Probing Low-Temperature OCM Performance over a Dual-Domain Catalyst Bed

: The Mn-Na 2 WO 4 /SiO 2 catalyst is regarded as the most promising catalyst for the oxidative coupling of methane (OCM). Despite its remarkable performance, the Mn-Na 2 WO 4 /SiO 2 catalyst requires a high reaction temperature (>750 ◦ C) to show signiﬁcant activity, a temperature regime that simultaneously causes quick deactivation. In the current work, we show that the beneﬁts of this catalyst can be leveraged even at lower reaction temperatures by a using a stacked catalyst bed, which includes also a small amount of 5% La 2 O 3 /MgO on-top-of the Mn-Na 2 WO 4 /SiO 2 catalyst. The simple stacking of the two catalysts provides >7-fold higher activity and ~1.4-fold higher C 2 yield at 705 ◦ C compared to Mn-Na 2 WO 4 /SiO 2 and La 2 O 3 /MgO, respectively. We speciﬁcally show that the enhanced OCM performance is associated with synergistic interactions between the two catalyst domains and study their origin.


Introduction
The oxidative coupling of methane (OCM) reaction was identified by Keller and Bhasin in 1983 as a promising route to convert methane directly into ethane and ethylene [1]. If this process can be accomplished in an effective small-scale and distributed way, it can be used to convert methane from natural gas, shale gas [2], biogas [3], and landfills into important value-added products such as polymers. However, due to the symmetric CH 4 and the strong C-H bonds, a high reaction temperature is usually required to promote OCM (>750 • C) [1]. The high reaction temperature also promotes the over-oxidation of methane into CO x (CO and CO 2 ) and affects the longevity of the catalyst over prolonged exposure [4]. Despite extensive efforts, a catalyst of industrially applicable yield and high stability has not yet been found.
The Mn-Na 2 WO 4 /SiO 2 (MnNaWSi) catalyst has been and is still regarded as the most promising catalyst for this reaction, with early reports showing CH 4 conversion~25% with a C 2 selectivity of~80% (>800 • C) over long periods of time (~500 h) [5,6]. Despite the remarkable performance of the MnNaWSi catalyst, more recent reports show that structural changes in the multi-metal oxide material occur under reaction conditions [6][7][8][9][10][11][12][13][14]. Hayek et al. have shown that these structural changes also occur in the presence of additional dopants and result in strong deactivation, which is masked by the use of large amounts of catalyst [15]. Distinctly, it is the melting of crystalline Na 2 WO 4 at 698 • C that is believed to impose instability in the catalyst structure due to its strong interaction with the silica support and the MnO x phase as well as its evaporation (~1000 • C) at hotspots in the catalyst bed [8,16]. An operating temperature below or close to 700 • C should in principle limit those structural changes and help suppress the over-oxidation of the desired C 2 products into CO 2 [17].
It is generally accepted that the activation of the reaction involves the cleavage of the C-H bond to form methyl radicals (CH 3 •). The CH 3 • then couples to form ethane in the gas phase near the catalyst's surface [4,14,16,18]. It has been shown that the production rate of methyl radicals is strongly correlated to the C 2 yield [19][20][21]. Ideally, a good OCM catalyst would be able to generate methyl radicals at low temperatures (below 700 • C) and allow those to desorb and couple in the gas phase [14,20,21]. Hence, it is the ability of the MnNaWSi catalyst to generate and release methyl radicals that enable its remarkably high C 2 selectivity [14]. However, the MnNaWSi catalyst is only able to effectively activate methane above 750-800 • C, which makes the produced C 2 products highly susceptible to further oxidation to unwanted CO x .
