2.1 Materials

Parent ZSM-5 zeolite, with SiO₂/Al₂O₃ ratio of 50, was supplied by Zeolyst (CBV5524G). Salicylic acid (SA) and atenolol (ATL) were supplied by pharmaceutical company “Galenika”, Belgrade, Serbia, and diclofenac-Na (DFC) was purchased from Sigma, Aldrich.

Mesopore formation in ZSM-5 zeolite was achieved via alkaline treatment, using the method of Groen et al. (2004a, 2004b). Zeolite samples (5 g) were subjected to NaOH solution (0.2 M. 150 ml) during 30 min at the temperature of 333 K. Subsequently, the samples were filtrated, thoroughly washed with deionized water until a neutral pH was reached and dried at 393 K overnight. Following the alkaline treatment, an acidic treatment was preformed, using hydrochloric acid (0.1 M, 150 ml) for 6 h at room temperature. The samples were then filtered, washed and dried again. Both parent and modified ZSM-5 were afterwards converted to their ammonium form by three-fold ion-exchange (150 ml, 0.1 M ammonium nitrate at 363 K during 1 h). As prepared materials were filtered, washed with deionized water and dried in an oven at 393 K. The resulting samples were calcined in air at 773 K, for 5 h.

2.2 Characterization

Powder X-ray diffraction patterns were recorded on a Bruker (Siemens) 5005 diffractometer at room temperature using Cu K_α radiation (0.154 nm), from 3 to 80 2θ in 0.02° steps, with 1 s per step. Nitrogen adsorption was performed at 77 K; zeolite samples were pretreated in vacuum at 673 K for 4 h. BET and t-plot methods were used to calculate the textural properties. Transmission electron Microscopy (TEM) was carried out in a JEOL JEM-1400 Plus microscope operated at 120 kV. The sample was dispersed from ethanol onto a carbon-coated copper grid which was subsequently dried at room temperature. Solid-state ²⁷Al MAS NMR spectra were acquired on a Bruker MSL-400 (9.39 T) spectrometer upgraded with Apollo console (Tecmag, USA) operating at 104.27 MHz. MAS probes (7 mm) were used and the spinning speeds were 4.7 kHz. Single-pulse excitations of 2 μs were used for the ²⁷Al MAS NMR experiments. All the spectra were accumulated with a 3 s recycle time. Obtained chemical shift was determined in ppm relative to AlCl₃ 6H₂O solution as an external standard. Point of zero charge for both samples was evaluated using batch equilibrium method (Dakovic et al., 2010). Two series of KNO₃ solutions (0.1 M and 0.01 M) were prepared, each solution having a different pH value (pH_i), which was adjusted by addition of 0.1 M HNO₃ or 0.1 M KOH. 100 mg of a zeolitic sample was added to 50 ml of each solution and the mixtures were then stirred for 24 h at room temperature. The solid was separated from the solutions by centrifugation at 13,400 rpm during 20 min and the final pH values (pH_f) of the solutions were measured. The point of zero charge was determined as the pH value at which a plateau is obtained on the graph pH_f = f(pH_i). Acidity of the samples was investigated by microcalorimetric – volumetric measurements of ammonia adsorption. The heats of ammonia adsorption were measured in a heat-flow Tian-Calvet calorimeter (Setaram), linked to a glass volumetric line. Heats of adsorption and adsorption isotherms were determined, using a well-established method (Solinas and Ferino, 1998).

2.3 Adsorption measurements and liquid phase microcalorimetry

Adsorption of salicylic acid, diclofenac-Na and atenolol was studied using batch method (Damjanovic et al., 2010): approximately 50 mg of a sample was mixed with an appropriate volume of adsorbate solution in a sealed container using a magnetic stirrer during 1 h at 303 K. Solid adsorbent was subsequently separated from aqueous solution by centrifugation (Minispin, Eppendorf, 13,400 rpm, 0.5 h). Equilibrium adsorbate concentrations were measured using spectrophotometry (Shimadzu UV-1650PC). The adsorbed amounts (mol per gram of zeolite) were determined as the difference between the initial and equilibrium concentrations. Every point of an isotherm was obtained from a separate experiment. Measurements were done in triplicates. Initial solution concentrations were in the range 10⁻⁵ M to 10⁻³ M. The one hour equilibrium adsorption time was determined experimentally: for a fixed dose of an investigated adsorbate, adsorbed amounts were measured for different solid-liquid contact durations (15 min, 30 min, 1 h, 2 h and 6 h). The values of concentrations that were found remaining in solutions after 2 h or 6 h of stirring were found to be the same as in the case of 1 h long contact time.

