Development of a High Temperature CO2 Sorbent Based on Hydrotalcite for a H2-Rich Syngas Production

To adapt hydrotalcite-based sorbents (also known as layered double hydroxides—LDHs) to high-temperature CO2 sorption compatible with tar steam reforming, the addition of CaO was investigated, maintaining the LHDs porosity and accessibility but mostly assuring the CO2 sorption stability during sorption/desorption cycles. In co-precipitation synthesis, the investigated parameters are (i) various interlayer anions with different sizes and valences (carbonate, oxalate, and stearate); (ii) various pH values; (iii) different Mg/Ca molar ratios. The characterization of these modified LDHs by TGA, XRD, N2 adsorption, SEM, sorption capacity, and sorption/desorption stability (cyclic TGA) allowed understanding the effect of the various synthesis parameters and highlighted the effect of oxalate use as the interlayer anion. After calcination of sorbent with Mg/Ca/Al ratio = 1/2/1, typical LDH sand roses were formed both with carbonate and oxalate anions: this former exhibited the highest sorption capacity and accessibility of CaO sites at 600 °C, higher than pure CaO. However, the best stability during cycles was obtained with the sorbent from oxalate and Mg/Ca/Al ratio = 1.5/1.5/1 at pH 10 for which comparable sorption results are reached. For these two samples, the observed macro-porosity was associated with the highest specific surface area and pore volume.