Lansford and coworkers showed that lanthanum oxide (La 2 O 3 ) is active for OCM and is able to generate CH 3 • even at temperatures as low as~500 • C [20,22]. However, the C 2 selectivity of the La 2 O 3 is relatively low compared to the MnNaWSi catalyst, which limits its applicability on its own. Hence, the combination of La 2 O 3 and the MnNaWSi catalyst should produce a superior low-temperature OCM catalyst. Wu et al. reported using La to dope the MnNaWSi catalyst to achieve enhanced dispersion of the Mn and W on the silica support [23], along with a significant improvement in C 2 selectivity and activity in OCM. Whereas Ghose et al. and Ismagilov et al., showed only a marginal enhancement in C 2 selectivity, following a similar catalyst doping procedure [24,25]. Neither of them reported low temperatures reaction performance. Nevertheless, at present, there is a consensus in the OCM literature that supports the high potential of La 2 O 3 to reduce OCM reaction temperature [22,[26][27][28][29]. Recently, Jaroenpanon et al. showed that Mn-Na 2 WO 4 supported on La 2 O 3 , rather than on the commonly used SiO 2 support, provided a C 2 yield of 5~10% between 500 and 650 • C [30]. Zou et al. reported a~32% C 2 selectivity and a C 2 yield of~10% at 550-700 • C over a catalyst bed composed of a 1:1 (%wt) mechanical mixture of La 2 O 3 and Na 2 WO 4 /SiO 2 (NaWSi) [31]. However, the economic benefit of using the rare earth La 2 O 3 as a bulk support material for OCM is questionable [9,14,15]. In this study, we evaluate the OCM performance over MnNaWSi mechanically mixed with a small amount (~0.6 wt.%) of La 2 O 3 nanoparticles supported on MgO (La/MgO). The small amount of La is shown to provide significant OCM enhancement with the MgO support mitigating the deactivation by limiting the direct interaction of the Na 2 WO 4 melt with the La 2 O 3 . We demonstrate that the La/MgO to MnNaWSi weight ratio in the catalyst bed has a strong effect on catalytic performance. We further compared the effect of mixing the two catalysts vs. stacking the La/MgO catalyst on-top of the MnNaWSi. The effect of the above configuration on catalytic performance and catalyst deactivation at low-temperature reaction is discussed in detail.

Catalyst Preparation
Mn 2 O 3 -Na 2 WO 4 /SiO 2 (MnNaWSi). The Mn 2 O 3 -Na 2 WO 4 /SiO 2 was synthesized following a multi-step incipient wetness impregnation (IWI) as described in detail elsewhere [12,15]. Prior to synthesis, fumed silica of 7 nm primary particle size (Sigma-Aldrich, St. Louis, MO, USA) was dried in an oven at 140 • C for 12 h. Firstly, 1.4 mL of aqueous solution (miliQ water) of Mn(Ac) 2 ·4H 2 O (Sigma-Aldrich) with an appropriate concentration was dropwise added to 1 g of silica while vortex mixing. After impregnation, the powder was sufficiently mixed and then dried in the oven at 140 • C overnight. After drying, the powder was calcined to obtain Mn 2 O 3 /SiO 2 . The calcination was conducted in a tubular furnace, under air flow of~60 mL·min −1 at 110 • C for 2 h and then at 500 • C for 3 h. The ramping rate was set to 2 • C per minute. Next, 1.4 mL solution of Na 2 WO 4 ·2H 2 O (Alfa Aesar, Haverhill, MA, USA) in miliQ water with appropriate concentration was impregnated in the same way on the obtained Mn 2 O 3 /SiO 2 . After drying in the oven at 140 • C overnight, the catalyst was calcined in a tubular furnace, under air flow of~60 mL·min −1 at 115 • C for 2 h and then at 850 • C for 5 h. The ramping rate was set to 5 • C per minute.
La/MgO. The La/MgO was synthesized by IWI. Prior to synthesis, the MgO of 325 mesh particle size (Strem Chemicals, Inc., Newburyport, MA, USA) was dried in an oven at 140 • C for 12 h. An amount of 1 mL of aqueous solution (miliQ water) of La(NO 3 ) 3 ·6H 2 O (Strem Chemicals, Inc.) with an appropriate concentration was dropwise added to 1 g of MgO while vortex mixing. After impregnation, the powder was sufficiently mixed and then dried in the oven at 140 • C overnight. After drying, the powder was calcined to obtain La/MgO. The calcination was conducted in a tubular furnace, under air flow of~60 mL·min −1 at 800 • C for 4 h. The ramping rate was set to 5 • C per minute. La 2 O 3 . The La 2 O 3 was synthesized by the thermal decomposition of La(NO 3 ) 3 ·6H 2 O (Strem Chemicals, Inc.) at 115 • C for 2 h and 700 • C for 4 h. The ramping rate was set at 5 • C per minute.

