Renewable Hydrogen Production with Steam Reforming of Ethanol Using Siliceous Mesocellular Foam‐Supported Nickel Catalysts

Nickel (Ni) catalysts loaded on siliceous mesocellular foam (MCF‐S) are synthesized via the wet impregnation method with 3, 5, 10, and 15 wt% NiO loadings and different aging levels (no, partial, and full ageing) to determine how both factors affect the progress of the ethanol steam reforming (ESR) reaction. After extensive material characterization testing to determine material porosity, crystallinity, and Ni metal particle size and spatial location, as well as reaction testing at 300–700 °C and 4 H2O: 1 EtOH molar ratio, the fully aged 10 wt% Ni/MCF‐S possesses the strongest structural stability and catalytic activity, reaching 100% EtOH conversion and 68% H2 selectivity at 700 °C. The aging disperses and embeds more Ni nanoparticles within the walls of the mesopores, which promote the ESR reaction from easier diffusion and more active site contact within the pores. Furthermore, the catalyst reveals little signs of deactivation, as the structure remains virtually unchanged, and any coke formed is on the silica support and not over the Ni nanoparticles after the ESR reaction. Such results have demonstrated a proven applicability for ESR and a further need to research about aging effects toward improving structural properties and the catalytic reaction activity.


Introduction
[3] Hence, there is an urgent need to look beyond fossil fuels as an energy source and shift the focus instead toward clean and renewable fuels.[3][4][5] A popular, stable, and welldeveloped renewable source of hydrogen is ethanol, [5][6][7] which can be produced from both biomass and crops and is consumed largely for energy and transportation purposes, especially as a hydrogen carrier. [5,8,9]he most common industrial method for hydrogen production from ethanol is known as catalytic ethanol steam reforming (ESR). [5,6]It involves the endothermic reaction of the ethanol feedstock with steam at atmospheric pressure to generate hydrogen gas, as shown in Equation ( 1) and (2). [7,10]ESR has many advantages over other current H 2 production methods.For example, it is more thermally efficient than electrolysis, biological methods, or gasification for H 2 production. [2,10]It has also been reported to require a lower activation temperature, also reaching higher H 2 : CO product ratios compared to partial oxidation or autothermal reforming methods. [7,10]However, ESR has some drawbacks in practice, such as requiring high operation temperature up to 800 °C and having extensive greenhouse gas discharges, for example, CO, CO 2 , and CH 4 . [7,11] To solve these problems, various transition metal-based catalysts have been designed and studied to activate the ESR reaction.][14] Blomberg et al. [15] tested different transition metals for C─C bond dissociation ability, finding that Ni has a lower activation energy (%1.73 eV) than other common transition metals such as Co (%1.84 eV) and Fe (%1.95 eV), though this is still higher than precious metals such as Pd (%1.64 eV) and Rh (%1.19 eV).As precious metals are generally more expensive, Ni is widely accepted as a strong catalyst candidate for C-C cleavage in the ESR reaction.
The ESR process over Ni catalysts comprises several possible reaction intermediates and additional side reactions.Zanchet et al. [16] summarized the intermediates as follows (N.B. '*' indicates an adsorbed species): 1) O─H bond cleaves to CH 3 -CH 2 -O*, 2) C─H bond cleaves (dehydrogenation reactions) to CH*-CO* with H 2 released, 3) C─C bond cleaves to CH* and CO*, and then either 4) CH* hydrogenation toward CH 4 (undesired) at lower temperatures or 5) C─H bond cleaves to release CO and H 2 (desired) at higher temperatures, via H 2 O activation from the steam to form C*, O*, and OH*.These trends are generally in agreement with the findings of other researchers in the literature. [5,17]Furthermore, the CO generated could be subjected to the water-gas shift reaction (WGS), as shown in Equation ( 3) to generate even more H 2 . [16]Given the high-temperature requirement to activate the reaction and the formation of carbonaceous products via C-C scission (such as via ethanol decomposition), these processes can expectedly favor sintering (i.e., metal particle compaction) and coke (i.e., solid carbon) formation to deactivate the Ni catalyst. [10,13,18,19]][23] Investigations by Alberton et al. [24] and Biswas and Kunzru [21] determined that increasing the loading reduced the incremental improvement in catalytic activity.Thus, to find a commercially viable catalyst, the overall catalyst should be designed to maximize performance using the least amount of metal loading possible.Density functional theory (DFT) analyses further indicated that a smaller active Ni particle size and therefore a larger active surface area improve the catalytic activity from the increased presence of low-coordination Ni sites, such as oxygen vacancies or surface defects, which can improve scission of bonds such as C─C, C─H, and O─H. [5,12,16]If the smaller active Ni particles are more dispersed over the surface, this will likely reduce the deactivation effects from an increased overall surface area. [12,22]From this, nickel-based catalysts can be highly active to promote the key ESR reaction pathways if the increased loading contains smaller and more dispersed active sites to break down EtOH more easily.
The Ni nanoparticles can be loaded onto a support material(s) to form a hybrid catalytic system, which can increase the overall dispersion over its surface, as well as providing enhanced thermal stability by preventing nanoparticle agglomeration at elevated temperatures. [12,22]Common support materials include silica and carbon (e.g., carbon nanotubes, spheres, and fibers).While carbon structures are easier to manipulate, [25,26] inert silica supports are also strong candidates, as they are superior in terms of both structural and thermal stability. [27][29][30][31] This also makes them more preferable than microporous silica with up to only 2 nm pore diameter, which normally would restrict fluid diffusion. [29,32][45][46] Ranjekar et al. [17] considered a Cu-Co-Ce/MCF-S catalyst to study the ESR reaction, which, to the best of our knowledge, remains the only example in the literature of MCF-S being applied to this reaction.Their MCF surface area (634 m 2 g À1 ) compared to that calculated by Widyaningrum et al. (825 m 2 g À1 ) [45] was much lower.The only difference in the synthesis method was that the former included ammonium fluoride as part of the dissolution process; at least this differentiating factor resulted in a smaller surface area and a larger pore size (18.23 nm vs 17 nm).The results by Ranjekar et al. [17] have nonetheless revealed a very strong 97.4% ethanol conversion and 82.1% H 2 selectivity at 400 °C due to the presence of smaller active metal particles (7.3 nm) embedded within the porous network.Widyaningrum et al. [45,47] studied Ni/MCF-S catalysts to generate H 2 from pyrolysis of cellulose, finding that the H 2 selectivity was much higher as compared to using SBA-15 of Al 2 O 3 supports, as the calculated Brunauer-Emmett-Teller (BET) surface area of 825 m 2 g À1 and pore diameter of 17 nm were higher.The boosted reaction performance was ascribed to smaller Ni particles having deposited into the pores and having dispersed over the surface to increase the overall Ni surface area.Given the low cost, adequate structural properties, and proven success of MCF-S for ESR and H 2 production, Ni/MCF-S would be a highly desirable catalyst candidate to study further for the ESR reaction.
The synthesis of MCF-S is a multistep process and can be tuned to achieve different structural outcomes, [45,48,49] a summary of which is discussed about in the Experimental Section.Schmidt-Winkel et al. [48] found that when trying to control the overall material porosity, increasing the amount of swelling agent increased the MCF cell size, increasing the aging temperature increased the pore volume but decreased the BET surface area, and removing the ageing step resulted in a much poorer overall porosity.Hermida et al. [49] found that increasing the aging time increased the pore sizes, but decreased the surface area and pore volume, as well as adding Ni resulted in more small particles being deposited into the porous network.However, neither paper considered the direct effect of aging (i.e., the extent of air exposure and ambient conditions) on the MCF structure or the Ni deposition.
In this study, Ni/MCF-S catalysts were synthesized via ageing and Ni impregnation for the ESR reaction to uncover three major identified gaps: 1) the effect of aging on catalytic activity; 2) the effect of Ni loading on catalytic activity, especially H 2 selectivity and EtOH conversion rate; and 3) the effect of both on the mechanisms involved at different temperatures.

