Na 2 WO 4 /Mn/SiO 2 Catalyst Pellets for Upgrading H 2 S-Containing Biogas via the Oxidative Coupling of Methane

: Biogas is a promising renewable energy source; however, it needs to be upgraded to increase its low caloriﬁc value. In this study, oxidative coupling of methane (OCM) was selected to convert it to a higher fuel standard. Prior to establishing the scaled-up OCM process, the effect of organic/inorganic binders on catalytic activity was examined. The selection of the binders and composition of the catalyst pellet inﬂuenced the pore structure, fracture strength, and catalytic activity of the catalyst pellets. It was also observed that the O 2 supply from the inorganic binder is a key factor in determining catalytic activity, based on which the composition of the catalyst pellets was optimized. The higher heating value increased from 39.9 (CH 4 , Wobbe index = 53.5 MJ/Nm 3 ) to 41.0 MJ/Nm 3 (OCM product mixture, Wobbe index = 54.2 MJ/Nm 3 ), achieving the fuel standard prescribed in many countries (Wobbe index = 45.5–55.0 MJ/Nm 3 ). The reaction parameters (temperature, gas hourly space velocity, size of the reaction system, and the CH 4 /O 2 ratio) were also optimized, followed by a sensitivity analysis. Furthermore, the catalyst was stable for a long-term (100 h) operation under the optimized conditions. and N(8)M(0.5)Mn(1.5) shifted to a higher binding energy. In particular, a signiﬁcant shift in N(8)M(0.5)Mg(1.5) toward a higher binding energy was observed. These observations indicate that inorganic binders Mont, MgO, and Mn 2 O 3 were signiﬁcantly modiﬁed during the H 2 S-containing OCM, which altered their inﬂuence on NWM. In contrast, the catalyst pellets without an inorganic binder and those containing P25 and Al 2 O 3 did not exhibit a signiﬁcant shift in the W 4f peaks, indicating fewer modiﬁcations in the catalyst pellets during the OCM. These observations indicated that the W in the catalyst pellets was signiﬁcantly affected by the H 2 S-containing OCM and transformed into electron-poor W, degrading their OCM activity. N(8)M(0.5)Mo(1.5), N(8)M(0.5)Mg(1.5), and N(8)M(0.5)Mn(1.5) shifted to a higher binding energy. In particular, a significant shift in N(8)M(0.5)Mg(1.5) toward a higher binding energy was observed. These observations indicate that inorganic binders Mont, MgO, and Mn 2 O 3 were significantly modified during the H 2 S-containing OCM, which altered their influence on NWM. In contrast, the catalyst pellets without an inorganic binder and those containing P25 and Al 2 O 3 did not exhibit a significant shift in the W 4f peaks, indicating fewer modifications in the catalyst pellets during the OCM. These observations indicated that the W in the catalyst pellets was significantly affected by the H 2 S-containing OCM and transformed into electron-poor W, degrading their OCM activity.


Introduction
Biogas is a renewable energy source produced through the anaerobic digestion of organic wastes, which can be used to produce heat and electricity [1][2][3]. Replacing fossil fuels with sustainable biogas can reduce greenhouse gas emissions [4][5][6]; therefore, the global demand for biogas is growing [7,8]. Biogas is a mixture of CH 4 , CO 2 , H 2 , H 2 S, and other compounds, and its composition depends on feedstock and digestion conditions [9,10]. Biologically produced CH 4 , a major component of biogas, has been purified or upgraded to replace natural gas, i.e., cleaning (removing unwanted compounds) and upgrading (increasing the heating value of biogas) have been attempted [2,11,12]. The mass production of biogas is globally achieved typically for heat generation, and the valorization of biogas into high-quality fuels and chemical feedstocks can also be performed to replace the current fossil fuel-based conventional chemical industry.
Among the possible methods of upgrading, oxidative coupling of methane (OCM) has been proposed to be promising for improving the heating value of biogas and produce valuable chemical feedstocks [13][14][15][16]. The OCM converts CH 4 to C 2 or higher carbon number hydrocarbons (C 2+ compounds), which have heating values two or three times higher than CH 4 (Table S1), to meet global fuel standards (Wobbe index = 45.5-55.0 MJ/Nm 3 ). However, the OCM is highly exothermic (∆H = −87 kJ/mol at 800 • C) and requires an O 2 reactant, producing unwanted byproducts in excess, including CO and CO 2 (CO x ).
For selectively producing paraffins and olefins via OCM, many catalysts have been reported [17,18]. The Na 2 WO 4 /Mn/SiO 2 (NWM) catalyst is particularly advantageous because of its high activity and stability [19][20][21][22][23][24][25][26][27]. In addition to its application in upgrading natural gas, the economically feasible upgrading of biogas via OCM using NWM has been suggested [13][14][15], which exhibits high sulfur resistance and high stability for a longterm OCM process using a simulated H 2 S-containing biogas reactant [16]. Based on the feasibility of OCM for biogas observed at the lab scale, its possible scale-up to the industrial scale has been investigated with a typical OCM using a methane reactant in our lab [28].
