3.1. Characterization of beads
Chitosan received great interest due to its cationic groups, high adsorption capacity, molecular structure, biodegradability, abundance and low cost. However, due to the protonation of free amino groups, the polymer is fully soluble below pH ∼5 (Crini and Badot 2008), low mechanical strength and low surface area (Vakili et al. 2015), thus limiting its capability of sorbate binding (Filipkowska and Jozwiak 2013). The latter can be overcome by modifying chitosan to increase its mechanical strength and adsorption capacity. Lignin, cellulose and hemicellulose are integral parts of the plant cell and provide mechanical strength. Almond and walnut shells were a choice of composite materials as the chemical composition (Table 1) of the shells provides the required mechanical strength to the beads as adsorbents. The chemical composition of shells varies: Walnut shells had 36.38% cellulose, 27.85% hemicellulose and 43.7 % lignin; Almond shells had 38.47% cellulose, 28.82% hemicellulose and 29.54% lignin.
Table 1
Proportion of cellulose, hemicellulose, lignin
Material
|
Cellulose (wt. %)
|
Hemicellulose (wt. %)
|
Lignin (wt. %)
|
Walnut shells
|
36.38
|
27.85
|
43.70
|
Almond shells
|
38.47
|
28.82
|
29.54
|
Scanning electron micrographs present the surface morphology of synthesized beads before and after adsorption. Before adsorption, the AWC beads (Fig. 2a) showed an irregular appearance with different gradients of scattered fissures and occasional pores. After adsorption (Fig. 2b), the SEM image was layered with fair surface topography and resembled clouds of an overcast sky. In the case of CAW beads (Fig. 2c), the SEM image showed multiple layers with the formation of chunks of varying sizes, fissures and conspicuous pores. The post-processed image (Fig. 2d) showed that the structure formed has reduced in size and adhesion pattern, indicating an activity. The WAC beads' surface seems rough and multiplies layered, giving an uneven outlook (Fig. 2e). After adsorption, the post-processed SEM image (Fig. 2f) showed that agglomerates became more condensed and tightly packed. A dispersive energy spectrum (EDX) showed that the beads contain a maximum carbon component followed by oxygen (Ahmad and Dutta 2020).
The FTIR spectra for the beads were taken in the range 450–4000 cm− 1, which was performed to distinguish functional groups present on the outside of the adsorbent that can conceivably support the adsorption of anti-infection agents as shown; it can be shown in Fig. 3, N–H, OH and C = O groups were found on the prepared beads, which are considered as a decent decision for adsorption measure (Kumar et al. 2018).
The C–H bonds at 3700–3800 cm− 1 indicate the presence of alkenes or aromatic compounds, while the peaks at 3273, 3255 and 3288 cm− 1indicate hydrogen bonding (O–H) stretch, suggesting the presence of phenols and alcohols in AWC (Fig. 3a), CAW (Fig. 3b) and WAC (Fig. 3c) beads. The sharp peaks at 2359, 2362 and 2355 cm− 1 confirm the presence of the N–H group for amino groups. When response happens between aldehyde groups of glutaraldehyde and some amino groups of beads, the amine groups may be formed (Migneault et al. 2004). The peaks support the transformation at 1643, 1647 and 1649 cm− 1(C = O stretch). After adsorption, the pinnacles have gotten limited in the event of AWC (a) beads. The pinnacles got heightened after adsorption if there should be an occurrence of CAW (b) beads. There was a minimal extensive distinction between them when adsorption tops on WAC (c) beads. It was observed that the interaction was due to hydrogen bonding between hydroxyl of alkali lignin with β-1, 4- glycosidic linkage of chitosan and alkyl-substituted ether of alkali lignin with the hydroxyl group of chitosan. There was no significant change in the peaks of primary amine, confirming no involvement of primary amine of chitosan in composite formation (Nair et al. 2014).
3.2. Effect of pH and adsorption mechanism
The pH of the solution is an important parameter affecting the adsorption of any pollutant, as it can influence the physiochemical properties, surface binding sites and charge of both adsorbents (Ahsan et al. 2018; Yadav et al. 2018) and adsorbate (Seo et al. 2017). Therefore, the adsorption efficiencies of beads were studied at different pH values (3–11) with the fixed initial concentration of antibiotics (30 mg/L) and contact time of 180 min. At lower pH values, the amino groups of FQs get protonated (Amin et al. 2007), and at higher pH values, they behave as anions due to the deprotonation of carboxylic groups. However, at neutral pH, zwitterion exists (Yadav et al. 2018); hence, the study was spread over a range from pH 3–11.
