Methane, the main component of natural gas and a well-known greenhouse gas (with 20 times the greenhouse effect of CO₂), has been widely used in power generation and other industrial heating applications. The incomplete combustion of natural gas not only wastes energy but also produces atmospheric pollution such as CO, NO_x, and unburned hydrocarbons. Compared with standard combustion, the catalytic combustion of CH₄ may provide an environmentally-friendly method of generating power with greater efficiency and lower emissions of air pollutants [1-4]. Because of the highly symmetrical structure of CH₄, it is challenging to decrease the complete oxidation temperature of this compound [5-7]. However, compared with transition metal oxide catalysts, supported Pd catalysts have exhibited better performance in this regard [8, 9].

It is known that the Pd particle size and chemical state as well as the support characteristics all play important roles in promoting CH₄catalytic combustion [10-12]. Hicks et al. [13, 14] determined that the activity of small PdO particles was almost two orders of magnitude higher than that of large Pd crystallites, and that the respective turnover frequency values of these materials were 0.02 and 1.3 s^-1, respectively. Although PdO is regarded as active for CH₄ combustion [15, 16], there are still contradictory data regarding this catalyst. As an example, Oh and coworkers [17, 18] found that both reduced Pd species in the form of Pd^δ+(0 < δ < 2) and high valence Pd⁴⁺ atoms [19] are active during CH₄ catalytic combustion.

Porous supports with high specific surface areas tend to improve the dispersion of active metal species [20, 21], while a support's acidity and redox properties have a more pronounced effect in terms of promoting CH₄ combustion. Lou et al. [22] confirmed that the surface Br nsted acid sites on H-ZSM-5 were able to anchor Pd species and modulate the morphologies of Pd structures. Cargnello et al. [23] designed a Pd@CeO₂/Al₂O₃ catalyst in which the close contract between Ce and Pd maximized the transformation of metallic Pd to PdO_x. SnO₂, which exhibited high oxygen transfer capability and thermal stability [24-26], was considered a good candidate for a support material. Sekizawa et al. [27-30] showed that the Pd/SnO₂ on SnO₂ calcined at 1200 ℃ exhibited excellent catalytic activity during methane oxidation, allowing complete oxidation at 440 ℃. Bratan et al. [31] also found that the oxygen storage capacity and the release properties of a SnO₂/TiO₂ support enhanced the mobility of lattice oxygen.

Improving the interactions between noble metal nanoparticles and supports can avoid the sintering and decomposition of the noble metal. Zou et al. [32] prepared well-crystallized PdO nanoparticles by epitaxial growth over NiAl₂O₄, thus avoiding particle agglomeration and obtaining a good distribution. Matching of lattice geometry was observed in trials with Pt/MgAl₂O₄, and this work also demonstrated that strong electronic attractions at the interface between the support and the epitaxial metal oxide stabilize the Pt particles [33, 34]. The above results prompted our own group to study Pd/SnO₂.

In the present work, a Pd/SnO₂ catalyst was prepared by the precipitation method and the effects of the support calcination temperature on CH₄ combustion were investigated. N₂ adsorption-desorption, X-ray diffraction (XRD), inductively coupled plasma atomic emission spectroscopy (ICP-AES), O₂ temperature-programmed desorption (O₂-TPD), CH₄ temperature-programmed surface reaction (CH₄-TPSR) and X-ray photoelectron spectroscopy (XPS) were all used to characterize the physicochemical properties of the samples. By studying the structure of these Pd/SnO₂ specimens, we hoped to understand the role of lattice matching between the noble metal and the support in promoting CH₄ combustion.

The SnO₂ support was synthesized by a chemical precipitation method. SnCl₄·5H₂O (99.7%, Sinopharm Chemical Reagent Co., Ltd.) and NaNO₃(99.7%, Aladdin) were dissolved in deionized water to form a mixed solution that was subsequently heated to 150 ℃ and refluxed for 16 h with vigorous stirring. The resulting white precipitate was filtered and washed with deionized water to remove chloride ions. After filtering, the solid powder was dried at 120℃ for 12 h and calcined at 600, 800, 1000 or 1200 ℃ for 4 h in air. The obtained SnO₂ powders are denoted herein as SnO₂(x), where x represents the calcination temperature.

