysts. Control of crystal orientation and shape can lead to the design of tailored nanomaterials with optimized catalytic function. The designated choice of a single oxide in the current review was a deliberate step taken to eliminate the influence of interactions among various phases or the support, which otherwise would complicate the analysis of catalytic performance metrics. This methodological approach is intended to ensure that the performance evaluation of the catalyst is based solely on its activity and not on any external factors that may interfere. It is noteworthy that the insights gained from this review can serve as a paradigm for other TMO-based catalysts that are commonly used in the field.