The carboxylic groups present in the GAT molecules get ionized to yield – COO- groups while –NH groups of the piperazine ring get protonated to –NH2+ when the drug is in its solution form of pH 5–7. The adsorption of GAT onto the beads is due to the interaction between the above-charged species with the charges on the beads. The chitosan composite beads grafted with plant materials like almond and walnut added lignin, cellulose and hemicellulose adsorption are mainly due to electrostatic attraction between negative moiety (COO-) of GAT and NH3+ of composite, polar π–π stacking dipole-dipole interplays with the cellulose and lignin fractions with GAT at pH 5–7. Maximum adsorption of gatifloxacin at pH 5.0 on CAW (85%). A similar observation was reported in other studies, in which the amoxicillin adsorption increases from pH 2 to 5 as the carboxyl functional groups (–COOH) on the amoxicillin readily dissociate to carboxylate (–COO−) ions, thus increasing the electrostatic attraction between amoxicillin and the adsorbent (Putra et al. 2009; Moussavi et al. 2013). The adsorption of GAT was maximized at pH 7.0 on AWC (84%) and WAC (82%) beads, respectively (zwitterionic form), which may be due to the protonated amino groups that still can favour the adsorption (Yadav et al. 2018). In even higher pH values, the decrease in the adsorption was attributed to the deprotonation of C = O groups on antibiotics and beads, which significantly caused electrostatic repulsion between negatively charged groups of beads and antibiotics (Jiang et al. 2013). WAC beads followed a lower adsorption level onto its surface than the other two compositions (Fig. 4). The FTIR analysis shows that all three different beads contain –O.H., and C = O groups, which give negative charges due to the oxygen-containing groups (Yadav et al. 2018). Studies have reported that the maximum removal of different antibiotics found to be optimum at a pH range from 5.0–7.0, like norfloxacin (Feng et al. 2018) Tetracycline (Fan et al. 2018), sulfamethoxazole (Soares et al. 2019), levofloxacin (Mahmoud et al. 2020) and ciprofloxacin (Fu et al. 2021).
3.3. Effect of adsorbent’s dosage
The removal of antibiotics was improved with the increase in adsorbent dose. The uptake value increases almost linearly as the adsorbent amount rises from 0.1 to 1.0 g. The increase was from 81 to 88% by AWC beads, 80 to 88% by CAW beads and 81 to 88% by WAC beads when the adsorbent dosage increased from 0.1 to1.0 g (Fig. 5a). This behavior was attributed to the increasing number of active sites, which causes the attraction of more antibiotic molecules on the bead's surface (Kakavandi et al. 2014; Chen et al. 2014; Ahsan et al. 2018) When the adsorbent dosage increased from 0.1 to 1 g/L, the gatifloxacin adsorption percentage increased slightly. Therefore, it can be concluded that 0.1 g as the adsorbent’s dosage was adequate because only a slight increase was presented (~ 7% increase) for all beads. However, the removal value decreases non- significantly for AWC beads, which later made up to the maximum value as other combinations. An increase in adsorbent dosage causes adsorption of ciprofloxacin (Danalıoğlu et al. 2017; Liang et al. 2018), tetracycline and chlortetracycline (Ma et al. 2019). In another study (adsorption of cephalexin onto walnut shell-based activated carbon), it was found that the adsorption capacity decreased considerably with the increase of adsorbent dose (Nazari et al. 2016).
3.4. Effect of initial concentration
The rate of adsorption is a function of the initial concentration of the adsorbate, which is an essential factor for effective adsorption. The effect of initial concentrations of adsorbed antibiotics onto beads is presented in Fig. 6. The removal of antibiotics was found to decrease with the increase of antibiotic concentration. In the case of AWC beads, the adsorption decreased from 96 to 63% with the rise of initial GAT concentration, while the adsorption decreased from 96 to 65% and 94 to 62% in the case of CAW and WAC beads, respectively. This can be explained by considering that all adsorbents have a fixed/limited number of active adsorption sites. The active sites become saturated under the same experimental conditions (Balarak et al. 2016). The results obtained in this study are in agreement with a decrease in the adsorption percentage of antibiotics like cephalexin (Khosravi et al. 2018), norfloxacin (Feng et al. 2015), ciprofloxacin (Fu et al. 2021) and—chlortetracycline antibiotic (Tunc et al. 2020) with their increase in initial concentration respectively.