Experimental data were fitted to the Langmuir, Freundlich, Sips and Toth equations using Microcal Origin 8 software; the corresponding constants were calculated. Eqs. (1)–(4) present Langmuir, Freundlich, Sips and Toth equations, respectively (Jaroniec and Marczewski, 1984):

The heats of adsorption of investigated adsorbates were determined using liquid calorimetric measurements. A differential heat flow reaction calorimeter (TITRYS, from Setaram) equipped with a stirring system was used. The calorimeter was linked by capillary tubes to a system for liquid injection (a programmable syringe pump PHD 2000, from Harvard Apparatus). Successive pulse injections of 0.2 ml of adsorbate solutions were sent to the sample in 2 h time intervals. The temperature of the calorimetric cell was maintained at 303 K. Furthermore, each dose of the liquid was equilibrated at the same temperature in a preheating furnace prior to being injected in order to avoid enthalpy effects due to mixing phenomena. All liquids used in this investigation were previously purged with flowing nitrogen as this procedure, based on experimental experience, prevents formation of air bubbles in the capillary tubes. At the beginning of the experiment, the reference and measurement cells contained water (1.5 cm³) and water with a weighted amount of the solid sample (ca. 50 mg), respectively. Initial solution concentrations of atenolol, diclofenac sodium and salicylic acid were the same as in the previously described adsorption experiments. Hence, the calorimetric experiments were used to directly measure the heats which evolved due to adsorption (expressed as J g⁻¹ of solid), for each admitted dose. Using previously determined corresponding adsorbed amounts, differential heats of adsorption (Q_dif, expressed in kJ mol⁻¹) were calculated.

3.1 Characterization

The alterations of pore structure of zeolite ZSM-5, caused by the alkaline treatment, were investigated using low temperature nitrogen adsorption (Fig. 2). The modified sample shows a more than two-fold increase in the total amount of nitrogen adsorbed compared to the parent zeolite. The isotherm of the parent sample consists of an adsorption branch of type I and a H4 type hysteresis, typical for microporous materials (Sing et al., 1985). The modified sample exhibits an isotherm of type IIb, with a steep slope at higher values of p/p⁰, and a H3 hysteresis, which is in accordance with the results obtained by other authors for hierarchical zeolites (Fernandez et al., 2010; Wu et al., 2012). Alkaline treatment of ZSM-5 resulted in an increase of the total pore volume (from 0.236 to 0.449 cm³ g⁻¹), mesopore volume (from 0.085 to 0.380 cm³ g⁻¹), BET surface (from 365 to 470 m² g⁻¹) and external surface (65 to 173 m² g⁻¹), which is attributed to mesopore formation (Groen et al., 2004b). Microporosity of the zeolite was preserved to a high degree: only a slight decrease of the micropore volume was measured (from 0.144 to 0.138 cm³ g⁻¹). The pore size distribution of both samples is presented in Fig. 2 (inset). The modified ZSM-5 sample exhibits a fairly broad peak centered at 14 nm. Formed mesopores and differences in pore structure between parent and modified zeolite are clearly visible in TEM micrographs (Fig. 3).

Close

Fig. 2 Isotherms of nitrogen adsorption and pore size distributions of parent and modified ZSM-5 zeolite.

Export to PPT

Close

Fig. 3 TEM micrographs of (a) microporous and (b) hierarchical ZSM-5.

Export to PPT

The influence of the alkaline treatment on the acidity of the zeolite structure was investigated by coupled microcalorimetry and volumetry of ammonia adsorption. Fig. 4 presents the site strength distributions for the investigated samples. The results show that the shapes of distributions are generally the same for both samples, consisting of two major domains, centered at ∼150 and 75–80 kJ mol⁻¹. Compared to the parent zeolite, there is a slight decrease of the number of strong (∼150 kJ mol⁻¹) and weak (∼75 kJ mol⁻¹) acid sites for the hierarchical sample, while an increase of the number of sites with intermediate strength (85–125 kJ mol⁻¹) is also noticeable. The observed changes in the distribution of strength of the acid sites are not drastic and it can be concluded that the acidity of ZSM-5 remained unaffected to a large extent by the modification procedure. A detailed analysis of the obtained calorimetric results is presented elsewhere (Rac et al., 2013).

Close

Fig. 4 Distribution of strength of acid sites in parent and modified ZSM-5 zeolite.

Export to PPT

The coordination state of Al in both zeolite samples was estimated by solid-state ²⁷Al MAS NMR spectra (not presented here). The obtained results show that both the parent and the modified sample predominantly contain tetrahedral Al (peak at 54 ppm), which is attributed to framework aluminum. Some extraframework Al species, octahedrally coordinated (peak at 0 ppm), were also found in the spectra of both samples. These results indicate that desilication did not result in significant extraframework Al buildup. It is here important to point out that the total SiO₂/Al₂O₃ ratio, measured by chemical analysis, changed from 50 to 43 after the alkaline treatment, but its value was restored to 50 again, after the subsequent acid wash. This is in accordance with data obtained by Mitchell and Pérez-Ramírez (2011), who emphasized the importance of acid wash treatment, which removes non-framework aluminum resulting from alkaline treatment and recovers Si/Al composition similar to that of the starting material.

