Use of structured systems as a strategy to minimize the deactivation of Ni-based catalysts applied in dry reforming of methane

The main challenge in the use of Ni based catalysts is the high deactivation rate of these catalysts. In this work, strategies aimed at improving this characteristic such as the use of structured catalysts were studied. In this work, the Ni/ γ -Al 2 O 3 (Ni/Al) and Ni/La 2 O 3 - γ -Al 2 O 3 (Ni/La-Al) catalysts were synthesized by the all-in-one method and used in the Dry Reforming of Methane combined with its application in structured systems to minimize the effects of deactivation. The catalysts were characterized and a smaller Ni crystallite size for the La-promoted catalyst was observed. The deactivation of the structured catalysts and application of residual activity deactivation models (DMRA) were evaluated by applying different weight hourly velocities (WHSV). Besides that, the regeneration of the catalysts was developed through the comparison of the treatment with CO 2 and H 2 atmospheres. Furthermore, the greatest and the lowest deactivation of the structured systems were identi�ed for the WHSV values of 40 and 20 L g cat−1 h − 1 , respectively. Finally, the regeneration treatment with CO 2 showed to be more e�cient than the treatment with H 2 . A deactivation model was predicted in the region of equilibrium in the catalytic activity, which is associated with the appearance of a residual activity, which decreases with increasing WHSV variable.


Introduction
Among the environmental impacts caused by the increase in fuel demand, the greenhouse effect is currently one of the most studied and important.This effect is promoted by several pollutant gases, although the main pollutants are carbon dioxide (CO 2 ) and methane (CH 4 ) [1,2].In this way, methane reforming with CO 2 arises as an interesting alternative to enhance the added value of these pollutants through syngas production.The syngas consists of a mixture of carbon monoxide (CO) and hydrogen (H 2 ).Besides the H 2 production, the combination of CO and H 2 can be applied to the Fischer-Tropsch synthesis for the production of liquid hydrocarbons [3].
Thus, the pursuit of catalysts that present high activity but also possess signi cant resistance to coke deposition is justi ed [6,7].Regarding the applicability, nickel catalysts are the most used due to their good cost-bene t, however, they do not present an elevated coke resistance.In face of such an issue, studies aiming at catalyst modi cation through the addition of lanthanum (La), cerium (Ce), and zirconium (Zr) oxides have been investigated to minimize the catalytic deactivation by coke deposition [8][9][10].La increases the active phase dispersion, besides providing oxygen to react with the formed coke, thus improving the catalyst activity and stability [11].Moreover, cerium also enhances the active phase dispersion and decreases the catalyst deactivation by coke [9].In addition, zirconium acts in Ni species reducibility, besides providing basic sites for CO 2 adsorption in DRM reaction [4].
Furthermore, other studies related to reactional medium modi cation have been developed since most studies about DRM reactions are performed in xed-bed reactors with powder catalysts [10,12,13].
Therefore, microchannel reactors arise as an alternative of great importance in the pursuit of modi cation in the reactional bed.The small dimension of the channels of the structured catalyst provides a high surface area and reduces the heat and mass transfer limitations [14,15].The monolithic reactors present high thermal conductivity and lower pressure drop, improving the catalyst application [15].Additionally, this system improves thermal integration and increases the DRM e ciency due to the surface area/volume ratio [14,39].Katheria, Deo and Kunzru (2018) [15] studied the effect of the application of the structured system through FeCralloy metal monolith performance compared with a packed bed.They observed that the FeCralloy monolith showed an increase in methane conversion compared to the other systems, thus evidencing the effects of the application of the structured system.In addition, the authors reported advantages regarding the lower pressure drop and catalytic stability over time for the Structured System, which they associate with higher e ciency to catalytic deactivation.
The aforementioned studies aim at the synergy between cost and e ciency, which is substantial for the processes.However, sometimes these factors are not su cient to mitigate the deactivation effect and increase the process e ciency.Thus, catalytic regeneration is an interesting alternative to maintain the catalytic activity of the catalyst since it is submitted to some regeneration cycles until its activity presents a signi cant decrease [16,17].The regeneration process is used to remove the coke deposited on the catalysts' surface and pores [16].Besides that, this step permits the recovery of the catalyst without modifying its structure [17].
Therefore, this work aimed at the development of alumina-supported and La 2 O 3 -promoted Ni-based catalysts applied to a conventional reactor ( xed-bed) and microreactor (structured system) and the impact of these modi cations on the activity, deactivation, and regeneration of the developed catalysts.
The effects of reactant feed ratio (CH 4 :CO 2 ), La (La 2 O 3 ) addition as a catalytic support promoter, and the application of a structured system in the dry reforming of methane were studied.Furthermore, the deactivation of the structured catalyst was investigated by applying different weight hourly space velocities (WHSV).Finally, the regeneration study was performed by the submission of the spent catalysts to CO 2 and H 2 atmospheres for the structured systems.
The feed composition manipulation aimed at the evaluation of the catalysts behavior in adverse conditions, thus seeking to maximize the hydrogen production by the increment of methane in the feed composition.Therefore, the deactivation and regeneration studies were carried out in unfavorable conditions, concerning the feed and temperature variables.These studies allowed us to verify the real impact of the structured system and to compare the effects of La addition to the catalyst.The lanthanum nitrate hexahydrate (La(NO 3 ) 3 .6H 2 O, 99.99%, NEON) was used as the precursor.
Subsequently, the support was dried in a mu e at 120°C for 12 h and then calcined at 750°C for 4 h with a heating ramp of 2°C min − 1 .
After the support preparation, the catalysts were synthesized by the all-in-one method [18].The La-Al and Al supports were impregnated with 15 w% of Ni using the nickel nitrate hexahydrate (Ni(NO 3 ) 2 .6H 2 O -99.99% SIGMA-ALDRICH) as the metal precursor.Besides, polyvinyl alcohol (PVA) and colloidal alumina (AL20 NYACOL Nano Technologies Inc.) were applied to emulsion stabilization.In the end, the suspension pH was adjusted to 4 with nitric acid (HNO 3 , 0.5 M -65% -MODERNA).After that, the suspensions remained under magnetic stirring at 200 rpm for 24 h before use.Both suspensions were used to obtain the powder catalysts and to coat the monoliths (section 2.2).To obtain the powder catalysts, the suspensions were heated in a water bath to evaporate most of the water and then dried in a mu e at 120°C for 12 h with a ramp of 10°C min − 1 .Finally, the catalyst precursors were calcined in a mu e at 550°C for 4 h with a ramp of 2°C min − 1 .