Characterization Methods
The loading of the metals in each sample was determined by Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) using a Thermo Scientific (iCAP 6000) spectrometer. Prior to analysis, each sample was dissolved in HNO 3 and HCl in 1 /4 volumetric mixture at 70 • C overnight and then diluted with miliQ water. Crystalline components of the catalysts were determined by Powder X-ray Diffraction (PXRD). Patterns were recorded at ambient conditions on a Rigaku SmartLab diffractometer using Cu Kα radiation (λ = 1.54 Å) with 3 • min −1 scan speed and 0.01 • angle step. The catalyst morphology was evaluated before and after the reaction using High-Resolution Scanning Electron Microscopy (HR-SEM). The analysis was performed on an Ultraplus (Carl Zeiss, Heidelberg, Germany) Microscope. Everhart-Thornley and Backscattering detectors were used. For analysis, the sample powder was sprinkled on a carbon film before imaging.

Catalytic Testing
The catalysts were tested using a single-pass fixed-bed flow reactor made of quartz with an inner diameter of 7 mm. In each run of reaction, 50 mg of catalyst was loaded and packed inside the reactor. Two pieces of quartz wool were loaded and packed on both ends of the catalyst bed like a sandwich configuration to stabilize the reactor bed. Prior to catalyst loading, the sample was sieved and collected within 60-100 mesh. The reactor bed was packed to be 5 mm long. As La 2 O 3 and MgO support were denser than SiO 2 , some amount of inert quartz sand (60-100 mesh) was added and mixed with the reactor bed to reach a consistent bed size. The temperature of the quartz reactor was controlled by a temperature-controlled split tubular furnace (Vecstar, Eurotherm, Suisse). The feed gas temperature was monitored by an external K-type thermocouple housed in a quartz capillary tube. The thermocouple was placed above the catalyst bed inside the reactor. It is noted that the measured furnace temperature was consistently higher by~10 • C; hence, for convenience, we note only the furnace temperature. The feed gas composition and flow rate were controlled by mass flow controllers (Brooks SLA5850). Typically, the feed gas mixture had a total flow of 50 mL min −1 STP with composition of 4:1:1:4 CH 4 :O 2 :N 2 :Ar, which resulted in GHSV (gas hourly space velocity) of 60,000 mL h −1 g −1 (unless otherwise stated upon total flow change). The absolute reactor pressure was kept constant at~1.3 bar using a back-pressure controller (Brooks SLA5820). The reaction-produced water was separated in an ice trap placed at the outlet of the reactor. The outlet stream of the reactor was analyzed by a Clarus 580 GC (PerkinElmer) equipped with TCD and FID detectors. The CH 4 and O 2 conversions were calculated according to Equation (1) (i denotes CH 4 or O 2 ), while C 2 selectivity and C 2 yield were calculated according to Equation (2) (F i is molar flow rate) and Equation (3). To consider the change in the molar flow following the reaction, N 2 was used as an internal standard. The mass balance was ensured to close with >95%.
The stacking and mixing of different catalytic beds were prepared as follows: Stacked Beds (s). The stacking (s) beds La 2 O 3 -MnNaWSi and La/MgO-MnNaWSi were prepared by packing the La-containing catalysts (La 2 O 3 or La/MgO) on top of the already packed MnNaWSi catalyst bed. Different bed ratios (x) between the Lacontaining bed and the entire bed (50 mg) were studied, denoted as La 2 O 3 -MnNaWSi_xs and La/MgO-MnNaWSi_xs (x = 0.1, 0.2, or 0.5), respectively. The catalyst bed size for all tests was kept constant at length of 5 mm. Since La 2 O 3 and La/MgO have higher density than the MnNaWSi catalyst, the MnNaWSi catalyst bed was diluted with inert quartz sand (60-100 mesh) to maintain a consistent bed size of 5 mm. The quartz sand was mixed with MnNaWSi bed instead of La catalyst beds to minimize the limitation of •CH 3 transport to the lower bed. Prior to the preparation of catalyst bed, all particles were sieved and collected in between 60 and 100 mesh.