Experimental Section 2.1. Catalyst Preparation
A microemulsion templating method followed by aging and calcination was used to synthesize MCF, according to the method proposed by Schmidt Winkel et al. [48] Generally in the synthesis stage, the first step is microemulsion templating, in which a surfactant is added to an acidic solution to generate an emulsion of micelles. [17,31,46,48]Adding a hydrophobic organic swelling agent causes the micelles to expand and generate the porous network as a result of decreased liquid-solid and liquid-liquid surface tensions. [17,31,46,48]A silica precursor is then added and merged into the network. [48]The resulting system is aged at higher temperatures of %100 °C under static conditions [50] to polymerase the hydrophobic regions of the micelles and condense the silica precursor to form a more rigid and porous network. [48,51]Finally, the system is filtered, dried, and calcined at up to 600 °C to remove excess moisture and the original template and thereby generate an inert catalytic structure. [17,45,46,48]luronic P123 [17] (poly(ethylene glycol)-block-poly(propylene glycol)-blockpoly(ethylene glycol)-block) (8.0000 g, 1.4 mmol) was dissolved in 300 mL of 0.7 M HCl solution.The dissolution was performed at 40 °C under stirring.Upon completion of the dissolution step, 1,3,5-trimethylbenzene (Mesitylene) (4.0000 g, 33.28 mmol) and NH 4 F (0.0934 g, 2.52 mmol) were both added to the solution under vigorous stirring.The solution was stirred for 12 h.Tetraethyl orthosilicate (17.054 g, 81.99 mmol), TEOS, was then added as the SiO 2 source.The solution was stirred for another 20 h.Three different pathways were then undertaken to synthesize the MCF structure: 1) no aging, where the solution was heated to 100 °C using an oil bath under atmospheric conditions and stirred for %3-4 h until 30 mL of the solution remained; 2) partial aging, where the solution was heated to 100 °C under partially static conditions (i.e., partially exposed to atmospheric conditions during synthesis) for 24 h; and 3) full aging, where the solution was heated to 100 °C and aged for 24 h under static conditions in an oven.
The remaining solution was washed and filtered with deionized water and then left to dry in an oven at 80 °C overnight.Finally, the powdered sample was ground and calcined in air at 550 °C (1 °C min À1 ) for 8 h.
A wet impregnation method, using ethanol as the solvent and nickel (II) nitrate hexahydrate (Ni (NO 3 ) 2 •6H 2 O) as the nickel source, was used to produce ten different Ni catalysts supported on MCF.Ni/MCF catalysts with 3, 5, and 10 wt% Ni loading with no aging and partial ageing, and 3, 5, 10, and 15 wt% Ni loading with full aging, were prepared.Larger Ni loadings than this would likely promote ethanol decomposition and the reverse WGS reactions further from pore blockages and particle agglomeration. [5,46]n appropriate amount of the support was weighed and immersed in 40 mL of ethanol.Nickel (II) nitrate hexahydrate was then weighed and added into the suspension.The solution was then covered and put into an ultrasonic bath for 2 h to mix the solids.The solution was left covered for 12 h at room temperature under continuous stirring.The temperature was then increased to 70 °C under stirring to evaporate ethanol and the sample was left until %10 mL of ethanol remained.
The remaining sediment was dried in an oven at 80 °C overnight.Finally, the powdered sample was ground and calcined in air at 550 °C (1 °C min À1 ) for 6 h.For easier reference, the MCF-S materials and Ni/MCF-S catalysts were labeled throughout the remainder of this article based on the aging method (see Table 1).

Characterization
The surface area, pore size, and pore volume of the MCF supports were determined by a 77 K N 2 sorption experiment using an autosorb IQ (Quantachrome).The BET model [52] was used to extract the surface area, while the Barrett-Joyner-Halenda (BJH) method [53] was used to determine the pore size distribution and volume.The crystallinity of the catalysts was analyzed by X-ray diffraction (XRD) analysis using the Bruker D5000 and Shimadzu S6000 instruments with Cu as the X-ray irradiation source (λ = 1.54056Å).The diffraction patterns were recorded in the range of 1.5-10°with Bruker D5000 and 10-80°with Shimadzu S6000.
The thermal reducibility and stability of the catalyst samples were determined through a H 2 -temperature programmed reduction (TPR) method (TGA-SA Q5000).Catalysts' sample with 5-10 mg was loaded onto a small platinum alloy pan and heated from ambient temperature to 1000 °C at a ramp rate of 10 °C min À1 in a reducing gaseous environment (5% H 2 /95% Ar, 30 mL min À1 ).The level of coke deposition over the catalyst surface was analyzed using a temperature programmed oxidation (TPO) method (TGA-SA Q5000) and applying the same conditions as the H 2 -TPR method, except air was added instead of H 2 .To visualize the morphology and dispersion of metal nanoparticles immobilized around and within the catalyst, highresolution transmission electron microscopy (HRTEM) (JEOL 2100 operated at EHT of 200 kV) was applied.