In an industrial process, the catalyst pellets in small, compressed, and hard chunk forms are used for the efficient operation of large-scale processes. Mechanical rigidity is a key parameter in a catalyst pellet because the abrasion of catalyst pellets results in a significant pressure drop, and the resulting overheating of a reactor creates low efficiency and a consequent plant shutdown [29].
In this study, OCM using NWM catalyst pellets to upgrade biogas was investigated to (i) elucidate the effects of the catalyst pellets on OCM activity and (ii) optimize the reaction conditions. Catalyst pellets containing NWM, one of the most stable and active OCM catalysts, were used for biogas conversion [22]. The mechanically stable catalyst pellets were prepared using an extrusion method. This manuscript will investigate whether the geometric (pore structure and mechanical strength) and electronic properties of the prepared catalyst pellets are highly dependent on organic and inorganic binders, as illustrated in the literature [30][31][32]. The composition and types of binders can be manipulated to achieve optimal OCM activity [29,33]. Additionally, the organic binders induce the formation of pore structures, whose removal by calcination creates spaces between the catalyst powder particles and disperses the active sites on the catalyst pellets. However, the use of organic binders may not influence catalytic activity if the pore structures collapse during hightemperature calcination typically used for OCM catalysts [34][35][36]. Further, the addition of inorganic binders can improve the mechanical strength of the pellets. The inorganic binders of metals or metal oxides can also manipulate the electronic states of the active components by supplying oxygen to the OCM catalyst systems [37][38][39][40][41][42]. The OCM using catalyst pellets can also be controlled by reactor design and operating conditions [37,43,44] because the catalysis results are highly dependent on the reaction conditions [44][45][46][47]. Finally, optimization of the gas hourly space velocity (GHSV), reaction temperature, reactant flow rate, and CH 4 /O 2 ratio was performed to determine the feasible operating range for the desired OCM activity represented by C 2+ selectivity, olefin selectivity, and C 2+ yield.

Preparation of Catalyst Pellets and Their Physical Properties
The pore structures and mechanical strength of the catalyst pellets are important for the reaction. This is because the pore structures can adjust the mass and heat transfer in the catalyst bed to manipulate catalytic activity, while brittle pellets cannot be used for industry-scale reaction processes; therefore, a high mechanical strength is required. To prepare the catalyst pellets in this study, organic binders, including methyl cellulose (MC or M), polyvinyl chloride (PVA or V), and starch, as well as inorganic binders, including TiO 2 P25 (P25 or P), Al 2 O 3 (A), montmorillonite (Mont or Mo), MgO (Mg), and Mn 2 O 3 (Mn), were mixed, and the physical properties of the prepared pellets were characterized prior to the OCM reaction (fresh catalyst pellets). The nomenclature used to refer to the catalyst pellets is described in Scheme 1. The effects of the organic binders, with P25 as the inorganic binder, on the pore structures and mechanical properties of the catalyst pellets were investigated by measuring their Brunauer-Emmett-Teller (BET) surface area and fracture strength (Table 1). For the MC, PVA, and starch binders with an inorganic binder P25, negligible BET surface areas (1-2 m 2 /g) were obtained, indicating the collapse of the pores during high-temperature calcination [34][35][36]. For pellets containing the organic binder MC and different inorganic binders, BET surface areas of 1-8 m 2 /g were obtained. Notably, the measured BET surface areas of the catalyst pellets in this study were not significantly different from those of the catalyst powder (3 m 2 /g), indicating that the pore formation in the pellets was not particularly significant during pelletization. The largest BET surface area (8 m 2 /g) was observed for N(80)M(15)A(15), a pellet containing MC (organic binder) and Al 2 O 3 (inorganic binder); this was attributed to the high BET surface area of Al 2 O 3 . and fracture strength (Table 1). For the MC, PVA, and starch binders with an inorganic binder P25, negligible BET surface areas (1-2 m 2 /g) were obtained, indicating the collapse of the pores during high-temperature calcination [34][35][36]. For pellets containing the organic binder MC and different inorganic binders, BET surface areas of 1-8 m 2 /g were obtained. Notably, the measured BET surface areas of the catalyst pellets in this study were not significantly different from those of the catalyst powder (3 m 2 /g), indicating that the pore formation in the pellets was not particularly significant during pelletization. The largest BET surface area (8 m 2 /g) was observed for N(80)M(15)A(15), a pellet containing MC (organic binder) and Al2O3 (inorganic binder); this was attributed to the high BET surface area of Al2O3. Scheme 1. Nomenclature of NWM pellets. When inorganic binder P25 was mixed with organic binders MC, PVA, and starch, a minimal change in the fracture strength of the catalyst pellets was observed (2.52-2.79 MPa), as compared to that of N80M5P15 (2.79 MPa). However, adding other inorganic binders (Mont, Mn2O3, MgO, and Al2O3) with organic binder MC significantly reduced the fracture strength (0.19-1.45 MPa). These observations indicate that the inorganic binder is the highest contributing factor toward the fracture strength of the pellets, and P25 is the optimum inorganic binder to produce pellets with the highest mechanical stability. Observations of the BET surface area and fracture strength also suggest that the organic binders help pelletize the powder particles, but do not significantly impact the physical properties upon removal during calcination.
