**Removal of Rhodamine B from Water Using a Solvent Impregnated Polymeric Dowex 5WX8 Resin: Statistical Optimization and Batch Adsorption Studies**

#### **Moonis Ali Khan 1,\* , Momina <sup>2</sup> , Masoom Raza Siddiqui <sup>1</sup> , Marta Otero <sup>3</sup> , Shareefa Ahmed Alshareef <sup>1</sup> and Mohd Rafatullah 4,\***


Received: 29 January 2020; Accepted: 19 February 2020; Published: 24 February 2020

**Abstract:** Herein, commercially available Dowex 5WX8, a cation exchange polymeric resin, was modified through solvent impregnation with t-butyl phosphate (TBP) to produce a solvent impregnated resin (SIR), which was tested for the removal of rhodamine B (RhB) from water in batch adsorption experiments. The effect of SIR dosage, contact time, and pH on RhB adsorption was studied and optimized by response surface methodology (RSM), interaction, Pareto, and surface plots. Scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FTIR) were respectively used for characterizing SIR surface morphology and identifying active binding sites before and after RhB adsorption. SEM showed that the pristine SIR surface was covered with irregular size and shape spots with some pores, while RhB saturated SIR surface was non-porous. FTIR revealed the involvement of electrostatic and π–π interactions during RhB adsorption on SIR. Dosage of SIR, contact time, and their interaction significantly affected RhB adsorption on SIR, while pH and its interaction with dosage and contact time did not. The optimum identified experimental conditions were 0.16 g of SIR dose and 27.66 min of contact time, which allowed for 98.45% color removal. Moreover, RhB adsorption equilibrium results fitted the Langmuir isotherm with a maximum monolayer capacity (*q*max) of 43.47 mg/g.

**Keywords:** modified polymeric resin; t-butyl phosphate impregnation; polymer based adsorbents; dye adsorption; response surface methodology

#### **1. Introduction**

Textile industries are among the largest consumers of water, dyes, and different types of chemicals, resulting in the generation of large volumes of highly toxic effluents. The discharge of these effluents without prior treatment can be lethal to the environment. Usually, the textile effluents are rich in color, pH, chemical oxygen demand, inorganic salts, turbidity, and temperature [1]. According to the United States Environmental Protection Agency (USEPA), textile waste is mainly divided into four principal classifications, namely hard-to-treat, high volume, dispersible, and hazardous and toxic wastes [2]. Rhodamine B (RhB) is one of the most widely used cationic water-soluble organic dyes, and it is toxic to

aquatic environments. It reduces sunlight penetration into water bodies, which can be lethal for aquatic life due to limited availability of oxygen for respiration [3–5]. Therefore, before effluent discharge, it is necessary to apply intensive treatment processes to minimize its concentration in water bodies.

Membrane filtration, flocculation, biological treatments, photocatalytic oxidation, and adsorption [6– 10] are some of the commonly used textile effluent treatment processes. On the other hand, the application of polymer materials in water treatment and selective sequestration has impressively developed in the last decades, with the production of novel materials and composites, post-polymerization modifications, introduction of functional groups, and development of supramolecular assemblies and nanomaterials [11–16]. However, these methods have their own limitations and efficiencies in terms of cost effectiveness. Recently, functionalized polymeric resins have become an alternative to commercial activated carbon and other adsorbents due to economic concerns and regeneration properties. Polymeric resins are characterized by their high surface area, moderate swelling, and narrow pore size distribution. In order to improve the adsorption characteristics of such resins, their surface properties can be modified using the advantage of adsorbate and adsorbent interaction. Mostly, ion-exchange resins are prepared by styrene divinylobenzene cross-linked co-polymer, which is comparatively lower in cost than activated carbon and serve several advantages as a matrix [17]. Additionally, the polystyrene based matrix has the potential to provide excellent chemical and physical stability together with resistance to degradation by oxidation or hydrolysis. Apart from ion-exchange resins, solvent extraction and liquid–liquid extraction, which work on the principle that solute distributes itself in a certain ratio with immiscible solvents, have attracted considerable attention in recent years [18]. The merits of solvent extraction include rapid and very selective separations that are usually highly efficient [18]. Solvent impregnated resins (SIRs) pose synergic merits of both ion-exchange and solvent extraction. A SIR is described as a liquid complexing agent dispersed homogeneously in a solid polymeric medium. In SIR removal processes, a specific solute is extracted from the aqueous phase to the organic phase inside the pores of the resin. The resin acts as carrier of the solvent and reduces the entrainment and irreversible emulsification that occur during solvent extraction [19]. Previously, a non-functional macroporous polymeric resin was used as a polymeric support for the removal of dyes from wastewater [20]. However, macroporous resins have lower retention capacity and slower kinetic diffusion compared to gel-type resins. Therefore, gel-type resins have higher removal efficiency than conventional macroporous polymeric resins. Moreover, the optimization of operational conditions for dye adsorption using SIR can improve the removal efficiency of dye. Thus, in this study, a gel-type Dowex cation exchange resin was used to remove RhB dye from aqueous solution. Moreover, the effect of interaction of operational conditions on the removal of RhB dye was also studied. Response surface methodology (RSM) is effective, reliable, and very comprehensive as compared to other conventional optimization processes [21]. It is a statistical tool that is very effective for designing, analyzing, and optimizing the effect of independent factors for the prediction of response output [21]. Therefore, the aim of this work was to investigate the operating conditions for removal of RhB dye using a 2<sup>3</sup> full factorial design. Moreover, an equilibrium batch study was performed at optimum conditions to study the mechanism of dye removal using SIR.

#### **2. Experimental**

#### *2.1. Chemicals, Reagents, and Adsorbent*

Dowex 5WX8 gel-type cation exchange resin (BDH, England, UK) with particle size 0.39–1.00 mm and 50%–58% moisture content was used as an adsorbent. The resin contained a styrene divinyl benzene matrix, having sulfonic acid as a matrix active functional group. Rhodamine B (RhB: C28H31ClN2O3) (Sigma-Aldrich, Darmstadt, Germany) with respective color index and molecular weight 45170 and 479.2 g/mol, synonymously known as basic violet 10, was used as an adsorbate. Tributyl phosphate (TBP: C12H27O4P) was obtained from Sigma-Aldrich, Darmstadt, Germany. All the chemicals and reagents used during the study were of analytical reagent (A.R) grade or as itemized. Ultra-pure deionized (D.I: Millipore, Burlington, MA, USA) water was used throughout the study.

#### *2.2. Synthesis of Solvent Impregnated Dowex 5WX8 Resin*

Initially, Dowex 5WX8 resin was washed with D.I water in order to remove inorganic impurities and monomeric material. Thereafter, the resin was overnight dried in an oven at 70 ◦C. The resin was impregnated with TBP (hydrophobic in nature) through the wet impregnation method reported elsewhere [22]. Briefly, undiluted TBP and resin in a volume to weight ratio of 6.0 was used to impregnate resin in a conical flask. Resin was aged for 24 h in TBP to achieve highest impregnation efficiency [20]. Further, the impregnated resin was separated from TBP through filtration and thoroughly rinsed with D.I water to remove unimpregnated traces of TBP. Then, solvent impregnated Dowex 5WX8 resin (SIR) was ready to use for adsorption studies.

#### *2.3. Characterization of Solvent Impregnated Dowex 5WX8 Resin*

Fourier transform infrared (FT-IR: Is10 Nicolet Thermo Scientific, Waltham, MA, USA) analysis was carried out to determine the available functional groups on Dowex 5WX8 resin and SIR (both pristine and RhB saturated) surfaces. The surface morphology of Dowex 5WX8 resin and SIR (both pristine and RhB saturated) was analyzed by scanning electron microscopy (SEM: Zeiss, model EVO Ma10, Oberkochen, Germany).

#### *2.4. Batch Scale Adsorption*

The RhB adsorption studies over SIR were performed at room temperature by varying operation parameters viz. initial pH (pH<sup>i</sup> : 2–8), SIR dose (*m*: 0.1–0.5 g), and contact time (*t*: 5–30 min). A series of 10 mL RhB solutions of initial concentration *Co*: 100 mg/L were equilibration with 0.1–0.5 g SIR in 25 mL conical flasks over a shaker at 230 rpm. At predetermined contact times, solid/solution phases were separated, and residual RhB concentrations in solutions were determined by using a Shimadzu UV-Visible Spectrophotometer at λmax: 554 nm. The amount of RhB adsorbed at any time *t* onto SIR was calculated as:

$$\text{Adsorbed concentration at time t } (q\_t \,\text{mg/g}) = (\text{C}\_o - \text{C}\_l) \times \frac{V}{m} \tag{1}$$

where V (L) is the volume of RhB solution, *C<sup>o</sup>* (mg/L) is the initial RhB concentration, *C<sup>t</sup>* (mg/L) is the remaining RhB concentration in solution at any time *t*, and *m* (g) is the mass of SIR.

The amount of RhB adsorbed on SIR at equilibrium, which was attained in 30 min under shaking, was calculated as:

$$\text{Adsorbed concentration at equilibrium } (q\_{\varepsilon\prime} \text{ mg/g}) = (\mathbb{C}\_o - \mathbb{C}\_e) \times \frac{V}{m} \tag{2}$$

where *C<sup>e</sup>* (mg/L) is the equilibrium concentration of RhB in solution.

The decolorization efficiency (D.E, %) was calculated as follows:

$$\text{Decolourization} \, efficiency \, (\text{D.E.}\,\%) = \frac{\text{C}\_o - \text{C}\_\varepsilon}{\text{C}\_o} \times 100\tag{3}$$

#### *2.5. Design of Experiments and Optimization of Parameters*

Two level (low level –1 and high level +1) factorial design (2<sup>3</sup> ) of response surface methodology (RSM) was applied for three independent variables (factors), namely the operational parameters dosage of SIR (A), contact time (B), and initial pH (C), to predict D.E. (%) (response factor, *y*) from RhB dye solution using solvent impregnated resin (Table 1). A total 12 runs including 4 times of replication for center point was carried out (2<sup>3</sup> = 8 runs; 8 runs + 4 replications for center point = 12 runs) using Design Expert (6.0.10) (Stat Ease, Minneapolis, MN, USA). Suitable approximation can be determined for the true functional relationship between the process response, *y*, and the set of factors by first order model or second order model. When the response linearly varies with the independent variable, then the first order model, which is given by Equation (4), is satisfied.

$$y = \beta\_0 + \beta\_1 \mathbf{x}\_1 + \dots \beta\_k \mathbf{x}\_k + \varepsilon \tag{4}$$

where, *y* is the response, β<sup>0</sup> is the offset term, β1, . . . , β*<sup>k</sup>* are the effect term, *x1,* . . . *, x<sup>k</sup>* are the independent variables, and ε is the random error term. When a curvature is detected in the system, second order model is selected and expressed by the following equation (Equation (5)):

$$y = \beta\_0 + \sum\_{i=1}^{k} \beta\_1 \mathbf{x}\_1 + \sum\_{i=1}^{k} \beta\_{ii} \mathbf{x}\_i^2 + \sum\_{i=1}^{n} \sum\_{i$$

where *y* is the predicted response, β<sup>0</sup> is the constant, β<sup>1</sup> is the linear effect, β*ii* is the square effect, and β*ij* is the interaction effect.


**Table 1.** Factors, levels, and ranges of the parameters considered for the factorial design.

For the optimization of experimental design, the statistical software Minitab 16 was used. The results were analyzed by estimating the response of the dependent response variable to obtain the effects, coefficient, and other statistical parameters. The conditions for optimization of adsorption process were obtained from Minitab 16 (Minitat LLC., Penn State University, PA, USA) as well. By using the analysis of variance (ANOVA), the determination coefficient (r<sup>2</sup> ) and statistical significance were determined.

#### *2.6. Isotherm Modeling*

Langmuir and Freundlich isotherm models in linearized forms were fitted to data on RhB adsorption onto SIR. The Langmuir isotherm model, which assumes monolayer adsorption over homogenous sites on the adsorbent surface and equal activation energy for each molecule, is given by Equation (6) in its linearized form:

$$\frac{1}{q\_e} = \frac{1}{q\_{\text{max}} K\_L \mathbb{C}\_e} + \frac{1}{q\_{\text{max}}} \tag{6}$$

where *q<sup>e</sup>* (mg/g) is the amount of RhB adsorbed on SIR, *C<sup>e</sup>* (mg/L) is the saturated amount of RhB adsorption at equilibrium concentration, and *qmax* (mg/g) is the maximum monolayer adsorption capacity of RhB on SIR. The constants *K<sup>L</sup>* and *qmax* can be calculated from a linear plot of 1/*q<sup>e</sup>* vs. 1/*C<sup>e</sup>* . The characteristics of the fitting to the Langmuir equation are given by a dimensionless number, *R<sup>L</sup>* (Equation (7)), which indicates the type of isotherm to be irreversible (*R<sup>L</sup>* = 0), favorable (0 < *R<sup>L</sup>* < 1), linear (*R<sup>L</sup>* = 1), or unfavorable (*R<sup>L</sup>* > 1) [23].

$$R\_L = \frac{1}{1 + K\_L \mathcal{C}\_o} \tag{7}$$

The Freundlich isotherm, which is an empirical model, is usually associated with multilayer adsorption of RhB molecules over heterogenous adsorption sites and can be expressed in linearized form as:

$$
\log q\_\ell = \log K\_F + \frac{1}{n} \text{ log } \mathbb{C}\_\varepsilon \tag{8}
$$

where *K<sup>F</sup>* is a Freundlich constant, and *n* is a parameter related to the binding strength changes with the adsorption density. If 1/*n* = 0, it indicates that the extent of adsorption is independent between two phase concentration; 1/*n* < 1 indicates favorable chemical adsorption; 1/*n* > 1 indicates a cooperative adsorption [23].

#### **3. Results and Discussion**

#### *3.1. Pre and Post-Adsorption Characterization*

The surface morphologies of Dowex 5WX8 resin and SIR (both pristine and RhB saturated) were analyzed using SEM with 300X magnification, as illustrated in Figure 1a–d. The raw Dowex 5WX8 resin has a smooth surface with some pores (Figure 1a). After TBP solvent impregnation over raw Dowex 5WX8 resin, the whole SIR surface was covered with spots of irregular size and shape (Figure 1b). This confirms successful impregnation of raw Dowex 5WX8 resin [24]. The structural pores after impregnation remained unchanged, as shown in Figure 1c. These pores were well occupied by RhB molecules during adsorption, displayed by protruding occupation of pores (Figure 1d). The FT-IR spectrum of raw Dowex 5WX8 resin (Figure 1e) showed a strong band centered at 3420 cm–1, ascribed to hydroxyl (–OH) group stretching, due to the presence of internal moisture in raw Dowex 5WX8 resin. The conjoint bands at 2927 and 2852 cm–1 were associated with C–H stretching vibrations for saturated aliphatic species. The bands between 1483 and 1510 cm–1 were due to CH<sup>3</sup> deformation in amino acid or hydrochloride compounds in raw Dowex 5WX8 resin. Moreover, the bands at 1010 and 1033 cm–1 represented C–O stretching of cyclic alcohol in raw resin. However, after impregnation of raw Dowex 5WX8 resin with TBP, the band at 1033 cm–1 became sharp and intense due to P–O–C stretching, thus confirming the attachment of phosphorous with C–O. A band at 1226–1237 cm–1 was due to P=O stretching in the phosphate group, and a band at 1383–1388 cm–1 showed CH<sup>3</sup> deformation in the t-butyl group. The presence of P=O and P–O–C groups, and CH<sup>3</sup> deformity indicate successful impregnation of raw Dowex 5WX8 resin to form SIR. After RhB adsorption on SIR, the band at 1033 cm–1 was displaced by a low intensity band at 1034 cm–1, confirming its involvement in binding dye molecules during adsorption.

**Figure 1.** *Cont*.

**Figure 1.** Scanning electron microscopic (SEM) images of raw Dowex 5WX8 polymeric resin (**a**), pristine SIR (**b, c**), RhB saturated SIR (**d**), and Fourier transform infrared (FT-IR) spectra of raw Dowex 5WX8 polymeric resin (**i**), pristine SIR (**ii**), and RhB saturated SIR (**iii**) (**e**)**.**

#### *3.2. Screening of Process Independent Variables*

In this batch adsorption study, the interaction of operational parameters for the removal of RhB using SIR was examined. A total of 12 experimental runs were optimized using three dominant parameters for the removal of RhB from aqueous solution, which was calculated using Equation (3). It was used to achieve improved adsorption capacity of SIR by possible interaction of operational parameters (Table S1, Supplementary Materials). The results showed that two of the considered operational parameters, specifically dosage of SIR and contact time, influenced the removal of RhB from aqueous solution. Additionally, the surface modification of Dowex 5WX8 resin by impregnation to SIR led to improved surface characteristics for the removal of RhB.

The effect of variable interaction during the adsorption process was carried out using analysis of variance (ANOVA). Then, 2<sup>2</sup> fractional factorial designs were used to study the selected factors. The purpose of carrying out the fractional factorial design was to determine the factors that had a significant effect on D.E. (%). ANOVA for the fractional factorial design is given in Table 2. The ANOVA and response surface regression of D.E. (%) is tabulated in Table 3. The effect of pH for aqueous phase was found to be insignificant due to the p-value of 0.314, which was greater than 0.05. Moreover, the negative coefficient (–1.08) of pH pointed towards a decrease in adsorption efficiency as the pH increased. The overall prediction of the output model in terms of operational parameters showed that the model was suitable for predicting the adsorption of RhB on SIR (p < 0.05). The respective p-values of SIR dosage and contact time were 0.063 (nearer to 0.05) and 0.012, which showed that both parameters significantly influenced D.E. (%). Furthermore, the linear effects of both SIR dosage and contact time indicated that they were suitable for improving adsorption efficiency.

 = 91.374 + 11.215 + 17.214 − 20.055 However, quadratic coefficients of SIR dosage and contact time inhibited the performance of RhB adsorption from aqueous solution (p > 0.05). Moreover, the negative coefficient of interaction of dosage and contact time was helpful in increasing adsorption efficiency. The polynomial first order and interactive regression model equation was developed using Minitab software. Therefore, the model equation is given as:

$$y = 91.374 + 11.215A + 17.214B - 20.055AB \tag{9}$$

where *y* is the D.E. (%) of RhB. In the aforementioned equation, a synergistic effect was indicated by the positive sign, while an antagonistic effect was indicated by the negative sign [25]. The ANOVA for the model is given in Table 4. It was deduced that the color removal was significant at 95% (*p* < 0.05) confidence level, which shows the validity of the model for RhB adsorption onto SIR.


**Table 2.** ANOVA analysis for the fractional factorial design carried out to determine the factors that have significant effect on the decolorization efficiency (D.E. (%)).

SE Coefficient = standard error of the coefficient.

**Table 3.** Estimated regression coefficients and ANOVA for optimization of decolorization efficiency (D.E. (%)).


SE Coefficient = standard error of the coefficient.

**Table 4.** ANOVA analysis of the model for the removal of RhB using the produced SIR.


The determination of coefficient (r<sup>2</sup> ) was used to evaluate the quality of the developed model [7,26]. The r<sup>2</sup> value of the color removal was 82.17%, which means that 0.8217 of total variation was explained by the model, while 17.83% of the variation was left unexplained. The importance of the effects of the operation variables and their interaction can be best described by the Pareto chart, as shown in Figure 2a. A student's t-test was performed to determine whether the calculated effects were significantly different from zero; these values for each effect are shown in the Pareto chart by horizontal columns [7]. The t-value for 95% confidence level was 2.013. The values exceeding the reference line were considered as significant for 95% confidence level, whereas values below the reference line were considered as insignificant. As shown in Figure 2a, the two parameters, dosage of SIR (A) and contact time (B), as well as their interaction (AB) were found to be significant at the 0.05 level. However, the effect of pH (C) and its interaction with dosage (AC) and contact time (BC) were below the reference line, which points to their insignificance for D.E. (%). According to previous studies, the effect of increasing pH was considered as favorable for the removal percentage of RhB [27,28]. However, in this

study the effect of pH was not found to be significant. Therefore, the effects of dosage, contact time, and their interaction, which resulted in 97.45% of dye adsorption efficiency, were studied.

**Figure 2.** Pareto chart showing the effects and interactions of operational variables, namely dosage of resin (g), contact time (min) and pH, on the decolorization efficiency (D.E. (%)) by SIR (**a**)**,** and interaction plot of decolorization efficiency (D.E. (%)) by SIR versus contact time (min) for the different dosages of resin, namely 0.1, 0.2, and 0.3 g of SIR (**b**).

The interaction of contact time and dosage of resin were described by interaction plots, illustrated in Figure 2b. The interaction plots show the D.E. (%) versus the contact time (min) for each dosage of SIR. It was found that as the contact time increased, the D.E. (%) increased and reached its maximum for a 0.3 g dosage of SIR. The interaction between contact time and SIR dosage improved the adsorption efficiency of RhB, as shown in Figure 3a. The response surface plots show the estimated value of D.E. (%) (the height of the surface represents the value of D.E. (%)) as a function of the independent variable. It must be highlighted that the surface plots represent the same results as observed in interaction plots.

**Figure 3.** Surface plot showing the decolorization efficiency (D.E. (%)) as a function of the dosage of resin (g) and contact time (min) under a shaking speed of 230 rpm and pH 3.6 (**a**), and optimization plot for the determination of the optimum conditions, namely dosage of resin (g) and contact time (min), for a maximum decolorization efficiency (D.E. (%)) by SIR (**b**).

#### *3.3. Optimization of Experiment*

A standard RSM design called central composite design (CCD) was used to optimize the operating parameters. Optimum conditions of effective parameters with minimum number of experiments can be determined by this statistical technique. This method is also suitable to analyze the interaction and relationship between each parameter. The optimum conditions for the removal of RhB dye were obtained from the screening of operational parameters (Table S1, Supplementary Materials). The optimum operational parameters for D.E. (97.45%) were achieved with 0.3 g SIR dosage and 30 min contact time. Since contact time played a major role in the extraction process [20], therefore extraction efficiency of TBP impregnated SIR increased steadily with time until it reached equilibrium. Moreover, an increasing amount of SIR increases the D.E. (%) because low amounts of SIR contain low amounts of extractant [20]. However, high amounts of SIR and TBP cause an increase of the dye solution acidity. Therefore, concentrated acidic medium causes back-extraction of extractant–dye complex, thus resulting in higher dye concentration [29].

The optimization plot, which is displayed in Figure 3b, was used to find optimum conditions for RhB removal by SIR. From the analysis of experimental data obtained, the optimum identified conditions were 0.16 g of SIR dose and 27.66 min of contact time. With the application of such optimum conditions, the predicted value of D.E. was 98.45%, which was experimentally verified to be fulfilled with a deviation of ± 0.1%.

#### *3.4. Adsorption Isotherm*

The adsorption isotherm is used to study the mechanism and pattern of adsorption at liquid-phase equilibrium [21,30,31]. Fittings of equilibrium data for the adsorption of RhB on SIR by Langmuir and Freundlich models were determined in this work. Linear plots of Langmuir and Freundlich isotherms and the respective parameters are shown in Figure S1 (Supplementary Materials). As can be seen, equilibrium results fitted the Langmuir isotherm (r<sup>2</sup> = 0.99) but not the Freundlich model (r <sup>2</sup> = 0.087). Therefore, it may be assumed that the adsorption of RhB on SIR was the monolayer on the surface of SIR, where the active sites and energies were homogenously distributed [32,33]. The fitted maximum monolayer adsorption capacity (*qmax*) of RhB on SIR was found to be 43.47 mg/g, and *K<sup>L</sup>* was 0.0126 L/mg. Therefore, adsorption of RhB on SIR was found to be favorable because R<sup>L</sup> was calculated (Equation (7)) to be 0.284, which is greater than 0 and smaller than 1.

#### *3.5. Adsorption Mechanism*

Dowex 50WX8, a cation exchange polymeric resin, was used in this study. It contains cross-linked styrene divinyl benzene co-polymer with sodium sulfonate groups as ion-exchange sites. The resin was impregnated with TBP, which contains certain functional groups that have significant influence on the adsorption of RhB dye (as revealed by FT-IR results in Figure 1e). Therefore, an increased RhB adsorption can be achieved using the interaction between the adsorbate (RhB) and modified resin (SIR). The structure and functional groups present on modified SIR resin were the main factors responsible for the adsorption of RhB dye on this resin. Rhodamine B dye has amino and carboxylic functional groups, which can be involved on its adsorption on modified SIR. According to FTIR results, the peaks for P–O–C disappeared after adsorption of dye on impregnated resin and were replaced by the C–O group of the RhB dye. On the other hand, phosphate was not detected in solution after the adsorptive removal of RhB. Therefore, phosphate groups might be responsible for binding the positively charged dye ions by modified SIR resin. Furthermore, the possible interaction that might be occurring between modified SIR resin and RhB dye can be electrostatic and π–π bonding, as shown in Figure 4.

**Figure 4.** Schematic representation of SIR production and RhB adsorption mechanism on SIR.

#### **4. Conclusions**

A TBP impregnated polymeric resin was produced in this work and tested for the removal of RhB from water. FT-IR analysis showed the presence of a phosphate functional group on the surface of the solvent impregnated resin (SIR), which was indicative of successful impregnation of TBP over the resin. The effect of operational conditions, namely pH, adsorbent dosage, and contact time, on the adsorption of RhB onto SIR was studied and optimized. It was found that pH does not have a significant effect on the D.E. (%). The maximum color removal obtained was 97.45% at 100 mg/L of initial dye concentration, 230 rpm of shaking speed, pH 3.6, 0.3 g of resin dosage, and 30 min of contact time. The optimum conditions for the adsorption of RhB by SIR were identified as 0.16 g of SIR resin and 27.66 min of contact time, which gave 98.45% of color removal. The adsorption data fitted well the Langmuir isotherm model, which pointed to monolayer adsorption on the SIR surface, with

π π

homogeneous distribution of active sites and energies. Furthermore, interaction of RhB and SIR was inferred to be electrostatic and π–π bonding.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/2/500/s1, Figure S1: Langmuir (a) and Freundlich (b) plots for the adsorption of RhB on SIR; Table S1: Results of response surface methodology design.

**Author Contributions:** Data curation, S.A.A.; Supervision, M.R.; Validation, M.R.S.; Visualization, M.O.; Writing—original draft, M.A.K. and M.; Revision, M.A.K. and M.O. All authors have read and agreed to the published version of the manuscript.

