Adsorption characteristics of high-silica zeolites (HSZSM-5) for two selected sulfonamide antibiotics (SAs) (sulfamethoxazole and sulfadiazine) were investigated. The SAs were almost completely (>90%) removed from the water by HSZSM-5. Adsorption followed second-order kinetics with liquid–film diffusion as the dominant mechanism. SA adsorption capacity on high-silica zeolites was examined in terms of pH, temperature, and the presence of natural organic matter (NOM). HSZSM-5 had better adsorption performance in acidic conditions, and the apparent distribution coefficient indicated that SA0 species were the major contribution to the overall adsorption at pH of 2–10. Adsorption of SAs on HSZSM-5 was a spontaneous and exothermic physisorption process. SA removal by HSZSM-5 was a mixed mechanism through ion-exchange and hydrophobic interaction. HSZSM-5 has potential application prospects in removing SAs from wastewater.
Sulfonamide antibiotics (SAs) were the earliest clinically used antibacterial drugs, and their application can be traced back to 1937. They are widely used to treat various bacterial, protozoan and fungal infections owing to their low cost and broad-spectrum antibacterial ability (Zhao et al. 2018). SAs are excreted through wastewater efﬂuents and expired drugs in their original or metabolized form into aquatic environments (Yang et al. 2011). The spread of sulfonamides in the environment has attracted considerable attention because of their potential toxicity and antibiotic resistance in microorganisms (Peiris et al. 2017). The traditional sewage treatment process cannot guarantee the elimination of SAs in sewage due to their low biodegradability (Zhao et al. 2018). Therefore, it is urgent to develop an efficient, economical and eco-friendly technology for SAs treatment.
Adsorption is a simple and effective SA removal technology (Wang et al. 2017; Zhao et al. 2018). The adsorbents that have shown promising adsorption properties include carbon materials (Peiris et al. 2017), clay minerals (Ahmed et al. 2015), hollow polymer nanorods (Xie et al. 2016), porous resin (Yang et al. 2011), graphene oxide (Nam et al. 2015) and zeolites (Braschi et al. 2016b). High silica (HS) zeolites with high affinity for organic molecules have been used to remove sulfonamides (de Ridder et al. 2012). The capacity to retain SAs including sulfamethoxazole (SMX) and sulfadiazine (SD) by zeolites mainly depends on the SiO2/Al2O3 (mol/mol) ratio (Braschi et al. 2016a). HS zeolite Y (SiO2/Al2O3 = 200) can completely and quickly remove sulfadiazine, sulfamethazine, and sulfachloropyridazine (Braschi et al. 2010a, 2010b). The adsorption kinetics of HS zeolite Y and MOR (SiO2/Al2O3 = 200) for SMX was favorable, while that of ZSM-5 (SiO2/Al2O3 = ∼500) was slow (Blasioli et al. 2014). HS zeolite Y (SiO2/Al2O3 = 200) exhibited an irreversible SAs removal (sulfadiazine, sulfamethazine and sulfachloropyridazine) due to the Van der Waals and weak H-bonding between zeolite and sulfa-drugs (Braschi et al. 2010a, 2010b). The adsorption mechanisms were the H-bonding and electrostatic interactions between FAU-type zeolites and SMX (de Sousa et al. 2018). Thus, different types of HS zeolites exhibited different adsorption characteristics of SAs.
SAs are rather water-soluble polar compounds whose ionization depends on pH. The adsorption of SMX on zeolite Y (SiO2/Al2O3 = 200) was pH-dependent but irreversible at pH from 5 to 8 as its bulky ‘V’ structure stabilized the embedded molecules (Braschi et al. 2016b). The low removal rate of SMX by FAU-type zeolites (SiO2/Al2O3 = ∼30 or 82) was obtained at pH ranging from 2.5 to 4.5 and 8.5 to 10.5. The adsorption capacity of selected sulfa drugs on Y-type zeolite (HSZ-385, SiO2/Al2O3 = 100) decreased when pH was lower and higher than their pKa1 and pKa2, respectively (Fukahori et al. 2011). To date, there is little known about the pH-dependent adsorption behavior of SAs by HSZSM-5 zeolites.
