Due to recent discoveries of natural/shale gas resources, the conversion of methane to ethane and ethylene (C₂-hydrocarbons), oxidative coupling of methane (OCM), has gained new research interest.¹ The use of O₂ as an oxidant makes this reaction more thermodynamically feasible. However, C₂-hydrocarbons are prone to oxidation to carbon oxides. Hindering these side reactions is a priority for OCM commercialization. The Mn/Na₂WO₄/SiO₂ system, one of the promising OCM catalysts, shows reasonable selectivity above 800 °C.² Catalysts based on oxides of lanthanides are also promising for OCM.^3-6 La₂O₃,⁷ Sm₂O₃⁸ or Nd₂O₃⁹ can activate CH₄ below 700 °C but have low selectivity to the desired products.

To control product selectivity, N₂O, CO₂, or S₂ have been used instead of O₂.¹⁰ N₂O improved C₂-hydrocarbons selectivity over various catalysts above 750 °C.^11-13 Sm₂O₃ was the only rare earth oxide tested in OCM with N₂O (N₂O−OCM), but showed low selectivity.¹⁴ The use of N₂O is advantageous from an environmental point of view, as the global warming potential of this compound is about 300 and 10 times higher than that of CO₂ and CH₄ respectively. One disadvantage of using N₂O as oxidant is its price. N₂O is, however, formed in large quantities as a by-product in the production of adipic acid,¹⁵ and can also be produced by the direct oxidation of NH₃,¹⁶ which may become the basis for low-cost N₂O production in the future. Motivated by these considerations, we introduce Gd₂O₃-based materials as a promising N₂O−OCM system that is highly selective at 650–700 °C. Kinetic tests revealed how the overall pathways of the formation of C₂-hydrocarbons and carbon oxides are affected by O₂ and N₂O. Temporal analyses of products enabled us to understand the role of promoter for Gd₂O₃ and type of oxygen species in the efficient conversion of CH₄ to the target products.

Oxides of Eu, Nd, Er, La, Gd, Ho, Dy or Sm were tested in O₂-, and N₂O−OCM at 750 °C (Figure 1). The selectivity to C₂₊-hydrocarbons (C₂−C₃ hydrocarbons) increased when O₂ was replaced by N₂O. The yield was almost unchanged over Sm₂O₃ and Dy₂O₃, decreased over Eu₂O₃ and Nd₂O₃ but increased over Ho₂O₃, La₂O₃, Er₂O₃ and Gd₂O₃ (to the highest extent). Gd₂O₃ was rarely used in OCM studies probably due to its low performance.^{6, 17} Ng and co-workers¹⁷ reported about 54 % selectivity to C₂₊-hydrocarbons at 32 % CH₄ conversion over Gd₂O₃ modified with BaO in O₂−OCM at 750 °C.

Inspired by the positive effects of Na, Sr and Ba promoters on the C₂-hydrocarbons selectivity of various catalysts,⁴ we modified Gd₂O₃ accordingly. The catalysts with 10 wt % promoter (0.1NaGd, 0.1SrGd and 0.1BaGd) consisted of the cubic Gd₂O₃ phase (Figure S1a). Ba and Sr were identified as carbonates. No Na-containing phase was detected. In situ and ex situ X-ray diffraction analysis of 0.1BaGd in O₂−OCM and N₂O−OCM revealed no change in phase composition (Figures S1b,c). Pseudo in situ X-ray photoelectron spectroscopy (XPS) tests also showed no obvious effect of the reaction feed on the catalyst surface composition. There are, however, some differences in the XP spectra of Gd₂O₃ and 0.1BaGd (Figures S1d–i). The XP C 1s spectrum of 01.BaGd is characterized by a signal at about 289 eV, which can be assigned to carbonates, probably barium carbonates in agreement with XRD data. The promoter should also be responsible for the presence of a signal at about 531 eV in the XP O1s spectrum.

The C₂₊-selectivity in O₂−OCM over Gd₂O₃, 0.1NaGd, 0.1SrGd and 0.1BaGd at 650 °C is 33.6 %, 15.1 %, 30.8 %, and 22.6 %, respectively (Figure 1b, Table S1). When the temperature increased, the selectivity increased, too. Such a dependence is consistent with previous O₂−OCM studies, which concluded that high temperatures are essential to achieve high C₂₊-selectivity.^{1, 2}

Compared with O₂−OCM, N₂O−OCM over 0.1BaGd reached the selectivity of 77 % at 0.44 % CH₄ conversion at 650 °C. The selectivity did not decrease even though the conversion increased to 7.5 % with increasing the reaction temperature to 800 °C (Figure 1c, Table S2). The selectivity of 0.1NaGd passed a maximum with increasing temperature while that of Gd₂O₃ and 0.1SrGd increased. Since the strongest positive effect was observed for 0.1BaGd, we further focused on this system. N₂O−OCM tests with BaCO₃ and BaO at 650 °C showed poor performance of these materials (Figure S2), so, Ba must be a promoter for Gd₂O₃. To understand the role of this promoter, BaGd catalysts with different Ba loading were further prepared and tested in O₂−OCM and N₂O−OCM.

