Protocol for cationic exchanges is the same used in recent studies [16] [18]. 3 g of NaX zeolite were placed under agitation in 250 mL during 24 hours (h) in an aqueous solution containing 0.02 mol∙L⁻¹ of the metal nitrate or chloride ([Ba²⁺] = [Ca²⁺]). The samples was then filtered, then washed with an ultrapure water (3 × 40 mL) to eliminate the free nitrate or chloride ions into the adsorbents. The solids was then dried during 24 h at 373 K in a furnace. Exchanged zeolites were been sieved to have a particle size between 0.2 and 0.4 mm. In this study, all samples obtained by cationic exchanges and the reference will be designated respectively by letters E and R. For example R-NaX (Reference NaX) and E-CaX and E-BaX (Exchanged zeolites by the calcium and Barium).

Characterization techniques and protocols used in this work are the same as those applied in recent studies [16] [18].

Thermal stability and mass losses of samples were obtained using an SDT-Q-600 TA instrument with a stream of dry argon (100 mL∙min⁻¹) in a temperature range between 298 and 873 K, with a 5 K∙min⁻¹ ramp.

The specific surface area and pore volumes were obtained using a Micromeritics TRISTAR 3000 instrument with automatic injection of nitrogen, where about 100 - 150 mg of sample was pre-treated during 1 h at 363 K then at 623 K during 10 h.

The N₂ physisorption isotherms were carried out at 77 K. The specific surface area was determined by the BET method, and the micropore volume was estimated through the t-plot method, applied to a layer thickness between 5 and 7 Å, and the Dubinin-Raduskevitch method. Therefore, the mesopore volume was determined by the difference between the total pore volume, determined at 0.97 P/P₀, and the total porous volume obtained by the Dubinin-Raduskevitch method.

X-ray diffraction (XRD) patterns were registered at room temperature using a Siemens D5005 low-resolution diffractometer using Cu Kα monochromatic radiation (λ = 0.154050 nm). The samples were scanned in the 3˚ - 70˚ (θ) range with a 0.01 (θ) step and a scanning speed of 0.1˚(θ)/min. Structural Rietveld refinements were carried out using the Fullprof program.

Protocols on carbon dioxide thermal desorption (CO₂-TPD) measurements are the same used by T. Belin et al. [18].

Protocols of Fourier transformed infrared (FTIR) measurements are the same used by T. Belin et al. [18].

Desorption energies of CO₂ on adsorbents were calculated using the method of Cvetanovic et al. [19]. The assumptions are that no readsorption occurs and no diffusional limitations of the desorbed species within the porosity during the process. Values were derived from the following equation [18]:

2 ln ( T m ) − ln ( β ) = E d e s R T max

where T_max (or T_m) is the temperature at the peak maximum (K), β is the heating rate (K∙min⁻¹), E_des is the desorption energy (J∙mol⁻¹) andR is the perfect gas constant (J∙mol⁻¹∙K⁻¹). Desorbed quantities of CO₂ are determined using a calibration curve, previously established, and representing the TCD signal in function of known CO₂ amounts.

In Table 1 are reported the ionic radius of Ca²⁺, Ba²⁺ and Na⁺, cationic exchange percentages and the global Si/Al molar ratio. The cationic exchange percentage is close to 100% after replacement. The same order of magnitude as those obtained for Chabazite, Beta, X and Y in the preparation of adsorbents for the elimination of CO₂ [16] [20] [21]. Cations are considered as centers of adsorption for CO₂ and that is why their numbers and distribution will have a direct influence on their performances [16] [22].

In contrast, the exchange of monovalent cations by divalent cations results in a decrease in the number of cations and of the electronic density in the cationic supercage [8] [16] [23]. This would cause a loss of the adsorption amounts of carbon dioxide onto zeolites modified with the divalent cations [16] [24].

Dehydration temperatures of adsorbents are determined by thermal gravimetric and differential analysis (TGA-TDA) and Figure 1 shows the curves obtained. On the three adsorbents, most of the weight is lost between 300 and 600 K related to the loss of water adsorbed into the pores of the zeolites [16] [24] [25].

After 600 and 1023 K, the weight losses are very low. The ratio between the total mass lost in 623 and 1073 K was calculated and the results are shown in Table 2. This ratio is very high as it is between 92% and 96%. Compared to reference zeolite NaX and E-CaX. The exchanged E-BaX zeolite shows a ratio more important.

