Removal of ionic liquids and ibuprofen by adsorption on a microporous activated carbon: Kinetics, isotherms, and pore sites

2.1 Adsorbate molecules

4-Tert-butyl-1-propylpyridinium bromide (named IL1, Fig. 1a) and 4-tert-butyl-1-(2-carboxyethyl)pyridinium bromide (named IL2, Fig. 1b) were synthesized in our laboratory by conventional methods. Stoichiometric amounts of 4-tert-butylpyridine (96%, Aldrich) and 1-bromopropane (99%, Aldrich) or 3-bromopropionic acid (>99%, Aldrich) were mixed under stirring (24 h, room temperature) in order to get about 25 g of IL1 or IL2, respectively. Reaction products were washed 5 times with ethyl acetate (>99.7%, Aldrich) and further dried under vacuum at 10⁻³ mbar. The final white powders obtained were stored in stoppered flasks under an argon atmosphere placed in a desiccator containing silica gel.

Close

Figure 1

Developed formulas of (a): 4-tert-butyl-1-propylpyridinium bromide (IL1), (b): 4-tert-butyl-1-(2-carboxyethyl)pyridinium bromide (IL2) and (c): ibuprofen.

Export to PPT

The purities of the synthesized ILs were examined by ¹H and ¹³C NMR spectroscopy using a Bruker 400 MHz spectrometer after dissolution in CDCl₃. Chemical shifts were reported to internal TMS and set at 0 ppm. A pKa of 4.1 was determined for IL2 by volumetric titration.

2-[4-(2-methylpropyl) phenyl]propanoic acid, also named ibuprofen (IBP, Fig. 1c), was purchased from Sigma-Aldrich (>98% purity, pKa = 4.91). As IBP is very slightly soluble in pure water (∼21 mg L⁻¹) (Yalkowsky and Dannenfelser, 1992), it was dissolved in a KH₂PO₄/Na₂HPO₄ buffer (pH = 7.5). A solubility of 300 mg L⁻¹ was determined in this buffered solution. To measure it, an excess of IBP was added to the buffer solution and stirred until dissolution equilibrium was reached. After filtration of the residual undissolved IBP, the concentration of the solution was determined by UV–visible absorption (λ = 220 nm) using a calibration curve obtained from IBP solutions prepared in a 10v/90v methanol/water solvent (Garzon and Martinez, 2004).

The molecular volumes of the adsorbates, determined using Turbomole software (Ahlrichs et al., 1989), were of 0.53 nm³ for IBP, 2.02 nm³ for IL1 and 2.27 nm³ for IL2.

The octanol/water partition coefficient values of IBP, IL1 and IL2 calculated by using Chemsketch software and model of ref. (Viswanadhan et al., 1989) are 3.72, −1.84 and −2.7, respectively. Thus, IL1 and IL2 are hydrophilic whereas IBP is hydrophobic. Owing to the presence of carboxylic group on IL2, it is more hydrophilic than IL1.

2.2 Activated carbon

The raw activated carbon cloth (ref. 900–20) referred to T0 was provided by Kuraray Chemical Co. Ltd. (Japan). Before adsorption experiments, it was washed with a 0.1 mol L⁻¹ hydrochloric acid solution under stirring for 24 h, in order to remove the metal salt impurities. After filtration, it was rinsed with distilled water (∼200 mL per g) until reaching a constant pH of the filtrate and dried (383 K, 24 h).

2.3 Textural characterization of the activated carbon cloth before and after adsorption

The porosities of T0 samples prior to and after adsorption were characterized by N₂ adsorption–desorption at 77 K and CO₂ adsorption at 273 K using an ASAP 2020 sorptometer (Micromeritics). The analyzed samples, recovered from the ILs or IBP adsorption isotherm studies, correspond to T0 loaded with the following adsorbate content: 56.7 mg g⁻¹ for IBP (Guedidi et al., 2017), 163.82 mg g⁻¹ for IL1 and 212.15 mg g⁻¹ for IL2. Loaded samples were respectively named T0(IBP), T0(IL1) and T0(IL2). Prior to N₂ or CO₂ adsorption, samples were outgassed for 4 days under vacuum (10⁻³ mmHg) at 323 K. In such conditions, the decomposition of the adsorbed molecules is avoided (Lerdkanchanaporn and Dollimore, 2000).

