Table 1 shows the chemical composition (wt%) data of the selected three samples. The data indicate that the percentage of SiO₂ in clay and Triassic sandstone samples was relatively high. This is consistent with the presence of quartz as shown by XRD [Figure 2(a)]. The higher concentration of Na ion in the evaporate sample confirmed the presence of the halite. The percentage of the K₂O in the clay and Triassic sandstone show that the illite is the major compound in both samples.

Figure 2 presents the XRD patterns of three powder samples. As illustrated by this figure, the clay sample [Figure 2(a)] essentially comprises illite (10 Å) and kaolinite at 7.16 Å; the associated minerals are dolomite (2.88 Å) and quartz (3.34 Å). X-ray diffraction analysis of the evaporate sample shows the presence of halite at 2.82 Å and 1.99 Å [Figure 2(b)]. The Triassic sandstone samples are composed of illitic and kaolinitic clay associated to quartz [Figure 2(c)].

The SEM analysis (Figure 3) exhibits the morphology of the three soils. The SEM imaging shows that the particle morphology is laminated. This is due to the clay fraction in the sample represented by the layers of argillaceous form. In the case of Jurassic evaporates, the particles have a spherical form with very small compared to the others samples.

In Table 2, we provide the textural characteristics of these different materials obtained with the Nitrogen adsorption technique. The specific surface area and the total pore volume are the most important characteristics of adsorbents. For all samples, the SBET is relatively lower and the clay sample has the highest specific surface area. The total pore volumes of these samples are comparable.

In this study, a gas chromatograph series IGC 10 M was used to study the adsorption of CO₂ on the following adsorbents: clay, Triassic sandstone and Jurassic evaporate. It is worth noting that the most important difference between these samples is the particle size and the outer surface.

After filling with a given mass of adsorbate, the adsorption column is conditioned with a stream flow of CO₂ and heated to a temperature of about 120˚C during 24 hours to remove any residual moisture. Then, the system is cooled to the ambient temperature. CO₂ is introduced directly into the chromatograph through a manual injection valve (Figure 1).

We balance the katharometer at each injection; check the flow rates of the gas in the reference column and that adsorbed. Then, a definite amount of CO₂ is adsorbed for a given injection pressure, depending on the flow rate value and injection temperature, an electrical signal in the form of a chromatographic peak is obtained on a recording paper.

At this stage, one can detect the coordinates in each chromatogram point using a Matlab program based on the trapezoidal method.

The adsorption experiments were conducted to estimate the gas flow rate and the pressure of CO₂ injected at the inlet in order to reach thermodynamic equilibrium inside the column. For this purpose, we inject a fixed amount of CO₂ in a column filled with a mass m_b of adsorbent, at a velocity V_inj and a specific temperature T_inj. Then, we vary the flow vector gas, D, setting the pressure of the injected CO₂ from the inlet. Table 3 summarizes the equilibrium conditions inside the column for each adsorbent. The adsorption experiments were conducted to estimate the gas flow rate and the pressure of CO₂ injected at the inlet in order to reach thermodynamic equilibrium inside the column. For this purpose, we inject a fixed amount of CO₂ in a column filled with a mass m_b of adsorbent, at a velocity V_inj and a specific temperature T_inj. Then, we vary the flow vector gas, D, setting the pressure of the injected CO₂ from the inlet. Table 3 summarizes the equilibrium conditions inside the column for each adsorbent.

For each pair CO₂/adsorbent, we have made several injections at different temperatures. The obtained experimental results are shown in Table 4.

The study of Figure 4 shows that all the CO₂ adsorption isotherms on the selected adsorbents have the same shape and are of type I according to the classification of BDDT. Therefore, we can conclude that basically a

monolayer adsorption takes place for all the adsorbents in the fixed operating conditions.

Furthermore, the isotherms show no flat curves, indicating that the adsorption sites are not saturated in the range of the introduced amount of CO₂. In all cases, the largest adsorption capacity occurs at the lowest temperature (room temperature) and decreases with the increase of the injection temperature. This is obvious since we are dealing with an exothermic adsorption phenomenon.

Langmuir and Freundlich models are extensively used to model adsorption isotherms of CO₂ on different adsorbents [11] -[13] .

The Langmuir model is used to calculate the maximum adsorbed amount q_m to form the monolayer as well as the adsorption constant K_L [14] [15] . These parameters are determined using the linearization of the model [equation (1)] by plotting 1/q vs. 1/P.

The linear form of Langmuir model equation allows us to determine the parameters of this model (i.e. q_m and k_L). This model is given by the following equation [16] [17] :

When we plot 1/q as a function of 1/P, one can obtain a straight line with a slope 1/q_mK_L that intercepts 1/p- axis at 1/q_m. Then, the two-parameters of this model can be deduced.

It is important to note that Langmuir model is a good representation of adsorption isotherms of type I.

The Freundlich model parameters n and K are also determined by the linearization of the empirical model [Equation (2)] by plotting as a function of. Freundlich model is represented by the following relationship:

where K is the Freundlich constant which is mainly related to the maximum adsorption capacity and n a constant characterizing the interaction between adsorbate and adsorbent. In general, this model is valid for the monolayer adsorption on heterogeneous surfaces, where the adsorption sites are different in nature [17] [18] .

By plotting as a function of, one can get a straight a line with a slope 1/n that intercepts-axis at.

The analysis of Table 5 shows that the Langmuir model describes correctly the adsorption isotherms of CO₂ on clay for all fixed temperatures with a correlation coefficient R² between 0.9995and 0.9998. The high values of the correlation coefficients obtained in the case of CO₂ adsorption on clay are in agreement with the basis of Freundlich model, involving surface heterogeneity.