Supported Pd catalysts were prepared by impregnating these SnO₂(x) powders with a CH₂Cl₂ (99.7%, Sinopharm Chemical Reagent Co., Ltd.) solution of Pd(CH₃COO)₂(Pd ≥ 47.32%, Kunming Institute of Precious Metals). After impregnation, the catalysts were dried and calcined in air at 600℃ for 4 h. The resulting catalysts are denoted as Pd/SnO₂(x), where x represents the calcination temperature applied to the SnO₂. The final Pd loadings were determined by ICP-AES.

The catalytic combustion of CH₄ over these supported Pd catalysts was carried out in a fixed-bed reactor at atmospheric pressure, using 200 mg catalyst samples (40-60 mesh) diluted with 1 g silica sand (20-40 mesh). A feed gas containing 1 vol% CH₄, 20 vol% O₂ and N₂as the balance flowed through the catalytic bed at a rate of 30 mL/min. The temperature of the catalyst bed was measured using a thermocouple inserted at the top of the bed. The bed heating rate was 4 ℃/min. An online gas chromatograph (GC, Agilent 7890) was used to monitor the composition of the outlet gas. The catalytic activity is expressed herein by the T₁₀ and T₁₀₀ values for CH₄, corresponding to the reaction temperatures required for 10% CH₄ and 100% CH₄ conversion, respectively.

The specific surface areas of the catalysts were determined from N₂ adsorption isotherms at -196 ℃ using a Micromeritics ASAP 2020 analyzer. The powder XRD patterns of the samples were obtained with a Rigaku D/max 2550 VB/PC diffractometer with Cu K_α radiation operating at 40 kV and 40 mA. The intensity data were collected at room temperature over a 2θ range from 10° to 70° at a scan rate of 6°/min. The average crystallite sizes of samples were calculated using the Debye-Scherrer formula based on the diffraction peaks. Transmission electron microscopy (TEM) images were obtained with a JOEL 2100 instrument operating at 200 kV. XPS was performed with an AXIS-Ultra-DLD spectrometer incorporating a monochromated Al K_α X-ray source (1486.6 eV). The base pressure inside the analysis chamber was 3×10^-11 kPa and the XPS spectra of selected elements were acquired at a constant analyzer pass energy of 40 eV. All binding energy (BE) values were determined relative to the adventitious C 1s peak (BE = 284.8 eV).

CO chemisorption measurements were performed using a pulse method with a chemisorption analyzer (Micromeritics Autochem 2920). The catalysts were pre-reduced under 10% H₂ in Ar at 150 ℃ for 1 h prior to testing, and the average CO-to-Pd molar ratio was assumed to be 1 when calculating the Pd dispersion [35].

O₂-TPD was carried out in a quartz flow microreactor using a chemisorption analyzer (Micromeritics Autochem 2920) equipped with an HPR-20 QIC mass spectrometer (MS). In these trials, a 200 mg quantity of the catalyst was pretreated under 3 vol% O₂ in He (50 mL/min) at 600 ℃ for 30 min, then cooled to room temperature and allowed to absorb O₂ for 1 h. After subsequent purging with He for 1 h, the sample was heated at a rate of 10 ℃/min from 30 to 850 ℃ and the desorbed O₂ signal (M_r/z = 32) was detected by the MS.

CH₄-TPSR was performed with the same reaction apparatus as used for O₂-TPD. In these trials, a 100 mg quantity of the catalyst was pretreated at 400 ℃ for 40 min in a 50 mL/min flow of 3 vol% O₂ in He, after which the catalyst was cooled to room temperature. The TPSR data were acquired by heating the catalyst from 30 to 650 ℃ at a rate of 10 ℃/min in 50 mL/min flow of 5 vol% CH₄ in He. The release of CH₄ (M_r/z = 16) and CO₂ (M_r/z = 44) was determined with an online MS (HPR-20 QIC Benchtop Gas Analysis System).