This result is in line with those of Moïse et al. [25] who showed that the amount of water adsorbed on MgX zeolite was higher than that adsorbed on NaX zeolite. The author attributed this behaviour to the radius of the cation Mg²⁺ which is higher than that of Na⁺ resulting in a radius of hydration higher for Mg²⁺.

Thus, for all adsorbents, the minimal required temperature for the activation of exchanged zeolites is of about 623 K [16]. At this temperature, an average of 85% of the adsorbed species have already been desorbed. Then, before each experiment, an outgassing thermal treatment will be carried out in two steps as indicated by T. Belin et al. [18].

N₂ adsorption/desorption isotherms obtained on R-NaX and the exchanged zeolites (E-CaX and E-BaX) are shown in Figure 2. According to IUPAC norms, the shape of the isotherms is of Type I for all the adsorbents indicating that they can be considered as microporous zeolites. A strong adsorption is observed at low P/P_saturation (less than 0.2) followed by a plateau at higher pressure indicating that

no more adsorption occurs. From these isotherms, BET surface areas and microporous and total porous volumes have been calculated (Table 2). Discussions of its results on the reference NaX and exchanged BaX have been widely discussed on previous studies [16] [18].

However after cationic exchange, there is a decrease of BET surface areas and the microporous volumes. It seems that the main cause of these losses BET surface areas and the microporous volumes is the decrease of cationic density into supercage of faujasite X or to a plugging of the porosity as indicated by Chandwadkar et al. in these studies [20] [21]. These losses are also related to the quality of cationic exchange. They increase when the cationic exchange rate is important (Table 1 and Table 2).

Ammoudi [20] also made the same observations during cationic exchanges of Cu²⁺ on zeolite X with different rates. However, since the CO₂ adsorption occurred mainly into the micropores, a significant decrease in the accessible pore volume would result in a loss of CO₂ amounts adsorbed into exchanged zeolites [18].

Results of this analysis do not show amorphous phase on these diffractograms. The crystalline structure of the faujasite X does not seem to be modified after cationic exchange since the significant XRD peaks of reference NaX remain visible for the XRD spectrum of the exchanged zeolite (Figure 3).

Moreover, the data of these XRD spectra indicates the absence of parasite phase and/or impurities within the detection limit of the apparatus (~2%) for E-BaX and E-CaX. On the other hand contrary to E-CaX, the diffractogram of the E-BaX zeolite shows some peaks more intense in the region between 10˚ and 45˚. This is probably due to new arrangement of cationic sites within of the X zeolite structure [16].

The FTIR analysis of all adsorbents was performed in the range between 4000 and 1000 cm⁻¹. Hydroxyl groups (HO⁻), the linearly adsorbed CO₂ on the cationic sites (CO₂-Na⁺), the bicarbonates ( HCO 3 − ), the unidentate and bidentate carbonates ( CO 3 2 − ) are the species studied in this region [18] [26].

The spectra obtained at 373 and 473 K on the reference zeolite (R-NaX) and exchanged zeolites (E-CaX and E-BaX) are shown in Figure 4. Differences

between the samples are visible and two main domains are studied: 2400 - 2200 cm⁻¹ and 1800 - 1000 cm⁻¹. These regions are respectively attributed to the physisorbed and chemisorbed CO₂ species [18] [23] [27] [28]. In the domain of the physisorbed CO₂ at 323 K (Figure 4(a)) all adsorbents show a band of physisorbed CO₂. These zeolites present the adsorption sites of CO₂ at this temperature. However, the intensity of the absorption band of exchanged E-CaX remains higher than that of two others adsorbents (Figure 4(a)). The amounts of physisorbed CO₂ are thus more important on this adsorbent at 373 K, which can translate strong interactions CO₂/cation.

At 473, a quasi-disappearance of the band of physisorbed CO₂ is observed on zeolites R-NaX (Figure 4(b)). Exchanged zeolites E-CaX and E-BaX show a band of low intensity. Bonenfant et al. [12] explained that a loss of the CO₂ adsorbed capacities is generally expected on the NaX zeolite when the temperature increases, due to the thermal agitation in the zeolite pores. Thus, a decrease of the microporous volume is only sufficient to explain the intensities of this band whatever the modification process [18]. However, the absorption band obtained on E-BaX is more intense than that of E-CaX (Figure 4(b)), which indicates an amount of physisorbed CO₂ more important present at this temperature. This remains in conformity with the basic force of exchanged cation (Ba²⁺ > Ca²⁺). It seems in this case, that a reinforcement of the alkalinity of the cationic sites in the reference zeolite NaX would increase the trapping of physisorbed CO₂ at high temperatures on the cationic sites.

