Understanding the Role of Mono and Ternary Alkali Metal Salts on CO2 Uptake of MgO Sorbents

CO2 uptake by MgO-based sorbents at intermediate temperatures is attractive for pre- and post-combustion CO2 capture applications. However, besides the high CO2 uptake potential of these materials (1.1 g CO2 g−1 sorbent), in practice, the realistic CO2 capture is far from that of the theorical values. In this work, the sol–gel method was used to synthetize unsupported and supported MgO sorbents (10% Ca− or 10% Ce− support, mol) that were impregnated with different fractions (15, 25, and 35; % mol) of a NaNO3 single salt or a ternary alkali salt (NaNO3, LiNO3 and KNO3 (18/30/52; % mol)). To understand the role of alkali metal salts (AMSs) in the MgO sorbents’ performance, the working and decomposition temperature ranges of AMS under different atmospheres (CO2 and air) were evaluated. The findings show that the CO2 uptake temperature range and maximum uptake (20–500 °C, CO2 atmosphere) of sorbents are correlated. The cyclic CO2 uptake of the most promising sorbents was tested along five carbonation–calcination cycles. For the first and fifth cycles, respectively, the 15 (Na, K, Li)-MgO sorbents showed the highest carrying capacity, i.e., 460–330 mg CO2 g−1 sorbent, while for the 15 (Na, K, Li)-MgO-Ca sorbents, it was 375–275 mg CO2 g−1. However, after the first cycle, the carbonation occurred faster for the 15 (Na, K, Li)-MgO-Ca sorbents, meaning that it can be a path to overpassing carbonation kinetics limitations of the MgO sorbent, making it viable for industrial applications.


