A batch study on the adsorption/desorption behavior of vancomycin on bentonite nanoparticles in aqueous solutions

Negative impacts caused by the presence of micro-pollutants in aqueous environments have become a severe dilemma for human and environmental health in recent years (Luo et al. 2014; Danalıoğlu et al. 2017). Higher consumption of pharmaceuticals due to the growth of population has led to the presence of emerging contaminants that have polluted surface and groundwater. Therefore, efficient and effective removal of these materials from water resources is considered necessary (Seo et al. 2017). Antibiotics are among the most widely used pharmaceuticals utilized extensively to prevent and treat infectious diseases in humans and livestock (Genç et al. 2013; Peng et al. 2016, 2019; Ahmed et al. 2017). These drugs, which are of chemical origin, inhibit the growth of microorganisms such as bacteria, fungi, and protozoa. A considerable amount of the consumed antibiotics (about 60–90 percent) may be excreted unchanged in urine or feces and finally enter aquatic and soil environments through livestock sewage, pharmaceutical wastewater, sewage sludge, and fertilizers (Wu et al. 2013; Peng et al. 2016). This has led to their presence in various environments, such as surface water, groundwater, soils, and sediments (Hamscher et al. 2002; Kolpin et al. 2002; Costanzo et al. 2005; Hernando et al. 2006).

Glycopeptides are among the oldest types of antibiotics and their activity mostly covers Gram-positive bacteria and anaerobic organisms. They were not widely used at first but this has changed with the development of bacterial resistance (Van Bambeke et al. 2004; Van Bambeke 2006). Vancomycin (VAN) is a tricyclic glycopeptide considered by medical communities as a last resort for coping with bacteria. It is used when other antibiotics do not exhibit acceptable efficacy. The chemical structure of VAN is depicted in Figure 1. The most crucial concern regarding this drug is resistance among natural bacterial communities such as VAN-resistant enterococci, known as the troublemaker during the last two decades (Qiu et al. 2016).

VAN has been detected at low concentrations (ng L⁻¹ to μg L⁻¹) in wastewater treatment plant effluent and high concentrations (mg L⁻¹) in wastewater from pharmaceutical industries. Moreover, low VAN concentrations have been reported in sediments and sludge (Laverman et al. 2015; Qiu et al. 2016).

Various methods have been used for the removal of antibiotics from aqueous environments, including biological treatment (Arikan 2008), chlorination (Navalon et al. 2008), advanced oxidation technologies (Lee et al. 2011), electrochemical treatment (Yu et al. 2016), membrane process (Koyuncu et al. 2008), adsorption (Zhang et al. 2016), and ultrasonic cavitation effect method (Hou et al. 2012). Among these methods, the adsorption process is considered to be a suitable option because of its advantages such as easy operation, high efficiency, low cost, and also the fact that there is no risk of highly toxic byproducts being created during the process (Yi et al. 2018; Davoodi et al. 2019).

In recent years, due to their high adsorption capacities and low cost compared to other adsorbents, clay minerals, such as bentonite, have been used in the removal of various contaminants (Wang et al. 2013; Yi et al. 2018). Bentonite is a common clay, comprised mostly of montmorillonite, and is used in various applications such as adsorption, cosmetics, and pharmaceuticals due to its high surface area and swelling property (Guo et al. 2019).

Although VAN presence in aqueous environments and its negative impacts have been studied and evaluated before (Laverman et al. 2015; Qiu et al. 2016), its removal using the adsorption process has not been studied. Moreover, the desorption of antibiotics is also considered a threat due to its potential for secondary pollution of water resources (Aga 2007). Consequently, acquiring information on desorption of these chemicals from soil is necessary to evaluate the threat of their presence in the environment. Montmorillonite is known as a major constituent of soil and sediments (Wu et al. 2013; Fijałkowska et al. 2019).