The specific surface areas of the activated carbons were calculated using the Brunauer–Emmett–Teller (BET) equation, assuming the area of the N₂ molecule to be 0.162 nm². The micropore and mesopore volumes were calculated from the pore size distributions (PSD) of the activated carbons determined using NLDFT (nonlocal density functional theory) models applied on both the adsorption isotherms of N₂ at 77 K and CO₂ at 273 K. For that, infinite slit pores were assumed for CO₂ adsorption (pore diameters from 0.35 to 0.8 nm), while finite slit pores having a diameter-to-width aspect ratio of 12 were used for N₂ adsorption simulations (pore diameters from ∼0.45 to 25 nm) (Jagiello and Olivier, 2009). N₂ adsorption data at P/P₀ < 0.01 were obtained using incremental fixed doses of 10 cm³ g⁻¹ (STP), setting the equilibration interval at 300 s. CO₂ adsorption data were obtained at P/P₀ ranging from 4 × 10⁻⁴ to 3.5 × 10⁻², using 45 s equilibration interval (Jagiello and Thommes, 2004).

2.4 Adsorption kinetics and isotherms

The adsorption kinetics of the ILs were studied at room temperature at pH 3 and 7.5. The pH of 3 was adjusted by adding 0.1 mol L⁻¹ HCl to the ILs solutions prepared at 0.75 mmol L⁻¹ in Ultra High Quality water (UHQ, 18.2 MΩ). Experiments at pH 7.5 were performed by dissolving ILs in a phosphate buffer (prepared by dissolving 1.1796 g of KH₂PO₄ and 4.304 g of Na₂HPO₄ in 1L of UHQ water) (Weast et al., 1990).

Samples of activated carbon cloth (10 mg) were cut off and immerged in flasks containing 10 mL of acidified UHQ water (pH 3) or the buffer solution (pH 7.5). Closed flasks were stirred for at least one hour at 250 rpm in an orbital shaker (New Brunswick Scientific, Innova 40) in order to assure complete wetting of the adsorbent. After this preliminary step, 20 mL of the 0.75 mmol L⁻¹ IL solution was added to each flask to obtain an initial concentration of 0.50 mmol L⁻¹. For the kinetics the contact time was set between 2 min and 72 h.

The adsorption isotherms of both ILs and IBP were performed at pH 3 and 7.5 at three different temperatures (286 K, 298 K, and 313 K). 10 mg of the activated carbon cloth was previously impregnated with the solvent as described above. Solutions of given concentrations (from 0.05 to 0.5 mmol L⁻¹), derived from the mother solutions by dilution (with UHQ water for pH 3 or phosphate buffer for pH 7.5) were then added to the flasks and kept under stirring for 24 h.

The adsorption kinetics of a mixture of IBP, IL1 and IL2 were also studied on T0. A disk of T0 (10 mg) was placed in a phosphate buffered aqueous solution (pH 7.5, 298 K, 250 rpm) containing 0.4 mmol L⁻¹ of each adsorbate.

All the initial and residual adsorbent concentrations were measured by high performance liquid chromatography (HPLC) using a Waters chromatograph equipped with a high pressure pump (Waters 515), a photodiode array detector (Waters 996) and a Sunfire C18 column (5 µm, 4.6 × 250 mm). For IBP analysis, a methanol/UHQ water solution (80/20, v/v), containing 0.1 vol.% of concentrated phosphoric acid (H₃PO_4, 95 wt.%) in isocratic mode at a flow rate of 1 mL min⁻¹ was used as mobile phase. The ILs were analyzed in the same mode using a mobile phase composed either of 88 vol.% of acidified UHQ water and 12 vol.% of methanol for IL1 (1 vol.% of 95 wt.% H₃PO₄ added to the total volume), or of 85 vol.% of acidified UHQ water and 15 vol.% of methanol for IL2. Detection was operated at 220 nm for the IBP and 222 nm for the ILs.

The equilibrium uptake of each molecule Q_ads (mg g⁻¹) was calculated from equation: $$Q_{\mathit{ads}}=\frac{(C_i-C_e)·V}{m}$$ where V is the solution volume (L), C_i and C_e are respectively the initial and equilibrium concentrations of the adsorbent (mmol L⁻¹) and m is the mass of the dry activated carbon cloth (mg).