Introduction
The research about sustainable energy and the production of chemical compounds from biomass is a major challenge for the scientific community and industry. The biomass feedstock valorization allows decarbonizing the energy sources and can relocate the production for more local autonomy. The optimization of biomass treatments is also important for the ecological resilience of all countries especially those of the global South. Lignocellulosic biomasses can be treated via thermochemical processes: combustion to generate heat and power, pyrolysis and gasification to produce mainly biooils and syngas, respectively. Residual biomasses and wastes gasification are an interesting route to produce a gas mixture (syngas) mainly containing hydrogen and carbon monoxide from a wide range of recoverable wastes (agricultural, industrial waste, or forestry residues, solid recovered fuel, etc.), terrestrial and aquatic biomass crops, but the process also leads to the formation of carbon dioxide, methane, water and impurities as tar, H 2 S, HCl, NH 3 . The content of these impurities varies depending on the biomass feedstock, reactor type, and gasification conditions [1][2][3] and an efficient and cost-effective process is necessary to purify the produced syngas. Furthermore, to assure a reduction of its environmental footprint [4] by decarbonization, the main undesirable product (CO 2 ) has to be separated for further use.
On the other hand, the in situ CO 2 capture (Reaction 1) improves steam reforming (of methane for example Reaction 2) by displacing the equilibrium of the water gas shift (Reaction 3) towards the hydrogen production. Both operations enrich syngas in hydrogen and then improve its energy capacity. This process is also called sorption enhanced hydrogen production and allows producing two separate streams of H 2 and CO 2 , the latter easily reusable and/or storable.
The Sorption Enhanced Steam Reforming (SESR) and Sorption Enhanced Water Gas Shift (SE-WGS) can be combined in a single reactor for an intensified syngas purification. Characteristics required to undertake this process configuration led us to develop sorbent materials to assure CO 2 sorption at high temperature with great thermal stability under cyclic (sorption/desorption) operating conditions. Hydrotalcite structure possesses well-known CO 2 sorbent properties [19,20], and the mixed oxides obtained by their calcination at moderate temperature lead to optimal CO 2 sorption capacity and regeneration capability [19,21].
Hydrotalcite is a magnesium-aluminum hydroxycarbonate double lamellar hydroxide (LDH) having the formula [Mg (1−x) Al x (OH) 2 ][CO 3x/2 ] Z H 2 O. The most advantages of this structure are: (1) The possibility to exchange alkaline earth (Ca 2+ instead of Mg 2+ ) to increase the sorption temperature and improve its sorption capacity [22]. The exchange of a part of Mg 2+ by Ca 2+ will allow adapting the CO 2 sorption/desorption temperature to the catalytic reforming effective limit conditions. Indeed, at atmospheric pressure, magnesium oxide-based hydrotalcite absorbs until 400 °C whereas calcium oxide-based hydrotalcite absorbs between 400 and 600 °C and desorbs beyond according to hydrotalcite Ca/Al ratio [23]. On the other hand, CaO sorbents, subject to lose their CO 2 sorption capacity during regeneration cycles, were effectively stabilized by Mg and/or Al incorporation as inert support materials [24]. These materials lead to the dispersion of CaO particles during the synthesis procedure, limiting the sintering phenomenon due to multiple carbonation/calcination cycles. Then, CO 2 sorption capacity can be maintained at a level of theoretical values during more than 20 cycles by adding MgO [25] and during 10 cycles by adding Al in an optimal Ca/Al molar ratio of 7/1 [26]. The maximum sorption capacity of CaO is theoretically 78.4 g CO2 /100 g sorbent . (2) A width interlayer space in which anions nature (CO 3 2− , SO 4 2− , or NO 3 − ) can be modified to impact the carbon dioxide sorption kinetics and capacity [27]. The spheroidal "sand rose" morphology observed with CO 3 2− anions explained the large BET surface area and the (Reaction 1) better CO 2 sorption capacity [28]. Stearate (and longer molecules) exchange was extensively studied as a longcarbon-chain organic anion and allowed improving the CO 2 capture performance due to more surface basicity sites (O 2− ) after calcination [29]. (3) The CO 2 sorption favored by the presence of water vapor also assists the desorption step [30,31]. According to Reddy et al. [32], the sorption capacity of an LDH material between 100 and 400 °C is lower under dry conditions compared to wet conditions (water saturation of 12%) and presents easier desorption. (4) The possibility to exchange Mg 2+ and Al 3+ ions by Ni 2+ and Fe 3+ , respectively, to add a catalytic effect to these sorbents for reforming reactions and water gas shift (WGS) [6]. This goal will be further studied and the active metal will be chosen among the most currently studied in steam reforming, namely Ni for its high reactivity [33][34][35], Co for its reasonably high activity for C-C cleavage [36,37], Fe to improve the conversion of the carbon formed on the surface of the catalyst [38][39][40], and Ce to reinforce the sorbent basicity [40][41][42]. (5) The H 2 S sorption on LDH partially irreversible, competitive with CO 2 , and stronger than CO 2 sorption but also slower [21]. Around 100 ppm to a few percent can be present in syngas from gasification. Van Dijk et al. [43] suggest that LDH sorption capacity decreases during CO 2 sorption/desorption cycles in the presence of H 2 S but becomes quite stable after few cycles.
Different synthesis methods of hydrotalcite are available and each one may bring benefit. Urea hydrolysis [44], hydrothermal [44,45], sol-gel process [46] or co-precipitation [47] are the more common methods. Among these possibilities, the chosen method must allow the substitution of magnesium by calcium, carbonate by a bigger anion and aluminum by a transition metal like iron. In the urea hydrolysis method, urea decomposes into a carbonate anion that cannot be changed by a bigger anion. In the hydrothermal method, the diffusion of metal ions should be difficult without available and efficient stirring equipment for the addition of calcium oxide. The sol-gel process is interesting to change carboxyl groups into hydrogel matrices and pH at the same time but needs many reactants. Finally, the co-precipitation method is not only the more used and easier to scale up, but also to change cations and pH regardless of the used anion and their solubility.
This paper presents the development of hydrotalcitederived mixed oxides as CO 2 sorbents. Several fundamental parameters were studied in this work. If the (Mg + Ca)/Al molar ratio is still equal to 3/1 as commonly in the literature [48] to have a higher content of basicity in the material, for effective and enhanced CO 2 sorption [49], variables Mg/Ca ratios were evaluated to improve CO 2 sorption properties and stability. Sodium carbonate typically used as interlayer space anion was changed by bigger anions. The literature reports studies of interlayer charge-compensating anions (Cl − , NO 3 − , CO 3 2− , HCO 3 − , SO 4 2− , …) [27,28,50] to increase materials porosity. On the other hand, better sorption sites accessibility of LDHs intercalated with long-carbon-chain carboxylic acids (from stearate to palmitic acid) [29,51] can create more surface basicity (O 2− ) sites, but no obvious difference in CO 2 capture capacity was observed for carbon chain lower than 10. Moreover, Yong et al. previously reported the positive effect of the higher valence of CO 3 2− compared to OH − on CO 2 sorption capacity [27]. Thereby, this work focuses on another sodium anion (oxalate) with the same valence of sodium carbonate (2), a shape and valence different from stearate (1), and not yet reported in the literature. Finally, two different pH values were evaluated affecting structural and morphological properties.
TGA, XRD, SEM, and N 2 physisorption tell us about structure, morphology, and porosity. Sorbents are evaluated in terms of their sorption capacity and of the stability of this property under sorption/desorption cycles by TGA.