Reactor Setup
A custom-designed fixed-bed reactor was used for the ESR reaction.A 50 mg catalyst sample was packed and placed in a quartz tube (7 mm inner diameter) and then loaded into a Labec cylindrical furnace.A mixture of EtOH/H 2 O liquid carried with N 2 gas flowed through the catalyst bed from the top of the furnace using an Adelab Scientific NE-1000 Syringe Pump.The flow rates of the N 2 gases were controlled using digital mass controllers (ALICAT Scientific mass controllers (M Series)).The gaseous product was quantified using an online gas chromatography analyzer (Agilent gas chromatography (GC) (7890B)), equipped with one column (ShinCarbon packed column  The experiments were performed at atmospheric pressure. Prior to conducting the ESR reaction testing, the catalyst was first pretreated at 700 °C under a reducing gaseous environment for 2 h (2.5 mL min À1 of gas, consisting of 20% H 2 % and 80% N 2 ).The reaction system was subsequently cooled to room temperature and N 2 gas was applied to purge the system to remove any residual gas and to protect the catalyst from oxidation.Then, the liquid feed was passed through the system at 1 mL h À1 with a molar ratio of H 2 O to EtOH of 4:1 (this ratio is common in the literature). [54,55]The composition of the output gases was analyzed by an online GC with five replicates.Thereafter, the temperature of the reaction system was ramped up from 300 to 700 °C with 50 °C intervals.EtOH conversion rate and selectivity of H 2 , CH 4 , CO 2 , and CO product were calculated from the GC data.Stability tests were conducted at 700 °C for 20 h.
Based on the GC signals, EtOH conversion rate (X, %) was calculated based on a carbon balance Selectivity (S, %) of component i, where i = H 2 , CO, CO 2, or CH 4 , was calculated as follows