In O₂−OCM at 700 °C, the rate of overall CH₄ consumption (r(CH₄)) as well as the rates of CH₄ conversion to C₂H₆ (r(C₂H₆)), C₂H₄ (r(C₂H₄)), CO₂ (r(CO₂)) and CO (r(CO)) decreased linearly with Ba loading (Figure 2a–c, Figure S3). Contrarily, an unexpected effect of Ba loading on these rates was observed in N₂O−OCM. Although Ba addition reduced r(CH₄) (Figure 2a), r(C₂H₆) followed a volcano-type dependence on the Ba/Gd ratio (Figure 2b). The r(C₂H₆) of all xBaGd was higher than that of Gd₂O₃. The r(C₂)/r(CO_x) ratio of xBaGd in O₂−OCM is close to 1, but is significantly higher in N₂O−OCM, reaching its highest value of about 9 for 0.1BaGd (Figure 2d). Increasing the Ba loading further has, however, a negative effect. Importantly, r(CO₂) and r(CO) decrease with Ba loading stronger in N₂O−OCM than in O₂−OCM (Figure S4).

The effects of O₂ and N₂O on the reaction pathways to CO, CO₂, C₂H₄, and C₂H₆ were elucidated by analyzing the selectivity-conversion relationships obtained through varying the space velocity at 700 °C (Figure 2). To distinguish between primary (formed from CH₄) and secondary (not formed from CH₄) products, the selectivity to each product was extrapolated to zero CH₄ conversion. The products with non-zero selectivity should be formed directly from CH₄, while those with a zero value are secondary products derived from primary products. Regardless of the oxidant, CO and C₂H₆ are primarily products formed over 0.1BaGd or Gd₂O₃ (Figure 2e–h). This is also true for CO₂ formation in N₂O- and O₂−OCM over 0.1BaGd and in N₂O−OCM over Gd₂O₃.

Replacing O₂ with N₂O decreased the primary selectivity to C₂H₆ and CO over Gd₂O₃ from 35 to 28 % and from 65 to 27 %, respectively, but increased the primary selectivity to CO₂ from 0 to 35 % (Figure 2e,f). Conversely, the primary selectivity to C₂H₆ over 0.1BaGd increased from 39 to 83 % (Figure 2g,h). This increase was accompanied by a decrease in the primary selectivity to CO and CO₂ from 30 to 4.5 % and from 31 to 12.5 %, respectively. Thus, the modification of Gd₂O₃ with Ba is pivotal for improving the efficiency of CH₄ conversion to C₂H₆ in the presence of N₂O.

In both O₂- and N₂O−OCM, primarily formed C₂H₆ is converted to C₂H₄ as reflected by the decreasing and increasing dependence of the selectivity to these products on CH₄ conversion (Figure 2). N₂O appears to be superior to O₂ in promoting this reaction as indicated by the higher selectivity ratio of C₂H₄ to C₂H₆ (Figure S5). The selectivity to CO and CO₂ decreased and increased respectively with increasing methane conversion due to oxidation of CO to CO₂. Since the selectivity to C₂₊-hydrocarbons or carbon oxides does not change significantly with increasing CH₄ conversion, the desired products are not involved in consecutive oxidation reactions (Figure S6).

Direct N₂O decomposition over Gd₂O₃ and 0.1BaGd was also investigated. Consistent with proposed mechanisms for this reaction over various catalysts,¹⁸ N₂O first reacts with an anion vacancy ([ ]) to form gas-phase N₂ and a monatomic adsorbed oxygen species ([O]) (eq 1. Gas-phase O₂ can be formed by a reversible recombination of two [O] (eq 2 or an irreversible reaction of N₂O with [O] (eq 3.

The dissimilar ability of Gd₂O₃ and its modified counterparts to form O₂ from N₂O was independently proved by analyzing the height-normalized responses of O₂ and N₂ (Figure 3a,b and Figure S8) recorded after pulsing of N₂O:Ne=1 at 700 °C in a temporal analysis of products reactor.¹⁹ Since the O₂ concentration is still high when the N₂ concentration is zero, recombination of [O] originated from N₂O should contribute to O₂ formation.

For a simple qualitative analysis of the rates of O₂ and N₂ formation, we use the time of maximum intensity (t_max) of the height-normalized O₂ and N₂ responses. For both unmodified and modified Gd₂O₃, O₂ formation should limit N₂O decomposition because t_max of O₂ is significantly higher than t_max of N₂ (Table S3). As the former value becomes larger in the presence of Ba promoter and with its rising concentration, the rate of [O] recombination should be hindered by the promoter probably due to an increase in the bonding strength of [O]. This statement is supported by the time required to reach zero O₂ concentration (Figure S9).