**Funding:** Deanship of Scientific Research, King Saud University: Research Group No. RG-1437-031.

**Acknowledgments:** The authors would like to extend their sincere appreciation to the Deanship of Scientific Research at King Saud University for funding this work through Research Group No. RG-1437-031. Furthermore, Marta Otero would like to thank FCT/MCTES for the financial support to CESAM (UID/AMB/50017/2019), through national funds and support by the FCT Investigator Program (IF/00314/2015).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Investigation of Biocidal E**ff**ect of Microfiltration Membranes Impregnated with Silver Nanoparticles by Sputtering Technique**

### **Aline M. F. Linhares 1,\*, Cristiano P. Borges <sup>2</sup> and Fabiana V. Fonseca <sup>1</sup>**


Received: 18 June 2020; Accepted: 24 July 2020; Published: 29 July 2020

**Abstract:** Silver nanoparticles were loaded in microfiltration membranes by sputtering technique for the development of biocidal properties and biofouling resistance. This technology allows good adhesion between silver nanoparticles and the membranes, and fast deposition rate. The microfiltration membranes (15 wt.% polyethersulfone and 7.5 wt.% polyvinylpyrrolidone in *N*,*N*-dimethylacetamide) were prepared by phase inversion method, and silver nanoparticles were deposited on their surface by the physical technique of vapor deposition in a sputtering chamber. The membranes were characterized by Field Emission Scanning Electron Microscopy, and the presence of silver was investigated by Energy-Dispersive Spectroscopy and X-ray Diffraction. Experiments of silver leaching were carried out through immersion and filtration tests. After 10 months of immersion in water, the membranes still presented ~90% of the initial silver, which confirms the efficiency of the sputtering technique. Moreover, convective experiments indicated that 98.8% of silver remained in the membrane after 24 h of operation. Biocidal analyses (disc diffusion method and biofouling resistance) were performed against *Pseudomonas aeruginosa* and confirmed the antibacterial activity of these membranes with 0.6 and 0.7 log reduction of viable planktonic and sessile cells, respectively. These results indicate the great potential of these new membranes to reduce biofouling effects.

**Keywords:** silver nanoparticles; microfiltration; membranes; biofouling; sputtering

#### **1. Introduction**

Microfiltration process (MF) has many consolidated advantages over conventional separation processes, mainly due to its ease of operation and low energy consumption, being widely used for disinfecting water. An evaluation of the global MF membrane market has indicated an expected annual growth rate of 9.0% from 2018 to 2023 [1]. However, membranes' fouling and biofouling are major drawbacks, reducing permeate flux and increasing operational costs. Usually, authors consider that biofouling can be one of the most difficult deposits on membranes to eliminate, highlighting the fast microbes growth, even at low nutrient concentrations [2–4].

To minimize such problems, one approach is to modify the properties of the surface of the membrane. For instance, silver nanoparticles (AgNps) are known for their bactericidal characteristics, which can work against microorganisms growth and, consequently, against biofouling on membranes [2,5–7]. Besides water disinfection, microfiltration membranes with biofouling resistance can be used for several applications such as membrane bioreactors and pretreatment for nanofiltration and reverse osmosis.

The biocidal mechanism of silver nanoparticles is not completely understood. However, the most accepted hypothesis is that silver ions interact with thiol groups in proteins, resulting in inactivation of enzymes and leading to the production of reactive oxygen species (ROS). Another important mechanism is the adhesion of AgNps to the surface of microorganisms, which alters the exchange of nutrients, salts, and water. DNA damage, resulting from AgNps penetrating the bacterial cell, can be highlighted as well [8–14].

On the other hand, the use of AgNps membranes for the disinfection of water should consider any possibility of risks to human health. It is known that levels of silver up to 0.1 mg L−<sup>1</sup> can be tolerated. However, long-term exposures to silver at high concentrations can generate skin darkening (argyria) [15–17]. Although the existing standards protect consumers from nanoproducts, there are still gaps in the assessment of risks for humans and some aspects need to be optimized, such as limits of toxicity, dose, and concentration to aquatic organisms and humans [14,16,18,19].

Silver nanoparticles may be synthesized by many methods and chemical reduction is the most commonly used. However, these routes present several reaction steps, and their residual solvents depict environmental concerns. Furthermore, AgNps tend to aggregate during their preparation in solution, thus the use of coating agents is necessary, which may reduce their antibacterial activity [2,20].

Physical techniques of vapor deposition, such as sputtering, allow the modeling of the size and distribution of the particles, as well as the high deposition rate and the good adhesion between AgNps and membranes. This technique uses high-energy ions to carry atoms from a target, which acts as a cathode, and deposit them onto a substrate. The last one acts as the anode in a sputtering chamber filled with inert gas. The releasing of Ag plasma ions moves with high kinetic energy and condenses as nanoparticles on the supporting material [21,22].

Previous studies reported concerning the inhibition of microorganism growth after contact between bacteria suspension and polymeric membranes impregnated with nanoparticles have been published. Among them, many reported the biocidal effect of AgNps membranes, but for short periods of time and with a continuous release of AgNps. This issue reduces the bactericidal properties of the membrane over time and can cause overestimations of the overall efficiency of the membrane [7,20,23]. For example, Dong et al. (2017) [7] showed 100% of mortality of *Escherichia coli* and *Bacillus subtilis* suspensions. Furthermore, Dong et al. (2019) [24] observed a significant suppression of *Serratia marcescens* using membranes loaded with AgNps. However, it should be considered that these studies evaluated the silver loss, either through membrane immersion or filtration process, only in short-term periods. On the other hand, Park et al. (2016) [20] verified the strong antibacterial activity against *E. coli*, *Pseudomonas aeruginosa*, and *Staphylococcus aureus*, even though the membrane was expected to last no longer than 97 days, while the estimation of Liu et al. (2015) [23] was of 340 days. Thus, nanoparticles leaching still poses a challenge to overcome, which suggests the development of new techniques to solve this problem [25,26].

In this work, silver nanoparticles were loaded on the surface of polymeric membranes by sputtering technique, aiming at the development of membranes with biocidal properties that would be resistant to biological fouling and capable of being used in the disinfection of water.

Silver leaching was extensively investigated through immersion and convective experiments.

#### **2. Materials and Methods**

#### *2.1. Materials*

Polyethersulfone (PES, MW 58 kDa) and polyvinylpyrrolidone (PVP, MW 360 kDa) were purchased from Basf, Ludwigshafen am Rhein, Germany and Sigma-Aldrich, St. Louis, MI, USA, respectively. The common solvent for both polymers was *N*,*N*-dimethylacetamide (DMAc, 99.5%), purchased from Tedia, Fairfield, Ohio, EUA. For microbiological experiments, deionized water was supplied by a Milli-Q apparatus (Merck KGaA, Darmstadt, Germany). Yeast extract, meat peptone, and agar were purchased from Kasvi, São José do Pinhais, Brazil. Magnesium sulfate, potassium phosphate

monobasic, potassium phosphate dibasic, and glycerol were purchased from Vetec, Duque de Caxias, Brazil. These reagents were used as culture medium. PES and PVP were dried at 60 ◦C overnight before being used and the other reagents were used as received.

#### *2.2. Microfiltration Membrane*

The microfiltration membranes were prepared by phase inversion method [27]. Briefly, a polymer solution with 15 wt.% PES and 7.5 wt.% PVP in DMAc was prepared by continuous stirring at room temperature. After degassing overnight, the solution was cast onto a glass plate with a 200 µm thick casting knife and exposed to the ambient atmosphere (60% RH) for 100 s prior to immersion in a precipitation bath composed of 80 wt.% DMAc and 20 wt.% deionized water. After complete precipitation (~15 min), the resulting membrane was immersed three times in a deionized water bath for 2 h to remove any residual solvent.

#### *2.3. Silver Nanoparticles Deposition*

Silver nanoparticles were directly deposited on the surface of the microfiltration membrane (4.7 cm in diameter) by sputtering (Quorum Q150R ES, Quorum Technologies, Laughton, UK) at room temperature and under a low-pressure argon atmosphere (0.1 Pa). A silver target (99.9% Ag, 57 mm in diameter and 0.1 mm thick, Sigma Aldrich, St. Louis, MI, USA) located in the center of the vacuum chamber acts as a cathode and the MF membranes are used as a substrate for deposition. The Ag plasma ions move with high kinetic energy to be condensate as nanoparticles on the surface of the membrane.

The MF-AgNps membranes were obtained with 15 mA (power = 6.0 W) and 50 mA (power = 26.0 W) of sputtering current and 15 and 120 s of deposition time, respectively, as described elsewhere [28]. These membranes are referred to as MF-15mA-15s and MF-50mA-120s, respectively. The chosen conditions aimed to investigate the biocidal properties of the membranes with different content of silver (8.22 and 317.78 mg m−<sup>2</sup> on MF-15mA-15s and MF-50mA-120s, respectively). For comparison purposes, a membrane without sputtering treatment was used as a control (MF-membrane).

#### *2.4. MF-AgNps Membrane Characterization*

The surfaces of MF-AgNps membranes were verified by Field Emission Scanning Electron Microscopy (FESEM, ZEISS Auriga 40, Ulm, Germany) and the presence of silver was evaluated by Energy-Dispersive Spectroscopy (EDS). Samples were coated with carbon using a metallizer (Quorum Emitech K550, Quorum Technologies, Kent, UK) for FESEM and with gold for EDS.

X-ray Diffraction (XRD, Rigaku Miniflex II, Rigaku, The Woodlands, TX, USA) was also used to verify the silver nanoparticles loaded by the sputtering technique and to assess their crystallinity. The data were collected in the 2-theta range of 5◦ to 90◦ and scanning speed of 0.05 s−<sup>1</sup> . The average size was calculated using the Debye-Scherrer formula, presented in Equation (1):

$$D\_p = \frac{K\lambda}{\beta\_{1/2}\cos\theta} \tag{1}$$

where *D<sup>p</sup>* is the crystallite size, *K* is a numerical factor referred to as the crystallite-shape factor (*K* = 0.9 is a good approximation), λ is the wavelength of the X-rays (λ = 1.5418 Å), β1/<sup>2</sup> is the full-width at half-maximum of the X-ray diffraction peak in radians, and θ is the diffraction angle.

Fourier Transform Infrared Spectroscopy (FTIR) analysis was conducted on Agilent Cary 630, Santa Clara, Califórnia, EUA (wavenumber range 500 to 4000 cm−<sup>1</sup> ; 32 scans at a resolution of 4 cm−<sup>1</sup> ).

Silver leaching from MF-AgNps membranes was evaluated through their immersion in water and their performance in convective experiments.

−

The immersion tests were performed following an established protocol that has been extensively reported in the literature [2,20,23,28,29]. The MF-AgNps membrane coupons (4.7 cm in diameter) were soaked in 50 mL of deionized water at 25 ◦C and stirred at 60 rpm in a shaker. Water samples were collected after 1, 4, and 24 h of immersion. Long duration tests were also conducted over 1 and 10 months of immersion without intermediate sampling. All samples were acidified to pH 2.0 with HNO<sup>3</sup> (2% v/v), and then analyzed by inductively coupled plasma-atomic emission spectrometry (ICP-AES) to quantify the amount of dissolved Ag. = ܬ ܸ ݐ × ܣ − − ⁻ −

Silver content in the membranes and silver loss percentage were also quantified by the digestion of other coupons of the MF-AgNps membranes that were immersed into HNO<sup>3</sup> (10% v/v) and sonicated for 3 h to ensure that all the AgNps present in the membrane were leached to solution. The total amount of Ag deposited onto the MF-AgNps membranes was assessed by ICP-AES [5,20]. ܴ ሺ%ሻ = 1 − ܥ ܥ × 100

The convective experiments were performed in a cross-flow membrane system, as illustrated in Figure 1, which had an effective membrane area of 45 cm<sup>2</sup> .

**Figure 1.** Cross-flow membrane system for silver leaching experiments.

The membrane system was designed to work with or without recirculation of permeate and retentate to the feed tank (TQ-01). The system consisted of a pump (BB-01) that transfers the solution from the feed tank to the permeation cell (C-01) with a flowrate meter (FI-01) between them. The permeate stream could be collected or recirculated by switching the valve (VE-01), and the retentate stream was continuously recirculated. The system's pressure was measured by a control pressure gauge (PI-01) and adjusted through a valve (VG-01).

The silver leaching experiments were conducted with the operational pressure set to 1.5 bar and a flow rate of 40 L h−<sup>1</sup> .

Additionally, the water flux was calculated using Equation (2) at specific transmembrane pressure, and the permeability was acquired by the slope of the curve *J<sup>p</sup>* × pressure (i.e., 0.5 to 1.5 bar).

$$J\_p = \frac{V}{A \times t} \tag{2}$$

where *J<sup>p</sup>* is the water flux (L h−1m−<sup>2</sup> ), *V* is the permeate volume (L), *A* is the effective membrane area (m−<sup>2</sup> ), and *t* is the filtration time (h).

*P. aeruginosa* suspension (10<sup>8</sup> UFC mL−<sup>1</sup> ) was used as an organism probe to evaluate the rejection capacity (Equation (3)) of each membrane at 1.5 bar.

$$R\ \left(\%\right) = 1 - \frac{\mathcal{C}\_p}{\mathcal{C}\_f} \times 100\tag{3}$$

where *C<sup>p</sup>* and *C<sup>f</sup>* is the final viable *P. aeruginosa* concentration of the permeate stream and feed solution, respectively.

#### *2.5. Antibacterial Activity Tests*

For the antibacterial activity tests, Gram-negative bacterium *P. aeruginosa* was selected as the model organism. *P. aeruginosa* is considered the paradigm organism for microbial biofilm studies due to its ability to quickly adhere to many different surfaces, its high reproduction rate, and its significance as a pathogen [25,30,31].

*P. aeruginosa* cells were inoculated into liquid culture medium and incubated with continuous stirring at 200 rpm overnight at 30 ◦C. This cell suspension served as a bacterial stock solution, which was further diluted to a specific concentration for each test.

The liquid culture medium was prepared with 5.0 g L−<sup>1</sup> yeast extract, 5.0 g L−<sup>1</sup> meat peptone, 0.2 g L−<sup>1</sup> magnesium sulfate, 7.0 g L−<sup>1</sup> potassium phosphate dibasic, 3.0 g L−<sup>1</sup> potassium phosphate monobasic, and 30 g L−<sup>1</sup> glycerol. The solid culture medium was produced from the same solution with the addition of 18 g L−<sup>1</sup> agar.

#### 2.5.1. The Disc Diffusion Method

The antibacterial activity of the MF-AgNps membranes was first investigated by a disc diffusion method against *P. aeruginosa*. Membrane coupons (17 mm in diameter) were previously sterilized by ultraviolet irradiation for 15 min. Then, the upper surface of the membranes, which holds the silver nanoparticles, was put in contact with the agar plates containing *P. aeruginosa* bacteria at a concentration of 10<sup>6</sup> colony forming units per mL (CFU mL−<sup>1</sup> ).

After incubation at 30 ◦C for 24, 48, and 72 h, the presence of inhibition zones was monitored and recorded by a digital camera. This inhibition ring, without microbial growth, served as an indicator of antibacterial activity. Furthermore, MF membrane (without AgNps) and an agar plate without membrane were also observed as control samples. All tests were made in triplicate.

#### 2.5.2. The Biofouling Resistance Tests

The biofouling resistance test was performed to evaluate the activity of AgNps in the prevention of bacterial adhesion on the membrane surface. MF-AgNps membrane coupons (1 cm<sup>2</sup> ) were immersed into 10<sup>7</sup> CFU mL−<sup>1</sup> *P. aeruginosa* suspensions and incubated for 24 h at 30 ◦C and 200 rpm stirring. Samples with MF membrane coupons and without membranes were also investigated as controls. All tests were made in triplicate.

After incubation, the planktonic cells in the supernatant and the sessile cells in the biofilm were counted. For total planktonic cells, the optical density at 600 nm (Shimatsu Mini 1240, Mumbai, India) was monitored and the bacterial concentration was determined. For viable planktonic cells, tenfold dilutions were spread onto agar plates and incubated overnight, and viable bacterial colonies were counted on the following day [5,31].

To measure the number of cells attached to the surface of the membrane (sessile cells), the coupon was rinsed with 20 mL of normal saline (0.9 wt.%) to ensure the removal of unattached cells. Then, the membrane coupon was placed in 15 mL of liquid culture medium and vortexed on the highest setting for 120 s in order to cause biofilm disruption. This supernatant was analyzed through optical density at 600 nm for total sessile cells. For the quantification of viable sessile cells, tenfold dilutions were spread onto agar plates, incubated overnight, and the CFU on the plates were counted on the following day.

The Log Reduction and the bacterial viability were calculated using Equations (4) and (5), respectively [20,32]:

$$\text{Log Reduction} = \log\_{10} \frac{N\_0}{N} \tag{4}$$

$$\text{Bacterial Velocity} \left( \% \right) = \frac{N}{N\_0} \times 100 \tag{5}$$

where *N* is the number of viable cells in contact with MF-AgNps membranes and *N<sup>0</sup>* is the number of viable cells in contact with MF membrane (control—without AgNps).

In order to investigate the occurrence of biofouling in the MF membrane and MF-50mA-120s, a filtration experiment was carried out in a cross-flow membrane system, as indicated in Figure 1. The tests were conducted with recirculation of both permeate and retentate streams at 1.5 bar, with a flow rate of feed of 40 L h−<sup>1</sup> . *P. aeruginosa* suspension of 10<sup>8</sup> CFU mL−<sup>1</sup> was prepared and placed in the feed tank, and after 4.5 h of permeation, the viable sessile cells were quantified as described before.

#### **3. Results and Discussion**

#### *3.1. Membrane Characterization*

Figure 2 presents the FESEM photomicrographs of the upper surface and cross section of MF membrane (A and D) and the surfaces of MF-15mA-15s (B) and MF-50mA-120s (C). No significant difference in the surface pores of MF membrane and membranes loaded with silver nanoparticles was observed. The cross section of MF membrane showed a sponge-like morphology with interconnected pores.

**Figure 2.** Field emission scanning electron microscopy (FESEM) photomicrographs of the surfaces (1000×) of (**A**) MF-membrane, (**B**) MF-15mA-15s, (**C**) MF-50mA-120s, and of (**D**) cross section of MF membrane (500×).

The EDS spectra of the membranes' surface are portrayed in Figure 3, where the presence of silver element is indicated by black arrows. In addition, Figure 4 shows their EDS mapping and FESEM images. FESEM images exhibit spherical particles on the surfaces of MF-15mA-15s (C) and MF-50mA-120s (D). These spherical particles present average diameters of 88 and 50 nm for MF-15mA-15s and MF-50mA-120s, respectively. The EDS mapping revealed uniform distribution of silver element and, as expected, a larger amount of AgNps in MF-50mA-120s, corroborating what was observed in FESEM images. These results suggest that the increases in sputtering time and sputtering current reduce the diameter of silver particles, which might be related to a higher nucleation rate for nanoparticle growth at higher sputtering current, and confirm that the silver nanoparticles were successfully impregnated on the surface of the MF membranes.

*θ*

**Figure 3.** The energy dispersive X-ray (EDS) spectra of membranes (**A**) MF membrane, (**B**) MF-15mA-15s, and (**C**) MF-50mA-120s.

**Figure 4.** EDS mapping for silver element from (**A**) MF-15mA-15s and (**B**) MF-50mA-120s, and FESEM photomicrographs of the surface of the membranes (100,000×) of (**C**) MF-15mA-15s and (**D**) MF-50mA-120s.

X-ray diffraction patterns are shown in Figure 5 from 5◦ to 90◦ . The MF membrane exhibited a broad peak that corresponds to the amorphous structure of PES. For MF-50mA-120s membrane, a sharp peak is observed at 2 θ = 38◦ , which is attributed to the crystallinity of silver (black arrow) and represents (111) Bragg's reflections of face-centered cubic (fcc) structure [33]. These observations are in agreement with values of silver nanoparticles reported in the literature [2,6,33–35]. For the MF-15mA-15s membrane, there was no observed peak related to crystalline domains, which may be attributed to the low amount of silver deposited in these sputtering conditions.

**Figure 5.** XRD patterns of Ag<sup>0</sup> , MF membrane, and MF-50mA-120s.

The broadening of peaks in the X-ray diffraction pattern can be related to the particle size [33]; thus, the average size of 17.7 nm was estimated using Scherrer's equation for the silver nanoparticles in MF-50mA-120s.

The synthesis of nanoparticles by sputtering deposition techniques and their mechanisms of nucleation and growth were investigated by other studies for several metals such as silver [36,37], copper [38], cobalt [39], niobium [40], and palladium [41]. However, the detailed description of these mechanisms is very complex [39,42].

− − It has been also described that, after long periods (e.g., more than 6 min) of plasma treatment, the surface of PES membranes can be damaged, for instance, the molecular bonds C-C and C-H can be cleaved by argon plasma. On the other hand, the application of short periods of time seams not affect the polymer chains of the membranes [43,44]. Therefore, in this current work, the FTIR spectra (Figure S1 in Supplementary Materials) of the membranes, before and after the impregnation of the nanoparticles, revealed no significant effects of the sputtering technique. The aromatic C-H stretches at 3094 and 3062 cm−<sup>1</sup> (Figure S1B), and aromatic C=C stretches at 1574 and 1481 cm−<sup>1</sup> (Figure S1C) remained the same for both membranes.

The water permeability of MF membrane and membranes impregnated with silver nanoparticles are presented in Table 1. The results indicate that there was no significant difference in water permeability between MF membrane and modified membranes (MF-15mA-15s and MF-50mA-120s). This result is in agreement with reported works for different membranes characteristics and AgNps synthesis techniques [7,23,24,45].

**− − −**

−

−

−


**Table 1.** Water permeability of MF membrane and membranes loaded with AgNps.

These membranes present large pore sizes, and they are fabricated with hydrophilic polymers (PES and PVP). Such polymers can input high water flux to the membrane, especially on phase inversion technique, where part of the additive, i.e., PVP, can be entrapped in the membrane matrix, as reported in the literature [46–48].

The PES membrane contains sulfone and ether groups alternated between aromatic rings [44,49]. The FTIR spectra show characteristic bands of PES: (1) 3094 and 3062 cm−<sup>1</sup> due to aromatic C–H stretch (Figure S1B) [46,49,50]; (2) 1574 and 1481 cm−<sup>1</sup> due to aromatic C=C asymmetric stretch (Figure S1C) [44,50]; and (3) 1320 and 1296 cm−<sup>1</sup> resulting from the anti-symmetric O=S=O stretch of the sulfone group (Figure S1D) [44,49,50]. Moreover, the FTIR spectra shows the presence of PVP with a characteristic band at 1650 cm−<sup>1</sup> due to carbonyl group (Figure S1E) [46,47,50]. μ <sup>−</sup> μ <sup>−</sup>

−

#### *3.2. AgNps Releasing Test*

In order to investigate the stability of silver nanoparticles loaded on MF-AgNps membranes, the concentration of silver leaching was also determined by immersion and filtration experiments (Figure 6). The percentage of the remaining Ag on the modified membranes was calculated based on the total amount of AgNps initially deposited on the surface of the membranes, corresponding to 8.22 and 317.78 mg m−<sup>2</sup> on MF-15mA-15s and MF-50mA-120s, respectively. μ <sup>−</sup>

− **Figure 6.** Percentage of the remaining Ag on MF membranes (**left axis**) over time (**upper axis**) in immersion test (**black square**) Ag loaded with 15 mA and 15 s of sputtering; (**red circle**) Ag loaded with 50 mA and 120 s of sputtering). Conditions: diameter of membrane = 4.7 cm, volume of deionized water = 50 mL, 60 rpm. Percentage of remaining Ag (**blue-dashed column**, **left axis**) and silver concentration leaching (**white column**, **right axis**) per volume of permeate (**lower axis**) in cross-flow experiment. Conditions: pressure = 1.5 bar, flow rate = 40 L h−<sup>1</sup> , membrane area = 45 cm<sup>2</sup> .

During the first hour of immersion, the MF-15mA-15s membrane lost 6.32 ± 2.98% silver content; however, considering the total period of 10 months (7200 h), a loss of 7.02% was observed, indicating

that the biggest part of silver release occurs at the beginning of immersion in the water bath. This fast decrease in the release of silver is qualitatively similar to other studies [7,20,24,51] and may be explained by probable unattached AgNps on the surface of the membrane.

Besides the percentage of silver releasing from the MF-15mA-15s membrane, its concentration in water after 10 months was 20.0 µg L−<sup>1</sup> , which is lower than the silver maximum contaminants limit of 100 µg L−<sup>1</sup> described by the World Health Organization Guideline for Drinking Water [17] and the U.S. Environmental Protection Agency (USEPA).

Not only did the MF-15mA-15s membrane present a feasible characteristic on the entrapment of silver nanoparticles, but also the MF-50mA-120s showed 1.34 ± 0.13% of silver loss after 24 h of immersion. Furthermore, after 1 month of immersion, the silver concentration in water was similar to the one after 24 h of immersion (0.1 mg L−<sup>1</sup> ), which indicates the same trend of silver releasing.

After 10 months of immersion, there was still approximately 93.0 and 87.9% of impregnated silver on MF-15mA-15s (Figure 6, black square) and MF-50mA-120s (Figure 6, red circle), respectively.

Figure 6 also shows the results of silver leaching from MF-50mA-120s membrane during the cross-flow experiment. After 2.75 L of water permeation, there was still 98.5% of silver impregnated on the membrane surface (Figure 6, blue-dashed column), and the concentration of silver leached for each 0.25 L of permeate was lower than 100 µg L−<sup>1</sup> (Figure 6, white column).

In addition, a test with recirculation of both permeate and retentate streams (full-recycle setup) was also performed during 24 h. The concentration of silver leached (in the feed tank) was 8.4 µg L−<sup>1</sup> at the end of the experiment, which indicates that 98.8% of the initial silver impregnated on MF-50mA-120s remained on its surface after permeation. The water permeation test corroborates the silver loss after 24 h of immersion in the water bath and indicates that, even at a flowrate of 40 L h−<sup>1</sup> , the MF-50mA-120s membrane showed a small loss of silver. The results of silver leaching in both experiments indicate that the sputtering technique is effective for impregnating and entrapping silver nanoparticles on the surface of membranes. This finding is important for the maintenance of the membrane's biocidal performance and for the minimization of silver leaching to the environment.

Evaluation of silver loss in previous studies of membranes impregnated with silver nanoparticles by chemical reduction method showed percentages of remaining silver of 99.38% and 98.75% after 24 h and six days of immersion, respectively [7]. Furthermore, a percentage of 97.0% was obtained after a cross-flow experiment [24]. However, in these studies, the immersion test was conducted for short periods and under reduced flowrate (cross-flow experiment) in comparison with this work. In addition, the chemical reduction method presents some disadvantages as several steps production and chemical reagents, the use of stabilizer agents and the residual solvents.

Another important fact is a continuous silver loss observed in other studies with physical methods and green synthesis [20,23]. In these cases, the authors indicated that the lifespans of their membranes were 97 and 340 days according to their silver leaching rate.

