Performance of NiAl2O4 spinel derived catalyst + dolomite in the sorption enhanced steam reforming (SESR) of raw bio-oil in cyclic operation

International Journal of Hydrogen Energy 58 (2024) 1526–1540 Available online 3 February 2024 0360-3199/© 2024 The Authors. Published by Elsevier Ltd on behalf of Hydrogen Energy Publications LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Performance of NiAl 2 O 4 spinel derived catalyst +dolomite in the sorption enhanced steam reforming (SESR) of raw bio-oil in cyclic operation Leire Landa * , Aingeru Remiro, Jos´ e Valecillos, Beatriz Valle, Javier Bilbao, Ana G. Gayubo Department of Chemical Engineering, University of the Basque Country (UPV/EHU), P.O. Box 644, Bilbao, 48080, Spain ARTICLE INFO Handling Editor: Dr A Bhatnagar Keywords: SESR Bio-oil Hydrogen Ni spinel Coke Reaction-regeneration cycles ABSTRACT The production of H 2 from raw bio-oil with high yield and purity requires the development of reforming technologies with low energy requirements, minimized CO 2 emissions, and stable and regenerable catalysts. This work studies the performance (activity, selectivity, stability and regenerability) in the sorption enhanced steam reforming (SESR) of raw bio-oil of a catalyst prepared by reduction of a NiAl 2 O 4 spinel together with dolomite as CO 2 sorbent. The reaction runs were carried out in a fluidized-bed reactor under the following conditions: 550–700 ◦C; space time, 0.15 and 0.30 g catalyst ⋅h/g oxygenates ; dolomite/catalyst mass ratio, 10 and 20; steam/ carbon (S/C) molar ratio, 3.4; time on stream, 50 and 300 min. The highest H 2 yield (>92 %) and purity (>99 %) in the CO 2 capture period are obtained in the 600–650 ◦C range and with a dolomite/catalyst mass ratio of 10, due to synergy between catalyst and sorbent activity. The catalyst/sorbent system can be regenerated (4 h in air at 850 ◦C and subsequent reduction at 900 ◦C) and used in the successive reaction-regeneration cycles. The results are of relevant interest to progress towards scale-up of this process, which combines sustainable production of high purity H 2 from biomass with CO 2 capture. 1. Introduction The production of renewable hydrogen is a priority goal in the global economic and political strategies for carbon neutrality and for a rapid transition to clean energy [1]. Hydrogen is a clean energy carrier, with a great potential to meet growing demand in the transport, industrial and power generation sectors. It is also a raw material in growing demand for fertilizer production, and its availability is key to the feasibility of producing fuels and synthetic commodities (such as methanol, dimethyl ether, olefins, aromatics or fuels) from CO 2 [2,3]. It is also important as a buffer to non-dispatchable renewable energy (wind, solar) [4]. The establishment of the H 2 economy through electrolysis technologies requires a huge multidisciplinary effort, including the application of advanced technologies, the development of stable and regenerable catalysts and advanced materials (nanostructured and composites) for H 2 production [5] and storage [6–10], combined with the efficient use of renewable energy [11]. In the transition period before the availability of green H 2, the production of H 2 from lignocellulosic biomass is of interest [12,13]. The cost of H 2 production from biomass with different technologies is higher than that of methane steam reforming [14], but emission taxes and CO 2 capture and storage costs are lower [13]. Among the different technologies, steam reforming (SR) of bio-oil is an alternative to pyrolysis/gasification of biomass and in-line SR of volatiles [15, 16]. Sorption Enhanced Steam Reforming (SESR) of bio-oil has great appeal for avoiding CO 2 emissions through its in situ capture [17–19] and produces H 2 of higher purity than in SR. Bio-oil is obtained from fast pyrolysis of lignocellulosic biomass with a high yield, its production can be carried out off-site, with simple technologies and reduced environmental impact [20] and its transport to the valorization unit (for its large-scale reforming) is easier than that of the original biomass. The SR or SESR of bio-oil avoids the costly elimination of the high water content required for its use as a fuel or for hydrocarbon production [21], and the by-product handling required in the separation of some of its oxygenates compounds [22]. The overall reaction for SR or SESR of bio-oil involves the reforming reaction to produce (CO +H 2 ) (Eq. (1)) and the subsequent water-gas-shift (WGS) reaction (Eq. (2)): CnHmOk+ (n-k)H2O → nCO +(n+m 2-k)H2(1) CO +H2O ↔ CO2+H2(2) Thus, the overall equation is: * Corresponding author. E-mail address: [email protected] (L. Landa). Contents lists available at ScienceDirect International Journal of Hydrogen Energy journal homepage: www.elsevier.com/locate/he https://doi.org/10.1016/j.ijhydene.2024.01.228 Received 18 October 2023; Received in revised form 3 January 2024; Accepted 20 January 2024 International Journal of Hydrogen Energy 58 (2024) 1526–1540 1527 CnHmOk+ (2n-k)H2O → nCO2+ (2n +m 2-k)H2(3) H 2 yield is also affected by secondary reactions such as decomposition/cracking (Eq. (4)), reforming of decomposition products (CH 4 and hydrocarbons, Eqs. (5) and (6), respectively), and interconversion of oxygenates (Eq. (7)). CnHmOk→ CxHyOz+gas (CO,CH4,CO2,CaHb,H2,…) + coke (4) CH4+H2O ↔ CO +3H2(5) CaHb+aH2O → aCO + (a+b/2)H2(6) CnHmOk→ CxHyOz(7) Moreover, the reactions of coke (C) formation and gasification (Eqs. (8)–(10)) should be considered as they affect catalyst stability. Methane decomposition, CH 4 → 2H 2 +C (8) Boudouard reaction, 2CO ↔ C +CO 2 (9) Coke gasification, C +H 2 O → CO +H 2 (10) Using a sorbent (taking CaO as example), the CO 2 is retained as: CaO +CO2↔ CaCO3(11) The existence of this reaction (Eq. (11)) in the SESR process, in addition to avoiding CO 2 emissions (with the consequent reduction of CO 2 separation and storage costs compared to the SR process), offers the following advantages over the SR process: i) improves the H 2 yield and selectivity by shifting the equilibrium of the WGS reaction (Eq. (2)), producing H 2 of higher purity; ii) reduces the energy requirement (since the decarbonation reaction is exothermic); iii) decreases the reaction temperature, thus reducing catalyst sintering (a critical factor in its selection). The release of high purity CO 2 in the regeneration of the sorbent by calcination (reverse of Eq. (11)) will facilitate its capture and subsequent valorization. A detailed review of sorbents has been carried out by Dou et al. [23], with calcium-based sorbents being the most widely used [24]. The high CO 2 sorption capacity at high temperatures and the low cost of CaO (easily prepared from limestone) make its use more interesting