Production of a syngas and CaO by desorption-enhanced reverse water–gas shift of CaCO3 with H2

Chemical Engineering Journal 493 (2024) 152191 Available online 16 May 2024 1385-8947/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Production of a syngas and CaO by desorption-enhanced reverse water–gas shift of CaCO 3 with H 2 J.C. Abanades a , G. Grasa b , * a CO 2 capture group, Instituto de Ciencia y Tecnología del Carbono (Spanish Research Council, INCAR-CSIC) Francisco Pintado Fe, 26, 33011 Oviedo, Spain b Environmental Research group, Instituto de Carboquímica (Spanish National Research Council, ICB-CSIC), Miguel Luesma Cast´ an 4, 50018 Zaragoza, Spain ARTICLE INFO Keywords: Syngas production Reverse water–gas shift Calcination CO 2 capture Packed bed reactor ABSTRACT A syngas production method is investigated that combines in a single reactor the enhanced decomposition of CaCO 3 with H 2 and the reverse water–gas shift (RWGS) of part of the CO 2 evolved during calcination. The method exploits Le Chatelier’s principle, to overcome RWGS equilibrium limitations by conducting such reactions with an excess of CaCO 3 and at sufficiently high temperatures to maintain the partial pressure of CO 2 close to the calcination equilibrium. The decomposition and RWGS reactions result in a ‘desorption-enhanced reverse water–gas shift’ (DERWGS) equilibrium of CaCO 3 on H 2 , observed in experiments performed in a packedbed reactor operated between 1023 and 1123 K and 1 and 5 atm when feeding H 2 to a mixture of CaCO 3 , with or without a RWGS catalyst. Product gases containing over 25 vol% CO, with an H 2 /CO molar ratio of 2 and below, were obtained. In experiments without the use of an RWGS catalyst, the DERWGS equilibrium was also approached thanks to the catalytic activity of CaO for RWGS. The syngas analogue obtained from these reactions opens the door to new processes for synthetic fuel production from CaCO 3 and renewable H 2 . 1. Introduction Sustainable carbon-containing fuels from renewable energy and renewable carbon have a key role to play in decarbonising the aviation and freight transport sectors, among others, although improvements to production processes and cost reductions are required [1]. Most existing process routes for the manufacture of synthetic gas and liquid hydrocarbons start from a syngas containing various proportions of CO +CO 2 +H 2 , accompanied by a diverse range of contaminants (i.e. H 2 S, HCl, hydrocarbons and alkalis), which requires deep purification before entering the catalytic synthesis process [2–6]. An alternative route for syngas production that has been receiving increasing attention in research is the hydrogenation of metal carbonates, also known as “direct reduction” or “reductive calcination”, as reviewed in [7] and explored in other recent works [8,9]. Among the different carbonate candidates this work is focused on CaCO 3 : CaCO3+H2→CaO +CO +H2OΔHr298K=220.3kJ/mol (1) The use of the expression “direct reduction” or “reductive calcination” in [8,9] refers to the possibility of obtaining in a single reactor the products of two separate reactions: as for example the CaCO 3 decomposition and the Reverse Water Gas Shift [8]. The use of these expressions does not necessarily imply a reaction mechanism with a single reaction step of direct reduction of CaCO 3 . Indeed, at the high temperatures of interest in this work (T >750 ◦C) the kinetics of calcination are known to be sufficiently fast [10] for reaction (1) to be the series of two reactions: CaCO3←→CaO +CO2ΔHr298K=178.8kJ/mol (2) CO2+H2←→CO +H2OΔHr298K=41.5kJ/mol (3) In such case, when reactions (2) and (3) occur in sufficient proximity as to allow them to progress according to the overall reaction (1), and the CaCO 3 acts as a direct source of pure CO 2 in the gas reaction medium, the reductive calcination of CaCO 3 provides opportunities for process intensification and efficiency gains [7,8]. The use of CaCO 3 as a source of CO 2 and CaO can contribute to GHG mitigation when the CaO is used as a regenerable CO 2 sorbent in a variety of CO 2 capture systems [11–14]. Also, in large scale emerging processes for direct CO 2 capture from air [15–18], that involve CO 2 extraction from a CaCO 3 calcination step at some point in their processes. Finally, CaO must be produced to meet the global demand of * Corresponding author at: Environmental Research group, Instituto de Carboquímica (Spanish National Research Council, ICB-CSIC), Miguel Luesma Cast´ an 4, 50018 Zaragoza, Spain. Tel.: +34 976 733 977. E-mail address: [email protected] (G. Grasa). Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej https://doi.org/10.1016/j.cej.2024.152191 Received 5 September 2023; Received in revised form 5 April 2024; Accepted 11 May 2024 Chemical Engineering Journal 493 (2024) 152191 2 cement and lime. Fig. 1 shows a simplified scheme of a CaCO 3 regeneration stage with H 2 integrated in the frame of a general CaL CO 2 capture process (from point sources or the atmosphere). Reaction (1) is highly endothermic, mainly due to the enthalpy needed for the CaCO 3 calcination reaction (178.8 kJ/mol at 298 K). This means that hydrocarbon synthesis routes starting from CaCO 3 are energetically unfavoured with respect to the equivalent routes starting from CO 2 (g). However, it could be argued that such a form of overall hydrogenation of CaCO 3 can still compete in energy terms with the equivalent route starting from CO 2 when the energy savings in the downstream uses of the CaO resulting from Eq. (1) are taken into consideration. The hydrogenation of CaCO 3 was first reported in 1968 [20], when the formation of CO and other hydrocarbons was identified within a range of temperature and pressures during experiments with an overwhelming presence of H 2 (>99 vol%). Subsequently, in a seminal paper in the field of carbonate reduction, Reller et al. [21] observed the complete calcination of small samples of CaCO 3 in an H 2 atmosphere, detected the presence CO as a gaseous reaction product, and reported catalytic effects and substantial reductions (>150 K) in CaCO 3 calcination temperatures, with these effects enhanced by the admixture at close to atomic level of transition metals with the CaCO 3 . In a recent review of these phenomena by R. Han et al. [13], a wide range of dual functional materials combining CaO as CO 2 sorbent and a transition metal as hydrogenation catalyst have been described. Examples