Introduction
MgO is an intermediate-temperature CO 2 sorbent (200-400 • C) that gained relevance in the scientific community during the last decade [1,2].This sorbent is suitable for both post-and pre-combustion CO 2 capture technologies.Most of the studies in the literature have been performed for post-combustion conditions [3,4], but the application in sorptionenhanced water gas shift (SEWGS) reactions [5,6], i.e., under pre-combustion conditions, is also relevant since it increases the H 2 production and purity.
CO 2 capture is based on the following reversible chemical reaction: MgO (s) + CO 2 (g) MgCO 3 (s), and since one mole of MgO can absorb one mole of CO 2 , its theoretical CO 2 capture capacity is high (1.1 g CO 2 g −1 MgO) [7].The calcination and carbonation reactions are both dependent on temperature, and its equilibrium is also related to the CO 2 partial pressure according to Equation (1) [8]: T eq = 13636 ln [bar] P CO 2 + 20.01 Calcination is an endothermic reaction (∆H 25 • C = +116.9kJ mol −1 ) that, ideally, should be performed in an atmosphere rich in CO 2 (>450 • C) to allow for a high CO 2 concentration at the exit of the calciner with an efficient compression and storage for the utilization of the captured CO 2 .Some advantages of this sorbent, which is available in nature as MgCO 3 or MgO, include its characteristics of being nontoxic and noncorrosive and its availability at a relatively low cost.Moreover, as mentioned above, its regeneration can occur at around 450-500 • C; then, the energy consumption during this step is reduced in comparison with that of CaO or Li 4 SiO 4 sorbents' regeneration that needs temperatures higher than 700 • C [5].
In addition to the interesting advantages of MgO as a CO 2 sorbent, MgO has a very low experimental CO 2 uptake capacity of 11-20 mg CO 2 g −1 and very poor experimental sorption kinetics at a very favorable temperature of 200 • C from a thermodynamic point of view [9].Some explanations have been proposed to justify the low performance of MgO [7]: (i) the low surface area does not allow for sufficient exposure of its basic sites for CO 2 carbonation; (ii) the low porosity can obstruct the CO 2 diffusion through the pores and delay the carbonation equilibrium; (iii) the MgO has a volume expansion of 2.49 due to the formation of MgCO 3 , while the formation of dense layers adjacent to basic active sites of the MgO sorbent can inhibit the CO 2 carbonation; (iv) the MgO has an intrinsically high lattice enthalpy.
For an upscale application of MgO sorbents, the kinetic and carrying capacity limitations need to be overcome.In this sense, different paths to enhance the MgO sorbents performance, based on their dependence on intrinsic and extrinsic factors, have been assessed.To upgrade the internal properties of sorbents, the effects of distinct factors and their synergies have been evaluated: different MgO precursors, the synthesis of mesoporous MgO, dispersion on inert supports, and doping with alkali metal salts (AMSs).Factors like the CO 2 concentration, pressure, presence of water, and presence of gas impurities are extrinsic factors that can also influence CO 2 uptake by sorbents.
The most common magnesium precursors include inorganic salts, organometallic salts, and natural minerals.In general, the organometallic salts of magnesium are better precursors than the other two types in producing efficient MgO sorbents.This is mainly because the former usually have a larger molecular weight, which results in MgO with a better pore size and porous structure [2].However, the organometallic precursors are expensive.On the contrary, natural Mg-containing minerals are abundant in nature and, thus, are costeffective.For instance, magnesium oxalate (MgC 2 O 4 ) is a better precursor than MgCO 3 because it generates MgO sorbents with better performance [3].Guo et al. [3] obtained a MgO sorbent derived from magnesium oxalate dihydrate (MgC 2 O 4 • 2H 2 O) with the highest CO 2 capture capacity of 194 mg CO 2 g −1 sorbent.This sorbent is characterized by excellent textural properties, a uniform surface, and abundant basic sites.Hanif et al. [10] synthesized mesoporous MgO with a BET surface area higher than 350 m 2 g −1 using the ammoniumhydroxide-assisted precipitation of Mg(OH) 2 from a Mg(NO 3 ) 2 aqueous solution followed by the thermal degradation of the precipitated Mg(OH) 2 in vacuum.This sample presented a CO 2 capture capacity of 75 mg CO 2 g −1 sorbent at 300 • C. The synthesis of mesoporous MgO is also an effective way to improve sorbents' CO 2 capture capacity.When the CO 2 reacts with MgO, it forms a covering layer of the thermodynamically stable MgCO 3 on the unreacted sorbents' surface, which delays the CO 2 molecules' diffusion through the product layer.Thus, the production of an MgO sorbent with a high surface area and porosity should facilitate the CO 2 diffusion, enhancing the capture capacity and kinetics reactivity of the sorbent [2].The dispersion of MgO on a porous support also increases the surface area availability of wellexposed basic active sites, which in turn improves the sorbent's capture capacity.Ideally, a good support must present a stable porous structure, which should enhance the sorbent performance and reduce its sintering after several carbonation-calcination cycles, contributing to the stability of the sorbent's CO 2 uptake capacity.
Among the approaches to improve the CO 2 uptake properties of MgO, the doping with alkali metal salts is the most widely recognized promising approach.The aim of the most recent experimental works is to improve the CO 2 uptake capacity of these materials up to 0.7-0.8g CO 2 g −1 sorbent [11].The main categories of alkali doping described in the literature [2] are the alkali carbonate doping, alkali nitrate/nitrite doping and binary or ternary alkali doping.
A US patent [12] described that alkali metal carbonate-doped MgO sorbents prepared by coprecipitation registered CO 2 capture capacities of 48-570 mg CO 2 g −1 sorbent at the temperature range of 350 to 400 • C and under 70% dry CO 2 .Moreover, it was observed that the performance of the doped sorbent was influenced by the doping precursor.For instance, the MgO sorbents doped with Na 2 CO 3 showed better performance than when doped with Li 2 CO 3 or K 2 CO 3 .
Zhang et al. [13] doped MgO sorbents with NaNO 3 .The authors achieve a good CO 2 carbonation kinetics and a MgO conversion of 75% against only 2% for an undoped MgO, both at 330 • C and ambient pressure.It was stated that molten NaNO 3 decreases the dissociation energy of Mg-O ionic bonds in bulk MgO; thus, the molten NaNO 3 acts as a phase transfer catalyst (PTC) between bulk MgO and CO 2 molecules which, in turn, facilitates the carbonation reaction.In addition, molten alkali metal nitrates prevent the formation of a rigid, CO 2 impermeable, and monodentate carbonate layer on the surface of MgO as it occurs with bare MgO and promote the rapid generation of carbonate ions to allow a high rate of CO 2 uptake [2].
Concerning to binary or ternary alkali doping, very promising results are found in the literature.Lee et al. [14] studied the Na 2 CO 3 /NaNO 3 -doped MgO sorbent that maintained a CO 2 sorption capacity of 153 mg CO 2 g −1 sorbent after seven cycles.The authors related this performance with the interaction established between MgO, CO 2 and the molten NaNO 3 that results in the formation of the double salt Na 2 Mg(CO 3 ) 2 .Identical results were registered for other cases of binary doping, such as K 2 CO 3 /KNO 3 -doped MgO sorbents.Zhang et al. [15] provided a detailed explanation for the enhancement of the CO 2 uptake capacity of MgO sorbents doped with the binary Na 2 CO 3 /NaNO 3 system.The mechanism is stated to be similar to that of the single alkali nitrate.The bulk MgO dissolves in the molten salt because Mg-O ionic bonds are easy to break.Na 2 CO 3 also dissolves in the same liquid medium because carbonate salts have good solubility in molten salts.Hence, ion pairs of Mg 2+ , O 2− and CO 3 2− are formed and react with the CO 2 molecules, generating the double salt (Na 2 Mg(CO 3 ) 2 ).The same mechanism is appropriate to binary alkali nitrate/carbonate doping, which forms the CaMg(CO 3 ) 2 double salt [2].
The binary doping with alkali nitrate/nitrite is also an interesting matter of study.Zhao et al. [16] compared the CO 2 sorption capacities of the single NaNO 3 and of the binary NaNO 3 /NaNO 2 -doped MgO sorbents.The latter showed higher CO 2 sorption capacity than the former.This new evidence was explained by the reduction in the melting temperature of the eutectic mixture.While single NaNO 3 and NaNO 2 present a theoretical melting point of 308 • C and 271 • C, respectively, the eutectic mixture of NaNO 3 /NaNO 2 exhibits a melting temperature of 185 • C. Thus, the eutectic mixture facilitates the carbonation process by providing a molten phase that works like a liquid channel.Ternary doping with NaNO 3 , LiNO 3 and KNO 3 (18/30/52; % mol) registered a sharper reduction in the eutectic mixture's melting point (120 • C) and an enhanced CO 2 sorption performance [2].
Summarizing, the doping with AMS improves the cyclic CO 2 sorption capacity of MgO sorbents when compared to that of commercial MgO, of mesoporous MgO and of MgO obtained from different precursors.Nevertheless, most of the AMS-doped MgO sorbents exhibit a CO 2 uptake capacity of less than 50% of the maximum theoretical capacity.Table 1 shows different studies found in the literature using unsupported and supported MgO-based sorbents doped with AMS and the corresponding CO 2 uptake capacity after multicyclic tests.
As mentioned above, the existence of a molten state generated by the melting of alkali metal salts is essential to increase the sorbent's CO 2 uptake and to improve the kinetic of MgO carbonation.In this work, the role of the NaNO 3 single salt and the amount of a ternary alkali salt (NaNO 3 , LiNO 3 and KNO 3 (18/30/52; % mol)) on the CO 2 uptake temperature range, as well as on the CO 2 maximum uptake, will be evaluated for unsupported and supported sorbents.To evaluate the calcination atmosphere effect on the working temperature range of the alkali salts, its decomposition was studied under air and CO 2 atmosphere.As innovative work in relation to the experiments shown in Table 1, the sorbent calcination will be performed under a CO 2 atmosphere to obtain a concentrated CO 2 stream useful for storage or utilization, which is more interesting for industrial applications.x-molar fraction (%) of AMS, Y-molar fraction (%) of support.

MgO-Based Sorbents Synthesis and Doping with Alkali Metal Salts
The Mg(NO 3 ) 2 •6H 2 O was dissolved in distilled water with a molar ratio of citric acid and distilled water to magnesium of 1:1 and 120:1, respectively.For the supported MgO sorbents, the same procedure was used but adding Ca(NO 3 ) 2 •4H 2 O or Ce(NO 3 ) 3 •6H 2 O.
As shown in Figure 1, each solution was continuously stirred at 80 • C for 6 h.Afterward, the wet gel was dried in the oven at 120 • C for 14 h, milled, and the sample obtained was calcined in the muffle by ramping the temperature to 500 • C at a rate of 2 • C min −1 plus 2 h more at 500 • C. The synthesized MgO-based sorbents were doped with mono AMS (NaNO 3 ) and with a ternary AMS mixture (NaNO 3 , LiNO 3 and KNO 3 (18/30/52; mol.%)) [11,22] using the wet-impregnation method.Prior to the impregnation, the AMS was dried in the oven at 120 • C for 24 h.First, the MgO-based sorbent and AMS have been dissolved in 20 mL of distillate water, using magnetic stirring for 1 h at room temperature.The obtained aqueous slurry was dried in the oven at 120 • C for 14 h.Afterward, the dried sample was placed in the muffle for calcination using a heating rate of 2 • C min −1 until it reached 450 • C, plus 4 h at 450 • C. Table 2 summarizes all the prepared AMS-MgO-based sorbents and the different ratios of AMS in each sorbent.The following nomenclature was used for the doped sorbent samples: x (Na or Na-Li-K)-MgO-Z, where x corresponds to the molar fraction (%) of AMS, and Z corresponds to the Ca or Ce support.