The objective of the present study was to investigate the conditions and mechanism of VAN adsorption/desorption using bentonite clay nanoparticles as an adsorbent. For this purpose, properties of bentonite were carefully studied using field emission scanning electron microscopy (FESEM), energy-dispersive X-ray spectroscopy (EDX), Brunauer–Emmett–Teller (BET) analysis, and Fourier transform infrared (FTIR) spectroscopy. Effects of pH, ionic strength, adsorbent dosage, and contact time on the adsorption/desorption process were also evaluated. Langmuir, Freundlich and Temkin isotherm equations were employed to describe adsorption equilibrium data. FTIR analysis was employed to understand the adsorption/desorption mechanism at the molecular level. To our knowledge, no similar research has been conducted before.

The nanoclay hydrophilic bentonite was purchased from Sigma-Aldrich. Vancomycin with a purity of 99% was obtained from Dana Pharmaceutical Co. Potassium chloride, sodium hydroxide, hydrochloride acid, aluminum chloride, calcium chloride, and sodium chloride were of analytical reagent grade and obtained from Merck.

Nitrogen adsorption/desorption analysis was performed on bentonite nanoparticles using a surface analysis device (Belsorp mini II, Japan) and the specific surface area and total pore volume were determined using the BET method. Surface morphology and structure of the adsorbent were determined using FESEM and EDX (Mira3, France) analyses. To determine the value of the zero point of charge (pH_zpc), a series of flasks containing 50 mL of distilled water with different initial pH values (2–11) were prepared. About 20 mg of bentonite was placed in each flask and left overnight at 25 °C. The final pH was then measured and plotted against the initial pH. The pH_zpc can be readily obtained from the point at which the initial pH versus final pH curve crossed the y = x line on the graph.

Concentrations of VAN in samples were measured using a UV–vis spectrophotometer (JENUS UV-1200, China) at a wavelength of 280 nm. For exploring the adsorption/desorption mechanism, FTIR analysis was performed on the adsorbent before and after the adsorption process (Perkin Elmer, USA).

In order to achieve VAN-preloaded nanoclay particles, 0.018 g bentonite was contacted with 30 mL aqueous VAN solution with two initial concentrations of 50 and 150 mg L⁻¹, and consequently VAN-preloaded nanoclay particles (80 and 236 mg g⁻¹, respectively) were obtained. For all desorption studies, the initial VAN loadings were 80 and 236 mg g⁻¹. In each 50 mL centrifuge tube, 0.018 g VAN-preloaded bentonite was combined with 25 mL of desorbing solution. The suspension was mixed in a laboratory mechanical shaker at 500 rpm. The resulting mixture was centrifuged at 6,000 rpm for 30 min to precipitate the majority of bentonite. A 10 mL aliquot of the supernatant was taken and 1.5 mL of a 1 M KCl solution was added to it to salt out any remaining bentonite particles from the liquid phase (the small increase in volume due to the addition of KCl solution was properly taken into account when calculating drug concentration). This was centrifuged down at 6,000 rpm for 15 min to ensure as much bentonite had been removed as possible before absorbance measurements. Finally, the VAN concentration in the supernatant was measured. The amount of VAN remaining after desorption was calculated from the difference between the amount of equilibrium adsorption and the VAN concentration in the supernatant. In experimenting the effect of cation concentration on VAN desorption, concentrations of NaCl, CaCl₂, and AlCl₃ were 10–50, 2–10 and 0.25–3 mM, respectively. For experiments investigating the influence of pH on VAN desorption, the equilibrium solution pH was adjusted from 3 to 10 with an increment of approximately 1.

Results of BET analysis indicate that the specific surface area and pore volume of bentonite nanoparticles were 12.45 m² g⁻¹ and 0.0996 cm³ g⁻¹, respectively. According to the results of EDX analysis on the adsorbent (Table 1), silicon, oxygen, and aluminum are the main constituents of bentonite. Moreover, a considerable amount of carbon and minimal quantities of magnesium, iron, sodium, and calcium are also found in bentonite.