than that of alkaline ceramic sorbents (such as Li 2 ZrO 3 ). The problem of losing CaO sorption capacity during cyclic carbonation/decarbonation operation has been addressed by incorporating CaO into other support materials and reactivating it with different thermal and chemical treatments [25]. SESR has been studied mainly for methane (SE-SMR process) [24–27], which is considered to be a more efficient technology than the combination of steam reforming of methane (SRM) and carbon capture and storage (CCS) units, commercially developed for blue H 2 production [28]. Using Ni and CaO based solids, SE-SMR is preferentially operated around 650 ◦C, at atmospheric pressure and with the S/C (steam/- carbon) molar ratio in the range of 3–4 [29]. Under these conditions, H 2 purity of over 95 % (dry basis) is achieved, compared to 76 % in SR. The scale-up of the process is studied by simulation with a dual fluidized-bed system combining a reformer and a calciner with catalyst circulation between the two units [30]. Thus, the activity of the sorbent is recovered in the calciner and the homogeneous temperature in each unit (higher in the calciner unit) favors its control and that of the heat transfer of the catalyst/sorbent stream (key to minimize the external energy input). Due to the complexity of bio-oil (a mixture of oxygenates of different nature) and its handling, SESR has generally been studied for pure model compounds, such as acetic acid [31–35], ethanol [36,37], phenol [38,39], glycerol [40,41] and toluene [42], and for mixtures of oxygenates [32,43]. In general, Ni-based catalysts have been used with different materials containing CaO as CO 2 sorbent, and also bifunctional composites (usually based on Ni over CaO and with different modifications) have been used [36,44–47]. The objectives of these works have focused on the optimization of the catalyst composition and the reaction conditions, although the results are different and the appropriate conditions to obtain a H 2 purity above 95 % depends on the nature of the oxygenates. Thus, this nature has a great influence on the stability of the catalysts, which has also been established in the SR of oxygenates in bio-oil [48]. Experimental studies on the SESR of the aqueous fraction of bio-oil are very scarce [49–51]. Remiro et al. [50], using a Ni/La 2 O 3 –Al 2 O 3 catalyst and calcined dolomite as sorbent, observed a significant role of the catalyst/sorbent mass ratio due to the active role of dolomite for oxygenates cracking and also for reforming reactions. A suitable balance was stricken at 600 ◦C for catalyst/dolomite mass ratios ≥0.17 with a H 2 yield around 99 %. The SESR of raw bio-oil offers a higher yield referred to biomass, as established by thermodynamic studies [52] and experimentally [35,53]. Landa et al. [53] obtain high H 2 purity using Ni/Al 2 O 3 (obtained by reduction of a NiAl 2 O 4 spinel) as catalyst and dolomite as sorbent, with better performance than other catalyst +sorbent systems, such as Ni/CeO 2 +dolomite or Ni/Al 2 O 3 +CaO/mayenite. Iliuta et al. [54] also determine the lower catalyst deactivation in SESR of bio-oil compared to SR by simulating the process in a packed-bed reactor, using the kinetic model of Gayubo et al. [55] for the reactions involved in SR of bio-oil (on a Ni/La 2 O 3 - α Al 2 O 3 catalyst) and the kinetics of CO 2 adsorption over Li 2 CuO 2 . Considering the interest of activity recovery by catalyst regeneration for the viability of the SESR process, Acha et al. [43] have verified the importance of the Ni catalyst support and the composition of the dolomite used as sorbent for the activity recovery of both materials used in the SESR of acetic acid and a mixture of oxygenates. The results highlight the importance of a support that minimizes sintering and coke deposition, while dolomites lose their capture capacity upon reactivation. Li et al. [35] investigated the stability of a Ni/Ce 1⋅2 Zr 1 Ca 5 bifunctional catalyst-sorbent (Ce/Ca ratio of 0.24) in the SESR of acetic acid and poplar sawdust pyrolysis oil in a packed-bed reactor over 15 reaction-regeneration cycles. While an excellent catalytic performance was reported after 15 consecutive cycles for the SESR of acetic acid (550 ◦C, S/C of 4 and LHSV of 0.48 ml/g catalyst ⋅h), a decrease in CO 2 adsorption capacity and H 2 yield was observed in the SESR of bio-oil, which was explained by the sintering of CaO particles and the formation of coke and sintering of Ni. These phenomena were more important in the bio-oil SESR, with the formation of more condensed coke, whose removal by gasification (Eq. (10)) is not complete at the reactivation conditions (750 ◦C). Considering the literature background and our above-mentioned previous results, the attention of this manuscript has been focused on the performance of the catalyst/sorbent system of (Ni/Al 2 O 3 catalyst derived from NiAl 2 O 4 spinel) +dolomite in the SESR of raw bio-oil operating in reaction-regeneration cycles. The process addresses the gaps in H 2 production from biomass using bio-oil without pre-treatment, producing high purity H 2 , no CO 2 emissions and lower energy requirements than other reforming processes. The catalyst was selected due to its complete regenerability in the SR of raw bio-oil under suitable conditions [56]. Likewise, dolomite, besides being a low-cost material, has a good performance as a CO 2 sorbent, with a positive synergy with Ni/Al 2 O 3 catalyst derived from NiAl 2 O 4 spinel in the SESR of bio-oil and as guard-bed, retaining a fraction of the deposited coke [57]. Furthermore, the valorization of raw bio-oil (with the attraction of high H 2 yield) and the low temperature required to favor the extent of CO 2 adsorption have the drawback of high coke deposition, leading to rapid catalyst deactivation, which is a challenge for catalyst and sorbent reactivation. Consequently, given the importance of deactivation, special attention has been paid to the study of its causes and mechanism, using several characterization techniques (Temperature Programmed Oxidation (TPO), adsorption-desorption of N 2 , Temperature Programmed Reduction (TPR), X-Ray Diffraction (XRD), Scanning Electron Microscopy (SEM) and X-Ray Photoelectron Spectroscopy (XPS)). The results in this manuscript are explained by considering the effect L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1528 of the reaction conditions on the extent of each reaction involved in the SESR process and on the content and nature of the coke deposited on the catalyst. These results show that SESR of raw bio-oil is an attractive route for the production of H 2 from biomass, avoiding the CO 2 emissions associated with other reforming technologies. Furthermore, the progress in knowledge of the process corresponds to key objectives for its viability, such as the use of raw bio-oil as feedstock, the proposal of a highly active catalyst with moderate deactivation by coke and regenerable, and the establishment of optimal conditions (reaction temperature, sorbent/catalyst mass ratio, space-time) for maximizing H 2 yield and purity and for stable operation in SESR of bio-oil in successive reaction-regeneration cycles. The use of a low-cost sorbent such as dolomite is another factor favoring the feasibility of the process. In addition, the results are based on experiments in a fluidized-bed reactor, which favors its use in scale-up where this type of reactor is likely to be required. 