Scanning electron microscopy (SEM) images of the prepared pellets exhibited different surface morphologies depending on the inorganic binder ( Figure 1). In the absence of Scheme 1. Nomenclature of NWM pellets.  When inorganic binder P25 was mixed with organic binders MC, PVA, and starch, a minimal change in the fracture strength of the catalyst pellets was observed (2.52-2.79 MPa), as compared to that of N 80 M 5 P 15 (2.79 MPa). However, adding other inorganic binders (Mont, Mn 2 O 3 , MgO, and Al 2 O 3 ) with organic binder MC significantly reduced the fracture strength (0.19-1.45 MPa). These observations indicate that the inorganic binder is the highest contributing factor toward the fracture strength of the pellets, and P25 is the optimum inorganic binder to produce pellets with the highest mechanical stability. Observations of the BET surface area and fracture strength also suggest that the organic binders help pelletize the powder particles, but do not significantly impact the physical properties upon removal during calcination.
Scanning electron microscopy (SEM) images of the prepared pellets exhibited different surface morphologies depending on the inorganic binder ( Figure 1). In the absence of inorganic binder (N(8)M(0.5)), which contained only the NWM catalyst after the organic binder was removed during calcination, the pellet formed large particles networked with others. Adding inorganic binders P25, Mont, MgO, and Mn 2 O 3 did not significantly change surface morphology. However, adding Al 2 O 3 led to the formation of small sub-micrometer particles on the surface, which may be small OCM-inert Al 2 O 3 particles. Barring the presence of small particles on the surface, the near-identical morphology of the pellets suggests that the mass and heat transfer in the catalyst pellets may not be influenced by the type of inorganic binder. inorganic binder (N(8)M(0.5)), which contained only the NWM catalyst after the organic binder was removed during calcination, the pellet formed large particles networked with others. Adding inorganic binders P25, Mont, MgO, and Mn2O3 did not significantly change surface morphology. However, adding Al2O3 led to the formation of small submicrometer particles on the surface, which may be small OCM-inert Al2O3 particles. Barring the presence of small particles on the surface, the near-identical morphology of the pellets suggests that the mass and heat transfer in the catalyst pellets may not be influenced by the type of inorganic binder. Organic binder MC and inorganic binder P25 were selected based on the observed physical properties of the catalyst pellets. Because the organic binder was used for agglomerating the catalyst powder only, and not for manipulating the physical properties of the pellets, optimization of the P25 fraction was performed ( Table 2). The highest fracture strength of the catalyst pellets was observed for 15 wt% P25 or N(8)M(0.5)P(1.5), while N(8)P(1.5) prepared without the organic binders exhibited the lowest BET surface area. A lower surface area may correlate with reduced catalytic activity and poor mass transport (Table 2). However, because the pore structure collapses during calcination, adjusting the amount of organic binder does not significantly change the BET surface area.  Organic binder MC and inorganic binder P25 were selected based on the observed physical properties of the catalyst pellets. Because the organic binder was used for agglomerating the catalyst powder only, and not for manipulating the physical properties of the pellets, optimization of the P25 fraction was performed ( Table 2). The highest fracture strength of the catalyst pellets was observed for 15 wt % P25 or N(8)M(0.5)P(1.5), while N(8)P(1.5) prepared without the organic binders exhibited the lowest BET surface area. A lower surface area may correlate with reduced catalytic activity and poor mass transport (Table 2). However, because the pore structure collapses during calcination, adjusting the amount of organic binder does not significantly change the BET surface area.

Biogas Upgrading Using Catalyst Pellets
An OCM of the biogas-simulating mixture composed of CH 4 , O 2 , N 2 , CO 2 , and H 2 S was performed using the prepared catalyst pellets (Tables 3 and S2). The OCM activity did not significantly change depending on the type of organic binder (Table S3), and the effects of the inorganic binders were the focus of this study. Compared to the NWM powder catalyst, the catalyst pellet without an inorganic binder (N(8)M(0.5)) exhibited a lower CH 4 conversion (8.20% to 7.85%) and C 2+ selectivity (63.5% to 51.0%). The absence of the inorganic binders enables the channeling of the reactants through the void space between the pellets. Further, the dehydrogenation of paraffins to olefins was also suppressed, exhibiting lower olefin selectivity (28.5% to 21.0%). For the pellets containing organic binder MC and inorganic binders Al 2 O 3 , Mont, or MgO, poor OCM activity was observed (low CH 4 conversion, low C 2+ selectivity, and low C 2+ yield). Furthermore, MgO alone exhibited a high C 2+ selectivity (Table S4), whereas its mixture with NWM did not. However, pellets with inorganic binders P25 (N(8)M(0.5)P(1.5)) and Mn 2 O 3 (N(8)M(0.5)Mn(1.5)) exhibited higher OCM activity compared to the NWM powder: CH 4 conversion (7.85% to 10.0% and 8.63%, respectively) and C 2+ selectivity (51.0% to 83.9% and 71.7%, respectively); however, it is to be noted that the pellets contained less NWM because of the inorganic binder fraction. The improved catalytic activity with the addition of P25 and Mn 2 O 3 can be attributed to the improved O 2 supply [37][38][39].
