Simulation of Biogas Upgrading by Sorption-Enhanced Methanation with CaO in a Dual Interconnected Fluidized Bed System

: In this work, ASPENplus was used to simulate biogas upgrading by sorption-enhanced methanation in a dual interconnected bubbling ﬂuidized bed conﬁguration using inexpensive, abundant, and eco-friendly CaO to remove H 2 O from the reaction environment. The chemical looping scheme consisted of two reactors: a methanator/hydrator, where the catalytic reactions occurred on a catalyst with 20% Ni supported on alumina as well as the steam removal by CaO, and a regenerator, where the Ca(OH) 2 was dehydrated back to CaO. The simulations were carried out to identify possible reactant compositions (H 2 and biogas), CaO amount, and the methanation temperature able to produce an outlet gas matching the speciﬁcations for direct grid injection. When considering a stoichiometric gas feed ratio at the methanator inlet, the unwanted CaO carbonation worsened the process performance, subtracting CO 2 from the desired methanation reaction. However, optimal conditions were found with hydrogen-lean gas feedings, balancing the limited H 2 amount with the capture of CO 2 due to the sorbent carbonation. Thermodynamic considerations pointed out the possibility of solid carbon formation induced by sorption-enhanced methanation conditions, especially for H 2 sub-stoichiometric feedings.