3.1 Characterization of the activated carbon cloth

The surface chemistry of the activated carbon cloth T0 was characterized by Boehm titration and the determination of pH and pHpzc. The result of Boehm titrations obtained in a previous study (Guedidi et al., 2017) revealed that T0 was mainly acidic with a pH (measured in boiled UHQ water) of 6.05. The basic value of pHpzc (point of zero charge) 8.40 might be assigned to the presence of high content of carbonyl groups (3.45 × 10⁻⁴) on its surface, as confirmed by Boehm titration (Table 1) (Guedidi et al., 2017). The porosity of the activated carbon cloth T0 was characterized by N₂ adsorption–desorption at 77 K and CO₂ adsorption at 273 K. As described in a previous study (Guedidi et al., 2017), the nitrogen adsorption isotherm is of type I and typical of microporous materials according to IUPAC classification (Fig. 2). The isotherm profile is characteristic of narrow microporous sample. As negative unrealistic C factors were obtained by applying the BET model in the 0.05–0.3 relative pressure range, the BET specific surface areas were then preferentially computed in the 0.01–0.05 relative pressure range, as for microporous materials. It is equal to 1910 m² g⁻¹. The pore size distribution (PSD) of T0 obtained by using a simulation method by NLDFT (Nonlocal density functional theory) model applied on the N₂ adsorption isotherm of T0 at 77 K (Fig. 7) confirms that T0 is mainly microporous in nature: the microporous volume is equal to 0.75 cm³ g⁻¹ whereas the mesoporous volume is only of ∼0.04 cm³ g⁻¹. As shown in our previous study, T0 contains two kinds of microporous: The amounts of ultramicropores (∅ < 0.7 nm) and supermicropores (0.7 < ∅ < 2 nm) determined by CO₂ and N₂ adsorption respectively (Guedidi et al., 2017) are roughly the same, i.e. of ∼0.31 cm³ g⁻¹.

Close

Figure 2

N₂ adsorption-desorption isotherms at 77 K of T0 (♢: adsorption, ♦: desorption).

Export to PPT

Close

Figure 7

Pore size distributions of T0 (♦), T0(IL1) (▴) and T0(IL2) (■) from DFT simulations by slit-pore models of N₂ (a) and CO₂ (b) isotherms.

Export to PPT

3.2 Chemical characterization of the ionic liquids

The ¹H NMR spectrum of IL1 exhibited peaks at the following chemical shift values from the TMS signal: 9.6 ppm (d, 2H, J = 7.6 Hz), 8.0 ppm (d, 2H, J = 7.2 Hz), 4.9 ppm (t, 2H, J = 7.4 Hz), 2.1 ppm (m, 2H), 1.4 ppm (m, 9H), 1.0 ppm (t, 3H, J = 6 Hz). A peak of very low intensity observed at 7.23 ppm was assigned to the presence of residual traces of CHCl₃ present in CDCl₃ (Gerothanassis et al., 2002). The ¹³C NMR spectrum of IL1 showed as expected 8 signals at the following chemical shifts: 145 ppm (2CH), 171.4 ppm (C), 124.7 ppm (2CH), 60.3 ppm (CH2), 34.2 ppm (C), 25.6 ppm (CH2), 31.3 ppm (3CH3) and 11.1 ppm (CH3).

The ¹H NMR spectrum of IL2 showed peaks at the following chemical shift values from the TMS signal: 9.4 ppm (d, 2H, J = 6.8 Hz), 8.0 ppm (d, 2H, J = 7.2 Hz), 5.1 ppm (t, 2H, J = 6.4 Hz), 3.3 ppm (t, 2H, J = 6.4 Hz), 1.4 ppm (m, 9H). A very low intensity signal at 7.23 ppm attributed to CHCl₃ was also observed on the IL2 spectrum. The absence of the expected singlet at ∼11 ppm that would confirm the presence of the carboxylic group can be explained by a quick exchange between the acidic proton and a deuterium from the CDCl₃ solvent. The 13C NMR spectrum of IL2 exhibited 8 peaks at the following chemical shifts: 171.9 ppm (C), 171,4 ppm (C), 145 ppm (2CH), 124.7 ppm (2CH), 50.3 ppm (C), 34.4 ppm (CH2), 34.2 ppm (C) and 31.3 ppm (3CH3).

The ¹H NMR and ¹³C NMR analyses have confirmed the structures and the sufficient purities of IL1 and IL2 (Williams and Fleming, 2007).