From the above results, the catalyst pellets containing P25 exhibit the highest catalytic activity. Thus, the effect of P25 content on OCM activity was studied (Table 4). CH 4 conversion, O 2 conversion, and C 2+ yield increased with increasing P25 fraction; however, as stated above, increasing the P25 fraction decreases the fraction of OCM-active NWM. N(55)M(4)P(41), with the largest fraction of P25 in this study, exhibited the highest CH 4 conversion (12.8%) and O 2 conversion (95.5%). However, C 2+ selectivity decreased with the increasing P25 fraction, indicating that the conversion of methane to paraffins and olefins decreased with decreasing fractions of OCM-active NWM. Dehydrogenation was also favored, with an increasing fraction of P25. The highest C 2+ selectivity (83.9%) and highest olefin selectivity (44.4%) were achieved for N(8)M(0.5)P(1.5).
In addition to the inorganic binders, the effect of varying the organic binder content on OCM activity was also determined. However, this was less significant with adjusting the MC-to-NWM ratio to 0-2/8 (w/w), exhibiting 10.0-11.5% CH 4 conversion, 73.1-83.9% C 2+ selectivity, and 7.81-8.66% C 2+ yield (Table 4). Notably, N(8)M(0.5)P(1.5) exhibited the highest C 2+ selectivity (83.9%) and olefin selectivity (44.4%). However, in the absence of organic binder MC(N(8)P(1.5)), a lower CH 4 conversion (5.72%) and a lower C 2+ selectivity (67.7%) were observed, indicating that, despite its less significant manipulation of the catalyst structure, an organic binder is required for improved inorganic binder-induced OCM activity. O 2 TPD of a fresh (as-prepared) catalyst pellet was performed to elucidate the role of the inorganic binders in supplying O 2 to the reaction system ( Figure 2). Among the pellets containing inorganic binders, N(8)M(0.5)P(1.5) and N(8)M(0.5)Mn(1.5), which exhibited good OCM activity, exhibited strong O 2 desorption peaks at 750 • C and above. Compared to the pellets without inorganic binders (N(8)M(0.5)), the O 2 desorption peaks of those containing P25 and Mn 2 O 3 decreased (867 to 797 and 861 • C, respectively), indicating an easier oxygen supply from these inorganic binders [37,40]. For the pellets containing inorganic binder P25, the O2 desorption temperature varied with the P25 fraction ( Figure S1). Increasing the fraction of P25 resulted in a higher desorption temperature with a lower desorption peak intensity, indicating a suppressed oxygen supply. These observations indicate that the oxygen supply in the catalyst is manipulated by the presence of inorganic binder P25. Smaller fractions of P25 improved the oxygen supply, while an excess P25 suppressed it. An excess amount of P25 rapidly oxi- For the pellets containing inorganic binder P25, the O 2 desorption temperature varied with the P25 fraction ( Figure S1). Increasing the fraction of P25 resulted in a higher desorption temperature with a lower desorption peak intensity, indicating a suppressed oxygen supply. These observations indicate that the oxygen supply in the catalyst is manipulated by the presence of inorganic binder P25. Smaller fractions of P25 improved the oxygen supply, while an excess P25 suppressed it. An excess amount of P25 rapidly oxidizes the reactant, which can overwhelm the NWM-catalyzed OCM activity. For the catalyst pellet without organic binder MC (N(8)P(1.5)), the O 2 desorption temperature increased to 867 • C, which was higher than that of pellets containing organic binder MC ( Figure S1). These observations confirm that both organic and inorganic binders are required for improved OCM activity of the catalyst pellets; the organic binder MC improves the contact between inorganic binder P25 and the NWM catalyst.