Results and Discussion
The Ni/MCF-S catalysts with different Ni loadings were prepared and characterized.The original MCF-S structures with different aging levels were characterized to determine their overall porosity and distribution.
The N 2 adsorption-desorption isotherms of each MCF-S structure are depicted in Figure 1a.It can be seen that samples 2 and 3 (i.e., partial aging and full aging) were classified as Type IV with Type H2 hysteresis loops beginning at p=p0 % 0.65, indicating the overall mesoporosity of the supports as according to IUPAC classification. [56]However, the hysteresis loop of sample 1 (i.e., no aging) revealed a narrower hysteresis loop compared to the other two samples, which was reminiscent of a Type IV isotherm, but instead with a Type H4 hysteresis loop beginning at p=p0 % 0.6, [5,56,57] indicating a smaller average pore diameter.Furthermore, the surface area of sample 3 was 827 m 2 g À1 , similar to another similarly synthesized sample reported in the literature. [45]The corresponding pore size distribution for each sample is shown in Figure 1b.It revealed that the aging of MCF-S increased the pore sizes to 7.86 nm for partial aging and 9.52 nm for full aging, respectively, compared to that of MCF-S with no aging (i.e., 3.82 and 6.52 nm). [48]This trend is consistent with the findings of Hermida et al. and Schmidt-Winkel et al. [48,49] At the same time, the fully aged MCF-S (sample 3) also had the highest pore volume and the highest surface area among the three MCF-S supports (Table 2).However, the surface areas do not correspondingly decrease with an increased aging level despite the increasing pore size.Hermida et al. and Schmidt-Winkel et al. [48,49] indicated that further ageing the MCF material packs and agglomerates the particles further and therefore reduces the surface area.The trend in our results could have been due to the aging level of the partially aged material being  more difficult to control and quantify during synthesis, resulting in the nonaged and fully aged materials having a larger number of large pores.Nonetheless, increasing the aging effect toward synthesizing the Ni/MCF-S catalyst increases the pore size distribution and surface area, which would provide more space within and around the support to disperse the Ni active sites and promote the ESR reaction. [45]he MCF-S supports and Ni/MCF-S catalysts were analyzed by the powder X-ray diffraction (PXRD) to establish the overall crystallinity.Ni-3 with 15 wt% was also prepared to further reveal the dispersion effect.The low-angle XRD analysis of the MCF-S supports is shown in Figure 2a, which exhibited a stronger peak at 1.5°representing MCF-S for Ni-1 [58] ; this peak is weakly present in Ni-2 and not present in Ni-3.Combined with the pore size distributions from Figure 1b, this indicates that aging the MCF-S structure makes it more ordered with hexagonal and symmetrical structures. [58]61] Expectedly, the increased peak intensities indicated a larger presence of NiO over the catalyst surface with the increase of Ni loading.
To further reveal NiO particle locations, TEM and H 2 -TPR characterizations were performed.The morphology of the 10 wt% Ni/MCF-S catalysts was directly visualized through HRTEM, as displayed in Figure 3.The dark lumps within the catalysts were NiO.Ni-1 (Figure 3a) exhibited both an irregular mesoporous structure and a random, yet dispersed distribution of NiO [48] with an average particle size of 4.09 nm over the support surface.Ni-2 (Figure 3b) contained slightly disordered mesoporous structures [48] with pore sizes ranging from 6 to 12 nm.It was noted that the NiO lumps with average particle size of 6.42 nm followed the curvature of the MCF walls, suggesting that some NiO was embedded within both the MCF-S pores and the walls. [45]Ni-3 (Figure 3c) instead contained a more consistent and ordered pore size of %10 nm, which was larger than the other two, as is consistent with the BJH and XRD results and  the literature. [62]At the same time, the NiO with an average particle size of 5.75 nm appeared again to follow the MCF-S curvature and was embedded within the pore walls. [45]Since the difference in average particle size between each catalyst is very small, it may suggest that particle location would affect the catalyst activity more strongly than particle size.From above, it is confirmed that a higher level of aging generates a more homogeneous porous MCF-S structure that can embed more NiO nanoparticles within the pore walls.
H 2 -TPR characterization (Figure 4) was then applied to determine the NiO reducibility of the catalysts.Generally, more intensive peaks represent increased reducibility of NiO species.It can be seen that two distinct regions are visible, one of which is located at %400 °C representing NiO weakly bound to the support (i.e., NiO-WSMI), while the other is located at %550 °C, representing NiO strongly bound to the support (i.e., NiO-SMSI). [12,37,38,62]The latter peak became more observable with increased aging of the catalyst (especially for Ni-3), indicating that more NiO was embedded within the pore walls, which is consistent with the HRTEM results.Expectedly, it was noted that the reducibility of the catalysts proportionally increased with NiO loading (the relative amounts of NiO based on the TPR peak areas are included in Table S2, Supporting Information). [38,39]o further reveal the effect of the increased NiO loading, 15 wt% Ni was tested for the fully aged catalyst.It was found that adding beyond 10 wt% NiO for the fully aged catalysts could only increase the amount of weakly bound NiO over the surface instead of within the pores, which can barely enhance the ESR reaction performance. [45]Thus, aging and using up to 10 wt% NiO maintain a sufficient amount of dispersed and strongly bound NiO around and within the MCF-S support to increase ESR reaction activity.At the same time, given the presence of the temperature of reduction peaks, 600 °C would be the lowest temperature to reduce the catalysts with H 2 prior to reaction testing.However, as the tests were conducted at up to 700 °C, 700 °C was selected as the reduction temperature instead.
Catalytic performances of the Ni/MCF-S catalysts were tested in the ESR reaction, as shown in Figure 5.The performance of the 15 wt% Ni loading catalyst was also tested, with the results provided in Figure S1, Supporting Information.Given that 10 wt% loading has been revealed as potentially optimal to drive the ESR reaction, the effects of aging were first examined at this loading.EtOH conversion (Figure 5a) and H 2 and CO selectivity (Figure S2, Supporting Information) generally increased with temperature as ESR is an endothermic reaction, during which the catalyst showed no obvious signs of deactivation.However, while at least partially aging the catalyst (i.e., Ni-2 and Ni-3) appeared to slightly enhance the conversion, this improved structure did not majorly affect the product selectivity to any of CH 4 , CO, CO 2 , or H 2 (see Figure S2, Supporting Information).Since average particle sizes were similar for all the catalysts, the slightly inferior activity of the nonaged catalyst (Ni-1) could be related to the slight differences in the structure.For aged catalysts, the aging process optimized the structure of the supports through improving the porosity and therefore the extent of diffusion and mass transfer throughout the catalysts, as well the insertion of dispersed NiO to increase active site exposure.Thus, the structural improvements from aging, including the larger porosity and the active site insertion, have slightly increased the reaction activity by boosting the EtOH conversion from more fluid contact within the pores.