For the isotherm modelling, Langmuir (Eqs. (2, 4)) and Freundlich (Eq. (3, 5)) (Ahmed and Theydan 2014) isotherm equations were used for fitting adsorption equilibrium data (in non-linear and linear form), and the following equations illustrate them:
\(\frac{{{{\text{C}}_{\text{e}}}}}{{{{\text{Q}}_{\text{e}}}}}{\text{ = }}\left( {\frac{1}{{{{\text{Q}}_m}}}} \right){{\text{C}}_{\text{e}}}+\frac{1}{{{{\text{K}}_{\text{L}}}{{\text{Q}}_m}}}\) (Linear form of Langmuir equation) (2)
\({\text{log}}\left( {{{\text{Q}}_{\text{e}}}} \right){\text{ = }}\frac{1}{n}\log \left( {{{\text{C}}_{\text{e}}}} \right){\text{+log}}\left( {{{\text{K}}_{\text{F}}}} \right)\) (Linear form of Freundlich equation) (3)
\({{\text{Q}}_{\text{e}}}{\text{ = }}\frac{{{{\text{Q}}_{\text{m}}}{{\text{K}}_{\text{L}}}{{\text{C}}_{\text{e}}}}}{{1{\text{+}}{{\text{K}}_{\text{L}}}{{\text{C}}_{\text{e}}}}}\) (Non-linear form of Langmuir equation) (4)
where Qm (mg/g) is the maximum amount of adsorption; K.L. (L/mg) is the Langmuir constant of adsorption equilibrium; KF (mg1 − 1/n L1/n/g) is the Freundlich constant which symbolizes the capacity of adsorption; n is the constant that represents the intensity of adsorption (dimensionless).
All the values of constants for the isotherms were compiled in Table 2. The data were analyzed to see whether the isotherm obeyed Langmuir and Freundlich isotherm model. Adsorption was considered satisfactory when the Freundlich constant n values were between 1–10 (Deng et al. 2011). The adsorption of gatifloxacin on chitosan and almond/ walnut composite beads and its equilibrium data were fitted to the Langmuir and Freundlich isotherms. The R2 calculated by Langmuir was the highest in ACW beads than the Freundlich equation. Therefore, Langmuir was the best fit for isotherm for gatifloxacin adsorption on ACW beads which showed the surfaces of these materials had nearly homogenous sites for gatifloxacin adsorption. Furthermore, the maximum monolayer coverage capacity (Qm) calculated by the Langmuir isotherm was obtained (35.84 ± 0.49 mg/g), and KL 0.45 L/mg and R2value are 0.999 showing sorption fitted to the Langmuir model of gatifloxacin on ACW beads.
Table 2
Isothermal parameters for the adsorption of gatifloxacin antibiotic compound on AWC, WAC, CAW.