include the intense work carried out on the design of dual functional materials (DFM) by Sun et al. [19,22], who evaluated the performance of a number of metals acting as catalysts for the reverse water–gas shift (RWGS) reaction, and who more recently [23] confirmed in independent experiments the catalytic activity of CaO for the RWGS that had been reported by Giammaria and Lefferts [24]. Such DFM materials have shown to have promising properties that enable them to operate in a wide range of process concepts, as reviewed by Lux et al. [7]. These processes usually involve means for circulating solids between a carbonator (where CO 2 contained in a gas reacts with CaO to form CaCO 3 ) and a hydrogenator (where the CaCO 3 will decompose in presence of H 2 to form CaO and a hydrocarbon-containing gas). Alternatively, it has been proposed that these processes operate according to principles of pressure and/or temperature swing adsorption, by switching between carbonation and hydrogenation conditions in the same vessel [8,19,22,23]. We have recently proposed a variant of this approach in order to accommodate the heat supply steps needed by including a redox loop, that provides the energy to drive the endothermic reactions (2) and (3) that undergoes the CaCO 3 contained in a packed bed of solids [25]. The aim of this work is to report on a phenomenon that has been found to occur in these systems when operating at conditions that allow both the calcination and RWGS equilibria to be fulfilled. To our knowledge, this is the first time that this phenomenon has been reported with experimental evidence, and if scaled, it could lead to new process routes for hydrocarbon production from CaCO 3 (formed in calcium looping systems shown in Fig. 1) and renewable H 2 . 2. Desorption-enhanced reverse water–gas shift of CaCO 3 The first reported experiments involving CaCO 3 decomposition in presence of H 2 [20] produced a variety of hydrocarbons and carbon deposits, depending on pressure and temperature conditions. More recently, Sun et al. [23] reported experiments of CaCO 3 calcination with H 2 at atmospheric pressure and temperatures of 873–973 K, observing that while CO 2 concentrations in the product gas where similar to those obtained in experiments using N 2 feed, the CO concentrations where over 3 times higher than those of CO 2 . They confirmed by dedicated experiments that CaO catalyses the RWGS reaction of the CO 2 evolved from CaCO 3 . However, despite the high selectivity attributed to CO, the actual concentrations of CO in the product gas were very low (<3% vol CO), requiring challenging gas separation steps downstream of the hydrogenation step before such a gas can be processed as a syngas. Shi et al. [26] analysed the effect of pressure on product gas distribution at a similar range of temperatures and H 2 pressures up to 60 atm. They concluded that pressures over 40 atm and temperatures below 975 K favoured the production of CH 4 with respect to CO when CaCO 3 was calcined with H 2 . Our interest in this work is narrowed down to a high temperature window of 1023–1223 K (for a pressure range 1–10 atm), where the intrinsic CaCO 3 calcination reaction is sufficiently fast [10,27] as to rapidly approach the equilibrium of CO 2 in CaO, with a partial pressure of CO 2 given by [28]: PCO2eq =e(16.3−19130 T(K))=Kcalc (4) Fig. 2 illustrates two different calcination equilibria in an original batch of CaCO 3 (red dots) in a control volume where the total number of carbon atoms in the system (as CaCO 3 , CO or CO 2 ) is maintained constant. The discussion that follows is only valid if there is an excess of CaCO 3 (i.e. with both CaO and CaCO 3 present in the solid phase, as shown) within the control volume. In Fig. 2 a), the partial pressure, P CO2_eq , and the molar conversion of CaCO 3 to CaO, X cal , are reached at equilibrium. If an inert gas is added to the control volume, then the total pressure, P T =P CO2_eq +P inert , will increase without altering either P CO2_eq or X cal . At a constant total pressure, the introduction of H 2 to the control volume instead of the inert gas first requires an increase in volume, as shown in Fig. 2 b), because there is a redistribution of the carbon atoms in the gas and solid phase caused by the RWGS (Eq. (3). Le Chatelier’s principle, represented in Fig. 2 b), dictates that a certain Nomenclature a, b Apparent reaction orders in Equation (12) DERWGS Desorption-enhanced reverse water–gas shift E a Apparent activation energy in Equation (12), kJ/mol k o Apparent pre-exponential factor in kinetic expression of Equation (12), s -1 bar –(a+b) K calc Calcination equilibrium constant, atm K H2O Absorption equilibrium constant in Equation (12) K WGS Water-gas shift equilibrium constant P CO , CO2 , H2O , H2 Gas partial pressures (subscript _eq at equilibrium), atm P T Total pressure (atm) R CO CO formation rate in Equation (12), s -1 T DERWGS Average temperature at the DERWGS reaction front or region, K X cal CaCO 3 molar conversion to CaO X N CaCO 3 content in a CaO/CaCO 3 material X RWGS CO 2 conversion via RWGS according to Equation (11) Δz Element volume height (m) in Figures 4 and 9 Fig. 1. Simplified CaL scheme, including sorbent regeneration through CaCO 3 calcination in H 2 , inspired by references in literature [7,8,19]. J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 3 fraction of CaCO 3 will decompose to CaO to compensate for the CO 2 that is removed from the gas phase owing to the RWGS reaction. In consequence, CaCO 3 conversion to CaO will increase, and the carbon content in the product gas (P CO_eq +P CO2_eq ) will be higher when compared with the reference case of Fig. 2 a) (i.e. P CO2_eq alone). At constant pressure, the control volume will increase as a result (Fig. 2 b)). In other words, the CaCO 3 calcination equilibrium and the RWGS equilibrium can be simultaneously fulfilled (as long as there is an excess of unconverted CaCO 3 in the control volume (Fig. 2 b)). Owing to its similarity to wellknown sorption-enhanced reactions [29–33] taking place in the reverse direction of these reactions, we have named this equilibrium ‘desorption-enhanced reverse water–gas shift’ (DERWGS). To allow a quantitative analysis of the DERWGS equilibrium, some further assumptions are adopted. When introducing H 2 into the control volume, several CO 2 reduction reactions can take place that lead to the formation of atomic carbon, CO