In fact, this current work highlights the evaluation of silver loss for long-term immersion and cross-flow experiments for membranes loaded with silver nanoparticles by a one-step production method with no residual reagents.

#### *3.3. Antibacterial Activity Tests*

#### 3.3.1. The Disc Diffusion Method

As illustrated by the disk tests (see Figure 7), MF-membrane (A) had no significant effect on the growth of *P. aeruginosa*, while MF-AgNps membranes showed a clear area with no evidence of bacterial growth. An inhibition zone around the membranes of 0.5 and 0.8 mm was observed for MF-15mA-15s (B) and MF-50mA-120s (C), respectively. These results are similar to many reported works, which also found the inhibition zone around the substrate containing AgNps, and demonstrate that this antibacterial activity comes mainly from the silver and not from PES [23,51,52].

**Figure 7.** Diffusion disc method against *Pseudomonas aeruginosa* of (**A**) MF membrane (control sample), (**B**) MF-15mA-15s, and (**C**) MF-50mA-120s. Inhibition zone are indicated by red arrows.

− In general, the inhibition zones reported by different authors are larger than the ones observed in this work, which can be explained to by the hypothesis that the release of silver nanoparticles is amplifying the biocidal zone. Once the nanoparticles are not well attached to the membranes, they can diffuse in the media and inhibit bacterial growth. However, in this current work, the silver release is reduced to a small region, corroborating this hypothesis. Thus, the sputtering technique was efficient at entrapping silver nanoparticles on the membrane, and their biocidal effect will be concentrated in this region, which may be noteworthy to inhibit biofouling formation.

#### 3.3.2. The Biofouling Resistance Test

−

The quantification of total planktonic cells and total sessile cells, shown in Figure 8, did not indicate any significant difference between the membranes, with the CFU mL−<sup>1</sup> being in the same order of magnitude for all membranes.

The quantification of viable planktonic cells grown in suspensions without membrane (control sample) and with MF membrane (without AgNps) did not show a significant difference (Figure 9). On the other hand, for MF-15mA-15s, there was a considerable reduction of ~0.6 log units in CFU mL−<sup>1</sup> compared to MF-membrane. In this case, after the exposure to MF-AgNps membranes for 24 h, the viability of the planktonic cells was only 24.2%. This decrease demonstrated that the MF-15mA-15s membrane causes an expressive *P. aeruginosa* growth inhibition.

Figure 9 also depicts the viable sessile cells on the different membranes. A comparison between MF membrane and MF-15mA-15s indicates that the bacterial growth showed a 0.7 log unit reduction in CFU mL−<sup>1</sup> , while the viability of the sessile cells was 19.8% after 24 h of exposure to AgNps. This result confirmed that the silver impregnated in MF-15mA-15s improved the bactericidal properties of the membranes, which is in agreement with other studies that investigated silver nanoparticles synthesized by chemical route [25,51].

−

− **Figure 8.** Total cells of *P. aeruginosa* after their exposure to MF membrane and MF-AgNps: (**A**) total planktonic cells and (**B**) total sessile cells. Number of bacteria initially inoculated on each sample: log CFU mL−<sup>1</sup> = 7.0 at 30 ◦C, 200 rpm for 24 h of incubation.

The adhesion of *P. aeruginosa* in the MF membrane and MF-50mA-120s was investigated during a cross-flow experiment in order to evaluate their performance in a filtration process. The results revealed 0.66 ± 0.02 log unit reduction of viable sessile cells, and as a consequence, 22% of bacterial viability for MF-50mA-120s (log MF-membrane = 6.63, log MF-50mA-120s = 5.97). These results show an outstanding maintenance of the effectiveness of the MF-50mA-120s in comparison to the MF membrane, even with a continuous flowrate through the membrane.

Liu et al. (2013) [53] observed the deposition of *E. coli* in polysulfone membranes impregnated with silver nanoparticles synthesized by chemical route. The authors concluded that the AgNps did not affect the kinetics of bacterial deposition. However, the bacterial detachment ratio during rising is large in the presence of silver nanoparticles because the bacteria become inactivated after the contact with these nanoparticles, enhancing the detachment rate.

Thereby, in this study, the adhesion of *P. aeruginosa* in MF-AgNps was verified and quantified as the total sessile cells. Nevertheless, the viable sessile cells decreased in comparison with the MF membrane, indicating that AgNps inactivated the microorganisms, which would facilitate the detachment.

Even though the exact mechanism of antibacterial activity of AgNps is not fully understood and further studies are needed to explain this gap, several researchers agree on a synergistic action between contact killing of nanoparticles and the releasing of silver ions from the membranes [15,54,55].

In this context, the antibacterial effect of MF-AgNps membranes produced in this work is affected by its lower concentration of silver leached; however, for applications in water treatment, for example, it is important to minimize this leaching due to the concern with human health and environment protection. Furthermore, the development of membranes with biocidal advantages associated with a long lifespan is needed.

Therefore, the MF-AgNps membrane demonstrated potential for these applications because of its antibacterial properties and its expected longer lifespan due to its low silver leaching.

−

−

− **Figure 9.** *P. aeruginosa* cells viabilities after their exposure to MF-membrane and MF-AgNps: (**A**) viable planktonic cells and (**B**) viable sessile cells. Number of bacteria initially inoculated on each sample: log CFU mL−<sup>1</sup> = 7.0 at 30 ◦C, 200 rpm for 24 h of incubation.

#### **4. Conclusions**

Impregnation of silver nanoparticles on the microfiltration membranes by sputtering technique was confirmed by different and complementary analyses. From FESEM images, average diameters of 88 and 50 nm were estimated for MF-15mA-15s and MF-50mA-120s, respectively, suggesting that the increase in sputtering time and sputtering current reduces the diameter of silver particles. The diffractogram of the MF-50mA-120s membrane exhibited a sharp peak at 2 θ = 38◦ , which is attributed to the crystallinity of silver, proving the presence of silver on the surface of this membrane and corroborating the FESEM and EDS observations.

One of the main advantages of the sputtering technique is the good adhesion between AgNps and membranes, which allows the release of silver ions lower than the maximum limit of silver in drinking water by World Health Organization and the U.S. Environmental Protection Agency. After 10 months of immersion, there was still approximately 93.0 and 87.9% of silver initially impregnated onto MF-15mA-15s and MF-50mA-120s membranes, respectively. Furthermore, after 24 h of filtration test in full-recycle set-up with high flowrate, MF-50mA-120s showed that 98.8% of silver remained on the membrane. These results indicate the efficiency of sputtering technique to entrapped silver nanoparticles on the membranes.

The disc diffusion method demonstrated the antibacterial activity of silver nanoparticles impregnated onto membranes by an inhibition zone of approximately 0.5 and 0.8 mm for MF-15mA-15s and MF-50mA-120s, respectively. The microbial inhibition occurs with the diffusion of silver nanoparticles from the membrane into the agar layer. Thereby, these small inhibition zones can be explained by the lower release of silver nanoparticles impregnated on the microfiltration membranes by sputtering method.

The biofouling resistance test for MF-15mA-15s showed 0.6 and 0.7 log unit reductions in CFU mL−<sup>1</sup> when compared with the MF membrane (without AgNps) for viable planktonic cells and viable sessile cells, respectively. After exposure to this membrane for 24 h, the planktonic cells and sessile cells viabilities were both under 25%, confirming the biocidal property of the membrane with silver nanoparticles.

Although there was adhesion of *P. aeruginosa* in microfiltration membranes impregnated with AgNps, the viable sessile cells decreased, indicating a great potential of silver nanoparticles to reduce biofouling and its consequences.

The highlight of this work is the occurrence of bacterial inhibition in membranes impregnated with AgNps with reduced silver leaching, which allows a longer lifespan for these membranes. Less silver loss was observed in this work, even though it was analyzed in long-term immersions and in filtration experiments with higher flowrate when compared with other studies in the literature.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/8/1686/s1, Figure S1: FTIR spectra of MF-membrane, MF-15mA-15s and MF-50mA-120s.

**Author Contributions:** Conceptualization, C.P.B. and F.V.F.; Methodology, A.M.F.L.; Validation, A.M.F.L.; Formal analysis, A.M.F.L.; Investigation, A.M.F.L.; Resources, C.P.B. and F.V.F.; Data curation, A.M.F.L.; Writing—original draft preparation, A.M.F.L.; writing—review and editing, C.P.B. and F.V.F.; Visualization, C.P.B. and F.V.F.; Supervision, C.P.B. and F.V.F.; Project administration, F.V.F.; Funding acquisition, C.P.B. and F.V.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico—Brasil (CNPq).

**Conflicts of Interest:** The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **A Biodegradable Magnetic Nanocomposite as a Superabsorbent for the Simultaneous Removal of Selected Fluoroquinolones from Environmental Water Matrices: Isotherm, Kinetics, Thermodynamic Studies and Cost Analysis**

**Geaneth Pertunia Mashile 1,2, Kgokgobi Mogolodi Dimpe 1,2 and Philiswa Nosizo Nomngongo 1,2,3,\***


Received: 1 March 2020; Accepted: 31 March 2020; Published: 12 May 2020

**Abstract:** The application of a magnetic mesoporous carbon/β-cyclodextrin–chitosan (MMPC/Cyc-Chit) nanocomposite for the adsorptive removal of danofloxacin (DANO), enrofloxacin (ENRO) and levofloxacin (LEVO) from aqueous and environmental samples is reported in this study. The morphology and surface characteristics of the magnetic nanocomposite were investigated by X-ray diffraction (XRD), Brunauer–Emmett–Teller (BET) adsorption–desorption and Fourier transform infrared spectroscopy (FTIR). The N<sup>2</sup> adsorption–desorption results revealed that the prepared nanocomposite was mesoporous and the BET surface area was 1435 m<sup>2</sup> g −1 . The equilibrium data for adsorption isotherms were analyzed using two and three isotherm parameters. Based on the correlation coefficients (R<sup>2</sup> ), the Langmuir and Sips isotherm described the data better than others. The maximum monolayer adsorption capacities of MMPC/Cyc-Chit nanocomposite for DANO, ENRO and LEVO were 130, 195 and 165 mg g−<sup>1</sup> , respectively. Adsorption thermodynamic studies performed proved that the adsorption process was endothermic and was dominated by chemisorption.

**Keywords:** fluoroquinolones; ultrasound radiation; mesoporous carbon; desirability function; thermodynamics; wastewater; cost analysis

#### **1. Introduction**

The presence of pharmaceuticals in aquatic environments has become a subject of interest for environmental chemists [1]. Their wide distribution owes itself to the growing need for treatments and cures for human and animals diseases [2]. They are introduced into the aquatic environments through effluents of urban wastewater treatment plants (WWTPs) [3]. This is a result of their extensive use and their ineffective removal processes by wastewater transport and treatment [4]. Among various pharmaceuticals, antibiotics residues have proved to be the most commonly detected in the aquatic environment for both surface and ground waters [5]. Although they may occur in fairly low concentrations in environmental waters, their different modes of action and particular chemical and physical characteristics may pose a risk to the aquatic system [6]. Thus, there is a need to monitor and evaluate their persistent presence, which even at a low level can further increase antibiotic

resistance [7]. The focus of this work is mainly on fluoroquinolones which are an important emerging group of synthetic antibacterials [8]. They have been used extensively for both human and veterinary medicine due to their effectiveness against both gram-positive and negative bacteria for the treatment of bacterial infections [2]. Moreover, different antibiotics have different half-lives; therefore, others may be more persistent in the environment which may result in increased levels of contamination to the environment [9].

Studies have shown that they are introduced to environmental bodies by either direct or indirect pathways [4,10,11]. Furthermore, they have been found to occur in surface waters at concentrations ranging from ng L−<sup>1</sup> to µg L−<sup>1</sup> [10,12]; ng L−<sup>1</sup> to mg L−<sup>1</sup> in groundwater [13]; and mg L−<sup>1</sup> in soil [14]. Since they are continuously introduced into the environment they have been identified as pseudo-persistent organic pollutants [11]. The greatest challenge is the removal of antibiotics from wastewater before discharge into the environment due to the high costs associated with it [9]. Techniques such as advanced oxidation processes (AOPs), multi-treatment processes, separation processes and biological processes have been applied in the removal of antibiotics from wastewater [15]. However, they prove to be very expensive and require high maintenance for the complete removal of compounds, including antibiotics, at a larger scale [16].

Adsorption processes are of significant interest in removal applications of organic compounds such as antibiotics due to their simplicity in design [17], flexibility, cost and friendliness towards potential the toxicity of biological base processes [18]. The adsorption is a technique based on the removal of contaminants from a matrix onto an adsorbent surface [19]. The effectiveness of the technique is highly dependent on the adsorbate properties, adsorbent type and composition of matrix analyzed [20]. To date, various adsorptive material has been used, such as zeolites [21], graphene oxide (GO) [22,23], activated carbon (AC) [24–28], metal-organic frameworks (MOFs) [29], carbon nanotubes (CNT) [30] and clay [31], amongst others for adsorption removal of pharmaceuticals [32]. However, for antibiotic removal, CNTs, ACs, mesoporous clay material, exchange resins and bentonite are the most widely reported adsorbents [9]. Despite their widespread use, these sorbents also present some limitations, such as inefficient extraction, low antibiotic adsorption properties and costliness (high generation costs) [9]. Mesoporous carbon from carbon-based material, on the other hand, can serve as an artery for adsorbates and also contribute greatly towards adsorption [33,34]. It can boast advantageous features, such as a large surface area; a high adsorption capacity; a large and ordered pore size and structure; and chemical and mechanical stability [28,33–39]. Furthermore, mesoporous carbon can be made from cheaper materials, such as starch and waste biomass [28,33,37–39]. In addition, the incorporation of magnetic nanoparticles to mesoporous carbon facilitates ease during separation, and functionalizing the material enables for reduction of its hydrophobic nature [38,39].

Furthermore, the natural polymers such as chitosan and beta-cyclodextrin have gained prominence in recent years due to their advantageous features [40–46]. They possess similar features, such as biocompatibility [44,47] and biodegradability [46]. Their non-toxicity has proven that they are less harmful to humans and the environment, and thus they are often selected as solid phase materials for adsorptions of various pollutants, including pharmaceutical ones [43,48–50]. Moreover, they are formed from environmentally friendly sources; chitosan is formed from naturally existing resources, such as the exoskeletons of anthropoids, like shellfish, crabs and prawns [51], whereas beta-cyclodextrins can be derived from enzymatic degradation of starch [46]. Great attention has been focused on the immobilization of cyclodextrins on chitosan; their combination improves the adsorption capacity of chitosan [42,44].

Recently, coupling of adsorption processes and ultrasound irradiation have gained considerable attention due to their numerous advantages [26,52–55]. These include faster chemical reactions and mass transfer as a result of acoustic cavitation with the establishment of new adsorption sites on the adsorbent surface [26,52–55]. The influences of ultrasonic irradiation on the adsorptive removals of numerous pollutants from aqueous solutions have been reported in the literature [26,52–58].

Therefore, the objective of the present study was to synthesise magnetic mesoporous carbon/β-cyclodextrin–chitosan (MMPC/Cyc-Chit) nanocomposite as a sorbent for the elimination of fluoroquinolones. Factors that play a role in the adsorptive removal of the fluoroquinolones by MMPC/Cyc-Chit nanocomposite were examined; namely, sonication power level, sample pH and initial concentration of DANO, LEVO and ENRO. The overall process was to utilize cheap and readily available material for nanocomposite synthesis and ultrasonic radiation for superior removal efficiency. The incorporation of biodegradable polymers such as chitosan and β-cyclodextrin to magnetic mesoporous carbon resulted in a nanocomposite with super-adsorbent activities considering high surface area and adsorption capacities. The application of MMPC/Cyc-Chit nanocomposite for removal of fluoroquinolones has been reported for the first time.

#### **2. Materials and Methods**

#### *2.1. Materials and Reagents*

Chemicals reagents used for this study were of analytical grade, and Ultra-pure water (Direct-Q® 3UV-R purifier system Millipore, Merck, Darmstadt, Germany) was used throughout the duration of the experiments. Danofloxacin (99.7%) (DANO), enrofloxacin (99.0%) (ENRO), levofloxacin (99%) (LEVO), HPLC grade ethanol, methanol and acetonitrile were used, along with acetic acid, sodium hydroxide, ammonium hydroxide, ferrous chloride, ferric chloride, starch, chitosan, β-cyclodextrin and ortho-phosphoric acid purchased from Sigma-Aldrich (St. Loius, MO, USA). A synthetic sample mixture of the fluoroquinolones (FQs) stock solution was prepared by dissolving appropriate amounts of DANO, ENRO and LEVO in small amounts of methanol. The mixture was then diluted with ultra-pure water to a final volume of 100 mL. The solution were stored in to refrigerator at 4–8 ◦C.

#### *2.2. Instrumentation*

The synthesized adsorbent material was analyzed utilizing different techniques of characterization in order to determine its structural suitability for adsorption of the fluoroquinolones (DANO, ENRO LEVO). X-ray diffraction (XRD) patterns were recorded using a PANalytical X'Pert X-ray diffractometer (PANalytical BV, Almelo, Netherlands) utilizing Cu Kα radiation (λ = 0.15406 nm) in the 2θ range 4–90 at room temperature. The Fourier transform infrared (FT-IR) Perking–Elmer spectrum 100 spectrometer (Perkin-Elmer, Shelton, CT, USA) using the potassium bromide (KBr) pellet technique in a region of 4000–400 cm−<sup>1</sup> was used to report the infrared spectrum for the prepared material. Surface characteristics such as porosity and area of the as-prepared material were analyzed by using the Brunauer–Emmett–Teller (BET) 77 K using an ASAP2020 porosity and surface area analyzer (Micrometrics Instrument Corp., Norcross, GA, USA).

The samples were degassed was at 100 ◦C for 3 h using N<sup>2</sup> gas before analysis. Adjustments for pH where necessary were performed using an OHAUS starter 2100 pH meter (Pine Brook, NJ, USA). The surface charge/point of zero charge was evaluated for the as-prepared material using a Nano-ZS Zetasizer (Malvern Instruments, Malvern, UK). The pH was adjusted within the range of 2.0–11.0 by the addition of 0.1 mol L−<sup>1</sup> acetic acid and ammonium solution to each solution with 37 mg of adsorbent material. A Scientech Ultrasonic cleaner (Labotec, Midrand, South Africa) with a volume of 5.7 L (internal dimensions: 300 × 153 × 150 mm) was used to facilitate the adsorption process. The ultrasonic system was equipped with a variable frequency and power setting. In this study, the frequency was fixed at 50 Hz and the emission power of 150 W. The system has 5 power levels (1 (weakest) to 5 (strongest)), this power setting is used to reduce or increase the size of the cavitation bubble implosion force. Therefore, the sonication power levels were varied. The analysis of the antibiotics was performed using an Agilent HPLC 1200 Infinity series, equipped with a photodiode array detector (Agilent Technologies, Waldbronn, Germany). Chromatograms were recorded at 290 nm. An Agilent Zorbax Eclipse Plus C18 column (3.5 µm × 150 mm × 4.6 mm) (Agilent, Newport, CA, USA) was operated at an oven temperature of 25 ◦C. The mobile phase (water with 10 mmol L−<sup>1</sup> of

phosphoric acid; the pH adjusted to 3.29 with triethylamine): acetonitrile (85.7:14.3, *v*/*v*) at a flow rate of 1.5 mL min−<sup>1</sup> . All chromatographic experiments were carried out 25 ± 3 ◦C while the injection volume was 10 L for all samples.

#### *2.3. Preparation of the Nanocomposite*

#### 2.3.1. Synthesis of Mesoporous Carbon (MPC)

Modified version of the hard templating method adapted from literature was used in the synthesis of mesoporous carbon [59]. Briefly, in a 100 mL beaker containing deionized water and equipped with magnetic stirrer for easy dissolving starch was used. The mixture was then heated over an oil bath at 120 ◦C to form a homogenous solution with continuous stirring at 200 rpm. Silica solution was added dropwise at approximately 1 drop per second using a burette with continuous stirring until the starch had completely dissolved. Thereafter, the solution was transferred onto a glass petri dish and left to cool at an ambient room temperature. A gel-like material was formed and dried at 60 ◦C in an oven for 1 h and further carbonized with the gentle flow of nitrogen gas at 500 ◦C for 3 h. Once carbonized the material was stirred for 24 h at 70 ◦C in a sodium hydroxide (30 wt %) solution to remove silica. The formed product was washed with a mixture of ethanol and water (1:1) and filtered under vacuum. The filtered product was then oven dried at 60 ◦C for 2 h.

#### 2.3.2. Preparation of Magnetic Mesoporous Carbon Coated with Chitosan and β-CD

Ferrous and ferric chloride solutions were dissolved in ultrapure water at a Fe2+/Fe3<sup>+</sup> ratio of 1:2 and stirred for 5 min. Then 3 g of β-CD and 4 g MPC were added into the iron solution with vigorous stirring along with the addition of diluted sodium hydroxide solution (1.0 mol L−<sup>1</sup> ) while heating at 80 ◦C for 1 h. That solution was then filtered by vacuum filter and washed with methanol plus water. The filtrate was then dried in an oven at 60 ◦C for 24 h. Chitosan flakes were modified based on a method described by [42]. Briefly, 3 g of chitosan flakes was dissolved in 50 mL of 3% acetic acid. Prepared magnetic material was then added to the solution of chitosan and this mixture was transferred to a round bottom flask. These were ultra-sonicated to facilitate dispersion were the pH of the prepared mixture was adjusted to 8.0–9.0 by means of diluted sodium hydroxide solution. Thereafter, it was filtered and washed with mixture of ethanol (50:50) plus water until the pH reached about 7, and oven dried at 40 ◦C.

#### *2.4. Batch Adsorption Studies*

Batch adsorption method was employed for adsorption studies. This was achieved by adding a specific mass of adsorbent (10–30 mg) to 25 mL synthetic sample solutions containing a mixture of FQ antibiotics (that is DANO, ENRO and LEVO) at a concentrations of 10 mg L−<sup>1</sup> . The pH of the synthetic sample solutions (5–9) were adjusted using 0.1 mol L−<sup>1</sup> HCl and 0.1 mol L−<sup>1</sup> NaOH. The adsorption process was carried out using an ultrasonic bath. The frequency of the ultrasonic bath was fixed at high 50 Hz while the sonication power level was varied between 2 (60 W or 40% of total power and 5 (150 W or 100% of total power). Once the adsorption processes was completed, the adsorbent and sample solution were separated using an external magnet. The supernatant was filtered by using 0.22 µm syringe filters and the residual FQ antibiotic concentration in the solution was determined HPLC-PDA.

A response surface methodology constructed by a central composite design (CCD) was used for the optimization of the most influential parameters for the removal of FQ antibiotics. These factors include sample pH, mass of adsorbent (MA) and sonication power level (SP). The removal efficiency (%RE) was used as an analytical response. The optimization process was carried out using Statistica version 13. When the optimal conditions were achieved, the adsorption isotherm and kinetics for the removal of FQ antibiotics were examined.

Under optimum conditions, Langmuir, Freundlich, Hill and Langmuir-Freundlich isotherm models (Table 1) were used to study the interaction between the prepared MMPC/Cyc-Chit nanocomposite and FQ antibiotic mixture. To achieve this, model solutions containing different concentrations selected FQs antibiotics mixture (5–80 mg L−<sup>1</sup> ) were used.


**Table 1.** Adsorption isotherms and kinetics models equations.

The kinetic studies performed using an initial concentration of 50 mg L−<sup>1</sup> were used to explain the rate and mechanism of the adsorption process. The kinetics models, such as pseudo-first-order, pseudo-second-order and intraparticle diffusion, were employed to analyze the equilibrium kinetic data. The thermodynamic studies were carried out using a concentration of 50 mg L−<sup>1</sup> at different temperatures: 25, 35 and 40 ◦C.

#### *2.5. Regeneration and Reusability (Recyclability) of the Nanocomposite*

To investigate the regeneration capability of the MMPC/Cyc-Chit nanocomposite, 36 mg of adsorbent was placed into 25 mL of 10 mg L−<sup>1</sup> FQ antibiotic solution. The mixture was sonicated for 30 min, and after the adsorption process had been completed, the separation of adsorbent by an external magnet was done. The adsorbent was then sonicated with a mixture of 10 mL of acidified water and acetonitrile mixture (55:45 ratio) for 10 min to remove the adsorbed FQs. The water was obtained by adjusted ultrapure deionized to pH 3 using *ortho*-phosphoric acid. It should be noted that 10 min desorption time was enough to remove all the analytes adsorbed. An external magnet was applied to facilitate the decantation of the desorption solvent. Desorption solution containing the FQs was analyzed using HPLC-PDA. After decantation, the adsorbent was washed with the desorption solvent; filtered; and finally, washed two times with ultrapure water and dried at 60 ◦C for 2 h. The above procedure was repeated 10 times.

#### *2.6. Application in Real Water Samples*

Wastewater (influent and effluent) samples were collected from a wastewater treatment plant (WWTP) in Pretoria, South Africa. River water and tap water were collected from the Apies River (Pretoria, South Africa) and the University of Johannesburg laboratory (Johannesburg, South Africa). The sample collection was performed during October 2019. The wastewater and river water samples were kept in 1 L glass amber bottles and transported to the laboratory to be stored at 4 ◦C before adsorption studies. The physicochemical characteristics, such as pH; conductivity; total dissolved solids (TDS); and dissolved organic carbon of wastewater, laboratory tap water and river water, are presented in summarized in Table A1. In addition, the concentrations of major elements such as calcium, magnesium, sodium and iron are presented in Table A1.

#### **3. Results and Discussion**

#### *3.1. Characterization*

#### 3.1.1. X-ray Diffraction Spectroscopy

Figure 1 shows the XRD patterns of the (a) mesoporous carbon, (b) β-cyclodextrin, (c) chitosan and (d) MMPC/Cyc-Chit nanocomposite. The XRD patterns for chitosan, β-cyclodextrin and mesoporous carbon are comparable with those reported in the literature [50,59,60]. The XRD pattern for MMPC/Cyc-Chit nanocomposite shows diffraction peaks at 2θ = 31.3◦ , 35.7◦ , 42.8◦ , 54.1◦ , 56.8◦ and 63.2◦ . These diffraction peaks correspond to the magnetite planes indexed to (220), (311), (400), (422), (511) and (440). These results confirmed the importation of iron oxide nanoparticles (Fe3O4) in the nanocomposites. Moreover, they were in agreement with other results in literature [42,45]. β β θ

**Figure 1.** XRD of (**A**) mesoporous carbon, (**B**) beta-cyclodextrin, (**C**) chitosan and (**D**) MMPC/Cyc-Chit nanocomposite composite.

