Effective Poly (Cyclotriphosphazene-Co-4,4′-Sulfonyldiphenol)@rGO Sheets for Tetracycline Adsorption: Fabrication, Characterization, Adsorption Kinetics and Thermodynamics

The development of excellent drug adsorbents and clarifying the interaction mechanisms between adsorbents and adsorbates are greatly desired for a clean environment. Herein, we report that a reduced graphene oxide modified sheeted polyphosphazene (rGO/poly (cyclotriphosphazene-co-4,4′-sulfonyldiphenol)) defined as PZS on rGO was used to remove the tetracycline (TC) drug from an aqueous solution. Compared to PZS microspheres, the adsorption capacity of sheeted PZS@rGO exhibited a high adsorption capacity of 496 mg/g. The adsorption equilibrium data well obeyed the Langmuir isotherm model, and the kinetics isotherm was fitted to the pseudo-second-order model. Thermodynamic analysis showed that the adsorption of TC was an exothermic, spontaneous process. Furthermore, we highlighted the importance of the surface modification of PZS by the introduction of rGO, which tremendously increased the surface area necessary for high adsorption. Along with high surface area, electrostatic attractions, H-bonding, π-π stacking and Lewis acid-base interactions were involved in the high adsorption capacity of PZS@rGO. Furthermore, we also proposed the mechanism of TC adsorption via PZS@rGO.


Introduction
Antibiotics have been used worldwide for the treatment of various diseases related to humans and animals [1]. Tetracycline (TC) is one of the most extensively used antibiotics, owing to its low toxicity with a broad spectrum of activity [2]. However, in recent years, increasing concern has been raised due to poor degradation through the metabolism. Consequently, residual TC is discharged to the environment through urine and feces, causing non-point pollution [2][3][4]. It also causes chronic toxicity, lower immunity, dysbacteriosis, liver damage and gastrointestinal reactions in human bodies; and also affects the aquatic photosynthetic species and indigenous microbial growth, which ultimately damages the food chain [5,6]. Therefore, it is the need of the hour to remove TC from the environment for a green future [7,8].
In this regard, various techniques such as electrochemical degradation, photochemical degradation, chemical oxidation, biodegradation and adsorption have been applied to remove TC [9][10][11][12][13]. Among these techniques, adsorption is a promising approach to eliminate the residual TC due to its efficiency and cost-effectiveness [14]. The adsorbents that have been successfully used for this purpose including polymeric materials [14], montmorillonite [7], smectite clay [15], alumina [16], diatomite [8] and carbon materials [17,18]. Polyphosphazenes are an important class of organic-inorganic hybrid polymers, comprising nitrogen and phosphorus atoms in a conjugated binding owing to their backbone stability, structural diversity, biocompatibility, biodegradability and ability to form hybrid molecules [19][20][21][22][23][24][25][26][27]. Poly (cyclotriphosphazene-co-4,4-sulfonyldiphenol) (PZS) is a copolymer of hybrid phosphazenes [28] containing an organic-inorganic structure, which makes it a promising candidate with an extensive range of potential applications, e.g., encapsulation [29], catalyst support [30], carbon material precursor [31], electrochemical features [32] and bio-medicinal applications [33]. Fu et al. [34] and Wei et al. [35] successfully adsorbed different toxic materials from aqueous media using PZS microspheres. The excellent adsorption capacity of these microspheres is related to their various binding modes such as H-bonding, electrostatic interactions, π-π stacking and group bindings. However, they exhibited less surface area while the adsorption process demands a high surface area of adsorbents. To overcome this issue, various carbon-based substrates such as multi-walled carbon nanotubes and graphene sheets were introduced to enhance the surface area [36].
Herein, for the first time, PZS@rGO (4-12%) was used to investigate the adsorption of TC. We further investigated the effect of various parameters such as pH, temperature, initial concentration of adsorbate and adsorbent dosage on its adsorption. The main objectives of the present study are as follows: (1) controlling the growth of the PZS sheet on rGO; (2) examine the adsorption performance of these materials for TC and find out the optimum adsorption measurements; (3) describe their kinetic and thermodynamic models; and (4) explore the possible adsorption mechanism of TC. The structure-property relationship was investigated, which included the H-bonding, electrostatic interactions, π-π stacking, and Lewis-acid base interactions between adsorbent and adsorbate.