Powder X-ray Diffraction (Powder XRD)
Fresh catalyst pellets, calcined at 800 • C, exhibited the crystal structures of α-cristobalite (PDF#39-1425), Na 2 WO 4 (PDF#12-0722), and Mn 7 SiO 12 (PDF#41-1367) regardless of the binder (Figure 3a). The type of organic binder did not induce any significant changes in the crystalline structure of the pellet ( Figure S2). However, the formation of quartz (PDF #46-1045) was observed with the addition of the inorganic binders P25, Mont, MgO, and Mn 2 O 3 , but not with Al 2 O 3 . Because the formation of cristobalite has been observed in silica supports of the NWM catalyst [34], and quartz was not observed in the absence of inorganic binders (N(8)M(0.5)), the presence of quartz can be attributed to the presence of the non-silica inorganic binders. Interestingly, the catalyst pellets containing P25, exhibiting the high diffraction peaks for quartz, achieved the highest OCM activity (Table 3), although the formation of cristobalite has been suggested to improve the OCM activity [22]. The formation of quartz was further investigated by calcining N(8)M(0.5)P(1.5) in air, N 2 , and 5% H 2 /Ar ( Figure S3). While quartz formed in the oxidizing (air) and inert (N 2 ) environments, the formation of cristobalite was observed in the reducing (5% H 2 /Ar) environment. These observations indicate that the oxidation of TiO 2 contributed to the formation of quartz. The possible oxygen supply from TiO 2 to the adjacent NWM may improve OCM activity, although the formation of cristobalite is suppressed with inorganic binder P25.    During the OCM reaction, the OCM-active Na 2 WO 4 transformed into MnWO 4 (PDF#13-0434) (Figure 3b) because of the redox cycle that supplies oxygen atoms to the active WO 4 [43]. Active WO 4 (PDF# 27-0789), which indicated a strong interaction between MgO and WO 4 , leading to lower OCM activity. Further, MnTiO 3 (PDF# 29-0902), which catalyzes O 2 activation at a lower temperature [38], was observed in spent N(8)M(0.5)P(1.5). The formation of MnTiO 3 was also confirmed by the TEM-EDS results ( Figure S4).
The change in the catalyst structure during the OCM reaction was further investigated using high-temperature powder XRD measurements (Figure 4). With increasing temperature, a new SiO 2 phase, tridymite (PDF #42-1401), formed, and the peaks of cristobalite shifted to a lower 2θ because of its transition to β-cristobalite [48]. Quartz was still observed at high temperatures.  Figure S3) and the oxidation state of Mn was higher in Mn 7 SiO 12 (2+ and 3+) than in MnWO 4 (2+), the catalysts were oxidized at higher temperatures, while the redox cycle supplied oxygen to the active WO 4 [42]. The presence of isolated Mn and W, which do not form Mn-W oxides, suggests the formation of highly dispersed Na-WO 4 surface sites [27]. Therefore, the improved OCM activity of the catalyst pellets containing inorganic binders is attributed to the facile oxygen supply and the formation of new active sites (MnTiO 3 ); however, quartz does not promote OCM activity.  The effect of P25 on the crystal structure of the catalyst pellets was also investigated ( Figure S5). Without binder (N(8)M(0.5)), the silica support exhibited α-cristobalite with a low intensity peak of quartz. Adding P25 increased the peaks of quartz and rutile TiO2, while those of Na2WO4, Mn7SiO12, and MnTiO3 decreased. On the other hand, the type and amount of the organic binders did not significantly affect the crystal structure (Figures S2 and S5). Quartz was observed as a major phase for SiO2, but MnTiO3 did not form in the absence of the organic binder. Since the organic binder acts as a glue for the inorganic binder and NWM, its absence limits their interaction, leading to the absence of  The effect of P25 on the crystal structure of the catalyst pellets was also investigated ( Figure S5). Without binder (N(8)M(0.5)), the silica support exhibited α-cristobalite with a low intensity peak of quartz. Adding P25 increased the peaks of quartz and rutile TiO 2 , while those of Na 2 WO 4 , Mn 7 SiO 12 , and MnTiO 3 decreased. On the other hand, the type and amount of the organic binders did not significantly affect the crystal structure ( Figures S2 and S5). Quartz was observed as a major phase for SiO 2 , but MnTiO 3 did not form in the absence of the organic binder. Since the organic binder acts as a glue for the inorganic binder and NWM, its absence limits their interaction, leading to the absence of MnTiO 3 .

Electronic Structures
The electronic structures of the NWM catalyst pellets were observed by X-ray photoelectron spectroscopy (XPS) (Figures 5 and S6-S10) [37][38][39][40][41]. The W 4f peaks of the fresh calcined N   Furthermore, significant changes in the XPS results were observed for Na 1s (Mont), Si 2p (Mont, Mn2O3), and O 1s (Mont, Mn2O3) during the OCM of simulated biogas (Figures S6-S10), which can be attributed to the degradation of the catalysts by the sulfur species. Notably, the S 2p peaks were not clearly observed for the spent catalyst pellets because of their low concentrations ( Figure S10).

Reaction Condition Optimization
The reaction conditions were optimized for the best OCM activity to selectively produce C2+ compounds (Figure 7) [37,[44][45][46][47]. The effects of the GHSV, temperature, flow rate, and CH4/O2 ratio on CH4 conversion, C2+ selectivity (and yield), olefin selectivity (and yield), and olefin/paraffin ratio were investigated. While the CH4 conversion depicts the activation of methane to form methyl radicals, the C2+ selectivity and the C2+ yield depict the formation of methyl radicals rather than the deep oxidation to COx. The olefin selectivity increases with improved formation of C2+ compounds and improved selective dehydrogenation of paraffins to olefins. The highest CH4 and O2 conversions (~16.5% and 100.0%, respectively) were observed at GHSV = 3333 h −1 , which did not significantly change at GHSV < 3333 h −1 (Figure 7a). The dehydrogenation of the paraffins to olefins was improved with increasing space time (decreasing GHSV from 20,000 to 1000 h −1 ), thereby increasing the olefin selectivity from 37.9% to 54.6%; however, the deep oxidation to COx (CO and CO2) improved, decreasing C2+ selectivity.