The catalytic activities of the SnO₂ and Pd/SnO₂ during CH₄ combustion are summarized in Fig. 1. The activity of the SnO₂ gradually decreased with increasing calcination temperature from 600 to 1200℃ (Fig. 1(a)). The SnO₂(600) exhibited the highest catalytic activity, giving T₁₀ and T₁₀₀ values of 410 and 580℃, respectively. In contrast, the T₁₀ and T₁₀₀ values obtained using the SnO₂(1200) were 490 and 736℃.

The deposition of Pd significantly improved the CH₄ combustion, as shown in Fig. 1(b), such that the T₁₀₀ for Pd/SnO₂(600) was 140 ℃ lower than that for SnO₂(600). Increasing the calcination temperature also lowered the temperature required for complete CH₄combustion. The T₁₀₀ for Pd/SnO₂(1200) was relatively low at 390℃ while the value for Pd/SnO₂(600) was 440 ℃. A stability test was also conducted and the results are presented in the insert to Fig. 1(b), showing that there was no obvious decline in activity even after 80 h at 370℃.

The specific surface areas of the samples varied with the calcination temperature, and so this effect was accounted for by calculating the relative reaction rate of CH₄, giving the results presented in Table 1. With increases in the calcination temperature, the reaction rate over Pd/SnO₂ can be seen to have increased. The relative reaction rate of CH₄ over Pd/SnO₂(1200) was 36 times that over Pd/SnO₂(600) at 300 ℃. These data also demonstrate that there was no obvious difference in the amount of Pd loading between samples, meaning that the observed variations in activity required further study.

The specific surface areas of samples are also provided in Table 1. The Brunauer Emmett Teller (BET) surface area of the SnO₂ decreased as expected with increasing calcination temperature from 600 to 1000 ℃, and the surface area of the Pd/SnO₂ presented the same trend. Thus, the Pd/SnO₂(600) had the highest surface area (19.8 m²/g). Temperatures above 800℃ did not affect the surface area, such that the surface area of the Pd/SnO₂(1200) was almost equal to that of the Pd/SnO₂(1000).

The XRD patterns of the SnO₂(x) and Pd/SnO₂(x) are shown in Fig. 2. The SnO₂ exhibits a classic rutile phase(JCPDS card 411445), and the SnO₂diffraction peaks intensify with increasing calcination temperature, accompanied by increases in the crystallite size from 12.6 to 72.3 nm (Table 1). The Pd/SnO₂ did not generate any Pd diffractions, indicating either that the Pd species were highly dispersed on the SnO₂or that the Pd concentration was too low to be detected by XRD [37, 38].

Compared with the XRD pattern of SnO₂(x), a slight shift to higher 2θ values was evident in the case of the Pd/SnO₂(600) and Pd/SnO₂(800) (Fig. 2(b)) as the result of lattice shrinkage. This effect is due to the insertion of smaller Pd ions into the SnO₂ lattice. At calcination temperatures above 1000 ℃, there was only a minimal shift in the diffraction peak positions. Higher calcination temperatures thus also increased the Pd particle size, in agreement with the CO adsorption analysis.

HRTEM images of the Pd/SnO₂(x) specimens are presented in Fig. 3. These images demonstrate that the average Pd particle size on the Pd/SnO₂(600) surface was 4 nm. As the support calcination temperature was raised from 600 to 1200 ℃, the Pd particle size increased from 4 to 8 nm. Both PdO and SnO₂lattice fringes are clearly seen in the HRTEM image of Pd/SnO₂(1000) in Fig. 3(b). The interplanar lattice spacings for PdO and SnO₂ are 0.2641 and 0.2638 nm, respectively, corresponding to the PdO(101) and SnO₂(101) lattice planes.