Characterization Methods
A D8 Advance diffractometer (Bruker AXS GmbH, Karlsruhe, Germany) using Cu Kα (λ = 0.15406) radiation and a Ni filter was used to identify the crystalline phases of the synthetized sorbents before and after the cyclic experiments on a thermogravimetric analyzer (Setsys Evolution TGA, Setaram Instruments, Caluire, France).The equipment was operated at 30 mA and 40 kV, scanned within the 2θ range of 5-80 • , and had a step size (scanning duration) of 0.03 • (0.5 s).The crystallography open database (COD) was used to identify the crystalline phases.The MgO average crystallite size of the sorbents was estimated using Scherrer's equation ( D = Kλ/bcos θ), based on the XRD data, where D is the crystallite size (nm), b corresponds to the full width at half maximum (FWHM) of the XRD peak considered, λ is the wavelength (0.15406 nm), θ is the Bragg angle (degree) and K is Scherrer's constant (K = 0.9, assuming that particles are spherical).
The specific surface area (S BET ) of the synthetized sorbents based on the BET method and the total pore volume (V p ), were determined by the N 2 sorption technique at −196 • C, (Quantachrome Instruments, Model autosorb IQ, Graz, Austria) at a relative pressure (p/p 0 ) of 0.95.The degasification procedure was performed in two steps: the first one at 90 • C for 1 h and the second one at 350 • C for 5 h.

Thermal Decomposition of Mono and Ternary Alkali Nitrates Salts
The thermal decomposition of the molten salts was assessed for the pure NaNO 3 and for the ternary mixture of 52% (mol) KNO 3 , 18% (mol) NaNO 3 and 30% (mol) LiNO 3 by thermogravimetric analysis with air and pure CO 2 atmospheres between 25 and 1000 • C at a heating rate of 10 • C min −1 .

Effect of Mono and Ternary AMS Molar Fraction and Inert Support Addition
The effect of mono or ternary alkali metal salts doping, as well as the best molar fraction of AMS-MgO were evaluated in a thermogravimetric analyzer (TGA) system TG-DSC Setsys Evo 16.The TGA studies were carried out with MgO-based sorbents for the range of temperature from 20 to 500 • C and a temperature ramp of 10 • C min −1 .During the experiments, a constant flow of 80 mL min −1 of CO 2 and ~10 mg of sample were used.
The best two mono and ternary AMS samples, i.e., with higher CO 2 uptake capacity, were identified and selected to proceed to the tests with five carbonation-calcination cycles in the TGA as described below.Then, using the same procedure, the effect of the addition of an inert support to the MgO sorbent was tested with Ca and Ce additions (10%, mol), considering the most promising AMS doping and the respective molar fraction.

Carbonation-Calcination Cycles
A TGA system TG-DSC Setsys Evo 16 was used to carry out the studies for the assessment of both the cyclic stability and the CO 2 uptake of the supported and unsupported MgO sorbents.These tests consisted of five carbonation-calcination cycles with carbonation and calcination temperatures of 300 • C (60 min) and 445 • C (60 min), respectively, using a constant flow of 80 mL min −1 of CO 2 , a ramp temperature of 10 • C min −1 and ~10 mg of sample.A blank experiment was performed to correct apparatus buoyancy effects like atmosphere density changes with temperature.The unsupported sample showing the best performance was identified and tested through 12 carbonation-calcination cycles using the same experimental conditions.
The TGA mass variation allows determining the mass of CO 2 captured by the sorbent during the carbonation step (Equation ( 2)), and subsequently, the carrying capacity and the MgO conversion to carbonate can be obtained using Equations ( 3), ( 4) and ( 5), respectively.
where m CO2captured is the mass of CO 2 captured in each cycle, m carbonation is the mass of the sample after the carbonation reaction, and m calcination is the mass of the sample after the calcination reaction.The carrying capacity (CC) of the sorbent is defined as the mass of CO 2 captured per mass unit of the sorbent (mg CO 2 g −1 sorbent).
where m sorbent is the mass of the sorbent.The theoretical carrying capacity (CC theor.) is calculated with Equation (4), where M CO 2 and M MgO represent the molar mass of CO 2 and MgO (g mol −1 ), respectively.The MgO conversion of the sorbent is defined as the percentage of the CC when compared to the CC theor.
where w MgO sorbent represents the nominal mass fraction of the MgO (%, wt.) in the sorbent.

