Removal of natural organic matter (NOM) from an aqueous solution by NaCl and surfactant-modified clinoptilolite

Natural organic matter (NOM) is considered a complex matrix of organic compounds present in many water resources, especially surface water (Wang et al. 2006; Zazouli et al. 2007; Brigante et al. 2008). NOM causes a wide variety of problems in drinking water treatment operations. The presence of NOM in water may be carcinogenic as a result of reactions between NOM and disinfectants. The presence of NOM in raw water sources is unfavorable and has long been a problem for water supply systems in industries. In particular, NOM affects the quality of drinking water in several ways. For example, color, taste, and odor are important esthetic drinking water quality problems caused by such compounds (Mahvi et al. 2009; Li et al. 2011; Ben-Sasson et al. 2013). In addition, NOM can bind with synthetic organic pollutants (such as pesticides) and heavy metals and result in their transport through treatment units and the distribution network (Anirudhan & Ramachandran 2007; Mahvi et al. 2009; Imyim & Prapalimrungsi 2010). Also, as a consequence of the presence of high concentrations of NOM in raw water, the chlorination of raw water can result in disinfection by-products such as trihalomethanes during the chlorination process (Anirudhan & Ramachandran 2007; Chiang et al. 2009). Therefore, the removal of NOM from drinking water resources is very important.

Different methods have been applied to remove NOM from drinking water including coagulation and flocculation, ion exchange, adsorption by activated carbon, and membrane filtration (Landrum et al. 1984; Cornelissen et al. 2008; Chiang et al. 2009; Zhan et al. 2011a). Among the above-mentioned processes, adsorption is a promising method. Many different adsorbents including activated carbon, resin, carbon nanotube, and chitosan have been used to remove NOM (Newcombe et al. 2002; Wu et al. 2002; Chen & Wu 2004; Anirudhan & Ramachandran 2007; Wang & Zhu 2007; Sun et al. 2012). Natural zeolites are hydrated aluminosilicates with the chemical composition of tetrahedrons of SiO₄ and AlO₄, and include water molecules, and alkali and alkaline earth metals in their structures (Wang et al. 2006; Zhan et al. 2011b). Surfactant-modified zeolite (SMZ) can be used for the adsorption of NOM from aqueous solutions. Various surfactants can be used for modification of natural zeolites such as cetylpyridinium bromide, hexadecyltrimethylammonium bromide (HDTMA), and cetyltrimethylammonium bromide (Wang et al. 2006; Karadag et al. 2007; Lin et al. 2011a). Natural zeolites can generally exchange cations instead of anions, which make them appropriate for surface modification by cationic surfactants (Anirudhan & Ramachandran 2007; Zazouli et al. 2008). Cationic surfactants comprise positive groups that can readily exchange their group with the exchangeable cations on the external surface of the zeolite, making a monolayer or bilayer surfactant depending on the conditions, which results in an improvement in the adsorption capacity (Wang et al. 2006; Zhan et al. 2011b). Some studies have investigated the evaluation of NOM adsorption capacity of natural and SMZ. Anirudhan's study showed that surfactant-modified bentonite can be used for the removal of humic acid (HA) from wastewaters (Anirudhan & Ramachandran 2007). HDTMA-modified zeolite was used in Wang et al.'s study on the adsorption of fulvic acid from aqueous solution (Wang et al. 2006). Adsorption of tannic acid from aqueous solution onto SMZ was examined by Lin et al. (2011a).

It is worth noting that each zeolite has intrinsic properties, and the efficiency with which a particular zeolite absorbs a pollutant may vary. Since the efficiency of Iranian zeolites has not been evaluated for the removal of NOM from aqueous solution, the present study aimed to modify a natural Iranian zeolite with NaCl and HDTMA, and investigate its efficiency as a low-cost and frequent adsorbent for the removal of NOM from water. The effects of operational parameters on the adsorption of NOM were evaluated.

The natural zeolite was obtained from Semnan province located in the central region of Iran via Afrazand Co. Ltd. The surface morphology of the natural and surfactant-modified zeolite was analyzed by scanning electron microscopy (SEM) (XL30 Philips Holland). X-ray fluorescence spectroscopy (PW 2404 Philips Holland) was used to analyze the chemical composition of the natural zeolite. The crystallinity phase of the natural and SMZ samples was analyzed using X-ray diffraction (XRD) (Philips-XPERT) techniques. HDTMA was purchased from Merck Company, Germany. Other chemicals used in this study were obtained from Merck Company, Germany. NOM solution samples were real and collected from Tehran Latiyan dam.