Construction and Fecralloy® monoliths coating
The structured support was made by plates of Fecralloy® from Goodfellow® company.Initially, Fecralloy® blades of 3 cm in width and blades with two different lengths were cut: 15 cm and 20.5 cm.
The former was used as the at blade, while the latter was used as the corrugated blade.Subsequently, the sheets were washed with distilled water, detergent, and acetone to remove the soil.After the drying of the sheets, the blades of 20.5 cm in length were inserted into a homemade machine displayed with two corrugated rolls to form the microchannels.At the end of this process, the at and corrugated sheets were rolled up and tied down with kanthal® wire, thus obtaining the monoliths.Finally, the monoliths were calcined at 900°C for 22 h with a heating ramp of 10°C min − 1 .
After the stabilization of the suspension prepared by the all-in-one method (section 2.1), the metallic substrate was coated by the washcoating technique with the aid of a homemade washcoating machine.
The monolith was immersed in the suspension and remained immersed for 1 min.Afterward, the suspension excess was removed by air ow centrifuged.After that, dried in a mu e at 120°C for 15 min.This procedure was repeated until the desired mass, 100 mg was reached.Finally, the monolith was calcined in a mu e at 550°C for 4 h with a heating ramp of 5°C min

Catalyst characterization
N 2 physisorption of the catalysts and supports was performed in the Quantachrome Autosorb-iQ Instruments apparatus.Prior to the analysis, the samples were degassed at 180°C for 3 h.The Brunauer-Emmett-Teller (BET) equation was used to calculate the surface area, while the pore volume and average size were obtained by the Barrett-Joyner-Halenda (BJH) method.
The crystalline structure of the supports and catalysts was determined by X-ray diffraction (XRD) in the D8 Advance diffractometer from Bruker with Cu-Kα radiation (λ = 1.5406Å) with a monochromator and Bragg-Brentano con guration with a step size of 0.05° and an acquisition interval of 5 s.The analysis of the catalysts was carried out in the range from 40° to 50° with a step size of 0.02° and an acquisition interval of 5 s.The average NiO crystallite size (D c ) of the Ni/Al and Ni/La-Al catalysts was calculated by the Scherrer equation (Eq. 1) [19].
The atomic absorption spectroscopy (AAS) was performed in a Shimadzu Atomic Absorption Spectrophotometer AA-6300 apparatus, with synthetic air-acetylene ame.To determine Ni content on the catalysts, the digestion of the samples was performed and the formed solutions were thermal treated in an ultrasound bath at 75°C for 1.5 h and then diluted to reach the theoretical concentration of 7.5 ppm for both catalysts.Finally, the analysis was performed in duplicate.
Thermogravimetric analysis (TG) of the spent catalysts to investigate coke deposition was carried out in Netzschb-Leading Thermal Analysis STA 449F3 equipment.The samples were submitted to a synthetic air ow of 50 mL min − 1 and exposed to the temperature range from room temperature to 1000°C with a heating rate of 10°C min − 1 .
Scanning electron microscopy (SEM) images were obtained by a TESCAN VEGA3 apparatus, corresponding to a tungsten thermionic emission system.Raman spectroscopy analysis was performed in a Horiba iHR320 spectrometer, with a laser of 671 nm, and a range of 100-4000 cm − 1 .The spectra were obtained in the region from 1000 to 3500 cm − 1 .
Temperature programmed reduction (TPR) was executed to investigate the reducible NiO species and the catalyst reducibility.The analysis was carried out in a Chemisorb 2720-TPx System from Micromeritics®.The supports and catalysts were heated from room temperature to 900°C with a heating ramp of 10°C min − 1 and submitted to a 10 v% H 2 /Ar ow of 20 mL min − 1 .
To evaluate the adhesion of catalysts on the monoliths was performed according to the methodology proposed by Aguero et al. (2011) [20].Two coated monoliths, previously dried at 200°C for 30 min, were immersed in hexane (P.A.) (99%, MODERNA) and subjected to an ultrasonic bath (650 W and 50-60 Hz -ELMASONIC EASY, ELMA) for 30 min.Subsequently, the monoliths were dried in an oven at 200°C for 60 min, weighed, and the adhesion (%) was calculated using Eq. 2. 2 Where, (g) is the initial mass of the covered monolith; (g) is the nal mass of the coated monolith and cat.(g) is the total mass of catalyst deposited on the monolith.