Mixed Beds (m). The mechanically mixed (m) beds of La 2 O 3 -MnNaWSi and La/MgO-MnNaWSi were prepared by mixing the La-containing catalysts (La 2 O 3 or La/MgO) with MnNaWSi. Prior to mixing, all particles were sieved and collected in between 60 and 100 mesh. Different weight ratios (x) between the La-containing catalysts and the entire bed (50 mg) were studied, denoted as La 2 O 3 -MnNaWSi_xm and La/MgO-MnNaWSi_xm (x = 0.1, 0.2 and 0.8), respectively. The catalyst beds for all tests were kept constant at mass of 50 mg and length of 5 mm. Since La 2 O 3 and La/MgO have higher density than the MnNaWSi catalyst, the MnNaWSi catalyst bed was diluted with inert quartz sand (60-100 mesh) to maintain a consistent bed size of 5 mm.

Results and Discussion
The catalysts in this study include La 2 O 3 , MgO-supported La 2 O 3 (La/MgO), and Mn-Na 2 WO 4 -SiO 2 (MnNaWSi) or a combination thereof. We examined two types of catalyst bed packing: (1) stacking (s), namely the La-containing catalyst (La 2 O 3 or La/MgO) placed on top of the MnNaWSi catalyst bed, and (2) mixed (m), namely a uniform mechanical mix of La-containing catalyst (La 2 O 3 or La/MgO) and the MnNaWSi. For clarity, the nomenclature of all the reactor bed combination tested was summarized in Table 1. Different loadings of the La-containing bed (La catalyst mass fraction in bed) were studied, as shown in Table 1. Each catalyst alone, La 2 O 3 , La/MgO, or MnNaWSi was tested separately as controls. Table 1. Nomenclature and the description of the catalyst reactor bed, including its reactor bed configuration and the weight proportion of the La bed.

Nomenclature La Catalyst Mass Fraction in Bed
La/MgO -

Catalyst Characterization
The loadings of the active metals in the La/MgO and the MnNaWSi catalysts were measured by ICP-OES, and the results are summarized in Table S1. The crystal phases identified for MnNaWSi were cubic Na 2 WO 4 , tetragonal α-cristobalite, and tetragonal Mn 7 O 8 SiO 4 , consistent with the literature of Mn 2 O 3 -Na 2 WO 4 -SiO 2 , as shown in Figure S1 [10,15,32,33]. The crystal phases of La/MgO were cubic MgO and hexagonal La(OH) 3 , as shown in Figure S1a. Notably, the La 2 O 3 phase is prone to react with the moisture in the air. Since the catalyst samples were exposed to ambient air containing water moisture, the spectra of the La-containing samples also showed the formation of La(OH) 3 , as shown in Figure S1b in Supplementary Materials. To validate the formation of La 2 O 3 , we analyzed a dedicated fresh catalyst sample, which was analyzed immediately as prepared using PXRD, as shown in Figure S1c. Previous work reported that at around 500 • C under air, the La(OH) 3 phase transforms into La 2 O 3 [34]. Hence, it is assumed that because of the in-situ calcination, the reaction temperature (>650 • C), and the oxidative conditions of OCM, the La in the active catalyst bed was initially present as La 2 O 3 in both the La/MgO and La 2 O 3 catalysts.