Synthesis Method
All the studied sorbents respect the (Mg + Ca)/Al molar ratio equal to 3/1, like the natural mineral composition. Magnesium, calcium, and aluminum nitrates (Mg(NO 3 ) 2 ·6H 2 O, Ca(NO 3 ) 2 ·6H 2 O, and Al(NO 3 ) 3 ·9H 2 O) were solubilized in 50 mL of deionized water and an aqueous solution of sodium carbonate (Na 2 CO 3 ), oxalate (Na 2 C 2 O 4 ) or stearate (NaC 18 H 35 O 2 ) was added dropwise at 60 °C then the mixture was vigorously stirred. The equivalent of precipitating agent (NaOH) was calculated according to the anion charge.
The valence of stearate being lower than that of other anions and the molecular weight of sodium stearate being the highest of studied anions, its weight necessary for the LDH synthesis was largely higher than for the sodium di-anions. The LDHs synthesis from sodium stearate is complicated by its very low solubility in water (140 mg/L at 20 °C) and the hydrotalcite structure was difficult to obtain. It was reported that at pH 10, Al 3+ and Mg 2+ are precipitated as Al(OH) 3 and Mg(OH) 2 nanoparticles, respectively, at the same time, that immediately converts into Mg 3 Al 2 -CO 3 LDH [52]. At higher pH, Mg 2+ precipitates first as Mg(OH) 2 nanoparticles. The precursor solubility was assure by a lower pH value (8). This latter was measured during the co-precipitation by a pHmeter (type HI2202-02 from Hanna Instruments) and controlled by a drip addition of sodium hydroxide (3 mol/L). After 1 h of maturation, the product was filtered, washed with deionized water at room temperature, and dried 1 3 overnight at 100 °C. Layered double oxides (LDOs) were obtained after calcination at 500 °C or 700 °C for 1 h after a 10 °C/min ramp. They were crushed to obtain a powder with grains smaller than 250 µm.

Materials Nomenclature
The for mula of LDH precursor is globally In the name of each Mg/Ca/Al studied sample, the molar ratio of these three elements is followed by the anion abbreviation, then the used pH are indicated as reported in Table 1. The 0/3/1 CO3.pH8 was not synthesized because the 0/3/1 CO3.pH10 was just studied as reference for the other compositions to prove the interest of Ca dispersion in Mg oxide.

Characterization Techniques
Firstly, characterization was performed by Thermal Gravimetric Analysis (TGA) on LDHs to evaluate their decomposition profile during further calcination.
LDOs samples were characterized by X-Ray Diffraction (XRD) to identify the present crystalline phases, N 2 adsorption for textural properties, Scanning Electron Microscopy (SEM) for morphology, sorption tests to evaluate their CO 2 sorption capacity, and Thermal Gravimetric Analysis (TGA) to study their stability during carbonation/calcination (sorption/desorption) cycles.

Thermal Gravimetric Analysis (TGA)
A Linseis L81 is used for TGA. The sample (from 1 to 4 mg, grain size smaller than 250 µm) is calcined under 25 mL/ min of air during a heating ramp of 10 °C/min up to 700 °C.

X-Ray Diffraction (XRD)
XRD spectra were recorded by X-ray diffractometer Brucker AXS D8 Advanced using CuKα radiation to detect crystalline phases. Phase identification is performed by comparison with JCPDS (Joint Committee on Powder Diffraction Standards) files. Spectra are recorded in a 2θ range from 10° to 90°, with a scanning step equal to 0.0158° and a sampling time of 1 s per step.

N 2 Physisorption
A Micromeritics ASAP 2420 analyzer was used to record N 2 adsorption and desorption isotherms at − 196 °C for each sample (40 mg, grain size smaller than 250 µm), previously degassed at 250 °C for at least 7 h. Properties as specific surface area and pore volume were calculated from desorption isotherm and BET and BJH methods, respectively.

Scanning Electron Microscopy (SEM)
Few grains of the sample are sticking on a cylindrical sample carrier using double-sided adhesive tape. Pictures were taken by a Zeiss Gemini SEM 500 equipped with an In-lens SE detector, an electron high tension usually at EHT = 2 kV, a sample-objective distance of WD = 4-5 mm, and different fields width values between 2 µm and 200 nm.

Sorption Tests
Sorption Capacity Technique 100 mg of the sorbent (grain size smaller than 250 µm) is charged into a U quartz reactor which a 0.6 cm diameter. Thanks to a 4-ways valve, the reactor can be supplied by a 15% CO 2 gas mixture (7.5 mL/ min of CO 2 in 42.5 mL/min of He) for the sorption step or by a pure He flow (50 mL/min) for the purge and desorption steps (Fig. 1).
The flow at the reactor outlet is sent to a CO 2 IR detector (Emerson X2GP IR). An additional He flow (200 mL/ min) permits the dilution of the analyzed flow before the analyzer. The analyzer possesses three CO 2 infrared detectors with sensibility limits of 3000 ppm, 10,000 ppm, and 200,000 ppm, respectively. In our case, the CO 2 concentrations were evaluated with the third detector with an experimental uncertainty of about ± 4% of the sorption capacity.
The sorption capacity test is composed of three steps ( Fig. 2): Step under He from room temperature until 800 °C (10 °C/min). The CO 2 desorbed during this step corresponds to the decomposition of CaCO 3 remaining after calcination at 500 °C (see section 'Calcination Temperature Effect') or formed during storage of the samples. Step begins once inlet CO 2 is purged, continues during a temperature increase (rate 5 °C/min) until 800 °C, and finishes when no more CO 2 is detected. Then, the reactor oven is turned off and the He flow is maintained through the rector until the room temperature is reached.