and hydrocarbons [7]. However, we assume that at the temperatures of interest in this work (>1023 K), only CO and H 2 O (g) are produced, in agreement with the water–gas shift equilibrium: KWGS =PCO2eq*PH2eq PCO eq*PH2O eq (5) To facilitate the equilibrium calculations below, an approximate equation is used to estimate K WGS [34]: ln(KWGS) = 5693.5 T+1.077ln(T) + 5.44*10−4T−1.125*10−7T2 −49170 T2−13.148 (6) Particular solutions of the DERWGS equilibrium can be obtained when the control volume is occupied by the appropriate initial quantity of pure H 2 and CaCO 3 to yield a final gas product at a certain target total pressure. Again, this means that there should be sufficient CaCO 3 in the initial control volume to supply all the necessary carbon (P CO_eq and P CO2_eq ) in order to fulfil both the calcination and RWGS equilibria in the gas phase. In these conditions: PT =PH2eq +PH2O eq +PCO eq +PCO2eq (7) PH2O eq =PCO eq (8) PH2eq =PT −2PCO eq −PCO2eq (9) Which allows the calculation of equilibrium partial pressures by solving the quadratic equation resulting in the mass balance above, here solved for P CO_eq : PCO eq =⎛ ⎜ ⎜ ⎝ −2+ 4+4KWGS Kcalc (PT −Kcalc) √2KWGS Kcalc ⎞ ⎟ ⎟ ⎠ (10) As can be seen in Fig. 3 a), at atmospheric pressure, the DERWGS equilibrium (solid lines) allows the partial pressure of CO to be higher than the partial pressure of CO 2 (for temperatures below T =1080 K in Fig. 3 a)), reaching a maximum of about 0.25 in the partial pressure of CO at 1088 K. To illustrate such enhancement in RWGS due to CO 2 desorption Fig. 3 a) also plots in dotted lines the RWGS equilibrium composition that would be obtained in a non-enhanced system where pCO 2 at reactor inlet is equal to that marked by the CaCO 3 calcination equilibrium with temperature. As it can be seen, CO yields are higher due to the DERWGS equilibrium until a temperature of 1103 K is reached, after which a maximum of 0.26 in CO partial pressure is predicted by the RWGS equilibrium. From such maxima, the CO yields decline towards zero as the steep equilibrium curve of CO 2 on CaO displaces all other gases in the control volume and approaches the total pressure (1 atm in Fig. 3 a). Fig. 3 b) illustrates the P CO _eq curves when the final total pressure at equilibrium reaches different values (note that the curve of P CO_eq at PT =1 atm is identical to that presented in Fig. 3 a), but the T scale has changed). Note that the ratio of experimental yields of CO and CO 2 reported by Sun et al. [23] at 973 K in their hydrogenation experiments of CaCO 3 was about 3.1, consistent with the value of 3.6 estimated from equations 4–10. As it will be experimentally demonstrated in the following paragraphs, the DERWGS equilibrium can be exploited in practice at higher temperatures, attempting the hydrogenation of a packed bed of CaCO 3 - containing particles (or moving beds moving relatively slowly with regard to gas velocities) when mixed with a suitable RWGS catalyst. These particles can be limestone (i.e. the hydrogenation reactor will resemble a kiln for the production of CaO from the hydrogenation of the CaCO 3 contained in limestone, with means to mechanically separate the catalyst and CaO particles at the outlet) or carbonated CaO particles (e.g. carbonated particles exiting a previous carbonation stage [19,23] in calcium looping processes, in particular when using adiabatic packed beds [32,33,35–39], which generate carbonated beds at temperatures over 1023 K). On the other hand, potential catalysts for use to enhance close-to-equilibrium conditions in the RWGS reactions could include conventional supported metal catalysts with Cu as the active phase [40], although other precious metals (such as Pt and Rh) and bimetallic systems have been proposed. As stated in the same reference [40], if the aim of the RWGS reaction is to play a major role in e-fuels production, a catalyst with earth-abundant materials, such as Fe, would be preferable. However, since the issue of catalyst development for RWGS is outside the scope of this work, for the sake of simplicity, all experiments were Fig. 2. a) calcination equilibrium; b) desorption-enhanced reverse water–gas shift equilibrium. J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 4 performed with the use of a commercial Cu-catalyst. In addition, also for simplicity purposes, we assume here that there is an option, when needed, to preheat the packed bed of solids to the target temperature region in order to achieve maximum yields of P CO +P CO2 , as given by the DERWGS equilibrium at T DERWGS , and that H 2 can be preheated to T DERWGS . At this stage, we can also ignore the adiabatic cooling that endothermic DERWGS reactions will cause in the reaction zone (which can be compensated by periodically firing a fuel in the bed to heat up the solids again to a temperature slightly above T DERWGS , as claimed in [25]). Under such conditions, that implicitly assumes fast kinetics for both calcination and RWGS reactions Fig. 4 represents a particular point in time in a DERWGS experiment when the bottom part of the bed of solids has been already calcined (zone 1) and the calcination front is progressing upwards towards the unconverted solid region in the bed (zone 4). A differential volume of H 2 fed into the reactor will first pass unreacted through a region of CaO (marked as region 1). When H 2 reaches the CaO/CaCO 3 interphase in the packed bed, calcination will take place as a result of the partial pressure swing imposed by the influx of pure H 2 into the bed of CaCO 3 , and CO 2 concentration will increase sharply from this point (indicated by the blue line in the Figure on the right), also causing an expansion of the control volume. Assuming that there is sufficient catalytic activity in the element of volume with Δz as height, marked in Fig. 4, the RWGS reaction will also progress from the point at which the CO 2 from calcination becomes available in the gas phase. The RWGS reaction rate will accelerate downstream in the z direction (region 2 in Fig. 4, and red line in the Figure on the right) as the CO 2 concentration increases owing to calcination. If RWGS reaction kinetics do not take place with sufficient speed to allow equilibrium to be achieved within the bed control volume (marked in the Figure by its height, Δ Z ), the gas stream containing H 2 , CO 2 , CO and H 2 O will continue reacting to approach K WGS equilibrium downstream of the initial calcination