2. Experimental 2.1. Bio-oil The raw bio-oil was supplied by BTG Bioliquids BV (The Netherlands) and was synthetized by fast pyrolysis of pine sawdust in a plant with a capacity to operate continuously with 5 ton/h of biomass (bio-oil yield of 70 %) and with a conical rotatory reactor. The physicochemical properties of the bio-oil are as follows: water content, determined by Karl-Fischer titration (KF-Titrino Plus 870), 24 wt%; density, 1.201 g/ml; viscosity at 40 ◦C, 250 cP (Brookfield DV2T Ametek); pH, 2.5–3.5, and; empirical formula, obtained by CHO analysis, using a Leco CHN-932 analyzer (water-free basis), resulted in C 4⋅6 H 6⋅2 O 2.4 . The detailed composition of the raw bio-oil was determined using a Shimadzu QP2010S gas chromatography/mass spectrometer (GC/MS) equipped with a BPX-5 column (50 m ×0.22 mm ×0.25 μ m) and massselective detector. The main compounds are ketones, acids, phenols (guaiacol), esters, saccharides (levoglucosan), aldehydes, furans/furanones, alcohols, and ethers (Table 1). 2.2. Preparation of catalyst and sorbent The NiAl 2 O 4 spinel (33 wt% nominal Ni content) was synthesized by the co-precipitation method [58], mixing at 25 ◦C aqueous solutions of hexahydrated nickel nitrate (Ni(NO 3 ) 2 ⋅6H 2 O, Scharlau, purity of 98 %) and nona-hydrated alumina nitrate (Al(NO 3 ) 3 ⋅9H 2 O, Honeywell Fluka, purity of 98 %) and adding dropwise a 0.6 M solution of ammonium hydroxide (NH 4 OH 5 M, Honeywell Fluka) as a precipitating agent until a pH of 8 is reached. The precipitate was then recovered by filtration and washed with distilled water to remove the remaining ammonium ions, dried overnight at 110 ◦C, calcined (850 ◦C for 4 h, with a heating ramp of 10 ◦C/min), crushed and sieved to obtain particle sizes in the range of 150–250 μ m. The NiAl 2 O 4 spinel precursor was reduced (conditions indicated in Section 2.4) in order to obtain the Ni/Al 2 O 3 catalyst active for the reforming reactions. A schematic of the catalyst synthesis process is shown in Fig. S1. Calcined dolomite, composed of calcium and magnesium oxide (CaO and MgO), is obtained from dolomitic minerals consisting of calcium and magnesium carbonate (CaMg(CO 3 ) 2 ) with Fe 2 O 3 impurities (supplied by Calcinor S.A. (Cantabria, Spain)) through a calcination process (Eq. (12)). CaMg(CO3)2+heat ↔ CaO +MgO +2CO2(12) The preparation of the sorbent consists of a first grinding to adjust the particle size between 90 and 125 μ m. Then, the natural dolomite is dried at 110 ◦C for 24 h. Finally, it is calcined at 850 ◦C for 5 h with a heating rate of 10 ◦C/min, so that the natural sorbent is thermally decomposed into CaO (the active phase for CO 2 adsorption) and MgO [59,60]. Afterwards, it is sieved again to ensure a particle size between 90 and 125 μ m, since during the calcination step the particles decrease in size due to decrepitation caused by the loss of CO 2 from the carbonates. A schematic of the calcined dolomite preparation procedure is shown in Fig. S2. Note that the particle size of the dolomite is different from that of the catalyst to facilitate their separation by sieving after each reaction, so that the catalyst used can be characterized separately from the sorbent (and also from the inert solid used in the fluidized-bed reactor, as described in Section 2.4). 2.3. Characterization of the catalyst The physical properties of the fresh-reduced and used catalysts (BET surface area, pore volume and mean pore diameter), were characterized by adsorption-desorption of N 2 in a Micromeritics ASAP 2010. Temperature Programmed Reduction (TPR) analysis was carried out in a Micromeritics AutoChem II 2920 apparatus for determining the reducibility of the metal species. The TPR measurement was also used to estimate the real Ni content in the spinel and after catalyst regeneration using a thermobalance (TA Instruments SDT-2960) to determine the O atoms removed as H 2 O. The structural properties of fresh-reduced and used catalyst were analyzed by X-Ray Diffraction (XRD), measured on a Bruker D8 Advance diffractometer with CuK α 1 radiation, in order to calculate the sintering dynamics of the Ni crystals (using the Scherrer equation) and the crystalline state of the coke deposits. The amount and nature of coke deposited on used catalyst samples has been determined by Temperature Programmed Oxidation (TPO) in a TA-Instruments TGA-Q5000IR thermobalance, coupled in line with a mass spectrometer (Thermostar Balzers instrument) to monitor the CO 2 signal. The coke content has been quantified from the CO 2 spectroscopic signal, since the oxidation of Ni crystals during the combustion process could mask the thermogravimetric signal. The surface mass content of the different elements on the catalyst surface and the possible migration of Ni during the deactivation or regeneration stages has been determined by X-Ray Photoelectron Spectroscopy (XPS) analysis in a SPECS system equipped with a Phoibos 150 1D-DLD analyzer, Al K α monochromatic radiation (1486.6 eV), an X-ray exciting source and a hemispherical electron analyzer. The binding energy of the C1s carbon peak was set at 284.6 eV to correct the material charging, and the analysis was carried out using an electron take off angle of 90◦. The scanning electron microscopy (SEM) images of the used catalysts were obtained in a Hitachi S-4800 N field emission gun scanning electron microscope with an accelerating voltage of 5 kV and a secondary electron detector (SE - SEM). Table 1 Composition (wt%) of raw bio-oil obtained from fast pyrolysis of pine sawdust. Compound wt% Acids 19.5 Acetic acid 16.6 Ketones 21.4 Linear 17.1 Acetone 5.2 Acetol 9.4 Cyclic 4.3 Esters 11.3 Furans/Furanones 5.0 Alcohols 3.2 Aldehydes 6.8 Ethers 0.8 Saccharides 13.7 Levoglucosan 11.1 Phenols 18.4 Alkylphenol 1.4 Guaiacol 11.1 Catechol 0.9 Syringol 0.5 Others 4.5 L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1529 2.4. Reaction equipment and operating conditions The kinetic runs have been performed in an automated reaction system (MicroActivity-Reference, PID Eng & Tech) provided with two units (thermal step and catalytic step) in series (Fig. S3 in