Properties of MgO-Based Sorbents
The crystalline structures of the MgO-SG sorbent impregnated with 15, 25 and 35% of either NaNO 3 (Figure 2a) or a ternary mixture of KNO 3 , NaNO 3 and LiNO 3 (Figure 2b) were analyzed by powder XRD.All the XRD-obtained patterns show the characteristic peaks of the MgO (42.9 • , 62.2 • , 78.5 • ) pattern and a main peak located at ~29.4 • that is referred to the NaNO 3 pattern.For the case of the sorbents impregnated with the ternary alkali metals salts, the main peak of KNO 3 (~23.6• ) was also observed, and the second main peak should be overlaid by the NaNO 3 peak (29.4 • ).The LiNO 3 was not identified, which can be justified by its high dispersion on the sorbent sample, leading to the Li concentration below the instrumental detection limit [23] or its partial decomposition to Li 2 O 2 or LiO 2 whose main peaks are around 32.9 • and 33.5 • .Wang et al. [24] studied a similar ternary system, LiNO 3 (25.9mol.%),NaNO 3 (20.06mol.%) and KNO 3 (54.1 mol.%), and concluded that unlike the sodium and potassium nitrate phases, lithium nitrate is an unstable phase in the ternary system that starts to decompose to Li 2 O 2 and LiO 2 at temperatures higher than 440 • C. The MgO peaks are significantly sharper in the impregnated samples, suggesting that the impregnation influences the average size of MgO crystallite.It is usually expected that the sorbents with smaller MgO crystallites have higher surface areas and an enhanced reactivity; thus, the effect of impregnation on the growth of MgO crystallites is evaluated and correlated with the sorbent's performance along the carbonation-calcination cycles in Section 3.3.For instance, the average crystallite size of the MgO-SG sample is 10 nm, while the 15 Na-MgO presents a crystallite with 15 nm, and the MgO-SG impregnated with 15% of (Na, Li, K)NO 3 achieves a crystallite size of 26 nm.Such results indicate that the impregnation with alkali metal nitrates boosts the growth of the MgO crystals, which is more pronounced in the case of the impregnation with the ternary mixture.On the other hand, the AMS molar percentage of the impregnation does not seem to influence significantly the growth of the MgO crystallite.The MgO-SG sorbent doped with NaNO 3 shows an average crystallite size of ca. 15 nm and the MgO-SG sorbent doped with (Na, Li, K)NO 3 of ca.28 nm regardless the molar percentage of AMS impregnation (Table 3).The supported sorbents, MgO-Ca-SG and MgO-Ce-SG, were doped with 15% of a ternary mixture of (Na, Li, K)NO 3 .Figure 3 shows the X-ray diffraction peaks and Table 4 compares the MgO crystallite size for the unsupported and supported sorbents before and after the AMS doping.
The 15 (Na, Li, K)-MgO, 15 (Na, Li, K)-MgO-Ca and 15 (Na, Li, K)-MgO-Ce sorbents exhibit the characteristic peaks of the MgO pattern (42.9 • , 62.2 • , 78.5 • ).Moreover, the 15 (Na, Li, K)-MgO-Ca and 15 (Na, Li, K)-MgO-Ce also present the main peaks of the XRD pattern of CaCO 3 (29.4• ) and CeO 2 (28.5 • ), respectively.Regarding the nitrate salts used in the impregnation, as justified above, the LiNO 3 was not detected by XRD.The main peaks of the KNO 3 pattern (23.5 • and 29.4 • ) were identified for the 15 (Na, Li, K)-MgO and of the 15 (Na, Li, K)-MgO-Ca as well as the main peaks of the NaNO 3 pattern (29.4 • , 31.9 • , 38.9 • , 47.9 • ).However, for the 15 (Na, Li, K)-MgO-Ce sorbent, as the characteristic peak of the CeO 2 XRD pattern is very wide, it is impossible to identify the peak at 29.4 • of the KNO 3 and NaNO 3 patterns.Still, a small peak at 27 • of the KNO 3 pattern is exhibited as well as the peaks at 31.9 • , 38.9 • , 47.9 • related to the NaNO 3 pattern.Table 4 shows that the AMS doping influenced the crystallite size of supported sorbents, causing it to increase but in a similar way for both supports.
The specific surface area (S BET ) and the pore size distribution of MgO-SG, MgO-SG-Ca and MgO-SG-Ce before and after doping with 15 of (Na, K, Li)NO 3 was evaluated (Figure 4a,b) by the N 2 sorption technique.
The MgO-SG-Ce shows the highest S BET (230 m 2 g −1 ) followed by MgO-SG (187 m 2 g −1 ) and the MgO-Ca-SG (144 m 2 g −1 ).However, after doping, all the sorbents show a severe reduction in their S BET , namely, 83%, 88% and 87% for 15 (Na, Li, K)-MgO, 15 (Na, Li, K)-MgO-Ca and 15 (Na, Li, K)-MgO-Ce sorbents, respectively.This evidence is in line with the previous conclusions from the XRD characterization, in which we observed an increase in the MgO crystallite size, indicating a reduced surface area available for the MgO sorbent to uptake CO 2 .
Dal Pozzo et al. [11] reported a S BET of 22 m 2 g −1 for a MgO sorbent promoted with 10% of (Na, Li, K)NO 3 , which is smaller than that of 15 (Na, Li, K)-MgO and 15 (Na, Li, K)-MgO-Ce sorbents but higher than that of 15 (Na, Li, K)-MgO-Ca.In addition, all the analyzed sorbents of the set also registered a decrease in the total pore volume after AMS doping.This is probably due to its partial occupation by the ternary mixture [17], which should affect the CO 2 uptake due to the higher occurrence of pores' blocking.The PSD of the MgO-SG and MgO-SG-Ce sorbents (Figure 4b) indicates a predominance of the presence of mesopores (2-50 nm).There is also a small percentage of macropores (>50 nm).The MgO-SG-Ca consists mainly of a macroporous sorbent, although the MgO-SG-Ca also exhibits a small share of mesopores.After doping with 15(Na, K, Li)NO 3 , all the sorbents are predominantly made up of macropores (>50 nm), but in case of 15 (Na, Li, K)-MgO-Ce, a small share of mesopores is still present.

Thermal Decomposition of Mono and Ternary Alkali Nitrates Salts
The mass loss (%), the first derivative of the mass change (DTG) and the heat flow profiles with temperature were obtained in the TGA, between 25 and 1000 • C (heating rate: 10 • C min −1 ) under 100% of air or 100% of CO 2 flow, for pure NaNO 3 and a ternary mixture of 52 mol.%KNO 3 , 18 mol.%NaNO 3 and 30 mol.% LiNO 3 (Figure 5a-c).The knowledge about the mono and ternary alkali nitrates thermal decomposition under different atmospheres contributes to the understanding of the AMS stability on the doped sorbents.Figure 5a shows a mass loss until 200 • C related with the release of the adsorbed water, since these salts are very hygroscopic [25].It stands out that all the samples remain almost stable up to a temperature of around 600 • C, where an abrupt loss of mass is observed regardless of the atmosphere.The atmosphere affects the extent of the salt's decomposition: under pure air atmosphere, higher decomposition temperatures (~800 • C) are achieved when compared to those achieved for the case of 100% CO 2 atmosphere (~715 • C) [26].Due to the formation of alkali carbonates with higher molar mass than the corresponding oxides, i.e., Na 2 O, K 2 O and Li 2 O/Li 2 O 2 , the observed mass loss is lower under a CO 2 atmosphere.
The alkali carbonates are very stable at high temperature, especially under CO 2 atmosphere; i.e., Na 2 CO 3 and K 2 CO 3 start to decompose slowly above 900 • C [27] and, in the case of Li 2 CO 3 , the decomposition starts at 737 • C, but only at temperatures near 900 • C is the decomposition significant [26].
The decomposition temperature of each sample can be defined as the upper stability limit of a mixture (T3): that is, the maximum temperature at which it loses 3% of the initial weight [25].The definition of the decomposition temperature allowed to set the working temperature range with the melting point and the T3 being the extremes of that interval [25].In the present study, since we identified the presence of water in all the samples, the initial weight was replaced by the weight at 200 • C. The decomposition temperature range was also determined.The results obtained are summarized in Table 5.The results show that the ternary mixture offers a larger temperature window to operate when compared with the single salt, NaNO 3 .This advantage relies mostly on that already described in the literature: the addition of the LiNO 3 , which significantly decreases the melting temperature of the mixture with relation to the sodium nitrate salt, that is, 120 • C [28] against 308 • C [11], and not in the T3, which do not vary as much between both cases.Nevertheless, in air atmosphere, the eutectic mixture presents a T3 lower than the single salt, which is associated with the presence of Li that prematurely decomposes as stated by Bauer [28], but it does not happen under the CO 2 flow, evidencing that the atmosphere type is also a relevant parameter.
The heat flow TG curves presented in Figure 5c exhibit several up and downward peaks, whether an endothermic or exothermic heat event occurs, respectively, which allow properly identifying that the temperature occurs at each event.The analysis of both the heat flow profile (Figure 5c) and corresponding DTG curves (Figure 5b) reveals the existence of corresponding peaks in the same temperature intervals for all the samples (T3 ≥ 575 • C) regardless of the atmosphere considered.In addition, it can be observed that the NaNO 3 melting (endothermic reaction) occurs around 310 • C (Figure 5c), which agrees with the literature [28].In the case of ternary salt, it was difficult to identify its melting.Due to the water vaporization, which was confirmed by the mass loss in both salts, only a small peak around 129 • C was visible for (Na, Li, K)NO 3 salt, meaning that probably the water vaporization is masking the endothermic reaction associated with the ternary salt melting.