Synthesis of PZS@rGO
The scheme for the synthesis of the PZS@rGO is shown in Figure 1. We dispersed the prepared rGO (4%, 8% and 12%) [37] in 100 mL methanol under ultrasonication (150 W, 40 kHz) for 2 h. A total of 40 mg of HCCP and 80 mg of BPS solution in 10 mL methanol was mixed with the rGO solution. Afterward, 100 µL TEA was added dropwise, and magnetic stirring occurred for 7 h. The solution was washed with ethanol and then dried overnight under a vacuum. The PZS microspheres were prepared through analogous conditions reported elsewhere [19].
Nanomaterials 2021, 11, x FOR PEER REVIEW 3 of Figure 1. Scheme for the formation of PZS@rGO.

Adsorption of PZS@rGO Materials for Tetracycline Hydrochloride
Adsorption of Tetracycline hydrochloride was performed through the as-fabricate layered PZS@rGO by adding them into a conical flask containing 100 mL of TC solutio (100 ppm). The conical flask was ultrasonically agitated for 40 s and placed in a thermost at 30 °C with a rotation rate of 120 rpm (round per minute). In order to calculate the r sidual concentration of TC, we subsequently pipetted out 3 mL of suspension after a sp cific time interval, filtered it through a membrane (pore size: 0.22 μm) and finally, adsor tion was measured at λmax = 357 nm. Afterward, the adsorption capacity was calculate using Equation ( where qt is the adsorption capacity, C0 and Ct are the TC concentrations before and aft adsorption (ppm), respectively, V is the volume of the solution (L), and m is the mass adsorbent (g).

Characterization
UV-visible absorption spectra were obtained through a Lambda 3600 UV-vis spe trophotometer (PerkinElmer, Inc., Waltham, MA, USA). Further, X-ray photoelectro spectroscopy (XPS) analyses were performed via VG-ESCALAB 250 under a high vacuu (2 × 10 −9 Pa), at a standard Al Kα source (1486.6 eV). The C 1s peak at 284.9 eV was take as a reference for the binding energy. Fourier transform infrared (30 co-added) spect were collected through a Bruker Vector-22 while implying KBr as a carrier to fix the sam ples. The surface morphology of the samples was examined via JEOL JSM-7610F fie emission scanning electron microscope (SEM) (Tokyo, Japan). Transmission electron m croscopy (TEM) images were obtained from the Hitachi-H-800 transmission electron m croscope (TEM) (Tokyo, Japan). The surface area was calculated from nitrogen adsor tion-desorption isotherms at 77 K through Micrometrics ASAP 2460 using th

Adsorption of PZS@rGO Materials for Tetracycline Hydrochloride
Adsorption of Tetracycline hydrochloride was performed through the as-fabricated layered PZS@rGO by adding them into a conical flask containing 100 mL of TC solution (100 ppm). The conical flask was ultrasonically agitated for 40 s and placed in a thermostat at 30 • C with a rotation rate of 120 rpm (round per minute). In order to calculate the residual concentration of TC, we subsequently pipetted out 3 mL of suspension after a specific time interval, filtered it through a membrane (pore size: 0.22 µm) and finally, adsorption was measured at λ max = 357 nm. Afterward, the adsorption capacity was calculated using Equation (1) where q t is the adsorption capacity, C 0 and C t are the TC concentrations before and after adsorption (ppm), respectively, V is the volume of the solution (L), and m is the mass of adsorbent (g).

Characterization
UV-visible absorption spectra were obtained through a Lambda 3600 UV-vis spectrophotometer (PerkinElmer, Inc., Waltham, MA, USA). Further, X-ray photoelectron spectroscopy (XPS) analyses were performed via VG-ESCALAB 250 under a high vacuum (2 × 10 −9 Pa), at a standard Al Kα source (1486.6 eV). The C 1s peak at 284.9 eV was taken as a reference for the binding energy. Fourier transform infrared (30 co-added) spectra were collected through a Bruker Vector-22 while implying KBr as a carrier to fix the samples. The surface morphology of the samples was examined via JEOL JSM-7610F field emission scanning electron microscope (SEM) (Tokyo, Japan). Transmission electron microscopy (TEM) images were obtained from the Hitachi-H-800 transmission electron microscope (TEM) (Tokyo, Japan). The surface area was calculated from nitrogen adsorption-desorption isotherms at 77 K through Micrometrics ASAP 2460 using the Brunauer−Emmett−Teller method. Prior to N 2 adsorption-desorption, samples were subjected to 300 • C under vacuum for 6 h.