Catalytic activity
The dry reforming of methane reaction was carried out in a lab-scale quartz tubular reactor with dimensions of 600 mm in length, 16 mm inner diameter, and 20 mm outer diameter.Besides the reactor, the reaction unit possesses a vertical tubular oven (SANCHIS®, 2585-k type model), a mass ow meter (ΩOmega® Engineering Inc.), a mass ow controller (ΩOmega® Engineering Inc.), and an Agilent Technologies® 6890N series 7890B gas chromatograph to an online analysis of the gases produced during the reaction.The reaction process was conducted in a bench-scale reaction unit which displays reactional feed gas cylinders (CH 4 , CO 2 , and H 2 ), and inert gas cylinders (Ar and N 2 ).The schematic owchart of the processing unit is presented in Fig. 1.
Before the reaction, the catalyst was reduced in situ under H 2 ow (200 mL min − 1 g cat −1 ) 1 bar and at 700°C for 2 h.The gases were set at a molar ux value for CH 4 and CO 2 in the various compositions in this study to determine the initial moles of each reaction.The reactions were then performed in the gaseous phase at 650°C, atmospheric pressure, and several compositions of the CH 4 :CO 2 mixture.
The catalytic performance was evaluated by the determination of weight hourly space velocity (WHSV) (Eq. 3) and conversion (Eq.4) [21][22][23]. ( Where i is CO 2 or CH 4 components, m cat is the catalyst mass (g), F in is the total volumetric gas ow in the reactor inlet (L h − 1 ), a i is the catalytic activity of the i component (dimensionless), C i in is the inlet molar concentration of the i reactant (mol L − 1 ), and C i out is the outlet molar concentration of the i reactant (mol A blank test was also performed and no conversion was observed, hence con rming that the conversions shown in section 3 are strictly related to the performance of the catalysts.

Effect of feed composition
The evaluation of the reaction gas composition is aimed at investigating the in uence of different CH 4 :CO 2 ratios in two reactional systems: powder and structured systems.Five different CH 4 :CO 2 volume ratios were employed, which vary from 0.7 to 4. Initially, the powder system was used and the results obtained for Ni/Al and Ni/La-Al catalysts were evaluated.The reactions were carried out at a total volumetric gas ow of 50 mL min − 1 , maintaining the WHSV constant (30 L g cat −1 h − 1 ) and a xed catalyst mass of 60 mg.To investigate the in uence of the reaction system, the monoliths coated with Ni/Al and Ni/La-Al catalysts were used and the assays were performed at a CH 4 :CO 2 ratio of 4, the composition that presented the lowest catalytic activity in the powder system, as presented in section 3.5.1.The tests were carried out in the same aforementioned operational variable values.

Evaluation of catalytic deactivation
The catalytic deactivation study was performed with Ni/La-Al and Ni/La-Al[S] (structured system) catalysts.To obtain the deactivation data, the catalysts were submitted to adverse reaction conditions regarding feed composition (CH 4 :CO 2 ratio of 4) and temperature (650°C).These assays were carried with a xed catalysts mass of 100 mg, and varying WHSV values (20, 30, and 40 L g cat −1 h − 1 ).

Evaluation of catalytic regeneration
In this step, Ni/La-Al[S] catalyst (structured system) and CH 4 :CO 2 ratio of 4 were employed.The catalyst mass of 100 mg coated onto the monoliths was maintained for all tests to obtain the experimental data.
The regeneration study was carried out at 700°C and a WHSV value of 20 L.g cat −1 .h− 1 since some papers [16,17,19] veri ed that this temperature is the optimal regeneration temperature.Two assays with different regenerating gas (CO 2 and H 2 ) were performed.Each assay was conducted with Ni/La-AL[S] with 3 cycles of use.The catalyst regeneration took place after 17 h of reaction in each cycle.The regeneration velocity WHSV reg of 20 L g cat −1 h − 1 was used.Therefore, the greatest regenerating gas in terms of catalytic activity recovery was determined [16, 19].