The scale-up of the reaction system was investigated by increasing the flow rate at a fixed GHSV (Figure 7c). The CH4 and O2 conversions increased from 10.0% and 77.8% to 13.8% and 100.0%, respectively, as the flow rate increased from 30 to 150 mL/min. However, the C2+ selectivity slightly decreased from 83.9% to 81.6%. The trends followed by the effects of the scale-up and reaction temperature are similar, which may be attributed  Figures S6-S10), which can be attributed to the degradation of the catalysts by the sulfur species. Notably, the S 2p peaks were not clearly observed for the spent catalyst pellets because of their low concentrations ( Figure S10).

Reaction Condition Optimization
The reaction conditions were optimized for the best OCM activity to selectively produce C 2+ compounds ( Figure 7) [37,[44][45][46][47]. The effects of the GHSV, temperature, flow rate, and CH 4 /O 2 ratio on CH 4 conversion, C 2+ selectivity (and yield), olefin selectivity (and yield), and olefin/paraffin ratio were investigated. While the CH 4 conversion depicts the activation of methane to form methyl radicals, the C 2+ selectivity and the C 2+ yield depict the formation of methyl radicals rather than the deep oxidation to CO x . The olefin selectivity increases with improved formation of C 2+ compounds and improved selective dehydrogenation of paraffins to olefins. The highest CH 4 and O 2 conversions (~16.5% and 100.0%, respectively) were observed at GHSV = 3333 h −1 , which did not significantly change at GHSV < 3333 h −1 (Figure 7a). The dehydrogenation of the paraffins to olefins was improved with increasing space time (decreasing GHSV from 20,000 to 1000 h −1 ), thereby increasing the olefin selectivity from 37.9% to 54.6%; however, the deep oxidation to CO x (CO and CO 2 ) improved, decreasing C 2+ selectivity. mol/mol; however, in the range of 3.74-8.10 mol/mol, the change was minimal. Further, the C2+ selectivity decreased from 83.9% to 49.5% with the decreasing CH4/O2 ratio, especially from 3.74 mol/mol to 1.57 mol/mol. The olefin selectivity exhibited a volcano-shaped curve, with the highest olefin selectivity (50.8%) at CH4/O2 = 3.74 mol/mol. Although improved dehydrogenation can be expected with a greater amount of the O2 reactant, the observed lower olefin selectivity at the lower CH4/O2 ratio (or the larger amount of O2) can be attributed to the gas-phase deep oxidation to COx.  The OCM activity of the catalyst pellets was significantly affected by the reaction temperature ( Figure 7b). As the temperature increased from 750 • C to 850 • C, the CH 4 conversion and O 2 conversion increased from 3.01% to 14.0% and 53.6% to 100.0%, respectively. The olefin selectivity also increased from 22.7% (750 • C) to 56.7% (850 • C), while the C 2+ selectivity decreased from 86.7% (750 • C) to 81.6% (850 • C).
The scale-up of the reaction system was investigated by increasing the flow rate at a fixed GHSV (Figure 7c). The CH 4 and O 2 conversions increased from 10.0% and 77.8% to 13.8% and 100.0%, respectively, as the flow rate increased from 30 to 150 mL/min. However, the C 2+ selectivity slightly decreased from 83.9% to 81.6%. The trends followed by the effects of the scale-up and reaction temperature are similar, which may be attributed to the formation of hot spots at higher amounts of the catalyst [28].
The OCM activity of the catalyst pellets significantly varied as a result of decreasing CH 4 /O 2 ratio (Figure 7d and Table S6). A significant change in catalytic activity was observed at CH 4 /O 2 < 3.74 mol/mol. The CH 4 and O 2 conversions increased from 18.9% and 88.4% to 41.6% and 96.6%, respectively, as the CH 4 /O 2 decreased from 8.10 mol/mol to 1.57 mol/mol; however, in the range of 3.74-8.10 mol/mol, the change was minimal. Further, the C 2+ selectivity decreased from 83.9% to 49.5% with the decreasing CH 4 /O 2 ratio, especially from 3.74 mol/mol to 1.57 mol/mol. The olefin selectivity exhibited a volcano-shaped curve, with the highest olefin selectivity (50.8%) at CH 4 /O 2 = 3.74 mol/mol. Although improved dehydrogenation can be expected with a greater amount of the O 2 reactant, the observed lower olefin selectivity at the lower CH 4 /O 2 ratio (or the larger amount of O 2 ) can be attributed to the gas-phase deep oxidation to CO x .