The point of zero charge (PZC) for both parent and mesoporous sample was determined by batch equilibrium method (Fig. 5). The obtained results show that the PZC value is 2.95 ± 0.05 for parent and 3.05 ± 0.05 for mesoporous ZSM-5, indicating that the introduction of mesoporosity in ZSM-5 did not lead to any significant change in surface charge.

Close

Fig. 5 pH_f versus pH_i for parent and modified ZSM-5 zeolite; PZC determination.

Export to PPT

XRD patterns were recorded in order to estimate the influence of desilication on the crystallinity of ZSM-5. The results (not presented here) show that the diffraction patterns typical for ZSM-5 structure were recorded for both samples and that the desilicated sample preserved its crystalline structure. However, the degree of crystallinity was altered in the modified zeolite. The variation in crystallinity of the modified zeolite was calculated as proposed in literature (Martınez et al., 2013; Auroux et al., 1996), from the intensities of the reflection at 23° 2θ, taking the crystallinity of the parent sample to be 100%. The crystallinity of the mesoporous ZSM-5 was found to be 76%.

3.2 Adsorption measurements and liquid phase microcalorimetry

Adsorption experiments revealed a distinct difference in the way mesopore creation in ZSM-5 affected its adsorptive behavior towards salicylic acid, atenolol and diclofenac. In the case of salicylic acid there is a slight decrease of the overall adsorptive capability for the mesoporous sample, compared to the parent one (Fig. 6). However, this decrease is more pronounced in the region of lower equilibrium concentrations while maximal amounts of SA adsorbed on microporous and hierarchical samples, calculated using Langmuir, Sips and Toth models are similar (Table 1). The best fit of the experimental values was obtained using Toth equation and, in this case, the extrapolated maximal adsorptive capacities for parent and modified sample are almost the same, 0.00021 mol g⁻¹ and 0.00022 mol g⁻¹, respectively. These results are in accordance with the fact that the molecular diameter of SA molecules, 4.6 Å (Kirumakki et al., 2003), enables them to enter into the micropores of ZSM-5. Consequently, no positive effect of mesopore existence in the structure on the adsorption capacity for SA should be expected. In fact, our results indicate that bonding of salicylic acid to zeolitic inner surface is somewhat favored in narrow micropores.

Close

Fig. 6 Adsorption isotherms of salicylic acid, diclofenac sodium and atenolol.

Export to PPT

In the case of atenolol and diclofenac sodium adsorption, the effect of mesopore formation is quite opposite: an increase of the maximum adsorbed amount of both molecules is observed for the alkaline treated sample compared to the parent (Fig. 6). This effect is moderate for diclofenac sodium and very pronounced for atenolol, where there is a more than threefold increase of the maximal adsorbed amount. Obviously, the existence of mesopores in the desilicated sample enables more of the adsorbate molecules to enter the zeolite structure and interact with the active sites to a much larger extent than in the case of the parent sample. The difference in the enhancement of the adsorption capability towards diclofenac and atenolol is probably caused by the structure and shape of the molecules in question. The difference in dimension between the two molecules (7 × 18.2 Å for ATL (Fagerholm et al., 1999) and 3.5 × 8.3 × 7.0 Å for DFC (Xie et al., 2012)) can also play an important role, considering the zigzag channel structure of ZSM-5 zeolite. Since the acidity of the zeolite was only mildly affected by the mesopore formation process and the surface charge remained unchanged, the presumption that the geometry of the pores and molecules is a crucial factor can be expressed.

Hence, the observed increases in adsorption capabilities toward atenolol and diclofenac seem to be the result of the availability of the adsorption sites to the adsorbing molecules, which was the parameter that was actually changed by the mesopore formation process. The molecules in the hierarchical zeolite are able to bond to the adsorption sites which are inaccessible to them in the parent sample. The extent of the adsorption capability enhancement depends on the shape and the diameter of both mesopores and the adsorbing molecules. Fitting of the experimental isotherms to theoretical models showed that the use of Toth equation resulted in best agreement, for all adsorbent – adsorbate combinations (Table 1). However, in some cases, Sips and Langmuir models also adequately fitted empirical data. It is, at this point, important to emphasize that good agreement between experimental values and a given model is very useful for interpretation and analysis of the results, but should not be used as definite proof of the theoretical postulates the model is based upon (Adamson, 1990). Nevertheless, further analysis of the results obtained using Toth equation allowed us to easily evaluate and compare adsorptive capabilities of the investigated zeolites at low concentrations. Toth equation reduces to linear shape in the region of low concentrations. Indeed, this is a property that adequately depicts the real situation: when the concentration of the adsorbate is very small, the available concentration of adsorption sites within the zeolite is, relatively, so large that it can be regarded as “infinite” in the sense that its change during the adsorption process is negligible. Hence, at these conditions, the adsorbed amount becomes solely dependent of the adsorbate concentration and the isotherm is theoretically expected to be linear. Toth equation becomes linear when K_Tc_eq ≪ 1, which reduces Eq (4) to