Supplementary Material). The first unit (thermal treatment) consists of a U-shaped steel tube (inner diameter =0.75 in) at 500 ◦C for controlled vaporization of bio-oil and repolimerization of some oxygenates (mainly phenolic compounds) which are retained as a carbonaceous solid called pyrolytic lignin (PL) [50]. In this way, the subsequent catalyst deactivation in the SESR and SR processes is attenuated, with a moderate decrease in the flow rate of oxygenates entering the catalytic reactor (unit 2). The volatile stream leaving the thermal treatment unit is converted (by catalytic SESR or SR) in a fluidized-bed reactor (stainless steel, with 22 mm of internal diameter and total length of 460 mm). In this reactor, the catalytic bed composed of NiAl 2 O 4 spinel (with a particle size of 150–250 μ m) and sorbent (calcined dolomite, with a particle size between 90 and 125 μ m) is mixed with inert solid (SiC, with a particle size of 75 μ m) in order to improve the fluid dynamics of the bed. The inert nature of SiC in the SR of bio-oil has been demonstrated in a previous work [53]. An injection pump (Harvard Apparatus 22) was used to feed the biooil (0.06 ml/min) and a 307 Gilson pump to co-feed the additional water required according to the desired steam to carbon (S/C) molar ratio. The reaction products were analyzed in a Micro GC (Varian CP-490) connected in-line to the reactor through an insulated line (130 ◦C) to avoid condensation of the products. The gas chromatograph is equipped with three analytical channels: molecular sieve MS5 for quantification of H 2 , O 2 , N 2 , CH 4 and CO; PPQ column for light hydrocarbons (C 2 –C 4 ), CO 2 and water; and Stabilwax for oxygenated compounds (C 2+ ) and water. Prior to each reaction, the Ni spinel precursor is reduced in situ under H 2 –N 2 (7 vol% H 2 ) at 850 ◦C for 4 h to obtain the active Ni/Al 2 O 3 catalyst with the Ni metallic phase well-dispersed on the alumina support [58]. The SESR runs of raw bio-oil were carried out in the catalytic reactor under the following conditions: 550–700 ◦C; bio-oil flow rate, 0.06 ml/min; space time, 0.15 and 0.30 g catalyst ⋅h/g oxygenates (corresponding to 0.5 g and 1 g of catalyst, respectively); sorbent/catalyst mass ratio, 10 and 20; steam/carbon (S/C) molar ratio at the reactor inlet, 3.4; time on stream (TOS), 50 and 300 min. Joint regeneration conditions of catalyst and dolomite were: 850 ◦C in air in an external oven, for 4 h. 2.5. Reaction indices The following reaction indices were used to quantify the results: The conversion of oxygenates in the treated bio-oil (that is, the volatile oxygenates exiting the Unit 1 used for thermal treatment) is expressed as the carbon units converted into gas: X=Fout,gas Fin (13) where F out,gas is the carbon-molar flow rate of the total carbon in gaseous product (CO 2 , CO, CH 4 and light hydrocarbons) at the reactor outlet, and F in is the carbon-molar flow rate of the oxygenates at the reactor inlet. Eq. (13) is suitable for quantifying the total conversion of oxygenates in the reforming reactor because the yield of solid carbon (coke) is low under the conditions studied (below 5 % as defined in the following Eq. (15)). H2yield is calculated as :YH2=FH2 Fo H2 (14) where F H2 is the H 2 molar flow rate in the product stream and Fo H2 is the stoichiometric molar flow rate, calculated as (2n +m/2 – k)/n F in , according to the global stoichiometry for bio-oil (C n H m O k ) steam reforming (including the WGS reaction) (Eq. (3)). Yield of the carbon products :Yi=Fi Fin (i=CO,CO2,CH4and HC)(15) where F i is the carbon-based molar flow rate of the i-product (CO, CO 2 , CH 4 and HC) in the effluent (out) stream of the second unit (catalytic reactor). 3. Results 3.1. Behavior of catalyst/sorbent under different conditions This section shows the results of the effect of different operating variables on the reaction indices in the SESR of bio-oil, both in the CO 2 capture period and after dolomite saturation, in order to analyze the activity and stability of the catalyst/sorbent system. The focus is on the effect of temperature, but the effect of sorbent/catalyst mass ratio and space time are also addressed, considering the relationship of the effect of these variables. The results shown in this section are mean values obtained from repeated experiments (with an average mean error below 5 %). Fig. 1 compares the evolution with time on stream (TOS) of the reaction indices (conversion and product yield) for different temperatures in the 550–700 ◦C range, with a sorbent/catalyst mass ratio of 10 and a S/C ratio of 3.4. A low space time (0.15 g catalyst ⋅h/g oxygenates ) was used to better appreciate the effect of temperature on catalyst stability in the post-saturation period. The results obtained in runs without sorbent (SR tests) and with a sorbent/catalyst mass ratio of 20 are shown in Figs. S4 and S5, respectively (Supplementary Material). As observed in Fig. 1 and Fig. S5, the SESR process can be divided into three regions: pre-breakthrough (with CO 2 capture, green zone in Fig. 1c), breakthrough (when dolomite saturation begins to be noticeable, blue zone in Fig. 1c) and post-breakthrough (after complete saturation of the dolomite). Under the conditions of Fig. 1 (with a dolomite/ catalyst mass ratio of 10), the duration of the pre-breakthrough period is about 30 min above 600 ◦C and is slightly longer at 550 ◦C (Fig. 1c). The duration of this period is doubled when twice the sorbent/catalyst ratio is used (Fig. S5). In the capture period, the higher H 2 yield and lower CO and CH 4 yields compared to the SR test (Fig. S4) are explained by the shift of the WGS reaction equilibrium by CO 2 capture (Eq. (2)) and, as a consequence, the shift of the SRM reaction equilibrium (Eq. (5)). The breakthrough period involves the depletion of the carbonation reaction efficiency, and therefore, a sudden increase in the yield of CO 2 (Fig. 1c and S5c) and a decrease in the yield of H 2 (Fig. 1b and S5b) is observed. In the post-breakthrough period, CO 2 removal is no longer effective and SR and WGS reactions occur together with other secondary reactions responsible for coke formation. Consequently, in this period, the reaction indices evolve with time on stream due to the deactivation of the catalyst and sorbent. Fig. 2 compares the product distribution results in the CO 2 capture period of SESR reactions (for sorbent/catalyst mass ratio values of 10 and 20) and SR reactions (sorbent/catalyst mass ratio of 0) at different temperatures. The values of SESR reactions are mean values during the CO 2 capture period, whereas those of SR (with a noticeable deactivation, Fig. S4) correspond to zero time on stream. Light hydrocarbons yield is not shown because it is insignificant for all the operating conditions