Stability (TGA Cycles)
The same Linseis L81 thermobalance previously described for TGA was used for TGA-multicycles carbonation/calcination tests. Between 1 and 2 mg of powdered (grain size smaller than 250 µm) sample is charged on the cradle.
The cycles began with a Purge Step to decompose CaCO 3 or Ca(OH) 2 formed during storage under room atmosphere. The Purge Step was composed of a treatment under N 2 (25 mL/min) during a ramp of temperature until 800 °C at 10 °C/min then a plateau at this temperature for 20 min. Then an alternation of the Carbonation Step and Desorption Step was repeated from 7 to 20 times for each sample.
The Carbonation Step took place at 600 °C under CO 2 flow (25 mL/min containing 20% CO 2 in He) for 90 min to saturate the sorbent (conversion of oxides into carbonates).
The Desorption Step was performed under the N 2 step (25 mL/min) during a ramp of temperature up to 800 °C at 5 °C/min and an isotherm during 20 min.
It should be noted three main differences between this multicycles test and sorption test: • The weight was around five hundred times less in TGA-multicycles tests than in sorption tests; • Gas flow passed through the sorbent bed during sorption tests contrary to TGA-multicycle tests in which gas flow breath above sample. So, the sorption will be more limited in the second case; • The Carbonation Step duration is the same for both sorption tests, but for TGA-multicycle tests it may be insufficient to achieve sorbent saturation.
The values of CO 2 sorption capacity for each cycle are calculated according to the lowest and weights highest (so after merely 40 min of sorption) (Eq. 3).
With m max and m min : highest and lowest weight respectively measured at the beginning and at the end of each cycle; m i : initial sorbent mass (g); M CO2 : CO 2 molar weight (g/mol); M sorbent : sorbent molar weight (g/mol).

LDH Characterization
The thermogravimetric analysis of 3/0/1 .pH10 samples shows the typical two-stage weight loss (Fig. 4). After a first step (between 50 and 250 °C) attributed to the water (physisorbed and inter-layer) removal, a second step (between 220 and 700 °C) is due to the removal of the octahedral layer hydroxyls and inter-layer anion groups [29].

Anion Effect
After water removal, the calcination of 3/0/1 .pH10 samples prepared from various sodium anions (Fig. 3) begins at the same temperature (around 250 °C) and their weight losses occur in one or two steps more or less distinguishable and become stable above 500 °C. Due to the molecular weight of the interlayer stearate, the remaining weight at the end of calcination is lower for stearate (22%) than for carbonate (62%) and oxalate (60%) anions.

pH Effect
The TGA profiles are similar for carbonate samples (Fig. 4a) with both pH values (3/0/1 CO3·pH8 and 3/0/1 CO3·pH10 ). The decomposition of the LDH structure begins around 200 °C and finishes around 500 °C. The same behavior is observed for oxalate (Fig. 4b) and stearate (Fig. 4c) samples for which the decomposition always finishes at a temperature lower than 500 °C. It is worth to note that for oxalate samples (3/0/1 Ox ) and stearate samples (3/0/1 St ), the rate of weight loss presents various distinct maximum values for the two different studied pH, whereas very similar curves are obtained for both carbonate samples. In fact, the study of thermal decomposition of LDH precursors under air revealed two or three steps thermogravimetric losses [22]. The first stage at low temperature is assigned to surface adsorbed or interlayer water molecules and the two other stages (between 200 and 500 °C) to the dihydroxylation of the layer hydroxyl groups and the decomposition of interlayer carbonate.