Introduction
In the framework of the efforts to address the energy transition, a renewed attention is growing for the catalytic or biological processes to produce synthetic renewable methane (synthetic or substitute natural gas-SNG).This is due to the crucial role played in numerous sectors by this important energy carrier, which benefits from a developed storage and distribution infrastructure in many countries, and a large public acceptance.Catalytic paths to renewable methane production can involve either valorization of residual biomass by gasification and subsequent syngas methanation, or CO 2 methanation combined with power-to-gas applications, converting surplus renewable electric energy into a grid-compatible gaseous fuel.This latter application includes using green hydrogen obtained by water electrolysis, either for methanation of CO 2 from streams of captured CO 2 (carbon capture and utilization-CCU), or for biogas upgrading to increase the methane content of streams from digestion of organic feedstock [1][2][3].Referring to this last solution, biogas, its production has received attention as a promising renewable path for energy valorization of residual biomass, and it mostly consists of CH 4 (generally ca.50-75%) and CO 2 (ca.25-45%), with percentages determined by feedstock and digestion operating conditions.The most used feedstocks are sewage, farming residues, and the organic fraction of domestic and industrial waste.The heterogeneity of these sources, besides the consequent technological challenges, also implies that the removal of contaminants in the gas phase is a sensitive issue, and there appears to be ample room for research, especially on the development of cost-effective materials [4].Among the mentioned technological issues of anaerobic digestion, there are challenges such as the methods to reduce the inhibition of anaerobic sludge digestion by emerging contaminants, especially pharmaceuticals, for example.These rising matters have been recently addressed in the literature [5].
Although even trace amounts of species like H 2 O, O 2 , NH 3 , and N 2 may need to be removed, the core upgrading stage for injection into the natural gas grid or usage as a fuel is CO 2 removal, with the separated CO 2 generally being released into the atmosphere.Conventional and well-established separation processes rely on absorption, adsorption, membrane, cryogenic, or biological removal [6].An interesting alternative to such separation processes may be the further methanation of CO 2 in raw biogas to produce additional CH 4 .
Catalytic CH 4 production, in addition to CO 2 hydrogenation (1), can occur via CO hydrogenation (2), according to the reactions discovered by Sabatier and Senderens in 1902 [7], with high exothermicity and a net decrease in moles: CO + 3H 2 CH 4 + H 2 O − 206 kJ/mol (@298K). ( Commercial processes are generally found in ammonia synthesis plants to eliminate carbon monoxide traces and rely, due to the high exothermicity of the reactions, on a series of adiabatic catalytic fixed beds operated at temperatures between 250 and 600 • C with intermediate cooling steps and recycles, and at high operational pressure.Apart from the above-mentioned original industrial application, studies of synthetic natural gas (SNG) production as a standalone process have led to other commercial solutions being developed over the past decades.In the 1970s and 1980s, Haldor Topsøe carried out a steam reforming of methane using nuclear energy and then transported the synthesized gas to a heat-consuming site to be reconverted in a cyclic process by means of methanation, exploiting the high reaction enthalpy; TREMP stands for Topsøe's Recycle Energy efficient Methanation Process.The scheme consists of three adiabatic fixed bed reactors with a recycle to the first reactor, temperatures from 250 to 700 • C, pressure up to 30 bar, and a near-stoichiometric hydrogen to carbon oxides ratio at the inlet.The original project was completed but the TREMP process is still used to produce SNG from synthesis gas.The process has been demonstrated in plants of a semi-commercial scale, producing between 200 and 2000 Nm 3 /h of SNG under realistic industrial conditions, and mainly relies on the purposely developed Topsøe MCR methanation catalysts which present high and stable activity in a wide temperature range [8].On the other hand, the promising power-to-gas solution has been applied on an industrial scale in Audi motor company's "e-gas" facility in Werlte, Germany, where 1000 metric tons/year of synthetic natural gas (SNG) are produced using concentrated CO 2 from a nearby biogas plant [9].
The difficult temperature control appears to be a sensitive issue related to, for example, catalysts' stability and their deactivation in the presence of high temperature hotspots.Regarding methanation catalysts, if considering the main active phases (Ru, Ni, Fe, Co, and Mo), nickel is in general found in industrial solutions since it is quite inexpensive despite being significantly active and selective [2].Unfortunately, in addition to thermal deactivation, commercial nickel catalysts are also subject to mechanical and chemical deactivation [10].The latter is mainly represented by carbon generation and accumulation on the catalyst surface [11].Because of the more efficient control of large-scale exothermic processes, fluidized beds have been suggested as a possible alternative to the consolidated fixed bed configurations [2].
Recently, among different research activities directed to innovate the traditional process, the sorption-enhanced methanation (SEM) concept has been proposed, based on Le Chatelier's principle, to improve reaction performance by selectively removing produced steam from the reaction environment [12,13].The local H 2 O adsorption, by shifting the equilibrium toward the products generation, would allow for overcoming of thermodynamic constraints, improving methanation performance at relatively low pressure.This solution may be applied to reduce the energy duty for gas compression work.Early studies proposed sorption-enhanced CO 2 methanation [12] at atmospheric pressure and low temperatures (250 • C) and investigated nickel as an active metal and zeolite 4A as a sorbent in a cyclic series of regenerative experiments.Although an almost-unitary conversion was detected, the authors highlighted a worsening in terms of water capture (15-20%) due to the parallel CO 2 adsorption.These studies were first carried out by Walspurger et al. [12], focusing on biogas upgrading.They analyzed the thermodynamics of CO 2 methanation in ASPEN Plus under typical conditions found in industrial processes (CO 2 captured and hydrogen produced by electrolysis available at nearly 1 bar and 40 • C).The target was to obtain an SNG quality suitable for a certain gas grid of specifications at a pressure of 60 bar.The target of an injectable stream will also be considered in our study, considering average standards in the EU on a dry molar basis: H 2 < 2%, CO 2 < 2.5%, CO < 0.5%.Walspurger et al. selected the traditional fixed bed configuration and modelled the influence of in situ water removal, considering the third reactor either as a conventional reactor or as a sorption-enhanced reactor.It was seen from the simulation that a water-removing reactor in place of the third conventional reactor led to a significantly lower H 2 level in the SNG and that the gas grid specifications could be matched at much lower operational pressure (nearly 30 bar compared to 60 bar of a conventional case).The authors pointed out that operating the methanation at relatively low pressure may determine a significant compression energy saving: more than 60% at pressures lower than 5 bar.The temperature increase aspect was also investigated by these authors, providing results that support the suggested application of fluidized bed reactors.In fact, at pressures below 10 bar the amount of CO 2 and H 2 to be converted in a sorption-enhanced reactor would lead to an excessive temperature rise, considering this process as a third step in a fixed bed reactor scheme.
Borgschulte et al. [13] demonstrated the enhancement using bifunctional materials combining sorption and catalyst properties by impregnating Ni on a zeolite 5A.Delmelle et al. [14] identified zeolite 13X as the sorbent with the highest steam adsorption capacity as a bifunctional material by adding Ni to its structure.Therefore, the key to the success of SEM technology relies on identifying a proper material to adsorb H 2 O at the temperatures of interest for methanation, which are significantly high for an adsorption process, and to allow for its dehydration at temperatures that, although higher than the adsorption ones, could be acceptable in terms of catalyst operation and energy requirements.Generally, most of the studies in the literature involve fixed bed operation conditions.So far, chemical looping configurations with sorption-enhanced applications have involved mostly CO 2 removal in the sorption-enhanced steam reforming of ethanol [15][16][17].
By taking advantage of the favorable features of fluidized beds, SEM chemical looping based on the technology of coupled fluidized reactors would allow for steady operation [1], enabling continuous sorbent transfer and regeneration (Figure 1).In fact, despite the compact design of fixed beds configurations, these reactors would imply discontinuous SEM process, and difficult heat control on a large scale [2].Coppola et al. [18] selected calcium oxide and a commercial zeolite 3A as sorbents, studying adsorption/desorption capacity over cycles in a chemical looping dual fluidized bed system.Zeolites, the most investigated sorbent in the literature, present a stable trend over the cycles; however, CO 2 competes with H 2 O during the sorption process.CaO shows a decline in steam capture capacity under cyclic operation, induced also by the undesired carbonation reaction.CaO was also tested by Agirre and coworkers [19]; this sorbent seems to be promising for SEM application, even if possible negative effects on the catalyst call for additional studies.Despite several drawbacks, CaO is an appealing possibility because it is inexpensive, abundant, and ecofriendly, unlike zeolites.
seems to be promising for SEM application, even if possible negative effects on the catalyst call for additional studies.Despite several drawbacks, CaO is an appealing possibility because it is inexpensive, abundant, and ecofriendly, unlike zeolites.Sorption-enhanced methanation using CaO to perform biogas upgrading in a dual interconnected fluidized bed (DIFB) configuration was modelled in ASPENplus with the idea of detecting working conditions for the process in terms of feeding gas and CaO amounts, as well as methanation temperatures, to obtain product compositions suitable for direct injection in the natural gas grid.The results were also subjected to thermodynamic considerations in terms of possible coke deposition enhanced by SEM conditions.In fact, although carbon deposition during the traditional process occurs exclusively during CO methanation [6,[20][21][22][23], thermodynamic studies demonstrated that SEM conditions may also enhance carbon formation during CO2 methanation [24].In particular, Catarina Faria et al. [6] analyzed the thermodynamics of CO2 methanation under SEM conditions applied to a biogas upgrading process, pointing out the increased importance of carbongenerating reactions with a much more extensive carbon formation than in the absence of H2O removal.