Results and Discussion
FTIR spectra of the PZS microsphere and PZS@rGO8% are shown in Figure 2. The characteristic peaks for hydroxylic (Phenolic group) could be observed at 3431 and 3093 cm −1 of the PZS microsphere ( Figure 2a). The peaks at 1292 and 1152 cm −1 were attributed to O = S = O and P = N peaks, respectively. A new peak at 942 cm −1 can be attributed to P-O-(Ph), which was evidence to prove the successful condensation of HCCP and BPS. FTIR spectra of PZS@rGO8% ( Figure 2b) followed the characteristic peaks of PZS microspheres and rGO. The new peaks emerged at 2920 and 2850 cm −1 due to the C-H stretching vibration, which belonged to rGO. The peaks at 1564 and 1466 cm −1 represented the N-H amide stretching and C-N bonding, respectively. The wide stretching of the peak at 3410 cm −1 is evidence of the presence of the abundant -OH group, which later proved to be the key site for adsorption.

Results and Discussion
FTIR spectra of the PZS microsphere and PZS@rGO8% are shown in Figure 2. T characteristic peaks for hydroxylic (Phenolic group) could be observed at 3431 and 30 cm −1 of the PZS microsphere ( Figure 2a). The peaks at 1292 and 1152 cm −1 were attribut to O = S = O and P = N peaks, respectively. A new peak at 942 cm −1 can be attributed to O-(Ph), which was evidence to prove the successful condensation of HCCP and BPS. FT spectra of PZS@rGO8% ( Figure 2b) followed the characteristic peaks of PZS microspher and rGO. The new peaks emerged at 2920 and 2850 cm −1 due to the C-H stretching vibr tion, which belonged to rGO. The peaks at 1564 and 1466 cm −1 represented the N-H ami stretching and C-N bonding, respectively. The wide stretching of the peak at 3410 cm −1 evidence of the presence of the abundant -OH group, which later proved to be the k site for adsorption.  Figure 3a-d, exhibited the layer morphology. Moreover, PZS@rGO8% exhibited the fabrication of a smooth PZS layer ov rGO sheets, while PZS@rGO4% showed the aggregation of PZS over rGO more likely the form of microspheres, possibly due to less surface being offered by rGO due to lower content. Meanwhile, the PZS@rGO12% morphology was more wrinkled due to t aggregation of rGO sheets. The TEM images of PZS@rGO (4%, 8% and 12%), Figure 4brespectively, were consistent with the SEM results. The light grey area was attributed the rGO sheets, while the dark area was indexed to PZS layers. PZS@rGO8% exhibited t thin layer of PZS equally spreading over rGO sheets. Meanwhile, PZS@rGO4% was a gregated in a bulk form, and PZS@rGO12% had excessive rGO that was aggregated. The morphology of PZS@rGO was examined through SEM ( Figure 3) and TEM ( Figure 4). SEM images of the PZS microspheres ( Figure S1) show the spherical morphology, while the introduction of rGO in the PZS, shown in Figure 3a-d, exhibited the layered morphology. Moreover, PZS@rGO8% exhibited the fabrication of a smooth PZS layer over rGO sheets, while PZS@rGO4% showed the aggregation of PZS over rGO more likely in the form of microspheres, possibly due to less surface being offered by rGO due to its