The zeolite tuff was ground and sieved using 0.21–0.3 mm American Standard Test Sieve Series (ASTM) sieves. The milled zeolitic tuff was washed several times to remove inorganic impurities; then, for dissolution of salts, 10 g of the zeolite was completely soaked in 1 L of deionized water for 24 hours. To eliminate organic materials in the zeolite, it was then dried at a temperature above 250 °C in an oven for 24 hours (Koh & Dixon 2001; Torabian et al. 2010).

Before the modification of the zeolite surface with the surfactant, in order to make homoionic adsorption sites on the zeolite surface with sodium, 10 g of the zeolite was added to each conical flask containing 100 mL sodium chloride solution of 2 M concentration. Then the conical flasks were shaken by an incubator shaker (150 rpm; 20 °C) for 24 hours (Torabian et al. 2010). In order to prevent the possible intervention of the chloride anions in the modification of the zeolite, samples were washed several times with distilled water to remove the remaining chloride. The residual chloride in the supernatant of the pre-modified samples was determined by AgNO₃ using the argentometry method (Ghiaci et al. 2004). Finally, the samples were dried in an oven at 50 °C for 24 hours (Zhan et al. 2011b).

Zeolite surfactant modification can improve the adsorption capacity. A monolayer is formed when the surfactant concentration is equal to or less than its critical micelle concentration (CMC). The surfactant molecules form a bilayer of surfactant on the surface of the zeolite if the surfactant concentration in a solution exceeds the CMC (Haggerty & Bowman 1994). The surfactant modification of the zeolite makes a hydrophobic environmental condition for the partitioning of the organic molecules with low polarity and high molecular weight (Wang & Peng 2010). The CMC of HDTMA is 1.8 mmol/L (Ghiaci et al. 2004; Torabian et al. 2010). In order to determine the effect of surfactant concentration on the adsorption of NOM, three initial concentrations of HDTMA surfactant (0.5, 1.8, and 20 mmol/L) were selected for modification, with their concentrations being lesser, equal, and higher than the CMC of the surfactant, respectively. Cation exchange capacity (CEC) and external cation exchange capacity (ECEC) of the adsorbents were determined according to the procedure of Haggerty & Bowman (1994). For the preparation of the modified zeolite, 10 g of the prepared cationic zeolite was added into 200 mL conical flasks containing 100 mL of different concentrations of HDTMA. Then, the flasks were shaken in an incubator shaker (150 rpm; 20 °C) for 24 hours. The solid was washed with distilled water repeatedly until no Br⁻ was detected by AgNO₃ solution, and then it was dried in an oven at 50 °C for 24 hours (Torabian et al. 2010). The zeolites modified with 0.5, 1.8, and 20 mmol/L HDTMA solutions were named SMZ1, SMZ2, and SMZ3, respectively.

The XRD patterns showed that clinoptilolite was the main component of the natural zeolite. The structural parameters of the SMZ were very close to that of the corresponding parent natural zeolite, which showed that the crystalline nature of the zeolites remained intact after chemical treatment with surfactant molecules (Figure 1). The surface of the natural and modified zeolite was observed using an SEM device and the observed results are presented in Figure 2. The crystal structure of the natural zeolite can be clearly seen (Figure 2(a)). However, the crystal structure of the SMZ shows different images (Figure 2(b)). This result indicates that an organic layer formed on the zeolite surface when natural zeolite was modified by the surfactant solution. The chemical composition of the natural zeolite is given in Table 1.

The CEC and ECEC of the natural zeolite were determined to be 1.85 and 0.205 meq/g, respectively.

The removal efficiencies of the different types of modified zeolites are shown in Figure 3. According to the figure, the results show that SMZ3 as the zeolite loaded with a surfactant concentration greater than CMC had a higher capacity to adsorb NOM. This result indicates that surface coverage as a bilayer rather than a monolayer has a strongly favorable influence on NOM adsorption onto HDTMA-modified zeolites. Therefore, among the types of adsorbents, SMZ3 was chosen as the best adsorbent and used for the next experiments. The results showed that natural zeolite had an insignificant affinity for NOM in aqueous solutions. This may be due to the strong dipole interaction between natural zeolite and water, which excludes HA from zeolites. Different mechanisms, such as hydrogen bonding, hydrophobic interaction, and partitioning mechanisms, have been suggested to account for NOM adsorption onto SMZ with monolayer coverage (Wang et al. 2006). HA is one of the main components of NOM and contains several kinds of functional groups, such as –COOH and –OH (Hartono et al. 2009). The formation of hydrogen bonds between the N of HDTMA and hydroxyl groups and carboxylic groups of HA may be favorable to HA adsorption onto modified zeolite. Koopal et al.'s study showed the binding of cationic surfactant cetylpyridinium chloride to HA was due to electrostatic and hydrophobic attractions (Koopal et al. 2004). This shows that hydrophobic interaction or electrostatic attraction may be the main mechanism in HA or other NOM adsorption onto cationic SMZ. When a surfactant monolayer forms on modified zeolite, NOM adsorption may be attributed to hydrophobic interaction and hydrogen bonding. NOM components adsorption may be driven by a collaboration of hydrophobic interaction, hydrogen bonding, and electrostatic interaction for modified zeolite with bilayer coverage (Ding & Shang 2010).

