Natural gas, which contains mostly methane, is abundant in our planet and is widely used in power generation and other heating applications because of its low cost [1-3]. However, unburned CH₄ is a more potent greenhouse gas because its global warming potential is twenty times higher than that of CO₂; CH₄ has become a serious environmental problem that has drawn the attention of many researchers [4-7]. Therefore, the development of catalysts for completely oxidizing unconverted CH₄ at low temperatures has held great promise in the pursuit of highly efficient engine combustion and turbine-based energy devices, prompting the urgent need for low-cost, efficient, and robust catalytic emission control devices [8].

However, it is difficult to initiate oxidation of CH₄ at low temperatures (less than 300 ℃) because CH₄ is a typical nonpolar molecule, and the C-H bond in CH₄ has the highest bond energy (434 kJ/mol) among all organic compounds [7, 9]. The activation of a C-H bond is the most critical step for the catalytic combustion of methane (CCM). The activation mechanism of the C-H bond can be summarized as follows: (1) homolytic activation mechanism over precious-metal-based catalysts and (2) heterolytic activation mechanism on oxides [10, 11]. Generally speaking, noble metals [12-16] (e.g., Pt, Pd), especially Pd-based catalysts [17, 18], can exhibit outstanding catalytic performance and complete oxidation of methane at low temperatures. In order to improve the catalytic efficiency of noble metals, three-dimensionally ordered micro/mesoporous (3-DOM) supported high dispersion noble metal nanoparticles have been reported by Dai' s group [19] and Li's group [20]. However, high costs have limited their wide application. Transition metal elements generally have multiple valence states and easily form a redox cycle between the high and low oxidation states, and lattice oxygen can be released and restored [21]. For example, LaMnO₃ with mild citric acid etching showed good catalytic activity due to higher surface area, abundant vacancies, higher surface Mn⁴⁺/Mn³⁺ ratio, and more active oxygen species [22]. Li's group obtained Co₃O₄ nanosheets with exposed high index (112) crystal planes, which showed a high catalytic performance for CCM because the unusually high index crystal planes were more reactive than common planes such as (001) and (011) [23]. Fe-substituted hexaaluminates with superior thermal stability synthesized by a two-step strategy were shown to enhance surface area and show high catalytic activity [24, 25]. Other metals oxides, such as Cu, Cr, In, and Sn oxides, have been used as catalysts in the past decades [21, 26, 27]. However, the catalytic activity of those oxides for complete oxidation of CH₄ above 450 ℃ occurred at relatively low gas hour space velocity (GHSV). Therefore, efforts to develop non-noble metal oxides catalysts for low temperature methane oxidation become more urgent.

Spinel oxides with catalytic activity even higher than the precious-metal-based catalysts are promising candidates for CCM at low temperature, although relatively little research has been conducted on this catalytic material [21]. Tao's group first reported that NiCo₂O₄ spinel oxides exhibited higher catalytic activity at relatively low temperature. This results from the integration of Ni³⁺, on which methane dissociates to methyl and couples with the surface lattice oxygen atoms to form -CH₃O, and CO₂ forms after dehydrogenation and oxidation [28]. McEwen et al. [29] also presented some evidence that a high oxidation state of Ni is more effective for the first C-H cleavage in methane using both experiments and density functional theory (DFT) calculations. In containing the Co spinel structure, however, most researchers agreed that Co³⁺ association with catalytic activity was essential [23, 30]. Lim et al. [3, 31] controlled the Co to Ni ratio and used mesoporous silica KIT-6 as the hard template to synthesize ordered mesoporous NiCo₂O₄, which showed higher activity due to the high amount of Co³⁺ and surface oxygen species on the surface. Furthermore, others also reported methane more easily dissociates to methyl on Co cations compared to Ni cations based on results from DFT calculations [8]. Therefore, the relative catalytic activities of Ni³⁺ and Co³⁺ is still a controversial issue.