3.3 Adsorption kinetics of IL1 and IL2 on T0: effect of pH

The adsorption kinetic of IL1 is faster at pH 3 than at pH 7.5 (Fig. 3). Indeed, the time to reach equilibrium is of about 600 min and 900 min at pH 3 and pH 7.5 respectively. Moreover, the IL1 uptake is increased as the pH increases: it is of about 0.37 mmol g⁻¹ at pH 3 compared to 0.65 mmol g⁻¹ at pH 7.5. The adsorption decrease at acidic pH could be explained by a positively charged activated carbon surface (pH < pH_PZC) repelling the ionic liquid cation. The adsorption of the IL1 cation at low pH has been only promoted by dispersive interactions. As the pH of the solution has increased, the surface charges of the adsorbent have been neutralized and the adsorption of the ionic cation has consequently increased due to the reduction in the repulsive interaction.

Close

Figure 3

Adsorption kinetics of IL1, IL2 on T0 at pH 3 and 7.5 (298 K, [IL] = 0.5 mmol L⁻¹, mT0 = 10 mg, V = 30 mL, 250 rpm).

Export to PPT

For the IL2 adsorption, the time required to reach adsorption equilibrium is significantly the same at pH 3 and 7.5, i.e. about 900 min (Fig. 3). As for IL1, the activated carbon cloth adsorbed a greater amount of IL2 at pH 7.5 than at pH 3 (∼0.5 mmol g⁻¹ and 0.7 mmol g⁻¹ were adsorbed at pH 3 and 7.5 respectively). The lowest adsorption at acidic pH could be explained by repulsive interactions between the positively charged adsorbent and adsorbate (pKa IL2 = 4.1; pH < pKa IL2 < pH_PZC). On the other hand, at pH 7.5, the surface of T0 remained positively charged (pH < pH_PZC) but IL2 was present under a zwitterionic form due to the deprotonation of the carboxylic acid function. At this pH, the higher adsorption capacity of the activated carbon could be explained by both the electrostatic attractions (between the carboxylate function of IL2 and the positive charges present on the carbon surface) and some π – π dispersive interactions (between the pyridinium ion and the aromatic layers of T0). In a previous study on the adsorption of IBP in a methanol/water (10%v/90%v) solution by the T0 activated carbon cloth, a decrease in the IBP adsorption was observed as increasing pH (Guedidi et al., 2017). Maximum uptakes are about 2.4 mmol g⁻¹ and 0.87 mmol g⁻¹ at pH 3 and 7, respectively (for IBP initial concentration of 100 ppm) and higher than for ILs. Times to reach IBP adsorption equilibrium are of the same order as for the ILs, i.e., about 600 min at pH 3 and 1400 min at pH 7. In comparison with the ILs, the IBP adsorption is promoted at acidic pH because of its strongest hydrophobic nature in relation to its highest K_ow (octanol/water partition coefficient) value compared to IL1 and IL2 (Section 2.1). This suggests that at acidic pH, IBP adsorption is mainly governed by hydrophobe-hydrophobe and dispersive interactions between the non-dissociated molecule and the positively charged activated carbon, whereas at pH 7, the hydrophobe-hydrophobe interaction decreases as IBP is dissociated and it is dominant with respect to the electrostatic attractions.

The kinetics of adsorption of the two ionic liquids by T0 was investigated by four models, namely the Lagergren pseudo-first-order model, the pseudo-second-order model and the Elovich model. Table 5 in the Appendix shows the adsorption kinetic parameters calculated from the kinetic models. For both ILs, whatever the pH value, the Elovich model (Table 5 in Appendix) was not in good agreement with the experimental kinetics. For IL1, the best fits of adsorption kinetics at pH 3 and 7.5 (R² = 0.954 and 0.978 respectively) were obtained using the pseudo-second-order model. The IL1 uptake increased while the pH increased from 3 to 7.5. The pseudo-second-order constant k₂ obtained at pH = 3 is higher than at pH = 7.5. As a consequence, adsorption kinetics of IL1 is quicker when pH decreases. On the other hand, the pseudo first order model was the best one for fitting IL2 adsorption at pH 3 and 7.5 (R² = 0.989 and 0.995 respectively). The pseudo first order constant k₁ founded at pH 3 is higher than at pH 7.5 indicating that the adsorption kinetics of IL2 adsorption is quicker when pH decreases.