lower content. Meanwhile, the PZS@rGO12% morphology was more wrinkled due to the aggregation of rGO sheets. The TEM images of PZS@rGO (4%, 8% and 12%), Figure 4b-d, respectively, were consistent with the SEM results. The light grey area was attributed to the rGO sheets, while the dark area was indexed to PZS layers. PZS@rGO8% exhibited the thin layer of PZS equally spreading over rGO sheets. Meanwhile, PZS@rGO4% was aggregated in a bulk form, and PZS@rGO12% had excessive rGO that was aggregated.   Furthermore, the successful fabrication of PZS over rGO was evidenced b ure 5). Figure 5a shows the complete XPS spectra of PZS@rGO8%, which reve necessary peaks of aforementioned adsorbent. The high-resolution C 1s XPS PZS@rGO (Figure 5b) was further deconvoluted into three components, C = C ized peak at 284.55 eV, C sp 3 peak at 285.22 eV and the π-π stacking peak at 28 to the conjugated system and aromatic structure [38]. The π-π stacking peak a suggested the successful fabrication of PZS to rGO. Figure 5c shows the N 1s   Furthermore, the successful fabrication of PZS over rGO was evidenced b ure 5). Figure 5a shows the complete XPS spectra of PZS@rGO8%, which revea necessary peaks of aforementioned adsorbent. The high-resolution C 1s XPS PZS@rGO (Figure 5b) was further deconvoluted into three components, C = C ized peak at 284.55 eV, C sp 3 peak at 285.22 eV and the π-π stacking peak at 28 to the conjugated system and aromatic structure [38]. The π-π stacking peak a suggested the successful fabrication of PZS to rGO. Figure 5c shows the N 1s X Furthermore, the successful fabrication of PZS over rGO was evidenced by XPS ( Figure 5). Figure 5a shows the complete XPS spectra of PZS@rGO8%, which revealed the all-necessary peaks of aforementioned adsorbent. The high-resolution C 1s XPS spectra of PZS@rGO (Figure 5b) was further deconvoluted into three components, C = C sp 2 hybridized peak at 284.55 eV, C sp 3 peak at 285.22 eV and the π-π stacking peak at 289.59 eV due to the conjugated system and aromatic structure [38]. The π-π stacking peak at  Figure 5c shows the N 1s XPS spectra distributed into types of N species. Peaks at 397.94 and 401.1 eV were assigned to pyridinic N of PZS and graphitic N, respectively [39][40][41]. The XPS spectra of P 2p and S 2p were displayed in Figure 5d,e. Peaks at 133.52 and 167.5 eV were assigned to oxidized P 2p and S 2p, respectively [42,43]. The XPS spectra of O 1s (532.48 eV) shown in Figure 5f confirmed the presence of −OH species. Meanwhile, it was noticed from the XPS atomic content (Table S1) that PZS@rGO8% has 93.25% rGO content. The surface area and structure of the PZS@rGO8% were explored by N2 a desorption isotherms ( Figure S2). PZS@rGO8% showed the type IV isotherm shaped hysteresis loops [44], which is a characteristic of mesoporous materials tive pressure of 0.451-1.0 is indexed to the mesoporous structure and the slit-sh in the material [45]. The adsorbent exhibited a high surface area of 443 m 2 /g considerably higher than the surface area of the PZS microsphere, 10.5 m 2 /g [3