Assessment of the Effect of Mono and Ternary AMS Molar Fraction
Temperature-programmed carbonation-calcination followed by TGA studies were carried out with samples of MgO sorbent undoped and doped with different molar fractions of mono or ternary AMS to assess the effect of the promotors ratio (AMS/MgO) on the CO 2 uptake (Figure 6a,b).
Undoped MgO-SG sorbent immediately captures CO 2 in the beginning of the TGA test, achieving its maximum CO 2 uptake at 50 • C, and gradually releases it with the increasing of the temperature from 75 to 325 • C. The corresponding CO 2 desorption peak is significantly wider than those of the MgO-doped samples, which is probably due to the type of surface of each sorbent and, in turn, with the type of species being formed during the CO 2 uptake.In contrast, the extent of CO 2 uptake is insignificant for all the doped MgO-SG sorbents at low temperatures until a certain value of temperature is reached, where it starts to uptake CO 2 significantly.According to the literature, in the case of the undoped MgO-SG sorbent, the CO 2 molecules were quickly adsorbed at the surface of the MgO particles to form an unidentate carbonate layer impermeable for gaseous reactants.On the other hand, in nitrate-promoted MgO, the molten layer dissolves the CO 2 .Since the diffusivities of the carbonate ions and the oxygen ions in MgCO 3 layers are higher in the absence of the unidentate layers, the formation of MgCO 3 occurs faster [11].This enhancement is justified by the melting of the nitrates, since the CO 2 solubility increases in the phase transition from the solid to liquid state of the nitrates.In this way, the AMS impregnation is proved to increase considerably the CO 2 uptake by the MgO sorbents.The mentioned temperature was determined as the inflection point of each CO 2 uptake curve, that is, where the corresponding second derivate changes its signal.The different threshold temperatures were obtained for each sorbent (Supplementary Materials, Figure S1), and the results are summarized in Table 6.The sharp increase in the CO 2 uptake by the doped sorbents, with the temperature, can be attributed to the physical state of the corresponding AMS [29], ranging from solid to melting.At low temperatures, the CO 2 uptake is hindered due to the formation of a solid layer of nitrates that covers the surface of the sorbent.Only with the increasing temperature until close to the melting point of the AMS, this covering solid layer starts to melt.It is called the pre-melting phenomenon, where an interfacial liquidlike film of the partially disordered surface of the promoted sorbent is formed at temperatures below the corresponding melting point, enhancing the kinetic of the CO 2 uptake [29].For instance, the melting point of the single salt NaNO 3 is 308 • C, but it suffers a solid-state transition from an ordered to a disordered rhombohedral structure at a lower temperature of 275 • C that, in turn, is believed to potentiate its pre-melting [11].The present experiments indicate that this transition occurs at even lower temperatures (234-240 • C).In addition, the MgO-SG sorbent doped with single NaNO 3 appears to have a higher threshold temperature than the corresponding MgO-SG sorbent doped with the ternary mixture.Hence, it is suggested that the eutectic ternary mixture lowers the CO 2 uptake temperature.One study in the literature [11] describes a CO 2 uptake temperature of 160 • C using the ternary eutectic mixture of (Na, Li, K)NO 3 .In this work, it was possible to reduce the temperature to around 175 • C. The difference between the inflection temperature and the temperature that corresponds to the maximum CO 2 uptake (mg CO 2 g −1 sorbent) achieved during the temperature-programmed experiment suggests that at low temperatures, the accelerating effect of the addition of alkali metal nitrates on the CO 2 uptake by MgO is reduced, which is justified by the thermodynamic equilibrium of MgO/MgCO 3 [11].
The maximum instantaneous rate of change in temperature with respect to CO 2 uptake was obtained based on the upward peak of the first derivative of each TG curve of the tested sorbents (Supplementary Materials, Figure S1).The results indicate that the maximum CO 2 uptake rate temperature does not significantly vary with the AMS impregnation percentage for both sorbents (Table 6).The samples of MgO sorbent impregnated with the ternary mixture show lower maximum conversion temperatures than those impregnated with the single salt.As the former presents lower inflection temperatures, it is also possible to associate them with a larger CO 2 uptake temperature range, which extends the carbonation region.Moreover, the highest CO 2 uptake of 292 mg CO 2 g −1 sorbent was obtained for the MgO-SG sorbent doped with 35% of NaNO 3 , which, in turn, corresponds to the highest percentage of NaNO 3 .This result suggests that the sorbents with higher percentages of NaNO 3 might favor the Mg 2+ ions diffusion on the molten NaNO 3 [11].For the ternary mixture, the highest CO 2 uptake was 239 mg CO 2 g −1 sorbent for the MgO-SG doped with 15% of AMS and decreased with the increasing AMS molar fraction.This suggests that a certain thickness of the molten layer covering the surface of the sorbent hinders the CO 2 uptake by increasing the mass transfer resistance.For the synthesized undoped MgO sorbent, there was found to be a maximum CO 2 uptake of 36 mg CO 2 g −1 sorbent, which is better than the value found in the literature (19 mg CO 2 g −1 sorbent) [11].Maximum instantaneous rate of change in temperature with respect to CO 2 uptake ( Figure 7a,b shows the variation of the MgO crystallite size (CS) of the sorbents before and after the TG tests.Summarizing, the AMS impregnation contributes to the growth of the MgO crystallites with a higher increase in the MgO crystallite size when the impregnation is carried out using the ternary mixture of KNO 3 , NaNO 3 and LiNO 3 than the single salt NaNO 3 .The variation of the MgO crystallite size ∆CS = CS dopedsorb − CS undopedsorb /CS undopedsorb × 100) of the sorbent doped with NaNO 3 varies between 50 and 60%, but it increases for values between 160 and 200% for the sorbent with higher (Na, Li, K)NO 3 content.Conversely, the samples impregnated with NaNO 3 have a greater increase in the crystallite size before and after the TGA tests, i.e., between 50 and 60%, and 7% to 15%, for NaNO 3 and (Na, Li, K)NO 3 , respectively.The AMS molar percentage of the impregnation does not affect the crystallite size regardless of using the single salt or the ternary mixture.
The CO 2 uptake and the cyclic stability of the most promising doped sorbents identified above for both AMS doping, i.e., 25% and 35% for NaNO 3 and 15% and 25% of (Na, Li, K)NO 3 , were tested in a TGA by performing five carbonation-calcination cycles.Figure 8a,b show the cyclic CO 2 carbonation-calcination profiles obtained in the TGA apparatus and Figure 9a,b show the respective carrying capacity and MgO conversion obtained for the MgO-SG sorbent impregnated with the single salt NaNO 3 (25 and 35%, mol) and with the ternary mixture of (Na, Li, K)NO 3 (15 and 25%, mol).Both MgO-SG sorbents impregnated with 35% of NaNO 3 and with 25% of NaNO 3 exhibit a sharp decay on the CO 2 uptake in the 2nd cycle followed by a gradual reduction until the 5th cycle.The 25 Na-MgO registered an MgO conversion of 55% in the 1st cycle, of 35% in the 2nd cycle and of only 7% in the last cycle.The 35 Na-MgO achieved even higher reductions with an MgO conversion of 78% in the 1st cycle, 10% in the 2nd cycle and only 2% in the 5th cycle.A qualitative comparison of Figure 8a,b immediately reveals that the MgO-SG samples impregnated with the ternary mixture of (Na, Li, K)NO 3 have a better performance than those impregnated with the single salt NaNO 3 .Regardless of the molar percentage of ternary impregnation used, the two tested samples exhibited consistent peaks after five carbonation-calcination cycles, showing good CO 2 uptakes and cyclic stability.The 15 (Na, Li, K)-MgO reached a MgO conversion of 58% and of 42% in the 1st and 5th cycles, respectively.Meanwhile, the 25 (Na, Li, K)-MgO presented a MgO conversion of 64% in the 1st cycle and of 37% in the last cycle.Although the 25 (Na, Li, K)-MgO has the highest MgO conversion in the 1st cycle, it shows the lowest MgO conversion in the last cycle.Hence, between all the studied sorbents, the 15 (Na, Li, K)-MgO sorbent was considered the most promising since the MgO conversion values remained stable over five carbonation-calcination cycles.Subsequently, a TG test was carried out with 12 carbonation-calcination cycles using the 15 (Na, Li, K)-MgO sorbent (Figure 10a).Figure 10b plots the carrying capacity and the MgO conversion of the cyclic CO 2 carbonation-calcination test of the 15 (Na, Li, K)-MgO sorbent.The literature [11] reports a study with a MgO sorbent doped with 10% of (Na, Li, K)NO 3 that registered carrying capacities of 480 and 410 mg CO 2 g −1 sorbent for the 1st and 5th cycles, respectively.Another article describes the carrying capacities of 440 and 352 mg CO 2 g −1 sorbent for the 1st and 5th cycles, respectively, of a MgO sorbent impregnated with 15% of the same ternary mixture [19].Since the 15 (Na, Li, K)-MgO sorbent synthesized in this work presents a carrying capacity between 460 and 330 mg CO 2 g −1 sorbent from the 1st to the 5th cycle, it is clear that these results are in agreement with the results found in the literature.Dal Pozzo et al. [11] justify the MgO-based sorbents' deactivation along cycles with the nitrate's segregation.As shown in Figure 5a,b, the NaNO 3 and (Na, Li, K)NO 3 decomposition only starts above 600 • C, so these sorbents remain stable under the temperatures used for the above carbonation-calcination tests.However, since the AMS melting starts at lower temperatures (Table 5 and Figure 5c), the segregation of AMS can occur as it was demonstrated by SEM images for NaNO 3 [11] and in our previous work (unpublished), where the segregation of K particles was also demonstrated.
Due to the AMS segregation, the dispersion of MgO in the melted phase of the nitrates will be spoiled, and the CO 2 carrying capacity decreases.Apparently, under the used conditions, since the CO 2 uptake during the 1st carbonation is more effective and faster (the CO 2 uptake temperature range is lower) for the NaNO 3 salt, the nitrates segregation is more pronounced, and the sorbent's deactivation after the 1st cycle is abrupt.
In general, the 15 (Na, Li, K)-MgO sorbent exhibited a stable carbonation-calcination profile over the 12 cycles.Its MgO conversion dropped from 58% to 34% from the 1st to the 2nd cycle, but it gradually increased up to 48% until the 6th cycle, from where it progressively decreased until the last cycle, reaching a final MgO conversion of 33%.The decrease in MgO conversion is more significant after the 1st cycle, which can be correlated with the fast initial MgO-CO 2 carbonation reaction, increasing the MgCO 3 content in the sorbent and contributing to the alkali's segregation [11].Indeed, due to the segregation, the alkali will be less dispersed on the MgO particles, and its role as a phase transfer catalyst between bulk MgO and CO 2 molecules decreases alongside the CO 2 capture capacity.Thus, between the 2nd and the 6th cycles, the carbonation reaction occurred slowly, and more time was required for the carbonation stabilization as confirmed by the narrow peaks shown in Figure 10a and Figure S2 in the Supplementary Materials.However, it seems that there is an arrangement involving the alkalis and the MgO that allows the sorbent to recover some CO 2 carrying capacity.After the 7th cycle, the carbonation reached the stabilization faster (evidenced in Figure 10a and Figure S2 in the Supplementary Materials), meaning that the alkalis were still enhancing the kinetic stage of carbonation reaction, but the sorbent started to lose its capture capacity.Harada et al. [19] reported carrying capacities of 440 and 310 mg CO 2 g −1 sorbent between the 1st and the 12th cycle for a MgO sorbent doped with 15% of (Na, Li, K)NO 3 .In line with the literature, the 15 (Na, Li, K)-MgO synthesized in this work achieved a carrying capacity of 460 mg CO 2 g −1 sorbent in the 1st cycle and 262 mg CO 2 g −1 sorbent in the last cycle.