studied. In the SR reactions (solid lines in Fig. 2), the H 2 yield increases with increasing temperature (from 58 % at 550 ◦C to 81 % at 700 ◦C) because the reforming of oxygenates and CH 4 is favored. Consequently, the yield of CH 4 is negligible at 700 ◦C in Fig. 2. The CO yield increases with increasing temperature as a consequence of favoring the reverseWGS reaction. In the SESR reactions (dashed and dotted lines in Fig. 2), temperature also affects the CO 2 adsorption equilibrium, which becomes unfavorable with increasing reaction temperature. In the range studied, the increase in temperature favors the formation of H 2 , whose yield increases from 74 % to 92 % between 550 and 600 ◦C for a sorbent/ L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1530 catalyst mass ratio of 10 (Fig. 2a). It is noteworthy that at 600 ◦C the reforming of CH 4 is almost complete with a yield of less than 1 %. Above this temperature, the yield of H 2 increases only slightly despite the increase in oxygenate conversion (Fig. S4a). This result is explained by the fact that CO 2 capture by dolomite is unfavorable above 600 ◦C [49]. Thus, at 650 and 700 ◦C, the sorbent capture capacity decreases slightly, Fig. 1. Effect of temperature on the evolution with time on stream of conversion (a) and yields of H 2 (b), CO 2 (c), CO (d), CH 4 (e) and C 2 –C 4 hydrocarbons (HCs) (f). Reaction conditions: space time, 0.15 g catalyst ⋅h/g oxygenates ; sorbent/catalyst mass ratio, 10; S/C, 3.4. Colored zones in graph c: CO 2 capture period (green shade) and breakthrough period (blue shade). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Fig. 2. Effect of temperature on product yield distribution (H 2 , CO 2 , CO and CH 4 ) (a) and H 2 purity (b) in the CO 2 capture period for different sorbent/catalyst mass ratio (0 (SR), 10, 20). Reaction conditions: space time, 0.15 g catalyst ⋅h/g oxygenates ; S/C, 3.4. L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1531 as observed in the pre-breakthrough period in Fig. 1c. Although dolomite promotes the formation of CH 4 by oxygenates cracking/decomposition reaction (Eq. (4)), the yield of CH 4 remains almost null in the CO 2 capture period above 600 ◦C due to the high activity of the Ni/Al 2 O 3 catalyst for its SR at these temperatures [61]. Nevertheless, the CO yield (which is less than 1 % at 600 ◦C) slightly increases to 6 % at 700 ◦C with a sorbent/catalyst mass ratio of 10 due to the promotion of the reverse-WGS reaction with increasing temperature, which leads to a decrease in H 2 purity at high temperatures (Fig. 2b). For a sorbent/- catalyst mass ratio of 20 (dotted lines in Fig. 2), the product yields in the CO 2 capture period are almost equal to those obtained with a sorbent/catalyst mass ratio of 10, except at 550 ◦C. At this low temperature, the H 2 yield for a sorbent/catalyst mass ratio of 20 is only 64 % (slightly higher than that obtained in the SR test), which can be attributed to the significant deactivation of the catalyst along the CO 2 capture period in the SESR run with high dolomite loading, as observed in Figs. S5a–c. This catalyst deactivation is due to the activity of dolomite for decomposition/cracking of oxygenates [59], which is favored by the low extent of SR reactions at this low temperature for this low space time. As mentioned above, the evolution of the product distribution with time on stream in the post-breakthrough period (after CO 2 capture) in Fig. 1 is a consequence of catalyst deactivation. Thus, the yields of H 2 and CO 2 decrease due to the attenuation of the SR and WGS reactions, which involves an increase in the yields of CO (Fig. 1d), CH 4 (Fig. 1e) and hydrocarbons (Fig. 1f). To explain the effect of temperature on these trends, it is necessary to take into account the different residual activity of the catalyst at the beginning of this period. Moreover, the relationship between the effects of temperature and oxygenate concentration on deactivation [62,63] and the role of dolomite on deactivation should be considered [50,53,57]. Thus, as mentioned above, at low temperature (550 ◦C) the catalyst has undergone significant deactivation during the CO 2 capture period because the space time was low to convert all the oxygenates, which explains the subsequent slow decrease in conversion (Fig. 1a) and yields of H 2 (Fig. 1b) and CO 2 (Fig. 1c), and the slow increase in CO yield (Fig. 1d). These trends indicate that the catalyst/sorbent system has reached a quasi-steady state at this temperature. Conversely, above 600 ◦C, the catalyst/sorbent system has a higher remaining activity after the CO 2 capture period and its deactivation in the post-breakthrough period is slightly more pronounced with increasing temperature, as observed by the faster decrease in H 2 yield with time on stream (Fig. 1b). The increase in the extent of decomposition/cracking reactions of oxygenates over dolomite (Eq. (4)) with increasing temperature contributes to the increase in the yield of carbonaceous products [59,60], which explains why the decay of oxygenates conversion with time (Fig. 1a) is less pronounced than the decrease in H 2 yield (Fig. 1b) at high temperature. Comparing the results of Fig. 1 (corresponding to SESR with a sorbent/catalyst mass ratio of 10), Fig. S4 (SR) and Fig. S5 (SESR with a sorbent/catalyst mass ratio of 20), it is observed that the effect of temperature on catalyst stability depends on the presence and content of dolomite with the catalyst. Thus, at low temperatures (550–600 ◦C), the presence and content of dolomite has a negative effect, leading to a faster catalyst deactivation. At 650 ◦C, the presence and content of dolomite prolongs the stability period of the catalyst, but after this period, the deactivation is faster than in the SR process, with a rapid decrease in the reaction indices. Consequently, the H 2 yields at 300 min on stream at this temperature are 33 %, 20 % and 17 % for SR (Fig. S4b), sorbent/catalyst mass ratios of 10 (Figs. 1b) and 20 (Fig. S5b), respectively. On the contrary, at 700 ◦C, the moderate presence of dolomite (sorbent/catalyst mass ratio of 10) has a positive effect, with the reaction indices decreasing more slowly with time on stream. This result is explained by the activity of dolomite for SR reactions at this temperature, and also because dolomite deactivates more slowly than the catalyst [60]. However, for a sorbent/catalyst mass ratio of 20, there is a faster catalyst deactivation. In fact, the use of a high dolomite/catalyst ratio of 20 (Fig. S5) results in faster catalyst deactivation at all temperatures. In addition to the variables studied above (reaction temperature and sorbent/catalyst mass ratio), the effect of space time has been studied, as the amount of available active catalyst sites influences