3.4 Adsorption isotherms

3.4.1

3.4.1 Effect of temperature

Whatever the initial pH value and for both ionic liquids, the adsorption capacity of T0 has decreased as increasing the temperature from 286 K to 313 K (Fig. 4a and b), indicating an exothermic phenomenon. The maximum IL1 uptake is reduced from 0.30 to 0.18 mmol g⁻¹ at pH 3 and from 0.65 to 0.60 mmol.g⁻¹ at pH 7.5 (Fig. 4a). A similar temperature variation is brought out for the adsorption of IL2 decreasing from 0.60 to 0.42 mmol g⁻¹ and from 0.73 to 0.63 mmol g⁻¹ at pH 3 and 7.5, respectively. However, the adsorption has been less influenced by the temperature at pH 7.5 than at pH 3 (Fig. 4a and b), maybe because it has been favored at pH 7.5 compared to pH 3. Indeed, the ILs’ uptakes were about two times higher at pH 7.5 than at acidic pH.

Close

Figure 4

Experimental adsorption isotherms of IL1 (a), IL2 (b) and IBP (c) on T0 at 286 K (♦, ♢), 298 K (■, □) and 313 K (▴, △) fitted by the Langmuir-Freundlich model (lines). Full and empty symbols correspond to isotherms at pH 7.5 and 3, respectively.

Export to PPT

The adsorption of IBP on T0 in the buffer solution (pH 7.5) has been also an exothermic phenomenon (Fig. 4c) as the IBP uptake has decreased from ∼2.30 mmol g⁻¹ at 286 K to ∼2.02 mmol g⁻¹ at 313 K. Exothermal phenomena were also evidenced for the adsorption of IBP from a methanol/water (10%v/90%v) solution on the same activated carbon cloth (T0) (Guedidi et al., 2017) as well as on a microporous-mesoporous granular activated carbon (Guedidi et al., 2013). The exothermic nature of the adsorption phenomenon remains consistent with the solute distribution between the aqueous phase and the activated carbon as the ILs and IBP solubilities increase together with the temperature. All the adsorption isotherms were tentatively reproduced with the Langmuir (Zhu et al., 2009), Freundlich, Langmuir-Freundlich (Yao, 2000) and Redlich-Peterson (Kumar and Porkodi, 2007) models, using an iterative procedure based on a nonlinear least-squares algorithm (with MATLAB® software).

Whatever the adsorbate and the working pH (Fig. 4), the experimental data were best fitted by the Langmuir-Freundlich model with R² ≥ 0.95 correlation coefficients (in Appendix Tables 6 and 7).

3.5 Ternary mixture adsorption kinetics of IBP, IL1 and IL2

The kinetic curves of the IBP and ILs mixture (pH 7.5, Fig. 5) display that the adsorption equilibrium is achieved much quicker for IL1 and IL2 (∼3000 min) than that for ibuprofen (∼7000 min). Times to reach equilibrium for each molecule of the ternary system are much longer than those previously determined for single adsorption (i.e., ∼900 min for IL1 and IL2). This difference indicates a competitive adsorption between the two ionic liquids and ibuprofen. However, T0 has had a higher affinity for ibuprofen than for both ILs due to its highest hydrophobicity as almost the whole initial amount of IBP has been adsorbed from the ternary mixture against only 60% and 10% of IL1 and IL2, respectively. This is also in agreement with the single-compound experiments for which the maximum adsorption uptakes at pH 7.5 are of about 2.20 mmol g⁻¹ for IBP (Fig. 4c) and only around 0.65 mmol g⁻¹ for both ILs (Fig. 4a and b). Adsorption behaviors of IL1 and IL2 in the ternary mixture have been quite different. Whereas a continuous decrease in C/C₀ was observed for IL1 until the plateau has been reached, IL2 was firstly adsorbed (during the first 500 min) and further desorbed to reach a plateau after about 3000 min. While IL2 was desorbed, IBP and IL1 continued diffusing toward the activated carbon adsorption sites (Fig. 5). This phenomenon might be related to the adsorbates’ molecular sizes which govern their ability to diffuse in the porous network of T0. Ibuprofen, which is the smallest molecule (0.53 nm³), could diffuse much faster in the porosity of T0, which is mainly microporous (Fig. 6), compared to the bulkiest IL1 (2.02 nm³) and IL2 (2.27 nm³) molecules. In addition, the lowest affinities of ILs for T0 at pH 7.5 can be explained by the electrostatic repulsions between the positive charge of the ILs cations and the positive surface of T0 (pH_PZC ∼ 8.4) on the one hand, and on the other by the hydrophilic nature of the ILs toward the hydrophobic T0 carbon surface. The presence of a carboxylate group on the IL2 molecule (pKa = 4.1) does not favor its adsorption at equilibrium suggesting that steric dimensions are key parameters for the adsorption. On the contrary, the highest IBP uptake could be justified by both the main contribution of hydrophobic-hydrophobic interactions and the one of the electrostatic attraction between this negatively charged adsorbate (pKa = 4.91) and the positive T0 surface. A recent study (Sellaoui et al., 2016) has demonstrated using the COSMO-RS model that the highest interaction energy of IL2 or IBP with the carbon surface is due to an additional contribution of hydrogen bonds and to the presence of a carboxyl group on these adsorbate molecules. Thus, IL2 and IBP could be adsorbed on similar sites having oxygen surfaces groups through a contribution of hydrogen bonding. This could explain that these adsorbates compete together and that IL2 was desorbed in the ternary kinetics study and mainly replaced by IBP. The strongest hydrophobicity of IBP might be the driving force that has promoted its substitution with IL2 (Fig. 5).