Assessment of the Effect of Inert Support Addition
Temperature-programmed carbonation-calcination studies followed by TGA were carried out on samples of supported MgO sorbents (Ca or Ce-based) undoped and doped with 15% of (Na, Li, K)NO 3 to assess the effect of the support on CO 2 uptake (Figure 11a,b).Undoped MgO-SG-Ca and MgO-SG-Ce sorbents have their maximum CO 2 uptake of 73 mg CO 2 g −1 sorbent and 86 mg CO 2 g −1 sorbent, respectively, at the beginning of the TGA test at 25 • C, and we observed a gradual CO 2 release with the increase in temperature until ~400 • C. Both supported sorbents exhibit CO 2 calcination peaks significantly wider than those of the corresponding unsupported AMS sorbents (Figure 6).Conversely, the CO 2 uptake of the 15 (Na, Li, K)-MgO-Ca and of the 15 (Na, Li, K)-MgO-Ce sorbents is negligible at low temperatures until it reaches the respective inflection temperature.This parameter was obtained as it was explained in Section 3.3.1,and the respective plots can be found in the Supplementary Materials, Figure S3.The maximum instantaneous rate of change in temperature with respect to CO 2 uptake ( • C) was also determined, as it was explained above.Both parameters are summarized in Table 7.The promoted MgO-SG-Ce sorbent showed a higher inflection temperature than the promoted MgO-SG-Ca, meaning that the latter starts the CO 2 uptake sooner.As expected, the AMS doping enhanced considerably the CO 2 uptake of both MgO sorbents, 15 (Na, Li, K)-MgO-Ca and 15 (Na, Li, K)-MgO-Ce sorbents achieving a maximum CO 2 uptake of 299 mg CO 2 g −1 sorbent and 252 mg CO 2 g −1 sorbent at 384 • C and 393 • C, respectively.Summarizing, undoped MgO-SG-Ce exhibited higher CO 2 uptake than MgO-SG-Ca, but AMS doping was more effective for the latter, with 15 (Na, Li, K)-MgO-Ca exhibiting higher CO 2 uptake than 15 (Na, Li, K)-MgO-Ce.Moreover, 15 (Na, Li, K)-MgO-Ca also provides a more extensive carbonation than 15 (Na, Li, K)-MgO-Ce, since the former presents a wider carbonation temperature range than the latter.Table 7. Properties of unsupported and supported MgO sorbents (Ca or Ce-based support) undoped and doped with 15% of (Na, Li, K)NO 3 : maximum instantaneous rate of change in temperature with respect to CO 2 uptake ( • C), inflection temperature ( • C), maximum CO 2 uptake temperature ( • C), CO 2 uptake range ( • C) and maximum CO 2 uptake (mg CO 2 g −1 sorbent).Papalas et al. [20] report a CO 2 uptake of 317 mg CO 2 g −1 sorbent for an MgO sorbent promoted with 5% (mol) of CaCO 3 and impregnated with 20% of ternary mixture (Na, Li, K)NO 3 at 300 • C for 30 min under an atmosphere consisting of 30% of CO 2 , which is in line with the CO 2 uptake obtained in the present work for the 15 (Na, Li, K)-MgO-Ca sorbent, despite the difference between the CO 2 atmospheres.Jin et al. [21] synthesized a supported sorbent with CeO 2 (10% mol) impregnated with a mixture of LiNO 3 :NaNO 3 :Na 2 CO 3 :K 2 CO 3 = 0.2:0.76:0.04:0.5 (17% mol) MgO sorbent that achieved a CO 2 uptake of 450 mg CO 2 g −1 sorbent at 325 • C for 120 min under 100% CO 2 , which is almost double the CO 2 uptake obtained for the 15 (Na, Li, K)-MgO-Ce.It is important to note that in the above example a temperature plateau of 120 min (325 • C) was applied, while the present study does not include a temperature plateau.The 15 (Na, Li, K)-MgO-Ca sorbent starts to uptake CO 2 at a lower temperature (153 • C) than the 15 (Na, Li, K)-MgO-Ce (210 • C).However, the last one reaches the maximum instantaneous rate of change in temperature with respect to CO 2 uptake at a lower temperature (294 vs. 308 • C).The higher inflection temperature of 15 (Na, Li, K)-MgO-Ce was unexpected, since the same alkali ternary mixture was used.An explanation can be assigned to the high dispersion of both the CeO 2 and AMS on the sorbent.As it can be observed in the XRD patterns of the samples supported with Ce (Figure 3), broader peaks are observed, meaning there is a high dispersion of these compounds in the sorbent, which probably delays the molten effect of the ternary alkali salts mixture on the CO 2 uptake.Yu et al. [30] synthesized CeO 2 -MgO and MgO sorbents and found that despite the similarity of their initial surface areas, the doped sorbent had an increased CO 2 uptake, which was justified by changes in the pore structures and by the increase in the basicity of the MgO phase induced by the addition of CeO 2 .In the present study, the corresponding values of S BET are also similar (31 vs. 32 m 2 g −1 ), and comparatively with the unsupported sorbent, the maximum CO 2 uptake of 15 (Na, Li, K)-MgO-Ce (239 vs. 252 mg CO 2 g −1 sorbent) was slightly improved.It must be noted that no temperature plateau was considered during this experiment.It is important to note that as shown in Table 7 and in accordance with the literature [2], the addition of both Ca and Ce-based supports improved the MgO sorbents' performance.