XPS was used to characterize the surface compositions and chemical states of the catalysts. The surface Pd concentrations on the Pd/SnO₂(x) samples are summarized in Table 1, while the Pd 3d XPS data are shown in Fig. 4. The presence of Pd species on the Pd/SnO₂(600) in two states is evident from the peaks at 336.8 and 338.1 eV, corresponding to Pd²⁺ and Pd⁴⁺, respectively [39-41], with a Pd²⁺/Pd⁴⁺ atomic ratio of 2.3 (Table 1). With increases in the calcination temperature, this ratio decreases in the order Pd/SnO₂(1200) > Pd/SnO₂(1000) > Pd/SnO₂(800) > Pd/SnO₂(600). These data are in accordance with the activity results shown in Fig. 1(b).

The XRD patterns in Fig. 2 demonstrate lattice shrinkage of the Pd/SnO₂(600) and Pd/SnO₂(800). The radius of Pd⁴⁺(61.5 pm) is smaller than that of Sn⁴⁺(69 pm), and so Pd⁴⁺ readily inserts into the SnO₂ lattice to generate a Sn_1-xPd_xO₂ solid solution during the preparation process. This results in high valence Pd⁴⁺ ions in the lattice [42, 43]. In contrast, in the case of the catalyst prepared from the support calcined at high temperature ( > 1000 ℃), there is no angle shift because Pd²⁺(86 pm) is too large to enter into the lattice.

Although the Pd loading amounts were almost equal, different surface Pd concentrations were determined by XPS analysis. Increases in the calcination temperature raised the Pd concentration from 0.76% to 1.56%. Considering the XRD and TEM results, this result can be explained. Varying the calcination temperature generates Pd species in different forms. The catalyst made from the supports calcined at low temperatures ( < 800 ℃) contained agglomerated Pd particles, thus reducing the surface Pd concentration. In contrast, higher temperature calcination resulted in elevated Pd concentrations.

O₂ activation is a crucial factor in redox reactions, and the bonding strength of oxygen species on the catalyst surface determines the catalytic activity during deep oxidation [44, 45]. In this work, the relationship between oxygen activation and the Pd/SnO₂structure was investigated by obtaining O₂-TPD data, as shown in Fig. 5. A broad O₂ desorption peak ranging from 540 to 680 ℃ was generated by each of the samples. The peak portion at lower temperatures (540-640 ℃) is assigned to the decomposition of oxygen species on PdO_x surfaces [46], while the portion in the range of 640-680 ℃ is attributed to lattice oxygen species driven from the PdO bulk phase [16, 47].

With increases in the calcination temperature, the O₂ desorption peak shifts to lower temperatures, which is in agreement with changes observed in the size of the PdO particles. Small PdO particles on the support tend to have stronger interactions, which shifts the O₂ desorption peak to higher temperature, and vice versa.

The nature of the active sites can be assessed based on the quantity of sites and their intrinsic activity [48]. The O₂ consumption amounts per unit Pd surface were obtained based on the areas of the O₂ desorption peaks, setting the desorption peak area for Pd/SnO₂(1200) to 1.00. These values are shown in Table 1. The O₂ desorption amount of the Pd/SnO₂(1200) was found to be three times that of the Pd/SnO₂(600). The larger area of the O₂ desorption peak generated by the former material implies a greater concentration of included oxygen species compared with the other catalysts.

To further examine the performance of the PdO species during the CH₄ reaction, CH₄-TPSR was carried out and the results are shown in Fig. 6. A CH₄ consumption peak was generated in the temperature range of 200 to 250 ℃ and the production of CO₂ was detected simultaneously. Low temperature CH₄ consumption peaks ( < 400℃) are typically attributed to the reaction between active oxygen species and CH₄, and so reflect the reduction of PdO species [49].