In this work, two widely used SAs, SMX and SD, were chosen as model pollutants. The specific objectives of this work were to: (i) investigate the effect of contact time and pH on SA adsorption by HSZSM-5; (ii) evaluate the adsorption capacity and reveal the adsorption mechanisms of HSZSM-5 towards SAs; and (iii) examine the effect of model natural organic materials (natural organic matter (NOM)). Understanding the mechanism of pH-dependent adsorption is crucial for the potential application of hydrophobic zeolite as an adsorbent for SA removal in wastewater treatment.
MATERIALS AND METHODS
Materials and chemicals
High-silica zeolites ZSM-5 (HSZSM-5, average pore size of 0.53–0.58 nm, SiO2/Al2O3 (mol/mol) ratio of 70, 170 and 500) were purchased from Nanjing Xfnano Materials Tech. Co., Ltd (Nanjing, China). SAs including SMX and SD were obtained from Sigma-Aldrich (USA). All other reagents were of analytical grade and used as received. Double distilled water was used for all preparations.
Adsorption of HSZSM-5 at various adsorbent doses and equilibrium times
The adsorption experiments were maintained at 298 K and pH 6. The initial concentration of both SD and SMX was 100 μg/L, respectively. Zeolite was immerged into SA solutions under continuous stirring. The zeolite dosage ranged from 0.1 to 1.0 g/L. At selected time intervals (0.083, 0.166, 0.333, 0.66, 1, 2, 4, 8, 12, 24, 48, 72 h), a 100 ml solution from the reaction system was taken as the sample. The concentration of SAs was determined by high performance liquid chromatography (HPLC) (1200, Agilent, USA). The sample concentration was measured in triplicates.
Adsorption of SAs at different initial pH
To investigate the effect of pH, it was adjusted in the range 2.0–10.0. The residual concentration of each equilibrium solution was determined. The influence of pH on the zeta potential of the zeolites with different SiO2/Al2O3 ratios was also examined. Other experimental conditions depended on the experimental results in context.
Adsorption of SAs at different NOM concentrations
The effect of NOM on SA adsorption was also investigated. The temperature of the batch experiments was maintained at 298 K, and the SiO2/Al2O3 ratio of high-silica zeolites was 500. The initial pH value of the solutions was 3.0, 6.0 and 9.0, respectively. Bovine serum albumin (BSA), sodium alginate (NaAlg) and humic acid (HA) were used as the model NOM for the adsorption experiment (Alresheedi et al. 2019), respectively. Initial NOM content ranged from 1 to 20 mg/L. The following sorption step was performed as single solute sorption experiments.
The adsorption experiments were performed at 288, 298 and 308 K, respectively. The SA concentrations were set as 0.1 to 15 mg/L at pH of 6.0. After 24 h to reach equilibrium, the samples were filtered by a 0.45 μm membrane before testing.
The experiments were performed with 100 μg/L of SAs and 0.2 g/L of zeolites. The mixture was stirred continuously at an initial pH of 6. At time intervals (0.083, 0.166, 0.333, 0.66, 1, 2, 4, 8, 12, 24, 48, 72 h), 100 mL sample solutions were obtained.
Analysis of sulfonamides
For the analysis of SAs, sample extraction based on solid-phase extraction (SPE) was used as a pre-concentration technique before their quantitative determination. The analysis consisted of carrying out an SPE of 100 mL samples using 60 mg OASIS HLB cartridges (Waters, Milford, MA, USA) and a final elution from the cartridge using 6 mL of methanol. An Agilent 1200 module (Agilent, Palo Alto, CA, USA) equipped with a 150 × 4.6 mm ZOBRAX Eclipse XDB-C18 column (5 mm, Agilent, Palo Alto, CA, USA) was employed to separate the analytes. A gradient with a flow rate of 0.8 mL/min was used. Mobile phase A contained HPLC-grade methanol and mobile phase B contained 0.5% (v/v) HPLC-grade acetic acid in water. The sample injection volume was 10 μL and the autosampler was operated at 40 °C.
