Studies on activation factors for oxidative coupling of methane over lithium-based silicate/germanate catalysts

Abstract

Several oxides containing group 14 elements were characterized to investigate catalytic activity factors and the reaction mechanism for oxidative coupling of methane (OCM). Lithium-based silicate Li₂SiO₃ and germanate Li₂GeO₃ showed much higher selectivities for C₂ hydrocarbons than their Sn-based counterpart Li₂SnO₃. Our XPS study suggested that strong metal–oxygen bonds in silicates and germanates could be of importance to effectively suppress the deep oxidation of methane at the catalyst surface, resulting in enhanced C₂ selectivities. In contrast, weakly bound surface oxygen of Sn-based oxides is subject to direct oxygenation of methane to form CO_x as by-products, emphasizing the importance of metal–oxygen bonding strength. To achieve high OCM catalytic activity, effective methane activation by strong basicity is also required, and the local environment around oxygen sites could be a key factor to enhance CH₄ conversion. Based on the chemical/structural viewpoints, we found that silicate and germanate with higher Li contents, Li₄SiO₄ and Li₄GeO₄, are significantly OCM active, comparable to the previously reported catalyst Li₂CaSiO₄.

---

1. Introduction

Nowadays, fossil fuels such as coal and crude oil have been mainly used for energy generation and chemical production. There is, however, a concern about the depletion of natural carbon resources, and the development of sustainable alternative processes will be necessary for future industries.¹ The effective utilization of methane as a primary component of natural gases is thus noteworthy.² Nevertheless, methane is the most stable and inactive alkane with no functional groups, limiting its use mostly to simple combustions.³ The oxidative coupling of methane (OCM) is fascinating because this reaction can directly convert methane into ethylene, one of the essential chemicals for manufacturing various chemical products. The OCM process to C₂ hydrocarbons proceeds viaeqn (1) and (2), but non-selective reactions, represented by eqn (3) and (4), occur simultaneously:

CH₄ + 1/4O₂ → 1/2C₂H₆ + 1/2H₂O ΔH = −89 kJ mol⁻¹ (1)

CH₄ + 1/2O₂ → 1/2C₂H₄ + H₂O ΔH = −141 kJ mol⁻¹ (2)

CH₄ + 3/2O₂ → CO + 2H₂O ΔH = −608 kJ mol⁻¹ (3)

CH₄ + 2O₂ → CO₂ + 2H₂O ΔH = −890 kJ mol⁻¹ (4)

Since the initial report by Keller and Bhasin in the early 1980s,⁴ approximately 2300 publications on OCM have been published. Among various catalysts investigated so far, a composite consisting of manganese oxides and sodium tungstates on a silica support (Mn–Na₂WO₄/SiO₂) is known to be one of the most active OCM catalysts.^3,5 Kiani and coworkers recently reported experimental evidence for Na-coordinated WO₄ surface sites acting as active centers.^6–8 However, the chemical/structural complexities of Mn–Na₂WO₄/SiO₂ prevent the establishment of catalytic design concepts towards developing practical OCM catalysts with sufficient activity and durability.

Previous mechanistic studies on OCM^9–20 indicated that the formation of C₂ hydrocarbons predominantly involves a gas-phase association of methyl radicals (·CH₃) generated at surface oxygen sites of the catalyst. Therefore, the methane activation by the catalyst is suggested as a dominating factor of the OCM activity, and it is crucial to trigger the hydrogen abstraction reaction effectively while suppressing side reactions. Several groups have claimed the importance of the electronic state of oxygen species on the catalyst surface.^15,21–24 However, there is no concrete view about the role of these oxygen species, and the activation factors of OCM at the catalyst surface are not completely understood.

Our recent study achieved the discovery of a highly active catalyst Li₂CaSiO₄ with a maximum C₂ yield of 26% (800 °C, CH₄/O₂ = 2/1) in a single-phase form, which essentially differs from conventional catalysts with chemical and structural complexities.²⁵ It should be noted that the C₂ yield of Li₂CaSiO₄ is comparable to or even higher than that of the known active catalyst Mn–Na₂WO₄/SiO₂ under the same reaction condition (Table S1†). The fact that Li₂CaSiO₄ exhibits chemical stability even in long-term catalytic tests ensures that any decomposition products are not likely reactive centers, but the excellent OCM activity stems from bulk nature or a specific surface structure of Li₂CaSiO₄. Indeed, the compound essentially acts as a single phase, facilitating mechanistic studies on OCM catalysts.