The effect of contact time on VAN adsorption at various initial concentrations (50, 100, 150 mg L⁻¹) was investigated. According to the results, the adsorption process for all the three initial concentrations had reached the equilibrium in the first minute (47.31, 94.03 and 138.1 mg g⁻¹, respectively).

Effect of contact time on VAN desorption at two initial loadings (80 and 236 mg g⁻¹) in deionized water and 50 mM Na⁺, Ca²⁺ and AL³⁺ solutions was investigated (Table 2). In general, VAN desorption from bentonite clay was almost independent of contact time and reached equilibrium in all the samples after 1 minute. As seen in Table 2, when Al³⁺ and Ca²⁺ were the desorbing reagents, desorption capacity was higher compared to Na⁺. This can be justified by the charge effect. It means in a cation exchange reaction, cations with smaller sizes and higher charges are preferred over cations with larger sizes and lower charges.

The results show that the pH_zpc for bentonite was 9.1. By definition, this indicated that the bentonite clay surface had a positive charge at pH values below 9.1 and a negative charge at pH values higher than 9.1 (Putra et al. 2009; Derakhshani & Naghizadeh 2018). The presence of trace amounts of carbonates in bentonite could be one of the reasons for its alkaline pH. The effect of the initial pH on VAN adsorption is shown in Figure 2(a). According to the results, the amount of VAN adsorbed by bentonite considerably declined with increasing the solution pH. Hence, the largest amount of VAN was adsorbed at pHs 3 and 4. Based on the obtained pH_zpc for bentonite, its surface has a positive charge in acidic solutions. Moreover, since the molecular formula of VAN contains two alkylamine groups, several amide nitrogen groups, and only one carboxylic acid group (Figure 1), it has an overall positive charge in acidic environments (Giammarco et al. 2016). Therefore, it can be concluded that under the experimental conditions, the electrostatic bond played a minor role in VAN adsorption and cation exchange was the dominant mechanism in VAN removal by bentonite in acidic condition. It is also possible that acidic pH of the solution increases the surface positive charge, and therefore, by attracting the negative section of VAN molecules, enhances the adsorption performance of bentonite. Other studies also reported that cation exchange played a significant role in the adsorption of antibiotics like ciprofloxacin that have a positive charge on bentonite (Genç et al. 2013).

In a similar study, adsorption of enrofloxacin on montmorillonite was investigated, and it was found that the adsorption process was also pH-dependent and cation exchange was its dominant mechanism in acidic environments (Yan et al. 2013). As shown in Figure 2(a), when the pH of the solution is 9 (which is higher than pKa = 7.75 and pKa = 8.89 that are related to VAN functional groups), almost no adsorption takes place. Considering the pH_zpc obtained for bentonite, this may be due to the repelling force between the adsorbent surface and VAN molecules at high pH values. Given the dependence of the adsorption process on pH, and also to achieve high efficiency, the rest of the experiments were carried out at pH = 4.

Figure 2(b) reveals that VAN desorption from the bentonite surface was strongly dependent on solution pH because of the pH effect on bentonite surface charge. The desorption process happens at a very low rate when pH is below 9. However, VAN desorption increases with further increases in pH values. This mechanism can be explained by the fact that pH_zpc for bentonite is 9.1. Therefore, the bentonite surface has a net negative charge at pH values higher than 9.1. Moreover, an electrostatic repelling force is generated between VAN molecules and the negatively charged surface of bentonite particles due to deprotonation of the amide and carboxylic acid groups. Consequently, considerable VAN desorption happens at pH values higher than 9.