In the chemisorbed CO₂ domain (1800 - 1200 cm⁻¹), R-NaX zeolite shows three bands well defined at 373 and 473 K. These bands are identified as unidentate carbonates at 1691 and 1360 cm⁻¹ ( CO 3 2 − : υ_asOCO et υ_sOCO respectively) and as water or bicarbonates at 1642 cm⁻¹ ( HCO 3 − : υ_asOCO ou υ_asH₂O) [18] [28] [29]. The presence of bands due to carbonate groups revealed that the framework oxygen atoms surrounding the cations in the reference NaX zeolite behave as stronger basic centers [30]. Contrary to R-NaX, at 373 K, exchanged E-CaX shows only the presence of monodentate and bidentate carbonates (1672 cm⁻¹ and 1628 cm⁻¹) and on the exchanged E-BaX, only the bicarbonates are show (1650 cm⁻¹). It seems that at 373 K (Figure 4(b)) the bicarbonates and bidentate carbonates are majority on exchanged zeolites (E-BaX and E-CaX respectively).

At 473 K, there is the appearance of two bands at 1480 cm⁻¹ and 1420 cm⁻¹ on reference R-NaX corresponding to bidentate carbonates. A decrease intensity of bicarbonates (1642 cm⁻¹) is observed. The monodentate carbonates (1700 and 1350 cm⁻¹) are always present. However, the amount of the carbonates is reduced by the temperature as evidenced with the slight decrease of bands intensities. The spectra obtained on exchanged E-BaX (Figure 4(b)) indicate that the bicarbonates (1642 cm⁻¹) are majority, in front of bidentate carbonates (1360 cm⁻¹). For the exchanged E-CaX, the majority species remain are the bidentate carbonates (1628 cm⁻¹). We will tend to associate the energies desorption of each type of CO₂ species.

The thermal desorption curve obtained show two regions (Figure 5) which are attributed to physisorbed (200 - 400 K) and chemisorbed CO₂ species (above 600 K) [18] [31]. For each region, these curves associated with the desorbed amount are deconvoluted using Gaussian functions and the peaks are indexed. These peaks represent the minimum number to avoid the errors between experimental and theoretical profile [18] in its study of physisorbed CO₂ on NaX zeolite. It was then possible to estimate the temperature of maximum desorption, to calculate the amounts and the desorption energies of physisorbed and chemisorbed species of adsorbed CO₂.

Table 3 recapitulates the total amounts adsorbed and desorbed of CO₂ obtained after exploitation of thermal desorption curves to 2, 4, 6 and 10 K∙min⁻¹ for R-NaX and of exchanged zeolites. The amounts of adsorbed and desorbed CO₂ on reference R-NaX zeolite (4.041 and 3.103 mol∙g⁻¹ respectively) are higher than exchanged zeolites (Table 3). On the other hand, it is interesting to note

that the CO₂ desorbed percentage increases with the alkalinity of the cation (Table 3). This must be also close to the specific loss of surface and to microporous volume observed during textural analysis.

The total amount adsorbed decreases with the reduction of porous volume but the difference between adsorbed and desorbed amount is also reduced. This observation must be relativized by the observation of the relationship between amounts desorbed and adsorbed.

For example, approximately 77% of the species are desorbed on the zeolite of R-NaX reference. The residue seems be irreversibly adsorbed. A decrease of physisorbed CO₂ amounts is also observed on exchanged zeolites E-CaX and E-BaX. These results also can be explained by the fact that the cationic density is decreased in the supercage during the cationic exchanges with bivalent cations (2Na⁺ for 1M²⁺) [10] [12] [32] and that specific surface is decreased. Desorbed amounts of chemisorbed CO₂ on exchanged zeolites are higher than those obtained on the R-NaX reference. This seems related to the basic character of the cations Ca²⁺ and Ba²⁺ which is higher than Na⁺. However, exchanged zeolite E-BaX shows a total desorbed amount near to 100% at 873 K (Table 3) and an average proportion of chemisorbed absorbed species (≈12%), contrary to E-CaX which shows an average proportion of 19.1 with a total desorbed amount (about 75%) near to reference R-NaX zeolite (Table 3). For the adsorption /desorption processes, the exchanged zeolite E-BaX can be interesting.