Figure 4 illustrates the effect of contact time and initial concentration on the adsorption of NOM onto SMZ3. For assessing the effect of contact time and initial concentration of NOM, a fixed dose of SMZ3 at pH = 7 was used for each concentration of NOM. The experiments showed that the amount of adsorbed NOM on the zeolite surface increased gradually; by increasing the contact time, the adsorption rate was initially rapid and then gradually decreased with time until it reached equilibrium beyond which there was no significant increase in the rate of adsorption. Equilibrium was nearly reached after 150 min for five different initial NOM concentrations. Hence, in the present work, 150 min was chosen as the equilibrium time. Figure 4 shows that the amount of adsorbed NOM depends on the initial concentration of NOM in the solution. By increasing the initial concentration of NOM from 4 to 30 mg/L, the amount of adsorbed NOM at the equilibrium time (q_e) increased from 4 to 16 mg/g. It was observed that the amount of adsorbed NOM increased with the increase in the contact time for all the initial NOM concentrations. These observations show that initial NOM concentration has no effect on required time for equilibrium. During the initial stages, many numbers of vacant sites are accessible on the surface of the adsorbent; therefore, the adsorption rate of NOM is relatively high. The adsorption process becomes less efficient due to the gradual occupancy of these sites. With increasing time, the occupation of the remaining vacant sites becomes more difficult owing to the increased repulsive forces between NOM components on the surface of the adsorbents. These results are consistent with other similar studies (Anirudhan & Ramachandran 2007).

Figure 5(a) shows that the NOM removal efficiency increases with decreasing pH. A low pH was best for NOM adsorption onto SMZ3, and we used pH 5 for isotherm experiments in the following stages. According to the literature, electrostatic interaction may be the significant mechanism for NOM adsorption onto SMZ, in addition to hydrophobic interaction and hydrogen bonding (Zhan et al. 2011b). HA, fulvic acid, and tannic acid are the principal components of NOM; they are anionic hydrophobic macromolecules and have weakly acidic functional groups. At lower pH, most of the weakly acidic functional groups in HA and fulvic acid are uncharged and are therefore more susceptible to adsorption by SMZ (Zhan et al. 2011c). NOM components are weak organic acids, and their ionization is strongly dependent on solution pH. At low solution pH, NOM molecules are present in a neutral form. The adsorption of unionized NOM onto the positively charged surface of SMZ is unlikely to be driven by electrostatic attraction. Therefore, hydrogen bonding and organic partitioning are responsible for the adsorption of NOM onto SMZ at lower solution pH. NOM molecules are almost completely ionized at higher solution pH. The increase of solution pH from 3 to 11 leads to decrease of hydrogen bonding between NOM and SMZ but increases electrostatic attraction between NOM and SMZ. At higher solution pH, completely ionized NOM molecules cannot supply hydroxyl hydrogen atoms to the nitrogen atom of the HDTMA bilayer of SMZ to form hydrogen bonding (Karadag et al. 2007; Lin et al. 2011a). In addition, the increase of the solution pH leads to the increase of competition between the hydroxyl ion and the ionized NOM molecules for the same positively charged adsorption sites on the surface of SMZ3, which causes a decreased NOM adsorption capacity (Wang & Zhu 2007; Zhan et al. 2011b). Furthermore, it is noticeable that the NOM molecular size may increase from a spherical structure form at lower pH to a linear structure form at higher pH, which also indicates that the adsorption of NOM is reduced at higher pH values (Wang & Zhu 2007; Monash & Pugazhenthi 2010). Similar trends of pH effect were observed for the adsorption of fulvic acid and HA on SMZ (Wang et al. 2006; Anirudhan & Ramachandran 2007). However, Lin's study showed that the tannic acid adsorption capacity for SMZ was relatively high at solution pH of 4.0–7.0, and decreased with increasing solution pH from 7.0 to 8.5 (Lin et al. 2011b).