The involvement of [O] in CH₄ activation was verified by pump-probe experiments with N₂O and CH₄. When these gases were pulsed with a time delay (Δt) of 0.5 s, the t_max value of the O₂ response did not change compared to that of single N₂O pulses (Figures 3d and S10a). Its intensity, however, decreased sharply when CH₄ entered the reactor because [O] oxidized this alkane to carbon oxides. C₂H₆ was not observed due to the high-vacuum conditions, which are unfavorable for recombination of methyl radicals. An increase in Δt caused a decrease in coverage by [O] due to recombination of these species to O₂ (eq.2), which is detrimental to methane conversion (Figure S10b). The lower conversion of CH₄ over Gd₂O₃ compared to 0.1BaGd is explained by the higher recombination rate of [O] to gas-phase O₂. Lattice oxygen of Gd₂O₃ or BaO is significantly less active for CH₄ oxidation than [O] as proven by pulse experiments with CH₄ only (Figure S10c–f).

Based on the above results, we propose that the different activity of Gd₂O₃ and its modified counterparts to form O₂ from N₂O is the origin of the dissimilar behavior of these catalysts in O₂- and N₂O−OCM. Since the recombination of [O] to gas-phase O₂ over the modified catalysts (eq 2) is slow, CH₄ should react mainly with [O] in N₂O−OCM. The much faster recombination of [O] over Gd₂O₃ favors the formation of gas-phase O₂. Thus, O₂−OCM could occur during N₂O−OCM over Gd₂O₃. This scenario can explain why in contrast to Gd₂O₃ the selectivity to C₂₊-hydrocarbons is improved over xBaGd, 0.1NaGd and 0.1SrGd when O₂ is replaced by N₂O (Figures 1 and 2g,h). This explanation is indirectly supported by a negative relationship between t_max of the O₂ response in N₂O pulse tests and r(CO_x) or r(CH₄) in N₂O−OCM (Figures 3e, S11a). The stronger the oxygen species are bound, the lower their ability to oxidize CH₄/C₂-hydrocarbons to carbon oxides. The volcano-type relationship between r(C₂H₆) and Ba loading in Figure 2b can now be rationalized. This rate also follows a volcano-type relationship with t_max (Figure 3f). Importantly, 0.1NaGd and 0.1SrGd fit this relationship. Thus, an appropriate binding strength of [O] to the catalyst surface is vital to achieve the maximum r(C₂H₆), while the r(CH₄) decreases with the strength due to the inhibition of CO_x formation (Figure S11b).

The impact of the performance of 0.1BaGd achieved in N₂O−OCM is demonstrated by Figure 4a, where data from previous N₂O−OCM studies with different catalysts are also presented (Table S4). The 0.1BaGd catalyst showed the selectivity to C₂₊ and C₂-hydrocarbons of 89 % and 82 % at 3.5 % CH₄ conversion, and 87 % and 80 % at 8 % CH₄ conversion at 650 and 700 °C, respectively. Performing N₂O−OCM with a feed containing 70 vol % CH₄ further improved the selectivity to 89.8 % and 81 % at 11 % CH₄ conversion. All catalysts tested so far were less efficient. Potassium lanthanide chlorides,^16c Ca-actinide oxides,²⁰ Ca-lanthanide oxides,⁶ Sm-based catalysts,^16a LiMg-based catalysts,¹³ and Na₂WO₄-based catalysts¹³ operated at higher temperatures and achieved similar C₂₊-hydrocarbon yields but lower selectivity. According to a techno-economic analysis,²¹ however, the final price of ethylene is more influenced by selectivity than by methane conversion.

The C₂₊-selectivity of 90% at 11% CH₄ conversion obtained in the present study in N₂O−OCM at 700 °C is also remarkable in view of the O₂−OCM data obtained over different catalysts summarized in the database of Ref.,⁴ which contains about 1800 selectivity-conversion data points. Using this database, we have selected catalysts showing the selectivity to C₂-hydrocarbons above 80% at CH₄ conversion degrees above 5% (Table S5). Only the most promising Mn/Na₂WO₄/SiO₂ catalysts achieved the selectivity between 81 and 89 % at methane conversion above 15 % but above 750 °C (Figure 4b).

In conclusion, we have unveiled Gd₂O₃ modified with Ba, Na or Sr as a highly efficient system for N₂O−OCM at 650–700 °C. The promoter enhances the binding strength of mono-atomically adsorbed oxygen species formed from N₂O, which is pivotal for hindering their recombination to a diatomic adsorbed oxygen species, which shows higher ability for the direct CH₄ oxidation to carbon oxides. The mono-atomic species appears to have a higher selectivity for the conversion of CH₄ to C₂H₆. The established fundamentals provide clarity on the role of oxygen species in controlling product selectivity and can contribute to the design of OCM catalysts that also operate selectively with O₂.