To study the effect of adsorbent dosage, the range of 0.2–1.2 g L⁻¹ was selected. Figure 3(a) presents VAN adsorption by bentonite at various adsorbent doses at initial VAN concentrations of 50, 100, and 150 mg L⁻¹. The adsorption capacity of bentonite decreased with increases in adsorbent dose at all concentrations. This can be explained by an increase in the number of active adsorption sites on the adsorbent surface. There will be more active adsorption sites for adsorbing VAN with increases in adsorbent dose. Because of the large number of these sites, some of them will be empty; hence, the entire adsorbent surface is not utilized in the adsorption process. This has also been reported in studies by other researchers (Genç et al. 2013; Derakhshani & Naghizadeh 2018). Figure 3(b) shows the effect of salt concentration on VAN adsorption by bentonite. Addition of NaCl greatly influenced the adsorption process so that increasing NaCl concentration from 0.001 to 0.1 M led to reduction in VAN adsorption from 95.69 to 15.96 mg g⁻¹. This decrease suggests that surface composition played a very small role in the adsorption process and also made it clear that cation exchange was probably the dominant mechanism in the adsorption process. Since VAN has a positive charge at pH = 4, the decrease in adsorption at higher Na⁺ concentrations could be due to competition between VAN and high Na⁺ concentrations for adsorption sites (Genç et al. 2013).

Initial concentrations of cations in desorption solution can play a significant role in the desorption process. Increases in concentrations of cations in the solution enhanced VAN desorption by bentonite (Figure 4).

VAN desorption from bentonite at both loadings of 80 and 236 mg g⁻¹ was improved when Na⁺, Ca²⁺, and Al³⁺ concentrations increased from 10 to 50, from 2 to 10, and from 0.5 to 3 mM, respectively. This indicates that the process of VAN desorption from bentonite mostly takes place through cation exchange, and agrees with other studies conducted on antibiotic desorption from clay surfaces (Wu et al. 2013). The low concentrations of Al³⁺ and Ca²⁺ compared to Na⁺, and the higher desorption rates in the presence of Al³⁺ and Ca²⁺ compared to Na⁺, suggest the dominance of the cation exchange mechanism in VAN adsorption. As further explained in the following section, Al³⁺ ions having smaller ionic radius compared to Ca²⁺ and Na⁺ ions, due to higher positive charges, justifies these results (Chang et al. 2014).

Pseudo-first- and pseudo-second-order models were employed to study the kinetics of VAN adsorption by bentonite nanoparticles. The parameters obtained from the kinetic models are listed in Table 3. Kinetic studies present valuable information on the effectiveness of the adsorption process and reaction rates. Since the pseudo-second-order kinetic model had higher R² values at all concentrations, it was clear that kinetics of VAN adsorption by bentonite followed the pseudo-second-order model. Furthermore, unlike the pseudo-first-order model, the values for adsorption capacity calculated by the equation (q_e-cal) and those obtained in the present research (q_e-exp) from the pseudo-second-order model were slightly different from each other. This also indicates that the kinetics of VAN adsorption followed the pseudo-second-order model. The results of VAN desorption were also fitted to the pseudo-first- and pseudo-second-order models. It was found that the pseudo-second-order model fitted the data best. Based on the results listed in Table 4, the R² values for the deionized water, NaCl, CaCl₂, and AlCl₃ were always higher than 0.978. According to this information, at the initial VAN loading of 236 mg g⁻¹, the initial desorption rates for the Al³⁺ and Ca²⁺ solutions and also the rate constant K were higher than those of the other two desorption solutions. This could be explained by the effect of electric charge. In cation exchange reactions, the smaller cations with higher electric charges show greater chemical affinity compared to larger cations with a lower electric charge (Chang et al. 2014).

The equilibrium relation between bentonite and VAN was examined using Langmuir, Freundlich, and Temkin isotherm models (Table 5).

Here, C_e is adsorbate equilibrium concentration (mg L⁻¹), and are Langmuir constants (L ⁠, the Freundlich isotherm constant (L ⁠), n the exponent in the Freundlich isotherm, the constant related to the heat of adsorption (R and T: the gas constant (8.314 J mol⁻¹ K⁻¹) and absolute temperature (K), respectively), and the equilibrium binding constant (L mol⁻¹) corresponding to the maximum binding energy.