In order to investigate the relationship between catalytic performance and number of high state valence cations, we designed a series of spinel oxides catalysts by exposing different amounts of Ni³⁺ and Co³⁺ deposited on the surface by hydrothermal method and co-precipitation method. We also focused on preparation methods and the relationship between structure and catalytic properties of spinel oxides [32, 33]. Through hydrothermal synthesis, we successfully obtained multiporous spinel oxide nanosheets with high surface area. There are few reports of complete methane oxidation under 350 ℃ using porous spinel oxides nanosheets. At the same time, these novel porous nanosheets may also be utilized in other systems, such as an electrode material in pseudo capacitors and in oxygen evolution reduction (OER) [33-35].

All the chemicals were of analytical grade and used as received without further purification.

Co(NO₃)₂·6H₂O (10 mmol, 2.91 g), Ni(NO₃)₂·6H₂O (5 mmol, 1.45 g) and urea (100 mmol, 6 g) were physically mixed and completely dissolved in 70 mL deionized water for hydrothermal synthesis of NiCo₂O₄. The obtained solutions were transferred into 100 mL autoclaves with a Teflon liner at 120 ℃ and kept for 10, 20, 36, 60, and 72 h. The resulting product was centrifuged and washed with deionized water 3 times, was then dried overnight at room temperature, and was finally calcined in air at 350 ℃ for 2 h. For synthesis of Co₃O₄, Co(NO₃)₂·6H₂O (10 mmol, 2.91 g) and urea (100 mmol, 6 g) were dissolved in 70 mL deionized water. The rest of the synthetic procedure was similar to that used for synthesis of the NiCo₂O₄.

Co(NO₃)₂·6H₂O(10 mmol, 2.91 g) and Ni(NO₃)₂·6H₂O (5 mmol, 1.45 g) were dissolved in 20 ml deionized water under magnetic stirring for 1 h to synthesize NiCo₂O₄ by co-deposition precipitation. Then 50 mL NaOH (1 mol/L) was added to the solution under continuous stirring. The precipitate was collected by centrifuging and washing 3 times, was dried overnight at room temperature, and was finally calcined in air at 350 ℃ for 2 h. For synthesis of Co₃O₄, the above procedure was used without the added Ni(NO₃)₂·6H₂O.

The surface areas of the samples were determined from N₂ adsorption isotherms at 197 ℃ with the Brunauer-Emmett-Teller (BET) method using a Micromeritics ASAP 2010 analyzer. Prior to the measurement, the samples were outgassed at 300 ℃ for 2 h under a reduced pressure of 10^-5 Torr.

The crystal structure of the catalysts was examined using powder X-ray diffraction (XRD). XRD analysis was carried out on a Bruker D8 Advance with Cu K_α radiation source (λ = 1.5406 Å). A scan rate of 10°/min from 10° to 80° allowed us to obtain the XRD pattern. The crystal structure parameters of the samples were calculated using the MDI Jade 6.5.

The hydrogen temperature-programmed reduction (H₂-TPR) experiments were performed on a Micromeritics AutoChem 2920 instrument. Prior to reduction, the samples were pretreated at 300 ℃ for 1 h in an argon flow of 30 mL/min and then cooled down to room temperature. They were then reduced with 10% H₂/Ar until the temperature reached 500 ℃ at a rate of 3 ℃/min.

X-ray photoelectron spectroscopy (XPS) was performed using a VG Thermo ESCALAB 250 spectrometer. The instrument uses a focused monochromatic Al K_α X-ray (1486.7 eV) source operated at 120 W and outputs a 500 μm-diameter beam. The binding energy scale was calibrated using the known standard carbon C 1s line at 284.6 eV. Deconvolution of the XPS peak was performed with XPSPEAK.

Transmission electron microscopy (TEM) was performed using an FEI Tecnai G2 with an accelerating voltage of 200 kV. The lattice spacing and lattice planes of the samples were analyzed using DigitalMicrograph.