#### 3.1.2. Fourier Transform Infrared Spectroscopy

β β FTIR spectra for mesoporous carbon (MPC), β-cyclodextrin (β-CD), chitosan (Chi) and MMPC/ Cyc-Chit nanocomposite are presented in Figure 2. The FTIR spectrum of MPC (Figure 2) reveals the peaks at 2924–2889 cm−<sup>1</sup> and 1384 cm−<sup>1</sup> which were ascribed to the stretching and bending of CH<sup>3</sup> and CH<sup>2</sup> stretching [37], whereas the broad peak at 3439 cm−<sup>1</sup> was attributed to the O–H stretching. The band at 1615 cm−<sup>1</sup> was assigned to the C=O vibration of carbonyl groups [39,61]. In addition, the CH<sup>3</sup> stretching and unsaturated sites were observed at 2361 cm−<sup>1</sup> [37]. The major bands for β-cyclodextrin and chitosan (Figure 2) were allocated as follows: 1024 cm−<sup>1</sup> for (R-1, 4-bond skeleton vibration of β-CD); 1649–1656 cm−<sup>1</sup> for C–N and C=O (NHCO (amide I)) stretching vibrations; and 3280–3353 cm−<sup>1</sup> (O–H and N–H stretching vibrations) [42,45]. In addition, the peaks at 1586 and 1153 cm−<sup>1</sup> were assigned to the N–H stretching vibration (primary amine) and antisymmetric

glycosidic linkages [42]. The MMPC/Cyc-Chit nanocomposite shows two characteristic absorbance bands centered at 1652 and 1597 cm−<sup>1</sup> , which correspond the C=O stretching vibration of NHCO (amide I) and N–H bending of NH2, respectively [42]. − −

−

−

− −

−

−

β <sup>−</sup>

**Figure 2.** FT-IR spectra of MPC, Chi, beta-CD and MMPC/Cyc-Chit composite.

#### 3.1.3. Nitrogen Adsorption–Desorption

β − The important textural properties that influence the quality and application of an adsorbent, especially for adsorptive removal of pollutants in matrices that are complex (such as wastewater and polluted river waters), are the porosity and specific surface area [24,28,36,62,63]. It has been reported that the two properties are significant because they are strongly related to the maximum adsorption capacity of the adsorbent [24,27,28,36,62–64]. Textural properties of the nanocomposite are presented in Table 2. The results confirmed that incorporating chitosan and β-cyclodextrin into magnetic mesoporous carbon resulted in a superabsorbent with high specific surface area (1264 m<sup>2</sup> g −1 ). The micropore and mesopore surface areas of the nanocomposite in comparison with mesoporous carbon were used to analyse the textual properties of the prepared material. As seen in Table 2, the percentage of the surface comprised of mesopores was 60%, suggesting that the nanocomposite is predominantly a mesoporous material [28,34–37,39]. These characteristics validate the applicability of the nanocomposite for adsorption processes. According to the results in Table 2, it was anticipated that during the adsorption process, the investigated FQ antibiotics would percolate through pores of the adsorbents. These findings were in agreement with SEM results, and they both confirm that the prepared adsorbent possesses outstanding characteristics which endorse it for wastewater treatment using adsorption technology.

**Table 2.** Characteristics of adsorbent material; BET surface area; pore volume parameters of MPC and MMPC/Cyc-Chit.


#### 3.1.4. Point of Zero Charge

−

−

− −

The pH of the FQs solution might have an effect on their adsorption on the surface of the MMPC/Cyc-Chit nanocomposite. Moreover, the pH of the sample solution was used to assess the distribution percentage of the investigated FQ species during their adsorption process. For example, subject to the pH of the sample solution, the surface of the nanocomposite could be protonated or deprotonated, thereby changing the surface charge of an adsorbent. Therefore, it is important to investigate the pH at which negative and positive charges are equal, also known as pH at point of zero charge (pHpzc). This point will as assist in the determination of the possible adsorption mechanism. Therefore, the influence of pH onto the zeta potential of MMPC/Cyc-Chit nanocomposite was evaluated and results are shown in Figure 3. The surface MMPC/Cyc-Chit nanocomposite was positively charged at pH values lower than 8 and the pHpzc value was estimated as 8.0. This implied that MMPC/Cyc-Chit nanocomposite has a negative charge above pH = 8.0.

**Figure 3.** Determination of pHpzc of nanocomposite.

#### *3.2. Optimization*

The batch adsorption process was optimized using the RSM-ased CCD, and the design matrix together with respective responses obtained at the equivalent experimental conditions are indicated in Table 3. Statistica software was used to generate second-order polynomial model which used to explain the adsorption process of FQ antibiotics onto the MMPC/Cyc-Chit nanocomposite. The removal efficiency was used as the dependent variable or analytical response. The *R* <sup>2</sup> values were used to assess the performance of the RSM model, and they were found to be 0.9985, 0.99876 and 0.9975 for DANO, ENRO and LEVO, respectively. These findings revealed the best agreement between the actual and predicted responses. Moreover, these results proposed that about 99% of the total variation in removal efficiency was attributable to the experimental factors.

The validity and appropriateness of the RSM model, as well as the estimation of the most significant independent variables and their interactions, were examined by analysis of variance (ANOVA). The ANOVA results are reproduced in the form of Pareto charts (Figure 4). The importance of an independent variable was evaluated by the magnitude of the bar length. If the length of the bar passes the red line (0.05 confidence level line), this phenomenon suggests that the corresponding independent factor is significant at a 95% confidence level. As seen in Figure 4, the mass of the adsorbent and sample pH were significant at the 95% confidence level for every sample investigated. This implied that they had more influence on the analytical response.


**Table 3.** The design matrix and the results of the two-level fractional factorial design.

**Figure 4.** Pareto chart of standardized effects for adsorption of (**A**) DANO, (**B**) ENRO and (**C**) LEVO). MA = mass of adsorbent; SP = sonication power; 1Lby2L shows the interaction between pH and MA; 2Lby3L shows the interaction between MA and SP; 1Lby3L shows the interaction between pH and SP.

#### 3.2.1. Response Surface Methodology

Three-dimensional (3D) response surface plots were constructed to investigate the effect of each variable on the removal efficiency, and their interactions (Figures 5, A1 and A2). The effects of sample pH, mass of adsorbent (MA) and sonication power level (SP) were concurrently examined for the adsorptive removal of FQs from synthetic samples. Figure 5A shows the 3D plot of sample pH versus mass of adsorbent. As seen in Figure 5, both mass of adsorbent and the sample pH played a critical role in removal of FQ antibiotics from aqueous solutions. This might be because sample pH affects the ionization of analytes and the charge on surface of adsorbent. Based on Figure 5A,B, the removal efficiency increased with increasing sample pH, and the maximum removal was achieved between pH 6 and 8. Below and above these values, a decrease in analytical response was observed. This is because DANO, ENDRO and LEVO can exist in three forms in aqueous systems, that is, cationic (pH > pKa2), zwitterionic (pKa1 ≤ pH ≤ pKa2) and anionic (pH < pKa1), and these forms are pH-dependent [65–68]. Consequently, the adsorption mechanism is also dependent on the adsorbent surface charge. For instance, the FQ antibiotics can be adsorbed by a negatively or positively charged adsorbent using cation exchange through protonation of amine group or electrostatic interaction due to the deprotonation of carboxylic groups [39,61,65–67,69–72].

**Figure 5.** The 3D surface response plots describing the interactions of the parameters investigated. (**A**) interaction between sample pH and mass of adsorbent (MA); (**B**) interactions between sonication power (SP) and sample pH and (**C**) interactions between SP and MA.

At lower pH values, the FQs are predominately in cationic forms due to a high concentration of hydronium ions [61,68,73,74]. This results in lower removal efficiencies due to the competition between the adsorbate and small molecules of hydronium ions which can fill the available active sites. Additionally, the pHpzc of the material was found to be 8, indicating that the charges on surface of the nanocomposite are positive charges. Therefore, lower removal efficiencies can also be attributed to electrostatic repulsion between positively charged nanocomposite and cationic forms of FQs. As the sample pH increases, the electrostatic interaction between the adsorbate/analytes and the surface of the adsorbent occurs, resulting in higher removal efficiencies. However, at pH values > pHpzc value of 8, a decline in the removal efficiency was observed. These could be attributed to electrostatic repulsion between negatively charged FQs and negatively charged nanocomposite. Several researchers in the

literature have observed similar findings with respect to the adsorption behavior of FQs at low and high pH values [61,65–71,73,74]. The results for the effect of sonication power level are shown in Figure 5B,C; it was not significant at the 95% confidence level. However, the 3D response surface plots reveal that as the sonication power levels increases, the removal efficiency also increases. As seen in Figure 5B,C, %RE values above 80% were obtained when the sonication power was 3 (90 W or 60% of the total power) and above. The increased removal efficiency can be attributed to the increase in adsorbate–adsorbent interactions due to turbulence produced by implosion of the cavitation bubbles.

#### 3.2.2. Desirability Function

The desirability profile was used to estimate the optimum experimental conditions obtained using RSM optimization approach (Figures 6, A3 and A4). The optimal conditions for the removal of fluoroquinolones were sample pH: 7.0, mass of adsorbent: 36 mg and sonication power level 3. The sonication or contact time, initial concentration and sonication frequency were fixed at 30 min, 10 mg L−<sup>1</sup> and 50 Hz. Under the abovementioned conditions, the predicted removal efficiencies of the model for the adsorption of DANO, ENRO and LEVO were 97.2%, 98.3 and 95.3%, respectively. To certify the acceptability of the RSM model and to confirm the agreement between the predicted and experimental removal efficiency, six replicates were carried out at the abovementioned conditions. The obtained experimental results showed removal efficiencies of 98.7 ± 1.3%, 99.1 ± 0.9% and 96.8 ± 1.2% for DANO, ENRO and LEVO, respectively. These results showed that the RSM model could be considered an accurate and valid procedure for the optimization of the adsorption process.

**Figure 6.** Profiles for predicated values and desirability function for removal of fluoroquinolones.

#### *3.3. Adsorption Kinetics*

The adsorption kinetics data (Figure 7) were used to study the adsorption process of FQ antibiotics onto the surface of the nanocomposite. The data were analysed using three commonly used kinetic models; namely, pseudo-first-order, pseudo-second-order and intraparticle diffusion. The equations of these kinetic models are widely reported (See Table 1), and they were adapted from the literature [66,67,69].

**Figure 7.** Effects of contact time on DANO, ENRO and LEVO by MMPC/Cyc-Chit kinetic modelling.

The estimated parameters are presented in Table 4. As it is indicated, the R<sup>2</sup> values achieved for pseudo-second-order were constantly higher compared to those of pseudo-first-order. In addition, the adsorption capacities obtained using the pseudo-second-order kinetic model were in agreement with experimental values. These outcomes suggested that the rate-determining step might be dominated by chemical interactions of FQ antibiotics with the homogenous surface of the adsorbent. The chemisorption mechanism might be driven by electrostatic attraction between the adsorbent and FQs. The dissociated forms FQ antibiotics have carboxylate and nitrogen functional groups that can bind on the positive or negative adsorbent surface.


**Table 4.** Parameters for the various kinetic models fitted onto data obtained for adsorbate solutions and results.

To further understand the adsorption mechanism and the rate-controlling step, the adsorption data were fitted to the intraparticle diffusion model [74]. The plots of q<sup>t</sup> versus *t* 1/2 for the investigated adsorbates showed multi-linearity (Figure A5). These plots indicated that there were two adsorption steps that took take place. According to the literature, the steeper first-step is due to diffusion of FQ antibiotics through the solution to the mesoporous nanososorbent. The second stage is attributed to transfer of the DANO, ENRO and LEVO charged molecules into intraparticle active sites or pores of the nanocomposite. Furthermore, it was noticed that the linear part of the first step did not pass through the origin. This signified that intraparticle diffusion was not the only rate-determining step [74,75]. Therefore, it can be concluded that adsorption processes were driven by both surface adsorption and intra-particle diffusion. The intraparticle diffusion rate constants for the first and second stages (*k*id1, *k*id2), correlation coefficients and intercept, C are indicated in Table 4. The *R* 2 values suggested that the adsorption of FQs on MMPC/Cyc-Chit nanocomposite may be dominated by intra-particle diffusion.

#### *3.4. Adsorption Isotherms*

To study the relationship between the concentration of FQs retained by the surface of the adsorbent and that of residual FQs in the bulk solution, the equilibrium studies were performed. The adsorption data were determined using Langmuir, Freundlich, Hill and Langmuir–Freundlich (Sips) isotherm models, and the model expressions are summarized in Table 1. The adsorption isotherms of FQs using the nanocomposite were carried out at 25 ◦C, and the pH of the solution, mass of adsorbent and contact time were set at 7, 30.0 mg and 30 min, respectively. Figure 8 demonstrates the adsorption isotherms of FQs onto nanocomposite from aqueous solutions. The isotherm models were used to derive various parameters related to the adsorption process.

**Figure 8.** Sorption isotherms—modeling with Langmuir, Freundlich, Sips and Hill (MA: 34 mg; sonication time: 25 min; pH 7; temperature: 25 ± 3 ◦C).

− Table 5 shows the summary of parameters derived from Langmuir, Freundlich, Hill and Langmuir–Freundlich (Sips) isotherm model plots. Comparing the correlation coefficients (R<sup>2</sup> ) values for the Langmuir and Freundlich isotherm models, it was detected that DANO, ENRO and LEVO were better suited to the Langmuir model. These findings demonstrated that the adsorption process took place as a monolayer of FQs on the surface of the adsorbent. The maximum DANO, ENRO and LEVO adsorption capacities for the adsorbent were 130, 196 and 194 mg g−<sup>1</sup> , respectively. As seen in Table 5, the Freundlich isotherm model was also used to some extent; however, it was not as good as the Langmuir isotherm model.

The separation factor (*R*L) values for each adsorbate (Table 5) were used to examine wherever the adsorption process was favourable. The values were calculated from the Langmuir isotherm and they suggested that the selected FQ antibiotics were easily adsorbed onto the nanocomposite because *R*<sup>L</sup> values were less than 1. In addition, the observation was also made that *R*<sup>L</sup> values decrease with an increase in the initial concentration, stipulating that the adsorption of FQs was more favourable at high concentrations [76]. The equilibrium adsorption data were also modelled using three-parameter isotherms expressions (Hill and Langmuir–Freundlich isotherm models) and parameter values are illustrated in Table 5. As seen, both models confirmed that the adsorption process assumes the homogeneous monolayer on the heterogeneous surface of the nanocomposite. In addition, the Hill model exponent *n*<sup>H</sup> values for DANO, ENRO and LEVO were greater than 1, indicating that the binding interaction between FQ antibiotics and nanocomposite was in the form of positive cooperativity [72].



#### *3.5. Adsorption Thermodynamics*

The effect of temperature in the removal of DANO, ENRO and LEVO using the nanocomposite was investigated. The thermodynamic parameters, such as Gibs energy (∆*G* ◦ ) enthalpy (∆*H*◦ ) and entropy (∆*S* ◦ ) are presented in Table 6. The values of ∆*G* ◦ were calculated using Equation (4), whereas the ∆*H*◦ and ∆*S* ◦ values were estimated from the slopes and intercepts of the plots that were obtained using Equation (4). As seen, the ∆*G* ◦ values were negative at all investigated temperatures. This phenomenon suggested that the adsorption was spontaneous [65,66,77]. Furthermore, the positive values of ∆*H*◦ demonstrated that the adsorption interaction between the antibiotics and the nanocomposite was characterised by endothermic nature [24,65,66,77]. The values of ∆*H*◦ were higher than 20.9 kJ/mol, confirming that the adsorption processes of FQ antibiotics were dominated by a chemisorption mechanism [78]. Moreover, the positive values of ∆*S* ◦ suggested that there is an increase in randomness at the boundary of solid/liquid phases, which might reveal the possible structural variations of the analyte and adsorbent [65,66,77].


**Table 6.** Thermodynamic parameters for DANO, ENRO and LEVO sorption on MMPC/Cyc-Chit.

#### *3.6. Comparison of Sorption Capacities for Various Adsorbents*

To compare the performance of the nanocomposite for the adsorption of FQ antibiotics, adsorption capacities of DANO, ENRO and LEVO on various adsorbents is presented in Table 7. As observed in Table 7, the adsorption capacity of the nanocomposite was comparable even better than other adsorbents reported elsewhere [25,61,65,67,69–71,74]. However, the adsorption capacity was lower than those reported by references [66,73].


**Table 7.** Comparison of sorption capacities for DANO, ENRO and LEVO fluoroquinolones with various composite sorbents at 25 ± 1 ◦C.

#### *3.7. Regeneration and Reusability Studies*

Regeneration and reusability for spent adsorbent are two of most crucial factors from the cost-effective point of view. This study investigated the possibility of regenerating and reusing the spent nanocomposites-loaded with FQ antibiotics. The regeneration and reusability process was performed according to the procedure described in Section 2.5. As seen in Figure 9, the regenerated nanocomposite retained 90–100% of its adsorption capacity toward the removal of DANO, ENRO and LEVO, after five cycles of the desorption–adsorption. Furthermore, the adsorption capacities of the spent adsorbent for removal of DANO, ENRO and LEVO remained at 88, 122 and 116 mg g−<sup>1</sup> , respectively, after the eighth cycle. Furthermore, the spent adsorbent after the eighth cycle was used for the removal of FQ antibiotics. It was observed that even though the adsorption capacities decreased, the removal efficiency remained above 95%. These results demonstrated that the nanocomposite can be reused several times without affecting its removal efficiency. Additionally, it was then concluded that the prepared nanocomposite had relatively high chemical and thermal stability. −

**Figure 9.** Regeneration of MMC/Cyc-Chit nanocomposite for eight successive adsorption–desorption cycles. Experimental conditions: *C*<sup>0</sup> = 50 mg L−<sup>1</sup> ; extraction time = 180 min; pH = 3.0; mass adsorbent = 36 mg; desorption solvent: acidified water and acetonitrile mixture (55:45 ratio); desorption time = 10 min.

#### *3.8. Application to Real Samples*

The applicability of the synthesized nanocomposite was assessed for the adsorptive removal DANO, ENRO and LEVO from real water samples; i.e., tap water, river water, influent and effluent wastewater. The river water, influent and wastewater samples were filtered using a 0.22 µm syringe filter. The target analytes were detected in influent and effluent wastewater and their concentrations ranged from 58 to 1230 µg L−<sup>1</sup> , whereas only traces of ENRO were detected in river water samples (Table 8). As seen, the overall removal efficiencies of DANO, DANO and LEVO in spiked water samples ranged from 90–99% and the concentration of the target analyte reduced significantly. These outcomes demonstrate the good performance of the adsorbent for water and wastewater treatment.



#### *3.9. Cost Analysis for the Preparation of Adsorbent*

The cost of the adsorption process is predominantly dependent on the cost of adsorbent used for the removal of organic and inorganic pollutants from wastewater [79]. Therefore, relatively low-cost materials with properties that are comparable to commercially available adsorbents are required. The cost estimation breakdown for the preparation of the mesoporous carbon and nanocomposite is presented in Table 9. In comparison with the other commercially available nanomaterials, such as multi-walled carbon nanotubes (R2354/g, Sigma-Aldrich), graphene oxide (R2163/g, Sigma-Aldrich) and mesoporous carbon (R2623/5 g, Sigma-Aldrich), the cost of mesoporous carbon and nanocomposite is much cheaper. A kilogram of the prepared nanocomposite will cost about R23262.50 (\$1324.85). The regeneration and reusability studies of the nanocomposite further reduce the cost of the adsorbent, since one batch can be reused at least five times. This confirms that the production of the MMPC/Cyc-Chit nanocomposite is economical and sustainable. Furthermore, regeneration and reusability are value-added properties of MMPC/Cyc-Chit nanocomposite as a promising adsorbent in the treatment of wastewater contaminated with emerging contaminants. The incorporation of magnetic nanoparticles led to the easy and fast separation (using external magnet) of adsorbent from aqueous solutions. The spent adsorbent can be first treated by the Fenton process (advanced oxidation processes, AOPs) before degrading the adsorbed pollutants. Furthermore, chitosan and β-cyclodextrin are types of fully biodegradable natural materials. This means that once the organic pollutants have been degraded by Fenton process, the adsorbent can be buried in the soil to allow biodegradation process.


**Table 9.** Cost estimation breakdown for the production of magnetic mesoporous carbon/β-cyclodextrin– chitosan (MMPC/Cyc-Chit) nanocomposite.

#### **4. Conclusions**

A magnetic mesoporous carbon/β-cyclodextrin–chitosan (MMPC/Cyc-Chit) hybrid nanocomposite adsorbent was synthesized by the facile hydrothermal method. The prepared MMPC/Cyc-Chit adsorbent was characterized using BET, XRD, TEM and FTIR. The adsorption capabilities of the synthesized nanocomposite were studied in a multicomponent system employing the ultrasound-aided removal process. The effects of independent variables (sample pH, mass of adsorbent and sonication power level) were investigated and optimized using RSM based on the CCD. The use of the ultrasound system led to rapid achievement of equilibrium and improved the adsorption process due to intensified mass transfer as well as the enhanced affinity between adsorbate and adsorbent due to acoustic cavitation effects. The adsorption isotherm equilibrium data followed the Langmuir model, suggesting that the surface of the adsorbent is coated as monolayer coverage by DANO, ENRO and LEVO molecules. Furthermore, the three-parameter models confirmed that the adsorption process assumes the homogeneous monolayer on the heterogeneous surface of the MMPC/Cyc-Chit nanocomposite. The kinetic data were best described by the pseudo-second-order model proposing that the adsorptive removal process was dominated by chemisorption. The thermodynamic parameters which include ∆G◦ , ∆H◦ , and ∆S ◦ indicated the adsorption process was feasible, spontaneous and endothermic in nature. In addition, the magnitude of ∆H◦ suggested that the removal of FQ antibiotics was via chemisorption and these findings agreed with the kinetic data. The synthesized MMPC/Cyc-Chit nanocomposite showed relatively high chemical and thermal stability and reusability over five adsorption–desorption cycles. The adsorption process was also applied in the removal of fluoroquinolones from real wastewater, tap water and river water samples. The results obtained demonstrated thatMMPC/Cyc-Chit nanocomposite can be applied in water and wastewater treatment process.

**Author Contributions:** Formulated the research idea, G.P.M., K.M.D. and P.N.N.; designed the experiments, G.P.M. and P.N.N.; performed the actual experiments and data collection, G.P.M.; carried out the analysis of data, G.P.M. and P.N.N.; wrote the first draft of the manuscript, G.P.M.; reviewed and edited the final version of the manuscript, P.N.N.; collected real water samples, G.P.M. and P.N.N.; supervision K.M.D. and P.N.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the University of Johannesburg, South Africa (Department of Chemical Sciences) and the National Research Foundation (grants 91230 and 99270, South Africa).

**Acknowledgments:** The authors wish to thank the University of Johannesburg, department of Chemical Sciences for providing laboratory space.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**


**Table A1.** Physicochemical properties of water samples.

**Figure A1.** The 3D surface response plots describing the interactions of investigated parameters (ENRO).

**Figure A2.** The 3D surface response plots describing the interactions of investigated parameters (LEVO).

**Figure A3.** Profiles for predicated values and desirability function for removal of ENRO.

**Figure A4.** Profiles for predicated values and desirability function for removal of LEVO.

**Figure A5.** Intraparticle diffusion model.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Dopamine Grafted Iron-Loaded Waste Silk for Fenton-Like Removal of Toxic Water Pollutants**

#### **Md Shipan Mia** † **, Biaobiao Yan** † **, Xiaowei Zhu, Tieling Xing \* and Guoqiang Chen**

National Engineering Laboratory for Modern Silk, Soochow University, Suzhou 215123, China; shipan0143@gmail.com (M.S.M.); yanbiao136@163.com (B.Y.); 18771093295@163.com (X.Z.); chenguojiang@suda.edu.cn (G.C.)

**\*** Correspondence: xingtieling@suda.edu.cn; Tel.: +86-512-6706-1175

† These authors contributed equally to this work.

Received: 24 October 2019; Accepted: 4 December 2019; Published: 9 December 2019

**Abstract:** Dispersion of iron was achieved on waste silk fibers (wSF) after grafting of polydopamine (PDA). The catalytic activity of the resulting material (wSF-DA/Fe) was investigated in Fenton-like removal of toxic aromatic dyes (Methylene Blue, Cationic Violet X-5BLN, and Reactive Orange GRN) water. The dye removal yield reached 98%, 99%, and 98% in 10–40 min for Methylene Blue, Cationic Violet X-5BLN, and Reactive Orange GRN, respectively. The catalytic activity was explained in terms of the effects of temperature, dyes, and electrolytes. In addition, the kinetic study showed that the removal of dyes followed pseudo-1st order adsorption kinetics. These findings allow envisaging the preparation of fiber-based catalysts for potential uses in environmental and green chemistry.

**Keywords:** waste silk; dopamine; iron particles; wastewater treatment

#### **1. Introduction**

The control of the water pollution is becoming one of the major challenges worldwide. As for extensive industrialization and subsequently the massive discharge of numerous types of organic toxicants, such as dyes, phenols, and nitroaromatic compounds [1–3] releasing in water supplies is becoming a serious issue. However, all of the pollutants are receiving tremendous attention from the water researchers, among them the aromatic dyes received one of the highest concerns due to their high toxic nature, chemical stability, and their resistance to conventional treatment methods [1]. Many approaches have been introduced by the several researchers where hazardous aromatic dyes were removed by various processes like biodegradation, chemical oxidation, adsorption, and so on [4–8]. Various heterogeneous catalytic systems with metal oxides or hydroxides like CuO, ZnO, TiO2, and FeOOH as catalysts, [5,9–11] as well as advanced oxidation systems have also been introduced all over the decades [12–14].

Fenton and Fenton-like treatments showed superior pollutant degradation/reduction efficiency compared with microbial oxidative metabolism, and physical adsorption also has some drawbacks [15–18]. Several new treatment configurations were introduced to optimize the Fenton process including a photo-assisting process to reduce iron supply [19,20], use of solid iron, or avoid modifying the natural pH value of wastewater [21], and so on. However, one of the drawbacks of this system still remained the separation of iron sludge in the solution after the treatment and recycling before discharging the treated wastewater. This is a time consuming and costly procedure [22]. Immobilization of iron particles (Fe) may become an appropriate and permanent remediation technique to solve this problem. However, it is important to immobilize Fe onto a carrier that can be easily separated from the contaminated water solution as well as provide no/less harm on subsequent management. The Fe immobilized inpolyacrylic acid(PAA) [23,24], starch [25], and polyglycol [26] reported in different

literature are difficult to recycle once they are used to treat contaminants. To avoid secondary water contamination, immobilization of zero-valent iron (ZVI) onto solid supports, for example, polymeric membranes [27] and activated carbon [28], could be an ideal option. Morshed et al. [29] immobilized and stabilized ZVI nanoparticles on fibrous polyester. Very few literatures have been found on the use of textile fibers or fabrics to immobilize iron particles aiming heterogeneous catalytic application.