Statistical analysis of the past catalytic data suggested that oxides containing strongly basic elements such as lithium tend to be high-performance catalysts.²⁶ However, the basicity alone seems insufficient to enhance both the CH₄ conversion and C₂ selectivity, suggesting that the coexistence of silicon plays an important role in the catalytic performance of Li₂CaSiO₄.²⁵ In the present work, fundamental characterization was conducted on several oxides containing group 14 elements, including Li₂MO₃ (M = Si, Ge, Sn; Fig. S1 of the ESI†), Li₄SiO₄, and Li₄GeO₄, to investigate the effects of metal–oxygen bonding strength on the C₂ selectivity. Furthermore, based on these findings, possible activation factors and the reaction mechanism are discussed.

Experimental section

Synthesis. Li₂MO₃ (M = Si, Ge, Sn), Li₄SiO₄, and Li₄GeO₄ catalysts were synthesized by a solid-state reaction route. MO₂ (99.9%, KOJUNDO CHEMICAL LABORATORY CO., LTD.) and Li₂CO₃ (99.99%, KOJUNDO CHEMICAL LABORATORY CO., LTD.) were used as starting materials. Mixtures of the reagents with 2% excess of lithium added to the stoichiometric amount were calcined at 600 °C for 12 hours in air. The calcined powders were then fired at 850 °C to 1100 °C for 12 hours in air.

Characterization. The phase purity of the resultant catalysts was checked by means of X-ray powder diffraction (XRD; Rigaku Ultima IV; Cu Kα radiation). The XRD data indicated that these catalysts are of phase pure (Fig. S1 and S19†). The specific surface areas were estimated with the Brunauer–Emmett–Teller (BET) method employing nitrogen gas adsorption/desorption isotherms at 77 K (Microtrac BEL, BELSORP-mini II). To investigate the basicity of the catalysts, CO₂ temperature-programmed desorption measurements (CO₂-TPD) were performed between 100 °C and 800 °C using a catalyst analyzer (Microtrac BEL, BELCAT-B). In the CO₂-TPD experiments, the sample (0.1 g) was heated to 800 °C in a He atmosphere and kept for 1 hour to remove the adsorbed gas. Next, the sample was cooled to 100 °C and exposed to a CO₂ flow (99.99%) for 1 h to saturate the surface completely. After purging with ultrahigh purity He for 30 minutes at 100 °C to make the baseline flat, a TPD curve was obtained from 100 to 800 °C (heating rate: 10 °C min⁻¹). X-ray photoelectron spectroscopy measurements (XPS; JEOL JPM-9010MC) were performed at room temperature using a monochromatic Mg Kα radiation. The binding energies of the spectra were calibrated with the C 1s peak of graphite at 284.5 eV. For each catalyst, the surface was Ar-etched for 10 seconds with an accelerating voltage of 400 V and an emission current of 6.4 mA. To identify loosely or tightly bound oxygen species, O₂ temperature-programmed desorption (O₂-TPD) was measured between 100 °C and 800 °C using a catalyst analyzer (Microtrac BEL, BELCAT-B). The sample (0.1 g) was heated to 800 °C in a flowing 5% O₂/He gas mixture, kept for 1 hour to remove the adsorbed gas, and then cooled to 100 °C. After purging with ultrahigh purity He for 30 minutes at 100 °C to make the baseline flat, a TPD curve was obtained from 100 to 800 °C (heating rate: 10 °C min⁻¹). DFT calculations were performed using a first-principles simulation program (CASTEP⁵⁰) based on a plane wave pseudopotential. The exchange–correlation interaction was treated by the generalized gradient approximation with GGA and PBE. Atomic positions and lattice constants were respectively optimized until the residual forces and stress became smaller than 0.03 eV Å⁻¹ and 0.05 GPa.

Catalytic activity tests. The OCM activity was examined for each catalyst. The sample (0.1–1.0 g) was filled in an alumina microreactor with an outer diameter of 6 mm and set in a tube furnace. The amount of catalyst was generally set at 0.3 g and increased to 1.0 g (with catalyst-bed widths of 2–3 cm) in several runs to investigate the influence of increasing catalyst amounts. The reactor was heated in a flowing gas mixture (10.0 mL min⁻¹) consisting of CH₄/O₂/N₂, and the outlet gas composition was analyzed by gas chromatography (INFICON, Micro GC 3000). The measurements were performed between 600 °C and 800 °C with 50 °C intervals. The sample was held for 1 hour prior to each run to ensure its steady state. The reaction temperature was controlled by a thermocouple that was located near the alumina reactor. Our preliminary tests using an additional thermocouple inserted in the catalyst bed indicated that the temperature increase due to the exothermic methane oxidation was relatively small (0.6–5.6 °C). For the catalysts investigated, no decomposition products were found in their XRD patterns after the activity tests (Fig. S2 and S24†), ensuring that these catalysts have sufficient chemical stability.