MgO-Based
Figure 12a,b shows the variation of the MgO crystallite size of the undoped and doped sorbents with the Ca-or Ce-based support before and after the TG tests.
Both undoped MgO-SG-Ca and MgO-SG-Ce sorbents maintained the MgO crystallite size after undergoing the TGA test.The same behavior was already observed for the unsupported sorbent (Figure 7).For 15 (Na, Li, K)-MgO-Ca and 15 (Na, Li, K)-MgO-Ce, the MgO crystallites size increased 150% and 125% (Table 4), which was mainly due to the AMS impregnation procedure, while after the TGA experiment, the increase was only 10% and 22%, respectively.Hence, the AMS impregnation with 15% (Na, Li, K)NO 3 affected identically the supported MgO-Ca-SG and MgO-Ce-SG sorbents.The 15 (Na, Li, K)-MgO-Ca sorbent exhibited a more promising carbonation-calcination profile along the five cycles than the 15 (Na, Li, K)-MgO-Ce sorbent, since a higher CO 2 carrying capacity was achieved over cycles.However, both supported sorbents present quite stable peaks until the end of the test.Quantitively, the MgO-SG-Ca impregnated with 15% of (Na, Li, K)NO 3 obtained a carrying capacity of 375 mg CO 2 g −1 sorbent in the 1st cycle and of 275 mg CO 2 g −1 sorbent in the 5th cycle.Papalas et al. [20] reports a Ca-based supported MgO promoted with 20% of ternary mixture (Na, Li, K)NO 3 that varied its carrying capacity from 396 to 330 mg CO 2 g −1 sorbent for the same number of cycles, which is in agreement with the obtained results.Jin et al. [21] report a Ce-based supported MgO sorbent promoted with 12% of ternary mixture of (Na, Li, K)NO 3 with carrying capacities of 340 and 180 mg CO 2 g −1 sorbent between the 1st and the 5th cycles.The same sorbent but with an impregnation molar percentage of 17% managed to achieve 420 mg CO 2 g −1 sorbent in the 1st cycle and 330 mg CO 2 g −1 sorbent in the 5th cycle.The 15 (Na, Li, K)-MgO-Ce sorbent presents an impregnation molar percentage of 15%, which is in between those mentioned in the literature, and it achieved a carrying capacity of 375 mg CO 2 g −1 sorbent, which is within the same range.In fact, the unsupported 15(Na, K, Li)-MgO sorbent performs better than the supported sorbents, since a higher carrying capacity was achieved for this sorbent, i.e., 460 and 330 mg CO 2 g −1 sorbent for the 1st and 5th cycle, respectively.Nevertheless, after the 1st carbonation-calcination cycle, the 15 (Na, K, Li)-MgO-Ca is the unique one that maintains a similar carbonation rate profile along the cycles.The carbonation profiles (Figures 8 and 13a) show that the 1st carbonation rate is similar for all the sorbents, but after the 2nd cycle, the carbonation slope decreases, meaning that the carbonation rate is delayed, except for (Na, K, Li)-MgO-Ca.The carbonation reaction kinetics is one of the MgO-based sorbent's drawbacks, and the faster carbonation of 15 (Na, K, Li)-MgO-Ca is a huge advantage, which can contribute to a comparative upscale of MgO sorbents.Similar results were found by Cui et al. [31]; the authors performed kinetic studies with Ca and Mg precursors and verified that the activation enthalpy of CO 2 sorption for doped AMS-MgO sorbent was lower than for the undoped AMS-MgO, indicating that the CaCO 3 decreases the activation energy of CO 2 sorption on MgO.The CaCO 3 should act as a spacer among MgO crystallites, enhancing the AMS dispersion on the MgO surface area and making the carbonation reaction easier.
The positive effect of the calcium support was already evidenced by the lower inflection temperature of the 15 (Na, K, Li)-MgO-Ca sorbent relatively to 15 (Na, K, Li)-MgO (153 vs. 175 • C, Table 7) and the maximum CO 2 uptake obtained when no temperature plateau was applied (299 vs. 239 mg CO 2 g −1 sorbent, Table 7), respectively.

