This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
The removal of five selected pharmaceuticals, viz., carbamazepine, clofibric acid, diclofenac, ibuprofen, and ketoprofen was examined by batch sorption experiments onto a synthesized mesoporous silica SBA-15. SBA-15 was synthesized and characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), N2 adsorption-desorption measurement, and point of zero charge (PZC) measurement. Pharmaceutical adsorption kinetics was rapid and occurred on a scale of minutes, following a pseudo-second order rate expression. Adsorption isotherms were best fitted by the Freundlich isotherm model. High removal rates of individual pharmaceuticals were achieved in acidic media (pH 3-5) and reached 85.2% for carbamazepine, 88.3% for diclofenac, 93.0% for ibuprofen, 94.3% for ketoprofen, and 49.0% for clofibric acid at pH 3 but decreased with increase in pH. SBA-15 also showed high efficiency for removal of a mixture of 5 pharmaceuticals. Except for clofibric acid (35.6%), the removal of pharmaceuticals in the mixture ranged from 75.2 to 89.3%. Based on adsorption and desorption results, the mechanism of the selected pharmaceuticals was found to be a hydrophilic interaction, providing valuable information for further studies to design materials for the purpose. The results of this study suggest that mesoporous-silica- based materials are promising adsorbents for removing pharmaceuticals not only from surface water but also wastewater of pharmaceutical industrial manufactures.
Key words: Removal, Adsorption, Mesoporous silica SBA-15, Pharmaceuticals, Surface water.
It has been almost 10 years since scientists paid attention to pharmaceuticals as new emerging contaminants [1,2]. Numerous pharmaceuticals were found ubiquitously up to µg L-1 in aqueous environment [1-5], in soil [1,6] and as high as 100 µg L-1 in effluent from drug manufactures . The presence of pharmaceuticals in the environment originates from the fact that these contaminants cannot be removed completely in sewage treatment plants (STPs) [2,8] (removal efficiency of carbamazepine and clofibric acid was 7%  and 0-30% , respectively) and are persistent in the environment [9,10]. In general, total removal rates of pharmaceuticals in current STPs were 40 - 60% [2,8]. In drinking water treatment plants (DWTP), incomplete elimination of some pharmaceuticals such as clofibric acid was also observed [11,12]. Additionally, adverse effects of pharmaceuticals at µg L-1 - level on living organisms have been recently confirmed  and possibly there exist unknown chronic effects of pharmaceuticals, which have not been investigated yet. There is a need to optimize and improve current technologies used in STP and DWTP for elimination of pharmaceutical residues, especially for treatment of pharmaceutical industrial wastewater and purifying surface water in DWTP.
Removal of pharmaceuticals by adsorption is one of the most promising techniques, due to its convenience once applied into current water treatment processes. So far, the removal of pharmaceuticals can be achieved by adsorption using activated carbon [11-13]. Although, activated carbon displayed efficient removal for a number of pharmaceuticals, especially for hydrophobic compounds, inefficient removal for pharmaceuticals which are either electrically charged or hydrophilic has been observed [11-13]. The working capacity of activated carbon greatly decreases in the presence of natural organic matter. Furthermore, regeneration of adsorbents is questionable.
Mesoporous silica such as MCM-41, MCM-48 , SBA-15  can be a promising adsorbent owing to its novel structure comprised of uniform ordered structure, high pore volume, and high surface area (up to 1000 m2/g). Mesoporous materials have been investigated for adsorptive removal of various organics such as cyanuric acid and p-chlorophenol , phenols , and dyes [18-20] in aqueous solution. Recently, removal of naproxen on a MCM-41 incorporated with nickel(II) complexes has been reported . However, only single component adsorption at mg L-1-level concentration was investigated, while adsorption kinetic, isotherm, desorption results as well as effect of pH and chemical factors such as competing anions, humic acid on adsorption have been not considered.