Mg/Ca Ratio Effect
TGA oxidation indicates (Fig. 5) that for the sample with the maximum of Ca (0/3/1 CO3.pH10 ), the only mass loss after water removal starts at high temperature (around 560 °C) with a maximum around 645 °C associated with the decomposition of CaCO 3. Conversely, for the sample without Ca (3/0/1 CO3.pH10 ), the mass loss starts just after water removal with a maximum of around 365 °C associated with the decomposition of MgCO 3 . For the intermediate samples (other Mg/Ca values), after water removal, the TGA profiles present two-stages weight loss combining MgCO 3 and CaCO 3 decompositions as previously observed [22].
TGA thermograms of various Mg/Ca/Al CO3.pH10 samples indicate that, if the 3/0/1 CO3.pH10 sample is decarbonized after calcination at 500 °C, the substitution of Mg by Ca needs a higher calcination temperature of at least 700 °C for CaCO 3 total decomposition. After calcination at 700 °C, around 40% of weight loss is observed for each sample (Fig. 5).  As far as the pH values and Mg/Ca ratio are concerned, the calcination of oxalate samples containing Ca passes through the decomposition of oxalate, then carbonate and finishes around 700 °C with a remaining weight between 50 and 60% (Fig. 6a) while the calcination of stearate samples is complete at 500 °C with a remaining weight around 25% (Fig. 6b).
As a conclusion of LDHs characterization, the calcination temperature of 500 °C is enough to obtain LDOs from LDHs prepared without Ca at pH 8 or 10 and with carbonate or oxalate anion (3/0/1 ·pH10 and 3/0/1 ·pH8 samples) and for those prepared with stearate interlayer anion with various compositions. On the contrary other compositions, e.g. with Ca and carbonate and oxalate interlayer anions, need a higher temperature of calcination (700 °C) to produce calcium oxide available for CO 2 sorption.

XRD
Anion Effect For all the 3/0/1 compounds calcined at 500 °C, the organic part of the material decomposes (dehydroxylation and decarbonation), leading to the formation of the LDH structure then of mixed oxides (LDO) with a 3D network [48]. As expected from TGA results, for all the 3/0/1 compounds prepared at pH 10 and calcined at 500 °C, the only phase produced is amorphous mixed oxides (MgAlO x ), with the features of periclase (JCDPS n° 00-045-0946) [29], regardless of the interlayer anion used for LDH synthesis and its size or its valence (Fig. 7).
pH Effect XRD permits the identification of the formation of MgO periclase as the main phase regardless of the pH used. However, for 3/0/1 Ox.pH8 sample, a spinel phase (MgAl 2 O 4 )  is also identified [53]. So, pH does not affect LDOs formation from carbonate interlayer anion nor from other anions (Fig. 7) for 3/0/1 composition except for 3/0/1 Ox.pH8 sample.

Mg/Ca Ratio Effect
For the sample containing calcium, XRD analysis allows attesting to the formation of a phase mixture after calcination at 500 °C. MgO periclase and a magnesium calcium carbonate (Mg 0.03 Ca 0.97 CO 3, JCPDS n° 01-089-1304) are observed in all the sorbents containing calcium from carbonate (Fig. 8a), oxalate (Fig. 8b), and stearate (not shown) interlayer anions. In all these samples, the magnesium calcium carbonate phase is visible and is much more crystallized than the periclase phase. The main periclase intense ray (2θ = 42.9°) is hidden by the magnesium calcium carbonate ray at 2θ = 43.3° and disappears for the largest Ca contents.
The evolution of the Mg/Ca ratio does not affect the phase distribution in the sample prepared from various anions: the XRD diffractograms are similar for materials using carbonate anion for 2/1/1 and 1.5/1.5/1 Mg/Ca/Al compositions (Fig. 8a) and using oxalate anion for 2/1/1, 1.5/1.5/1 and 1/2/1 Mg/Ca/Al compositions (Fig. 8b). If there is an effect of pH, it is undetectable by XRD because the same phase distribution is observed (not shown).

Calcination Temperature Effect
The calcination temperature between 500 and 700 °C does not affect the formation of LDOs without calcium (3/0/1 ratio) from various interlayer anions. XRD analysis shows the higher intensity of MgO periclase (as the only phase) with higher calcination temperature (not shown), indicating a better crystallinity of this phase as previously reported [54].
However, calcination temperature has a noticeable effect on the formation of LDOs containing calcium from carbonate or oxalate interlayer anions for all the other Mg/Ca/Al ratios as presented, for example, for 1/2/1 CO3.pH8 (Fig. 9a) and 1/2/1 Ox.pH10 (Fig. 9b). The modification of the sorbents structure is due to the decomposition of a magnesium calcium carbonate containing large part of calcium for the benefit of the formation of CaO and is expected by TGA results of LDHs (section 'Mg/Ca Ratio Effect'). For LDOs  X-ray diffractograms of LDOs samples prepared from a 1/2/1 CO3·pH8 , and b 1/2/1 Ox·pH10 LDH according to calcination temperature containing calcium from stearate interlayer anion and for all the other Mg/Ca/Al studied ratios, the increase in calcination temperature from 500 to 700 °C does not affect the structure (XRD not shown), because of the complete decomposition of stearate in oxide at a temperature below 500 °C as presented by TGA (Fig. 6b).