RESULTS AND DISCUSSION
Effect of adsorbent dose and contact time
Removal of SAs was enhanced at higher zeolite dosage as indicated in Figure 1. With an increase of zeolite dosage from 0.1 to 0.6 g/L, the removal rates of SMX and SD increased from 42.1% to 99%, and 51.8% to 70.0%, respectively.
The increase of zeolite concentration led to the increment of the effective surface area and available adsorption sites (Pan et al. 2009), which led to the increase of the adsorbed amount of SAs. The removal rates of SMX and SD increased to 100% and 90% when zeolite dosage further increased to 1.0 g/L, respectively. The result suggested that SMX in the solution could be removed completely. In addition, with further increase of zeolite dosage, the adsorption capacity decreased from 0.42 to 0.10 mg/g and 0.51 to 0.089 mg/g for SMX and SD, respectively. Considering cost and in order to better understand the effect of other experimental conditions, the optimum adsorbent dosage was kept as 0.2 g/L for further experiments in this study.
Equilibrium time is important for understanding the distribution of adsorbents and adsorbate in wastewater treatment processes. Figure 2 illustrates the effect of contact time on SA adsorption by zeolites. The initial higher adsorption efficiency was attributed to the availability of active binding sites for SAs in the first 60 min. The removal efficiencies of SMX and SD were limited after 1 h with the lack of reactive adsorption sites at a fixed adsorbent dosage. The equilibrium time for SA adsorption on HSZSM-5 was 24 h.
Effect of pH
Adsorption removal analysis
The protonation–deprotonation conversion of functional groups of ionizable organic compounds depends on the pH of the solution. The effect of pH on SA adsorption on zeolites is shown in Figure 3.
Moreover, HSZSM-5 with an SiO2/Al2O3 ratio of 500 was positively charged in the experimental pH range as indicated in Figure 3. This suggested that the electrostatic interaction between SD+ and ZSM-5 did not affect the adsorption of SD+, meaning that it was not good for ion-exchange between SD+ and HSZSM-5. At 3 < pH < 6.4 (pKa2), SD0 was the main SD species in the solution, while the removal efficiency of SD decreased. This was ascribed to the decreased hydrophobicity of SAs as the pH increased (Yu et al. 2015). This was consistent with the finding that the change of SAs adsorbed on HS zeolite Z-385 was due to hydrophobic interaction (Fukahori et al. 2011). At 6.4 (pKa2) < pH < 10.0, the removal efficiency of SD continuously decreased. The electrostatic interaction between SD− and HSZSM-5 did not favor the adsorption of SD species. When the pH was lower than 3.0, SMX0 was the main SMX species, and the removal rate of SMX was approximately 100%. The removal rate of SMX decreased with increasing pH from 4.0 to 10.0. Thus, the increasing percentage of SA− did not increase the removal rate of SAs. It further confirmed that electrostatic interaction did not favor SA adsorption by HSZSM-5. A possible explanation for the sharp decrease could be that increasing pH led to a decrease in the hydrophobic character of HSZSM-5 as well as the ion exchange, and then resulted in the reduced SA adsorption. In addition, another reason for the increase of SA adsorption on HSZSM-5 at lower pH range may be the decrease of hydrophobicity in the ionized form and the inhibition of Lewis acid–base interaction (Zhao et al. 2016).
Surface charge of the absorbent is an important parameter affecting adsorption. The effect of SiO2/Al2O3 ratio on the zeta potential of HSZSM-5 at different pH was investigated and the experimental data are shown in Figure 4.