the extent of reforming and WGS reactions, and consequently the distribution of products and catalyst deactivation. Fig. 3 shows the evolution of the reaction indices with time on stream for a high space time of 0.3 g catalyst ⋅h/g oxygenates (twice that of Fig. 1). The results correspond to 600 ◦C and a sorbent/catalyst mass ratio of 10 (the conditions leading to the highest H 2 purity according to Fig. 2b). The duration of the CO 2 capture period in Fig. 3 (about 50 min) is longer than that obtained in Fig. 1 at the same temperature (about 30 min, red dots) due to the higher content of dolomite in the bed (10 g, compared to 5 g in Fig. 1). It should be noted that this dolomite content is the same as that used in the runs with a sorbent/catalyst mass ratio of 20 (Fig. S5), but now with twice the amount of catalyst. Comparing the product yields in the CO 2 capture period in Fig. 3 and the red curves in Fig. 1 (those corresponding to 600 ◦C with half the space time as in Fig. 3), a significant improvement in H 2 yield is observed with increasing space time, with a value of 95 % in Fig. 3 and almost the same H 2 purity (around 99 %) in both runs. The combined effect of increasing the CO 2 capture capacity by increasing the amount of dolomite and increasing the extent of reforming and WGS reactions by increasing the amount of catalyst contributes to this result. Furthermore, after the CO 2 capture period, it is observed that the stability of the catalyst has increased with increasing space time value. As a result, the H 2 yield remains high (about 72 %) after 5 h of operation, compared to about 25 wt% under the same conditions but with half the space time (red curve in Fig. 1b). This higher stability is consistent with previous studies on bio-oil SR, that show an improvement in catalyst stability with increasing space time values due to the lower concentration of oxygenates in the reaction medium, as a result of the higher extent of oxygenate reforming reactions [63,64]. 3.2. Characterization of catalyst and sorbent used under different conditions The results in section 3.1 show that the stability of the catalyst depends on the reaction conditions. In order to identify the causes responsible for the deactivation and its dynamics during the CO 2 capture periods in Fig. 1, samples of the catalyst used after 50 min on stream (end of the CO 2 capture period) and after 300 min on stream (end of the runs in Fig. 1, comprising the post-capture period) were analyzed using different techniques (TPO, TPR, XRD, N 2 adsorption-desorption and SEM). After each reaction, the three components of the catalytic bed (catalyst/sorbent/inert solid) were separated by sieving, in order to Fig. 3. Evolution with time on stream of carbon conversion and product yields. Reaction condition: 600 ◦C; space time, 0.3 g catalyst ⋅h/g oxygenates ; dolomite/ catalyst mass ratio, 10; S/C, 3.4. Colored zones: CO 2 capture period (green shade) and breakthrough period (blue shade). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1532 characterize the used catalyst separated from the sorbent or inert solids. 3.2.1. Analysis of coke deposits The Temperature Programmed Oxidation (TPO) of the used catalyst provides information on the total content of coke deposited, and the nature and/or location of the coke on the catalyst [63,65]. Fig. 4 shows the TPO profiles and the corresponding coke content (wt%) for the catalyst used at different temperatures. It should be noted that the scale of the two graphs is different. Graph (a) corresponds to the runs performed up to the breakthrough regime with CO 2 capture (50 min time on stream), whereas graph (b) corresponds to the catalyst used in the 300 min duration tests (corresponding to the reactions in Fig. 1). For comparison, the TPO profiles for the catalysts used in the SR tests (without sorbent) are plotted in Fig. S6. In line with previous results in the literature on coke deposition on supported Ni-based catalysts used in the SR of oxygenates and bio-oil [58,63,65], the TPO profiles in Fig. 4 show two different combustion domains, corresponding to coke burning below 500 ◦C or above 500 ◦C. These fractions can be attributed to the different nature of the coke: amorphous carbon (burning at low temperature) and carbon filaments or graphitic structures (burning at high temperature) [66,67]. It has also been claimed that the location of the combustion peaks depends on the location of the coke, with coke deposited on the metal surface burning at a lower temperature than coke deposited on the support [61,65]. The results in Fig. 4 reveal a remarkable effect of the reforming temperature on the coke content and nature in the two groups of runs. For the catalysts used only in the CO 2 capture period (Fig. 4a), the coke burning below 500 ◦C (presumably amorphous coke deposited on the Ni sites) is the only coke fraction at a reforming temperature of 550 ◦C. As the reaction temperature increases, the total coke content decreases and so does this amorphous coke fraction, whereas a fraction burning above 500 ◦C appears at 600 ◦C and becomes the predominant fraction above 650 ◦C. These trends indicate that the deposition of coke is attenuated, but the deposited coke is more refractory as the reforming temperature increases. Comparing the TPO profiles in Fig. 4a and b, it can be seen that at a reforming temperature of 550 ◦C the nature of the deposited coke is similar for both values of time on stream, although the coke content increases for the higher time on stream (Fig. 4b). For the reforming temperatures of 650 and 700 ◦C, the TPO profiles of the coke deposited during the CO 2 capture period (Fig. 4a) and at 300 min reaction time (Fig. 4b) are also similar, especially for the reactions at 700 ◦C. But interestingly, the TPO profiles of the catalyst samples used at 600 ◦C in the CO 2 capture period and at the end of the reaction (after saturation of the dolomite) differ significantly, so that the structured coke burning at high temperature is the major coke fraction for the latter sample (used for longer reaction time after dolomite saturation). Moreover, the TPO profiles of the catalyst used at 600, 650 and 700 ◦C for 300 min in the presence of saturated dolomite (Fig. 4b) differ significantly from those in Fig. S6, corresponding to the SR (without dolomite) of the bio-oil. All these results show a remarkable role of dolomite in coke deposition, with the role of dolomite at a reforming temperature of 600 ◦C being different from that at higher reforming temperatures. The results on the nature of the coke (amorphous or filamentous) obtained by TPO analysis have been corroborated by SEM images of the catalyst used at selected operating conditions, shown in Fig. S7. Thus, in the SEM images of the catalyst particles