Silk fibers are one of the most discussed fibers derived from natural sources due to mechanical and thermomechanical properties. However, due to their exceptional properties and complex production procedures, raw silk fiber spun out from silkworm cocoons is a luxurious material in many applications. Hence, the material constructed by silk would not be cost effective compared with other plant based natural fibers. On the contrary, during industrial process, lots of waste silk fibers or scrap silk fibers are produced, which are considered as waste and available at extremely low cost. Therefore, the increasing challenge of recycling them is another concerning issue. So utilizing waste silk into a sustainable remediation application can be an ultimate alternative.

Due to strong adhesion property, dopamine, a component of marine mussel gaining tremendous attention for surface modification of different materials [17–19], is being extensively used to prepare super-hydrophobic, antimicrobial, UV-blocking, conductive, as well as dye adsorption materials. Our previous study reported use of dopamine to prepare silk fabric with hydrophobicity, flame retardancy, and UV shielding properties [30]. However, no study reported the effectiveness of dopamine grafted waste silk fibers to prepare heterogeneous catalyst for Fenton-like wastewater treatment application.

In this work, dopamine was grafted on the surface of silk fibers by rapid oxidative polymerization followed by loading/immobilization of Fe particles on PDA grafted waste silk fibers. Hence, a fiber-based catalyst was prepared for heterogeneous Fenton-like removal of toxic aromatic dyes. The surface morphology and chemical composition of resultant catalyst were characterized by scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), and Fourier transform infrared (FT-IR) spectroscopy. The effectiveness of catalytic property was assessed under the influence of H2O<sup>2</sup> concentration, dye concentration, temperature, and electrolytes.

#### **2. Experimental**

#### *2.1. Materials*

Analytical grade dopamine hydrochloride (98.5%), sodium perborate(NaBO3·4H2O), ferric chloride hexahydrate (FeCl3·6H2O), hydrogen peroxide (H2O2, 33%), and ferrous sulfate heptahydrate (FeSO4·7H2O) were purchased from Shanghai Lingfeng Chemical Reagent Co., Ltd. (Shanghai, China) and used as received. Waste silk (average fiber diameter 10 µm) were obtained from Nantong Nafuer Clothing Co., Ltd (Nantong, China). All the dyes (Methylene Blue, Cationic Violet X-5BLN and Reactive Orange GRN) were purchased from Tianjin Tianshun Chemical Dyestuff Co., Ltd (Tianjin, China).

#### *2.2. Methods*

#### 2.2.1. Preparation of Dopamine Grafted Waste Silk

Scheme 1 illustrates the pathways used for grafting of dopamine on waste silk. Typically, 1.60 g of waste silk was wetted by deionized water as a pre-treatment. Then, 0.6 g of dopamine hydrochloride and 0.16 g of FeCl3·6H2O were dispersed in 300 mL deionized water. Wet waste silk fibers were added in dopamine solution and placed in a shaking water bath for 20 min followed by addition of 0.55 g of NaBO3·4H2O, the final solution was stirred at 50 ◦C for 50 min. The resultant dopamine grafted waste silk fibers (wSF-PDA) were rinsed and dried overnight in ambient condition.

**Scheme 1.** Schematic illustration of the preparation process of (**i**) grafting of polydopamine (PDA) on waste silk fibers (wSF) and (**ii**) loading of iron on dopamine grafted waste silk fibers (wSF-PDA).

#### 2.2.2. Preparation of wSF-PDA/Fe

− The loading of iron on dopamine grafted waste silk fibers were carried out using chemical adsorption and incorporation method as illustrated in Scheme 1. Here, 15.54 g of FeSO4·7H2O were dissolved in 1400 mL deionized water in a ceramic tray and stirred until uniform dispersion. Then, the prepared (as Section 2.2.1 described) dopamine-grafted waste silk fibers were immersed in the iron solution and kept at 25 ◦C for 24 h. The resultant wSF-PDA/Fe undergone successive rinse and stored in a desiccator before analysis and use.

#### *2.3. Material Characterizations*

μ Fourier transform IR spectroscopy of pristine and iron-loaded wSF was carried out using a Nicolet-5700 Fourier transform infrared spectrometer (MA, USA). The sample to be tested was cut into powder and sampled by potassium bromide. The scanning range was 600–4000 cm−<sup>1</sup> and the number of scans was 120 times. The samples were drilled before IR analysis, and background spectra were recorded on air. Surface morphologies of all wSF (before and after dopamine grafting and iron loading) were analyzed using desktop scanning electron microscope (Hitachi Ltd., Tokyo, Japan) at an accelerating voltage of 15 kV. Prior to SEM, samples were sprayed with a conducting resin followed by sputter coating with carbon films having a deposition depth of about 10 nm. Energy dispersive spectroscopy (EDS) was carried out using BRUKNER axes EDS analyzer mounted with SEM.

#### *2.4. Catalytic Activity of wSF-DA*/*Fe*

− To study the catalytic performance of wSF-PDA/Fe, the heterogeneous Fenton-like removal of dyes (methylene blue, cationic violet X-5BLN, and reactive orange GRN) in presence of hydrogen peroxide (H2O2) was investigated. Typically, 70 mL of dye solution (10–81 mg/L) were treated by using 0.1 mg wSF-PDA/Fe 8 µL (0.05–5 mmol/L) of H2O<sup>2</sup> at a specific temperature (25, 50, or 75 ◦C). A small amount of solution was taken through a Lab Sphere UV-1000F transmission analyzer (Lab sphere, Inc., North Sutton, VA, USA) to determine the specific absorbance at the characteristic peak of the dye, and the corresponding dye residual rate is calculated as per the Equation (1) [31]:

$$\text{Removal rate}\% = (\text{C}\_0 - \text{C}) / \text{C}\_0 \times 100 \tag{1}$$

where *C* is the concentration of the dye during removal at different time interval and *C*<sup>0</sup> is the initial concentration of the dye. *C*/*C*<sup>0</sup> was calculated at the maximum absorption wavelength of the dye in visible region. For example, methylene blue at 665 nm, cationic violet X-5BLN at 590 nm, and reactive orange GRN at 480 nm. Control experiments using wSF and wSF-PDA with/or without hydrogen peroxide were conducted to investigate the adsorption and catalytic property of the catalyst.

#### 2.4.1. Effect of H2O<sup>2</sup> Concentration on Removal Performance

The effects of different H2O<sup>2</sup> concentrations on dye degradation were investigated. The reaction conditions were as follows: the reaction temperature was 50 ◦C, the pH of the reaction was about 7, the concentration of the dye solution was 20 mg/L, and the wSF-PDA/Fe was 0.1 g. The concentration of hydrogen peroxide studied was 0.05, 0.1, 0.5, 1, 3, and 5 mmol/L.

#### 2.4.2. Effect of Pollutant Concentration on Removal Performance

The effects of dye concentration on removal efficiency using wSF-PDA/Fe catalyst were examined. All the experiments were carried out at 50 ◦C, pH = 7 in a shaking water bath where the initial concentration of H2O<sup>2</sup> reagent was 1 mmol/L, and the dopamine grafted iron-loaded waste silk (wSF-PDA/Fe) was 0.1 g. The dye concentrations were studied at 10, 20, 40, 60, and 80 mg/L, respectively.

#### 2.4.3. Effect of Reaction Temperature on Pollutant Removal Performance

The pollutant removal performances of wSF-PDA/Fe as function of different reaction temperatures were studied. Standard dye (20 mg/L) and H2O<sup>2</sup> (1 mmol/L) concentration was used for 0.1 mg of wSF-PDA/Fe catalyst. Three-reaction temperatures (25, 50, and 75 ◦C) were studied.

#### 2.4.4. Effect of Different Electrolytes (NaCl, Na2SO4) of Pollutant Removal Performance

In industrial production, dyeing wastewater contains not only dyes, but also many other inorganic and organic additives. Therefore, the effects of different electrolytes (NaCl, Na2SO4) on the removal efficiency of wSF-PDA/Fe were also investigated. All the experiments were carried out at 50 ◦C, pH = 7 in a shaking water bath for the 20 mg/L dye using 1 mmol/L H2O2, and 0.1 g wSF-PDA/Fe catalyst.

#### **3. Results and Discussion**

The results were presented and discussed in two separate parts, where the first part focused on analysis of waste silk fibers before and after PDA incorporation and iron loading, and the second part conferred the catalytic behavior of prepared Fe loaded waste silk fibers towards removal of various dyes.

Part 1: Analysis of Waste Silk Fibers Before and After Dopamine Incorporation and Iron Loading

Morphological analysis using SEM, as well as elemental and functional group analysis by means of EDS X-ray and infrared spectroscopy was used to characterize waste silk samples before and after PDA grafting and iron loading immobilization.

#### *3.1. Morphological Analysis*

The changes in surface morphology of the waste silk after different treatment were investigated (see Figure 1). The untreated waste silk (Figure 1a) exhibits a network of a randomly overlapping fibers with a smooth surface. However, a uniform layer of cluster can be observed in PDA grafted waste silk (see Figure 1b), indicating that the PDA is successfully grafted onto the fibers surface through rapid oxidative polymerization. A large amount of particles are found successfully loaded on the surface of wSF-PDA and formed wSF-PDA/Fe. In the case of Fe loaded waste silk fibers (see Figure 1c), several patches appear as regularly shaped clusters [30,32] in the form of layers as demonstrated in Scheme 2. The presence of functional groups on treated waste silk has been further established by EDS X-ray and infrared analysis (see proceeding sections).

**Scheme 2.** Schematic illustration of iron loading on waste silk.

#### *3.2. Fourier Transform Infrared (FT-IR) Analysis*

− − − − − In order to further investigate the changes in the structure of the waste silk surface after various surface modifications, the infrared spectrum analysis was carried out as shown in Figure 2. The untreated and treated waste silk demonstrated rough bands in the region 1660–1630 cm−<sup>1</sup> , which were attributed to the stretching vibration of amide I. Other characteristics peaks found at 1545–1525 cm-1 are the amide II and 1265–1235 cm−<sup>1</sup> is the amide III [33,34]. The FT-IR spectra of sample wSF-PDA and wSF-PDA/Fe showed a characteristic bending at 570 cm−<sup>1</sup> [35,36], indicating the formation of Fe–O due to the presence of PDA [30] and subsequent loading of iron particles. − − −

**Figure 2.** Infrared spectrum of untreated and functionalized waste silk.

#### *3.3. Loading Analysis*

EDS X-ray analysis further confirmed the changes in the chemical composition of the fabric surface before and after treatment (see Figures 3 and 4). The surface elemental composition of untreated and functionalized silk fibers is presented in Table 1. It can be concluded that the surface element ratio of N/C of the treated fabric is significantly smaller than the untreated sample [30], indicating successful grafting of PDA. After loading of iron, the peak intensity of Fe element (4.23%) can be noticed in wSF-PDA/Fe. A small quantity of iron element was noticed in wSF-PDA derived from the presence of some Fe-oxides in PDA.

**Figure 3.** Energy dispersive spectra of (**a**) wSF and (**b**) wSF-PDA/Fe.

**Table 1.** Surface elemental composition and loading analysis of samples (energy dispersive spectroscopy (EDS) analysis).


**Figure 4.** Energy Dispersive X-ray Spectroscopy (**A**) wS, (**B**)wSF-PDA, and (**C**) wSF-PDA/Fe.

Part 2: Analysis of Catalytic Fenton-Like Removal Toxic Water Pollutants

Three different dyes (methylene blue, cationic violet X-5BLN, and reactive orange GRN) were adopted to analyze the catalytic Fenton-like property of prepared fibrous catalyst (wSF-PDA/Fe). Dye removal as a function of time and influence of several factors, such as concentration of pollutants, reactants, temperature, and presence of electrolytes, has been investigated by means of UV-Vis colorimetric analysis.

#### *3.4. Degradation of Dyes as a Function of Time*

In order to investigate catalytic behavior of prepared iron loaded waste silk fibrous catalyst, Fenton-like degradation of methylene blue, cationic violet X-5BLN, and reactive orange GRN dye as a function of time was studied and analyzed by UV-Vis spectrophotometer (see Figure 5). Fenton reaction involves the combined use of ferrous ions and hydrogen peroxide (H2O2) to produce advanced oxidation potential active oxygen species capable of degrading organic contaminants. The dyes selected in this study are common industrial dyes containing an aromatic ring, which are resistant to traditional chemical and biological removal methods. As shown in Figure 5, the characteristic absorption peak intensity of methylene blue at 285 and 665 nm [35] decreased with the function of the reaction time. Once the reaction reached 10 min, the color almost completely disappeared, hence most degradation was achieved by 40 min of catalytic oxidation. Similar phenomena were found in the other two (cationic violet X-5BLN and reactive orange GRN) dyes, where characteristic absorption peak intensity of cationic violet X-5BLN at 275 and 590 nm and reactive orange GRN at 260 and 480 nm disappeared in 20 and 40 min, respectively. Iron loaded wSF fibrous catalyst showed good color removal efficiency, but no visible color removal was observed in control experiments performed using wSF-PDA or wSF-PDA/Fe without one or more reactants (Fe or H2O2) necessary to initiate Fenton reaction. Nonetheless, trivial reduction in color concentration was observed in various sample (wSF-PDA, wSF-PDA + H2O2, and wSF-PDA/Fe), which can be due to the adsorption characteristics of heterogeneous catalysts consistent with the literature [29,37,38].

**Figure 5.** UV-visible absorption spectrum of (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) reactive orange GRN removal as a function of time. (wSF-PDA/Fe 0.1 g, dye concentration 20 mg/L, H2O<sup>2</sup> concentration 1.0 mmol/L, T = 50 ◦C, pH = 7).

#### 3.4.1. Kinetics of Dye Degradation

λ The [instant/initial] absorbance ratio of the methylene blue, cationic violet X-5BLN, and reactive orange GRN dye at λ = 665, 590, and 480 nm (*A*<sup>t</sup> /*A*0) respectively, which accounts for the corresponding concentration ratio (*C*/*C*0) and allows plotting of ln(*C*/*C*0) as a function of time. Model validation of the pseudo-first-order kinetics for color removal with the catalyst is obtained by the linear evolution in time of ln(*C*/*C*0), as supported by *R* <sup>2</sup> values 0.98, 0.99, and 0.98, respectively (see Table 2). Plots summarized in Table 2 show that dye degradation exhibited good linear relationships of ln(*C*/*C*0) versus reaction time up to a certain time where maximum number of dyes was degraded following pseudo-first-order kinetics. These results are consistent with those found in previous reports [10,24,33,39–42].


**Table 2.** Pseudo First-order kinetics for the Fenton-like removal of methylene blue, cationic violet X-5BLN, and reactive orange GRN dye.

<sup>a</sup>Reaction time required for color removal; <sup>b</sup> *k*: rate constant for the 1st order kinetics and is expressed in min−<sup>1</sup> ; c*R* 2 : correlation coefficient of the linear regression. −

#### 3.4.2. Postulated Mechanism of Dye Removal

Based on the results explained in the above sections, a plausible mechanism for removal of all three dyes has been postulated. The underlying mechanisms of removal of methylene blue, cationic violet X-5BLN, and reactive orange GRN dyes in the system of iron loaded PDA grafted waste silk fibrous catalyst were considered for Fenton-like reaction in presence of hydrogen peroxide. The removal of all dyes was attributed to the synergistic effect caused by free radicals and other reactive species formed though a heterogeneous Fenton reaction [29]. The produced free radicals oxidize the dyes into colorless nontoxic substances as illustrated in Scheme 3.

**Scheme 3.** Schematic postulated main mechanism of degradation of methylene blue, cationic violet X-5BLN, and reactive orange GRN dyes.

The postulated main reaction mechanism involves three steps as follows:

→

(i) The process of producing reactive species

$$\rm{Fe^{2+} + H\_2O\_2 \to Fe^{3+} + HO^- + \bullet OH} \tag{1}$$

(ii) The process of color removal of dyes

$$\bullet \text{ types} + \bullet \text{OH} \rightarrow \text{Reaction intermediates} \tag{2}$$

(iii) The process of degradation

$$\text{Reaction intermediates} + \bullet \text{OH} \rightarrow \text{CO}\_2 + \text{H}\_2\text{O} \tag{3}$$

→

#### *3.5. Factors Influencing the Removal of Dyes*

#### 3.5.1. Effect of Different Samples

It can be seen from Figure 6 that different samples have a great influence on the degradation dye, and wSF-PDA/Fe-H2O<sup>2</sup> is the most effective for degrading dyes. The addition of wSF-PDA mainly brought about the adsorption of dyes by silk fibers and dopamine dominantly, and the dye removal rate could not meet the requirements. The addition of wSF-PDA and H2O<sup>2</sup> resulted in a lower degradation rate due to the weaker oxidative decomposition of H2O<sup>2</sup> itself and the weak adsorption of wSF-PDA. The wSF-PDA/Fe also played a dominant role in adsorption, and the dye removal rate was not significant. The addition wSF-PDA/Fe and H2O<sup>2</sup> generated a strong reaction system, which effectively increased the release rate of hydroxyl radicals and degraded most of the dyes in a short time.

In addition, it is interesting that wSF-PDA and wSF-PDA-H2O<sup>2</sup> have better degradation for methylene blue and cationic violet X-5BLN. This may be due to the negative charge on the surface of the wSF-PDA sample, which can better absorb cationic dyes and achieve effective degradation.

**Figure 6.** The *C*/*C*<sup>0</sup> of different samples in the three dye solutions (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) active orange GRN. (wSF-PDA/Fe 0.1 g, dye concentration 20 mg/L, H2O<sup>2</sup> concentration 1.0 mmol/L, T = 50 ◦C, pH = 7).

#### 3.5.2. Effect of H2O<sup>2</sup> Concentration

The effect of H2O<sup>2</sup> concentration (0.05, 0.1, 0.5, 1, 3, and 5 mmol/L) on the removal of dyes was studied. The evaluation of color removal in terms *C*/*C*<sup>0</sup> as shown in Figure 7 shows that, H2O<sup>2</sup> concentration below 0.5 mmol/L showed poor or insufficient dye removal. However, 0.5 to 5 mmol/L of H2O<sup>2</sup> concentration showed most removal of dyes in similar experimental condition. When the concentration of H2O<sup>2</sup> was 0.05 and 0.1 mmol/L, the dye degradation rates after reaction for 40 min were 48.1% and 57.5%, respectively. At this point, it can be clearly seen that the low concentration of H2O<sup>2</sup> was too diluted to react with the mineralized iron to form sufficient hydroxyl radicals to fully degrade methylene blue. When the concentration of H2O<sup>2</sup> was gradually increased to 0.5, 1, 3, and 5 mmol/L, the final degradation rate reached nearly 97% to 98%, but the efficiency of the reaction was also quite different. The H2O<sup>2</sup> concentration of 1 mmol/L was the appropriate, and the degradation rate was the faster and saturated within 15–20 min, while the reaction rate of high concentration H2O<sup>2</sup> was decreased. This phenomenon indicated that high concentration of H2O<sup>2</sup> cannot increase the degradation rate as linear regression rather showed no improvement after saturation. The similar phenomena can be noticed for other two dyes (cationic violet X-5BLN and reactive orange GRN) as well (see Figure 7b,c).

**Figure 7.** Evolution in time of *C*/*C*<sup>0</sup> of (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) reactive orange GRN removal as a function of H2O<sup>2</sup> concentration. (wSF-PDA/Fe 0.1 g, dye concentration 20 mg/L, T = 50 ◦C, pH = 7).

#### 3.5.3. Effect of Dye Concentration

The concentration of dyes is a precarious parameter of color removal rate, which influences the effectiveness of the removal process. The effect of dye concentration on the removal of dyes were studied in terms of different concentration (10, 20, 40, 60, and 80 mg/L) of dye solutions. Results presented in Figure 8 show a certain influence on the removal rate of all three dyes due to the variation in concentration of the dye solution. The results show (see Figure 8) the most removal of dyes at the initial concentrations from 10–20 mg/L, which moderates upon increase in dye concentration for a specific reaction time. However, for the maximum concentration (80 mg/L) of dye solution, the removal percentage remains above 80%. The highest 90% of removal rate on the maximum concentration are recorded for methylene blue. This may be due to the fact that in the case of high dye concentration, PDA grafted iron loaded waste silk catalyst forms a stable reaction system with H2O2, at which time the hydroxyl radical is not easily deactivated, and the probability of contact with the dye increases, eventually resulting in a high removal.

**Figure 8.** Evolution in time of *C*/*C*<sup>0</sup> of (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) reactive orange GRN removal as a function of dye concentration. (wSF-PDA/Fe 0.1 g, H2O<sup>2</sup> concentration 1.0 mmol/L, T = 50 ◦C, pH = 7).

#### 3.5.4. Effect of Reaction Temperature

Temperature in a catalytic reaction might increase or decrease the reaction rate, thus the effect of reaction temperature (25, 50, and 75 ◦C) on the removal of methylene blue, cationic violet X-5BLN, and reactive orange GRN dyes were studied and the results are shown in Figure 9. It can be seen from Figure 9a–c that the reaction temperature has a significant influence on the removal of dyes. In the range of 25–75 ◦C, the dye removal rate increases with increasing temperature. A slight increase in removal rate can be found in methylene blue removal (see Figure 9a). Whereas, a significant increase has been noticed in cationic violet X-5BLN and reactive orange GRN dyes. Figure 9c shows that, when the reaction temperature is 25 ◦C, the final dye removal rate of active orange GRN is 77.2% after 40 min reaction; when the reaction temperature is 50 and 75 ◦C, the final dye removal rate of active orange GRN is about 98.5% after 40 min reaction. However, it is clear that the reaction rate at 75 ◦C is significantly faster than 50 ◦C. This phenomenon indicates that the reaction rate under high

temperature conditions accelerates the decomposition of H2O<sup>2</sup> into hydroxyl radicals, which increases the velocity of dye degradation reaction. Similar phenomena are found in removal of cationic violet X-5BLN (see Figure 9b).

**Figure 9.** Evolution in time of *C*/*C*<sup>0</sup> of (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) reactive orange GRN removal as a function of reaction temperature. (wSF-PDA/Fe was 0.1 g, dye concentration 20 mg/L, H2O<sup>2</sup> concentrations 1.0 mmol/L, pH = 7).

#### 3.5.5. Effect of Different Electrolytes

The traditional dye house uses significant amounts of electrolytes during textile processing, so that considerable amounts of electrolytes are present in wastewater. Consequently, the effect of different electrolytes (NaCl, Na2SO4) on effectiveness of removal of dyes were studied. Stimulating phenomena are noticed in the results as shown in Figure 10. It can be seen that the electrolyte does influence the removal of degradation of methylene blue in a limited way. For cationic violet X-5BLN and reactive orange GRN, the effect of electrolyte is significant. This may be due to the fact that methylene blue belongs to the class of easily degradable dyes, and the structure of cationic violet X-5BLN and reactive orange GRN are more complicated. Therefore, the electrolyte has a greater influence on the latter two dyes. For different electrolytes, Cl<sup>−</sup> and SO<sup>4</sup> <sup>2</sup><sup>−</sup> can capture and destroy •OH [43,44], affecting the degradation process of dyes. In general, wSF-PDA/Fe still plays an important role in the degradation of dyes. − −

**Figure 10.** Evolution in time of *C*/*C*<sup>0</sup> of (**a**) methylene blue, (**b**) cationic violet X-5BLN, and (**c**) reactive orange GRN removal as a function of electrolyte. (wSF-PDA/Fe 0.1 g, dye concentration 20 mg/L, H2O<sup>2</sup> concentration 1.0 mmol/L, T = 50 ◦C, NaCl, Na2SO<sup>4</sup> concentration 40 g/L, pH = 7).

#### **4. Conclusions**

In this report, the loading of iron particle on dopamine grafted waste silk fibers were achieved successfully. The resultant material was investigated by means of physicochemical and catalytic property. SEM and EDS analysis confirmed the loading of PDA and iron particles on the surface of waste silk fibers, which revealed the potentiality of the prepared material as a catalyst in possible Fenton-like removal of aromatic dyes. The results of investigation of H2O<sup>2</sup> concentration, dye concentration, temperature, and electrolytes on dye removal indicated the prepared catalyst showed significant 98%–99% removal dyes in 10–40 min depending on the concentration of H2O<sup>2</sup> used. Postulate mechanism showed the most degradation of pollutants into nontoxic substance. Thus, this study can provide rational foundation towards further advancement in fiber based catalyst for Fenton-like removal of toxic pollutants in water.

**Author Contributions:** T.X. and G.C. conceived and designed the experiments; M.S.M., B.Y., and X.Z. performed the experiments and analyzed the data; M.S.M., B.Y., and T.X. wrote the paper. All authors discussed the results and improved the final text of the paper.

**Funding:** This work was supported by the National Natural Science Foundation of China (51973144, 51741301); the Major Program of Natural Science Research of Jiangsu Higher Education Institutions of China (18KJA540002); the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).

**Conflicts of Interest:** Authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Polystyrene Magnetic Nanocomposites as Antibiotic Adsorbents**

**Leili Mohammadi <sup>1</sup> , Abbas Rahdar 2,\* , Razieh Khaksefidi <sup>3</sup> , Aliyeh Ghamkhari <sup>4</sup> , Georgios Fytianos <sup>5</sup> and George Z. Kyzas 5,\***


Received: 16 May 2020; Accepted: 5 June 2020; Published: 9 June 2020

**Abstract:** There are different ways for antibiotics to enter the aquatic environment, with wastewater treatment plants (WWTP) considered to be one of the main points of entrance. Even treated wastewater effluent can contain antibiotics, since WWTP cannot eliminate the presence of antibiotics. Therefore, adsorption can be a sustainable option, compared to other tertiary treatments. In this direction, a versatile synthesis of poly(styrene-block-acrylic acid) diblock copolymer/Fe3O<sup>4</sup> magnetic nanocomposite (abbreviated as P(St-*b*-AAc)/Fe3O4)) was achieved for environmental applications, and particularly for the removal of antibiotic compounds. For this reason, the synthesis of the P(St-*b*-AAc) diblock copolymer was conducted with a reversible addition fragmentation transfer (RAFT) method. Monodisperse superparamagnetic nanocomposite with carboxylic acid groups of acrylic acid was adsorbed on the surface of Fe3O<sup>4</sup> nanoparticles. The nanocomposites were characterized with scanning electron microscopy (SEM), X-ray diffraction (XRD) and vibrating sample magnetometer (VSM) analysis. Then, the nanoparticles were applied to remove ciprofloxacin (antibiotic drug compound) from aqueous solutions. The effects of various parameters, such as initial drug concentration, solution pH, adsorbent dosage, and contact time on the process were extensively studied. Operational parameters and their efficacy in the removal of Ciprofloxacin were studied. Kinetic and adsorption isothermal studies were also carried out. The maximum removal efficiency of ciprofloxacin (97.5%) was found at an initial concentration of 5 mg/L, pH 7, adsorbent's dosage 2 mg/L, contact time equal to 37.5 min. The initial concentration of antibiotic and the dose of the adsorbent presented the highest impact on efficiency. The adsorption of ciprofloxacin was better fitted to Langmuir isotherm (R<sup>2</sup> = 0.9995), while the kinetics were better fitted to second-order kinetic equation (R<sup>2</sup> = 0.9973).