Catalytic Performance
The reaction onset temperature was~700 • C for the MnNaWSi catalyst and <650 • C for the La/MgO catalyst in Figure 1a. Interestingly, by simply stacking a small amount of the La/MgO catalyst (10 wt.% of the bed) over the MnNaWSi catalyst (La/MgO-MnNaWSi_0.1s), the reaction started already below 650 • C, as shown in Figure 1a. Even more interesting is the higher C 2 selectivity and C 2 yield obtained by the La/MgO-MnNaWSi_0.1s, as shown in Figure 1c,d. We found that the CH4 conversion over the La/MgO catalyst increased only slightly (~16-21%) as compared to the more significant increase in MnNaWSi (1-16%), but the C2 selectivity of La/MgO did increase with temperature, as shown in Figure 1a,c, respectively. The La/MgO-MnNaWSi_0.1s catalyst showed C2 selectivity greater or similar to that of the MnNaWSi, which comprises 90 wt.% of the catalyst bed, as shown in Figure 1c. The low activity of the MnNaWSi catalyst below 750 • C is consistent with previously published observations [35][36][37]. The drastic increase in activity of the MnNaWSi above 750 • C was previously associated with its ability to desorb molecular oxygen [37], facilitated by the interactions between the MnO x and the Na 2 WO 4 phases [10,38]. Gordienko et al. showed that the onset temperature for O 2 desorption from the MnNaWSi catalyst was 700 • C and reached a maximum of around 750-800 • C [37]. Using operando Raman and XRD-CT, Werny and Matras et al. demonstrated that at temperatures > 650 • C, the Mn 7 O 8 SiO 4 phase starts losing its crystallinity due to reaction with the molting Na 2 WO 4 and the simultaneous formation of MnWO 4 [10,38]. These coincide with the strong deactivation of the MnNaWSi catalyst, as was previously addressed by Lunsford and coworkers [39] and later on emphasized by Hayek et al. [15]. Hence, from the perspective of the longevity of the MnNaWSi catalyst, it is beneficial to run the OCM reaction at a temperature lower than the melting temperature of Na 2 WO 4 (698 • C).
We found that the CH 4 conversion over the La/MgO catalyst increased only slightly (~16-21%) as compared to the more significant increase in MnNaWSi (1-16%), but the C 2 selectivity of La/MgO did increase with temperature, as shown in Figure 1a,c, respectively. The La/MgO-MnNaWSi_0.1s catalyst showed C 2 selectivity greater or similar to that of the MnNaWSi, which comprises 90 wt.% of the catalyst bed, as shown in Figure 1c. However, its conversion was comparable to that of the La/MgO catalyst, which comprises only 10 wt.% of the La/MgO-MnNaWSi_0.1s catalyst bed. Specifically, at 725 • C, we find that the CH 4 and O 2 conversions of the La/MgO-MnNaWSi_0.1s were the same as those of the La/MgO while the selectivity of the former was higher by 14% and even slightly higher than that of the MnNaWSi. The reaction results at 725 • C with time on stream for 7 h demonstrate that the performance of all three catalysts was stable, with the C 2 yield being consistently highest over the La/MgO-MnNaWSi_0.1s, as shown in Figure 2d. The benefit of mixing bulk La2O3 with Na2WO4/SiO2 (NaWSi) at a 50:50 wt.% was previously identified by Zou et al. [31]. Their results at ~725 °C show similar p mance between the La2O3 and the La2O3 mixed with Na2WO4/SiO2, which indicate the La2O3 phase at that temperature range was governing the catalytic performanc therefore, tested OCM using unsupported La2O3 stacked on MnNaWSi at a 10:90 The benefit of mixing bulk La 2 O 3 with Na 2 WO 4 /SiO 2 (NaWSi) at a 50:50 wt.% ratio was previously identified by Zou et al. [31]. Their results at~725 • C show similar performance between the La 2 O 3 and the La 2 O 3 mixed with Na 2 WO 4 /SiO 2 , which indi-cates that the La 2 O 3 phase at that temperature range was governing the catalytic performance. We, therefore, tested OCM using unsupported La 2 O 3 stacked on MnNaWSi at a 10:90 wt.