front, aided by the presence of the RWGS catalyst (region 3 in the Figure) in the bed of solids. As CO 2 is being converted to CO along the bed length in region 3, some additional calcination of the solids may take place, and both CO 2 and CO concentrations will evolve further towards the DERWGS equilibrium, where the reactions will be negligibly slow (from point 4 onwards). In principle, the DERWGS reaction front, marked by regions 2–3, will advance towards the end of the reactor, as shown in Fig. 4. Depending on the kinetics of all reactions involved, the flow rates of H 2 to the reactor can be adjusted to reach suitable space times in order to achieve a product gas composition close to the DERWGS equilibrium at the reactor exit until a breakthrough occurs when the initial batch of CaCO 3 has been calcined and the reaction front marked by regions 2–3 reaches the end of the bed. The following section provides the first experimental proof of concept of the practical viability of the DERWGS equilibrium and a discussion of the conditions of temperature and pressure (as well as kinetic characteristics of the RWGS catalyst) needed to ensure that both the kinetics of calcination and RWGS reactions are sufficiently fast in order to exploit the benefits of the DERWGS equilibrium as shown in Figs. 3 and 4. 3. Experimental set up and materials The main set up used to investigate DERWGS reactions was that described in other studies of high temperature solid looping cycles [37] and is schematically represented in Fig. 5. It consisted of a packed bed of solids confined in a Kanthal tube with an internal diameter of 18 mm, and a maximum solid bed height of 325 mm. The bed was externally heated by an electric heating wire (power 1.25 kW and 5 m in length). As shown in the photograph in Fig. 5, an external insulation layer with a thickness of about 200 mm covered both the reactor and heating wire. The system was equipped with gas mass flow controllers that fed the reacting gases (H 2 , CO 2 and N 2 to close the mass balance) downwards from the top of the reactor. At the reactor exit, downstream of the water condenser, there was a back-pressure regulator valve that allowed the operation to be performed at pressures up to 5 atm. Gas analysis was Fig. 3. a) solid lines: solution to the derwgs equilibrium equations (4–10) when PT =1 atm. Dotted lines: RWGS equilibrium when pCO 2 at reactor inlet is equal to CaCO 3 calcination equilibrium CO 2 partial pressure. b) dependency of P CO_eq with temperature for different total pressures according to DERWGS equilibrium. Fig. 4. Representation of the progress of the DERWGS front through a packedbed reactor initially containing CaCO 3 and an RWGS catalyst. Graph on the right represents the evolution of CO 2 and CO with bed length referred to equilibrium. J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 5 performed on an aliquot of the product gas using a micro-GC (Varian CP4900) apparatus. The axial temperature profile during operation was measured and logged at 15 different points placed along the length of the reactor. An important detail for discussions below is that the control of the power input to the heating wire was based on the bed temperature as measured by the thermocouple immersed at the bottom part of the bed (i.e. the last portion of bed in contact with the gas). In addition to the set up in Fig. 5, a quartz micro packed-bed reactor (4 mm i.d.) placed inside a temperature-controlled furnace (capable of operating up to 1123 K), and fed by gas mass flow controllers for N 2 , H 2 and CO 2 , was used for two purposes: to rule out significant catalytic effects from the Kanthal wall in the experiments of Fig. 5 [41]; and to separately assess the catalytic activity for the RWGS of the Ca-materials used in the experiments, as there has been recent evidence reporting catalytic activity of CaO for RWGS [24] and [23]. In such experiments, the product gas composition was analysed with a Varian CP-3800, equipped with a HayesepQ and a Molsieve 13x, with a TCD detector. A more detailed study on the catalytic activity for RWGS of the solids used in this work is considered beyond its scope. With regards to materials, a standard high purity natural limestone (98.96 wt% CaCO 3 with MgO as main impurity) with a 1–2 mm size cut was used in the packed bed reactor experiments, while a finer size cut of 100–200 μ m was used to assess catalytic effects in the micro packed-bed reactor. Experiments with cycled CaO/CaCO 3 materials (i.e. with decreasing CaCO 3 content due to the carbonation/calcination cycling [42]) were carried out at different gas velocities between 0.04 m/s and 0.2 m/s in order to give a gas–solid contact time of between 7.5 and 0.74 s. To ensure fast RWGS reaction rates, as required to approach the DERWGS equilibrium, the initial tests were carried out by mixing solids containing CaCO 3 with a commercial Cu-based WGS catalyst known to have RWGS activity. As CaO has recently been reported to behave as a RWGS catalyst [23,24], experiments using only CaO/CaCO 3 with different carbonation contents were also carried out. The system operated cyclically during these experiments, and once the CaCO 3 was completely calcined in a DERWGS stage, a CaO carbonation stage was performed by introducing a CO 2 /N 2 gas stream with known composition into the reactor. By solving the CO 2 mass balance during the carbonation stages, it was possible to calculate the CaO carbonation conversion or CaCO 3 molar content , X N , at the beginning of the following DERWGS cycle. The evolution of X N with increasing number of cycles was consistent with the standard CO 2 carrying capacity decay curve presented by limestone when subjected to repetitive calcinationcarbonation cycles [39,42]. Furthermore, the average calcination rate and average CO production rate were determined in all the experiments from the mass balance applied to the system. The CO 2 and CO flows at the reactor exit were calculated based on product gas composition and N 2 flow at the inlet. In this way the average CaCO 3 calcination rate was calculated as the sum of CO 2 and CO flows divided by the reactor cross section, while the CO production rate was estimated from the CO flow at the reactor exit divided by reactor cross section. 