In this study, SBA-15 was first attempted to apply for removal of five selected pharmaceuticals at µg L-1 concentrations. SBA-15 was selected because of its benign synthesis pathway (due to employing biodegradable non-ionic surfactant), easily reproducible method and large uniform pore structure . The objective of the study is to (i) determine the removal efficiency of individual pharmaceuticals as well as their mixture on SBA-15, (ii) investigate the characteristics of pharmaceutical adsorption via kinetic, isotherm, desorption, dose effect, and pH effect results and (ii) understand the adsorption mechanism which is beneficial for further studies. As a part of this study, effect of anions (nitrate, phosphate, sulfate, silicate), cations (aluminum, ferric, calcium, magnesium), and humic acid on pharmaceutical adsorption have been also investigated and will be mentioned in another report.
2. Materials and Methods
The reagent grade chemicals used in the study (KCl, NaCl, HCl, NaOH) were purchased from Aldrich and Junsei chemical companies. Tetraethyl orthosilicate (TEOS) and Pluronic P123 (EO20PO70EO20, Mav = 5800) were obtained from Junsei, BASF, respectively.
2.2. Adsorbent and target compounds
SBA-15 was synthesized by the method described elsewhere . Briefly, a mixture solution of Pluronic P123 and TEOS was prepared in acidic media. Subsequently, the mixture was kept at 35oC for 20 h and aged at 100oC for 1 day. Finally, the sample was washed with de-ionized (DI) water and calcined in air at 550oC for 6 h. The synthesized material was characterized by X-ray powder diffraction (XRD) using a Rigaku D/MAX Uitima III diffractometer with Cu KÎ± radiation. Transmission electron microscopy (TEM) images of the materials were acquired on a JEOL 2010 electron microscope operating at 200 kV. Nitrogen adsorption-desorption measurement were carried out using a Micromeritics ASAP 2020 analyzer at 77 K. The point of zero charge (PZC) of SBA-15 was estimated by the titration method as described elsewhere .
Five pharmaceuticals were selected in this research as target compounds: carbamazepine (an antiepileptic), clofibric acid (a lipid regulator), diclofenac, ibuprofen, and ketoprofen (nonsteroidal anti-inflammatory drugs) and two compounds were utilized as surrogate standards: cloprop (2-(3-chlorophenoxy) propionic acid) and 10,11-dihydrocarbamazepine (DHC); all were purchased from Aldrich. These pharmaceuticals were selected because they are either widely used and detected [1-5], resilient during treatment processes [2,8,11,12], or persistent [9,10] in the aqueous environment. Moreover, they have been evaluated to possess high environmental risk (Risk Quotient > 1) . The molecular structure and physicochemical properties of these pharmaceuticals are summarized in Table 1.
Stock solutions of pharmaceuticals were prepared as 10 mg L-1 by first dissolved in methanol and then added with DI water. The amount of methanol was less than 0.1% in the final samples and thus had negligible impact on sorption. Stock solutions were stored at 4oC, and used within one month after preparation.
2.3. Batch adsorption and desorption experiments
Batch adsorption of individual pharmaceuticals was studied in 40-mL amber vials containing 10 mg of SBA-15 in 10 mL of pharmaceuticals solution in 0.01 M KCl. Reactions proceeded for 2 h at 25oC in an incubator (200 rpm). The 2 h reaction time was shown to be adequate by preliminary experiments for equilibrium to be attained (Fig. 1). After 2 h, the supernatant solution was filtered through a 0.45-µm cellulose acetate membrane filter (MFS), subsequently extracted by a solid phase extraction (SPE) procedure and analyzed by a LC-MS/MS system. All experiments were performed in duplicate, and the solutions were analyzed triplicate for pharmaceuticals by LC-MS/MS.
Adsorption kinetics was resulted by taking the samples after different intervals (1 - 120 min) with a fixed pharmaceutical starting concentration (100 µg L-1). Adsorption isotherms were produced by varying the pharmaceutical starting concentration from 10 - 300 µg L-1 and the pH-dependent adsorption were generated with a fixed pharmaceuticals concentration (100 µg L-1) and varying pH from 3 - 9. The solution pH was adjusted using 0.1 M NaOH or HCl to the desired pH value, measured by Orion (model 3 Star) pH meter. Except for the experiment to estimate the effect of pH, the solution pH was kept at pH 3 for clofibric acid and pH 5 for the other ones (unless otherwise specified). Regarding effect of dosage, mass of SBA-15 was varied from 1 mg to 20 mg in 10 mL of pharmaceutical solution at a fixed pharmaceutical concentration (100 µg L-1).