According to Carstens et al. [50], the crystalline form of PdO is more easily reduced than the amorphous forms. The HRTEM analysis (Fig. 3(b)) determined that the catalysts calcined at high temperatures ( > 1000℃) contained highly crystalline PdO (Fig. 3(b)). Thus, it is not difficult to understand why the O₂ desorption temperatures in the O₂-TPD profiles are in the same order as the reduction temperatures.

The activity data presented above show that the sequences of the CH₄ reduction temperatures and O₂ desorption temperatures both coincide with that of the activity during methane combustion.

The activities of unit Pd atoms on the surfaces of the catalysts during CH₄-TPSR were characterized by calculating the amounts of CH₄ consumed per unit Pd atom surface area, with the results shown in Table 1. It is evident that the CH₄ consumption order follows the same trend as the O₂ desorption data, although the CH₄ consumption of the Pd/SnO₂(1200) was 45% higher than that of the Pd/SnO₂(600).

The support calcination temperature had a pronounced effect on the progress of the CH₄ catalytic combustion. The Pd/SnO₂(1200) exhibited the highest activity, and the reaction rate during CH₄ combustion over this material was 36 times that over the Pd/SnO₂(600) at 300 ℃. Although there were differences in the Pd surface concentrations and chemical states between these two materials, it is difficult to explain the significant activity gap between them. The XRD and XPS data suggest that the catalysts made from the supports calcined at low temperatures ( < 800 ℃) contained some Pd species adsorbed in the support because of the similar diameters of Pd⁴⁺ and Sn⁴⁺, while some small, amorphous Pd particles were present on the support surface. In contrast, the catalysts fabricated using the supports calcined at high temperatures ( > 800 ℃) showed Pd species only on the catalyst surface, in the form of highly crystalline PdO.

The SnO₂(101) and PdO(101) lattice fringes were found to exhibit close spacing during HRTEM observations. A specific spinel oxide crystal plane family is known to interact with PdO via epitaxy [51]. In addition, intimate contact (with a small contact angle) between the PdO(110) and NiAl₂O₄(400) planes in Pd/0.5NiO/Al₂O₃ has been shown to improve the performance of the catalyst and inhibit the migration and growth of Pd particles over time [32]. This same study determined that intimate contact lowers both the O₂-TPD temperature and activation energy. Cai's group [52] found that oxygen vacancies in the CeO₂ lattice of Pd-Ce/ZSM-5 provide a pathway for oxygen atoms to easily approach or leave from the Pd particles, implying significant oxygen transfer from the support to the metal. Martin and Duprez [53] also demonstrated exceptional mobility of both surface and bulk oxygen over CeO₂, and found that the presence of metal particles on the CeO₂ markedly increases the oxygen mobility between 300 and 450 ℃.

Based on these literature data and our own experimental results, we suggest that the close spacing of the SnO₂(101) and PdO(101) lattice fringes is the main factor explaining the low O₂-TPD temperature and the high amount of O₂ desorption associated with the Pd/SnO₂(1200). The matching lattice geometries of the PdO and SnO₂ allows oxygen at the Pd-SnO₂ interface to be activated, following which the Pd acts as an oxygen pump to drive CH₄ combustion, similar to the back-spillover of oxygen from CeO₂ to oxygen vacancy sites on the PdO/Pd surface [54-56]. Therefore, we can reasonably conclude that the close spacing of the SnO₂(101) and PdO(101) lattice fringes also explains the remarkably enhanced CH₄ combustion activity of the Pd/SnO₂(1200).

The support calcination temperature plays a key role in the behavior of Pd/SnO₂ during CH₄ catalytic combustion. A catalyst prepared from SnO₂ calcined at 1200 ℃ exhibited excellent activity and stability; the reaction rate of CH₄ on the Pd/SnO₂(1200) specimen was 35 times higher than that over the Pd/SnO₂(600), and the former remained stable over an 80 h test. Lattice matching based on the geometry of PdO and SnO₂ facilitates oxygen activation from the SnO₂ to vacant oxygen sites on the PdO/Pd surface via the back-spillover of oxygen mechanism, which promotes CH₄ catalytic combustion.