4. Results and discussion Fig. 6 shows a first example of experimental results exhibiting DERWGS phenomena. The experiment was carried out by loading the packed-bed reactor with a batch comprising 0.04 kg of the 1–2 mm carbonated material (with CaCO 3 carbonate content, X N =0.74) and pellets of a Cu-based RWGS catalyst within the same particle size interval, to reach a Cu concentration in the bed of 8 wt%. The solids were first preheated to an average temperature of 1053 K in an atmosphere of pure CO 2 to prevent CaCO 3 decomposition. A fast switching of valves allowed the sudden feed of H 2 into the preheated solids (for convenience, in order to trace existing gas concentrations inside the analyser, 7.5 vol% N 2 was used to dilute the H 2 feed), at an inlet gas velocity of 0.19 m/s, resulting in a residence time for the gas of 1.7 s. The dashed lines, marked as T DERWGS , were estimated as the temperature at which P CO2_eq equals the experimental P CO2 measured in the product gas, which is consistent with the experimental temperature profiles in the bed, as it will be discussed below. As it can be seen in Fig. 6 a), during the first 1000 s of the experiment, the product gas composition is fully consistent with the DERWGS equilibrium at T DERGWS of 1043 K, calculated with Equations (4)–(10) and represented with dotted lines. Experimental P H2O is assumed to be equal to P CO as there are no other reactions apart from those in Fig. 2 b) that could consume H 2 O (i.e. no Ca(OH) 2 formation [43]). As of that point, a breakthrough appears in all gas compositions that coincides with the time at which the bed is approaching full calcination and the concentration of H 2 at the outlet is the same as at the inlet. Note that T DERWGS in this case is about 10 K lower than the initial temperature of the bed of solids. This can be explained by the endothermic nature of the DERWGS reaction, which generates temperature profiles in the bed that are unavoidable in the experimental set up used for these experiments, as it will be discussed below. According to mass balance applied to the product gas, 0.29 mol CaCO 3 was calcined in this test at an average calcination rate during the DERWGS period of 1.07 mol CaCO 3 /(m 2 •s) and a CO production rate of 0.62 mol CO/(m 2 •s). Since the initial CaCO 3 mass present in the bed was 0.3 mol, there was good closure of the carbon mass balance in this experiment. Fig. 6 b) confirms the feasibility of producing a syngas with the maximum CO content predicted by the DERWGS equilibrium when operating the system at different temperatures and pressures. In the example of Fig. 6 b), the packed-bed reactor had been operated at 5 atm total pressure and preheated at an average temperature of 1123 K, with an inlet gas flow containing 95 vol% H 2 and 5 vol% N 2 , resulting in a linear gas velocity of Fig. 5. Schematic diagram a) and picture of the experimental packed bed installation b). J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 6 0.043 m/s (gas/solid contact time 7.5 s). A breakthrough is again observed in the gas composition curves that corresponds to the complete consumption of the CaCO 3 in bed; however, the curves are less sharp than in Fig. 6 a) as the molar flow rate of H 2 into the reactor increased by 1.43 times. According to carbon mass balances, 0.32 mol CaCO 3 were calcined in the bed at an average rate of 1.07 mol CaCO 3 /(m 2 •s) during the DERWGS period, with an average CO formation rate of 0.65 mol CO/ (m 2 •s). The average experimental H 2 /CO ratio was 2.1, which was very close to the one predicted by the DERWGS equilibrium at 5 atm and 1113 K (H 2 /CO =2.2). As noted previously, Giammaria and Lefferts, [24], and later Sun et al. [23] recently reported the RWGS catalytic activity of the CaO surface resulting from CaCO 3 calcination. Therefore, additional set of experiments were performed in a bed without Cucatalyst. For this purpose, 70 g limestone (approximately 0.7 mol CaCO 3 ) was introduced in the reactor, resulting in a bed length of 150 mm. Nine consecutive DERWGS stages, each followed by a CaO carbonation stage in 90 vol% CO 2 , were performed in the reactor. The H 2 inlet gas flow rates were varied from 10 lN/h to 50 lN/h, resulting in linear gas velocities between 0.062 and 0.2 m/s. Fig. 7 a) shows experimental results from a test starting with limestone only in the bed. The reactor was preheated to an average bed temperature of 1035 K in pure CO 2 to prevent any CaCO 3 decomposition, after which 10 lN/h H 2 and 3 lN/h N 2 were introduced into the reactor, resulting in a linear gas velocity of 0.062 m/s and a gas–solid contact time of 2.47 s. A stable product gas with 56 vol% H 2 , 17 vol% CO 2 and 27 vol% CO was obtained at the reactor outlet, which was fully consistent with the DERWGS equilibrium at 1028 K. Fig. 7 b) shows the temperature evolution as measured by the six thermocouples immersed in the bed of solids. Because the gas entered from the top, the thermocouple at 25 mm was the first to face the reaction front. As previously mentioned, the thermocouple at 150 mm controlled the power input to the system. As can be observed in the Fig. 7 b) once the test started, the temperature at 25 mm decreased as CaCO 3 calcination and RWGS reactions were taking place. The temperature in this first bed slice decreased for approximately one hour. At this point, the CaCO 3 in the slice must have been calcined, and the reaction progressed to the following bed slice (note the T decrease at 50 mm from the gas inlet). The fact that no variations in temperature occurred downstream of the slice being calcined indicates that both CaCO 3 calcination and RWGS reaction took place in proximity. Otherwise, the gas composition would have varied along the bed length, and the bed temperature would have been affected by the reactions. According to mass balances, the calcination rate in the reactor was determined as 0.351 mol CO 2 /(m 2 •s) and the CO production was estimated as 0.213 mol CO/(m 2 •s). Furthermore, the obtained ratio H 2 /CO was 2.1, consistent with equilibrium predictions as a ratio H 2 /CO equal to 2.15 was calculated for a T DRWGS of 1028 K. The results of Fig. 7 confirm catalytic activity in the CaO bed for RWGS, since the kinetics of the homogeneous RWGS reaction at these Fig. 6. Product gas composition over time for DERWGS stages performed in a reactor containing CaCO 3 /CaO and RWGS catalyst at: a) 1 atm total pressure, with inlet gas velocity of 0.19 m/s at T DERWGS 1043 K. b) 5 atm total pressure, with inlet gas velocity of 0.043 m/s at T DERWGS 1113 K. Dotted-dashed