The gas composition was set as CH₄/O₂/N₂ = 1.00/0.25/8.75 unless otherwise noted. A blank test resulted in small methane conversion even at 800 °C (Table S8†). CH₄ conversion rate X_CH4, C₂ selectivity S_C2, and C₂ yield Y_C2 were calculated by the following formulas. Slight amounts of higher hydrocarbons such as C₃H₆ and C₃H₈ were often detected, and these products were taken into account for calculations of Y_C2 and S_C2.

The carbon balance and mass balance were calculated by the following equations. These values sometimes exceeded 100% due to inevitable quantitative errors, especially when the CH₄ conversion was small. Alternatively, the emission of CO₂ adsorbed on the catalyst surface would overestimate these values.

2. Results and discussion

2.1. Catalytic activity of Li₂MO₃ (M = Si, Ge, Sn)

The OCM catalytic activity of Li₂MO₃ (M = Si, Ge, Sn) is presented in Fig. 1. The C₂ selectivity and C₂ yield reach 52.0% and 12.3%, respectively, for Li₂SiO₃ at 800 °C. While Li₂GeO₃ exhibits similar catalytic performance (50.3% and 11.7%), the values for Li₂SnO₃ fall off to 36.6% and 7.1%. Li₂SnO₃ is featured with its much lower C₂ selectivity, suggesting that the choice of constituent group 14 elements is crucial for the catalyst design. It has been generally recognized that oxides with weak metal–oxygen bonds tend to give active oxygen species that promote deep oxidation reactions.^27–29 Therefore, the much lower C₂ selectivity of Li₂SnO₃ may be explained by weakly bound surface oxygen, which preferentially produces CO₂ and CO as by-products. The relatively high C₂ selectivity of Li₂SiO₃ and Li₂GeO₃ is attributed to strong metal–oxygen bonds with Si and Ge that moderately deactivate the surface oxygen to suppress deep oxidation.

Fig. 1 OCM catalytic activity of Li₂MO₃ (M = Si, Ge, Sn) in a flowing gas mixture of CH₄ : O₂ : N₂ = 1.0 : 0.25 : 8.75. The red curves denote contour lines of the yield. BET surface areas (m² g⁻¹) are as follows. Li₂SiO₃: 1.1, Li₂GeO₃: 0.9, Li₂SnO₃: 1.2.

Our previous work revealed that the C₂ selectivity of Li₂CaSiO₄ gets enhanced with smaller specific surface areas.²⁵ Therefore, we attempted to synthesize and characterize Li₂MO₃ samples with varying their surface areas (Fig. S3 and Table S2†). As seen in Fig. S4,† the samples with smaller surface areas show inferior OCM performance due to the lowered CH₄ conversion, in sharp contrast to Li₂CaSiO₄ for which high values of CH₄ conversion are retained irrespective of the reduced surface area. Also noticeably, the C₂ selectivity of Li₂SnO₃ is again smaller than those of the Si and Ge counterparts. It thus turns out that catalysts containing lithium are not necessarily highly OCM active, and the combination of constituent metals is essential.

2.2. Catalyst surface properties

O 1s XPS spectra were measured for reference samples SiO₂, GeO₂, and SnO₂ to investigate the catalyst surface, because the M–O bonding strength could be one of the key factors of the OCM catalytic activity. As shown in Fig. 2, the O 1s spectra are deconvoluted with multiple Gaussian components. Previous works indicated various types of oxygen species on the catalyst surface.^{15,21–24,30–35} These oxygen species are discriminable with XPS as lower and higher energy peaks (Table 1). The spectra are deconvoluted into two peaks for SiO₂ and GeO₂: the lower energy peak is assigned to lattice oxygen (M–O–M) that bridges multiple metal atoms, while the higher energy peak as oxygen without bridging metal atoms, presumably mainly hydroxyl groups (M–OH).^30–35 For SnO₂, three peaks are observed in which the highest energy peak would be attributed to adsorbed water.^31,32 For all the three oxides, the relative intensities of the higher energy peaks decreased upon etching, consistent with the assumption that weakly bound surface oxygen species are preferentially removed by the etching process. It should be noted that the O 1s spectra of SiO₂ and GeO₂ are similar to each other, whereas that of SnO₂ differs significantly. This feature is reminiscent that Sn-based catalysts tend to show poor OCM performance compared to the Si and Ge-counterparts.

Fig. 2 XPS O 1s spectra of SiO₂, GeO₂, and SnO₂.