With increasing pH, the removal rate of SD and SMX by HSZSM-5 at different SiO2/Al2O3 ratios followed a similar trend as indicated in Figure 3. The adsorbed amount of SAs by HSZSM-5 increased with the increase of SiO2/Al2O3 ratio, suggesting that the stronger the hydrophobicity, the better the adsorption capacity. Moreover, the zeta potential of HSZSM-5 with SiO2/Al2O3 = 70 was higher than that of HSZSM-5 with SiO2/Al2O3 = 170 at the same pH, as shown in Figure 4. The adsorbed amount of SAs on HSZSM-5 (SiO2/Al2O3 = 170) was higher than that of HSZSM-5 (SiO2/Al2O3 = 70) in the experimental pH range as shown in Figure 3(a) and 3(b). This confirmed again that the mechanism of SAs adsorbed by HSZSM-5 was not electrostatic interaction.
Adsorption affinity of different SA species
The model (Equation (9)) provided a better fit for the experimental data. For the adsorption of SD and SMX by HSZSM-5 (SiO2/Al2O3 ratio = 500), the order of adsorption coefficients was , suggesting that the SA0 species was the main contributor to the overall adsorption. Yu et al. (2015) also found that SA0 was the main species adsorbed on carbon nanotubes. The of SMX was much higher than that of SD because of its higher . Additionally, the -NH2 and -SO2NH- groups of SAs can interact with oxygen-containing functional groups on the zeolite surfaces by hydrogen bonding (Ng & Mintova 2008). The π–π electron donor–acceptor (EDA) interaction and hydrogen bonding for SA species followed the order SA+ > SA0 > SA−. This was different from the order of . Thus, π–π EDA interaction and hydrogen bonding were not the mechanism for the adsorption affinity.
It is clear from Figure 5 that SA0 contributed more to the total adsorption than its species fractions. SA+ and SA− contributed less than their species fractions, and especially SA−, meaning that there was a stronger repulsion interaction between SA− and HSZSM-5. However, the HSZSM-5 with SiO2/Al2O3 ratio = 500 was positively charged. These results reveal the importance of the SA0 species for adsorption. The main contribution of the SA0 species over the whole pH range indicated that the main mechanism was hydrophobic interaction rather than electrostatic interactions.
Effect of NOM
Micropollutants coexisting with NOM would affect adsorption removal efficiency in aquatic environments. The effect of NOM on SMX and SD adsorption on zeolites is shown in Figure 6.
HA increased SD− and SMX− adsorption, suggesting that a complexation reaction between SAs and HA was responsible for the sorption changes (Figueroa et al. 2004). BSA decreased SD− and SMX− adsorption. The dissociated amino-group of BSA competed with SA− for π electron donor sites on the adsorbent (Jia et al. 2017), which inhibited the sorption of SAs to HSZSM-5. NaAlg decreased the SD adsorption due to the dissociated carboxyl-group of NaAlg, while no difference was observed on SMX adsorption possibly due to its faster adsorption rate, which will be discussed in the following section. The effect of NOM on SA adsorption was also dependent on its concentration and solution pH.
In addition, as the molecule of the representative NOM was larger than the pores of the zeolites, it could not penetrate the pores of the zeolites to block the pores and compete for adsorption sites (Qin et al. 2007). The interaction between NOM and SAs may be neglected compared with their sorption to the adsorbents (Lian et al. 2015).
The adsorption equilibrium data are usually represented by adsorption isotherms. The equilibrium adsorption of SAs on zeolites decreased with the increment of temperature, indicating that low temperature was favorable for SA removal. The Langmuir and Freundlich models are the commonly used isotherms for determining adsorption of phenomena (Liu et al. 2013). The Langmuir model could not provide a good fit for the adsorption data (not shown). The fitted parameters of the other isotherm models are listed in Table 2.