Methane combustion catalytic activity was measured using a conventional quartz tubular reactor (i.d., 6 mm; length, 300 mm) at atmospheric pressure. The reagent gas mixture (1 %CH₄/air vol%) was flowed over 100 mg catalysts at a gas hourly space velocity (GHSV) of 24000 mL g^-1h^-1. The reaction of methane combustion was stabilized for 20 min, and activity data were obtained at steady state conditions from 200 ℃ to 500 ℃, with temperature increased in increments of 25 ℃. The reactants and reaction products were analyzed by an on-line gas chromatograph. Kinetic studies of complete oxidation of methane were conducted in the kinetics control regime (conversion < 15%). The CCM reaction rates were calculated using Eq. (1):

(1)

X_f is the feed concentration of CH₄ and G is the gas flow rate (mL/h). C is the CH₄ conversion and W_cat is the weight of the spinel oxides catalysts (g). The number of (Co³⁺ + Ni³⁺) atoms of surface were calculated using Eq. (2):

(2)

Q is the (Ni³⁺ + Co³⁺) to (Ni + Co + O) ratio on the surface obtained from XPS, A_BET is the surface area (m²/g) determined from the BET method, and K is the number of atoms per unit area (atom/m²), which we can estimate from the solid surface model (see Figs. S1 and S2 in supporting information)

Fig. 1 shows the XRD patterns of the samples synthesized at different hydrothermal times. Obviously, the five patterns indicate pure NiCo₂O₄ because all the diffraction peaks can be indexed to NiCo₂O₄ and match well with a cubic crystal structure (JCPDS No. 73-1702). However, the differences, such as in lattice parameters and crystallite sizes, obtained from the Jade refinement of the XRD patterns are listed in Table 1. The lattice parameter decreased with hydrothermal time increasing up to in 60 h (8.093 Å). Until the hydrothermal time up to 72 h, the lattice parameter increased to 8.104 Å. The lattice parameter associated with ionic size and ionic radius of Co ³⁺ (0.055 nm) is larger than that of Ni³⁺ (0.053 nm), but is smaller than Ni²⁺ (0.069 nm) [3, 36]. Previous studies show that Ni ions occupy octahedral sites of the spinel structure, while Co ions occupy both tetrahedral and octahedral sites [37, 38], and we can reasonably infer that the large size of Ni²⁺ occupy octahedral sites. Thus, there is a very low amount of Ni³⁺ in the spinel structure at the beginning of the hydrothermal process. The amount of Ni³⁺ in the lattice cell increases with hydrothermal time, and the number of Ni³⁺ is maximum at 60 h hydrothermal time. The XRD patterns of the of NiCo₂O₄ and Co₃O₄ prepared by hydrothermal method and the co-precipitation method are presented in Fig. 2, and the lattice parameter and crystalline size also listed in Table 1. For 60 h hydrothermal NiCo₂O₄, a slight shift in the diffraction peak toward higher angles is observed in comparison to 60 h hydrothermal Co₃O₄. This result matched well with the spinel structure of Co₃O₄ (JCPDS No. 80-1542), resulting in smaller lattice parameter (8.093 Å) than that of 60 h hydrothermal Co₃O₄ (8.110 Å). Some Ni³⁺ ions insert themselves into the Co₃O₄ spinel structure, leading to a decrease in the lattice parameter. For co-precipitation of NiCo₂O₄, however, the diffraction peak shifted towards low angles compared to co-precipitated Co₃O₄, which matched with another spinel structure of Co₃O₄ (JCPDS 71-0816) with a larger lattice parameter (8.071 Å) than that of co-precipitated Co₃O₄ (8.064 Å). The reason is that large Ni²⁺ ions occupy the octahedral sites in the Co₃O₄ spinel structure. It is worthy to note that co-precipitation synthesized spinel oxides have relatively small crystalline cells due to the large number of defects in the crystalline cell. The crystal formed by the hydrothermal method tends to crystallize better and form fewer defects [39].