BET and BJH Methods: N 2 Adsorption
The LDOs samples present type IVa isotherms [55] associated with mesoporous adsorbents (Fig. 10). CO 2 capture capacity depends on the number of CaO sites and their accessibility. Therefore, to optimize the sorption efficiency of calcium, the porosity properties (specific surface area and pore size) of sorbents must be improved [28].

Anion and pH Effects
In the adsorption/desorption curves (Fig. 10), a difference appears for stearate samples that present the flat line characteristic to mesoporous materials, whereas for carbonate and oxalate samples, adsorption curves do not end up with a plateau and have a nearly vertical segment revealing the presence of some macro-porosity. This reduces the calculations significance of average pore size [55].
Similar results (Fig. 11 left) of BET surface area (SSA) were observed for the samples prepared at pH 10 with various interlayer anions (214-242 m 2 /g) but at pH 8 stearate sample presented the highest SSA value (350 m 2 /g) while oxalate sample SSA was deteriorated (117 m 2 /g). However, for carbonate and oxalate samples, the Pv (0.8-1.0 cm 3 /g) is twice that of stearate samples (0.4-0.5 cm 3 /g) whatever the pH of synthesis (Fig. 11 right).
The greatest SSA is obtained at pH 8 with sodium stearate (350 m 2 /g) but with a middle Pv value compared to the highest Pv at pH 10 with sodium oxalate (1.00 cm 3 /g). Depending on interlayer anion, mesoporosity is associated with the highest SSA value for samples from stearate, and macroporosity is correlated to the greatest Pv values for samples from carbonate and oxalate. So, whatever the pH used, the best anions seem to be carbonate and oxalate for CO 2 sorption application. For stearate, the most appropriate pH (pH 8) is kept for the following studies.

Mg/Ca Ratio and pH Effects
All carbonate samples (Fig. 12a) seem to present some macroporosity because of their nearly vertical segment. If oxalate samples with or without calcium (not shown), and stearate samples (Fig. 12b) with calcium have the same behavior, a very different adsorption curve is observed for 3/0/1 St·pH8 sample with a flat line characteristic to mesoporous materials. Then, it seems to be confirmed that the best pH value for sorbents from stearate is the lowest one.
When Mg/Ca ratio decreases, carbonate samples SSA and Pv progressively fall (from 214-242 to 30 m 2 /g and from 0.8-1.0 to less than 0.2 cm 3 /g, respectively) whatever the pH value (Fig. 13). If oxalate samples give comparable SSA results as carbonate LDO whatever Ca content (Fig. 13), their Pv is greater especially for pH 10 (0.2-0.8 cm 3 /g compared to 0.4-1.0 cm 3 /g for various Mg/Ca ratios). The decrease of pH appears unfavorable for oxalate compounds.
Stearate samples containing Ca are not interesting from the point of view of their SSAs which match the values observed with the other anions at the same compositions (106-143 m 2 /g). The same behavior is observed for pore volume (Pv) values that reach similar or lower levels as corresponding carbonate and oxalate containing Ca compounds (Fig. 13). For example, 2/1/1 St.pH8 presents a very similar value to all the interlayer anions (between 0.4 and 0.5 cm 3 /g). So, N 2 adsorption analysis results attested to the increasing of SSA by BET method with anion size (from carbonate to stearate) as previously reported [29], but only for samples without Ca. Stearate is an overweight anion and could then collapse the layered structure during calcination, but in presence of Ca, this phenomenon does not appear when samples are calcined at 500 °C. The presence of magnesium calcium carbonate phase (Mg 0.03 Ca 0.97 CO 3 ) observed by XRD for all Ca-containing samples keeps the strength of the structure until 550 °C. However, this phase disappears before 700 °C for the benefit of oxide phases (CaO and MgO). Thus, the effect of calcination temperature up to 700 °C has to be studied to evaluate samples after the decarbonation phase. Because of forthcoming calcium addition, the anion and pH parameters can be fixed to carbonate or oxalate and pH 10, respectively.

Calcination Temperature Effect
The temperature increasing from 500 to 700 °C generally decreases SSA for carbonate and oxalate samples for all Mg/Ca ratios (Fig. 14).
Pv follows the same decreasing tendency for oxalate samples, with an increase in temperature. However, an opposite behavior is observed for carbonate samples for which Pv increases with calcination temperature, regardless of the composition. This difference of behavior between carbonate and oxalate samples can be explained as follow: calcium carbonate decomposes into calcium oxide between 500 and 700 °C, leading to higher pore volume left by calcium carbonate elimination, calcium carbonate volume being higher than calcium oxide one. In oxalate samples, a spinel phase (observed by XRD) after calcination at 500 °C could sinter during calcination between 500 and 700 °C and lead more easily to sorbent collapsing.