Desorption of pharmaceuticals from SBA-15 was investigated to study the interaction between pharmaceuticals and silica surface. First, adsorption of individual pharmaceuticals (100 µg L-1) on SBA-15 (1 g L-1) in 0.01 M KCl was conducted. After 2 h, the pharmaceutical-adsorbed SBA-15 was centrifuged, separated and gently washed with water to remove aqueous pharmaceuticals. The pharmaceutical-adsorbed SBA-15 was agitated for 12 h with 10 mL of 1 mM phosphate buffer (pH 7 and 9). Aliquots of supernatant solution were filtered (0.45 µm), extracted and analyzed by the LC-MS/MS system.
Adsorption of a mixture of 5 selected pharmaceuticals was conducted by introducing 10 mg SBA-15 into 10 mL solution of 5 pharmaceuticals at the same concentration (25 µg L-1) at pH 5. The samples were taken after 5, 15, 30, 60, 120, 480, and 1440 min. Aliquots of supernatant solution were filtered (0.45 µm), extracted and analyzed by the LC-MS/MS system.
2.4. Sample preparation and analytical methods
Before analysis, the samples obtained after filtration were extracted further by a SPE procedure using 60 mg Oasis HLB cartridges. Then, pharmaceuticals were analyzed on a LC-MS/MS system equipped with a HPLC Waters 2695 Separations Module accompanied with Waters 2996 PAD, a mass spectrometer (Waters (Micromass) Quattro micro API), and a XTerra MS C18 5 µm column (Waters, 2.1 Ã- 50 mm). (The experimental details of sample preparation and analytical methods are present in Supplementary Information.)
3. Results and discussions
3.1. Characterization of SBA-15
The XRD pattern of SBA-15 sample is shown in Fig. S1 (Supplementary Information), which exhibits three well-resolved peaks at 2q between 0.5 and 3o, characterizing a well-ordered mesoporous structure with P6mm hexagonal symmetry . The well-ordered hexagonal arrays of mesopores are also confirmed by the TEM image (Fig. S2, Supplementary Information) with high uniformity and the pore size evaluated ~ 8 nm. Based on N2 adsorption-desorption measurement, the Brunauer-Emmett-Teller (BET) specific surface area, pore volume, and pore size of SBA-15 was estimated at 737 m2/g, 1.03 cm3/g, and ~ 8 nm, respectively, which is consistent with the TEM image and the reported data . PZC of SBA-15 was estimated from the intersection point of three titration curves obtained at different concentrations of the inert electrolyte NaCl. PZC of SBA-15 was determined as 4.0 ± 0.1, which is similar to the isoelectric point of a mesoporous silica MCM-41 (3.6±0.18) .
3.2. Kinetics of pharmaceutical adsorption
Adsorption of all pharmaceuticals reaches equilibrium in very short time (< 15 min), even in 1 min (Fig. 1). The contact time needed for mesoporous silica is significantly shorter than the adequate time (4 h) for adsorption of pharmaceuticals on activated carbon [12,25], which can be resulted from the well-ordered mesoporous structure of SBA-15 rather than a disordered microporous structure of activated carbons. As mentioned above, SBA-15 possesses a system of uniform and large pores that may allow pharmaceutical molecules to diffuse easily into the pores and adsorb on the surface in a short time.
The adsorption kinetic data were examined using a pseudo-second-order reaction kinetics expression :
where k2 (g mg-1 min-1) is the equilibrium rate constant of pseudo-second order chemical sorption; qe is the amount of adsorbate sorbed at equilibrium (mg g-1); q is the amount of adsorbate sorbed at t (min). The straight line plots of (t/q) versus t have been tested to obtain rate parameters. Fig. S3 (Supplementary Information) shows good compliance with the pseudo-second order equation for all pharmaceuticals, which is reflected by the extremely high determination coefficients (R2 > 0.997) of the linear plots. Interestingly, adsorption of ibuprofen on activated carbon from aqueous phase also obeyed the pseudo-second order equation .