Obviously, the TEM pictures shown in Fig. 3, also demonstrated that 60 h hydrothermal spinel oxides had relatively smaller particles than co-precipitation spinel oxides. The crystalline sizes of 60 h hydrothermal NiCo₂O₄ and 60 h hydrothermal Co₃O₄ almost have the same about of 10-20 nm. There is a good correlation between crystalline size of the hydrothermal spinel oxide samples calculated by XRD and TEM. Generally speaking, the (220) interplanar lattice spacing in the NiCo₂O₄ spinel structure (JCPDS No. 73-1702) is 0.287 nm and (311) is 0.244 nm. However, lattice fringes with an interplanar lattice spacing of 0.281 nm corresponding to the (220) atomic planes and 0.238 nm corresponding to the (311) atomic planes were observed in HRTEM image (Fig. 3(b)), and all the lattice fringe spacings of the 60 h hydrothermal NiCo₂O₄ decrease. One possible cause is that the partial Co³⁺ ions are replaced by smaller Ni³⁺ ions. This indicates consistency with lattice contraction, which leads to a decrease in the lattice parameter by the calculated XRD results. However, there is no obvious change in the lattice fringe spacing of the 60 h hydrothermal Co₃O₄. Moreover, the selected area fast Fourier transform (FFT) confirms that 60 h hydrothermal NiCo₂O₄ and 60 h hydrothermal Co₃O₄ consist of the single-crystalline particles (insert of Fig. 3(b) and (d)). The angle labeled in the FFT pattern is 65°, which is in agreement with the theoretical angle between the (220) and (311) planes. The set of diffraction spots can be indexed as the (114), which also indicated dominant exposed planes of 60 h hydrothermal NiCo₂O₄ are (114). Meanwhile, the dominant exposed planes of the Co₃O₄ nanosheets are (112), which is consistent with Li's results [23]. From the above analysis of the exposed planes, we can build up a solid surface model (Fig. S1 and S2) with XPS and H₂-TPR results to understand the correlation between structure and catalytic activity.

Furthermore, the morphology of samples is clearly visible in the TEM images. Hydrothermal synthesis of spinel oxide nanosheets occurs by assembling by a large number of small particles, resulting in a disordered mesoporous structure, while co-precipitated samples have no definite morphology. SEM images of the 60 h hydrothermal NiCo₂O₄ (Fig. S3(a)) and Co₃O₄ (Fig. S3(b)) samples further confirm their sheet-shaped structures. Generally speaking, the nanosheet has a larger surface area than other nanoshapes. The BET surface area of the samples are summarized in Table 2. Obviously, the hydrothermal samples have larger surface area than the co-precipitated sample, and 60 h hydrothermal NiCo₂O₄ catalysts have the highest surface area among all samples.