Based on the improved structure by aging, the active loading was increased from 3, 5 to 10 wt% Ni over Ni-3 to evaluate the optimum amount to successfully drive the ESR reaction.Once again, EtOH conversion and H 2 and CO selectivity expectedly increased with the loading amount with no major signs of deactivation (Figure 5b and S3, Supporting Information), although the difference between using 10 and 15 wt% loading was negligible.As the conversion rate never declined with an increase in temperature, the pores in the catalysts were presumably never blocked to limit ESR reaction progress.This is generally the same for the selectivity.Thus, structural changes to the catalyst only slightly improved the overall activity, though the aged 10 wt% catalyst was revealed to possess the strongest EtOH conversion and H 2 selectivity from having a sufficient number of Ni sites embedded within the larger pores.
The selectivity trends and the structure of the aged 10 wt% Ni/MCF-S catalyst were then used to infer the surface reaction mechanisms (see Figure 5c).At high temperatures (%650-700 °C), CO (20%) and H 2 (68%) selectivity values were higher from the ESR reaction, [16] although the decrease in both CO 2 and CH 4 at an equivalent rate suggested that ethanol decomposition to methane, [5,16,17] followed by steam and dry reforming of methane, could have occurred and contributed to this trend. [5,17,63,64]At medium temperatures (%400-550 °C), H 2 (59-69%) increased and CO (9-12%) selectivity only slightly increased, indicating both the ESR and WGS reactions. [5,16,17]At the same time, CO 2 and CH 4 decreased slightly as well, also suggesting the effects of dry methane reforming.At low temperatures (%300-450 °C), CO (21-9%) and CH 4 (17-9%) selectivity decreased, while that of CO 2 (8-18%) and H 2 (55-65%) increased, suggesting both the effects of the exothermic WGS and the declining effect of the exothermic CO methanation reaction. [5,16,17,65]Besides this, it was noted that no H 2 was formed at 300 °C for Ni-1, while there was no activity beneath 350 and 400 °C for Ni-3 catalysts with 3 and 5 wt% Ni, respectively.This may indicate that increasing the porosity and number of embedded particles from aging could promote WGS further at lower temperatures and using more NiO loading enables more reaction activity from a higher surface area of active sites enabling easier fluid contact.However, this only applies to lower reaction temperatures.As the temperature was increased, the selectivity trends remains the same.From this, as Ni-3 resulted in more conversion of EtOH from containing more Ni active metal embedded within the pores, these nanoparticles were therefore more involved in driving the ESR reaction forward.Moreover, as the 15 wt% Ni-3 catalyst contains more Ni sites over the surface and the catalytic activity remained largely unchanged, surface Ni does not promote the main reaction pathways as much.Since the selectivity values are not differentiated when changing either NiO loading or aging level as indicated above, it can be deduced that the location of the smaller-sized active sites within the pores does not change the mechanism pathways.Therefore, the key mechanisms for the overall ESR process are controlled over the Ni sites embedded within the MCF-S pores, although the overall structure does not change the main reaction pathways.
The key results of this analysis are summarized in Table 3 and compared with similarly investigated Ni/silica-based catalysts in the literature.Our results have revealed higher ethanol conversion and hydrogen selectivity values under generally similar reaction conditions and smaller Ni particle sizes.The ethanol conversion and H 2 selectivity results by Ranjekar and Yadav [17] are higher, as they used additional promoter materials to reduce the active Cu-Co bimetal particle size to promote ESR more easily.Nonetheless, the results confirm that incorporating MCF-S and material aging can improve the ESR catalytic activity.
Further encouraged by the potential of the 10 wt% Ni-3 catalyst, it was subjected to a 20 h stability test by performing the ESR reaction at 700 °C (Figure 6), as EtOH conversion reached 100% and H 2 selectivity (%66%) was relatively stable at this temperature (Figure 5).Over 20 h, the catalytic activity (Figure 6a) was mostly stagnant.EtOH conversion virtually remained at 100%,  while H 2 selectivity slightly decreased (from 68% to 64%), suggesting a very slight deactivation over time.To confirm the extent of deactivation, TPO and HRTEM characterization tests were performed.TPO (Figure 6b) was incorporated to check the oxidation of the material at elevated temperatures and therefore to determine the extent of coking by attempting to oxidize the surface carbon.The results revealed a very small peak at 375 °C, which was ascribed to amorphous coke, and a very large and identifiable peak at 692 °C, which was ascribed to graphitic/ filamentous coke. [66]Given the large amount of identified coke, the catalytic deactivation more likely occurred from coking rather than sintering.However, Wang and Lu [67] determined that the graphitic coke forms over the silica support instead of Ni, as the Ni particles are more embedded within the mesoporous framework.][70][71] Thus, the oxidation of amorphous coke at 375 °C was catalyzed by the surface Ni particles, while the oxidation of graphitic coke at 692 °C would have been catalyzed over the support surface by Ni particles close to or embedded within the porous support surface.As the catalytic activity was hardly affected over 20 h, the TPO analysis verified that the formed carbon was sufficiently distant from the Ni within the pores and correspondingly did not severely affect the activity of the Ni-3 catalyst.The HRTEM image (Figure 6c) further revealed a similar structure, pore size (%10 nm), and particle size distribution (%5.5 nm) compared to before the reaction and virtually no visible surface carbon.The calculated particle size was slightly smaller (5.32 compared to 5.75 nm), suggesting that some Ni was further deposited into the pores or pore walls. [72]hus, the catalytic activity for the 10 wt% Ni-3 catalyst was proved to be resistant to deactivating effects (coking and sintering) from the enhanced pore sizing and NiO pore insertion.