Modeling of catalytic deactivation
The deactivation study of the Ni/La-Al[S] catalyst was evaluated using residual activity modeling, DMRA (Deactivation Model of Residual Activity), proposed by De la Cruz, Martinez and Gracia (2020) [23] and Zambrano et al. (2019) [22].Two models were implemented (Eqs.s considered constant when investigating the reaction under isothermal conditions and under continuous ow and composition of the reaction gases [22].For both the models, the activity was determined over time, following (Eq.7), thus it was possible to perform an optimization of the data that best ts the proposed equation.In addition, statistical parameters were used for the evaluation of the DRMA1 and DRMA2 models: p_Value, Fstatistic, R2 and RMSE.Thus, enabling assertive conclusions regarding the proposed modeling.

Catalyst characterization
Figure 2a illustrates the N 2 adsorption/desorption isotherms of the supports and catalysts.According to IUPAC (International Union of Pure and Applied Chemistry), both support and catalyst isotherms are characteristics of IV-type isotherm, regarding mesoporous materials [24].Moreover, the hysteresis loop can be observed by the adsorption and desorption curves deviation due to the capillary condensation.This phenomenon takes place from the change of the gas phase to the liquid phase inside the pore since its pressure is lower than its saturation pressure, which depends on pore shape and size [25].
It can also be observed in Fig. 2a that both support and catalyst isotherms show the H1-type hysteresis, according to the IUPAC classi cation.This hysteresis type is characteristic of narrow pore distribution [25].This is corroborated by the pore distribution illustrated inside Fig. 2a.The surface area (S BET ), pore volume (V p ), and size (d p ) were determined from the obtained data and isotherms, which are presented in Table 1. Figure 2b shows the H 2 -TPR pro les of the La-modi ed support, and Ni/Al and Ni/La-Al catalysts.Both catalysts present four reduction events and the TPR pro les were divided into 3 regions.In region I, a sharp peak for Ni/La-Al catalyst and a spread peak for Ni/Al catalyst at 350°C are observed.These reduction peaks can be attributed to more accessible NiO species, which present weak interaction with the catalytic support.Region II comprehends the reduction peaks between 450°C and 700°C, which can be related to NiO species with moderate interaction with the support [7,27].
Moreover, the presence of one reduction peak was observed in region III for both catalysts [27,28].The reduction peak above 800°C regards the nickel aluminate (NiAl 2 O 4 ) species, thus presenting strong interaction with the catalytic support [29].It can also be observed in region III that the reduction peak areas are different for the catalysts, indicating that Ni/La-Al catalyst presented lower H 2 consumption than Ni/Al catalyst.This suggests the lower presence of NiAl 2 O 4 in the La-promoted catalyst.This could also highlight the bene cial effect of the lanthanum, regarding catalyst stabilization.Figure 2b also shows that most NiO species are reduced in region II, hence justifying the in situ reduction at 700°C employed in this work.
Finally, it is noted that La-Al support did not present any reduction peak, thus indicating no presence of any reducible species.Therefore, the absence of reducible La 2 O 3 species below 900°C can be justi ed by the strong bond between La and O 2 [11].
XRD spectra of the supports and catalysts are shown in Fig. 2c.The qualitative analysis of the compounds presented in all samples was performed based on the crystallographic charts from Inorganic Crystal Structure Database (ICSD): 30025 (Al 2 O 3 ); 11262 (NiAl 2 O 4 ); 24693 (La 2 O 3 ) and 9866 (NiO).
It can be observed in Fig. 2c the diffraction peaks of γ-Al 2 O 3 at 37.5°, 45.6°, and 66.7° for all samples.In to the Scherrer equation (Eq.5) and the NiO crystallite size was estimated (Table 1).Table 1 also indicates that Ni/La-Al catalysts present a smaller NiO crystallite size.Therefore, La addition to the catalytic support implies a bene cial effect since the small crystallite size is directly related to the high dispersion of the active phase [29,31].
The Ni content in the catalysts was determined by atomic absorption spectroscopy analysis.Both catalysts presented similar values of Ni content and close to 14 w%.The real value is softly below the nominal value (15 w%) and it can be attributed to a systematic error in the Ni impregnation step, which can be related to humidity present in the active phase precursor (Ni(NO 3 ) 2 .6H 2 O).Nevertheless, the percentage deviation is approximately 8%, which is an acceptable value.Besides, this does not cause any signi cant effect on other catalyst evaluations.In order, to evaluate the adherence of the coating to the monoliths, the adherence test was performed for each catalyst in duplicate, following the methodology described in item 2.3.The Ni/Al catalyst showed 93% of adherence in the metallic substrate, while the Ni/La-Al presented 91%.
The results show that the formulations used for catalytic suspensions led to optimal adhesion on the Fecralloy metal monoliths, presenting values above 90%, for both catalysts [18].These values revealed that, even when the system is exposed to aggressive conditions (ultrasonic bath with 60w power), when compared to the handling in the reaction tests, there is no signi cant detachment of the coated catalyst.Thus, it is possible to state that during the catalytic activity tests, which occur under extremely milder conditions, there was no loss of material by detachment due to the handling and ow of the gaseous reactants.