Adsorption Kinetics and Adsorption Capacity
Adsorption kinetics is a very important tool to probe adsorption behavior date the rate of adsorption. Kinetics data also describes significant information be helpful for the evaluation and understanding of adsorption mechanisms [4 The surface area and structure of the PZS@rGO8% were explored by N 2 absorptiondesorption isotherms ( Figure S2). PZS@rGO8% showed the type IV isotherms with H3shaped hysteresis loops [44], which is a characteristic of mesoporous materials. The relative pressure of 0.451-1.0 is indexed to the mesoporous structure and the slit-shaped hole in the material [45]. The adsorbent exhibited a high surface area of 443 m 2 /g, which is considerably higher than the surface area of the PZS microsphere, 10.5 m 2 /g [35].

Adsorption Kinetics and Adsorption Capacity
Adsorption kinetics is a very important tool to probe adsorption behavior and elucidate the rate of adsorption. Kinetics data also describes significant information that might be helpful for the evaluation and understanding of adsorption mechanisms [46]. The adsorption capacity of the as-prepared PZS@rGOx (x = 4%, 8% 12%) and PZS microsphere versus time is disclosed in Figure 6. PZS@rGO8% exhibited the highest adsorption capacity of 496 mg/g, and equilibrium was attained in 40 minutes, which was higher than many reported studies Table S1. However, there were descending trends of 265 and 156 mg/g for PZS@rGO12% and PZS@rGO4%, respectively. Although the PZS microsphere contains all the elemental components required to perform adsorption at the same rate as its composites with rGO, it was the surface area that played a significant role in the adsorption capacity. The PZS@rGO8% adsorbent shows a higher surface area of 443 m 2 /g, which is a lot higher than the reported surface area of 10.5 m 2 /g of the PZS microsphere [35]. However, in the case of PZS@rGO4%, the PZS turned into an agglomerated morphology due to fewer rGO contents, and the rGO sheets tend to agglomerate in PZS@rGO12% (Figure 4). The pseudo-first-order and pseudo-second-order kinetic models were use plore the kinetic mechanisms of TC adsorption, and the linear forms of these mo shown in Equations (2) and (3) where qe and qt (mg g −1 ) are the adsorption capacity at equilibrium and at any time respectively. k1 (min −1 ) is the pseudo-first-order rate constant and k2 (g mg −1 min − pseudo-second-order rate constant. The kinetic parameters and the correlation coe (R 2 ) obtained by linear regression are listed in Table 1. According to Figure 7, a pseudo-second-order kinetic model better explains sorption of TC on PZS@rGO with a correlation coefficient approaching one (Table 1 qe,exp is defined as the adsorption capacity at the equilibrium of any adsorbent fro time experiments. While qe,cal is known as the adsorption capacity at equilibrium The pseudo-first-order and pseudo-second-order kinetic models were used to explore the kinetic mechanisms of TC adsorption, and the linear forms of these models are shown in Equations (2) and (3), respectively.
where q e and q t (mg g −1 ) are the adsorption capacity at equilibrium and at any time (t; min), respectively. k 1 (min −1 ) is the pseudo-first-order rate constant and k 2 (g mg −1 min −1 ) is the pseudo-second-order rate constant. The kinetic parameters and the correlation coefficients (R 2 ) obtained by linear regression are listed in Table 1. According to Figure 7, a pseudo-second-order kinetic model better explains the adsorption of TC on PZS@rGO with a correlation coefficient approaching one (Table 1). Term q e,exp is defined as the adsorption capacity at the equilibrium of any adsorbent from real-time experiments. While q e,cal is known as the adsorption capacity at equilibrium calculated through different kinetic models, we may also refer to the latter term as theoretical adsorption. The theoretical adsorption capacity value differs with different applied kinetic models. The closer the theoretical adsorption capacity with real-time experimental results are, the more probable the adsorption system follows that particular kinetic model. In this case, theoretical adsorption capacity values suggest the adsorption system follows the pseudo-second-order reaction. Furthermore, the correlation R 2 values of pseudo-secondorder reaction are close to unity. The fact that the adsorption of TC through PZS@rGO followed the pseudo-second-order reaction confirms that there are different kinds of interactions between adsorbent and adsorbate due to the microporosity of PZS@rGO and surface functional groups of both adsorbent and adsorbate [47].

Adsorption Isotherms
The Langmuir and Freundlich isotherm models were applied to reveal insights into the adsorption phenomena. The Langmuir isotherm model assumes that the molecules are adsorbed in a monolayer configuration on the adsorbent surface where all the adsorption sites possess the same energy, so the adsorption of a species is identical [48][49][50].In

Adsorption Isotherms
The Langmuir and Freundlich isotherm models were applied to reveal insights into the adsorption phenomena. The Langmuir isotherm model assumes that the molecules are adsorbed in a monolayer configuration on the adsorbent surface where all the adsorption sites possess the same energy, so the adsorption of a species is identical [48][49][50]. In comparison, the Freundlich isotherm model states multilayer, non-ideal adsorption, and the adsorption sites are unevenly distributed, involving a different affinity for adsorbing molecules. Equations (4) and (5) illustrated linear forms of Langmuir and Freundlich isotherm models, respectively.
log q e = log K f + 1 n log C e (5) where b is the adsorption energy constant (L mg −1 ), q m is the theoretically calculated maximum Langmuir adsorption capacity (mg g −1 ), C e is related to the equilibrium adsorption concentration (mg L −1 ) and K f is associated to the Freundlich constant [(mg/g)·(L/mg) 1/n ] while 1/n is the adsorption strength.
The adsorption of TC on PZS@rGO followed the Langmuir isotherm in comparison to Freundlich isotherm (Figure 8). At 303 K, the R 2 value of adsorption of TC on PZS@rGOx was almost equal for the Langmuir and Freundlich equations, Table 2. This indicated that the adsorption of TC is more consistent with the Langmuir isotherm in comparison to the Freundlich isotherm and the adsorption of TC was monolayered.