Table 1 Relative peak intensities in the XPS O 1s spectra of SiO₂, GeO₂, and SnO₂

	H2O [%]	OH^− [%]	O^2− [%]
SiO2		88.5	11.5
SiO2 (Ar-etched)		79.6	20.4
GeO2		75.9	24.1
GeO2 (Ar-etched)		70.2	29.8
SnO2	10.2	57.8	32.0
SnO2 (Ar-etched)		16.9	83.1

---

The XPS analysis reveals a larger M–OH fraction for SiO₂ and GeO₂ than SnO₂. This tendency may be attributed to their different chemical bonding nature. The bandgap widths are theoretically estimated at 5.4, 3.4, and 0.7 eV for SiO₂, GeO₂, and SnO₂, respectively. Electronic excitation beyond the bandgap involves the decreased/increased numbers of electrons in the valence/conduction bands. Since the valence/conduction bands have bonding/antibonding characters, the excitation of valence electrons into the conduction band results in weakened bonding strength between metal and oxygen atoms. Thus, the bonding strength depends on the bandgap width, and oxygen atoms are bound more tightly in SiO₂ and GeO₂ than in SnO₂. Taking into account that M–OH has lower stability than M–O–M, the M–OH species of SnO₂ would be too reactive to exist in large quantities. This situation is in sharp contrast to that in SiO₂ and GeO₂, where the M–OH peak intensity remains large and almost unchanged upon etching. The XPS result supports the above discussion and suggests that the stronger metal–oxygen bond is a dominant factor in suppressing the deep oxidation of methane, i.e., enhancing OCM catalytic activity. We note that the XPS results only represent the chemical properties of the catalyst surface and are not directly related to the catalytic reaction. In situ and operando analyses are indispensable to specify actual species that oxidize methane molecules during the OCM reaction.

Temperature programmed O₂ desorption (O₂-TPD) was performed on the Li₂MO₃ catalysts (M = Si, Ge, Sn) to discuss their different catalytic activity. As shown in Fig. S5,† all the catalysts show large desorption peaks above 200 °C. The onset temperature of O₂ desorption increases in the order of Li₂SnO₃ (200 °C), Li₂GeO₃ (450 °C), and Li₂SiO₃ (600 °C). This result indicates that oxygen in Sn-based oxides is bound more loosely than in the Ge- and Si-counterparts, consistent with the systematic trend in bandgap widths of SnO₂, GeO₂, and SiO₂. We thus conclude that the weakly bound oxygen species in Li₂SnO₃ causes deep methane oxidation, resulting in the reduced C₂ selectivity.

For Li₂CaSiO₄, O 1s XPS spectra were also measured to investigate the electronic state at the catalyst surface (Fig. 3 and Table 2). The O 1s spectra of Li₂CaSiO₄ are deconvoluted with three components located at E_B = 529.9 eV, 532.1 eV, and 533.8 eV. The peaks at 529.9 eV and 533.8 eV would come from oxygen species bound to metal atoms. The peak at 532.1 eV is attributable to carbonates derived from chemisorbed carbon dioxide (carbonate-derived peaks are present in the C 1s region at 288–290 eV).^{21–24,33–39} The two oxygen species bound to metal atoms are similarly assigned to lattice oxygen and hydroxyl groups, and the latter is dominant for Li₂CaSiO₄. The result indicates that the metal–oxygen bonding force is strong for this compound enough to suppress the deep oxidation of methane. Besides, Li₂CaSiO₄ shows a Li 1s peak around 56–57 eV even after a durability test for 50 hours (Fig. S6 and Table S3†), implying the stable presence of Li atoms at the outmost surface. This feature would originate from the robust Si–O covalent framework that prevents surface reconstruction, leading to the high catalytic durability of Li₂CaSiO₄.

Fig. 3 XPS O 1s spectra of Li₂CaSiO₄.

Table 2 Relative peak intensities in the XPS O 1s spectra of Li₂CaSiO₄

	OH^− [%]	CO3^2− [%]	O^2− [%]
Li2CaSiO4	61.0	36.5	2.5
Li2CaSiO4 (Ar-etched)	43.2	38.9	17.9