used at 600 ◦C and with a catalyst/mass ratio of 10 (Figs. S7c and d), a high presence of carbon filaments is observed. Conversely, there are no carbon filaments in the catalyst used at 550 ◦C with a sorbent/catalyst mass ratio of 10 (Figs. S7a and b), and the presence of carbon filaments is significantly lower than at 600 ◦C in the catalyst particles used at 700 ◦C with a sorbent/catalyst mass ratio of 20 (Figs. S7e and f). Concerning the effect of space time on coke deposition, the comparison of the TPO profiles for two space time values at 600 ◦C (continuous and dashed red lines in Fig. 4b) confirms the well-known fact that coke deposition in bio-oil reforming is significantly attenuated with decreasing oxygenate concentration (precursors of amorphous coke) [61,64]. In this case, both amorphous and filamentous coke formation are notably attenuated by doubling the space time. 3.2.2. Deterioration of the used catalyst properties To assess the sintering of Ni crystals and the crystalline state of the Fig. 4. TPO profiles and coke content (wt%) for the catalyst used at different temperatures in the CO 2 capture period (50 min on stream) (a), and after 300 min on stream (b). Reaction conditions: space time, 0.15 g catalyst ⋅h/g oxygenates; dolomite/catalyst mass ratio, 10; S/C, 3.4. *space time, 0.30 g catalyst ⋅h/g oxygenates for dashed red line in graph b. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Fig. 5. XRD patterns of fresh-reduced and used catalyst samples in the SESR runs at different temperatures. Reaction conditions: space time, 0.15 g catalyst ⋅h/ g oxygenates (*0.30 g catalyst ⋅h/g oxygenates ); sorbent/catalyst mass ratio, 10; S/C, 3.4. L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1533 coke deposits, Fig. 5 shows the XRD patterns of the fresh-reduced and used catalyst samples at different temperatures under the conditions of Fig. 1 (sorbent/catalyst mass ratio of 10). XRD results corresponding to a sorbent/catalyst mass ratio of 20 are shown in Fig. S8. Peaks corresponding to Ni 0 (diffraction angle at 44.5◦in (111) plane, 51.8◦in (200) plane and 75.5◦in (110) plane) and Al 2 O 3 crystals (37.3◦, 45.6◦and 66.8◦) are identified in the fresh-reduced catalyst. The same diffraction peaks are observed in the catalyst used in the SESR of bio-oil. Similarly, the presence of NiO is not detected, confirming the high reducing capacity of the reaction medium to maintain the active metal in a reduced state. The average size of the Ni crystals of the fresh-reduced and used catalysts (Table 2) was calculated from the diffraction peak at 2θ =51.8◦ using the Scherrer equation. The estimated value is 15.0 nm for the fresh-reduced catalyst and increases slightly in the range of 15–20 nm for the used catalyst samples. Based on these results, there is no evidence for a significant sintering phenomenon of Ni crystals. The XRD patterns of the used catalyst samples also provide information about the coke deposits. The presence of a broad peak at a diffraction angle 2θ =26◦is observed in the catalyst used at 600 ◦C (more pronounced at the lowest space time). This peak corresponds to high crystallinity coke (graphitic carbon), usually identified in catalysts used in the SR of hydrocarbons [68] and pure oxygenate compounds, such as ethanol and phenol [69], acetone [70], acetic acid [71]. This result is coherent with the TPO profiles in Fig. 4b and the SEM images in Figs. S7c and d, and confirms the structured nature of the main combustion peak in the coke deposited at 600 ◦C, whose amount decreases significantly with increasing space time. The absence of the diffraction peak at 26◦in the coke deposited at 550 ◦C corroborates its amorphous nature. For the catalysts used in the SESR above 650 ◦C, although some structured coke is deposited according to the TPO profiles (Fig. 4b), its small amount does not allow it to be detected in the corresponding XRD patterns in Fig. 5. In order to evaluate the possible deterioration of the porous structure of the catalyst and its contribution to the deactivation, N 2 adsorptiondesorption isotherms have been obtained from fresh-reduced and used samples (Fig. S9), from which the physical properties (BET surface area, mean pore diameter and pore volume) have been determined (Table 2). All samples in Fig. S9 show a type IV isotherm, characteristic of mesoporous materials, associated with capillary (pore) condensation taking place in mesopores and the occurrence of multilayer adsorption. The isotherms for fresh-reduced catalyst and samples used at the lower temperature (Fig. S9a) present a hysteresis of the type H2, which is attributed to a difference in mechanism between condensation and evaporation processes occurring in pores with narrow neck and wide bodies. Materials with H2 hysteresis are often disordered and the distribution of pore size and shape is not well defined. Conversely, a H3type hysteresis cycle is observed in the isotherms of the catalysts used above 600 ◦C (Figs. S9c,e,g), which shows no limiting adsorption at high P/P 0 , and is associated to aggregated plate-like particles that give rise to slit-shaped pores [72]. Overall, there is an evolution of mesoporosity with a slight progressive decrease from the highest temperature to 550 ◦C. For most of the used samples, the total volume adsorbed at high pressures (P/P 0 ≈1) is lower than that of the fresh-reduced sample, demonstrating the partial blockage of the mesopores, which is more pronounced at the lowest reforming temperature. For all reforming temperatures studied, the BET surface area of the used catalyst decreases with increasing dolomite presence, evidencing a partial blockage of the mesoporous structure of the catalysts, which is promoted at higher sorbent loadings, probably due to the high deposition of a low porosity coke. On the other hand, the increase in BET surface area observed for the used samples at 600 ◦C (more pronounced at high space time) compared to the fresh-reduced catalyst, suggests that there is an additional porosity created by the filamentous coke deposited on these samples. Overall, the average pore diameter and pore volume decrease compared to the fresh-reduced sample (more noticeable at low temperatures, 550 and 600 ◦C), probably due to the blockage of part of the porous surface. To sum up, the different BET surface area values (higher or lower than those of the fresh catalyst, depending on the reaction conditions) indicate the formation of both: i) filamentous coke (at 600 ◦C), which is probably stacked on the surface of the catalyst, causing an increase in the BET surface area, and ii) coke clogging the porous structure, causing a decrease in the BET surface area. 3.3. Cyclic operation with joint regeneration of the catalyst and dolomite The efficiency of the joint reactivation