**Keywords:** ciprofloxacin; Polystyrene nanocomposite; modifications; adsorption; characterizations

#### **1. Introduction**

Aside from the well-known pollutants and contaminants in the aquatic environment, compounds of emerging concern (CECs) may impact aquatic life even in very low concentrations [1]. Wastewater influents and effluents can contain CECs, due to their presence in everyday products, such as detergents, fabric coatings, pharmaceuticals, cosmetics, beverages and food packaging [2]. Pharmaceuticals are

being detected in drinking and surface water, and although not very persistent, the continuous re-entering increases their abundance, and renders them pseudo-persistent [3]. Pharmaceuticals, include diverse types of compounds, e.g., antibiotics and show low biodegradability. CECs cannot be removed completely by wastewater treatment plants (WWTPs) [2], since WWTP were not designed to treat CECS. In some cases, even less than 10% of CECs is removed, making WWTP effluents a major factor for introducing CECs into the environment [2]. Recently, great attention is given to adsorption technique [4–19], which is easily applied to the last stage of wastewater treatment plants (WWTPs), with the aim of removing all residues that were not separated and removed from the previous stages.

In particular, special attention is given to find appropriate ways to effectively treat antibiotics from effluents, due to their strong resistance to various decontamination techniques [20]. Available statistics indicate that 100–200 tn of antibiotics are used annually worldwide. As a result, the risk of water resources contamination by these compounds is very high. The residue of those antibiotics in the form of major constituents or metabolites has also been observed in WWTP.

It is noteworthy to mention that the inability of WWTP to remove antibiotics leads to the discharge of those compounds into surface water and underground waters. The inadequate and incorrect use of those compounds, and their continuous entry into the environment, leads to biodistribution and faulty resistance [21,22]. Of the large antibiotic classes, fluoroquinolones are worth mentioning. Antibiotics in this family include Ciprofloxacin (CIP), epinephrine, and norfloxacin. The presence of fluorine atoms in combination with these antibiotics makes these compounds particularly stable, so they are considered to be very dangerous and toxic pollutants in the environment. CIP is detected in sewage and surface water in medical effluents and pharmaceutical plants. The antibiotic can be adsorbed into the sludge and, if applied as fertilizer, it is accumulated in the soil and enters into plants [23]. CIP was observed in surface waters and wastewaters at concentrations below 1 µg/L, while in medical wastewaters in 150 µg/L. Therefore, it is mandatory to find and apply an efficient method for ciprofloxacin removal.

The most important methods used to remove and separate the drug compounds from water and sewage include ozonation, nanofiltration, electron radiation, ion exchange, chemical coagulation and photocatalytic oxidation, all of which have high performance and operation costs [24–29]. Nowadays, nanotechnologies are mainly used in water and wastewater treatment, using materials like iron nanoparticles, zeolites and magnetic nanomaterials [30,31]. Among the various methods, adsorption is a simple, environmental friendly, fast, highly efficient and low-cost solution, making it one of the most favorable methods [32–37].

The removal of pharmaceuticals by adsorption has been the focus of many studies. So far, as adsorptive materials, activated carbon [38–41] or zeolites [42,43] have been widely used in wastewater treatment. The removal of pharmaceuticals by adsorption shows great potential, due to its easy application into existing water treatment processes. On the other hand, issues regarding adsorbent stability and regeneration costs lead to R&D of innovative and effective adsorbents from polymeric materials. Adsorption processes, such as activated carbon-based have high capital cost, and ineffectiveness and non-selectivity against vat and disperse dyes. Furthermore, saturated carbon regeneration is expensive and leads to adsorbent loss. Depending on the demand, cost, and the nature of the pollutant to be adsorbed, the adsorbents are either disposed or regenerated for future use. The regeneration process of adsorbents needs to be cheap and environmental friendly by recovering valuable adsorbates while reducing the need of virgin adsorbents.

In this study, a versatile synthesis of poly(styrene-block-acrylic acid) diblock copolymer/Fe3O<sup>4</sup> magnetic nanocomposite (abbreviated as P(St-*b*-AAc)/Fe3O4)) was achieved for environmental applications with a focus on the removal of ciprofloxacin. The nanocomposites were characterized with SEM, XRD and VSM analysis. The nanoparticles were then applied to remove ciprofloxacin (antibiotic drug compound) from aqueous solutions, evaluating the effect of certain important parameters such as the solution's pH, initial ciprofloxacin concentration, adsorbent dosage and contact time.

146

#### **2. Materials and Methods**

#### *2.1. Materials*

To begin, 4-cyano-4-[(phenylcarbothioyl) sulfanyl] pentanoic acid, as a RAFT agent, was synthesized [32]. Acrylic acid (AAc), styrene (St) monomers, 2, 2-azobisisobutyronitrile (AIBN), and dimethylformamide (DMF), FeCl2·4H2O, 99% and FeCl3·6H2O, 98% were purchased from Merck (Darmstadt, Germany). ‧ ‧

The antibiotic model compound used in the present study is ciprofloxacin, purchased from Merck (Germany). Its molecular structure is presented in Figure 1. When it comes to ciprofloxacin's dissociation and isoelectric constants, the isoelectric point has a value of pI = 7.14, which is calculated by the average of pKa<sup>1</sup> = 6.09 and pKa<sup>2</sup> = 8.62. This portrays the two ionizable functional groups of ciprofloxacin; the 6-carboxylic group and the N-4 of the piperazine substituent. pKa1 corresponds to the dissociation of a proton from the carboxyl group, and pKa<sup>2</sup> corresponds to the dissociation of a proton from the N-4 in the piperazinyl group [44].

**Figure 1.** Chemical structure of CIP and its ionizable forms.

#### *2.2. Synthesis of Poly(styrene) Homopolymer*

RAFT agents (10 mg, 0.036 mmol), styrene monomer (4 mL, 34.96 mmol) and AIBN (3.0 mg, mmol) were added in a 100-mL flask; the reaction was achieved with three freeze pump-thaw cycles under a nitrogen atmosphere. The solution was put to an oil-bath with a temperature of 75 ◦C for 24 h. The flask was then quenched by cooling. The polystyrene homopolymer was precipitated in methanol. Finally, drying of the product under vacuum at 25 ◦C for 24 h took place [22].

#### *2.3. Synthesis of Poly(styrene-block-acrylic acid), Sphere Superparamagnetic Iron Oxide Nanoparticles (SPIONs) and Poly(St-b-AAc)*/*Fe3O<sup>4</sup> Supermagnetic Nanocomposite*

Macro-RAFT agent (PSt, 200 mg, 19.8 mmol), AAc monomer (1.56 mL, 28.24 mmol), AIBN (3 mg, mmol) and DMF (10 mL) were charged in a two-neck reactor. The reaction was induced using three freeze pump-thaw cycles under a nitrogen atmosphere. The reaction solution was put to an oil-bath with a temperature of 75 ◦C for 24 h. The reaction mixture was then precipitated in cold diethyl ether (150 mL) and dried under vacuum at 25 ◦C. The SPIONs were synthesized using a co-precipitation method, as described in literature [22,45]. The poly(St-*b*-AAc)/Fe3O<sup>4</sup> supermagnetic nanocomposite was synthesized as described in a previous study [46]. The final product is magnetic nanocomposite [22] and its structure is illustrated in Figure 2.

**Figure 2.** Structure of the prepared poly(St-*b*-AAc)/Fe3O<sup>4</sup> .

It merits clarification that the objective of using magnetic nanoadsorbents and not common adsorbents was the easier separation of solid adsorbent particles from the solution at the end of the process. Due to magnetic particles, using an external magnetic field, poly(St-*b*-AAc)/Fe3O<sup>4</sup> was easily and fast separated from the aqueous solution after adsorption experiments. Also, the preparation of the polysterene nanocomposites was possible (instead of single magnetic particles Fe3O4), because it contains functional groups which increase its adsorption capacity.

#### *2.4. Characterization of Nanoadsorbents*

α For the XRD patterns, a Bruker XRD diffractometer (Billerica, MA, USA) with CuKα radiation was used. SEM (model Mira 3XMU, TESCAN company, Brno, Czech Republic) was used to study the morphology of nanoparticles.

#### *2.5. Preparation of CIP Solutions*

λ Ciprofloxacin hydrochloride (purity 99.8%) was purchased from Alborz Pharmaceutical Company of Qazvin (Qazvin, Iran), and used to prepare the stock CIP solution (100 mg/L prepared with the fixed pre-weighted amount of CIP and the respective volume of Milli-Q ultra-pure water). The residual concentration of CIP after the adsorption experiments was analyzed by (using a) UV-vis spectrophotometer (model Hach DR5000, Duesseldorf, Germany). The concentration of CIP was measured based on previous studies at a wavelength of λmax = 274 nm [47].

#### *2.6. Adsorption Experimental Design Method and Data Analysis*

In this study, the 7.0.1 Design Expert software was used to determine the number of experiments and the amount of parameters, and to perform the final analysis of the data obtained after the process (Table 1). The measurement of the level of pollutant removal was carried out with the standard design of the statistical model of the CCD (RSM). The main parameters affecting the process are: the initial pH of the medium in the range of 4 to 10, the amount of nanoparticles used in the reaction of 1 to 3 mg/L, the initial concentration of antibiotic ranging from 5 to 25 mg/L, and the reaction time (15 to 60 min).

**−α α**


**Table 1.** Design parameters together with the values and regions selected.

After the determination of optimal conditions and modeling of the process, the rate of CIP removal was investigated. Finally, the process efficiency in CIP removal was determined using the following equation. The removal (R, %) was also calculated based on the following formula: 0 f C C Re moval 100% 

$$\text{Removal} = \left(\frac{\text{C}\_0 - \text{C}\_f}{\text{C}\_0}\right) \times 100\%. \tag{1}$$

In this relation, R is the efficiency, C<sup>0</sup> (mg/L) is the initial concentration of CIP, and C<sup>f</sup> (mg/L) denotes the CIP concentration at the time of t. The amount of adsorbed CIP at equilibrium Q<sup>e</sup> (mg/g) was calculated from the following equation. In this relation, C<sup>0</sup> (mg/L)is the initial concentration of CIP, C<sup>e</sup> (mg/L) denotes the CIP concentration at the time of t, m (g) is the adsorbent mass, and V (L) is the sample volume: 0 e C CV

$$\mathbf{Q}\_{\mathbf{e}} = \frac{(\mathbf{C}\_0 - \mathbf{C}\_{\mathbf{e}})\mathbf{V}}{\mathbf{m}}.\tag{2}$$

#### **3. Results**

#### *3.1. Characterizations*

The morphologies of the P(St-*b*-AAc)/Fe3O<sup>4</sup> nanocomposite are spherical, with Daverage of 30 nm (Figure 3). It is obvious that the size of spheres is not the same for all particles, due to possible aggregation, but the uniformity regarding the shape is almost the same (spherical).

**Figure 3.** SEM image of P(St-*b*-AAc)/Fe3O<sup>4</sup> nanocomposite.

θ The X-ray diffraction patterns (XRD) resulting from the P(St-*b*-AAc)/Fe3O<sup>4</sup> superparamagnetic nanocomposite are indicated in Figure 4. The resulting peaks at 2θ equal to 30.28, 35.48, 43, 53.4, 57.16, and 63.04◦ correspond to (221), (312), (400), (421), (512), and (440) prisms of P(St-*b*-AAc)/Fe3O<sup>4</sup> nanocomposite crystalline structure, respectively (Figure 4) [22].

**Figure 4.** XRD patterns P(St-*b*-AAc)/Fe3O<sup>4</sup> magnetic nanocomposite.

The super paramagnetic behavior is demonstrated in Figure 5 with a VSM plot. The saturation magnetization of the P(St-*b*-AAc)/Fe3O<sup>4</sup> supermagnetic nanocomposite was around 26 emu/g, which shows that the synthesized magnetic nanocomposite is superparamagnetic.

**Figure 5.** Magnetization curve of P(St-*b*-AAc)/Fe3O<sup>4</sup> supermagnetic nanocomposite.

− − − − − − − − − − A Fourier-transform infrared spectroscopy of nanocomposites was conducted both prior and after adsorption of CIP, and the spectra are presented in Figure 6. Regarding the FTIR spectrum of CIP, a band around 3400 cm−<sup>1</sup> represent the vibrational frequency of stretching of the N–H bond of the imino moiety on the piperazine group of CIP. Absorption bands at 1633 cm−<sup>1</sup> and 1080 cm−<sup>1</sup> represent a primary amine (N–H) bend of the pyridone moiety and the C–F functional group, respectively. On the other hand, the FTIR spectrum related to the CIP-adsorbed nanoadsorbent is, in turn, related to the addition of the nanocomposite to the CIP solution. The broad peaks at 3463 cm−<sup>1</sup> are attributed to the stretching vibration of O–H bonds. O–H bonds were weaker and shifted down in the presence of ferrite nanoparticles. Similarly, the slight shift at around 1641 cm−<sup>1</sup> may be related to the interaction of carboxylic groups of polymer with the amino group of CIP (Figure 6). Also, by comparing the FTIR spectra, the intensity of the peaks after adsorption has increased in comparison to those before adsorption, due to the presence of ferrite nanostructures in the CIP solution.

**Figure 6.** FTIR spectra of CIP, and P(St-*b*-AAc)/Fe3O<sup>4</sup> (before and after adsorption).

#### *3.2. Data Analysis*

For the efficacy evaluation of antibiotic removal of ciprofloxacin using composite P(St-*b*-AAc), a composite design (one of the response surface methods) was used, and the effects of initial antibiotic concentration parameters, pH, adsorbance dose and reaction time were investigated. The response rate is presented in Table 2. The validity of the presented models was analyzed by ANOVA.



− − − −

In Table 3, the parameters A, B, C and D are the main effect of independent variables, which are the initial concentration of ciprofloxacin, pH, adsorbent dose, and contact time, respectively. The variable AB represents the effect of the initial concentration of ciprofloxacin (factor A) and pH (factor B), and variable A<sup>2</sup> represents the square effect of factor A on the desired response.


**Table 3.** ANOVA for Response Surface Quadratic Model.

The proposed model is presented as a modified model by removing non-significant variables via preserving the main effects of variables from the model for the antibiotic elimination efficacy in the following equation:

$$\text{Y(\%)} = 67.11 - 9.43\text{X}\_1 - 4.7\text{X}\_2 + 6.26\text{X}\_3 + 3.71\text{X}\_4 - 3.9\text{X}\_2\text{2}\tag{3}$$

In this regard, X1, X2, X<sup>3</sup> and X<sup>4</sup> are coded values of the initial concentrations of antibiotics, pH, adsorbent dose and reaction time. The linear regression is another test that was used to validate the model [48]. In this test, the coefficient of determination (R<sup>2</sup> = 0.8753), the adjusted coefficient of determination (R<sup>2</sup> adj = 0.8493) and the prediction coefficient (R<sup>2</sup> pred = 0.7955) were calculated and reported. Also, in each model, there is very little difference between the values of R<sup>2</sup> , R<sup>2</sup> adj and R<sup>2</sup> pred is observed.

#### The Effect of Variables on the Process

In order to study the effects of each variable and the interactions or duplicate effects of variables on the response generated by the model, the graphs were based on the polynomial model of the model, using the test design software. According to Equation (3), the initial concentration of antibiotics has the most significant effect on the removal process, with a coefficient equal to 9.94 and the reaction time smallest effect than other parameters with a coefficient of 3.71. The effect of independent variables on the efficacy of antibiotic removal is shown in Figures 7–9. Figure 7 shows the effect of the initial concentration of antibiotic and the pH of the solution. As shown from Figure 7, with the increase of antibiotic concentration, the removal efficiency decreases. In particular, with an increase of the antibiotic composition from 16.25 to 25 mg/L, the removal efficiency is reduced from 76.13 to 57.34%, respectively. The ideal efficiency was found to be at pH 6, while at pH > 6 and/or pH < 6, the efficiency is reduced.

**Figure 7.** The simultaneous effect of two variables; initial concentration of antibiotic and pH of solution; adsorbent dose of 2 mg/L and reaction time of 37.5 min.

**Figure 8.** The simultaneous effect of two primary antibiotic and adsorbent dose variables: pH = 7 and reaction time of 37.5 min.

**Figure 9.** The simultaneous effect of the two initial variables of antibiotic concentration and reaction time: pH = 7 and the adsorbent dose is 2 mg/L.

The effect of the initial concentration of antibiotics, and the amount of adsorbent, are presented in Figure 8. By increasing the amount of adsorbent effluent, the efficiency increases; for example, when the antibiotic concentration is used at a minimum level and the adsorbent content is 5.5 mg/L, the removal efficiency is approximately 76%. If the concentration is constant and the adsorbent amount is equal to 2 mg/L, the removal efficiency is 95.9%.

The effect of initial CIP concentration and the reaction time are shown in Figure 9. According to the graph, when the antibiotic concentration reaches maximum, and in after 26.5 min, 70.11% of the antibiotic is removed. But when the response time reaches the 2+ level (equal to 48.75 min), removal efficiency is increased to 98.77%.

#### **4. Discussion**

The contact time is an important factor that directly influences the whole process. In the present work, for a concentration of 5 mg/L, the adsorption process reaches equilibrium at about 37 min, and then shows a relatively stable trend. The effect of the pollutant's initial concentration is affecting a lot the adsorption process. In this paper, the pollutant's initial concentration was studied, ranging from 5 to 50 mg/L. As shown in Figure 7, the initial CIP concentration had a negative effect on the elimination efficiency, and by increasing the ciprofloxacin concentration from 16.25 to 25 mg/L, the elimination efficiency decreased from 84 to 57%. The decrease in removal efficiency when increasing initial concentration can be explained by the fact that the active sites are constant with a constant amount of adsorbent dose, but as the concentration of the adsorbent increases, the pollutant molecules (in the medium—water) saturate the available adsorption sites, thereby, the removal efficiency is lowered [49]. Bajpai et al. observed that by increasing the initial concentration of ciprofloxacin from 10 to 20 mg/L, the adsorption capacity increased from 3.74 to 11.32 mg/g [50].

#### *4.1. E*ff*ect of pH Solution*

In the purification processes, including adsorption, pH plays an important role. The Solution's pH can affect the adsorbent's surface load, the degree of ionization of various pollutants, the separation of functional groups on active adsorbent sites, as well as the structure of the antibiotic molecule; in effect, the solution's pH affects the chemical environment of the aqueous and adsorption surface bonds. The pH changes were applied to the range of 4–10, and its effect on the removal efficiency was then analyzed. The removal process had the highest percentage at pH 6.2–7, while with the increase of pH, the removal efficiency decreased.

The effect of pH on the ciprofloxacin molecule has shown that in pH less than 6.2, the surface of the molecule appears cationic and positive due to the protonation of amino groups. At pH values higher than 8.6, the ciprofloxacin molecule is converted into anionic form, due to the loss of the proton from the carboxylic group in the antibiotic structure. In the range of 6.2 to 8.6, the deprotonation of carboxyl groups leads to negative carboxylate production. However, the amino group of proteins has a positive charge. In other words, it has a positive and a negative "head". The stabilization and behavior of ciprofloxacin molecule from 6.2 to 7.8 have also been investigated [51]. Since the pH value at pHpzc at the isoelectric absorption point is 7.5, and is negatively charged at higher pH values, given that at pH values above 7.5, both the adsorbent and the antibiotic molecule are both negatively charged. At a pH of less than 6.2, the adsorbent and the antibiotic have positive charge, so in this range, the adsorption process occurs slower and reaches at minimum removal rate at pH = 6.2-6.8, because the unnamed bands reach the maximum electrostatic gravity.

#### *4.2. E*ff*ect of Adsorbent's Dose*

Based on the findings of this study, the adsorbent dose was the most important factor affecting the efficiency of ciprofloxacin elimination. The study of the effect of adsorbent mass on adsorption processes is one of the most important issues to be considered. Adsorption dose was applied to the range of 1 to 3 mg/L, and its effect on the effectiveness of ciprofloxacin antibiotic removal was measured. Depending on the results obtained using constant concentrations of antibiotics, the increase in the dose of the adsorbent improves the removal efficiency. As shown in Figure 8, when the concentration of antibiotic is constant and equal to 16.25 mg/L, and the amount of adsorbent is 1.5 mg/L, the removal efficiency is 75.97%—and when the amount of adsorbent reaches 2 mg/L, the removal efficiency is improved, reaching 95.91%; at a constant concentration of antibiotic, by increasing the dose of adsorbent, the ratio of active sites on the adsorbent's surface is high relative to the adsorbing molecules (pollutants), resulting in increased elimination efficiency. On the contrary, in low adsorbent amounts, the ratio of active sites to the adsorbent molecules is lower, and the adsorption decreases.

On the other hand, with the increase of adsorbent above the optimal amount, the adsorption capacity decreased below the maximum level of 15.25 mg/g, which is also due to the fact that by increasing the adsorbent dose, the total capacity of the active sites present in the adsorbent level is completely covered. If not, its adsorption capacity is reduced. This can be the use of available surface in the form of unsaturated attributed adsorbent. The results show that the adsorption pattern in the non-saturable adsorbent form causes undesirable use of existing spaces; this issue is very important in the design of the process economics, particularly in scaling-up.

In this study, 5 mg/L of antibiotic and 2 mg/L of adsorbent were introduced as the optimum amount, at maximum efficiency, with application of 2 mg/L of adsorbent, despite the increase in adsorbent content, other increase in cleavage removal efficiency has not shown any increase. In other words, the removal rate remains constant. It can be concluded that this amount of adsorbent adsorbs all the antibiotics in the solution. Therefore, the antibiotic concentration in the solution is so low that it is no longer "able to be adsorbed" easily. A study by Peasant et al. also showed that with the increase in the adsorbent dose (chitosan/zeolite composite), the dye removal increases, due to the increasing number of adsorption sites, while the increase of adsorbent's dose reduces the adsorption capacity (from the maximum of 17.77 mg/g) [52].

#### *4.3. E*ff*ect of Contact Time*

An important issue when using the adsorption system is providing an effective contact time under specific conditions. In this paper, contact time was applied to the range of 15 to 60 min, and its effect on the ciprofloxacin antibiotic removal. Figure 7 shows that the adsorption process reaches equilibrium at different times. For a concentration of 5 mg/L, the adsorption process reaches equilibrium at about 37 min, and then shows a relatively stable trend. By increasing contact time, the probability of colliding with adsorbent molecules is also increased, and the efficiency of removal increased. Chang et al. (2012) obtained the equilibrium time for tetracycline removal by Monte Myrnolite for 8 h [53]. In another study by Liu et al. who removed tetracycline using zeolite= by increasing contact time, resulted in the removal efficiency also increased, and the time of equilibrium was 120 min [54].

#### *4.4. Kinetics and Adsorption Isotherms*

The adsorption kinetics depends on the adsorbent chemical and physical properties, which influence the adsorption mechanism. In this study, we have used different kinetic and isotherm adsorption models such as pseudo-first order, pseudo-second order, Langmuir, and Freundlich (Table 4).

The pseudo-first and pseudo-second order kinetic equations are shown in Figure 10. Adsorption kinetics were used to determine the control mechanism of adsorption processes. Thus, in this figure, the experimental points were not shown, and only theoretical ones are presented. Based on Table 5, the best fitting was achieved with pseudo-second order equation (R<sup>2</sup> = 0.9984).


**Table 4.** Equations used in this study.

1

2.303

k

et e

log q q log q t

**Figure 10.** (Up triangle): pseudo-first order kinetic equation; (down triangle) pseudo-second order kinetic equation.

**Table 5.** Parameters and related kinetic coefficients.


− −

The isotherm of adsorption describes how the adsorbent and adsorbate interact. In this study, the experimental results were fitted to Freundlich and Langmuir isotherms. The Langmuir model is valid for single-layer adsorption on adsorbent surface, with limited and uniform adsorption locations, while the Freundlich isotherm is based on single-layer adsorption on heterogeneous adsorption sites with unequal and non-uniform energies. Figure 11 shows the relative isotherms.

**Figure 11.** Equilibrium results fitted to (**a**) Langmuir isotherm and (**b**) Freundlich isotherm.

In Freundlich isotherm, when K<sup>F</sup> increases, the adsorbent material adsorbed higher amounts of pollutant, and the value of n between 1 and 10 reflects the proper adsorption process. The parameters and coefficients are briefly summarized in Table 5. In this study, the calculated K<sup>F</sup> value is 4.75, and the value of n is 2.79, which is within the specified range. Therefore, the adsorption of ciprofloxacin on the adsorbent is well fitted to Langmuir model (Table 6), but it is fact that the data may suggest the presence of non-specific or multi-type interactions between the adsorbate molecules and the adsorptive sites.

**Table 6.** Parameters and correlation coefficients of isotherm models.


A major concern regarding any synthesized adsorbent material is answering why this material was synthesized instead of another structure-type material? To respond, it is of fundamental importance to mention some facts. Nanoparticles have a unique combination of properties, such as small size, large surface area, catalytic potential, large number of active sites, high chemical reactivity; all of the above give nanoparticles high adsorption capacity [59]. Also, magnetic nanoadsorbents can be applied as cost-saving and effective materials to separate the materials (solid) from the liquid-phase (water) after the end of the adsorption process. Moreover, the relatively simple isolation of magnetic materials from the solution can aid to their regeneration and reuse [60]. Therefore, the magnetic nanoadsorbents can be good candidates for water/wastewater treatment. Based on the above, Poly(vinylimidazole-co-divinylbenzene) magnetic nanoparticles have been used for the adsorption of fluoroquinolones from aqueous environments [61]. Wang et al. also synthesized the easy to separate magnetic chalcogenide composite KMS-1/L-Cystein/Fe3O<sup>4</sup> using L-cystein to connect KMS-1 and Fe3O<sup>4</sup> nanoparticles for ciprofloxacin removal from aqueous solutions [62]. Table 7 shows a brief comparison of some other adsorbent materials tested for the removal of CIP. However, similar experimental conditions should be kept in order to compare two adsorbents (even for the treatment of the same pollutant). Parameters affecting adsorption are the contact time, the solutions' pH, the initial concentration of the pollutant, temperature, adsorbate volume, agitation speed, the solution's ionic strength, and adsorbent dosage. Any change to the abovementioned conditions will lead to different results, and the comparison can be made for adsorbent/adsorbate systems of the same study. Also, based on the interaction groups, a possible mechanism of adsorption is illustrated in Figure 12.