The Freundlich model is suitable for the description of the adsorption isotherms as indicated by the larger coefficients of R2 (see supporting materials Figures S1 and S2, available with the online version of this paper). The values suggest that SAs had a stronger adsorption affinity with HSZSM-5, which decreased with an increase of temperature, revealing that the interaction between SAs and HSZSM-5 was exothermic. The 1/n values for SAs are between 0.4 and 0.8, which suggest good adsorption and high affinity of ZSM-5 to SAs. If n is above 1, the adsorption is a physical process (Jiang et al. 2002). It is noted that all n exceeded 1 in this work, meaning that SA adsorption on HSZSM-5 was a physical process.
The D-R model was used to observe the physical or chemical properties of the adsorption process. The E values of the D-R isotherm model were obtained from the plot of vs (see supporting materials Figures S3 and S4, available online). If E is less than 8 kJ/mol, the adsorption is of physical type (Wang et al. 2008). It was found again that the process of SMX and SD adsorption on HSZSM-5 was physisorption (1–8 kJ/mol). The apparent energy of adsorption decreased with increasing temperature.
The negative value of (−48.68 to −54.63 kJ/mol for SMX, and −40.00 to 59.71 kJ/mol for SD) suggests that the process was exothermic. The negative values indicate the feasible and spontaneous adsorption of SAs on HSZSM-5. If is less than 40 kJ/mol, the adsorption is physisorption (Kara et al. 2003). Thus, the adsorption of SAs by HSZSM-5 was a physisorption process. The negative value of shows the randomness of absorbent (HSZSM-5) and adsorbate (SA) decreased at the solid–liquid interface, which tended to be more in an ordered state.
To better understand the mechanism of the adsorption process, the adsorption kinetics models were examined and the fitting parameters are illustrated in Table 3.
The plot of vs t has no linear relationship (see supporting materials Figure S5, available online). The calculated values determined from the model basically deviate from the experimentally determined values. The pseudo-first-order equation does not match the adsorption data better. The pseudo-second-order equation is a proposition on the basis of the linear relationship between the adsorption kinetics and the square of available sites. The calculated values (for example, 0.14 mg/g for SD and 0.43 mg/g for SMX at pH = 6) are very close to the experimental data (0.14 mg/g for SD and 0.42 mg/g for SMX at pH = 6). Compared with the coefficient R2 of other models, the pseudo-second-order could better describe the adsorption behavior than the pseudo-first-order.
The control mechanism of the SA adsorption process was studied by intra-particle diffusion. If the plot of vs t0.5 is linear and passes through the origin, then intra-particle diffusion is the rate control step (Mahmoud et al. 2012). The plot shows the adsorption process was divided into three stages, including a sharp initial curve, linear curve and a plateau (see supporting materials Figure S6). The vs t0.5 plot is not linear with zero intercept, which implies that the intra-particle diffusion model might not be the only rate control step, while boundary layer diffusion controlled the adsorption to certain extent (Cheung et al. 2007). The linear liquid–film diffusion plots suggest that the diffusion of liquid film around HSZSM-5 was the total rate control step.
High-silica zeolites ZSM-5 were demonstrated to have good potential to removal SAs from water. The removal of SAs from water was strongly dependent on pH and temperature. The decrease of pH and temperature led to an increase in SA adsorption capacity on high-silica ZSM-5. The adsorption followed the Freundlich model. Pseudo-second-order was the accurate description of the adsorption kinetics of SAs on high-silica ZSM-5, and liquid–film diffusion was the overall rate control step. Negative thermodynamic constant values showed that the process of SAs adsorbed by high-silica ZSM-5 was physical, spontaneous and exothermic adsorption. The adsorption behaviors of SAs were ion-exchange process and hydrophobicity interaction. The present results imply that high-silica ZSM-5 may be an adsorbent to removal SAs from wastewater.
This work was supported by the Fundamental Research Funds for the Central Universities of China (No. 2662017JC019, 2662018JC013).