Using Eq (5) and data in Table 1, it can be shown that, for a given c_eq, S_e of hierarchical sample is approximately 8 times larger than S_e of microporous sample, in the case of atenolol adsorption. Hence, superior adsorptive capabilities of modified ZSM-5 towards atenolol are expected to be even more pronounced in low concentration situations.

Aside from Eq (5), which is specific for a given adsorbate – adsorbent couple, another mathematical relation between adsorbed amount (S_e) and equilibrium concentration (c_eq) exists:

Close

Fig. 7 Removal efficiencies for atenolol at low initial concentrations for different m/V.

Export to PPT

To further investigate the interaction between the active sites of the zeolite samples and the adsorbing molecules, liquid calorimetry measurements were conducted. For each of the molecules, calorimetric experiments were performed in the concentration range corresponding to the part of the isotherm before the saturation plateau, where the strongest binding is expected. In the case of atenolol adsorption, a remarkable increase of the strength of interaction was detected for the hierarchical sample compared to the parent zeolite (Fig. 8). The heats of adsorption in the initial regions of the two calorimetric curves show almost a 30 kJ mol⁻¹ difference. Considering the previously described distribution of strength of the active sites of the two zeolite samples, this increase of the atenolol – hierarchical ZSM-5 interactions can be interpreted as a result of enhanced accessibility of strong adsorption sites within the zeolite structure to atenolol molecules. However, this increase is limited to the initial region of the calorimetric curve after which the strength of the interaction becomes equal to that detected for the parent sample. It seems that, in the case of parent microporous sample, atenolol can be adsorbed only at its surface; since the dimensions of its molecules prevent them from entering three-dimensional interconnected channel system of ZSM-5 zeolite. Evidently, the formation of mesopores enabled ATL molecules to reach adsorption sites stronger than those existing at the surface of zeolite.

Close

Fig. 8 Differential heats of adsorption of atenolol obtained by liquid calorimetry.

Export to PPT

For the two other investigated molecules, such a drastic change in the strength of interactions with the zeolite samples (microporous and hierarchical) was not detected. The measured heats of adsorption for diclofenac sodium were approximately 45 kJ mol⁻¹. The calorimetric effects for salicylic acid adsorption were below detection limit, since in this case, we did not obtain signals which differed significantly from the baseline of the calorimeter. Generally, the existence of basic or acidic groups in the adsorbing molecules determines their strength of interaction with ZSM-5. Evidently, the interactions between zeolitic acid sites and basic groups of atenolol molecules result in the strongest bonds. The values of differential heats of DFC adsorption indicate somewhat weaker interaction with ZSM-5 structure, which is in agreement with weakly basic character of the molecule. In the case of salicylic acid, it is reasonable to assume that its acidity (pKa = 2.97) prevents strong adsorption. Obviously, salicylic acid interacts with ZSM-5 structure forming very weak bonds so the heats of adsorption could not be detected.

It has to be underlined here that the charge of an adsorbate in a solution (i.e. weather the molecules are dissociated, protonated or not) as well as the surface charge of the adsorbent are expected to be important factors in adsorption processes. However, although the state of a molecule in a solution is defined by its pK_a value and the pH of the solution, the situation is not as straightforward when conditions within the zeolite are concerned, where we can have no certainty about the local pH (Chen and Sachtler, 1998). Yousef and El-Eswed (2009) have even found increasing capacities for phenol adsorption on natural zeolite with increasing pH, even though the phenolate anion and the zeolite surface were both negatively charged. Considering the previously discussed results obtained for PZC values of the investigated samples, it can be concluded that the differences in adsorption capacities measured for the two zeolitic samples cannot be attributed to changes in total surface charge of the solid. Since the PZC values of both zeolites are almost identical, it is safe to state that the adsorbing molecules are exposed to the same surface charge during adsorption on either parent or hierarchical ZSM-5. Hence, electrostatic attraction/repulsion cannot be considered to be the reason behind the observed differences in adsorption capacities between the two samples. The work presented in this paper shows that the presence and strength of basic groups in a pollutant molecule and its size and geometry are the main factors governing adsorption on ZSM-5 zeolite and that induced mesoporosity can significantly improve ZSM-5 usability in depollution processes, when molecules with diameters comparable to or larger than micropore openings are concerned.