Aspen Flowsheet
ASPENplus V10 software was used to simulate the SEM process in a chemical looping DIFB configuration.The Fluidbed' blocks, as isothermal and one-dimensional reactors, were used, and the kinetics of all the main reactions was implemented in them.The flowsheet of the plant, operated at atmospheric pressure, is shown in Figure 2, where METH' and REGEN' indicate the methanator and the regenerator, respectively.For the METH' block, the methanation temperature was varied in a certain range (250-300-350 °C), while the REGEN' temperature was fixed at 450 °C; both reactors were simulated at nearly atmospheric pressure.Sorption-enhanced methanation using CaO to perform biogas upgrading in a dual interconnected fluidized bed (DIFB) configuration was modelled in ASPENplus with the idea of detecting working conditions for the process in terms of feeding gas and CaO amounts, as well as methanation temperatures, to obtain product compositions suitable for direct injection in the natural gas grid.The results were also subjected to thermodynamic considerations in terms of possible coke deposition enhanced by SEM conditions.In fact, although carbon deposition during the traditional process occurs exclusively during CO methanation [6,[20][21][22][23], thermodynamic studies demonstrated that SEM conditions may also enhance carbon formation during CO 2 methanation [24].In particular, Catarina Faria et al. [6] analyzed the thermodynamics of CO 2 methanation under SEM conditions applied to a biogas upgrading process, pointing out the increased importance of carbon-generating reactions with a much more extensive carbon formation than in the absence of H 2 O removal.

Aspen Flowsheet
ASPENplus V10 software was used to simulate the SEM process in a chemical looping DIFB configuration.The 'Fluidbed' blocks, as isothermal and one-dimensional reactors, were used, and the kinetics of all the main reactions was implemented in them.The flowsheet of the plant, operated at atmospheric pressure, is shown in Figure 2, where 'METH' and 'REGEN' indicate the methanator and the regenerator, respectively.For the 'METH' block, the methanation temperature was varied in a certain range (250-300-350 • C), while the 'REGEN' temperature was fixed at 450 • C; both reactors were simulated at nearly atmospheric pressure.Table 1 summarizes several parameters implemented for the SEM DIFB.The diameter of the methanator was varied to set a fluidization velocity equal to 1 m/s for the different inlet biogas feeding conditions (with a fixed biogas inlet of 0.4 kmol/h) at the three methanation temperatures considered.Table 1 summarizes several parameters implemented for the SEM DIFB.The diameter of the methanator was varied to set a fluidization velocity equal to 1 m/s for the differ-ent inlet biogas feeding conditions (with a fixed biogas inlet of 0.4 kmol/h) at the three methanation temperatures considered.The default or selectable correlations used in Aspen for modeling the bottom and freeboard zones, and particle size distribution (PSD); calculating the transport disengaging height (TDH); and determining the minimum fluidization velocity and the elutriation, are those found in the literature [25][26][27][28].