Conclusion
In this study, Ni/MCF-S catalysts with 3%, 5%, 10%, and 15% Ni loadings under different aging levels were synthesized via a wet impregnation method.The catalysts were subjected to the ESR reaction for their performance, activity, and stability.The characterization results revealed that a higher level of aging creates more ordered mesopores with increased overall porosity and embeds more NiO into the pore walls, leading to a strong SMSI.The small NiO nanoparticles embedded within the pores drive the key reactions rather than over the surface, which slightly increased the EtOH conversion, with intact selectivity.As such, the proposed reaction pathways, including ESR, WGS, ethanol decomposition to methane, methane reforming, and CO methanation, were not altered by the catalyst structure or active site location.Though NiO loading affects the catalytic activity, beyond 10 wt% NiO loading should not improve the catalytic activity significantly.Among the prepared catalysts, the fully aged 10 wt% Ni/MCF-S (Ni-3) demonstrated the highest catalytic activity with 100% EtOH conversion and 68% H 2 selectivity at 700 °C.At the same time, a 20 h stability test showed that this catalyst had high stability with little affected activity from hardly any coking over the Ni particles.Our research is the first known application of aged Ni/MCF-S catalysts specifically for ESR, which proved its great potential as a catalytic candidate for high structure stability and activity.In the future study, the effects of aging and Ni loading can be explored to optimize the reaction performance.