Based on the results depicted in Figure 7, optimized reaction conditions were determined ( Figure 8). Increasing the reaction temperature and decreasing the CH 4 /O 2 ratio increased the CH 4 conversion but decreased C 2+ selectivity. The C 2+ yield, a product of CH 4 conversion and C 2+ selectivity, also increased with increasing the reaction temperature and decreasing the CH 4 /O 2 ratio. This suggests that CH 4 conversion is a more decisive variable for the C 2+ yield. The olefin selectivity exhibited a complex dependence on the temperature and CH 4 /O 2 ratio because the formation of olefins requires the production of paraffins (C 2+ yield). Based on the suggested facile process operation, the optimal reaction conditions of 800 • C and CH 4 /O 2 = 8.10 mol/mol were selected, and a long-term OCM of simulated biogas was performed for 100 h. No significant catalyst deactivation (Figure 9) or change in the catalyst pellets after 100 h of reaction were observed ( Figure S11). Based on the results depicted in Figure 7, optimized reaction conditions were determined ( Figure 8). Increasing the reaction temperature and decreasing the CH4/O2 ratio increased the CH4 conversion but decreased C2+ selectivity. The C2+ yield, a product of CH4 conversion and C2+ selectivity, also increased with increasing the reaction temperature and decreasing the CH4/O2 ratio. This suggests that CH4 conversion is a more decisive variable for the C2+ yield. The olefin selectivity exhibited a complex dependence on the temperature and CH4/O2 ratio because the formation of olefins requires the production of paraffins (C2+ yield). Based on the suggested facile process operation, the optimal reaction conditions of 800 °C and CH4/O2 = 8.10 mol/mol were selected, and a long-term OCM of simulated biogas was performed for 100 h. No significant catalyst deactivation (Figure 9) or change in the catalyst pellets after 100 h of reaction were observed ( Figure S11).

Catalyst Pellets
The catalyst pellets were prepared using an extrusion method. The catalyst powder (8 g) was mixed with inorganic binders (0-6 g), organic binders (0-2 g), and DI water (6-10 mL). The mixture was kneaded for 5-10 min, and the prepared paste was extruded with a diameter of 2 mm. The extrudates were cut into 3-5 mm long pellets, which were dried at 105 • C for 16 h and calcined at 800 • C for 6 h. Details of the pellet preparation method are provided in Table S6.

Catalytic Activity Measurement
Catalytic activity was measured using a packed-bed quartz I-tube reactor with an internal diameter of 6 mm and a 250 mm straight cylindrical tubing. A mid-scale reaction was conducted using a reactor with an internal diameter of 1.54 cm. The catalyst bed was placed between the quartz wool in the middle of the reactor. Zirconia-silica ceramic beads, which are inert during OCM, filled the remainder of the reactor volume. The flow rates were controlled using mass-flow controllers. The total flow rate was 30-150 mL/min. The GHSV was controlled by increasing the amount of the catalyst from 1000 to 20,000 h −1 . Prior to the reaction, the catalyst was pretreated under N 2 flow by heating to 700 • C at a heating rate of 10 • C/min and maintained for 60 min. The reaction was performed at 700-850 • C. The temperature was maintained for 30 min prior to GC injection, followed by ramping at 5 • C/min. The flow rates were controlled using mass-flow controllers. The reaction mixture, including CH 4 , O 2 , CO, CO 2 , C 2 H 2 , C 2 H 4 , C 2 H 6 , C 3 H 6 , and C 3 H 8 , was identified using a flame ionization detector (FID) and a thermal conductivity detector (TCD). O 2 , CO, and CO 2 were quantified using a TCD with a 60/80 Carboxen-1000 packed column, and CH 4 , C 2 , and C 3 hydrocarbons were analyzed using the FID with an Agilent 19091P-S15 column. N 2 was used as the internal standard. Selectivity, conversion, and yield were calculated based on the data collected after 30 min of reaction (Equations (1)- (3)). The reaction was performed three times and averaged, which exhibited a deviation of 3.7% CH 4 conversion. Conversion (%) = (consumed moles of methane)/(initial moles of methane) × 100 (1) Selectivity of C x H y O z (%) = x × (produced moles of C x H y O z )/(consumed moles of methane) × 100 Yield of C x H y O z (%) = x × (produced moles of C x H y O z )/(initial moles of methane) × 100 (3)

Catalyst Characterization
N 2 physisorption was performed using an ASAP 2020 device (Micromeritics, Norcross, GA, USA). The fracture strength of the pellets was measured using an AFG-100N digital force gauge (Mecmesin, Slinfold, England). The pellet was loaded between the horizontally placed anvils with each end clamped. The magnitude of load where fracture of the pellet occurred was recorded. The value was divided by the cross-sectional area of pellet. Because the length of extruded pellets was not the same but ranged 3 to 5 mm, 20 catalyst pellets were measured, and the results were averaged. SEM images were collected using an Inspect F field-emission scanning electron microscope (Thermo Fisher Scientific, Waltham, MA, USA). O 2 TPD was performed using a BELCAT-B catalyst analyzer (Mi-crotracBEL, Osaka, Japan) equipped with a TCD and a mass spectrometer. Powder XRD was performed using a Dmax2500-PC (RIGAKU, Tokyo, Japan) with Cu Kα 1 radiation (λ = 1.5406 Å, 40 kV, and 200 mA). All samples were crushed into powder form prior to the XRD measurement. XRD data were collected using a quartz holder at 2θ = 5-90 • with a step scan rate of 4 • /min. XPS was performed using a Theta Probe AR-XPS system (Thermo Fisher Scientific, Waltham, MA, USA) with monochromated Al Kα excitation (hν = 1486.6 eV) operated at 15 kV and 150 W at the Korea Basic Science Institute (Busan, Korea). The measured binding energy was calibrated using the C 1s peak at 284.6 eV. High-temperature powder XRD was conducted using an X'Pert PRO (Philips, Amsterdam, Netherlands) with Cu Kα 1 radiation (λ = 1.5406 Å, 60 kV, and 60 mA).