Figure 5(b) presents the effect of adsorbent dose on NOM removal at different initial concentrations of NOM. The experiments showed that with increasing the amount of adsorbent dose from 0.2 to 1.2 g/L, % removal of NOM increased for all NOM concentrations. However, q_e for different NOM concentrations decreased with increasing the adsorbent dose. For example, according to Figure 5(b), q_e (mg/g) for the initial NOM concentration of 4 mg/L decreased from 10 to 4.1 mg/g, and for the initial concentration of 30 mg/L decreased from 55 to 19 mg/g when the adsorbent dose was increased from 0.2 to 1.2 g/L. It is easily understood that the number of available adsorption sites is increased by increasing the adsorbent dose, which results in the increased percentage of NOM removal. On the other hand, the experiments showed that q_e decreased when the adsorbent dose was increased, which is due to unsaturated sites on the adsorbent during the adsorption process (Wang & Peng 2010; Lin et al. 2011b).

Figure 6(a) shows the plots of linear form of pseudo-first-order kinetics for all concentrations of NOM. The intercept and slope of the plots were used to determine the values of q_e and the pseudo-first-order rate constant (k₁), respectively (Hartono et al. 2009). The calculated rate constant, experimental and calculated (q_e), and corresponding correlation coefficient values are presented in Table 2. It was also observed that correlation coefficients (r²) were relatively low for all NOM concentrations. The calculated q_e values for the pseudo-first-order kinetics do not agree with experimental values very well (Table 2). These results show no suitability of the pseudo-first-order model in foretelling the kinetics of NOM adsorption onto the SMZ3.

where k₂ is the equilibrium rate constant calculated for the pseudo-second-order adsorption model (g/mg/min).

The q_e is obtained from the slope of the plot of t/q_t against t (Figure 6(b)), k₂ can be calculated from the value of the initial sorption rate. The calculated q_e values agree with the experimental data very well (Table 3). The correlation coefficients (r²) for the pseudo-second-order kinetic model are higher than 0.99 for all concentrations. These indicate that the adsorption of NOM from solution onto SMZ3 obeys the pseudo-second-order kinetic model.

where C_e (mg/L) is the equilibrium concentration of NOM in the solution, q_e (mg/g) is the amount of NOM adsorbed onto SMZ3 at equilibrium, q_m (mg/g) is the maximum adsorption capacity, and K_l (L/mg) is the Langmuir constant related to the sorption energy. The plot of 1/q_e versus 1/C_e gives a straight line (Figure 7(a)), and the slope and intercept of this line correspond to q_m and K_l, respectively (Fytianos et al. 2000; Shirmardi et al. 2013; Vosoughi Niri et al. 2014). The calculated correlation coefficient (r²) and the Langmuir constant are presented in Table 3. From the table, it is evident that the maximum NOM sorption capacity of SMZ3 (q_m) and the r² value are 147 mg/g and 0.97 at 25 °C, respectively. The r² of the Langmuir isotherm was greater than that of the Freundlich isotherm for the adsorption of NOM. This implies that the Langmuir model better depicts the adsorption of NOM onto SMZ3 than the Freundlich model. Some investigations of NOM adsorption using several adsorbents have been conducted. For example, Ferro-Garcia et al. (1998) reported that the adsorption capacity of HA on a commercially available activated carbon was 6.9 mg/g. Daifullah et al. (2004) prepared activated carbon from rice husks by chemical activation using phosphoric acid and determined that the adsorption capacity was 45 mg/g based on a Langmuir isotherm. Table 1 summarizes the results of some experiments using different kinds of zeolites for removal of NOM. According to data provided in Table 1, SMZ has been shown to be an efficient adsorbent for the elimination of NOM. As shown in Table 1, the zeolite examined in this work has a greater capacity for eliminating NOM, compared to those of previously tested adsorbents. Considering its high adsorption capacity, as well as its profound availability and ease of accessibility in nature, it can be concluded that SMZ is an efficient and economic adsorbent for removing NOM, one of the most prevalent contaminants in most surface water. These characteristics, along with the simplicity of operation, introduce adsorption onto SMZ as a suitable technology for treating NOM-contaminated waters, particularly for rural and small communities.

The results show that SMZ can be used as a low-cost adsorbent for the adsorption of NOM from aqueous solution, especially in areas where there are many zeolite mines. The adsorption of NOM by SMZ was enhanced at acidic pH levels. For SMZ3, the NOM adsorption process developed reasonably well with the pseudo-second-order kinetic model. Results from this study indicate that HDTMA-modified natural zeolite is a favorable adsorbent for the removal of NOM from aqueous solutions. The equilibrium data fitted well with the Langmuir isotherm. The maximum adsorption capacity of SMZ3 with cationic surfactant for NOM was 147 mg/g at 25 °C and pH 5.

The authors appreciate the financial support of the Deputy of Research, Tehran University of Medical Sciences.