Kinetics
Table 2 reports the kinetics implemented for the simulations.In addition to the reactions previously introduced as ( 1) and ( 2), and the water-gas shift reaction kinetics, for which a 20% Ni supported on alumina catalyst was considered, CaO hydration and undesired irreversible carbonation were also implemented in the methanator, since CO 2 was present.The details of the kinetic equations and their literature sources are reported in previous work by the authors [28].
In the 'REGEN' block, N 2 was fed (about 40 Nm 3 /h), and Ca(OH) 2 was dehydrated and then fed back to the methanator.Only the reverse reaction, R4 (Table 2), was therefore implemented in this block.
Lastly, the flowsheet was completed by a 'PURGE' (Figure 2) and a fresh sorbent make-up ('SIN' in Figure 2) to avoid calcium carbonate accumulation in the loop, since the sorbent deactivated via carbonation in 'METH' cannot be regenerated through calcination in the 'REGEN' block due to the low temperature.The solid stream at the inlet of the methanator ('FEED') was fixed at 100 kg/h with the 'RECYCLE', calculated by means of the Aspen calculator block to maintain this value.

Key Parameters
The performances of the SEM process were evaluated by the definition of three different parameters that describe the compositions of the inlet gas in the methanator and the amount of the inlet flow of the sorbent CaO in relation to the gaseous reactants.The first parameter adopted in this work is α, which defines the methanation reactants feed ratio; when α = 3 the gas feed ratio is stoichiometric with respect to reactions R1 and R2 (Table 2), while lower values indicate a sub-stoichiometric gas feed with respect to H 2 : where all the species are expressed in molar flows.In this work, as the biogas is simply considered as a mixture of CO 2 and CH 4 , the expression of α can be simplified as follows: Specifically, three different values have been investigated: α = 1,2,3.
The parameter ϕ characterizes the biogas in terms of CO 2 and CH 4 : Two specific values of ϕ were explored in order to simulate two opposite conditions of a biogas poor and rich in methane (ϕ = 0.5 and 0.7), respectively.
Lastly, the third parameter is strictly related to the SEM process, because it is a measure of the inlet sorbent amount of H 2 O removal with respect to the reactant gases of the methanation process.The latter, indicated as θ, and already introduced in a previous work by the same authors [28], is defined as follows: where CaO st is the stoichiometric sorbent molar flow required for the total removal of the maximum theoretical amount of H 2 O that the methanation reactions can generate, considering that for each mole of steam one mole of CaO (R4) is necessary.From the R1 reaction stoichiometry, CaO st is equal to 2CO 2 if the inlet gas ratio is stoichiometric (α = 3), or sub-stoichiometric with respect to CO 2 (α ≥ 3).Conversely, since CO is totally absent in the feeding gas with an over-stoichiometric composition with respect to CO 2 (α < 3), the maximum amount of H 2 O is linked exclusively to the quantity of H 2 as follows: A complete description of the calculation of CaO st in the case when both CO and CO 2 are present in the inlet gas can be found elsewhere [28].
Four distinct values of θ have been investigated: θ = 0 (no SEM condition) used as a benchmark condition; θ = 0.5, a sub-stoichiometric condition of CaO with respect to the quantity of steam that can be produced at most by the inlet gas in the system; θ = 1, a stoichiometric condition of CaO; and θ = 2 an over-stoichiometric condition of CaO.

Effect of CaO on the Product Gas Quality
The effect of CaO on the product gas composition is reported in Figure 3 at different α and ϕ and for a fixed temperature of 300 • C. The compositions considered to be acceptable for direct injection in the grid are indicated in green, taking into account the average limits in Europe (H 2 < 2%, CO 2 < 2.5%, CO < 0.5% on a dry molar basis).On the other hand, the product gas compositions highlighted in red are those corresponding to operating conditions resulting in a gas quality which does not match the grid specifications.