---

2.3. Effect of specific surface area on catalytic activity

For Li₂MO₃ (M = Si, Ge, Sn), the magnitude of the specific surface area is found to affect the catalytic activity significantly. Although the Li₂MO₃ catalysts with small surface areas hardly catalyze methane because of their weak basicity, they show much higher CH₄ conversion with the increasing surface area. This activity increment is too drastic to attribute only to the increased surface area, suggesting that different oxygen species have been involved. Differences in the specific surface area may change the number of physisorbed oxygen molecules, thereby affecting the constitution of oxygen species at the catalyst surface. Then, the OCM activity of Li₂CaSiO₄ is compared with those of Li₂SiO₃ and SiO₂, focusing on the specific surface area. As presented in Fig. 4, the Li₂CaSiO₄ catalyst with a larger surface area shows lower C₂ selectivity than the smaller-surface-area catalyst. For the larger-surface-area Li₂CaSiO₄ catalyst, the surface roughness and the increased defect density promote the adsorption of oxygen molecules, being subject to deep oxidation and lowered C₂ selectivity.⁴⁰ On the other hand, the oxygenation of methane is effectively suppressed for the smaller-surface-area catalyst because of the dominance of strongly bound oxygen species. A similar trend is observed for Li₂SiO₃ and SiO₂ (Fig. S7†). However, the reduced surface area concomitantly leads to extremely low CH₄ conversion due to small amounts of basic sites for Li₂SiO₃ and SiO₂ (Fig. S8 and Table S2†). The high CH₄ conversion of the large-surface-area SiO₂ having weak basicity implies that oxygen molecules adsorbed on the surface are highly reactive to cause deep oxidation of methane.

Fig. 4 (a) OCM catalytic activity of Li₂CaSiO₄ with specific surface areas of 0.2 and 0.5 m² g⁻¹ in a flowing gas mixture of CH₄ : O₂ : N₂ = 1.0 : 0.25: 8.75. (b) XRD patterns for the two Li₂CaSiO₄ catalysts.

Next, OCM activity tests upon varying the catalyst amount were conducted. The specific surface areas were determined by triplicate measurements for each sample to be 1.1 m² g⁻¹ for Li₂SiO₃ and 0.2 m² g⁻¹ for Li₂CaSiO₄ (Table S4†). For Li₂SiO₃, the CH₄ conversion is enhanced with the increasing catalyst amount from 0.1 g to 0.3 g (Fig. 5). The strongly mass-dependent activity of Li₂SiO₃ indicates that the catalyst surface allows for the physisorption of many oxygen molecules. Meanwhile, a similar performance for 0.3 g and 1.0 g implies that the supplied oxygen (reactant) has been mostly consumed in these conditions. Irrespective of the catalyst-weight vs. flow-rate (W/F) ratio, the C₂ selectivity is enhanced as the temperature increases (Fig. S9†). This behavior is interpreted as a consequence that the surface oxygen desorbs at elevated temperatures. At temperatures below 750 °C, many oxygen molecules adsorb at the catalyst surface such that the deep oxidation of methane is predominant, and the OCM reaction cannot be promoted even though the catalyst amount is optimized.

Fig. 5 OCM catalytic activity of (a) Li₂SiO₃ and (b) Li₂CaSiO₄ with various catalyst amounts in a flowing gas mixture of CH₄ : O₂ : N₂ = 1.0 : 0.25 : 8.75.

In contrast, Li₂CaSiO₄ with a small surface area shows superb performance regardless of the catalyst amount. Rather, the low-temperature activity of Li₂CaSiO₄ gets enhanced as the catalyst amount increases. It should be noted that the C₂ yield for the 1.0 g catalyst at 700 °C is comparable to that for 0.1 g at 800 °C (Table S5†). Regarding the catalytic activity with the 1.0 g catalyst, both the CH₄ and O₂ conversions slightly decrease when the feed-gas flow rate increases from 10 mL min⁻¹ to 50 mL min⁻¹, but the OCM performance is not influenced significantly (Tables S6 and S9†). We suggest that the small surface area of the Li₂CaSiO₄ catalyst effectively suppresses the physisorption of oxygen molecules, enabling us to lower the reaction temperature upon increasing the catalyst amount while preventing the direct oxygenation of methane.

2.4. Activation factors and reaction mechanism

To improve the C₂ selectivity, it is now evident that oxygen species at the catalyst surface must be tightly bound to prevent direct oxygenation of methane. Besides, a high degree of basicity is crucial for activating C–H bonds to accelerate the CH₄ conversion.^22,41–45 The hydroxyl group M–OH observed by XPS would not directly induce the CH₄ conversion because of its low electron density and weak basicity. Our previous work revealed a significant difference in the CH₄ conversion capability of four oxides having the same elemental combination: Li₂CaSiO₄, Li₂Ca₂Si₂O₇, Li₂Ca₂Si₅O₁₃, and Li₂Ca₄Si₄O₁₃.²⁵ The XPS analysis reveals similar O 1s spectra of these oxides with a predominant M–OH component (Fig. S10†). This result supports our view that the hydroxyl group does not contribute to the activation of C–H bonds.