Batch Experiments
For further experiments, PZS@rGO8% is used as the adsorbent as it has the maximum adsorption capacity, as shown in Figure 6.

Effect of pH
The pH of a solution is an important factor that greatly influences the adsorption process because pH influences the surface charge and ionization behavior of materials. Figure 9a illustrates the effect of the initial pH of the solution on the adsorption capacity of PZS@rGO. The adsorption of TC was immensely lowered at pH = 2.0, and it could be related to the ion competition between H + and TC + for the adsorption sites of PZS@rGO [51]. The adsorption of TC increased with the increasing pH of the solution. The adsorption capacity reaches a maximum point at pH = 6.0. The ionization and hydration of TC may reduce, which is advantageous to the adsorption process via H-bonding and π-π stacking effect. TC may exist in a zwitter-ion formation at pH = 6-7. However, a further increase in pH slightly reduced the adsorption of TC. An increase in pH above a neutral value may affect H-bonding via OHgeneration [52]. Further experiments are conducted

Batch Experiments
For further experiments, PZS@rGO8% is used as the adsorbent as it has the maximum adsorption capacity, as shown in Figure 6.

Effect of pH
The pH of a solution is an important factor that greatly influences the adsorption process because pH influences the surface charge and ionization behavior of materials. Figure 9a illustrates the effect of the initial pH of the solution on the adsorption capacity of PZS@rGO. The adsorption of TC was immensely lowered at pH = 2.0, and it could be related to the ion competition between H + and TC + for the adsorption sites of PZS@rGO [51]. The adsorption of TC increased with the increasing pH of the solution. The adsorption capacity reaches a maximum point at pH = 6.0. The ionization and hydration of TC may reduce, which is advantageous to the adsorption process via H-bonding and π-π stacking effect. TC may exist in a zwitter-ion formation at pH = 6-7. However, a further increase in pH slightly reduced the adsorption of TC. An increase in pH above a neutral value may affect H-bonding via OHgeneration [52]. Further experiments are conducted at the optimized pH = 6. of molecules, as the temperature is directly proportional to the kinetic energy, which ultimately enhances the mobility of molecules, and therefore, adsorption is increased [53]. Therefore, these results facilitate to opt for the best conditions to pursue the adsorption of TC via PZS@rGO for practical implications.

Effect of Dosage
At the initial concentration, the effects of dosage on the amount of adsorption were studied, and the results are shown in Figure 9c. The removal of TC significantly increased with an increase in dosage amount. When the dosage of the adsorbent increased from 2 to 10 mg for 100 ppm initial concentration, the removal efficiency of the adsorbent increased from 90 to 496 mg/g, respectively. This is due to the adsorbent's more active adsorption sites at high dosage concentrations.

Effect of Initial Concentrations
The variation in adsorption capacity of PZS@rGO due to different initial concentrations of tetracycline was examined at pH = 6.0, 303 K and dosage of 0.1 g L −1 . Adsorption of TC decreases with the increase in initial concentration of TC, as shown in Figure 9d. Typically, this phenomenon is related to the saturation of available adsorption sites, and with the increase in concentrate, PZS@rGO does not provide enough binding sites for TC Figure 9. Effect of initial pH (10 mg PZS@rGO8% adsorbent, 100 ppm TC, volume of solution 100 mL, Temperature 30 • C) (a), effect of Temperature (10 mg PZS@rGO8% adsorbent, 100 ppm TC, volume of solution 100 mL, pH = 6) (b), effect of PZS@rGO8% adsorbent dosage (100 ppm TC, volume of solution 100 mL, pH = 6, temperature 30 • C) (c), effect of initial concentration (10 mg PZS@rGO8% adsorbent, volume of solution 100 mL, pH = 6, temperature 30 • C) (d).