To investigate the reducibility of the Ni and Co species in the spinel oxides catalysts, we performed H₂-TPR measurements at slow heating rate 3 ℃/min. Through slow heating, the Ni³⁺ and Co³⁺ peaks can be obviously distinguished in the H₂-TPR profiles. As shown in Fig. 4(c), there are two reduction peaks both in the 60 h hydrothermal and co-precipitated Co₃O₄, which correspond to stepwise reduction of Co₃O₄. The first peak position at 224 ℃ is related to the Co³⁺ to Co²⁺ transition, and the second is attributed to full reduction of Co²⁺ to Co⁰. There are a few peaks in the NiCo₂O₄ profile (Fig. 4(a)), which is different from others previous studies [40, 41]. In order to verify the first peak in the 60 h hydrothermal NiCo₂O₄ corresponding to Ni³⁺ reduction, we compared with NiO, where calcination at low temperatures (< 350 ℃) can form a few Ni³⁺ on the surface [42]. It is noteworthy that there are two peaks in the case of NiO. The large reduction peak at 224 ℃ overlaps with the Co³⁺ reduction peak, and a smaller peak at 169 ℃ is due to a small number of Ni³⁺ on the surface that reduce quickly compared to NiCo₂O₄. Thus, we confirmed that the first peak is attributed to the reduction of Ni³⁺ to Ni²⁺. Obviously, the reduction peak intensity at 177 ℃ for 60 h hydrothermal NiCo₂O₄ is greater than the intensity for co-precipitated NiCo₂O₄. This indicates that 60 h hydrothermal NiCo₂O₄ possesses a greater quantity of Ni³⁺. H₂-TPR spectra from hydrothermal NiCo₂O₄ samples are shown in Fig. S4. The results indicate that the 60 h hydrothermal NiCo₂O₄ contains the largest amount of Co²⁺ among all samples, and it has the largest amount of Ni³⁺. This agrees with the XRD and TEM results. In co-precipitated NiCo₂O₄ and Co₃O₄, the first reduction peak occurs earlier because there are a larger number of adsorbed oxygen species on the solid surface, resulting in very easy reduction, which is consistent with the observed O 1s XPS spectra.

We performed XPS to derive insights into the adsorbed species and metal oxidation states of catalysts, which is an effective technique to gain surface information [43]. Fig. 5(c) illustrates the XPS results from Ni 2p_3/2 for 60 h hydrothermal NiCo₂O₄ and co-precipitated NiCo₂O₄ (other hydrothermal samples are shown in Fig. S5). The coexistence of Ni(Ⅱ) and Ni(Ⅲ) in 60 h hydrothermal and co-precipitated NiCo₂O₄ catalysts is indicated by the shoulder observed on the main peak at 853.6 eV, which is assigned to the Ni(Ⅱ) species. The higher binding energy (BE) peak at 861.0 eV and its shakeup satellites are due to overlapping Ni(Ⅱ) and Ni(Ⅲ) satellites. The surface Ni³⁺/(Ni + Co + O) molar ratio in 60 h hydrothermal NiCo₂O₄ is approximately 8.22%, which is the largest among all samples. The lowest ratio occurs in the co-precipitated NiCo₂O₄ with 3.82% (Table 2). This result agrees well with the H₂-TPR results. Co 2p XPS spectra of four spinel oxide samples are shown in Fig. 5(a). One should note that the Co³⁺/(Ni + Co + O) ratio for 60 h hydrothermal Co₃O₄ is the highest in all samples, and the value in the co-precipitated NiCo₂O₄ sample is higher than in the co-precipitated Co₃O₄ sample, which agrees well with results reported by others [31].

In the O 1s XPS spectra (Fig. 5(b)), the broad peaks can be decomposed into four major components by deconvolution. These peaks correspond to lattice oxygen, a surface OH group, defective oxygen, and adsorbed moisture. Quantitative analysis in Fig. 5(d) reveals the lattice oxygen percentage obviously increases for hydrothermal synthesis of spinel oxides compared to the co-precipitation method, and co-precipitated spinel oxides have more adsorbed oxygen species on the surface. Generally speaking, surface oxygen is very important in catalytic oxidation because surface oxygen is more reactive than gaseous oxygen [44]. As is well known, the adsorbed oxygen species increase with increasing oxygen pressure. Therefore, in order to investigate the effect of adsorbed oxygen species on the catalytic reaction, we measured the CH₄ reaction rate under different oxygen pressures, as shown in Fig. 6(a). It suggests that the reaction rate is unaffected by O₂ pressure when O₂ pressure is greater than 15 kPa, which means adsorbed oxygen species are not crucial for the whole catalytic system under catalytic reaction conditions (O₂ pressure about 20 kPa). Moreover, the CH₄ reaction rate is proportional to the CH₄ pressure shown in Fig. 6(b), which is consistent with previous findings that activation and dissociation of the C-H bond is the rate-limiting step [28, 44].