% (La-MnNaWSi_0.1s) at 725 • C and compared it to the performance of the La/MgO-MnNaWSi_0.1s catalyst, as shown in Figure S2. Given that MgO gives a very low C 2 yield ( Figure S2), it can be expected that the La-MnNaWSi_0.1s, having 30 fold more La 2 O 3 , would give significantly higher activity and C 2 yield compared to the La/MgO-MnNaWSi_0.1s catalyst. Interestingly, it was observed that over the relatively short reaction time (7 h), the C 2 yield of the former started dropping while the latter remained fairly stable ( Figure S2d). The most evident difference between the two catalysts was in the sharper decrease in O 2 conversion of the La/MgO-MnNaWSi_0.1s compared to La-MnNaWSi_0.1s. We associate the leveling in the C 2 yield in the former with the decrease in O 2 gas phase concentration, which maintains a more stable C 2 selectivity in the former. Speculating that the direct interaction of the La 2 O 3 with the Na 2 WO 4 is promoting the deactivation, we analyzed the spent La 2 O 3 -NaWSi_0.8m using PXRD, as shown in Figure S3a. From the XRD data, we noticed that the Na 2 WO 4 disappeared completely (after 14 h on stream), the peaks associated with the La(OH) 3 phase diminished, and a new inactive NaLa(WO 4 ) 2 phase emerged, as shown in Figure S3. These peaks were found to be mild in the La/MgO-MnNaWSi_0.1m catalyst, as shown in Figure S4. The formation of this phase was previously reported by Matras et al. using a La-containing membrane reactor [40]. These results clearly show the instability of the La 2 O 3 when in direct contact with the Na 2 WO 4 . Presumably, in the work of Zou et al., the effect of Na 2 WO 4 -induced deactivation was not evident, which was likely masked by the use of much larger amounts of catalysts, as we previously reported for the MnNaWSi catalyst [15]. Importantly, no such interactions were evidenced for MgO and Na 2 WO 4 . To further monitor the deactivation process, we examined the HR-SEM images of bulk La 2 O 3, La/MgO, and MnNaWSi, as shown in Figure S5. The images clearly show that the spent La/MgO in the presence of the MnNaWSi phase undergoes sintering. Notably, the XRD data in Figures S3 and S4 show that the La-MnNaWSi_0.1s is more prone to deactivate compared to the La/MgO-MnNaWSi_0.1s, which highlights the benefit of the MgO support.
To examine the effect of residence time on catalytic performance, we tested the La/MgO-MnNaWSi_0.1s catalyst under several GHSV values at a temperature below the melting temperature of Na 2 WO 4 (675 • C), which also means that the MnNaWSi phase is inactive, and at 725 • C, where both catalyst domains are active, as shown in Figure 3. Interestingly, at 725 • C, increasing the residence time (lower GHSV) did not significantly affect the methane conversion, C 2 selectivity, or C 2 yield, whereas the O 2 conversion showed a consistent decrease. It can therefore be concluded that the trend seen in CH 4 conversion at 725 • C is attributed to active site saturation at all tested GHSV values.
At 675 • C, we find that the C 2 selectivity and C 2 yield increased with GHSV. The CH 4 conversion remained similar at 20 and 40 L/g cat /h and increased slightly at 60 and 80 L/g cat /h. Moreover, the C 2 selectivity was found to increase monotonically while the O 2 conversion decreased with the increase in GHSV. It is noted that the MnNaWSi catalyst is largely inactive at 675 • C (Figure 1a). To better understand the catalytic behavior of the stacked catalysts below the melting temperature of Na 2 WO 4 , we tested the bare La/MgO and MnNaWSi at the same GHSV values, keeping the stacked structure of the catalyst bed and its size the same while replacing one of the domains with quartz particles, as shown in Figure S6. Consistently, we find that in all the tested GHSV values, the activity, selectivity, and yield of the La/MgO-MnNaWSi_0.1s were much higher than the respective levels in the stacked single component bed of La/MgO and MnNaWSi. The stacked MnNaWSi alone was found to mildly consume both methane and O 2 with no measurable formation of C 2 products. We find that decrease in O 2 conversion of the La/MgO-MnNaWSi_0.1s catalyst as a function of GHSV is correlated to that of the La/MgO. In contrast, the stacked La/MgO showed a decrease in C 2 selectivity, whereas the La/MgO-MnNaWSi_0.1s showed an increase in C 2 selectivity and yield with the increase in GHSV. It is postulated that the high reactivity of the La/MgO phase leads to the depletion of oxygen in the upper part of the bed. Hence, the generated methyl radicals and ethane are channeled to the bottom part of the catalyst bed, which, as previously reported [14,41,42] and shown here in Figure S6, affords the higher C 2 selectivity