3.3. Adsorption isotherms
From the various isotherm equations that may be used to analyze adsorption data in aqueous phase, the Langmuir  - the theoretical equilibrium isotherm and the Freundlich  - the empirical equilibrium isotherm are the most common models. The linear forms of these equations are displayed as equation 2 (Langmuir) and 3 (Freundlich):
where qm (mg g-1) is the maximum adsorption capacity, qe (mg g-1) is the amount of adsorbed pharmaceuticals, Ce (mg L-1) is the equilibrium pharmaceutical concentration, KF and n are the Freundlich constants, and KL (L mg-1) is the Langmuir constant. The linear Langmuir and Freundlich isotherms were fitted to the experimental data. The Langmuir and Freundlich parameters, along with the coefficients of determination (R2) of the linear plots, are presented in Table 2. Adsorption of pharmaceuticals on SBA-15 can be fitted by both Langmuir and Freundlich model with R2 > 0.91; however, only Freundlich curves are shown in Fig. 2. In a recent study, Ho  discussed the advantages of using non-linear Chi-square analysis (c2) to compare the fitting of experimental data to isotherm models. c2 (mg g-1) determined using the following equation:
where qe (mg g-1) is the experimental equilibrium capacity and qe,m (mg g-1) is the equilibrium capacity obtained by calculated from the models. As Ho suggested, the smaller c2 values display the better match of the model . In Table 2, the results of c2 are also presented. c2 values of the Freundlich isotherm for all pharmaceuticals were lower than those of the Langmuir isotherm, indicating that the Freundlich isotherm is the better fitting isotherm for adsorption of pharmaceuticals on SBA-15. As reported earlier, adsorption of antibiotic ofloxacin and tetracycline onto mesoporous silica followed Langmuir-Freundlich model  and a dual site adsorption model , respectively.
In general, as the KF value increases, the adsorption capacity of the adsorbent for a given pharmaceuticals increases, reflecting partially the affinity of pharmaceuticals toward the adsorbent. From that view point, affinity of pharmaceuticals with SBA-15 follows the order: ibuprofen > carbamazepine ~ ketoprofen > diclofenac > clofibric acid. For acidic pharmaceuticals, this affinity order obeys well their pKa values: ibuprofen (pKa = 4.91) > ketoprofen (pKa = 4.45) > diclofenac (pKa = 4.15) > clofibric acid (pKa = 2.84). More explanation about adsorption mechanism will be discussed in detail later.
3.4. Effect of dosage
The removal efficiency of pharmaceuticals increases significantly as the adsorbent concentration was increased from 0.1 to 2.0 g/L (Fig. 3). At 2.0 g/L SBA-15 dosage, removal of clofibric acid, diclofenac, carbamazepine, ibuprofen, and ketoprofen can be achieved 62.5, 66.7, 84.4, 95.1, and 91.2%, respectively. In the range of 0.1 - 1.0 g/L SBA-15 dosage, pharmaceutical adsorption increased almost linearly with adsorbent concentration, but the increase of adsorbent dosage from 1.0 to 2.0 g/L does not make a significant effect. By increasing the dosage of SBA-15 from 1.0 to 2.0 g/L, the removal of pharmaceuticals was increased only from 6 - 16%.
3.5. Effect of pH on pharmaceutical adsorption
The removal efficiencies of pharmaceuticals on SBA-15 presented in Fig. 4 shows that, as expected, adsorption of pharmaceuticals is strongly dependent on pH. Adsorption of pharmaceuticals onto SBA-15 can be formed and controlled by both non-electrostatic and electrostatic interaction. In the case of carbamazepine (pKa = 13.90, Table 1), a neutral compound in the tested pH range, its binding onto SBA-15 is solely responsible by a non-electrostatic interaction involving hydrogen bonding . This hydrogen bonding is preferred at pH lower than 4.0 (PZC of SBA-15), at which silica surface has a net positive charge; however, it becomes less favorable on negative silica surface as pH higher than 4.0, resulting a decrease in adsorption of carbamazepine. Meanwhile, uptake of acidic pharmaceuticals (clofibric acid, diclofenac, ibuprofen, and ketoprofen) on silica surface is attributed to both kinds of interaction, dependent on solution pH. At pH below the pKa, acidic pharmaceuticals are neutral molecules, interacting with silica surface via non-electrostatic interaction and its behavior is similar to carbamazepine. When the pH is above the pKa (pH > 3 with clofibric acid and 5 with the others), acidic pharmaceuticals have negative charge while surface of silica gradually becomes more negatively charged, leading to an electrostatic repulsion between them. Consequently, adsorption of acidic pharmaceuticals dropped sharply. Furthermore, when pH is above 7, acidic pharmaceutical molecules and surface groups on SBA-15 completely become negatively charged, leading to no significant adsorption.