Despite the absence of redox metal species, Li₂CaSiO₄ efficiently catalyzes partial oxidation of methane. Regarding a possible catalytic mechanism for OCM, Aika et al. reported that the generation of methyl radicals at the catalyst surface could be the initial step, followed by radical coupling in the gas phase.^46–48 OCM activity tests with/without Li₂CaSiO₄ were carried out at various CH₄/O₂ ratios at 800 °C (Fig. S11 and S12†). The OCM reaction proceeds even under a non-catalytic condition in the absence of oxygen, resulting in the formation of ethane and hydrogen (Fig. S11(a)†). The hydrogen is also detected in the presence of oxygen (Fig. S11(b) and (c) and S12†). This experimental fact suggests that hydrogen atoms are taken from methane to produce methyl radicals. The resultant methyl radicals are coupled in the gas phase to form ethane molecules. Under an anaerobic condition with the Li₂CaSiO₄ catalyst, the amount of C₂ hydrocarbons produced is larger than that in the non-catalytic reaction (Fig. S13†). Furthermore, Li₂CaSiO₄ shows stable OCM activity over 30 hours at 800 °C (Fig. S14 and S15†) without phase decomposition (Fig. S16 and S17†). These features indicate that the lattice oxygen works as an activator of methane dehydrogenation. Based on this mechanistic view, the surface oxygen would involve the catalytic cycle: the surface oxygen receives hydrogen from methane to produce water molecules, and the gaseous oxygen subsequently fills the oxygen-deficient sites, as reported previously.^{9–20,46–49}

2.5. Materials design for high-performance OCM catalysts

Based on the mechanistic aspect described above, we suggest that high-performance OCM catalysts must contain electropositive elements such as alkaline metals, as well as elements that form robust frameworks with strong metal–oxygen bonds for suppressing the deep oxidation with surface oxygen species. Here, we report that silicate and germanate with higher alkaline metal contents, Li₄SiO₄ and Li₄GeO₄ (Fig. S19 and S20†), are also potential OCM catalysts. Their OCM activities are compared with the Li₂CaSiO₄ catalyst (Fig. 6 and Table S10†). Both Li₄SiO₄ and Li₄GeO₄ exhibit significant OCM activity comparable to Li₂CaSiO₄, reaching C₂ yields of 26.2% and 24.7%, respectively, under the stoichiometric CH₄/O₂ = 2/1 condition at 800 °C. The CH₄ consumption rate per unit surface area is much faster for Li₄SiO₄ and Li₄GeO₄ than for Li₂MO₃, demonstrating that the Li-rich silicate/germanate can effectively activate CH₄ molecules (Fig. S22†). The C₂ hydrocarbons production rate is similarly fast, contributing to their superior C₂ selectivity. In addition, Li₄SiO₄ shows stable OCM performance in a durability test over 30 hours at 800 °C (Fig. S23†) without phase decomposition (Fig. S24†).

Fig. 6 OCM catalytic activity of Li₄SiO₄ and Li₄GeO₄ in flowing gas mixtures of (a) CH₄ : O₂ : N₂ = 1.0 : 0.25 : 8.75, and (b) CH₄ : O₂ : N₂ = 1.0 : 0.5 : 8.5. BET surface areas (m² g⁻¹) are as follows. Li₄SiO₄: 0.3, Li₄GeO₄: 0.3, Li₂CaSiO₄: 0.2.

Interestingly, Li₄SiO₄ and Li₄GeO₄ have crystallographic similarities to Li₂CaSiO₄. As schematically drawn in Fig. S25,† every oxygen site in these oxides is bound to one Si/Ge atom and multiple Li atoms, similar to the Li₂CaSiO₄ crystal lattice.⁵¹ While Li atoms adjacent to each lattice oxygen induce strong basicity and promote methane activation, robust Si–O and Ge–O bonds suppress the deep oxidation of methane, resulting in simultaneously enhanced CH₄ conversion and C₂ selectivity.^22,24 From this structural point of view, the relatively low OCM activity of Li₂SiO₃ and Li₂GeO₃ may also be explained. These oxides contain oxygen sites bound to two Si/Ge atoms (in other words, Si–O–Si or Ge–O–Ge bonds; see Table S11†), which would accordingly reduce the basicity at the oxygen site, thereby lowering CH₄ conversion. It should be noted that Li₂MO₃ and MO₂ (M = Si and Ge) possibly exhibit high values of CH₄ conversion when their surface areas are sufficiently large, but the simultaneous increases in CH₄ conversion and C₂ selectivity seem impossible because the increased surface area inevitably induces deep methane oxidation by physisorbed oxygen molecules. These results demonstrate that the local environment around oxygen sites plays an important role in the OCM activity, and precise structural control is the key to improving catalytic performance.