|
Langmuir equation
|
Freundlich equation
|
|
Qm
|
KL
|
R2
|
KF
|
n
|
R2
|
Material
|
mg/g
|
L/mg
|
|
mg1 − 1/n L1/n g− 1
|
|
|
ACW
|
35.84 ± 0.49
|
0.45 ± 0.02
|
0.999
|
12.63 ± 1.93
|
2.90 ± 0.55
|
0.949
|
CAW
|
35.81 ± 0.48
|
0.47 ± 0.03
|
0.996
|
14.69 ± 0.56
|
3.39 ± 0.19
|
0.995
|
WAC
|
34.22 ± 1.82
|
0.48 ± 0.10
|
0.983
|
13.03 ± 0.77
|
3.14 ± 0.25
|
0.991
|
3.5. Effect of contact time
The effect of contact time on the adsorption of GAT is a function of time, as shown in Fig. 7. Initially, the antibiotic removal rate is higher, but it becomes slow as it reaches equilibrium (plateau). Similar results were obtained when amoxicillin was adsorbed onto the almond shell (Homem et al. 2015). The graph depicts that the adsorption percentage reached 84%, 88%, and 88% in 180 minutes for AWC, CAW, and WAC beads. The high availability may explain the high adsorption rate at the initial stage in the number of active binding sites on the adsorbent surface, which gradually is occupied by the antibiotic molecules and tends to become almost constant. Similar results were obtained in the adsorption of tetracycline and sulfamethoxazole onto nanomagnetic walnut shell-rice husk (Popoola 2020). After a definite time, the empty sites are challenging to be covered due to the repulsive forces between solute molecules available in the solid and bulk phase. The adsorption is likely an attraction-controlled process in the later stage due to fewer available sorption sites (Azarpira and Balarak 2016). As the surface adsorption sites become exhausted, the uptake rate is controlled by the rate at which the adsorbate is transported from the exterior to the interior sites of the adsorbent particles. Gatifloxacin adsorption on all the combination beads shows maximum adsorption at 120 mins, followed by a plateau phase. This study is in agreement with the adsorption of tetracycline and chlortetracycline on chitosan (Liang et al. 2018), Ciprofloxacin and enrofloxacin antibiotics adsorption onto chitosan hydrogels (Wang et al. 2019), and chlortetracycline adsorption on chitin from aqueous solution (Tunc et al. 2020) and ciprofloxacin on guava leaves (Tay and Ong 2019).
Pseudo–first (Eq. (6)) and –second-order (Simonin 2016) (Eq. (7)) equations (linear forms) were selected to fit the experimental kinetic data.
where k1 (min− 1) and k2 (g mg− 1 min− 1) are the rate constants for the pseudo-first and –second-order kinetic equation, respectively.
Table 3 presents the kinetic parameters resulting from the fitting. Fig. 7a shows the plot of linearization of the pseudo-first-order model (Tay and Ong 2019), where the slope (–k1/2.303) and intercept log(Qe) of plot log(Qe–Qt) versus t were used to determine the pseudo-first-order constant k1 and the equilibrium adsorption density Qe, cal.
Table 3
Kinetic constants for the adsorption of GAT (C0=30 mg/L).
Table 3. Kinetic constants for the adsorption of GAT (C0 = 30 mg/L).
|
|
|
Pseudo–first-order model
|
Pseudo–second-order model
|
|
Qe,exp
|
k1
|
Qe,cal
|
R2
|
k2
|
Qe,cal
|
R2
|
Adsorbent
|
(mg/g)
|
(min− 1)
|
(mg/g)
|
|
(g mg− 1 min− 1)
|
(mg/g)
|
|
AWC
|
12.68
|
0.0298
|
12.96
|
0.952
|
7.8×10− 5
|
38.62
|
0.886
|
CAW
|
12.93
|
0.0269
|
10.95
|
0.964
|
8.2×10− 5
|
37.31
|
0.874
|
WAC
|
12.95
|
0.0296
|
13.00
|
0.971
|
5.3×10− 5
|
44.92
|
0.818
|
There is a good agreement of experimental data with this model. The obtained correlation coefficients (R2) were in the range of 0.952–0.971. Also, the adsorption equilibrium values (Qe, cal) follow the experimental values. In the case of AWC, the difference was 0.28 mg/g (Qe,cal= 12.96 mg/g and Qe,exp= 12.68 mg/g), while for other samples the deviation was also very small (CAW: Qe,cal= 10.95 mg/g and Qe,exp= 12.93 mg/g; WAC: Qe,cal= 13.00 mg/g and Qe,exp= 12.95 mg/g). These findings suggest that this adsorption system belongs to the pseudo-first-order reaction.
Furthermore, the experimental data fitted the pseudo-second-order equation (Fig. 7b), calculating the respective parameters (Simonin 2016). The slope (1/Qe) and intercept (1/k2Qe2) of the plot (t/Qt) versus t were used to calculate the parameters of k2 and Qe, cal. The straight lines in plots of Fig. 8b showed an unsuccessful fitting with smaller correlation coefficients than those of the pseudo-first-order model. The correlation coefficients was in the range 0.818–0.876, with big differences of the calculated Qe,cal values compared to the experimental ones (AWC: Qe,cal = 12.68 mg/g and Qe,exp = 38.62 mg/g; CAW: Qe,cal = 10.95 mg/g and Qe,exp = 37.31 mg/g; WAC: Qe,cal = 13.00 mg/g and Qe,exp = 44.92 mg/g).