The experimental results also displayed that, for carbamazepine, diclofenac, ibuprofen, and ketoprofen, the removal efficiency is as high as 90% at pH 3 and about 75.6 - 81.0% (except for diclofenac) at pH 5. The removal of clofibric acid is only 49.0% at pH 3, however it is still higher than that in case of using high ozone doses (2.5-3.0 mg/L) (< 40%) . When pH was increased from 3 to 9, the removal of carbamazepine reduced gradually from 85.2% to 40.5%, while removal of acidic pharmaceuticals declined significantly and was almost negligible at pH 9.
In comparison among acidic pharmaceuticals, clofibric acid (pKa = 2.84) showed the lowest removal rates, followed by diclofenac (pKa = 4.15) (Fig. 4), because of their lower pKa values, compared to that of ibuprofen (pKa = 4.91) and ketoprofen (pKa = 4.45). The lower - pKa compounds are more readily deprotonated to become negatively charged. Therefore, more molecules exist in anion forms at the same pH, leading less favorable adsorption on silica surface.
At pH 3, carbamazepine, diclofenac, ibuprofen, and ketoprofen are all neutral molecules and showed quite similar adsorption (see Fig. 4), in spite of their differences in hydrophobicity (see Table 1). This result suggests that adsorption of the studied pharmaceuticals on mesoporous silica is mainly controlled by hydrophilic interactions between functional groups in pharmaceutical molecules and silanol groups on silica surface, rather than a hydrophobic sorption .
3.6. Desorption studies
Desorption of pharmaceuticals adsorbed on SBA-15 was performed in neutral (phosphate buffer at pH 7) and alkaline media (phosphate buffer at pH 9). Aforementioned, adsorption percentage of acidic pharmaceuticals on SBA-15 is almost negligible at pH 9 and less than 10% at pH 7, as such desorption of pharmaceuticals under these conditions expected to be almost complete. However, as shown in Fig. 5, desorption percentage of 5 pharmaceuticals varied from 20 - 40%. The results suggest that the adsorption is not completely reversible. In the other words, pharmaceuticals can be both reversibly and irreversibly adsorbed on the silica surface , which can be the result of heterogeneity of the silica surface (several types of silanol groups) and multifunctional nature of the adsorbates. As a result, there might be strong interactions between pharmaceuticals and silica surface, corresponding to irreversible adsorbed pharmaceuticals.
3.7. Mechanism of pharmaceutical adsorption onto SBA-15
The strongly pH-dependent adsorption of pharmaceuticals proposes that the interaction between pharmaceuticals and surface of the mesoporous silica is a hydrophilic case. Moreover, the low desorption percentages of pharmaceuticals from silica surface in alkaline media suggest that a fraction of pharmaceuticals are strongly adsorbed on SBA-15. One should remember that there are several types of functional groups on the silica surface, for example: °SiOH, °Si-O&H&O-Si°, °SiOH2+, °SiO-, etc. The strength of interaction between pharmaceuticals and silica surface is always variable and dependent on adsorption active sites (silanol groups). There exist adsorbed pharmaceuticals which bind tightly on silica surface and are difficult to desorb. Such strong bindings between pharmaceuticals and silica surface can be explained by hydrogen bonding or chemical bonding. In cases of acidic pharmaceuticals which possess carboxylic groups (-COOH), the interaction between pharmaceuticals and silanol groups can be postulated as follows:
°Si-OH + HOOC-B °Si-OOC-B + H2O (5)
Or °Si-OH + HOOC-B °Si-O&H&OOC-B + H+ (6)
where B is the remaining part of acidic pharmaceuticals. Generally, a hydrogen bonding reaction (eq. 6) is easier to occur, compared to a ligand-exchange reaction (eq. 5), due to its lower demand of activation energy. Sorption of the antibiotic ofloxacin to Al2O3 solids through the ketone and carboxylate functional groups via a ligand-exchange mechanism was reported . In the case of carbamazepine - a neutral compound, the hydrogen bonding between the amide groups of carbamazepine and the silanol groups of silica is mainly responsible for its adsorption .