The R_L value decreased from 0.238 to less than 0.043 when the initial concentration increased from 25 to 175 mg L⁻¹. This indicated that the adsorption process was desirable in the mentioned concentration range. The value of n in the Freundlich isotherm was 2.515, which proved the desirability of VAN adsorption on the bentonite (Ebrahimi et al. 2015). Moreover, the Temkin constant (⁠(Huang et al. 2017), which in a way determines whether the adsorption reaction is endothermic (⁠ < 1) or exothermic (⁠⁠, was 73.31 at 25 °C. This indicated that VAN adsorption on the bentonite was intrinsically an exothermic reaction.

Figure 5 shows the surface morphology of the adsorbent and its changes following adsorption. Based on FESEM images, bentonite nanoparticles have a laminar structure that includes a large number of pores on the surface. These sheets are estimated to be about 33 nm thick. Following VAN adsorption, the pores on the surface were occupied and the surface became more regular and smoother.

FTIR spectra of bentonite, VAN, bentonite after VAN adsorption, and bentonite after VAN desorption by Na⁺, Ca²⁺, and Al³⁺ solutions are presented in Figure 6. For bentonite clay, the tensile O-H peak at 3,630.15 is that of the Mg- or Al-enriched octahedral layer. The strong peak at 3,413.88 indicates the possible existence of hydroxyl bonding (Taha et al. 2017). The peak observed at 1,636.85 is related to the tensile alkenyl C=C band (Derakhshani & Naghizadeh 2018). The tensile Si-O frequency bands at 1,046.76 and 1,384 are related to hydrous silica and match the frequency band of montmorillonite (Taha et al. 2017). The peaks observed at 798.59 and 466.29 are those of Si-O-Si, and these frequencies prove the presence of quartz. The relatively weak peak at 623 may be introduced as vertical vibrations in the octahedral layer-like Al-O-Si and Mg-O-Si, and the peak at 525 is that of the flexural behavior of the Al-O-Si bond (Zhao et al. 2010; Taha et al. 2017; Derakhshani & Naghizadeh 2018). The specific tensile C-O-C band at 1,230, which is present in VAN, is also observed in bentonite clay after adsorption. The same trend is repeated in the case of the frequency band at 1,510 related to C=C. Furthermore, two displacements have also happened: frequency band 3,414 shifted to 3,422 and frequency band 1,637 shifted to 1,656. Altogether, these observations point to VAN adsorption by bentonite.

The study of the FTIR spectrum after desorption indicates that the peaks at 1,234 and 1,516 formed due to VAN adsorption disappeared following the desorption process. Of course, the presence of the very weak peak at 1,516 in the spectrum related to Na⁺ suggests that some VAN remained after desorption by Na⁺. The disappearance of the 1,516 and 1,234 bands may be due to the generation of surface complexes between aluminum and calcium cations and VAN (something that eventually results in increased VAN desorption from the bentonite clay surface). The behavioral study of the VAN desorption process from bentonite once again confirmed that cation exchange was the dominant mechanism in VAN adsorption.

Adsorption/desorption of VAN from aqueous solutions using bentonite was studied. Results indicate that both processes were strongly pH-dependent. Highest amounts of adsorption and desorption were reached in acidic and alkaline conditions, respectively. This was due to the surface charge of bentonite clay particles and pKa of VAN functional groups. Moreover, the results indicate that cation exchange may be the dominant mechanism of adsorption/desorption of VAN. Furthermore, it was found that VAN adsorption on bentonite followed the Langmuir adsorption model. The presence of Na⁺, Ca²⁺, and Al³⁺ in the solution increased VAN desorption from the bentonite surface. Desorption power of the solutions was in the following order: Al³⁺ > Ca²⁺ > Na⁺ > H₂O. Equilibrium was reached very quickly in both adsorption and desorption processes. After matching the data to the kinetic and isotherm models, it was found that they matched the pseudo-second-order kinetic and Langmuir isotherm models, respectively.

All relevant data are included in the paper or its Supplementary Information.