Figure 1 .
Figure 1.a) Nitrogen sorption isotherm at 77 K of each MCF-S sample.Prior to analysis, samples were degassed at 150 °C under dynamic vacuum.Closed and open circles represent adsorption and desorption isotherms, respectively.b) BJH pore size distributions of each MCF-S sample.

Figure 3 .
Figure 3. HRTEM images of 10 wt% Ni/MCF-S samples having undergone a) no aging (Ni-1), b) partial aging (Ni-2), and c) aging (Ni-3).NiO particle size distributions and the calculated average size (in nm) are provided within each TEM image.The smallest average NiO particle size among the catalysts is marked in red.

Figure 4 .
Figure 4. H 2 -TPR analysis of the Ni/MCF-S catalysts, featuring NiO-WSMI(weak metal-support interaction) and NiO-SMSI (strong metal-support interaction).The data was collected using a TGA-SA Q5000 device, at a ramp rate of 10 °C min À1 in a 5% H 2 /95% N 2 gaseous environment.

Table 1 .
Simplified labels of each MCF material and Ni-loaded catalyst based on the aging method incorporated during the synthesis stage.

Table 2 .
Surface area, pore volume, and pore diameter of the MCF-S supports exposed to different aging levels based on the BET and BJH results.

Table 3 .
Key results from the literature of Ni/silica-based or MCF-based catalysts applied to the ESR reaction.