Effect of Temperature
The temperature has a significant impact on the adsorption of TC, as shown in Figure 9b. We observe that at equilibrium, the adsorption capacity of PZS@rGO increases from 365 to 455 mg/g when the temperature is raised from 293 to 298 K. The results explain that the adsorption of TC is ideal at 303 K. As the temperature increases, the kinetic energy of molecules increases resulting in a higher proportion of molecular collision, ultimately increasing the amount of adsorption. According to Le Chateliar's principle, when all the other conditions are constant, an increase in temperature shifts the equilibrium in the forward direction when the reaction is endothermic in nature, which was later confirmed by the thermodynamic data (Table 3). An increase in temperature enhances the kinetic energy of molecules, as the temperature is directly proportional to the kinetic energy, which ultimately enhances the mobility of molecules, and therefore, adsorption is increased [53]. Therefore, these results facilitate to opt for the best conditions to pursue the adsorption of TC via PZS@rGO for practical implications.

Effect of Dosage
At the initial concentration, the effects of dosage on the amount of adsorption were studied, and the results are shown in Figure 9c. The removal of TC significantly increased with an increase in dosage amount. When the dosage of the adsorbent increased from 2 to 10 mg for 100 ppm initial concentration, the removal efficiency of the adsorbent increased from 90 to 496 mg/g, respectively. This is due to the adsorbent's more active adsorption sites at high dosage concentrations.

Effect of Initial Concentrations
The variation in adsorption capacity of PZS@rGO due to different initial concentrations of tetracycline was examined at pH = 6.0, 303 K and dosage of 0.1 g L −1 . Adsorption of TC decreases with the increase in initial concentration of TC, as shown in Figure 9d. Typically, this phenomenon is related to the saturation of available adsorption sites, and with the increase in concentrate, PZS@rGO does not provide enough binding sites for TC adsorption. As the initial concentration increases from 100 to 300 ppm, the removal percentage of TC decreases from 98% to 77%, respectively, suggesting at lower initial concentrations of TC, the removal percentage is higher.

Adsorption Thermodynamics
The thermodynamic parameters of PZS@rGO are determined at 20, 25 and 30 • C, and adsorption isotherm constants were obtained. Gibbs free energy was used as a criterion to determine the spontaneity of reaction and was calculated by Equation (6).
where ∆G is the change in Gibbs free energy, R is the general gas constant (8.314 J/K mol), T is the temperature (K) and K L is the distribution coefficient of Langmuir equilibrium isotherm (L/mol). Change in Enthalpy (∆H) and entropy (∆S) were calculated using Equation (7): All the thermodynamics data are shown in Table 3. Negative values of ∆G suggest the adsorption reaction is spontaneous while 30 • C is the most favorable temperature for adsorption reaction. A rise in temperature increases the diffusion rate and a reduction in the viscosity of the solution [54]. The low ∆S values exhibit no massive changes in the structure of the adsorption system, suggesting that there might be physical interactions more influential than chemical interactions [55]. Positive values of ∆H indicate that the adsorption process has to be endothermic in nature, and these results suggest the probability of bonding between adsorbent and dye [56]. If an endothermic reaction takes place, then for the process to remain spontaneous, the entropy change should not only be positive, but T∆S must exceed ∆H, numerically, so that the net Gibbs free energy change as a whole is negative. The trend of fewer enthalpy changes than entropy has been noticed in TC adsorption. Generally, the low enthalpy change is related to physisorption (<40 kJ/mol). An increase in temperature accelerates the amount of TC adsorption, which is consistent with the experimental results. The thermodynamics of TC adsorption suggests adsorption may occur at mild conditions due to lesser enthalpy change, low entropy and negative Gibbs free energy change.