**Table 7.** CIP adsorption capacities comparison from aqueous solutions using various adsorbents.

**Figure 12.** Possible adsorption interaction.

#### *4.5. Aspects*

It is known that activated carbon is a very popular adsorbent material, with the demand for virgin activated carbon expanding, since demand from water and wastewater treatment facilities has been steadily increasing. Together with the increase of wastewater treatment applications, the demand and production of activated carbon is also increasing. The largest quantities of activated carbon consumption are observed the U.S.A, Japan and then Europe [72]. Antibiotics are being detected in the aquatic environment. There are different ways for antibiotics to enter the aquatic environment with WWTP considered to be one of the main points of entrance. Even treated wastewater effluent can contain antibiotics, since WWTP cannot eliminate the presence of antibiotics. Compared to other tertiary treatments, adsorption can be a sustainable option for antibiotic removal from wastewaters. Activated carbon is used in the pharmaceutical for the removal of unwanted compounds [72]. Activated carbon possesses a plethora of disadvantages [73], such as high capital cost, ineffectiveness and non-selectivity against vat/disperse dyes. Furthermore, saturated carbon regeneration is expensive and leads to adsorbent loss. Depending on the demand, cost, and the nature of the pollutant to be adsorbed, the adsorbents are either disposed or regenerated for future use. Used adsorbents are considered hazardous waste, causing environmental and societal problems in various countries [74]. Heat accumulation and toxic adsorbates desorption could create hazardous conditions. In addition, odor can be caused by the dumping of adsorbents.

Since regeneration costs can be quite high, the reduction of consumption costs is the key to sustainable and industrial benefits. Substantial studies regarding the activated carbon-based adsorption of pollutants onto have been conducted, but research on regeneration methodologies remains limited [75]. Adsorbent regeneration capability cost analysis is necessary for the economic and environmental assessment of the adsorption process. For the spent adsorbent stabilizing or proper disposal seem to be difficult. The regeneration process of adsorbents from the points of view of sustainability and the environmental involves recovering valuable adsorbates, while reducing the

need of virgin adsorbents, and this is extremely important. Studies on novel adsorbents, at full-scale adsorption systems, should be considered for potential industrial applications.

#### **5. Conclusions**

Antibiotics are still being detected in the effluents of WWTP, and adsorption seems to be a sustainable option for antibiotics removal from waters. Poly(St-*b*-AAc) diblock copolymers were prepared using the RAFT technique. This copolymer with acrylic acid group was adsorbed onto the surface of Fe3O<sup>4</sup> nanoparticles, through the interaction with hydroxyl groups on the Fe3O<sup>4</sup> nanoparticles' surface. A magnetic nanocomposite ranged in 30 nm was then prepared. The VSM analysis showed the saturation magnetization (26 emu/g for P(St-*b*-AAc)/Fe3O4). The removal process was performed using P(St-*b*-AAc)/Fe3O<sup>4</sup> to remove ciprofloxacin antibiotic from synthetic sewage. The effects of parameters such as initial concentration of antibiotic, pH, soluble dose and reaction time were studied. The primary concentration of antibiotics with the highest negative effect and adsorbent dose showed the most positive effect in the removal process. The results also indicated that 97.5% of antibiotics were removed under optimal conditions, which include an initial antibiotic concentration of 5 mg/L, pH 7, and an adsorbent dose of 2 mg/L for 37.5 min. The adsorption of CIP was better fitted to Langmuir isotherm (R<sup>2</sup> = 0.9995), while the kinetics were better fitted to second-order kinetic equation (R<sup>2</sup> = 0.9973). Future work should include multi-component pharmaceutical adsorption with continuous adsorption of wastewaters, taking into account adsorbent regeneration.

**Author Contributions:** Methodology, L.M., A.G., G.F. and R.K.; A.R. and G.Z.K. writing—original draft preparation and supervision. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*

### **Core**−**Shell Molecularly Imprinted Polymers on Magnetic Yeast for the Removal of Sulfamethoxazole from Water**

**Liang Qiu 1,2, Guilaine Jaria <sup>2</sup> , María Victoria Gil <sup>3</sup> , Jundong Feng <sup>1</sup> , Yaodong Dai <sup>1</sup> , Valdemar I. Esteves <sup>2</sup> , Marta Otero 4,\* and Vânia Calisto <sup>2</sup>**


Received: 28 May 2020; Accepted: 17 June 2020; Published: 20 June 2020

**Abstract:** In this work, magnetic yeast (MY) was produced through an in situ one-step method. Then, MY was used as the core and the antibiotic sulfamethoxazole (SMX) as the template to produce highly selective magnetic yeast-molecularly imprinted polymers (MY@MIPs). The physicochemical properties of MY@MIPs were assessed by Fourier-transform infrared spectroscopy (FT-IR), a vibrating sample magnetometer (VSM), X-ray diffraction (XRD), thermogravimetric analysis (TGA), specific surface area (*S*BET) determination, and scanning electron microscopy (SEM). Batch adsorption experiments were carried out to compare MY@MIPs with MY and MY@NIPs (magnetic yeast-molecularly imprinted polymers without template), with MY@MIPs showing a better performance in the removal of SMX from water. Adsorption of SMX onto MY@MIPs was described by the pseudo-second-order kinetic model and the Langmuir isotherm, with maximum adsorption capacities of 77 and 24 mg g−<sup>1</sup> from ultrapure and wastewater, respectively. Furthermore, MY@MIPs displayed a highly selective adsorption toward SMX in the presence of other pharmaceuticals, namely diclofenac (DCF) and carbamazepine (CBZ). Finally, regeneration experiments showed that SMX adsorption decreased 21 and 34% after the first and second regeneration cycles, respectively. This work demonstrates that MY@MIPs are promising sorbent materials for the selective removal of SMX from wastewater.

**Keywords:** antibiotics; emerging contaminants; pharmaceuticals; wastewater treatment; polymeric adsorbents; magnetization

#### **1. Introduction**

Antibiotics are intensively used as human and veterinary medicines for the treatment and prevention of infectious diseases [1]. Among them, sulfamethoxazole (SMX) is a sulfonamide bacteriostatic antibiotic that has been commonly used during the last 80 years to treat urinary tract infections due to its low cost and broad spectrum of activity to treat bacterial diseases [2,3]. However, the widespread and indiscriminate use of SMX, as of other antibiotics, constitutes a huge potential threat to human health and contaminates natural ecosystems by affecting aquatic and soil organisms [4,5]. Recently, SMX has been detected in effluents of sewage treatment plants (STP), and also in surface and groundwater [6,7]. Indeed, it is known that pharmaceuticals (including SMX) can reach the aquatic environment in their unchanged or transformed forms mainly through discharge of effluents from municipal STP [7]. According to the statistics, more than 20,000 tons of SMX enter the environment worldwide every year, resulting in concentrations that range from 0.001 to 5.0 µg L−<sup>1</sup> in untreated or treated wastewater [8–10]. Therefore, the problem of environmental contamination by SMX is of great concern as pathogen resistance is highly documented and has been induced even by low levels of antibiotics [11].

To solve the above-mentioned problems, substantial research efforts have been directed worldwide to develop sustainable treatments for the removal of antibiotics, including SMX, from contaminated waters, such as membrane separation, adsorption processes, photocatalysis, and chemical oxidation [12]. Among these treatments, adsorption-based processes have been highlighted to be efficient, easy to implement and, furthermore, avoid the generation of transformation products [13–15]. However, the application of these processes is quite challenging due to the characteristic features of contaminated wastewaters, namely, large discharge flux, complex composition, and very low antibiotic concentrations [16]. Increasing the adsorbent specificity has been proposed as a strategy to address these challenges and improve the efficiency of the adsorptive removal of antibiotics from such complex matrices [17].

Molecularly imprinted technology (MIT) involves the creation of tailor-made selective binding sites in a polymeric matrix with memory of the shape, size, and functional groups of the template. Thus, molecularly imprinted polymers (MIPs) have become increasingly attractive as adsorbent materials due to their capacity to selectively bind specific targets and to their promising characteristics, such as low cost, easy synthesis, high stability to harsh chemical and physical conditions, and excellent reusability [18,19]. In recent years, MIPs, whose application of the extraction and analysis of organic contaminants in environmental water samples is well-established [20], have been successfully used for the adsorptive removal of pharmaceuticals, including antibiotics, from contaminated water [21–24]. In the specific case of SMX adsorption by MIPs, few works have been published, with most of them aiming at the analytic quantification of this antibiotic. For example, Qin et al. [5] used Fe3-O4-chitosan MIPs for SMX selective extraction and determination in aqueous samples, with the produced materials having attained a maximum adsorption capacity of 4.32 mg g−<sup>1</sup> . Zhao et al. [25] prepared core–shell MIPs on the surface of magnetic carbon nanotubes (MCNTs@MIP) for SMX, the resulting material having a maximum SMX adsorption capacity from aqueous solution of 864.9 µg g−<sup>1</sup> . However, to the best of our knowledge, the removal of SMX from complex wastewaters using MIPs has just been assessed by Valtech et al. [19]. Among the materials produced by these authors [18], those having the largest maximum adsorption capacity (6.5 × 10−<sup>5</sup> mol g−<sup>1</sup> (16.5 mg g−<sup>1</sup> )) performed similarly to a commercial activated carbon in terms of removal, but presented higher selectivity toward SMX in the presence of other pharmaceuticals and better regeneration ability.

Despite the above-mentioned advantages and applications, the preparation of MIPs by conventional MIT has two main drawbacks: (1) The imprinted polymer matrices are thick and, thus, hold a small number of recognition sites per unit volume; and (2) the template molecules are deeply embedded in the matrix, so there is a diffusion barrier for them, the mass transfer rate is low, and binding to the recognition sites is somehow hampered [26]. Surface molecular imprinting has been proved to improve mass transfer, recognition, and binding ability relative to MIT [27]. Among solid-support substrates used for the surface molecular imprinting process, microbial nano-magnetic materials are alternative supporters that have many advantages compared to inorganic materials [28]: (1) They are easy to obtain and short generations can be artificially cultured [29]; (2) there are many surface chemical functional groups and so modification steps can be avoided, reducing secondary pollution; (3) cells can guide the regulation of the growth process of inorganic materials [30]; (4) microbial cells have a variety of structures and can provide a rich array of templates for nanomaterials by template-assisted synthesis; and (5) magnetic properties allow for a simple after-use separation of the materials.

Yeasts, which belong to the fungus kingdom, are relatively large eukaryotic and single-celled microorganisms (diameters typically measuring 2.0–4.0 µm). Their cell wall includes glucan, mannan, chitin protein, and a small amount of lipids, and it has many surface chemical groups such as carboxyl (–COOH), carbonyl (–C=O), amino (–NH2), hydroxyl (–OH), and phosphoryl (–P=O) groups. Moreover, yeast is very cheap, easy to obtain, and environmentally friendly. These advantages make yeasts appropriate and widely used as supports for bio-nanocomposites [31].

In the above-described context, the objectives of this study were to: (1) Prepare a bio-nanocomposite of yeast-Fe3O<sup>4</sup> (magnetic yeast, MY) using an in situ one-step preparation of nano-Fe3O4; (2) use MY as the core to synthesize magnetic yeast-molecularly imprinted polymers (MY@MIPs) by a surface-imprinted polymerization method with MIPs as the shell and SMX as the template molecule; (3) characterize the resulting materials by Fourier-transform infrared spectroscopy (FT-IR), a vibrating sample magnetometer (VSM), X-ray diffraction (XRD), thermogravimetric analysis (TGA), specific surface area (*S*BET) determination, and scanning electron microscopy (SEM); (4) test the removal performance of MY@MIPs toward SMX and compare it with those of MY and MY@NIPs (magnetic molecularly imprinted polymers without template); and (5) explore the selective sorption capacity of MY@MIP in a real complex matrix (wastewater collected at a STP) and in the presence of other pharmaceuticals (diclofenac and carbamazepine).

#### **2. Materials and Methods**

#### *2.1. Chemicals and Materials*

Yeast cells (CICC 30225) were obtained from the China Center of Industrial Culture Collection (CICC). Iron salts used to produce MY were ferric chloride hexahydrate (FeCl3·6H2O) and ferrous chloride tetrahydrate (FeCl2·4H2O), purchased from Sigma-Aldrich (Stenheim, Germany). In addition, 2-vinyl pyridine (2-vpy), ethylene glycol dimethacrylate (EGDMA), acetonitrile (ACN), and azo-bis-isobutyronitrile (AIBN), which were also purchased from Sigma-Aldrich (Stenheim, Germany), were used for MIT. Other reagents used in this work included ammonium hydroxide, toluene (99.8%, Aldrich), ethanol (99.9%, Riedel-de Haën), methanol (99.99%, Fischer Chemical), and acetic acid (p.a., Merck). Ultrapure water was obtained from a Milli-Q water purification system (Millipore). SMX was purchased from TCI Europe (>98%); carbamazepine (CBZ; Sigma-Aldrich, 99%); diclofenac (DCF, TCI Europe, >98%). All solutions were stored at 4 ◦C immediately after preparation.

#### *2.2. Materials Preparation*

#### 2.2.1. Preparation of Magnetic Yeast (MY)

Nano-Fe3O<sup>4</sup> was loaded onto the yeast cell surface by a one-step method as described by Tian et al. [32]. Briefly, the yeast cells were cultured in ultrapure water with glucose. After reaching the exponential growth phase (6–10 h), the yeast cells were collected by centrifugation (4000 rpm). Then, collected cells (1.0 g) were suspended in 40 mL of 0.125 M FeCl<sup>3</sup> solution in a three-necked flask and stirred for 1 h at room temperature. After that, 0.6 g of FeCl2·4H2O was added under nitrogen atmosphere and stirred for another 1 h. The mixture was then heated in a water bath at 80 ◦C for 15 min, and the pH was adjusted to approximately 11 with 25% (*w*/*v*) ammonium hydroxide. Stirring was kept for 30 min and then stopped to age for 1 h. The resulting magnetic yeast (MY) was then washed, separated by applying a magnetic field, and then dried in an oven (35 ◦C, 4 h).

#### 2.2.2. Preparation of Magnetic Yeast-Based Molecularly Imprinted Polymer (MY@MIPs)

MY was treated as the core and the MIPs as the shell. The process used for the production of MY@MIPs was as follows: 1 mg of SMX (template molecule) and 4 mmol of 2-vpy (monomer) were dissolved in 60 mL of ACN/toluene (3/1; *v*/*v*). This solution was then self-polymerized for 8 h at room temperature (25 ◦C). Subsequently, 100 mg of MY (polymer supporter), 0.36 mmol of AIBN (initiator), and EGDMA (crosslinker) were added into the polymerized solution (template:monomer:crosslinker, 1:4:20), which was ultrasonicated for 10 min. The mixture was heated and maintained at 60 ◦C for 24 h under stirring with nitrogen protection. At last, the MY@MIPs were washed with methanol/formic acid (9/1; *v*/*v*) for 12 h and purified for 24 h by a Soxhlet extraction method (the extraction solution was methanol). Meanwhile, the MY@NIPs were also produced by following the above-described procedure but in the absence of the template.

#### *2.3. Characterization of MY, MY@MIPs, and MY@NIPs*

Fourier-transform infrared spectra of the produced materials were obtained in a Shimadzu-IRaffinity-1 equipment, using an ATR module (FTIR-ATR), under a nitrogen purge. The measurements were recorded in the range 500–4000 cm−<sup>1</sup> , 4.0 of resolution, 256 scans, and applying atmosphere and background correction.

A vibrating sample magnetometer (VSM EV9) with an oscillatory applied magnetic field (*H*) to a maximum of 22 kOe was used to determine the saturation magnetization (*M*S). The *M*<sup>S</sup> was calculated by plotting the magnetic moment versus the applied magnetic field, and it corresponded to the plateau value of the magnetic moment reached divided by the sample mass (10 mg). The sample was encapsulated in an acrylic cylindrical container (5.85 mm of diameter and 2.60 mm of height), which was coupled to the lineal motor of the VSM EV9 instrument, centered between the two polar heads of the electromagnet used to fluctuate the magnetic field. The instrument was calibrated with a disk of pure nickel (8 mm of diameter) using a procedure that establishes the determination of the magnetic field, applied at around 1 Oe, while the dispersion of the magnetic moment is inferior to 0.5%.

X-ray diffraction (XRD, 5–90◦ ) was measured on a D8-Focus X-ray diffractometer (Bruker Optics) with a test rate of 10◦ ·min-1. The results were analyzed by Jade program (9.0) and Origin (9.0).

Thermogravimetric analysis (TGA) was performed in a thermogravimetric balance Setsys Evolution 1750, Setaram, TGA mode (S type sensor). The samples were heated at a heating rate of 10 ◦C min−<sup>1</sup> , under nitrogen atmosphere, from room temperature to 105 ◦C and from 105 ◦C to 900 ◦C, maintaining constant temperature until total stabilization of the sample mass at the end of both stages (approximately 30 min).

The *S*BET and micropore volume (*W*0) were determined by nitrogen adsorption isotherms, acquired at 77 K using a Micromeritics Instrument, Gemini VII 2380, after outgassing the materials overnight at 120 ◦C. *S*BET was calculated from the Brunauer–Emmett–Teller equation in the relative pressure range 0.01–0.1. Pore volume (*V*p) was estimated from the amount of nitrogen adsorbed at a relative pressure of 0.99.

The surface morphology of the materials was analyzed by scanning electron microscopy (SEM) using a Hitachi S4100. The images were obtained at magnifications of 500, 3000, and 10,000×.

#### *2.4. Adsorptive Removal of SMX by the Produced Materials*

The produced materials (MY, MY@MIPs, MY@NIPs) were used as adsorbents for the removal of SMX under batch operation conditions. Summarizing, the materials were put in contact with a 5 mg L−<sup>1</sup> SMX solution in polypropylene tubes, which were shaken in a head-over-head shaker (80 rpm) for a predetermined period of time at controlled temperature (32 ◦C). The corresponding adsorbent material was separated from the suspension liquid by an external magnetic field. At last, the concentration of SMX in the liquid phase was measured by micellar electrokinetic chromatography (MEKC), using a methodology adapted from Silva et al. (2019) [33]. The experiments were conducted in triplicate, and control experiments without adsorbent were run in parallel. The performance of the materials was evaluated by carrying out kinetic, equilibrium, pH, selectivity, and regeneration/reutilization studies, described in detail in the next subsections.

#### 2.4.1. Kinetic Adsorption Studies in Ultrapure Water

In the kinetic study, tubes containing 250 mg of adsorbent material (MY, MY@NIPs, or MY@MIPs), together with 10 mL of a 5 mg L−<sup>1</sup> SMX solution in ultrapure water, were incubated and shaken as described above. After shaking during defined periods of time (*t*, min), at intervals from 0 to 24 h, the materials were separated from the aqueous phase and the remaining SMX concentration in solution was measured by MEKC. At each time, the corresponding value of the adsorbed concentration (*q*t , mg·g −1 ) was determined as follows:

$$q\_t = \frac{\mathcal{C}\_0 - \mathcal{C}\_t}{\mathcal{C}\_m} \tag{1}$$

where *C*<sup>t</sup> (mg L−<sup>1</sup> ) is the residual SMX concentration at time *t*, *C*<sup>0</sup> is the initial SMX concentration (mg L−<sup>1</sup> ), and *C*<sup>m</sup> is the adsorbent dosage (mg·L −1 ).

When adsorption equilibrium was attained, the percentage of adsorption R (%) was determined as:

$$R\left(\%\right) = \frac{\mathcal{C}\_0 - \mathcal{C}\_\varepsilon}{\mathcal{C}\_0} \times 100\% \tag{2}$$

where *C*<sup>e</sup> (mg·L −1 ) is the residual SMX concentration at equilibrium.

#### 2.4.2. Equilibrium Adsorption Studies in Ultrapure Water

For the equilibrium studies, the corresponding adsorbent material (MY, MY@NIPs, or MY@MIPs), with doses ranging from 50 to 2000 mg L−<sup>1</sup> , was added to 10 mL of a 5 mg L−<sup>1</sup> solution of SMX in ultrapure water. Tubes with the mixtures were shaken for 16 h, which allowed equilibrium to be reached. The materials were recovered from the suspension by the application of a magnetic field and the residual concentration of SMX was determined by MEKC. Then, for the different doses of material, the adsorbed concentration at the equilibrium (*q*e, mg·g −1 ) was determined as follows:

$$q\_{\varepsilon} = \frac{\mathcal{C}\_0 - \mathcal{C}\_{\varepsilon}}{\mathcal{C}\_m} \tag{3}$$

where *C*<sup>e</sup> (mg L−<sup>1</sup> ) is the SMX concentration in the liquid phase at equilibrium.

#### *2.5. Adsorptive Performance of MY@MIPs*

From the results of the above-mentioned kinetic and equilibrium studies in ultra-pure water, the most efficient material for removal of SMX was MY@MIPs. Thus, in order to assess the practical application of this material, further studies were carried out on the adsorptive performance of MY@MIPs under different experimental conditions.

#### 2.5.1. Kinetic and Equilibrium Adsorption Studies in STP Effluent

The kinetic and equilibrium procedures described in Section 2.4. were carried out using MY@MIPs for the adsorptive removal of SMX from a real matrix, namely the effluent from a STP. In this case, 5 mg L−<sup>1</sup> solutions of SMX were prepared using a STP effluent instead of ultrapure water. The effluent was collected from an urban STP in Aveiro (Portugal) that is designed to serve 159,700 population equivalents. This STP consists of primary and biological treatment stages. For this work, water was collected at the outlet of the biological decanter, as this is the final treated effluent that is discharged from the STP into the aquatic environment. Immediately after collection, the effluent was filtered through 0.45 µm, 293 mm Supor® membrane disk filters (Gelman Sciences) and stored at 4 ◦C until use, which occurred within a maximum of 15 days. The collected effluent had a pH of 7.99, conductivity of 3.03 mS cm−<sup>1</sup> , and total organic carbon content of 21.5 mg L−<sup>1</sup> .

#### 2.5.2. pH Study

Adsorption studies on the effect of pH were carried out at 32 ◦C with the initial conditions of *C*<sup>0</sup> = 5 mg L−<sup>1</sup> in ultrapure water and *C*<sup>m</sup> = 300 mg L−<sup>1</sup> . Experiments were carried out at three different pHs, namely 4, 7, and 8 (pH was adjusted by adding HCl or NaOH, 1 M). After shaking during 16 h, MY@MIPs were separated from the liquid suspensions, the residual concentration of SMX was analyzed by MEKC, and the corresponding *q*<sup>e</sup> (mg g−<sup>1</sup> ) at each pH was determined using Equation (3).

#### 2.5.3. Selective Adsorption

To study the selective capacity of MY@MIPs toward SMX, diclofenac (DCF) and carbamazepine (CBZ) were used as competing species in the adsorption experiments. These pharmaceuticals were selected due to their high global frequency of occurrence in wastewater, surface water, and groundwater and their recalcitrant properties, with low removal rates after conventional STP treatments [34]. The concentration of DCF and CBZ in ultrapure water solution was the same as that of SMX (5 mg L−<sup>1</sup> ), the *C*<sup>m</sup> was 300 mg L−<sup>1</sup> , the incubation temperature was 32 ◦C, the pH was 4, and shaking was maintained during 16 h. Then, the residual concentration of SMX at equilibrium was analyzed and the corresponding *q*<sup>e</sup> (mg g−<sup>1</sup> ) was determined with Equation (3).

#### 2.5.4. Regeneration and Reutilization

In order to evaluate the adsorptive performance after regeneration, after SMX saturation in ultrapure water, MY@MIPs were regenerated and then tested for the adsorption of SMX in four subsequent cycles. For the regeneration, saturated MY@MIPs were washed by methanol/acetic acid (9/1, *v*/*v*) through Soxhlet extraction during 72 h. Then, the regenerated material was used in adsorption experiments as described in previous sections (shaking during 16 h at 32 ◦C with the initial conditions of *C*<sup>0</sup> = 5 mg L−<sup>1</sup> in ultrapure water and *C*<sup>m</sup> = 300 mg L−<sup>1</sup> ). The residual concentration of SMX at the equilibrium was analyzed and the corresponding R (%) was determined as for Equation (2).

#### **3. Results**

#### *3.1. Preparation of MY*

In this study, an in situ one-step method was carried out to load nano-Fe3O<sup>4</sup> particles on the surface of yeast, which was used as a biological solid support. Under alkaline conditions, Fe2<sup>+</sup> and Fe3<sup>+</sup> co-precipitated on the surface of yeast and then Fe(OH)<sup>2</sup> or Fe(OH)<sup>3</sup> was converted to nano-Fe3O<sup>4</sup> at 80 ◦C, according to the following chemical reactions:

$$\text{Fe}^{2+} + 2\text{OH}^- \rightarrow \text{Fe(OH)}\_2\downarrow$$

$$\text{Fe}^{3+} + 3\text{OH}^- \rightarrow \text{Fe(OH)}\_3\downarrow$$

$$\text{Fe(OH)}\_2 + 2\text{Fe(OH)}\_3 \xrightarrow{80} \text{Fe(Fe}\_2\text{O}\_4)\downarrow + 4\text{H}\_2\text{O}$$

The observation of MY by a high-power optical microscope (Olympus CX22, Japan) clearly showed the loading of magnetic nanoparticles over yeast cells, as shown in Figure 1a. Meanwhile, Figure 1b represents the picture of MY at the actual size. Compared to other methods used to anchor Fe3O<sup>4</sup> nanoparticles on the surface of yeast biomass, such as cross-linking or electrostatic-interaction-driven hypercoagulation, the one-step method applied here only took approximately 3.5 h, in opposition to the referred methods, which can take up to 13.5 and 6 h, respectively [35], without considering the time of washing and drying. Hence, the results suggest that the one-step method is an interesting synthesis option.

(**a**) (**b**)

**Figure 1.** Optical microscopy photograph (500×) of magnetic yeast (MY) (**a**); actual size photograph of MY (**b**).

#### *3.2. Preparation of MY@MIPs*

MY was used as support of a MIP-based material for the selective adsorption of SMX. Compared to other support materials such as SiO2, carbon nanotubes, or Fe3O4-SiO<sup>2</sup> used in the literature [36,37], MY particles can act as support without the need of an intermediate chemical modification step and can distinctly improve grafting efficiency.