of catalyst and sorbent has been studied running 7 reaction-regeneration cycles. The conditions in the reaction step (established by the results in section 3.1) are 600 ◦C, space time of 0.30 g catalyst ⋅h/g oxygenates (corresponding to 1 g of catalyst), sorbent/catalyst mass ratio of 10, S/C molar ratio of 3.4, and time on stream of 120 min. These conditions allow to operate with a sufficiently high CO 2 capture time (about 60 min), with the corresponding sorbent/ catalyst mass ratio of 10. The regeneration step consisted of coke combustion and CO 2 removal in an external oven at 850 ◦C for 4 h in air atmosphere (sufficient time for complete removal of coke deposits and CO 2 ). After each regeneration step, the catalyst +sorbent were subjected to a reduction step (at 850 ◦C for 4 h in H 2 –N 2 (7 vol% H 2 )) to obtain the active Ni 0 particles. Fig. 6 shows the evolution with time of the reaction indices in the 1st reaction step (fresh material) (graph a) and in the 7th reaction step (graph b), and Fig. 7 shows the evolution of the H 2 yield in the CO 2 capture period in the successive reaction steps. When comparing the 1st and 7th reaction steps, a rather similar performance of the catalysts/sorbent bed is observed, indicating a good overall recovery in the regeneration step of both solids, the catalyst activity and the CO 2 sorption capacity of the dolomite. Nevertheless, the H 2 yield in the 7th reaction step is slightly lower than that obtained with the fresh catalyst in the 1st reaction, and a small CH 4 yield is observed in the CO 2 capture period. As observed in Fig. 7, this slight decrease in H 2 yield in the CO 2 capture period occurs in the first two reaction-regeneration cycles, but the catalyst/sorbent bed reaches a stable performance from the 3rd third reaction step onwards. The possible causes of the decrease in H 2 yield in the CO 2 capture period observed in the first two cycles are the incomplete recovery of the CO 2 sorption capacity of the sorbent and/or of the activity for the reforming and WGS reactions of the catalyst. It is interesting to note in Fig. 6 that the duration of the capture period in the 7th reaction step is even slightly longer than in the 1st step, suggesting that the dolomite has fully recovered its CO 2 capture capacity. This result can be observed more clearly in Fig. 8, which shows the evolution over time of the CO 2 yield in different reaction steps (1st, 2nd, 3rd and 7th). As observed, the duration of the CO 2 capture period is shorter in the first cycles and remains constant and longer after the third cycle. Table 2 Physico-chemical properties (average Ni 0 crystal size (d Ni ), BET surface area (S BET ), pore volume (V pore ) and mean pore diameter (d pore ) of fresh-reduced catalyst and used in the SESR of raw bio-oil at different temperatures and sorbent/catalyst mass ratio. Reaction conditions: space time, 0.15 g catalyst ⋅h/ g oxygenates ; S/C, 3.4. Catalyst Temperature (◦C) d Ni (nm) S BET (m 2 /g) V pore (cm 3 /g) d pore (nm) Fresh-reduced – 15 65.1 0.24 13.1 Used with sorbent/ catalyst mass ratio of 10 550 15 62.3 0.11 7.1 600 16 69.9 0.18 13.4 600 a 20 79.6 0.25 13.4 650 16 60.6 0.20 14.4 700 16 65.0 0.26 15.4 Used with sorbent/ catalyst mass ratio of 20 550 16 58.9 0.13 8.3 600 16 52.1 0.13 10.8 650 16 55.5 0.16 12.0 700 17 57.9 0.21 13.1 a Space time, 0.30 g catalyst ⋅h/g oxygenates . L. Landa et al. International Journal of Hydrogen Energy 58 (2024) 1526–1540 1534 Interestingly, the mesoporosity of the dolomite decreases along the reaction-regeneration cycles, as shown by the results in Table S1. The BET surface area and pore volume of the saturated dolomite after the 7th reaction decrease significantly compared to the fresh-calcined dolomite sample, while the mean pore diameter increases, indicating complete sorbent saturation due to CO 2 capture with a selective blockage of the narrower pores. After the regeneration of the dolomite by decarbonation, the mesoporosity is only partially restored, so that the surface area and the mean pore volume are about half of those of the fresh-calcined sample, which shows the partial blockage of the sorbent mesopores after the regeneration period, not being able to fully recover the initial structure of the fresh-calcined dolomite. Nevertheless, the dolomite retains its CO 2 capture capacity. Consequently, the decrease in H 2 yield during the CO 2 capture period in the first cycles (Fig. 7) should be attributed to the irreversible deactivation of the catalyst in the first two reaction-regeneration cycles and not to the deterioration of the CO 2 capture capacity of the dolomite, as studied in section 3.4. 3.4. Characterization of catalyst used in the cyclic operation The catalyst used in some reaction steps, and also the catalyst regenerated after the 7th reaction, have been characterized by several techniques (TPO, TPR, XRD, N 2 adsorption-desorption and XPS) (section 2.2) in order to analyze the possible changes in the physico-chemical properties of the catalyst along the cyclic operation and to correlate them with the kinetic performance observed in section 3.3. 3.4.1. Coke deposition Fig. 9 depicts the TPO profiles of the catalyst used after the 1st and 7th reaction steps. For both samples, there is a major combustion peak burning above 500 ◦C corresponding to filamentous and structured coke, with a low amount of amorphous coke burning below 500 ◦C. The coke content decreases in the successive reaction steps (from 17.2 wt% in the first reaction to 12.8 wt% in the seventh reaction), especially the filamentous coke, whereas the amorphous carbon remains fairly constant. The difference in coke deposition in Fig. 9 is consistent with the slightly lower catalyst activity observed after the first two reaction-regeneration Fig. 6. Evolution with time on stream of carbon conversion and products yields in the 1st (a) and 7th (b) reaction steps (with intermediate regenerations). Reaction conditions: 600 ◦C; space time, 0.3 g catalyst ⋅h/g oxygenates ; sorbent/catalyst mass ratio, 10; S/C, 3.4. Regeneration conditions: calcination with air in external oven at 850 ◦C for 4 h. Colored zones: CO 2 capture period (green shade) and breakthrough period (blue shade). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Fig. 7. Evolution of H 2 yield in the CO 2 capture period in successive reaction steps (with intermediate regenerations). Reaction and regeneration conditions of Fig. 6. Fig. 8. CO 2 yield evolution with time on stream in successive reaction steps (with intermediate regeneration). Reaction and regeneration conditions of Fig. 6. Fig. 9. Comparison of the TPO profiles and coke content (wt%) deposited on samples used in the 1st and 7th reaction steps. Reaction and regeneration conditions of Fig. 6. L. Landa et al.