Effect of CaO on the Product Gas Quality
The effect of CaO on the product gas composition is reported in Figure 3 at different α and φ and for a fixed temperature of 300 °C.The compositions considered to be acceptable for direct injection in the grid are indicated in green, taking into account the average limits in Europe (H2 < 2%, CO2 < 2.5%, CO < 0.5% on a dry molar basis).On the other hand, the product gas compositions highlighted in red are those corresponding to operating conditions resulting in a gas quality which does not match the grid specifications.At 300 °C, for a stoichiometric composition of the reactant gas (α = 3), starting from the benchmark case corresponding to a traditional methanation process, where θ = 0 since no CaO is fed, the outlet CO and CO2 are lower than the maximum concentrations accepted.The critical value, as also reported in the literature from purely thermodynamic considerations [24], is the H2 at the outlet, which is above 2% for both the φ considered.The dependence of the outlet gas concentration on φ is quite low.The H2 concentration slightly approaches the grid standard with φ increasing from 0.5 to 0.7.Although the H2 conversion decreases in the case of an inlet feeding gas richer in CH4 (φ = 0.7) as expected, the lower H2 concentration under this case is due to the better inlet gas quality in terms of methane and the lower stoichiometric H2 amount required for CO2 methanation.Overall, no condition fulfills the required specifications for a stoichiometric (α = 3) feed at 300 °C.Under a sub-stoichiometric ratio when α = 2, an overturning of the product gas behavior occurs compared to the α = 3 case.When α = 2, the CO2 outlet composition now does not match the standards.This concentration strongly increases with respect to the stoichiometric feed, moving from less than 1% to 12.64% and 7.73% for φ = 0.5 and φ = 0.7, respectively.On the contrary, the H2 output decreases at α = 2, as expected, but is more weakly affected by α than CO2.In particular, if considering the value of H2 corresponding to the φ = 0.7 case, it would respect the limit concentration for injection under this sub- At 300 • C, for a stoichiometric composition of the reactant gas (α = 3), starting from the benchmark case corresponding to a traditional methanation process, where θ = 0 since no CaO is fed, the outlet CO and CO 2 are lower than the maximum concentrations accepted.The critical value, as also reported in the literature from purely thermodynamic considerations [24], is the H 2 at the outlet, which is above 2% for both the ϕ considered.The dependence of the outlet gas concentration on ϕ is quite low.The H 2 concentration slightly approaches the grid standard with ϕ increasing from 0.5 to 0.7.Although the H 2 conversion decreases in the case of an inlet feeding gas richer in CH 4 (ϕ = 0.7) as expected, the lower H 2 concentration under this case is due to the better inlet gas quality in terms of methane and the lower stoichiometric H 2 amount required for CO 2 methanation.Overall, no condition fulfills the required specifications for a stoichiometric (α = 3) feed at 300 • C.Under a sub-stoichiometric ratio when α = 2, an overturning of the product gas behavior occurs compared to the α = 3 case.When α = 2, the CO 2 outlet composition now does not match the standards.This concentration strongly increases with respect to the stoichiometric feed, moving from less than 1% to 12.64% and 7.73% for ϕ = 0.5 and ϕ = 0.7, respectively.On the contrary, the H 2 output decreases at α = 2, as expected, but is more weakly affected by α than CO 2 .In particular, if considering the value of H 2 corresponding to the ϕ = 0.7 case, it would respect the limit concentration for injection under this sub-stoichiometric condition.CO also increases compared to the α = 3 condition, but its concentration is still much lower than the limit.However, once again, no condition can be detected for the direct injection into the grid.The same conclusion can be drawn for α = 1, where the CO 2 concentrations are higher than in the previous case; with the lowest H 2 amount fed to the system (α = 1), its concentration also fulfills the limit for ϕ = 0.5, unlike the α = 2 case where only at ϕ = 0.7 was its concentration below 2%.Although CO increases as α decreases, this species still remains below the limit.
When considering the results of the actual SEM process, with a sub-stoichiometric CaO injection in the reactive system (θ = 0.5), and for the stoichiometric gas feed case (α = 3), the H 2 outlet dramatically increases with respect to the benchmark case and a detrimental effect of the CaO can thus be identified.In particular, the undesired CaO carbonation reaction results in a removal of CO 2 , subtracting it from the methanation reaction and accelerating the consumption of CO in the WGS reaction, implying the production of more H 2 .Indeed, unlike the θ = 0 case, no carbon-containing contaminants are detected at the outlet.When considering sub-stoichiometric feed with respect to H 2 (α = 2), the presence of CaO keeps the CO 2 at the outlet below the limit unlike the traditional methanation case (θ = 0).However, under this condition the H 2 amount is still above the limit for the two ϕ considered.For α = 1, the outlet condition is reversed: the decreased H 2 feeding involves a sufficient decrease in H 2 at the outlet for injection in the grid, but the effect of CaO cannot counteract the strong increase in CO 2 for such a H 2 -lean feeding.The CO contaminant slightly increases at the outlet with α decreasing, similarly to what happens for the θ = 0 case.Again, under θ = 0.5 conditions, no gas outlet matches the grid specifications.However, given the reversal behavior between H 2 and CO 2 moving from α = 2 to α = 1, it could be possible that for an intermediate value of α both species meet grid restrictions.When considering stoichiometric CaO feeding (θ = 1), the same trend observed with θ = can be summarized.For a stoichiometric gas feed (α =3), the composition in terms of H 2 clearly worsens by increasing the sorbent feed; however, for the sub-stoichiometric value of α = 2, an optimal working condition can be identified for the direct injection into the grid corresponding to the ϕ = 0.7 case.Lastly, for the over-stoichiometric CaO fed to the system (θ = 2) and for the lowest alfa considered (alfa = 1), the excess of CO 2 is also offset by carbonation for ϕ = 0.5.Under these working conditions, the contaminants at the outlet meet the grid specifications.