In CO₂ physisorbed region (210 K - 450 K), the deconvoluted curves using Gaussian functions obtained to 2 K∙min⁻¹ on exchanged E-BaX, E-CaX zeolites and on R-NaX are show on Figure 6. The maximum desorption temperature T_m determined is also indicated.

Desorption energies calculated for the CO₂ physisorbed (and chemisorbed) to the maximal desorption temperature (T_m) after exploitation of curves to 2, 4, 6 and 10 K∙min⁻¹ des adsorbents studied are recapitulated in Table 4. Reference NaX zeolite indicates a desorption energy of CO₂ physisorbed equal to 13.47

u.d: undefined.

kJ∙mol⁻¹. These obtained on exchanged zeolites E-CaX and E-BaX are comparable (15.77 and 15.07 kJ∙mol⁻¹ respectively) and higher than those obtained on R-NaX.

It seems that the cationic exchanged with bivalent cations (Ba²⁺ and Ca²⁺) increase CO₂ interactions at low temperature with adsorbents. This slight increase of energy can be attributed to basic character of the cation (Na⁺ < M²⁺ or M = alkaline earth metal ions) [12]. CO₂ molecule will be, in cases, linked more strongly on the cation most basic. In addition, a study realizes by Aguilar et al. on the CO₂ adsorption on the clinoptilolite has indicated that the substitution of Na⁺ and K⁺ by Ca²⁺ caused a rise of the basicity of the framework oxygen acting as basic center [33]. Others studies carried out by Krista et al. [11], on the CO₂ adsorption by X and Y zeolites exchanged with alkaline metal, have shown an increase of CO₂/zeolite when the cation size increases. And according to them, this is related with an increase of basicity of the framework compared to the smaller cation forms.

Contrary to desorption profiles obtained in physisorbed region, the desorption profiles in CO₂ chemisorbed region between 523 - 813 K of exchanged zeolites (E-BaX and E-CaX) are very different compared to that obtained on reference NaX zeolite (Figure 7). In the region of the exploited temperatures, the desorption profile of R-NaX shows one massif with a maximal desorption temperature (T_m1) about 673 K.

The E-BaX desorption profile shows two massifs. One with a maximal desorption temperature (T_m1) belong 673 K and another with temperature (T_m2) above 673 K. This adsorbent seems developed of adsorption sites, at high temperature, lower than those of reference NaX in the temperature region studied. On exchanged E-CaX zeolite, one massif only is observed with a maximal desorption temperature (T_{m 1}) higher to that determined on R-NaX and comparable to that determined on E-BaX zeolite (Figure 7).

The infrared spectra results at 473 K (Figure 4(b)) for the exchanged E-CaX, the majority species remain mainly the bidentate carbonates with the CO₂ adsorption sites more homogeneous on the surface than other exchanged. It

appears therefore that the species of CO₂ chemisorbed to peak of maximal temperature desorption (T_m > 673 K), on exchanged E-CaX and E-BaX zeolites (T_{m 1} and T_{m 2} respectively) are bidentate carbonates (Figure 7). Thus, desorption energies related of these species on E-CaX and E-BaX zeolites are about 81.1 and 80.8 kJ∙mol⁻¹ respectively (Table 4). It seems that cationic exchanges no change the desorption energies of bidentate carbonates on faujasite X zeolites.

The infrared spectra obtained on exchanged E-BaX (Figure 4(b)) indicate that the bicarbonates are majority, in front of bidentate carbonates on the surface of the zeolite. Thus, CO₂ chemisorbed species with a maximal temperature desorption belong to 623 K (T_{m 1} < 623 K) correspond to bicarbonates (Figure 4(b)) and the associated energies on E-BaX are about 35.3 kJ∙mol⁻¹, corresponding to secondary adsorption sites (Table 4). On reference NaX zeolite, related energies bicarbonates are about of 62.3 kJ∙mol⁻¹. Compared to R-NaX zeolite, on E-BaX zeolite a decrease of CO₂ desorption energy of bicarbonates is observed (35.27 kJ∙mol⁻¹). It seems that the desorption energies of bicarbonates decrease and depending on the type of exchanged cation.

On other hand, in the literature, the isosteric heat of adsorption (Q_st) varies between 40 and 46 kJ∙mol⁻¹ [10] [32].

Desorption energy of bicarbonate does not correspond to the Q_st value found in the literature, but seems to be near. Then, the determination of Q_st value on NaX zeolite will be principally assigned to chemisorbed species. Amine khelifa et al. [10] have also observed a decrease of Q_st values when Na⁺ is exchanged with Mg²⁺ cation within the NaX zeolite.