In uence of feed composition
The produced catalysts were evaluated in terms of H 2 :CO ratio; CH 4 and CO 2 conversion (Fig. 3).this decrease in the CH 4 :CO 2 ratio favors the RWGS reaction due to the possible formation of traces of water, which tends to suppress catalytic deactivation since it acts as a gasi er.This is also observed in methane steam reforming [14,15].Furthermore, the CO 2 excess present in the CH 4 :CO 2 ratio of 0.7 could have provided a regenerating atmosphere to the catalyst and thus minimizing the catalyst deactivation by coke deposition [21,32].
However, Figs. 3 also illustrate that the increase in the CH 4 :CO 2 ratio reduces CH 4 conversion and decreases CO 2 conversion, respectively, for both catalysts.The CH 4 conversion decrease is strongly related to the main deactivation cause reported in the literature [5,6,10,16], which is coke deposition due to CH 4 decomposition (Eq. 3) and Boudouard reaction (Eq.4).The CO 2 conversion increment can be attributed to its function as a sequestering agent of carbonaceous material formed on the catalyst surface.
Finally, the evaluation of the performance of the catalysts makes evident that Ni/La-Al catalyst (Fig. 3) present the greatest catalytic activity for all the tested gas feed proportions.In the most adverse condition (CH 4 :CO 2 ratio of 4), this catalyst showed activity approximately 10% higher than Ni/Al catalyst.
Therefore, the bene cial effect of the La 2 O 3 impregnation on the catalyst is evident since it provides O 2 that assists the oxidation of the coke formed during the reaction [11,33].
The evolution of the H 2 :CO ratio is also presented in Fig. 3

Catalytic evaluation of structured catalyst
Figure 4 displays the reaction data of powder and structured Ni/Al and Ni/La-Al catalysts.The catalytic tests were carried out in the CH4:CO2 ratio of 1 over 17 h of reaction.It is noted that the structured system showed greater performance for both catalysts, presenting a catalytic activity 20% higher than the powder system.Nonetheless, the difference in the catalytic activity between the powder and structured La-promoted catalysts decayed to approximately 10% in the reaction.This can be justi ed by the presence of the La promoter, which leads the catalyst to a greater deactivation resistance by coke formation, as mentioned before.
To prove the e ciency of the structured system, reactions with Ni/La-Al catalysts in powder and monolithic systems were performed with the CH 4 :CO 2 ratio = 4, which is the most aggressive condition for catalyst deactivation.This condition provides an excessive coke formation in a short time, an effect typically observed when CH 4 :CO 2 ratios higher than one unit are employed [34] (Yasyerli et al. 2011).The results of these tests are exhibited in Fig. 5.
It can be veri ed in Fig. 5 that both systems present similar behavior during the catalyst deactivation.Nevertheless, the monolithic system shows slightly higher activity than the powder system, even after 17 h of reaction.Therefore, this corroborates the results of Fig. 4, thus making evident the bene ts of conversion of the monolithic system.In addition, the structured system showed no bed compaction while the powdered system showed high compaction of the catalyst bed after the reactions.
The slightly higher catalytic activity of the structured system can be justi ed by some properties of the referred system.The rst property consists of the intensi cation of heat transfer, avoiding cold spots over the reactional bed, which is of great importance for endothermic reactions.Besides, the microchannel system also provides the intensi cation of mass transfer and its application prevents the blockage of the reactional bed due to coke formation, hence enabling the reaction occurrence [15,35].

Evaluation of catalytic deactivation
The catalytic deactivation study was carried out with Ni/La-Al structured catalysts with the CH 4 :CO 2 ratio of 4, and using the WHSV values of 20, 30, and 40 L g cat −1 h − 1 .This gas feed ratio provides the greatest deactivation rate, thus allowing the best investigation of the catalyst and system behavior under harsh deactivation conditions.
The conversion values obtained for WHSV = 20, 30, and 40 L g cat −1 h − 1 , after 17h reaction, were 31%, 25% and 18%, respectively.A decrease in the deactivation rate was also observed for all spatial velocities when comparing the beginning and the end of the reaction.This decrease in the deactivation rate can be associated with the La 2 O 3 effect that hinders the catalyst deactivation by coke formation due to O 2 availability [10,11].It was also seen that the WHSV of the highest value (40 L g cat −1 h − 1 ) exhibited the greatest deactivation.This can be attributed to the increase in WHSV in the reactional bed, which reduces the residence time of the reactants and, consequently, provides little time for possible partial carbon gasi cation by CO 2 .By contrast, the lowest WHSV value (20 L g cat −1 h − 1 ) showed the lowest deactivation at the beginning of the reaction, which is justi ed by the highest residence time, hence enabling the carbon gasi cation.Additionally, the lower deactivation of the lowest WHSV can also be related to the higher contact time between the reactant and the catalytic surface, leading to higher conversion values.