SEM
SEM results can be correlated to the porosity trends obtained by N 2 adsorption associated with BET and BJH methods, respectively, for all the studied parameters (anion, pH, and Mg/Ca ratio).
Indeed, for the samples from carbonate, the morphology of layers appeared in the typical form of sand roses [28] which were also just as apparent at the lower pH value ( Fig. 15a and b). However, for pH 8, sand roses morphology was most evident in the absence of Ca (Fig. 15b) and occasionally when Mg/Ca decreases (not shown). For pH 10, if the morphology was maintained by decreasing the Mg/Ca, the thickness of the floral pattern was reduced (Fig. 15c).  (Fig. 15a). Here again, the finest sand roses were observed for the lower Mg/Ca ratio. Vice versa, the floral pattern disappeared for pH 8 and SEM analysis showed stone-like morphology without apparent layer and porosity (Fig. 15b) [27].
Finally, for stearate samples, sand roses were only present in the absence of Ca at pH 8 (Fig. 15b) and in dense form. At pH 10, stearate sorbents looked bulkier.
Then in all cases, the apparent porosities and the presence of sand roses observed by SEM were fully correlated to the highest SSA and Pv values obtained by N 2 adsorption (Fig. 13) as previously reported [27] but the finesse of sand roses was linked to the lowest SSA and Pv values (Fig. 13).

Sorption Capacity
The sorption capacity (in g CO2 /100 g sorbent ) was evaluated for several samples ( Table 2). Considering that not CO 2 sorption is possible on MgO sites at 600 °C, the theoretical value of this property (indicated below each Mg/Ca/Al A global increase in CO 2 sorption capacity with calcium content rise was observed for each anion at the same pH, despite the decrease of porosity previously observed by N 2 physisorption analysis. However, the upper limit of sorption capacity should be given by the higher Ca content (0/3/1 CO3·pH10 ), but its sorption capacity (32.8 g CO2 /100 g sorbent ) is lower than that of 1/2/1CO 3·pH10 (37.2 g CO2 /100 g sorbent ). This behavior can be explained by the absence of Ca dispersant (Mg) and lower accessibility of CaO by CO 2 . On the other side, 0/3/1CO 3·pH10 also exhibited a lower sorption capacity than pure CaO (49.3 g CO2 /100 g sorbent ) that cannot be justified in the same way, but by Ca involvement in another phase as calcium aluminate or the low accessibility due to low surface area of this sample, associated to its sintering during calcination (Fig. 13).
The accessibility of CaO sites (given by the sorption capacity in g CO2 /100 g CaO in the sample) permits the comparison between sorbents with various compositions (Table 2). These values can be compared both to the theoretical sorption capacity (78.4 g CO2 /100 g CaO ) and to the experimental one for pure CaO (49.3 g CO2 /100 g CaO ). The improvement of this property when CaO is dispersed in LDOs from 2/1/1 Mg/Ca/Al composition, starting from carbonate and pH 10 (67.6 g CO2 /100 g CaO ) is highlighted. The maximum value of 74.4 g CO2 /100 g CaO , very close to the theoretical one, can be reached by the LDOs originated from oxalate precursors.
For the most of Mg/Ca/Al ratio, the effect of pH decrease is a strong loss of sorption capacity proved with carbonate for 2/1/1 and 1.5/1.5/1, with oxalate for 2/1/1, and with stearate for 2/1/1 and 1.5/1.5/1. An exception was revealed for 1/2/1 CO3·pH8 which reached the highest sorption capacity (39.2 g CO2 /100 g sorbent ). However, the comparison of accessibility of CaO sites revealed the highest value for 2/1/1 Ox·pH10 and more generally, for all samples prepared at pH 10.
An optimal Mg/Ca/Al ratio of 1/2/1 could be found for carbonate at pH 8 which possessed both high sorption capacity (39.2 g CO2 /100 g sorbent ) and CaO sites accessibility (71.1 g CO2 /100 g CaO ) to avail on best its CO 2 sorption capacity. Besides, the CaO sites accessibility of 1.5/1.5/1 CO3·pH10 was (48.1 g CO2 /100 g CaO ) very close to pure CaO (49.3 g CO2 /100 g CaO ), while the corresponding oxalate sorbent (1.5/1.5/1 Ox·pH10 ) led to 69.7 g CO2 /100 g CaO one of the most appreciable values. Oxalate use also allowed obtaining the highest accessibility of CaO sites (74.4 g CO2 /100 g CaO ) for 2/1/1 Mg/Ca/Al composition and pH 10. Finally, stearate samples always had the lowest CO 2 sorption capacities and accessibilities CaO sites whatever the composition or the pH.
To conclude, the accessibility of CaO sites and the CO 2 sorption capacity are directly linked to the interlayer anion, pH, and calcium content, but not so clearly related to the porosity properties measured by N 2 adsorption.
As previously mentioned (section Sorption Tests), the values of sorption capacity differed from those obtained by the sorption test (section 'Sorption Capacity') because of the different conditions used: weight of samples, type of solid/gas contact, and carbonation state of the samples.
Absorption/desorption cycles on calcium oxide sorbents are well known to reduce CO 2 sorption capacity because of the formation of calcium carbonate (bulkier than CaO) and its possible sintering [5].
Sorbents from carbonate presented much worse stability than sorbents from oxalate. Their sorption capacity decreased very rapidly for the first cycles, still relevant until the 9th cycle reaching a loss of at least 50%, stabilized for the next 10 cycles (Fig. 16).
If an important sorption capacity difference can be noticed for the 1st cycle between 1/2/1 CO3·pH10 and 1/2/1 CO3·pH8 samples, a slight influence of pH on its stability was observed for these carbonate samples. Indeed, 1/2/1 CO3·pH10 had a higher initial sorption capacity, but this value is decreased at the same rate than 1/2/1 CO3·pH8 and reached the same level beyond 10 cycles.
Oxalate samples presented lower loss of sorption capacity (25% and 16% for 1/2/1 Ox·pH10 and 1.5/1.5/1 Ox·pH10 , respectively) than carbonate samples after 7 cycles depending on their composition: the higher the Ca content (1/2/1 Ox·pH10 ), the higher the loss of sorption capacity (Fig. 16) because of the sintering of calcium carbonate. Furthermore, this loss started to stabilize as soon as the 6th cycle.
Then, the most influential parameter on sorption capacity appeared to be the interlayer anion (oxalate versus