under lean oxygen conditions. formation of C2 products. We find that decrease in O2 conversion of the La/MgO-MnNaWSi_0.1s catalyst as a function of GHSV is correlated to that of the La/MgO. In contrast, the stacked La/MgO showed a decrease in C2 selectivity, whereas the La/MgO-MnNaWSi_0.1s showed an increase in C2 selectivity and yield with the increase in GHSV. It is postulated that the high reactivity of the La/MgO phase leads to the depletion of oxygen in the upper part of the bed. Hence, the generated methyl radicals and ethane are channeled to the bottom part of the catalyst bed, which, as previously reported [14,41,42] and shown here in Figure S6, affords the higher C2 selectivity under lean oxygen conditions. The trends above strongly indicate that the two catalyst domains "communicate" to affect the kinetic mechanism. This observation is further supported by looking at the change in C 2 H 4 /C 2 H 6 ratio as a function of temperature, as shown in Figure 4. We observe that above 700 • C, when the MnNaWSi becomes significantly active, the ratio starts to deviate from that obtained by the La/MgO catalyst to align with that of the MnNaWSi at 750 • C. The trends above strongly indicate that the two catalyst domains "communicate" to affect the kinetic mechanism. This observation is further supported by looking at the change in C2H4/C2H6 ratio as a function of temperature, as shown in Figure 4. We observe that above 700 °C, when the MnNaWSi becomes significantly active, the ratio starts to deviate from that obtained by the La/MgO catalyst to align with that of the MnNaWSi at 750 °C. To evaluate the effect of the transport of ethane and methyl radicals between the two catalyst phases, we tested the effect of catalytic bed composition with different mass ratio mechanical mixing (instead of stacking), as shown in Figure 5. Interestingly, the catalysts with the higher bed fractions of La/MgO (20 wt.% and 50 wt.%) showed higher and more stable oxygen consumption in conjunction with a proportionally lower C2 selectivity and, in turn, a lower C2 yield. Overall, the 10 wt.% stacked catalyst showed both the highest and most stable C2 selectivity and C2 yield with TOS. With the increase in the bed volume To evaluate the effect of the transport of ethane and methyl radicals between the two catalyst phases, we tested the effect of catalytic bed composition with different mass ratio mechanical mixing (instead of stacking), as shown in Figure 5. Interestingly, the catalysts with the higher bed fractions of La/MgO (20 wt.% and 50 wt.%) showed higher and more stable oxygen consumption in conjunction with a proportionally lower C 2 selectivity and, in turn, a lower C 2 yield. Overall, the 10 wt.% stacked catalyst showed both the highest and most stable C 2 selectivity and C 2 yield with TOS. With the increase in the bed volume to 20 wt.% and 50 wt.% we find that the C 2 selectivity drops. The drop in selectivity and, in turn, the C 2 yield is attributed to the short lifetime of methyl radicals [43], which most likely terminated before traversing to the bottom domain of the catalytic bed. This observation is further supported by the likeliness of the catalytic performance between the La/MgO-MnNaWSi_0.5s and the La/MgO-only bed. Explicitly, in the two catalyst bed compositions, the La/MgO dominated the catalytic performance, and the MnNaWSi was barely active in the test. The mechanically mixed 10 wt.% catalyst bed (La/MgO-MnNaWSi_0.1m) demonstrated low catalytic performance, similar to that of the MnNaWSi, as shown in Figure 5. Moreover, the mixed catalyst shows a drop in the catalytic activity and selectivity with TOS, which highlights the continuous occurrence of La 2 O 3 deactivation due to interaction with the molten Na 2 WO 4 phase, as previously reported for other oxides as well [15]. Plotting the C2 yield vs. the ethylene/ethane ratio at 725 °C for all the catalysts in this work, we find that the efficient stacking of the catalysts promotes both a higher C2 yield and a higher olefin-to-paraffin ratio, as shown in Figure 6. Interestingly, we find that the increase in the La/MgO amount, in the La/MgO-MnNaWSi_xs (x = 0.1, 0.2, 0.5) catalyst bed is a correlated to a decrease in both the C2 yield and the C2H4/C2H6 ratio with the La/MgO-MnNaWSi_0.5s