Study of the regenerating atmosphere
The Ni/La-Al structured catalyst was used in the regeneration study under CO 2 and H 2 atmospheres in the same operating conditions and for 3 cycles of use, which are displayed in Fig. 6.
Figure 6 shows, in the rst cycle, the catalyst fall from the conversion of 70% to a value close to 30% after 17 h of reaction.However, it is noted that the CH 4 conversion reaches a threshold of approximately 15% in the last hours of the second cycle.This is observed for both atmospheres, although the regeneration with CO 2 exhibits a slight decrease in the deactivation rate.The greater e ciency of the CO 2 regeneration in comparison with other atmospheres is also reported by other authors [16,17].However, the similar results presented in this work for both regeneration atmospheres can be attributed to the process intensi cation, provided by the structured system.
It is worthy to mention that the use of H 2 in the regeneration step consists of a hydrogenation reaction, which is an exothermic reaction.This additional energy can promote a greater catalyst sintering, thus decreasing its activity, which is a drawback of the use of this gas in this operation.Besides, two other factors can be considered to justify the use of CO 2 to detriment of the use of hydrogen.The rst is related to CO 2 reuse through the carbon capture, storage, and utilization (CCUS) technique which aims at the application of this pollutant in a transformation and/or reuse route.The second regards the H 2 to be a product of high value with an important role in the chemical industry as the application in fuel cells and its use in processes that require hydroprocessing [1,2,36].

Post-reaction characterization
The spent catalysts from the tests with CH 4 :CO 2 ratios of 0.7 and 4 were characterized via XRD analysis and the diffractograms are shown in Fig. 7.The qualitative analysis of the crystalline structure of the catalysts was performed based on the ICSD crystallographic charts: 30025 (Al 2 O 3 ), 646085 (Ni 0 ), 24693 (La 2 O 3 ), and 230104 (C).The XRD spectra of the fresh catalysts are also plotted for comparison.
Figure 9 illustrates a characteristic peak of carbon (C) for all spent catalysts at the Bragg re ection of 2θ = 26.2°,corresponding to the crystalline plane (0 0 2).This carbon peak is characteristic of carbonaceous species with a speci c degree of crystallinity [10].Other authors also associate this peak with a carbon species with elevated crystallinity [21].
It can also be observed in Fig. 7 the presence of two characteristic peaks of metallic nickel (Ni 0 ) at the Bragg re ection of 44.5° (1 1 1) and 52° (0 2 0).However, these peaks were not observed for the fresh catalysts.This demonstrates that the in situ reduction carried out at 700°C was e cient, corroborating the TPR results (Fig. 3), which shows that most NiO species are reduced in region II (400-700°C) [28].
Figure 8 exhibits the spectra obtained by the Raman spectroscopy to investigate the carbonaceous material deposited on the spent catalysts.
In Fig. 8, carbonaceous species with different degrees of crystallinity can be identi ed.This can be a rmed due to the presence of the peaks in 1320 and 1580 cm − 1 , which represent the D and G band, respectively, being generally attributed to carbon species with sp 2 bonding.In addition, D and G bands indicate different characteristics of carbon species.The D band indicates imperfections in the materials and the G band represents vibrations on C-C (carbon-carbon) bonding [10].Furthermore, the ratio between D and G bands intensity (I D /I G ) indicates the graphitization degree of the sample and, consequently, its hardness.Thus, the lower the I D /I G ratio the higher the graphitization/crystallinity degree of the sample.
However, the intensity of the peak associated with the 2D band is related to the graphene layers.In other words, a bigger 2D band implies a greater possibility of graphitic material [10].
The Table 2 shows the I D /I G ratio for each sample was evaluated, as well as the characteristic intensity of the 2D band.To calculate the I D /I G ratios, the adjustment of each characteristic peak through the deconvolution of the curves was performed.It can also be noted in Table 2 that the La-promoted catalyst presents lower I D /I G ratios for the same CH 4 :CO 2 ratio, which is attributed to the higher graphitization degree of the Ni/La-Al catalyst.
Nevertheless, this effect can also be explained by the performance of lanthanum as an oxygen supplier, thus promoting partial oxidation of malleable carbon present in the sample [11].
To investigate the carbon deposition, TG analysis of the spent catalysts applied to methane dry reforming reaction was carried out (Fig. 9).In total, 10 samples were analyzed at different feed ratios (CH The scanning electron microscopy was used to obtain the textural and morphological characteristics of the fresh and spent catalysts (Ni/Al and Ni/La-Al) in DRM reaction as presented in Fig. 10.Figures 10a   and 10b show that fresh Ni/Al and Ni/La-Al catalysts present a uniform morphology, characterized by the small agglomerates of particles [10,37].It is observed that both catalysts showed the growth of lamentous carbon over the surface of the catalysts, Fig. 12.These laments characterize a certain crystallinity degree for both catalysts [10,16].This is corroborated by the XRD results that showed carbon with higher crystallinity and, consequently, more resistance towards oxidation.In addition, the reaction with the CH 4 :CO 2 ratio of 4 (Figs.10c and 10d) presented a great amount of lamentous carbon on the surface of both catalysts.This is associated with the fact that this feed ratio was the most aggressive condition, among those studied in this work, regarding catalyst deactivation.Therefore, it is veri ed that the oxidizing function of La 2 O 3 is not relevant against coke deposition for this adverse condition.