Effect of Temperature on the Product Gas Quality
Regarding the methanation temperature effect, when the temperature is increased, thermodynamic conditions become more unfavorable and, predictably, the performance becomes worse (Figure 4).By increasing the temperature from 300 to 350 °C, under traditional methanation conditions (θ = 0) and α = 3, the outlet CO2 for both φ is only slightly below the limit value for the injection, demonstrating a worsened performance compared to 300 °C.Indeed, unlike the 300 °C case, the H2 concentration remains above the limit even when considering substoichiometric feeds (7.16% and 5.69% for α = 2 and α = 1, respectively).By increasing the temperature from 300 to 350 • C, under traditional methanation conditions (θ = 0) and α = 3, the outlet CO 2 for both ϕ is only slightly below the limit value for the injection, demonstrating a worsened performance compared to 300 • C. Indeed, unlike the 300 • C case, the H 2 concentration remains above the limit even when considering sub-stoichiometric feeds (7.16% and 5.69% for α = 2 and α = 1, respectively).
As for the SEM process with different CaO amounts fed to the methanator (θ varying from 0.5 to 2), results may be summarized as follows: • CO, which is not a reactant in this biogas upgrading scheme but can be formed by other reaction paths such as the Reverse Water Gas Shift reaction, increases with α decreasing and tends to decrease for the highest values of θ.However, it is always below the limit; • CO 2 is always below the limit except for when the sub-stoichiometric value of CaO (θ = 0.5) and α = 1; under this condition the sorbent fed is still not sufficient to compensate CO 2 excess in the feed gas; • H 2 represents, again, the critical species, always being above the limit value even for the best conditions in terms of its consumption, and namely under the limit case with minimum α and maximum θ, where the H 2 detected is about 4% for both values of ϕ.
Overall, at 350 • C no parameter values can be identified as optimal working conditions for the direct injection of the product gas.
At the lowest temperature investigated (250 • C), the process is clearly thermodynamically favored (Figure 5).For a stoichiometric feed and conventional methanation, the product stream is already injectable in the grid.However, when considering an α decrease, i.e., the realistic conditions for an industrial process at sub-stoichiometric feed in terms of H2, an increase in CO2 occurs, as was also the case at 300 °C and 350 °C.Again, the molar CO2 percentage is around 13% and 25% at φ = 0.5 and 8%, and 15% at φ = 0.7 for α equal to 2 and 1, respectively.The values under these conditions are similar for all the temperatures considered.Unlike traditional methanation, no conditions ensure a directly injectable outlet gas when considering α = 3 with the introduction of CaO due to the negative outcome on the outlet H2 amount.As for α < 3, the H2 composition is always suitable for injection even at θ = 0.5, while for this sub- However, when considering an α decrease, i.e., the realistic conditions for an industrial process at sub-stoichiometric feed in terms of H 2 , an increase in CO 2 occurs, as was also the case at 300 • C and 350 • C. Again, the molar CO 2 percentage is around 13% and 25% at ϕ = 0.5 and 8%, and 15% at ϕ = 0.7 for α equal to 2 and 1, respectively.The values under these conditions are similar for all the temperatures considered.Unlike traditional methanation, no conditions ensure a directly injectable outlet gas when considering α = 3 with the introduction of CaO due to the negative outcome on the outlet H 2 amount.As for α < 3, the H 2 composition is always suitable for injection even at θ = 0.5, while for this sub-stoichiometric value of CaO feed, the CO 2 percentage is still above the limit.At α = 2, since the CO 2 amount remarkably decreases with θ, working conditions acceptable in terms of contaminants can be found from the stoichiometric value of CaO (θ = 1) onward.Lastly, for θ = 2, an optimal condition is also detected for α = 1.