The shell of MIPs was co-polymerized on the surface of MY. Hence, to synthetize MIPs with affinity, selectivity, and appreciable removal capacity toward the target compound, the monomer and crosslinker types and the ratio of the reagents should be taken into account. Normally, if the template molecule has an alkaline chemical group, the monomer should be methacrylate (MAA), but if it has an acidic group, the monomer should be vinyl pyrimidine (vpy) [38]. As SMX has an oxazole moiety that displays acidity, 2-vpy was chosen as it has both a hydrogen-bond acceptor (N atom of pyridine) and alkalinity [39]. In this work, the molar ratio of the mixture of template and monomer was 1:4, as the monomer and template were in dynamic equilibrium, and it is not useful to add the monomer indiscriminately. Indeed, an excessive monomer may increase the non-selective sites, resulting in a selectivity decrease. On the other hand, during the synthesis of MIPs, in order to immobilize the template into the polymer without changing the spatial configuration of pores in the polymer, this must have a high rigidity. Therefore, it was necessary to use a crosslinker for increasing rigidity, EDGMA being selected due to its appropriate cost and solubility. However, if the ratio of monomer to crosslinker is too high, it will make the extraction of the template difficult due to the excessive rigidity of the MIP. Considering the referred considerations and conclusions from other studies [40,41], the ratio of monomer and crosslinker was selected to be 1:5.

#### *3.3. Characterization of MY, MY@MIPs, and MY@NIPs*

− − − − − − FTIR spectra of the produced materials (MY, MY@MIPs, and MY@NIPs), which were obtained in order to shed some light about the chemical groups present on their surface, are depicted in Figure 2. At 548 cm−<sup>1</sup> , a characteristic adsorption peak belonging to the Fe-O chemical bond was observed for all materials. Compared to MY, MY@MIPs and MY@NIPs had some new peaks. Among them were the absorption bands at 2363 and 2328 cm−<sup>1</sup> (MY@MIPs) and at 2377 and 2337 cm−<sup>1</sup> (MY@NIPs), which were attributed to the stretching vibrations of -CN or -NC, respectively. The peak at 1758 cm−<sup>1</sup> (MY@MIPs) or at 1727 cm−<sup>1</sup> (MY@NIPs) belongs to the stretching vibration of C=O in the EGDMA ester group and the carboxyl group, suggesting that EGDMA worked on the surface. Moreover, MY@MIPs had new peaks at 1118 and 955 cm−<sup>1</sup> , which belonged to the symmetrical and asymmetric stretching vibration of C-O in EGDMA, respectively, and reflected that it had a cross-linking polymerization on the surface of MY@MIPs. In the spectra of MY@MIPs and MY@NIPs, adsorption peaks at 1350 or

1340 cm−<sup>1</sup> were due to N–H bending vibrations, while this peak was very weak in MY, indicating the N–H bond of 2-vpy. Peaks at 1595 cm−<sup>1</sup> (MY@MIPs) or 1572 cm−<sup>1</sup> (MY@NIPs) were due to bending vibrations of N-H. − − −

**Figure 2.** FTIR spectra of magnetic yeast (MY), magnetic yeast-molecularly imprinted polymers without template (MY@NIPs), and magnetic yeast-molecularly imprinted polymers (MY@MIPs).

− The magnetic properties of the MY, MY@MIPs, and MY@NIPs were studied by VSM at room temperature, the *M*<sup>S</sup> of each material being shown in Table 1. The *M*<sup>S</sup> values were determined to be between 26 and 34 emu g−<sup>1</sup> , which were compatible with good magnetization. Indeed, Figure S1, within Supplementary Information, shows that MY@MIPs can be easily separated by an external magnetic field, which is beneficial for the after-use separation of the saturated MY@MIPs from treated water, achieving one of the major goals of this study.


**Table 1.** Physical characterization of the produced materials.

N<sup>2</sup> adsorption at −196 ◦C; *V*<sup>p</sup> = total pore volume; *D* = average pore diameter; *M*<sup>S</sup> = saturation magnetization.

− θ θ − − The XRD spectrum of MY in the 2θ range of 20 to 80◦ is shown in Figure 3, where the (220), (311), (400), (422), (511), and (440) planes of Fe3O<sup>4</sup> may be observed at 2θ = 30.22◦ , 35.40◦ , 43.36◦ , 53.68◦ , 57.21◦ , and 62.43◦ . This pattern is consistent with the standard XRD data of Fe3O<sup>4</sup> in the JCPDS-International Centre for Diffraction Data (JCPDS Card: PDF#75-0033). Therefore, XRD results evidenced that Fe3O<sup>4</sup> was successfully loaded onto the yeast surface during the production of MY and that, subsequently, surface molecular imprinting did not change the crystalline structure of magnetic nanoparticles. Similar patterns confirming the effective loading of magnetite have been reported in the literature on magnetic MIPs (MMIPs), including MMIPs produced for melamine analysis in milk [42], PEGylated magnetic core−shell structure-molecularly imprinted polymers (PMMIPs) for the specific adsorption of bovine serum albumin (BSA) [43], or core−shell MMIPs for the selective adsorption of tetracycline [44].

**Figure 3.** X-ray diffraction spectrum of MY.

The thermogravimetric (TG) and derivative thermogravimetric (DTG) curves of MY, MY@NIPs, and MY@MIPs are shown in Figure 4. All the materials evidenced three main weight loss peaks: The first at ~100 ◦C related to moisture; the second at ~300 ◦C related to the most thermolabile organic fraction; and the third centered at ~700 ◦C related to less thermolabile organic or inorganic fractions. For MY, a weight loss of approximately 62% was reached at 900 ◦C (Figure 4a); the second weight loss peak is particularly accentuated in this material, as it is the one with the highest amount of yeast per unit mass of material and, thus, yeast cells carbonized with increasing temperature. Meanwhile, MY@NIPs (Figure 4b) and MY@MIPs (Figure 4c) suffered, globally, a lower weight loss than MY, reaching 30% and 50% of weight loss, respectively, at 900 ◦C. This might be due to the introduction of less thermolabile structures in the composition of these materials (such as the magnetic nanoparticles and polymers).

**Figure 4.** TG (full line) and DTG (dashed lines) curves of MY (**a**), MY@NIPs (**b**), and MY@MIPs (**c**).

− − − − − − − − − − The results of *S*BET are shown in Table 1. The *S*BET of each material was as follows: MY—38.8 m<sup>2</sup> g −1 , MY@MIPs—43.2 m<sup>2</sup> g −1 , and MY@NIPs—39.2 m<sup>2</sup> g −1 . Similar *S*BET (47 m<sup>2</sup> g −1 ) were determined for magnetic sorbents with a metal–organic framework core and MIP shell [45]. Meanwhile, lower *S*BET, between 6 and 11 m<sup>2</sup> g −1 , have been measured for magnetic sorbents based on the iron oxide (Fe3O4) core and MIP shell [46,47]. Regarding the average pore diameter (*D*), it was 5.71, 5.41, and 5.06 nm respectively for MY, MY@MIPs, and MY@NIPs. Therefore, the three produced materials are mesoporous with no significant differences between them in terms of porosity.

The surface of MY, MY@MIPs, and MY@NIPs was examined by SEM (Figure 5). All the figures suggested that the particles (either MY, MY@NIPs, or MY@MIPs) were elliptical, which is due to the use of yeast as support, as it has been observed to have an ellipsoid shape with uniform size [48]. As it may be seen, MY@NIPs and MY@MIPs have a dispersed and comparatively smoother appearance than MY, which is rough-faced due to the magnetic nanoparticles coating the smooth-faced yeast [49]. Moreover, under 3000× magnification, results showed that MY@MIPs had a better dispersion compared to MY and MY@NIPs. Under 10,000×, MY@MIPs showed a bigger porosity than the other materials, which may benefit the adsorption of SMX and improve the mass transfer rate from the aqueous phase.

Globally, the characterization results demonstrated the successful loading of Fe3O<sup>4</sup> on the yeast surface and the preparation of MIPs on the surface of MY.

**Figure 5.** Scanning electron microscopy (SEM) images of MY (**a**–**c**), MY@NIPs (**d**–**f**), and MY@MIPs (**g**–**i**).

#### *3.4. Adsorptive Removal of SMX by the Produced Materials*

#### 3.4.1. Adsorption Kinetics

− The kinetic results on the adsorption of SMX onto the produced materials are shown in Figure 6, which evidences that, in all cases, the adsorbed concentration *q*<sup>t</sup> (mg g−<sup>1</sup> ) rapidly increased until 360 min of contact and then slowly increased until becoming stable. Moreover, all the materials performed quite similarly from a kinetic point of view.

− Comparing the results obtained here to those reported in the literature, it may be said that a shorter equilibrium time (around 20 min, at room temperature) was determined for the adsorption of SMX onto core−shell MIPs on the surface of magnetic carbon nanotubes (MCNTs@MIP) by Zhao et al. [25]. Meanwhile, using MIPs on the surface of yeast (yeast@MIPs), Wang et al. [16] found that (at 298 to 318 K) 200 min were necessary to attain equilibrium for the adsorption of ciprofloxacin (CIP), and Pan et al. [50] observed an equilibrium time around 375 min for the adsorption (at 303 K) of cephalexin. In any case, it has been noticed that surface-imprinting improves the binding kinetics as

compared to traditionally imprinted materials, which take longer (usually around 12–24 h) to attain adsorption equilibrium [51].

**Figure 6.** Experimental kinetic results together with pseudo-first- and pseudo-second-order model fittings for the adsorption of sulfamethoxazole (SMX) onto MY (**a**), MY@NIPs (**b**), and MY@MIPs (**c**) in ultrapure water.

Pseudo-first-order [52] and pseudo-second-order [53] kinetic models were applied to describe the adsorption kinetics of SMX onto the produced materials. The formulation of the models is as follows:

Pseudo-first-order

$$q\_t = q\_\ell \times \left(1 - e^{\left(-k\_1 t\right)}\right) \tag{4}$$

Pseudo-second-order

$$q\_t = \frac{k\_2 \times q\_\varepsilon^2 \times t}{1 + k\_2 \times q\_\varepsilon \times t} \tag{5}$$

q<sup>୲</sup> = k<sup>ଶ</sup> × qୣ <sup>ଶ</sup> × t 1 + k<sup>ଶ</sup> × qୣ × t where *k*<sup>1</sup> (min−<sup>1</sup> ) and *k*<sup>2</sup> (g mg−<sup>1</sup> min−<sup>1</sup> ) are the pseudo-first-order and the pseudo-second-order rate constants.

− − − The non-linear fitting kinetic parameters are summarized in Table 2. According to the correlation coefficient (R<sup>2</sup> ) and concordance between experimental and fitted *q*<sup>e</sup> values, both models described the SMX adsorption onto the produced materials, with the pseudo-second-order model describing slightly better the results onto MY and MY@NIPs and the pseudo-first-order model onto MY@MIPs.


**Table 2.** Kinetic parameters corresponding to the adsorption of SMX onto MY, MY@NIPs, and MY@MIPs in ultrapure water.

#### 3.4.2. Adsorption Isotherm

Equilibrium results on the adsorption of SMX onto the produced materials are shown in Figure 7. With the aim of describing these results, four isotherm models were used: Langmuir [54] and Freundlich [55] isotherm models for the adsorption of SMX onto MY@MIPs; BET isotherm [56] for the adsorption onto MY@NIPs; and Zhu−Gu isotherm [57] for the adsorption onto MY. The equations of these models are as follows:

Langmuir isotherm

$$q\_{\varepsilon} = \frac{q\_m \times b \times \mathbb{C}\_{\varepsilon}}{1 + b \times \mathbb{C}\_{\varepsilon}} \tag{6}$$

Freundlich isotherm

$$q\_{\varepsilon} = k\_f \times c\_{\varepsilon}^{\frac{1}{n}} \tag{7}$$

BET isotherm

$$q\_{\varepsilon} = \frac{q\_{\text{fl}} \times c \times \mathbb{C}\_{\varepsilon}}{(1 - c \times \mathbb{C}\_{\varepsilon}) \times (1 - c \times \mathbb{C}\_{\varepsilon} + c \times \mathbb{C}\_{\varepsilon})} \tag{8}$$

Zhu−Gu isotherm

$$q\_{\varepsilon} = \frac{q\_{\text{fl}} \times \left(\mathbf{g} \times \mathbb{C}\_{\varepsilon} \times \left(\frac{1}{r} + \varepsilon \times \mathbb{C}\_{\varepsilon}{}^{r-1}\right)\right)}{(1 + \mathbf{g} \times \mathbb{C}\_{\varepsilon}) \times \left(1 + \varepsilon \times \mathbb{C}\_{\varepsilon}{}^{r-1}\right)} \tag{9}$$

where *q*<sup>m</sup> is the maximum adsorption capacity (mg g−<sup>1</sup> ); *b* (L mg−<sup>1</sup> ) is the Langmuir equilibrium constant; *k*<sup>f</sup> (mg g−<sup>1</sup> (mg L−<sup>1</sup> ) −1/n ) is the Freundlich constant; *n* is the degree of non-linearity in the Freundlich isotherm; *c* is the BET constant, related to the energy of adsorption in the first adsorbed layer; *g* is the Zhu−Gu constant related to the first adsorption step (the first layer of molecules on the materials); *e* is the Zhu−Gu constant related to the subsequent layers adsorbed; and *r* is the aggregation number in the Zhu−Gu isotherm.

Experimental results on the equilibrium of SMX adsorption onto the produced materials are depicted in Figure 7 together with fittings to the above-mentioned isotherm models. From Figure 7, it is evident that, contrarily to the adsorption onto MY@MIPs, in the case of MY and MY@NIPs, the *q*<sup>e</sup> did not tend to stabilization. Furthermore, in the *C*<sup>e</sup> range between 0 and 3 mg L−<sup>1</sup> , a lower *q*<sup>e</sup> occurred for MY@NIPs than for MY. This may be related to the presence of chemical groups on the surface of MY, which were able to bind SMX groups, but became inaccessible in MY@NIPs due to molecular imprinting. In addition, a first stage with stabilization of *q*<sup>e</sup> at *C*<sup>e</sup> around 3 mg L−<sup>1</sup> may be observed in the MY isotherm, which could be associated with the saturation of the chemical adsorption sites. In the case of MY@MIPs, the isotherm showed an increase in *q*<sup>e</sup> with *C*<sup>e</sup> with a stabilization trend from *C*<sup>e</sup> ~ 3 mg L−<sup>1</sup> . Furthermore, it should be noted that, at relatively low *C*e, the *q*<sup>e</sup> values determined for MY@MIPs are higher than those for MY and MY@NIPs, which points to their larger affinity for SMX.

**Figure 7.** Experimental equilibrium results together with fittings to the considered models for the adsorption isotherm of SMX onto MY (**a**), MY@NIPs (**b**), and MY@MIPs (**c**) in ultrapure water.

− − The fitted equilibrium parameters are shown in Table 3 together with the correlation coefficients of the fittings (*R 2* ). In the case of SMX adsorption onto MY@MIPs, the Langmuir isotherm provided the best fitting of equilibrium results with a higher R<sup>2</sup> than the Freundlich model. From the results of non-linear fittings for the different isotherm models, the values of *q*<sup>m</sup> for MY, MY@NIPs, and MY@MIPs were, respectively, 23 ± 1, 3.8 ± 0.3, and 77 ± 3 mg g−<sup>1</sup> . These values indicate that molecularly imprinted polymers with the template resulted in a substantial increase in the monolayer adsorption capacity, SMX adsorbing onto the surface of MY@MIPs in a homogeneous distribution by occupying specific sites. Similarly, equilibrium results on the adsorption of SMX onto the MCNTs@MIP produced by Zhao et al. [25] also fitted the Langmuir isotherm, but with a considerably lower *q*<sup>m</sup> (0.87 mg g−<sup>1</sup> ). Indeed, compared to other materials used for the adsorption of SMX (Table 4), MY@MIPs are competitive in terms of SMX adsorption capacity.

−

− <sup>−</sup>


**Table 3.** Equilibrium parameters corresponding to the adsorption of SMX onto MY, MY@NIPs, and MY@MIPs from ultrapure water.

**Table 4.** Maximum Langmuir adsorption capacities (*q*m, mg g−<sup>1</sup> ) of different MIPs used for the adsorption of SMX.


CSMX = Initial concentration of SMX.

#### 3.4.3. Kinetic and Equilibrium Adsorption Studies from STP Effluent

In order to assess the practical applicability of MY@MIPs, kinetic and equilibrium experiments were carried out in a real matrix, namely the effluent from a STP. The obtained results together with fittings to the considered kinetic and equilibrium models are in shown in Figure 8, and the fitted parameters are depicted in Table 5. As it may be seen, the pseudo-second-order and the Langmuir isotherm models were those that best described the kinetic and equilibrium experimental results, respectively. On the other hand, it is evident in Figure 8 that, under identical experimental conditions, the adsorption velocity was slower and the *q*<sup>e</sup> values were lower for the STP effluent than they were for ultrapure water. This was confirmed by the parameters in Table 5, especially by the comparatively lower *q*<sup>m</sup> (24 ± 2 mg g−<sup>1</sup> ) than in ultrapure water, which might be related to interferences due to the complex composition of the STP effluent.

−

**−**

− − −

− −

−

**Figure 8.** Experimental results together with fittings to the considered models for the adsorption kinetics (**a**) and adsorption equilibrium isotherm (**b**) of SMX onto MY@MIPs in sewage treatment plant (STP) effluent.



#### *3.5. pH Study*

Results from the study of pH effects on the adsorption of SMX onto MY@MIPs are shown in Figure 9. Under identical experimental conditions, except for the pH, decreasing *q*<sup>e</sup> values were obtained at pH 4 > pH 7 > pH 9, thus indicating that SMX adsorption onto MY@MIPs was favored under acidic conditions. This may be related to the pH influence on the status of not only the adsorbate (by protonation/deprotonation) but also the adsorbent (by surface charge). For SMX, the p*K*<sup>a</sup> values are 1.97 and 6.16 (Table S1, as Supplementary Information), which means that SMX is mostly positively charged (protonated NH<sup>2</sup> groups, NH<sup>3</sup> <sup>+</sup> groups) at pH < 1.97 but predominantly negatively charged (deprotonated NH groups, N−) at pH > 6.16. Therefore, at the experimental pH 4, adsorption of SMX in the non-ionic form was favored, while at pH 7 and 9, SMX was mostly present in the anionic form, which partially hindered its adsorption. Indeed, the decrease in SMX adsorption from wastewater (with pH > 7) has already been related to electrostatic repulsion between the negatively charged SMX and the negatively charged surface of the waste-based adsorbents [59]. Moreover, the monomer used in the synthesis of MIPs was 2-vpa and the p*K*<sup>a</sup> of pyridine was 5.21, which is, therefore, negatively

charged when pH > 5.21. Thus, at the experimental pH 7, electrostatic repulsion forces between SMX and MY@MIPs cannot be disregarded, these increasing at pH 9.

−

− −

− − −

− − −

− −

**Figure 9.** Effect of pH on the percentage of SMX adsorption (R (%)) onto MY@MIPs.

#### *3.6. Selective Adsorption*

In order to find out the selectivity of MY@MIPs toward SMX, its adsorption was compared to those of DCF and CBZ from their single solution and then from their ternary solution. The values of percentage of adsorption (R (%)) for the single adsorption of each pharmaceutical are shown in Figure 10a and for adsorption from their ternary solution in Figure 10b.

**Figure 10.** Percentage of adsorption (R (%)) of SMX, diclofenac (DCF), and carbamazepine (CBZ) onto MY@MIPs from single solution (**a**) and ternary solution (**b**).

The results in Figure 10a evidence that MY@MIPs have a larger R (%) for SMX than for DCF or CBZ. Furthermore, under identical experimental conditions but from the ternary solution of the considered pharmaceuticals (Figure 10b), selective adsorption of SMX onto MY@MIPs occurred. Indeed, the R (%) determined for SMX from the ternary solution was just slightly lower than from its single solution, which points to the selectivity of MY@MIPs. Moreover, the results reflected that the adsorption of MY@MIPs was SMX > DCF > CBZ.

In this work, 2-vpy was the monomer and it was combined through –NH2. Moreover, considering the structure and properties of SMX, DCF, and CBZ, which are depicted in Table S1, they all have –NH<sup>2</sup> and/or –NH groups. However, SMX has two amino groups: -NH and –NH2, which is probably the main reason for its selective adsorption onto MY@MIPs under the presence of DCF and CBZ. For DCF and CBZ, the p*Ka* was 4 and 15.96 (Table S1), respectively. Meanwhile, the p*K*<sup>a</sup> value of 2-vpy (monomer) is 5.21, which may explain why MY@MIPs had a better removal ability for DCF than for CBZ. Selectivity toward SMX was also verified by Zhao et al. [25], who prepared MCNTs@MIP by using SMX as the template molecule and copolymerization of vinyl end groups on the surface of MCNTs [25]. These authors demonstrated the selective adsorption of SMX under the presence of other sulfonamides (SAs), namely sulfamethazine (SMZ), sulfamerazine (SMR), sulfadimethoxine (SDM), and sulfameter (SME). Still, the adsorbed concentration of SMX from the quinary solution was lower than from its single solution, which was ascribed to the close structure of the other SAs, which, therefore, could competitively occupy the imprinted sites.

#### *3.7. Regeneration and Reutilization*

Saturation of the produced MY@MIPs with SMX was carried out as described in Section 2.4.1. At that moment, R (%) calculated by Equation (2) was 92 ± 4%. Then, saturated MY@MIPs was regenerated as indicated in Section 2.5.4. and reused for the adsorption of SMX until saturation. A total of four regeneration/reutilization cycles were performed and the R (%) calculated for each of them are shown in Table 6.



As it may be seen, after cycles 1 and 2, the R (%) values decreased to 73 ± 3 and 61 ± 2%, respectively. Such decreases (21 and 34%, respectively, in cycle 1 and 2) indicate that the regeneration procedure affected the adsorption sites on the surface of MY@MIPs, which lost efficiency in the removal of SMX. After cycle 2, just a slight decrease in R (%) occurred, its value being similar in cycles 3 and 4 (58 ± 1 and 55 ± 2%, respectively). Therefore, deterioration of MY@MIPs adsorptive properties was not progressive with successive regenerations but occurred initially, with the performance remaining stable after cycle 2. MIPs sorbents are known to be easily regenerated by washing with organic solvents, with mixtures of methanol and acetic acid having been successfully employed to remove adsorbed pharmaceuticals [60]. Using the same regeneration agent as in this work, namely methanol/acetic acid (9/1, *v*/*v*), Dai et al. [61] regenerated MIPs synthesized for the adsorption of diclofenac and carried out thirty cycles with ≥95% recovery. Likewise, Duan et al. [62] also used this mixture for the regeneration of a multitemplate MIP, which was used in twenty regeneration/reutilization cycles, giving ≥95% removal of ibuprofen, naproxen, ketoprofen, diclofenac, and clofibric acid. Wang et al. [48], who used MIPs on the surface of yeast (yeast@MIPs), desorbed ciprofloxacin using the same mixture with losses of only about 8.5% of initial capacity after five cycles. Therefore, the relatively larger deterioration of the adsorptive performance observed in the present work may be related to the fact that magnetic yeast was used here as MIPs support. Thus, further work is to be carried out on the regeneration of the produced MY@MIPs, to maintain a high R (%) upon cyclic operation.

#### **4. Conclusions**

This work developed an efficient strategy to prepare yeast-Fe3O<sup>4</sup> (magnetic yeast, MY) and then used molecularly imprinted technology (MIT) to modify MY. The characterization of the produced magnetic yeast-molecularly imprinted polymers (MY@MIPs) showed that elliptical and monosized imprinted polymeric nanospheres with a surface area of about 43.2 m<sup>2</sup> g <sup>−</sup><sup>1</sup> were successfully produced. Sulfamethoxazole (SMX) adsorption studies using MY@MIPs indicated that the equilibrium was attained in 360 min either in ultrapure water or in a sewage treatment plant (STP) effluent. The Langmuir isotherm model provided the best fitting of equilibrium results and pointed to the monolayer and favorable adsorption of SMX onto MY@MIPs. In addition, the fitted parameters of the Langmuir isotherm model indicated that the maximum SMX adsorption capacity of MY@MIPs was 77 and 24 mg g−<sup>1</sup> in ultrapure water and STP effluent, respectively. The pH study pointed out that hydrogen binding was underneath the SMX adsorption onto MY@MIPs. Moreover, MY@MIPs showed successful selective adsorption of SMX from ternary solution under competition by other pharmaceuticals, namely diclofenac (DCF) and carbamazepine (CBZ). Finally, regeneration implied a reduction in SMX removal by MY@MIPs in the first two cycles, then tending to stabilization. Overall, it may be concluded that the MIPs-coated magnetic yeast designed here could be an alternative adsorbent for the selective removal of SMX from complex matrices such as wastewaters.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4360/12/6/1385/s1, Table S1: Physico-chemical properties of the pharmaceuticals used in this study (Source: Drugbank), Figure S1: MY@MIPs in the presence (**a**) and absence (**b**) of an external magnetic field.

**Author Contributions:** Conceptualization, J.F., Y.D., V.I.E., M.O. and V.C.; methodology, L.Q., J.F., Y.D., V.I.E., M.O. and V.C.; materials characterization, L.Q. and M.V.G.; experimental work, L.Q. and G.J.; data analysis, Q.L., V.I.E., M.O. and V.C.; writing, L.Q., M.O. and V.C.; supervision, J.F., Y.D. and V.I.E. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is a contribution to the research project WasteMAC (POCI-01-0145-FEDER-028598) funded by FCT – Fundação para a Ciência e a Tecnologia, I.P., through national funds, and the co-funding by the FEDER, within the PT2020 Partnership Agreement and Compete 2020. Thanks are due to FCT/ Ministério da Ciência, Tecnologia e Ensino Superior (MCTES), for the financial support to CESAM (UIDP/50017/2020+UIDB/50017/2020), through national funds. The work was also sponsored by the Natural Science Foundation of Jiangsu Province (SBK201404182); the environmental protection scientific research subject in Jiangsu province (Grant NO.2016003); and the A Project Fund by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD). Marta Otero and Vânia Calisto are thankful to FCT for the Investigator Program (IF/00314/2015) and for the Scientific Employment Stimulus (CEECIND/00007/2017), respectively. Guilaine Jaria thanks for her FCT PhD grant (SFRH/BD/138388/2018) supported by the National Funds and FSE, POCH (Programa Operacional Capital Humano), and the European Union. María V. Gil acknowledges support from a Ramón y Cajal grant (RYC-2017-21937) of the Spanish Government, co-financed by the European Social Fund (ESF).

**Conflicts of Interest:** The authors declare no conflict of interest. Furthermore, funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