nearing those of the mixed-bed La/MgO-MnNaWSi_0.1m. The mechanical mixing of the two catalysts leads to relatively low C2 yield and low C2H4/C2H6 ratios. Both of these trends are attributed to the deactivation process, as discussed above. Under lean oxygen conditions, Yoon et al. demonstrated that the conversion of ethane to Plotting the C 2 yield vs. the ethylene/ethane ratio at 725 • C for all the catalysts in this work, we find that the efficient stacking of the catalysts promotes both a higher C 2 yield and a higher olefin-to-paraffin ratio, as shown in Figure 6. Interestingly, we find that the increase in the La/MgO amount, in the La/MgO-MnNaWSi_xs (x = 0.1, 0.2, 0.5) catalyst bed is a correlated to a decrease in both the C 2 yield and the C 2 H 4 /C 2 H 6 ratio with the La/MgO-MnNaWSi_0.5s nearing those of the mixed-bed La/MgO-MnNaWSi_0.1m. The mechanical mixing of the two catalysts leads to relatively low C 2 yield and low C 2 H 4 /C 2 H 6 ratios. Both of these trends are attributed to the deactivation process, as discussed above. Under lean oxygen conditions, Yoon et al. demonstrated that the conversion of ethane to ethylene occurs over surface lattice oxygens in MnNaWSi [41]. This coincides well with the higher C 2 H 4 /C 2 H 6 ratios measured for the stacked beds. In these cases, the depletion of gas-phase oxygen in the top part of the bed results in the promotion of OCM in the bottom part of the bed under lean oxygen conditions, as opposed to the joint consumption of oxygen in the mixed catalyst bed.

Conclusions
In this work, we demonstrate the promotion of OCM over a dual-domain catalyst bed composed of a La/MgO catalyst and the MnNaWSi catalyst. We show that the OCM mechanism is better promoted when the bed is composed of only 10 wt.% La/MgO stacked on top of 90 wt.% MnNaWSi (La/MgO-MnNaWSi_0.1s). Testing the latter bed composition at ~650 °C and 60 L/gcat/h, we obtain ~10% CH4 conversion and ~19% C2 selectivity, whereas the MnNaWSi alone is inactive, and the La/MgO gives only 13% selectivity. Increasing the temperature to ~705 °C, the La/MgO-MnNaWSi_0.1s catalyst is found to give ~six-fold higher activity compared to the MnNaWSi catalyst or similar compared to the La/MgO catalyst, while the selectivity of the La/MgO-MnNaWSi_0.1s was 41%, which is higher than both individual catalysts. We further find that increasing the portion of the La/MgO or mixing it with the MnNaWSi showed diminished OCM performance, which is attributed to the deactivation of the La2O3 by reaction with the Na2WO4. Finally, we show that the dual-domain catalyst bed is promoting OCM through a synergistic kinetic mechanism that is active even below 700 °C.  Figure S6: Catalytic performance over a range of GHSV (20-80 L g -1 h -1 ) Table S1: Compositions of different pure catalysts.; Table S2: Expected catalytic performance of stacked 10 wt% catalyst based on performance of single components and their mass composition.

Conclusions
In this work, we demonstrate the promotion of OCM over a dual-domain catalyst bed composed of a La/MgO catalyst and the MnNaWSi catalyst. We show that the OCM mechanism is better promoted when the bed is composed of only 10 wt.% La/MgO stacked on top of 90 wt.% MnNaWSi (La/MgO-MnNaWSi_0.1s). Testing the latter bed composition at~650 • C and 60 L/g cat /h, we obtain~10% CH 4 conversion and~19% C 2 selectivity, whereas the MnNaWSi alone is inactive, and the La/MgO gives only 13% selectivity. Increasing the temperature to~705 • C, the La/MgO-MnNaWSi_0.1s catalyst is found to give~six-fold higher activity compared to the MnNaWSi catalyst or similar compared to the La/MgO catalyst, while the selectivity of the La/MgO-MnNaWSi_0.1s was 41%, which is higher than both individual catalysts. We further find that increasing the portion of the La/MgO or mixing it with the MnNaWSi showed diminished OCM performance, which is attributed to the deactivation of the La 2 O 3 by reaction with the Na 2 WO 4 . Finally, we show that the dual-domain catalyst bed is promoting OCM through a synergistic kinetic mechanism that is active even below 700 • C.  Table S1: Compositions of different pure catalysts.; Table S2: Expected catalytic performance of stacked 10 wt% catalyst based on performance of single components and their mass composition.