SEM Thermodynamics and Possible Carbon Formation
Ternary diagrams C-H-O have been used to detect possible thermodynamic conditions for carbon generation for the system under scrutiny.Figure 6 reports in triangular coordinates the molar fraction of C-H-O elements in the gas phase and the calculated carbon formation boundaries for the different temperatures [24].The points constituting the isotherms (green, blue, and red dashed lines in Figure 6) are the combinations of values for C-H-O elements in the gas phase in equilibrium with solid carbon.Above the lines coke deposition is possible from a thermodynamic point of view.The dots in the ternary diagram indicate the different feeding gas ratios.Evidently, with conventional methanation (θ = 0), the feeding gas and the outlet compositions coincide in terms of C-H-O molar fractions.On the other hand, H2O removal from the system (SEM conditions) makes the feeding points move toward the upper zone of the graph, leading to an enrichment of C in the system as indicated by the straight lines in Figure 6 ( steam capture' lines).The complete removal of steam from the reaction environment in the stoichiometric case' (α = 3) would determine pure CH4 as the outlet composition point (on the H-C axis).Differently, for α < 3, the final value would correspond to a mixture also containing oxygen due to the excess carbon oxide compounds (CO and CO2).Moreover, it can be noted that stoichiometric' feeding dots (α = 3) fall outside the carbon generation area, in particular, when φ is equal to 0.5.When the feeding is characterized by φ = 0.7, it falls just on the boundary line of carbon formation at 350 °C (red line in Figure 6).On the contrary, all feeding gases with α < 3 are located within the carbon deposition area.Under these conditions, even if considering traditional (no-SEM) methanation, carbon generation is favored from a thermodynamic point of view.For all the locations of the feeding dots (inside or outside the carbon formation zone), looking at Figure 6 it is clear that the capture of only steam always leads to the region favored for Evidently, with conventional methanation (θ = 0), the feeding gas and the outlet compositions coincide in terms of C-H-O molar fractions.On the other hand, H 2 O removal from the system (SEM conditions) makes the feeding points move toward the upper zone of the graph, leading to an enrichment of C in the system as indicated by the straight lines in Figure 6 ('steam capture' lines).The complete removal of steam from the reaction environment in the 'stoichiometric case' (α = 3) would determine pure CH 4 as the outlet composition point (on the H-C axis).Differently, for α < 3, the final value would correspond to a mixture also containing oxygen due to the excess carbon oxide compounds (CO and CO 2 ).Moreover, it can be noted that 'stoichiometric' feeding dots (α = 3) fall outside the carbon generation area, in particular, when ϕ is equal to 0.5.When the feeding is characterized by ϕ = 0.7, it falls just on the boundary line of carbon formation at 350 • C (red line in Figure 6).On the contrary, all feeding gases with α < 3 are located within the carbon deposition area.Under these conditions, even if considering traditional (no-SEM) methanation, carbon generation is favored from a thermodynamic point of view.For all the locations of the feeding dots (inside or outside the carbon formation zone), looking at Figure 6 it is clear that the capture of only steam always leads to the region favored for carbon formation.However, the situation is more complicated using CaO as a sorbent due to its ability to react with both H 2 O and CO 2 .In this last case, a subtraction of C will occur besides H and O and the 'migration paths' of feeding dots would be different from the conditions under which only H 2 O is removed.These paths would depend on the amount of sorbent θ, its different intrinsic affinity with CO 2 and H 2 O, and how the temperature influences the kinetics of the reactions.In Figure 7  CaO SEM curves tend to achieve compositions richer in H and C, since a signifi consumption of O occurs under these conditions.The higher the temperature, the m the migration paths move away from those relative to only steam removal since hi temperatures favor higher CO2 capture.Regarding the temperature range of this w the C deposition is likely to be quite slow [29], especially when considering the effe temperature control of a fluidized bed configuration.In addition, the combined CO2 H2O removal keeps the system closer to the boundaries of the carbon deposition which could be a further benefit for reducing the carbon generation.However, since C Ca(OH)2, and CaCO3 could also affect the carbon generation kinetics, further experim should confirm these predictions.

Conclusions
Biogas upgrading by SEM in DIFB considering CaO as sorbent was simulated in PENplus, with the aim of analyzing the worsening effect of the parallel carbonation tion and detecting possible working conditions to reach suitable product gas, in term contaminants, for direct injection it into the natural gas grid.The simulations involve study of the effect of the feed composition (in terms of α and φ) and CaO fed to the sy (θ), as well as the effect of the methanation temperature in the range 250-350 °C.
A traditional methanation process would entail the possible direct injection e sively for the lowest methanation temperature of 250 °C and only for a stoichiometri feeding (α = 3).However, it may be worth underlining that a 250 °C operating tempera is more of a theoretical limit, while the catalyst would be sufficiently active only higher temperatures.
As for the SEM process, the extent of the detrimental effect of CaO carbonation CaO SEM curves tend to achieve compositions richer in H and C, since a significant consumption of O occurs under these conditions.The higher the temperature, the more the migration paths move away from those relative to only steam removal since higher temperatures favor higher CO 2 capture.Regarding the temperature range of this work, the C deposition is likely to be quite slow [29], especially when considering the effective temperature control of a fluidized bed configuration.In addition, the combined CO 2 and H 2 O removal keeps the system closer to the boundaries of the carbon deposition area, which could be a further benefit for reducing the carbon generation.However, since CaO, Ca(OH) 2 , and CaCO 3 could also affect the carbon generation kinetics, further experiments should confirm these predictions.