Conclusions
This study evaluated the role of a NaNO 3 salt and a ternary mixture (NaNO 3 , LiNO 3 and KNO 3 (18/30/52; % mol.)) molar fraction (15, 25 and 35%) on the MgO carbonation temperature range and on the maximum CO 2 uptake temperature.Regarding the carbonation temperature range, the ternary mixture offers a larger temperature window of operation when compared to that of the single salt NaNO 3 , which is justified by the lower melting temperature of the ternary mixture (120 vs. 308 • C).Also, it is observed that the increase in the percentage of the NaNO 3 single salt between 15 and 35% allows enhancing the CO 2 uptake by the sorbent, but for the case of the ternary mixture sorbents, the opposite behavior is observed.This should be explained by the lower melting temperature of the ternary mixture that promotes the enhancement of the AMS dispersion on the MgO surface at the selected carbonation temperature (300 • C).Probably, the increase in the thickness of the molten layer covering the surface of the sorbent beyond a certain limit hinders the CO 2 uptake by increasing the mass transfer resistance, which justifies the enhanced results obtained with 15% of the ternary mixture.Despite the lower melting temperature of the ternary mixture, the inflection temperature only starts around 175 • C, suggesting that at low temperatures, the accelerating effect of the addition of the ternary alkali metal nitrates on the CO 2 uptake by MgO is reduced, which is justified by the thermodynamic equilibrium of MgO/MgCO 3 .
Regarding the supported sorbents (Ca and Ce-based), both show a lower carrying capacity along the carbonation-calcination cycles than the unsupported sorbent.However, the 15 (Na, K, Li)-MgO-Ca sorbent shows a stable slope for the carbonation, while for the other sorbents, it decreases after the 1st cycle, meaning that the maximum carbonation rate was delayed.Thus, the carbonation performs faster for the 15 (Na, K, Li)-MgO-Ca sorbent, overcoming some kinetic limitations, which is considered as a huge advantage, and it can contribute for a sooner upscale of MgO sorbents for industrial CO 2 capture applications.

Figure 1 .
Figure 1.Synthesis of MgO based sorbents by the sol-gel method.

Figure 4 .
Figure 4. Textural properties of MgO-SG, MgO-SG-Ca and MgO-SG-Ce before and after doping with 15(Na, K, Li)NO 3 : (a) specific surface area (S BET ) and total pore volume (V p ), (b) pore size distribution (PSD) estimated by the BJH desorption.

Figure 5 .
Figure 5. Thermogravimetric analysis: (a) thermal decomposition in mass loss (%), (b) first derivative of mass change and (c) heat flow profiles of NaNO 3 and (Na, Li, K)NO 3 with temperature.

Figure 11 .
Figure 11.Profile of CO 2 uptake of supported MgO sorbents undoped and doped with 15% of (Na, Li, K)NO 3 : (a) Ca-based support or (b) Ce-based support.

Figure 12 .
Figure 12.Average crystallite size of supported MgO sorbents undoped and doped with 15% of (Na, Li, K)NO 3 before (yellow, blue, orange and mustard bars) and after (green bars) TG test between 20 and 500 • C: (a) Ca-based support or (b) Ce-based support.The performance of supported sorbents along the cyclic carbonation-calcination cycles was assessed in TGA tests (Figure 13a,b).

Table 1 .
CO 2 uptake and experimental conditions of unsupported and supported MgO-based sorbents impregnated with AMS.

Table 2 .
Summary of the molar composition of unsupported and supported MgO sorbents.

Table 5 .
Thermal decomposition of mono and ternary alkali salts under air and CO 2 atmosphere: working range and decomposition range ( • C).