3.8. Adsorption of a mixture of pharmaceuticals
As forementioned, SBA-15 showed efficient removal for individual pharmaceuticals in acidic media. However, one should remember that pharmaceuticals usually occur in the environment as a mixture and though individual pharmaceuticals were generally detected at low-levels, total concentrations commonly exceed 1 mg L-1 . The behavior of pharmaceuticals in a mixture is definitely different and complicated, not as individual cases. Therefore, investigation of adsorption of a mixture of pharmaceuticals is required as a succeeding step for further application.
Fig. 6 presents interesting results for adsorption of a 5-pharmaceutical mixture onto SBA-15 as a function of contact time. Adsorption of the mixture can reach steady state in 30 min which is close to the equilibrium time needed for the cases of individual pharmaceuticals (see Fig. 1). Except for clofibric acid, removal of 4 remaining pharmaceuticals is very effective. Removal efficiency is 88.4% for carbamazepine, 77.8% for diclofenac, 75.2% for ibuprofen, 89.3% for ketoprofen, and 35.6% for clofibric acid. In comparison with single component case at the same initial concentration (25 mg L-1) where the removal is 79.6, 42.4, and 86.5% for carbamazepine, diclofenac, ketoprofen, respectively, the removal of these compounds in the mixture is surprisingly higher, especially for diclofenac. Considering two remaining pharmaceuticals, removal of ibuprofen (75.2%) and clofibric acid (35.6%) in the mixture is somewhat smaller than that for the individual cases (87.6% for ibuprofen and 44.8% for clofibric acid). However, it should be mentioned that in the individual case the removal of clofibric acid was estimated at pH 3, instead of pH 5 in the mixture.
It was expected that the removal of every pharmaceuticals in the mixture would be much lower than that in the individual case, due to competitive adsorption among pharmaceuticals. In contrast, the adsorption of diclofenac, carbamazepine, and ketoprofen displayed higher values in the mixture compared to the single case. This synergistic effect can be explained by uncompetitive co-adsorption or succeeding adsorption to form a multilayer of these pharmaceuticals. It can be visualized that there exist some kinds of binding, interaction or even reaction among diclofenac, carbamazepine, and ketoprofen, which induce the adsorption of the three compounds at the same time on surface of mesoporous silica. As reported earlier , in aqueous medium of pH 5.33, penicillin interacts with berberine, forming ion associates through electrostatic attraction of the polar head groups of penicillin and berberine. Whereas, ibuprofen and clofibric acid compete with the remaining compounds for the adsorption, leading to a decrease in their removal efficiency.
This study showed that mesoporous silica SBA-15 is a promising adsorbent for removal of pharmaceuticals not only for surface water but also for wastewater from pharmaceutical industrial manufactures where high-level concentrations of pharmaceuticals have been observed. SBA-15 presented very effective removals for tested pharmaceuticals in acidic media as individual and mixture forms in a short time. The low desorption percentages were confirmed, which implies that pharmaceuticals are not easily detached from adsorbents to get back into the treated water. The kinetics and equilibrium of pharmaceutical adsorption were best described using pseudo-second order equation and Freundlich isotherm model, respectively. Moreover, information about adsorption mechanisms is valuable for further development of adsorbent materials. One should remember that pharmaceutical-adsorbed SBA-15 and mesoporous silica - based materials can be regenerated by combustion without loss, due to stability of mesoporous silica structure up to 850oC . Further results about effect of anions, humic acid, and cations on pharmaceutical adsorption will be provided for evaluating feasibility of application of SBA-15 in real situations.