Calculation of Higher Heat Values (HHV)
The higher heating value (HHV) of the product stream was calculated using the molar fractions of the hydrocarbons and the corresponding HHV values in MJ/Nm 3 under standard conditions for temperature and pressure (Table S7). The molar fractions (x i ) of the hydrocarbons were measured using the GC results. Only the hydrocarbon mixtures without CO, CO 2 , O 2 , and other compounds were used to calculate the total HHV (HHV total ) (Equation (4)) [49].

Conclusions
The catalytic OCM reaction was observed in different reaction systems for developing scaled-up biogas upgrading processes. The HHV increased from 39.9 (CH 4 , Wobbe index = 53.5 MJ/Nm 3 ) to 41.0 MJ/Nm 3 (OCM product mixture, Wobbe index = 54.2 MJ/Nm 3 ), achieving the fuel standard prescribed in many countries (Wobbe index = 45.5-55.0 MJ/Nm 3 ). The type and compositions of the organic/inorganic binders widely influenced the physical properties, such as morphology, specific surface area, and fracture strength, of the catalyst pellets. The effect of the organic/inorganic binders on the catalytic activity was also examined. The inorganic binder affected the catalytic activity or crystalline structure of the pellets by supplying O 2 to the active sites, while the organic binder affected the catalytic activity of NWM by mediating the interaction between the catalyst and the inorganic binder. The sensitivity of the reaction parameters (temperature, GHSV, total flow rate, and CH 4 /O 2 ratio) was also examined. The reaction conditions were optimized based on the results of the sensitivity analysis. Furthermore, the stable OCM activity of NWM was confirmed by the results of a long-term (100 h) operation. Because this study focused on the preparation of catalyst pellets and the process condition optimization using these pellets, we provided the preliminary knowledge to design the scaled-up biogas upgrading process. Although the observed reaction results indicated that the scale up process did not significantly adjust the catalytic activity, an experimental confirmation was required to verify the optimized conditions of biogas upgrading processes. A better understanding of scaled-up OCM processes of biogas upgrades may be achieved based on the findings in this study because the process analyses in the literature are based on the OCM results of a powder catalyst without the sulfur-containing compounds in the reaction feed.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/catal11111301/s1, Table S1: Higher heating values of OCM products in STP conditions, Table S2: Compositions of OCM products, Table S3: OCM results at 800 • C using catalyst pellets depending on organic binders, Table S4: OCM results at 800 • C using inorganic binders, Table S5: OCM results at 800 • C using catalyst pellets depending on inorganic binders with different feed composition, Table S6: Preparation of catalyst pellets, Table S7: Higher heating values of products at STP conditions, Figure S1: O 2 TPD results of fresh catalyst pellets depending on the fractions of (a) inorganic binder P25 and (b) organic binder MC, Figure S2: Powder XRD results of (a) fresh and (b) spent catalyst pellets depending on organic binders, Figure S3: Powder XRD results of fresh N(8)M(0.5)P(1.5) depending on the calcination environments of air, N 2 , and 5% H 2 /Ar, Figure S4: TEM-EDS result of spent N(8)M(0.5)P(1.5), Figure S5: Powder XRD results of (a) fresh and (b) spent catalyst pellets depending on the amount of inorganic binder P25, Figure S6: Na 1s XPS results of (a) fresh and (b) spent catalyst pellets depending on inorganic binders, Figure S7: Mn 2p XPS results of (a) fresh and (b) spent catalyst pellets depending on inorganic binders, Figure S8: Si 2p XPS results of (a) fresh and (b) spent catalyst pellets depending on inorganic binders, Figure S9: O 1s XPS results of (a) fresh and (b) spent catalyst pellets depending on inorganic binders, Figure S10: S 2p XPS results of (a) fresh and (b) spent catalyst pellets depending on inorganic binders, Figure S11: Powder XRD results of spent catalysts after 100 h reaction.