Conclusions
Biogas upgrading by SEM in DIFB considering CaO as sorbent was simulated in ASPENplus, with the aim of analyzing the worsening effect of the parallel carbonation reaction and detecting possible working conditions to reach suitable product gas, in terms of contaminants, for direct injection it into the natural gas grid.The simulations involved the study of the effect of the feed composition (in terms of α and ϕ) and CaO fed to the system (θ), as well as the effect of the methanation temperature in the range 250-350 • C.
A traditional methanation process would entail the possible direct injection exclusively for the lowest methanation temperature of 250 • C and only for a stoichiometric gas feeding (α = 3).However, it may be worth underlining that a 250 • C operating temperature is more of a theoretical limit, while the catalyst would be sufficiently active only from higher temperatures.
As for the SEM process, the extent of the detrimental effect of CaO carbonation, resulting in increasing H 2 concentration, determines that with stoichiometric gas (α = 3) no conditions can fulfill the grid specifications even at the lowest temperature of 250 • C. As for sub-stoichiometric conditions, with respect to H 2 (α < 3), the outlet gas quality improves by increasing CaO, with the CO 2 excess being compensated by CaO carbonation.The detected gas and CaO feedings able to produce a suitable gas for injection can be summarized as a function of the process temperature.For example, at 300 • C for α = 2 and ϕ = 0.7 (CH 4 -rich biogas), the produced gas composition is suitable for the gas grid from a stoichiometric value of CaO (θ = 1) onward, while if considering ϕ = 0.5 (CH 4 -lean biogas), this is true only for α = 1 and an over-stoichiometric value of CaO (θ = 2).At 250 • C, a suitable gas is obtained for both biogas compositions considered: for α = 2 from θ = 1 onward, and for α =1 at θ =2.On the contrary, for the higher methanation temperature of 350 • C, no conditions are detected for a proper outlet gas quality.
In conclusion, as long as a sub-stoichiometric gas with respect to H 2 is fed to the process, it is possible to select CaO as a possible sorbent material to carry out SEM.In the framework of the power-to-methane processes, H 2 -lean feedings are of interest considering the wide range of possible conditions and flexibility required to manage the fluctuations associated with such applications.Since a thermodynamic study indicated that with H 2 O capture carbon formation cannot be excluded, especially under sub-stoichiometric H 2 feed, an experimental study may be required to identify the actual extent of carbon deposition under such conditions.

Figure 3 .
Figure 3. H 2 , CO 2 , and CO molar percentages on dry basis (rest CH 4 ) at the outlet of methanator as a function of α, Φ, and θ at T = 300 • C. Colors: green = acceptable, red = non-acceptable (for direct grid injection).

Figure 4 .
Figure 4. H 2 , CO 2 , and CO molar percentages on dry basis (rest CH 4 ) at the outlet of methanator as a function of α, Φ, and θ at T = 350 • C. Colors: green = acceptable, red = non-acceptable (for direct grid injection).

Figure 5 .
Figure 5. H 2 , CO 2 , and CO molar percentages on dry basis (rest CH 4 ) at the outlet of methanator as a function of α, Φ, and θ at T = 250 • C. Colors: green = acceptable, red = non-acceptable (for direct grid injection).

Figure 6 .
Figure 6.C-H-O ternary diagram reporting carbon deposition boundaries (at p = 1 atm and temperatures in the range 250-350 °C) and the feed compositions at different values of α and Φ.The straight lines reproduce the reaction environment composition after progressive steam removal.

Figure 6 .
Figure 6.C-H-O ternary diagram reporting carbon deposition boundaries (at p = 1 atm and temperatures in the range 250-350 • C) and the feed compositions at different values of α and Φ.The straight lines reproduce the reaction environment composition after progressive steam removal.

12 Figure 7 .
Figure 7. C-H-O ternary diagram with computed C-H-O points from SEM simulations with Ca = 3, Φ = 0.5, and for different values of θ).

Table 1 .
Main parameters for the reactors.

Table 1 .
Main parameters for the reactors.

Table 2 .
Reactions implemented in Aspen.