Fig. 7(a) shows the light-off curves for methane combustion in samples with various hydrothermal synthesis times. We can conclude that the catalytic activity improves as hydrothermal time increases, and 60h hydrothermal synthesis time resulted in 100% methane conversion at 340 ℃. From the catalytic activity and XRD results, we can reasonably infer that exposing many high valence state ions on the surfaces of spinel oxide can be attributed to a proper hydrothermal synthesis time, which results in more active centers for CCM. Moreover, as can be seen from Fig. 7(b), the hydrothermal and co-precipitation syntheses of spinel oxides show different light-off curves during CCM. The catalytic performance of hydrothermally synthesized spinel oxides is much better than their co-precipitated counterparts. The co-precipitated sample containing NiCo₂O₄ performs better than co-precipitated Co₃O₄. This result agrees well with previous studies, where results show that the concentration of surface Ni³⁺ is responsible for high catalytic performance [28]. However, it also should be noted that 60 h hydrothermal NiCo₂O₄ catalytic activity corresponds closely to 60 h hydrothermal Co₃O₄ with 90% methane conversion at 325 ℃, which is different from previous reports [3, 8, 28]. In order to understand the difference, we compared 60 h hydrothermal NiCo₂O₄ with 60 h hydrothermal Co₃O₄ at different GHSV values (Fig. 7(c)). Clearly, when working at relatively high GHSV (e.g., 84000 mL h^-1 g^-1), hydrothermal NiCo₂O₄ catalytic performance is higher than that of Co₃O₄. We try to use previous studies on Co³⁺ or Ni³⁺ as an active center for CCM to explain our experimental results. If Co³⁺ ions of the spinel oxides surface were active centers, 60 h hydrothermal Co₃O₄ and co-precipitated Co₃O₄ both would have better catalytic performance than NiCo₂O₄. If the active sites were Ni³⁺, co-precipitated NiCo₂O₄ would be superior to 60 h hydrothermal Co₃O₄. Therefore, we infer that Co³⁺ and Ni³⁺ maybe both are active sites in the spinel structure. Then we calculated the (Ni³⁺ + Co³⁺) to (Ni + Co + O) ratio on the surface from the XPS spectra. Table 2 clearly indicates that the (Ni³⁺ + Co³⁺) to (Ni + Co + O) ratio in the 60 h hydrothermal Co₃O₄ is the highest among all samples. It seems that 60 h hydrothermal Co₃O₄ would have the best catalytic activity. However, this was not the case. While we take the surface area and crystal plane into account to calculate the number of (Ni³⁺ + Co³⁺) ions on the surface, the methane combustion light-off curves give a relatively rational explanation. As indicated in Fig. 8(b), as the (Ni³⁺ + Co³⁺) content increases on the surface, the reaction rate also increases. Thus, this is direct evidence that Ni³⁺ and Co³⁺ both serve as active centers on the surface of spinel oxides. The apparent activation energy of all the samples are very similar in Fig. 8(a), which indicate they have similar activation modes and the same active center [45]. This indirect evidence also supports our conclusion that Ni³⁺ and Co³⁺ both serve as active centers.