It is important to highlight that above 46˚C, the maximum capacity q_m predicted by the Langmuir model is systematically lower than the adsorbed amounts [12.0482 ´ 10⁻³ (g/g)]. However, the capacity q_m must be greater than or equal to the adsorbed quantities calculated from the isotherm curves. Although, the correlation coefficients are satisfactory (between 0.9971 and 0.9996), the values of q_m are meaningless.

The Freundlich model parameters n and K are also provided (Table 6). As it is well known, lower values of

n imply a higher heterogeneity of sites [19] . In our case, n is higher than 1 which implies that heterogeneity of sites is lower. In addition, the K values vary slightly with temperature.

From Table 6, it comes out that the Langmuir model describes adequately the experimental data of adsorption of CO₂ on the Triassic sandstone for all the tested temperatures with a correlation coefficient R² between 0.9992 and 0.9998. Therefore, the experimental adsorption isotherms can be linearized according to the Langmuir model. In addition, the CO₂ adsorption isotherms of the Triassic sandstones can be correctly described by the Freundlich model. The obtained correlation coefficients imply a high surface heterogeneity. Furthermore, the value of n is greater than 1 indicating that the heterogeneity of sites is low. Results show also that K values are strongly subordinate with temperature and that adsorption sites are not energetically homogeneous.

It is worth noting that for a temperature above 27˚C, the maximum capacity q_m predicted by the Langmuir model is below the maximum capacity q_m for the same temperature obtained from the adsorption isotherms. Thus, q_m values seem to be coherent and meaningful. According to Table 7, the Langmuir model describes well the experimental data. It reflects a localized adsorption on homogeneously-distributed sites. The Freundlich model describes appropriately the adsorption isotherms for the couple CO₂/(Liassic evaporates) for different temperatures values. Moreover, the changes in the Langmuir parameters are not regular with temperature.

It is important to note that for all temperatures, the maximum adsorbed amounts of CO₂ (i.e. q_m) predicted by the Langmuir model are systematically lower than the adsorbed amounts corresponding to an injected amount of CO₂ equal to 0.42 cm³. However, the values of q_m should logically be higher or equal to the quantities adsorbed when plotting isotherms [7 ´ 10⁻³(g/g)]. Thus, the obtained values of q_m are meaningless, even if the Langmuir curves are satisfactory.

Then, we can conclude that the Langmuir model does not allow us to describe correctly the adsorption isotherms of CO₂ on the Jurassic evaporate. In addition, the values of K_L indicate a low affinity of CO₂ on evaporate with relatively-homogeneous sites.

In order to determine the adsorption nature for each adsorbent, we have calculated the enthalpy of adsorption based on the Clausius-Clapeyron equation.

Obviously, the isosteric enthalpy of adsorption ΔH_ads is an important thermodynamic property in gas-solid adsorption problems. Then, it is necessary to draw the isosteric isotherms obtained at different temperatures. To this end, we have plotted the variation of as a function of 1/T [20] . These results are not shown here for the sake of brevity. The slopes of these isosteres represent the adsorption enthalpy values divided by the ideal gas constant R.

We note that the curves ΔH_ads as a function of q for all adsorbents (Clay, Triassic Sandstone and Jurassic evaporates) [21] , not shown here for the sake of brevity, present negative values which confirm the exothermic nature of the adsorption phenomenon. In addition, the comparison between enthalpy adsorption values against available literature values are within the magnitude of the physical adsorption [20] [22] . We note also that the variation of isosteric enthalpy is small for clays and evaporates, showing that the heterogeneity of their sites is low.

In Table 8, the adsorption enthalpies of the different three couples studied here are depicted. In terms of CO₂ adsorption on clay, the average isosteric heat of adsorption is equal to 3.2904 kJ/mol. Variations in enthalpy as a function of the adsorbed amount are quite small for this material, suggesting that the heterogeneity of its surface is low, for the Triassic sandstone, the enthalpy of adsorption appears as an increasing function of the adsorbed amount or recovery rate and the isosteric heat of adsorption is equal to 6.5007 kJ/mol. For Jurassic evaporates, one can see that the enthalpy of adsorption decreases when the amount adsorbed increases. This can be explained by the fact that the adsorption of molecules on the first layer is on the first parts of the surface having the highest interaction energy. The adsorption of the molecules results in heat generation with an average isosteric adsorption heat equal to 5.6990 kJ/mol.

In this section, we aim to develop a comparative study between the adsorbents in order to choose the most efficient one for the adsorption of carbon dioxide CO₂.

Based on the previous discussions, the CO₂ adsorption on different materials is exothermic. Indeed, the temperature plays an important role as it influences the adsorption capacity by affecting the amount of the adsorbed CO₂ molecules. It has been observed that the number of molecules in the column at the equilibrium decreases with temperature. Therefore, the best adsorption capacity of CO₂ is obtained at a low temperature, (i.e. for a temperature close to room one). The adsorbed amount of CO₂ on different adsorbents at low temperatures is illustrated in Table 9.

Increasing the capacity of adsorption of various adsorbents is in the following order: q_clay(15) > q_{Triassic sandstone}(14) > q_{Jurassic evaporates}(7). Moreover, the largest adsorption capacity was found with clay due to its larger surface area (10.97 m²/g) followed by the Triassic sandstone (6.09 m²/g). The highest percentages of SiO₂ group in these two-adsorbents promote a good adsorption of CO₂.

Hence, the choice of an adsorbent depends on several factors such as:

-Chemical composition of the adsorbent surface;

-The textural characteristics of the adsorbent (specific surface area, pore size distribution, connected microporous...);

-The operating conditions of adsorption.