Modelling catalytic deactivation
Figures 11 and 12 show the results of tting the DMRA1 and DMRA2 models to the experimental CH4 activity data, respectively.In Figs.11 and 12 it was observed a good representation of the experimental data from DMRA 1 and DMRA 2 models.Furthermore, it was possible to verify that the models, for the different spatial velocities employed, predict a sharp drop in catalytic activity followed by stability (equilibrium region).This region was associated with the appearance of a residual activity that can be attributed to a reversible deactivation process such as coke, or even the presence of non-deactivated active sites [38].However, as the WHSV increases, a deviation from the DMRA 1 and DMRA 2 models is observed, indicating that the deactivation continues, but at a slower rate.
Tables S1 and S2 show the parameter values obtained from tting the experimental data to the DMRA 1 and DMRA 2 models, respectively.DMRA 1 (Table S1) a growth of the deactivation constant (φd) was observed with increasing WHSV.This behavior may be related to higher deactivation due to the higher ow rates employed.On the same hand, similar results were observed by De la Cruz, Martinez and Gracia (2020) [23], in which under conditions of lower deactivation, they obtained values for the deactivation function in the same orders of magnitude presented in this work.Moreover, the value of the residual activity (as), which decreases as the WHSV increases, is in accordance with the expected behavior, since higher as occur under conditions of less deactivation [23].
The results of the MDAR 1 model and Table S1 obtained R² values ≥ 0.90, indicating good ts of the experimental data.A low RMSE value was observed for the lower WHSV, indicating better representation of the experimental data under these conditions.Whereas, the values of Fstatistic < Fcritical (Fcritical = 1.841; degrees of freedom = 30) indicated that there were no signi cant differences between the experimental data and those predicted by the model, however, the 0.0588 ≤ p-value ≤ 0.1021 indicated that the con dence level of the model ranged from 89.79-94.12%,below 95%.
In Table S2, it was also possible to observe that, like DMRA1, that DMRA2 showed good ts of the experimental data in terms of R² ≥0.90.However, it was possible to verify that R

Conclusions
The study of activity, deactivation and regeneration of nickel-based catalysts in powder and monolithic systems in the DRM reaction has been investigated.Regarding the catalytic evaluation of the powder system, it was concluded that the decrease in CH 4 :CO 2 ratio increased CH 4 conversion for both catalysts, and the CO 2 excess in feed ratio performed in the oxidation of the formed coke.The structured catalysts presented better catalytic performance than the powder catalysts.The deactivation study of Ni/La-Al structured catalyst showed that La addition decreased the deactivation rate, minimizing the deactivation effect of coke formation.In the regeneration study, it was noted that the CO 2 atmosphere provided superior performance in detriment to the H 2 atmosphere, being the former the best atmosphere for the activity recovery of the studied catalysts.Regarding the post-reaction characterization, the XRD spectra indicated a characteristic peak of a carbon species of high crystallinity degree.The Raman spectroscopy veri ed that La 2 O 3 addition decreased the carbon hardness.In the deactivation modeling, observed that the data was best tted to the MDAR1 model according to the best statistical values, such as R 2 _DMRA1 > R 2 _DMRA2.While the behavior of the growth of φ d value and the decrease of a s towards the increase of WHSV could be associated with higher deactivation.

Supplementary Files
This is a list supplementary les associated with this preprint.Click to download. supplementarymaterial.docx

1
Catalyst preparation Two Ni-based catalysts for application in the dry reforming of CH 4 were synthesized: Ni/γ-Al 2 O 3 (Ni/Al) and Ni/La 2 O 3 -γ-Al 2 O 3 (Ni/La-Al).Prior to Ni impregnation on Ni/La-Al catalyst, the La-Al support was prepared by the incipient wetness impregnation of 10 w% of lanthanum oxide (La 2 O 3 ) on γ-Al 2 O 3 support.