lines in the figures correspond to DERWGS equilibrium compositions according to Equations (4–10) calculated at T DERWGS . Inlet gas composition marked by H 2 in the feed gas (being the rest N 2 ). Fig. 7. A) product gas composition over time for a derwgs stage performed in a reactor containing caco 3 /CaO at 1 atm total pressure, with inlet gas velocity of 0.062 m/s at T DERWGS 1028 K. Dotted lines in the figure correspond to DERWGS equilibrium compositions according to Equations (4–10) calculated at T DERWGS . Inlet gas composition marked by H 2 in the feed gas (being the rest N 2 ). b) Bed temperature profile evolution over time for the same test. J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 7 conditions would not have been able to justify CO 2 conversion of higher than 1 % in the absence of catalyst [41]. Although the detailed analysis of such catalytic effects for RWGS is outside the scope of this work, experiments were carried out in the previously described quartz micro packed-bed reactor to elucidate on the main source of such catalytic activity. For this purpose, 280 mg limestone was introduced in the micro packed-bed reactor. The experimental routine to determine CaCO 3 catalytic activity was to heat up the material in CO 2 to 1073 K to prevent calcination; then a mixture of 52 vol% H 2 and 41 vol% CO 2 balanced in N 2 (50 Nml/min total gas flow) was introduced into the reactor and the product gas analysed. A similar test with no solids was carried out to determine RWGS conversion in the empty reactor. After 30 min of recording stable product gas composition, the extent of RWGS reaction was determined according to Eq. (11) as the ratio between the equilibrium constant at these initial reacting conditions according to Eq. (6) [34] and the observed K WGS, obs calculated through Eq. (5): XRWGS =KWGS KWGS,obs (11) The value of X RWGS was lower than 10 % in experiments with CaCO 3 , and very similar to that obtained in the empty reactor. In contrast, and in agreement with the recent findings by Giammaria and Lefferts [24], the equivalent test conducted with CaO in the bed resulting from the calcination in N 2 of the carbonated material at 1073 K yielded a X RWGS value of 87 % (in this case, when feeding a mixture of 30 vol% H 2 and 19 vol% CO 2 balanced in N 2 to prevent any CaO carbonation). Therefore, CaO presented a catalytic activity of 0.11 mol CO produced/g CaO/h at 1073 K and 0.048 s residence time in the reactor. Such level of activity is consistent with the RWGS kinetic model in the presence of CaO as described by Giammaria et al. [24], from which the equation for CO formation reaction rates can be rewritten using the units and notation of this work as: RCO = koe(−Ea RT )PCO2aPH2b(1−KWGS KWGS,obs) 1+PH2OKH2O (12) with kinetic parameters within the range of those proposed in the referenced work (ln ko (s-1 bar-(a +b) =21; E a (kJ/mol) =135; a =0.6; b =0.8 and ln K H2O =6.7). From additional tests (some of which are shown in Figs. 8 and 9) with different values of the fraction of active CaO in the bed (varying from X N =1 to 0.24), it was observed that the catalytic activity for RWGS decreased with decreasing values of X N . Furthermore, for a given value of X N , increasing the H 2 flow resulted in a proportional reduction in breakthrough time and the sharpness of the DERWGS reaction front. As an example, Fig. 8 shows the evolution of the product gas composition and bed temperature profile over time for two experiments: a) at an inlet gas velocity of 0.14 m/s and average bed temperature of 1003 K in a bed containing 36.6 % moles CaCO 3 , being the rest CaO; b) for an inlet gas velocity of 0.2 m/s, at an average bed T of 1038 K in a bed containing 31.6 % moles CaCO 3 . From the temperature profiles in these two experiments, it can be observed that increasing the inlet H 2 flow increased the velocity of the DERWGS front, which reached the last portion of bed at about 2200–2400 s for the test performed at the lower gas velocity, compared to 900–1000 s for the experiment performed at 0.2 m/s gas velocity. These trends are consistent with the description of the DERWGS reaction front given in Fig. 4. A calcination rate of 0.346 mol CO 2 /(m 2 •s) with a CO production rate of 0.24 mol CO/(m 2 •s) was determined from the test in Fig. 8 a), while a calcination rate of 0.79 mol CO 2 /(m 2 •s) and the production of 0.454 mol CO/(m 2 •s) were determined for the test at the highest inlet gas flow. As can be observed, a gas with stable composition was obtained in both cases while there was still CaCO 3 in the bed; however higher H 2 /CO ratios than predicted by Fig. 8. Product gas composition, and bed temperature profile over time for a DERWGS stage performed in a reactor containing CaCO 3 /CaO at 1 atm total pressure: a) inlet gas velocity of 0.14 m/s at T DERWGS 1001 K; b) inlet gas velocity 0.2 m/s at T DERWGS 1030 K. Dotted lines in the figure correspond to DERWGS equilibrium compositions according to Equations (4–10) calculated at T DERWGS . Inlet gas composition marked by H 2 in the feed gas (being the rest N 2 ). J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 8 equilibrium were obtained (i.e. Fig. 8 a) H 2 /CO ratio =4.49 vs 3.67 predicted by equilibrium; Fig. 8 b) H 2 /CO ratio =4.05 vs 2.35 predicted by equilibrium). Fig. 9 shows an example with experimental conditions as in Fig. 7, but with a lower CaCO 3 content in the bed (25.6 wt%), being the rest inactive CaO resulting from the consecutive calcination-carbonation cycles experienced by this batch of material. As can be observed in Fig. 9 a), both the gas concentration profiles and temperatures show that the reaction front is occurring in a wider bed length and that the gas at the reactor outlet does not reach the DERWGS equilibrium (only CO 2 gas is at equilibrium). This indicates that the modest fraction of CaO generated by calcination does not provide sufficient catalytic activity for the RWGS reaction, as it seems reasonable to assume that the fraction of inactive CaO (i.e. the (1-X N ) fraction of CaO unable to react with CO 2 in carbonation experiments because it is covered by a CaCO 3 product layer [44,45]) cannot contribute to any catalytic activity. Therefore, in the absence of an effective RWGS catalyst downstream of the calcination front (see Fig. 9 b), the product gas composition at the reactor outlet (zone 4) should be close to that produced at the exit of the calcination