Our catalysts are compared with some previously reported catalysts, including noble metal catalysts, as summarized in Table 3. To examine the catalytic performance of NiCo₂O₄ in the complete oxidation of CH₄ from the exhaust of a natural gas engine, we measured the catalytic performance of complete CH₄ oxidation on NiCo₂O₄ in two different gas mixtures (Fig. 9). First, 1% CH₄ and 10% H₂O balanced with air at a flow rate 40 mL/min (Fig. 9(a)) is used to evaluate the influence of H₂O in the catalytic system. It is clear that the addition of H₂O causes the T₉₀ to lag behind about 50 ℃ compared to the case without H₂O. We also noted that the initial activity starts at a low temperature of about 200 ℃, which indicates the added H₂O will not cover the active sites and maybe only take more heat from the catalyst surface, resulting in a slight increase in complete oxidation temperature at relatively low GHSV. Second, results with 0.2% CH₄, 5% O₂, 10% H₂O, and 15% CO₂ balanced with N₂ at high GHSV (60000 mL h^-1 g^-1) (Fig. 9(b)) show that H₂O has a minor impact on catalytic performance, especially for 60 h hydrothermal Co₃O₄, and CH₄ can be completely oxidized to CO₂ and H₂O at 450 ℃. Generally, H₂O is easily absorbed on the catalyst surface, which tends to reduce catalytic activity. However, it is worth noting that the activity of 60 h hydrothermal NiCo₂O₄ and Co₃O₄, the first C-H cleavage remains at 200 ℃, which means H₂O cannot completely cover the active sites. On the other hand, our catalysts prepared by long hydrothermal treatments showed a very low content of adsorbed moisture on the catalyst surface, as indicated by the O 1s XPS results. This means that H₂O is weakly adsorbed on the catalyst surface, especially at high temperature. These gas compositions are typically used in the evaluation of catalytic performance of Pd- and Pt-based catalysts for complete oxidation of CH₄ in the exhaust of natural gas engines [46-48]. Thus, it is evident that 60 h hydrothermal spinel oxide catalysts show higher activity than some noble metal catalysts.

Furthermore, the stability performance shown in Fig. 7(d) indicates that both 60 h hydrothermal NiCo₂O₄ and Co₃O₄ are able to completely oxidize CH₄ to CO₂, and catalytic activity does not obviously decline for 12 h under feed gas including 10 vol% H₂O at 425 ℃ (GHSV = 24000 mL h^-1 g^-1). On the other hand, our catalysts were repeatedly used several times at different GHSV and different feed gas compositions, which also exhibit good stability. We can confirm that the stability of spinel oxides improves as the extensive hydrothermal time results in highly crystalline structures. This is supported by the appearance of a larger lattice oxygen peak in the O 1s XPS spectra.

The correlation between structure and catalytic activity is established for 60 h hydrothermal spinel oxides by combining the activity measurements with the characterization results. As observed in TEM, the 60 h hydrothermal spinel oxides investigated here both exhibit disordered nanosheet structure due to their disordered assembly from small nanoparticles. 60 h hydrothermal NiCo₂O₄ have higher surface area as confirmed with BET measurements. In general, a high surface area often results in greater catalytic activity because more active sites are exposed during the catalytic reaction [49]. For hydrothermal synthesis of spinel oxides, both have high valence ion content on the surface based on analysis of the XPS, H₂-TPR, XRD, and TEM results, as well as the solid surface model (Fig. S1 and S2), which leads to excellent catalytic performance. It can be concluded that a strong correlation exists between activity and structure and, more significantly, that a high concentration of valence ions and high surface area are responsible for high methane combustion activity.

In summary, we synthesized a series of spinel oxide catalysts with different amounts of Ni³⁺ and Co³⁺ cations by hydrothermal and co-precipitation methods. As the amount of (Ni³⁺ + Co³⁺) on the surface increases, the catalytic activity for methane combustion also increases. The (Ni³⁺ + Co³⁺) content on the 60 h hydrothermal NiCo₂O₄ surface is the highest among all samples, thus it exhibits the best catalytic performance and results in complete CH₄ conversion at 340 ℃. Therefore, we are convinced that Co³⁺ and Ni³⁺ can both function as active sites on the surface of spinel oxides. Moreover, multiporous spinel oxide nanosheets without any surfactants and templates were obtained from hydrothermal synthesis and were nearly unaffected by water in low-temperature methane combustion. This occurred because the long hydrothermal treatment results in less adsorbed oxygen on spinel oxides. Finally, multiporous nanosheet spinel oxides with high surface areas and our conclusion that both Co³⁺ and Ni³⁺ function as active sites may be applicable in other catalytic oxidation systems in the future.