As discussed in our previous report on Li₂CaSiO₄,²⁵ this aspect is valid only when the atomic arrangement at the crystalline surface is intimately related to the bulk structure. We believe that our Li-based silicate/germanate catalysts meet this condition, taking into account the significantly different OCM activity between the Li₄MO₄ and Li₂MO₃ systems. In fact, the chemical analysis by X-ray photoelectron spectroscopy evidences similar surface atomic ratios for these systems, implying that their catalytic activity is hardly accounted for in terms of surface Li concentrations (Table S12†).

3. Conclusions

Through systematic characterization of several oxides containing group 14 elements, we demonstrate catalytic activity factors and the reaction mechanism for the oxidative coupling of methane (OCM). Lithium-based silicate Li₂SiO₃ and germanate Li₂GeO₃ showed much higher selectivities of C₂ hydrocarbons than their Sn-based counterpart Li₂SnO₃. The XPS study suggested that strong metal–oxygen bonds in silicates and germanates could be of importance to effectively suppress the deep oxidation of methane at the catalyst surface, resulting in enhanced C₂ selectivities. In contrast, weakly bound surface oxygen of Sn-based oxides is subject to direct oxygenation of methane to form CO_x as by-products, emphasizing the importance of metal–oxygen bonding strength. To achieve high OCM catalytic activity, the basicity of the catalyst is another primary factor for promoting methane activation, which is likely related to the local environment around oxygen sites.

Based on the chemical/structural viewpoints in the present study, we found that silicate and germanate with higher Li contents, Li₄SiO₄ and Li₄GeO₄, are significantly OCM active, comparable to the potential catalyst Li₂CaSiO₄. The significant OCM activity of Li₄SiO₄ and Li₄GeO₄ is believed to originate from the crystallographic similarities to Li₂CaSiO₄, where every oxygen site is bound to one Si/Ge atom and multiple Li atoms, resulting in simultaneously enhanced CH₄ conversion and C₂ selectivity.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors thank Prof. Ryo Maezono and Prof. Kenta Hongo of Japan Advanced Institute of Science and Technology for their help in DFT calculations. This work was supported in part by JSPS Grant-in-Aid for Science Research JP20H02827 (B) and JP19H04707 (Innovative Area “Mixed Anion”).