front (zone 3). The previous observations highlight the need for an adequate characterisation of the catalytic activity for RWGS of all the solids located downstream of the calcination front generated when H 2 reaches the section of the bed containing CaCO 3 (region 2 in Fig. 9 b). However, the experimental data obtained in Figs. 7–9 show that catalytic activity for CaO in the short gas residence time expected within the calcination front (which can be estimated from temperature profile evolution) is coherent with the catalytic activity measured in the quartz reactor, which is in turn consistent with the results obtained by Giammaria and Lefferts [24] despite their different field of application. Future work should expand on these findings to other materials, such as combined CaO and RWGS catalyst materials (as the materials developed in other DFM applications reviewed in [13]). From the process design perspective, it will be important to explore the implications of scaling up the DERWGS process to the 10–20 m height reactor set up that is characteristic of similar high-temperature solid looping cycles using packed beds [46]. Two major process alternatives can be envisaged. The first is the production of CaO and syngas from the calcination of preheated limestone (or a solid with a high CaCO 3 content i.e. from a DAC system) with an H 2 stream, where the CaO RWGS catalytic activity may be sufficient, and no additional catalyst might be needed. The second major process route involves the integration of these reactions as part of a larger CaL CO 2 capture system (i.e. by carbonating CaO with CO 2 and relying on the modest Ca-conversion to CaCO 3 characteristic of these CaL systems) as has been conceptually described when using dual functional material operating at lower temperatures [13,19,22]. In this case, the low catalytic activity of the sintered CaO, characteristic of highly cycled Ca-materials, requires an RWGS catalyst in the bed of solids. However, since carbonation and DERWGS reactions would operate cyclically, no solid separation is required. In this case, in addition to the carbonation stage, an in-situ chemical looping combustion stage is needed to restore initial bed temperature in the carbonated bed required to launch a new DERWGS reaction stage [25]. The findings presented in this work open up a promising route for the calcination of CaCO 3 with H 2 followed by RWGS, according to reactions (2) and (3), to manufacture hydrocarbons and CaO. A DERWGS process can be seen as the reverse reaction of what has been investigated over many decades as sorption-enhanced reactions for H 2 production from fuels and CaO to form CaCO 3 . This paradigm shift is a sign of the renewable energy revolution in which the world is engaged today. 5. Conclusions The calcination of CaCO 3 with renewable H 2 (following reactions (2) and (3), while generating CaO as an industrial commodity or as a CO 2 capture sorbent, may become a viable route for the production of a syngas with suitable (CO 2 +CO)/H 2 molar ratios for hydrocarbon manufacture. Fig. 9. a) product gas composition, and bed temperature profile over time for a derwgs stage performed in a reactor containing caco 3 /CaO at 1 atm total pressure, same conditions as in Fig. 6. CaCO 3 content 25.6 wt%. Inlet gas composition marked by H 2 in the feed gas (being the rest N 2 ). b) Progress of the DERWGS front along a packed-bed reactor containing CaO/CaCO 3 , with calcination and RWGS occurring in a wide reaction front. Graph on the right represents the evolution of CO 2 and CO with bed length referred to equilibrium. J.C. Abanades and G. Grasa Chemical Engineering Journal 493 (2024) 152191 9 The equilibrium limitations imposed by the RWGS reaction, can be overcome at temperatures between 1023 and 1173 K by conducting such a reaction with an excess of CaCO 3 in order to keep the partial pressure of CO 2 close to the equilibrium of CO 2 in CaO, resulting in a desorptionenhanced reverse water–gas shift equilibrium of CaCO 3 on H 2 , DERWGS, which has been observed in a wide range of temperatures (between 1003 K to 1113 K) pressures (between 1 to 5 atm), carbonate content (X N from 1 to 0.24) and suitable gas–solid contact times for large scale applications. The first necessary condition to enable DERWGS to be experimentally observed in packed-bed reactors containing CaCO 3 solids is the development of a sufficiently fast calcination reaction front when H 2 reaches the bed of CaCO 3 solids (i.e. temperatures exceeding 1000 K at 1 atm in the reaction front are required). The second condition for DERWGS is that there is sufficient catalytic RWGS activity in the bed of solids located downstream of the calcination reaction front. This can be achieved by mechanically mixing an RWGS catalyst with the bed of solids containing CaCO 3 , or by using CaO materials with sufficient activity in relation to CO 2 capture, which have been shown to display catalytic activity for RWGS consistent with the model by Giammaria and Lefferts [24]. Product gases containing CO, CO 2 and H 2 , and H 2 O (v) with an H 2 /CO molar ratio of 2 or lower are obtained in a wide range of operating conditions that can form the basis for scaling up this process to direct routes for syngas and hydrocarbon production by the reductive calcination of CaCO 3 with renewable H 2 (through equations (2) and (3). CRediT authorship contribution statement J.C. Abanades: Writing – review & editing, Methodology, Conceptualization. G. Grasa: Writing – original draft, Investigation, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. Acknowledgements This research is part of the CSIC programme for the Spanish Recovery, Transformation and Resilience Plan (PTI +TRANSENER) funded by the Recovery and Resilience Facility of the European Union, established by the Regulation (EU) 2020/2094. References [1] H.C. Lee, K. Calvin, D. Dasgupta, G. Krinner, A. Mukherji, P. Thorne, SYNTHESIS REPORT OF THE IPCC SIXTH ASSESSMENT REPORT (AR6), Summary for Policymakers., in: IPCC (Ed.) IPCC, 2023. [2] V. Dieterich, A. Buttler, A. Hanel, H. Spliethoff, S. Fendt, Power-to-liquid via synthesis of methanol, DME or Fischer–Tropsch-fuels: a review, Energy and Environmental Science 13 (2020) 3207–3252, https://doi.org/10.1039/ d0ee01187h. [3] M. G¨ otz, J. Lefebvre, F. M¨ ors, A. McDaniel Koch, F. Graf, S. Bajohr, R. Reimert, T. Kolb, Renewable power-to-gas: A technological and economic review, Renewable