Close

Figure 5

Adsorption kinetics of IL1 (■), IL2 (▴) and IBP (♦) from a ternary mixture on T0 (pH 7.5, 298 K, initial concentration of 0.4 mmol L⁻¹ for each molecule, m_T0 = 10 mg, V = 30 mL, 250 rpm).

Export to PPT

Close

Figure 6

N₂ adsorption-desorption isotherms at 77 K of T0 (♦), T0 (IL1) (▴) and T0(IL2) (■). Curves of T0(IL1) and T0(IL2) are quite superimposed.

Export to PPT

3.6 Pore size required for ILs adsorption

A previous study has demonstrated that the IBP adsorption occurs in both the large ultramicropores and the supermicropores of the T0 activated carbon cloth (Guedidi et al., 2017). Indeed, gas adsorption experiments have suggested that the ultramicropore volume (0.190 cm³ g⁻¹, representing ∼61 vol.%) and supermicropores (0.070 cm³ g⁻¹, representing ∼21 vol.%) have been occupied by the IBP molecule (Table 4). Moreover, the occupied volumes deduced from the gas (N₂ and CO₂) adsorption isotherms and the IBP adsorption isotherm at pH 7.5 are quite similar (0.260 cm³ g⁻¹ and 0.200 cm³ g⁻¹, respectively, Table 3). The locations of the ILs molecules in the T0 porous network were investigated by comparing the porous volumes from the N₂ and CO₂ adsorption isotherms of T0 with the volumes occupied by ILs obtained from the loading content (Q_ads = 0.63 mmol g⁻¹ for IL1 and 0.73 mmol g⁻¹ for IL2). The N₂ adsorption-desorption isotherms of T0, T0(IL1) and T0(IL2) are all of type I, and typical of microporous materials according to the IUPAC classification (Fig. 6). The N₂-isotherms of the ILs loaded samples match the one of T0 but are shifted toward the lower adsorption volumes. Similar observation can be made for the CO₂ adsorption isotherms (not shown). The measured BET surface area decreased from 1910 m² g⁻¹ for T0 to ∼1320 m² g⁻¹ for T0(IL1) or T0(IL2), confirming the presence of the organic molecules in the porosity. A decrease in the total microporous volume from 0.75 cm³ g⁻¹ to ∼0.50 cm³ g⁻¹ was also measured after ILs adsorption (Table 4, Figs. 6 and 7a). The comparison of the ultra and supermicropore volumes before and after ILs’ adsorption has allowed calculating the volumes occupied by each IL in the porosity. They were roughly the same for IL1 and IL2 and represented about 0.125 cm³ g⁻¹ of the ultramicropores and 0.07 cm³ g⁻¹ of the supermicropores (Table 4). These values suggest that about 67% and 28% of the ultramicropore and supermicropore volume respectively, are occupied by the ILs (Fig. 7). However, for the two ILs, their occupied volumes calculated from their adsorption isotherms (at pH 7.5) were much higher (0.790 cm³ g⁻¹ for IL1 and 0.950 cm³ g⁻¹ for IL2) than the microporous occupied volumes determined by gas adsorption (0.514 cm³ g⁻¹ for IL1 and 0.503 cm³ g⁻¹ for IL2). This difference suggests that the micropores are not completely accessible by these large molecules. Another reason to explain these discrepancies is that the volumes determined for the ILs (2.02 nm³ for IL1 and 2.27 nm³ for IL2) were overestimated. ILs might have another configuration inside the pores due to electrostatic and dispersive interactions with the pore walls than the one taken for the calculation of their volume by modelization of the isolated molecules.