Conclusions
LDHs decomposition was evaluated by TGA analysis under air to understand the formation of the corresponding LDOs as a function of interlayer anion, pH of synthesis, and Ca content, but also calcination temperature. It appeared that the decomposition behavior depends on anion choice (carbonate, oxalate, or stearate) and takes place for everyone between 200 and 500 °C, if samples are free of calcium. As far as LDOs properties are concerned, the same parameters (interlayer anion, pH, Ca content, and calcination temperature) were evaluated thanks to XRD, N 2 adsorption, and SEM. It appeared different behaviors for Ca-containing samples and Ca-free samples.
For LDOs without Ca, XRD revealed that the periclase (MgO) phase is always the main phase, whatever the interlayer anion and pH value, except for the sample from oxalate and pH 8 for which a spinel phase (MgAl 2 O 4 ) was also observed. N 2 adsorption indicated different types of porosity versus interlayer anions: mesoporosity for samples from stearate and macroporosity for samples from carbonate and oxalate, associated with the highest SSA value and the greatest Pv values, respectively. This difference of behavior is also observed by SEM with the presence of sand roses for all anions produced at pH 8 for stearate anion and pH 10 for oxalate one.
For LDOs with various Ca contents, XRD indicated the formation of an Mg 1−x Ca x CO 3 phase with the MgO phase after calcination at 500 °C that decomposed in the CaO phase after calcination at 700 °C, for all the anions studied. A macroporosity was observed for all the samples and associated with a decrease in SSA and Pv values when Ca content rose, the best values being associated with samples from oxalate and pH 10. The presence of sand roses morphology observed by SEM confirmed the tendencies obtained by N 2 adsorption.
For CO 2 sorption capacity results, it seemed important to evaluate both sorption capacity (g CO2 /100 g sorbent ) and accessibility of CaO sites (g CO2 /100 g CaO ). Thus, the best results were obtained for 1.5/1.5/1 Ox·pH10 (31.1 g CO2 /100 g sorbent and 69.7 g CO2 /100 g CaO ), 1/2/1 Ox·pH10 (31.2 g CO2 /100 g sorbent and 56.6 g CO2 /100 g CaO ), 1/2/1 CO3·pH10 (37.2 g CO2 /100 g sorbent and 67.5 g CO2 /100 g CaO ) and finally for 1/2/1 CO3·pH8 (39.2 g CO2 /100 g sorbent and 71.1 g CO2 /100 g CaO ). The stability of these samples was then evaluated in cyclic adsorption/desorption (carbonation/calcination) test and gave the advantage to the oxalate samples and, without surprise, to the highest Mg/Ca ratio. Despite the samples have not been evaluated for a high number of cycles, the stability of absorption/desorption cycles is a promising signal for associating them to a metallic phase (Ni or Fe) to produce bi-functional catalyst and sorbent materials for Sorption Enhanced Steam Reforming processes (SESR).