Energy 85 (C) (2016) 1371–1390, https://doi.org/10.1016/j. renene.2015.07.066. [4] A.D.N. Kamkeng, M. Wang, J. Hu, W. Du, F. Qian, Transformation technologies for CO 2 utilisation: current status, challenges and future prospects, Chem. Eng. J. 409 (2021) 128138, https://doi.org/10.1016/j.cej.2020.128138. [5] L.J. Müller, A. K¨ atelh¨ on, M. Bachmann, A. Zimmermann, A. Sternberg, A. Bardow, A guideline for life cycle assessment of carbon capture and utilization, Frontiers in Energy Research 8 (2020) 1–20, https://doi.org/10.3389/fenrg.2020.00015. [6] J.C. Abanades, E.S. Rubin, M. Mazzoti, H.J. Herzog, On the climate change mitigation potential of CO 2 conversion to fuels, Energy & Environmental Science 10 (2017) 2491–2499, https://doi.org/10.1039/c7ee02819a. [7] S. Lux, G. Baldauf-Sommerbauer, M. Siebenhofer, Hydrogenation of inorganic metal carbonates: a review on its potential for carbon dioxide utilization and emission reduction, CHEMSUSCHEM 11 (2018) 3357–3375, https://doi.org/ 10.1002/cssc.201801356. [8] P. Jiang, L. Li, G. Zhao, H. Zhang, T. Ji, L. Mu, X. Lu, J. Zhu, Reductive calcination of calcium carbonate in hydrogen and methane: A thermodynamic analysis on different reaction routes and evaluation of carbon dioxide mitigation potential, Chem. Eng. Sci. 276 (2023) 118823, https://doi.org/10.1016/j.ces.2023.118823. [9] B. Shao, Y. Zhu, J. Hu, Y. Zong, Z. Xie, S. Li, W. Du, M. Wang, H. Liu, F. Qian, Chemical engineering solution for carbon neutrality in cement industry: tailor a pathway from inevitable CO 2 emission into syngas, Chem. Eng. J. 483 (2024) 149098, https://doi.org/10.1016/j.cej.2024.149098. [10] R.H. Borgwardt, Calcination kinetics and surface area of dispersed limestone particles, AIChE J. 31 (1985) 103–111, https://doi.org/10.1002/aic.690310112. [11] M.E. Boot-Handford, J.C. Abanades, E.J. Anthony, M.J. Blunt, S. Brandani, N. Mac Dowell, J.R. Fern´ andez, M.-C. Ferrari, R. Gross, J.P. Hallett, R.S. Haszeldine, P. Heptonstall, A. Lyngfelt, Z. Makuch, E. Mangano, R.T.J. Porter, M. Pourkashanian, G.T. Rochelle, N. Shah, J.G. Yao, P.S. Fennell, Carbon capture and storage update., Energy and Environmental Science 7 (2014) 130-189. https://doi.org/10.1039/ c3ee42350f. [12] M. Voldsund, S.O. Gardarsdottir, E. De Lena, J.F. P´ erez-Calvo, A. Jamali, D. Bersad, C. Fu, M. Romano, S. Roussanaly, R. Anantharaman, H. Hoppe, D. Sutter, M. Mazzoti, M. Gazzani, G. Cinti, K. Jordal, Comparison of technologies for CO 2 capture from cement production—Part 1: technical evaluation, Energies 12 (2019) 559, https://doi.org/10.3390/en12030559. [13] R. Han, Y. Wang, S. Xing, C. Pang, Y. Hao, C. Song, Progress in reducing calcination reaction temperature of Calcium-Looping CO 2 capture technology: A critical review, Chem. Eng. J. 450 (2022) 137952, https://doi.org/10.1016/j. cej.2022.137952. [14] J. Chen, L. Duan, Y. Ma, Y. Jiang, A. Huang, H. Zhu, H. Jiao, M. Li, Y. Hu, H. Zhou, Y. Xu, F. Donat, M.A. Naeem, O. Kr¨ ocher, Recent progress in calcium looping integrated with chemical looping combustion (CaL-CLC) using bifunctional CaO/ CuO composites for CO2 capture: A state-of-the-art review, Fuel 334 (2023) 126630, https://doi.org/10.1016/j.fuel.2022.126630. [15] D.W. Keith, G. Holmes, D.M. St. Angelo, K. Heidel, A Process for Capturing CO 2 from the Atmosphere., Joule 2(8) (2018) 1573-1594. https://doi.org/10.1016/j. joule.2018.05.006. [16] M. Erans, S.A. Nabavi, V. Manovic, Carbonation of lime-based materials under ambient conditions for direct air capture, Journal of Cleaner Production 242 (2020) 118330, https://doi.org/10.1016/j.jclepro.2019.118330. [17] J.C. Abanades, Y.A. Criado, H. White, Direct capture of carbon dioxide from the atmosphere using bricks of calcium hydroxide, Cell Reports, Physical Science 4 (2023) 101339, https://doi.org/10.1016/j.xcrp.2023.101339. [18] https://www.heirloomcarbon.com/technology. [19] H.S. Sun, J. Wang, J. Zhao, B. Shen, J. Shi, J. Huang, C. Wu, Dual functional catalytic materials of Ni over Ce-modified CaO sorbents for integrated CO 2 capture and conversion, Applied Catalysis b: Environmental 244 (2019) 63–75, https://doi. org/10.1016/j.apcatb.2018.11.040. [20] A.A. Giardini, J.F. Lakner, Synthesis of graphite and hydrocarbons by reaction between calcite and hydrogen, Science 159 (1968) 317–319, https://doi.org/ 10.1126/science.159.3812.317. [21] A. Reller, C. Padeste, P. Hug, Formation of organic carbon compounds from metal carbonates, Letters to Nature 329 (1987) 527–529. [22] S. Sun, C. Zhang, S. Chen, X. Zhao, Y. Wang, S. Xu, C. Wu, Integrated CO 2 capture and reverse water–gas shift reaction over CeO2-CaO dual functional materials, Royal Society Open Science 10 (2023) 230067, https://doi.org/10.1098/ rsos.230067. [23] S. Sun, Z. Lv, Y. Qiao, C. Qin, S. Xu, C. Wu, Integrated CO 2 capture and utilization with CaO-alone for high purity syngas production, Carbon Capture Science & Technology 1 (2021) 100001, https://doi.org/10.1016/j.ccst.2021.100001. [24] G. Giammaria, L. Lefferts, Synergy between dielectric barrier discharge plasma and calcium oxide for reverse water gas shift, Chem. Eng. J. 392 (2020) 123806, https://doi.org/10.1016/j.cej.2019.123806. [25] J.C. Abanades, G. Grasa, Method of hydrogenation and calcination of CaCO 3 with H 2, EP22383163.7 1/12/2022. [26] S. Shi, J. Yu, Y. Pan, Y. Zhang, H. Yang, T. Shen, Q. Liu, Z. Liu, Hydrogenation of calcium carbonate to carbon monoxide and methane, Fuel 354 (2023), https://doi. org/10.1016/j.fuel.2023.129385. [27] D. Beruto, A.W. Searcy, Use of the Langmuir Method for Kinetic Studies of Decompostion Reactions: Calcite (CaCO 3 ). Jornal of the Chemistry Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases. (0) (1974) 21452153. https://doi.org/10.1039/F19747002145. [28] E.H. Baker, The calcium oxide-calcium dioxide system in the pressure range 1–300 atm, J. Chem. Soc. (1962) 464–470, https://doi.org/10.1039/JR9620000464. [29] C. Han, D.P. Harrison, Simultaneous shift reaction and carbon dioxide separation for the direct production of hydrogen, Chem. Eng. Sci. 49 (24) (1994) 5875–5883, https://doi.org/10.1016/0009-2509(94)00266-5. [30] A. Lopez, D.P. Harrison, Hydrogen Production Using Sorption Enhanced Reaction. , Ind. Eng. Chem. Res. 40(23) (2001) 5102-5109. https://doi.org/10.1021/ ie001009c. [31] D.P. Harrison, Sorption-enhanced hydrogen production: A review, Ind. Eng. Chem. Res. 47 (2008) 6486–6501, https://doi.org/10.1021/ie800298z. J.C. Abanades and G. Grasa