− 1 .
The structured catalysts were named Ni/Al[S] for the monolith coated with Ni/γ-Al 2 O 3 catalyst and Ni/La-Al[S] for the monolith coated with Ni/La 2 O 3 -γ-Al 2 O 3 catalyst.
recovery function functions are dependent on the amount of component (pi) inducing deactivation and the temperature (T).The parameter   [dimensionless] represents the residual activity and is dependent on the same variables as .However, and  can be

Fig. 2c (
Fig.2c(III), the nickel aluminate peaks are noted in the Bragg angles of 19.4° (1 0 1) and 60° (2 3 1), which are related to the strong metal-support interaction.However, Fig.2c(IV) shows a reduction in the intensity of NiAl 2 O 4 peaks after La 2 O 3 addition, thus emphasizing that its addition increases the active phase reduction.In Figs.2c (II) and 2c (IV), the diffraction peak characteristic of La 2 O 3 can be identi ed at 39.5° (1 0 2), although the other peaks of this species could not be identi ed due to the overlapping of the γ-Al 2 O 3 peaks (Santamaria et al. 2020).In addition, two diffraction peaks at 43.5° (0 0 2) and 62.9° (0 2 2) regarding NiO species are noted for Ni/Al and Ni/La-Al catalysts (Figs.4c and 4d, respectively) [28, 30].The NiO crystallite size of Ni/Al and Ni/La-Al catalysts were calculated for the Bragg re ection 2θ = 43.5°.The full width of half maximum (β) was obtained by the Origin software.The obtained values of β were 1.5814° and 1.8754° for Ni/Al and Ni/La-Al catalysts, respectively.The obtained data were applied

Figure 2 a
Figure 2

Figure 4 Comparison
Figure 4

Table 1
Textural properties of the supports and catalysts.Table1demonstrates that La 2 O 3 and active phase (Ni species) addition to the support decreased the surface area by 43% and 34%, respectively.This can be attributed to the presence of these species in alumina mesopores, thus limiting the access of N 2 molecules to the support pores[26].It can also be noted that the nickel oxide addition to La-Al support decreased the catalyst surface area by 5%, which can be due to the prior lling of the great part of the support mesopores by La 2 O 3 .This could prevent the posterior access of Ni species.
Figures 3 demonstrate the experimental conversion of the catalysts in 17h of reaction.The WHSV of 30 L.g cat −1 .h− 1 was used and ve different feed ratios (CH 4 :CO 2 = 0.7, 1, 1.5, 2.3, and 4) were investigated.It can be observed in Figs. 3 that the decrease in CH 4 :CO 2 ratio increases CH 4 conversion for both catalysts.On the other hand, CO 2 conversion tends to decrease with the decrement of the CH 4 :CO 2 ratio,

Table 2
Intensity of the D, G, and 2D bands of fresh and spent catalysts at different CH 4 :CO 2 feed ratios.

Table 2
results indicate that the increase of CO 2 proportion in the feed ratio for the same catalyst provides a decrement in the I D /I G ratio, hence suggesting an increase in the graphitization degree of the sample.This can be justi ed by the CO 2 excess during the DRM reaction oxidizing the amorphous available carbon (more malleable), thus remaining more graphitic carbon to be quanti ed, carbon with elevated hardness.
Damyanova et al. (2017)2.3, and 4) for both Ni/Al and Ni/La-Al catalysts.It is observed in Figs.9a and 9b that both catalysts showed a signi cant mass loss (approximately 85%), indicating an elevated carbon formation for all mixture compositions.Nonetheless, the data illustrate the lowest mass loss for the CH 4 :CO 2 ratio of 0.7, presenting value of 78% for Ni/Al and 73% for CH 4 :CO 2 ratio of 4 in Ni/La-Al catalyst.Thus, the lowest amount of deposited carbon, in Ni/Al catalyst, is attributed to 0.7 feed ratio, corroborating the results presented before which indicated that CO 2 excess of the referred feed composition provides the partial oxidation of the deposited carbon.This is in accordance with the discussion of Raman results (Fig.8and Table2) that associates the possibility of part of the malleable carbon being oxidized by CO 2 excess.A predominance of graphitic carbon is observed in all the samples.It is also noted that CH 4 :CO 2 composition shows no signi cant difference in TG curves.However, different behavior in TG curves of Ni/Al and Ni/La-Al catalysts can be seen.Therefore, the TG and DTG curves of both catalysts are shown in Fig.9cto evaluate the quantitative effect of La 2 O 3 as a support promoter at the CH 4 :CO 2 ratio of 4. It can be seen in Fig.9cthat Ni/La-Al catalyst presented a mass loss of 77%, while Ni/Al catalyst showed 84% of mass loss, thus indicating that Ni/La-Al catalyst presented a lower amount of deposited coke in these adverse reaction condition.Thus, La 2 O 3 addition to the catalyst minimizes the catalytic deactivation by coke formation.Similar results were reported byDamyanova et al. (2017)[6].The authors observed that cerium oxide addition to the catalyst decreased carbon deposition.
2 _DMRA1 > R 2 _DMRA2, indicating a better t with respect to DMRA2.In the DMRA2 model, again the lowest RMSE value was found for the lowest WHSV.However, the RMSE values of DMRA 1 were lower compared to DMRA2, indicating a better representation of the experimental data by this model.The DMRA 2 model also showed values of FStatistic < Fcritical.It was still possible to verify that DMRA2 presented 0.0560 ≤ pvalue ≤ 0.0984 indicating the con dence level between 90.16% and 94.40%, below 95%, but very close to those obtained for DMRA1.