Mechanism
Based on these findings, we illustrated the adsorption mechanism of TC. The excellent adsorption capacity of PZS@rGO is dedicated to its high surface area and unique chemical composition. The BET surface area obtained by nitrogen adsorption-desorption (443 m 2 /g) was ca. 42 times higher than its microsphere [35]. The high surface area is the main difference between the PZS microsphere and PZS sheet's adsorption capacity, as the chemical composition of both microspheres and sheets is similar. The results suggest that the introduction of rGO into PZS changed the morphology from microspheres into sheets and improved the adsorption capacity of the adsorbent. Considering the structural characterization of both PZS@rGO and TC, there were multiple factors involved during adsorption. The PZS sheets prepared in our study contain abundant hydroxyl groups on their surface, which was confirmed by FT-IR results (Figure 2). These hydroxyl groups have a tendency to produce an H-bonding interaction and an electrostatic interaction with TC. Further, it is a well-known electrostatic screening effect that the addition of salt inhibits the electrostatic interactions. The addition of NaCl in the adsorbent-adsorbate solution reduced the adsorption capacity shown in Figure 10. It is an obvious hint at the existence of electrostatic attraction. Both adsorbent and adsorbate possess a benzene ring in their structures, which enhances the π-π interaction [35].
aterials 2021, 11, x FOR PEER REVIEW TC. Further, it is a well-known electrostatic screening effect that th its the electrostatic interactions. The addition of NaCl in the adsor reduced the adsorption capacity shown in Figure 10. It is an obvi of electrostatic attraction. Both adsorbent and adsorbate possess structures, which enhances the π-π interaction [35]. Another route for adsorption mechanisms might be on the ba tions in Figure 11. TC contains a number of anionic, cationic an structure, e.g., organic ammonium cations as the Lewis acid, wh abundant nitrogen with a lone pair as the Lewis base. Later, it was ing an experiment by treating adsorbent with ZnCl2. Cyclotriphos ity to engage in chelation with Zn(II) [35]. Figure 10 showed the capacity of PZS@rGO after treating with ZnCl2 due to the occupat pair by Zn (II) ions. Therefore, excellent adsorption of TC can be surface area of PZS@rGO, which facilitates adsorption by provid trostatic interactions, π-π stacking and acid-base interactions. Another route for adsorption mechanisms might be on the basis of acid-base interactions in Figure 11. TC contains a number of anionic, cationic and neutral groups in its structure, e.g., organic ammonium cations as the Lewis acid, while the PZS sheets have abundant nitrogen with a lone pair as the Lewis base. Later, it was confirmed by conducting an experiment by treating adsorbent with ZnCl 2 . Cyclotriphosphazenes have the ability to engage in chelation with Zn(II) [35]. Figure 10 showed the reduction in adsorption capacity of PZS@rGO after treating with ZnCl 2 due to the occupation of the nitrogen lone pair by Zn (II) ions. Therefore, excellent adsorption of TC can be speculated to the high surface area of PZS@rGO, which facilitates adsorption by providing more sites for electrostatic interactions, π-π stacking and acid-base interactions. Nanomaterials 2021, 11, x FOR PEER REVIEW 13 of 16 Figure 11. Proposed mechanism of TC decontamination through PZS@rGO.

Conclusions
In conclusion, the PZS layers modified with reduced graphene (PZS@rGO) have been successfully fabricated by simple co-precipitation methods under mild conditions. PZS@rGO successfully removed the tetracycline antibiotic from an aqueous solution and exhibited an excellent adsorption capacity of 496 mg/g in comparison to the PZS microsphere's 102 mg/g. Further studies revealed that the adsorption of TC followed the Langmuir isotherm model and pseudo-second-order kinetics. Moreover, the removal of TC was spontaneous. The high adsorption capacity was attributed to a high surface area of the adsorbent. Along with a high surface area, electrostatic attractions, H-bonding, π-π stacking and Lewis acid-base interactions were involved for the high adsorption capacity of PZS@rGO. This study not only provides efficient drug removal but also highlights the effect of different parameters on the adsorption of TC over PZS@rGO.

Conclusions
In conclusion, the PZS layers modified with reduced graphene (PZS@rGO) have been successfully fabricated by simple co-precipitation methods under mild conditions. PZS@rGO successfully removed the tetracycline antibiotic from an aqueous solution and exhibited an excellent adsorption capacity of 496 mg/g in comparison to the PZS microsphere's 102 mg/g. Further studies revealed that the adsorption of TC followed the Langmuir isotherm model and pseudo-second-order kinetics. Moreover, the removal of TC was spontaneous. The high adsorption capacity was attributed to a high surface area of the adsorbent. Along with a high surface area, electrostatic attractions, H-bonding, π-π stacking and Lewis acid-base interactions were involved for the high adsorption capacity of PZS@rGO. This study not only provides efficient drug removal but also highlights the effect of different parameters on the adsorption of TC over PZS@rGO.