References

1. M. Zhao, S. Ke, H. Wu, W. Xia and H. Wan, Ind. Eng. Chem. Res., 2019, 58, 22847.
2. J. H. Lunsford, Catal. Today, 2000, 63, 165.
3. E. V. Kondratenko, T. Peppel, D. Seeburg, V. A. Kondratenko, N. Kalevaru, A. Martin and S. Wohlrab, Catal. Sci. Technol., 2017, 7, 366.
4. G. E. Keller and M. M. Bhasin, J. Catal., 1982, 73, 9.
5. S. Arndt, T. Otremba, U. Simon, M. Yildiz, H. Schubert and R. Schomäcker, Appl. Catal., A, 2012, 425, 53.
6. D. Kiani, S. Sourav, J. Baltrusaitis and I. E. Wachs, ACS Catal., 2019, 9(7), 5912.
7. D. Kiani, S. Sourav, W. Taifan, M. Calatayud, F. Tielens, I. E. Wachs and J. Baltrusaitis, ACS Catal., 2020, 10(8), 4580.
8. D. Kiani, S. Sourav, I. E. Wachs and J. Baltrusaitis, Catal. Sci. Technol., 2020, 10, 3334.
9. D. J. Driscoll, W. Martir, J. Wang and J. H. Lunsford, J. Am. Chem. Soc., 1985, 107, 58.
10. D. J. Driscoll, K. D. Campbell and J. H. Lunsford, Adv. Catal., 1987, 35, 139.
11. K. D. Campbell and J. H. Lunsford, J. Phys. Chem., 1988, 92, 5792.
12. Y. Tong, M. P. Rosynek and J. H. Lunsford, J. Phys. Chem., 1989, 93, 2896.
13. Y. Feng, J. Niiranen and D. Cutman, J. Phys. Chem., 1991, 95, 6558.
14. Y. Feng, J. Niiranen and D. Cutman, J. Phys. Chem., 1991, 95, 6564.
15. Y. Gambo, A. Jalila, S. Triwahyono and A. Abdulrasheed, J. Ind. Eng. Chem., 2018, 59, 218.
16. L. Luo, X. Tang, W. Wang, Y. Wang, S. Sun, F. Qi and W. Huang, Sci. Rep., 2013, 3, 1625.
17. L. Luo, R. You, Y. Liu, J. Yang, Y. Zhu, W. Wen, Y. Pan, F. Qi and W. Huang, ACS Catal., 2019, 9, 2514.
18. A. M. Maitra, Appl. Catal., A, 1993, 104, 11.
19. Z. Zhang, X. E. Verykios and M. Baerns, Catal. Rev.: Sci. Eng., 1994, 36, 507.
20. M. C. Alvarez-Galvan, N. Mota, M. Ojeda, S. Rojas, R. M. Navarro and J. L. G. Fierro, Catal. Today, 2011, 171, 15.
21. V. J. Ferreira, P. Tavares, J. L. Figueiredo and J. L. Faria, Ind. Eng. Chem. Res., 2012, 51, 10535.
22. J. Xu, L. Peng, X. Fang, Z. Fu, W. Liu, X. Xu, H. Peng, R. Zheng and X. Wang, Appl. Catal., A, 2018, 552, 117.
23. P. Huang, Y. Zhao, J. Zhang, Y. Zhu and Y. Sun, Nanoscale, 2013, 5, 10844.
24. Y. Hou, W. Han, W. Xia and H. Wan, ACS Catal., 2015, 5, 1663.
25. T. Matsumoto, M. Saito, S. Ishikawa, K. Fujii, M. Yashima, W. Ueda and T. Motohashi, ChemCatChem, 2020, 12, 1968.
26. U. Zavyalova, M. Holena, R. Schlçgl and M. Baerns, ChemCatChem, 2011, 3, 1935.
27. H. Chang, E. Bjørgum, O. Mihai, J. Yang, H. L. Lein, T. Grande, S. Raaen, Y. Zhu, A. Holmen and D. Chen, ACS Catal., 2020, 10(6), 3707.
28. R. K. Grasselli, Top. Catal., 2002, 21, 79.
29. Y. Moro-oka, Y. Morikawa and A. Ozaki, J. Catal., 1967, 7, 23.
30. A. Hashemi and A. Bahari, Appl. Phys. A: Mater. Sci. Process., 2017, 123, 535.
31. H. Jadhav, S. Suryawanshi, M. A. More and S. Sinha, Appl. Surf. Sci., 2017, 419, 764.
32. G. Gaggiotti, A. Galdikas, S. Kačiulis, G. Mattogno and A. Setkus, J. Appl. Phys., 1994, 76, 4467.
33. G. M. Bancroft, H. W. Nesbitt, R. Ho, D. M. Shaw, J. S. Tse and M. C. Biesinger, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 075405.
34. K. Prabhakaran and T. Ogino, Surf. Sci., 1995, 325, 263.
35. L. Peng, J. Xu, X. Fang, W. Liu, X. Xu, L. Liu, Z. Li, H. Peng, R. Zheng and X. Wang, Eur. J. Inorg. Chem., 2018, 17, 1787.
36. J. Xu, Y. Zhang, X. Xu, X. Fang, R. Xi, Y. Liu, R. Zheng and X. Wang, ACS Catal., 2019, 9, 4030.
37. J. Xu, Y. Zhang, Y. Liu, X. Fang, X. Xu, W. Liu, R. Zheng and X. Wang, Eur. J. Inorg. Chem., 2019, 2, 183.
38. C. S. Gopinath, S. G. Hegde, A. V. Ramaswamy and S. M. Mahapatra, Mater. Res. Bull., 2002, 37, 1323.
39. X. D. Peng, D. A. Richards and P. C. Stair, J. Catal., 1990, 121, 99.
40. E. Iwamatsu and K. Aika, J. Catal., 1989, 117, 416.
41. L. Peng, J. Xu, X. Fang, W. Liu, X. Xu, L. Liu, Z. Li, H. Peng, R. Zheng and X. Wang, Eur. J. Inorg. Chem., 2018, 17, 1787.
42. T. W. Elkins, S. J. Roberts and H. E. Hagelin-Weaver, Appl. Catal., A, 2016, 528, 175.
43. A. A. Davydov, M. L. Shepotko and A. A. Budneva, Catal. Today, 1995, 24, 225.
44. V. R. Choudhary and V. H. Rane, J. Catal., 1991, 130, 411.
45. K. Tanaka and A. Ozaki, J. Catal., 1967, 8, 1.
46. E. Iwamatsu and K. Aika, J. Catal., 1989, 117, 416.
47. K. Aika, N. Fujimoto, M. Kobayashi and E. Iwamatsu, J. Catal., 1991, 127, 1.
48. E. Iwamatsu, T. Moriyama, N. Takasaki and K. Aika, J. Catal., 1988, 113, 25.
49. V. I. Lomonosov and M. Yu. Sinev, Kinet. Catal., 2016, 57, 647.
50. S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. I. J. Probert, K. Refson and M. C. Payne, Z. Kristallogr., 2005, 220, 567.
51. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cy01641e

---