Application of Ni 0 . 5 Zn 0 . 5 Fe 2 O 4 magnetic nanoparticles for diclofenac adsorption : isotherm , kinetic and thermodynamic investigation

Ni0.5Zn0.5Fe2O4 magnetic nanoparticles were synthesized to obtain a new efficient adsorbent for diclofenac sodium (DF) removal. Fourier Transform Infrared (FTIR), Energy Dispersive Spectrometer (EDS), scanning electron microscope (SEM), Brunauer–Emmett–Teller (BET) and vibrating sample magnetometer (VSM) were applied to characterize the prepared adsorbent. These analyses revealed that adsorbent was successfully prepared with average particle diameter of about 50 nm and a BET surface area of 168.09 m/g. The saturation magnetization value of magnetic nanoparticles (MNPs) was found to be 24.90 emu/g, thus, adsorbent was efficiently separated from the solution by a facile and rapid magnetic separation process. The effect of adsorption time, amount of adsorbent, initial pH of the solution, initial diclofenac concentration and temperature on the removal of DF were evaluated. Also, the adsorption data were best fitted to the pseudo-first-order kinetic model and Langmuir isotherm model. The thermodynamics studies suggested spontaneous and exothermic adsorption. The maximum diclofenac adsorption amount of the synthesized nanoadsorbent was 52.91 mg/g, which is higher than many recently studied adsorbents.


GRAPHICAL ABSTRACT INTRODUCTION
During the last several years, the existence of pharmaceutical compounds in aquatic media has been widely detected. These compounds are discharged into the aquatic systems through hospitals and drug factories (Calza et  Diclofenac (DCF) is one of the widely used pharmaceutical compounds for reduction of inflammation and pain relief (Soares et al. ; Shayesteh et al. ). Diclofenac consumption was found to be about 1,000 tons per year (Pylypchuk et al. ). Diclofenac is a non-degradable toxic substance and can be absorbed and enriched by organisms and human bodies (Xiong et al. ; Xu et al. ). Its long term exposure has adverse results such as thyroid disease, hemodynamic and renal effects (Hasan et al. ; Bickley et al. ). With these deleterious effects, the removal of diclofenac from effluents is extremely important.
Thus, various methods including flocculation, biodegradation, electrochemical degradation, chlorination, ozonation and adsorption, have been used to treat contaminated wastewaters (Finkbeiner et  Among a range of adsorbents, nano-sized adsorbents have recently attracted great interest as pharmaceutical adsorbents because of their high adsorption capacity based on their small size. In particular, magnetic nanoparticles (MNPs) attracted far more attention for use as magnetically separable nanoadsorbents. It had been proved that the magnetic property of nanoparticles allows them to be recovered rapidly and effectively by an external magnetic field from aquatic media in comparison to conventional approaches of centrifugation or filtration (Pashai Gatabi  This work is an effort to remove diclofenac pollutant using Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs from aqueous solution. For this purpose, Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic nanoparticles were prepared. Then, the characterization of adsorbents was conducted by some analytical techniques. After that, the adsorption process of this magnetic nanoadsorbent was carried out. Several parameters affecting the adsorption method, like contact time, initial diclofenac concentration, amount of nanoadsorbent, initial solution pH and temperature, were investigated. Also, equilibrium isotherm and kinetic model studies were done.

Equipment
Fourier Transform Infrared (FTIR) of the synthesized sample was done through a potassium bromide (KBr) pellet with an FTIR spectrometer (Perkin Elmer Spectrum, Germany). The adsorbent surface area has been analyzed by Brunauer-Emmett-Teller (BET) method (NOVA ® Station B, Quantachrome). To investigate the morphological features of magnetic nanoadsorbent, scanning electron microscope (SEM) images were obtained (VEGA\ \TESCAN SEM). Qualitative composition analysis of MNPs was performed using an Energy Dispersive Spectrometer (EDS) attached to an SEM instrument. The magnetic features of the sample were determined from magnetization measurement using a vibrating sample magnetometer (VSM) (Lakeshore, USA). The diclofenac absorbance was determined using a UV-vis spectrometer (Shimadzu, Japan). An ultrasonic bath was used for the mixing of samples (DT31H, BANDELIN electronic, Germany). MNPs formation (approximately 90 min). Then, a strong magnet was applied to collect synthesized MNPs from solution. Finally, the solid residue was washed repeatedly with distilled water and acetone, followed by drying at 50 C for 12 h, to obtain nano-scale Ni 0.5 Zn 0.5 Fe 2 O 4 powder.

Adsorption studies
The adsorption of diclofenac on Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs was investigated through batch technique. A stock solution (1,000 mg/L) was prepared by dissolving the appropriate amount of DF in distilled water, which was diluted to desired concentrations (15-150 mg/L). Various parameters affecting the adsorption process, namely contact time (15-90 min), adsorbent concentration (0.1-1.8 g/L), initial pH (3.2-10.3), initial diclofenac concentration (15-150 mg/L) and temperature (25-55 C), were evaluated. The effect of each parameter was assessed by varying that parameter while the other parameters were kept constant. All batch experiments were performed in an ultrasonic bath, using 250 mL Erlenmeyer flasks including 50 mL diclofenac solution. For each experiment, 60 mg of MNPs was added to 50 mL of desired diclofenac solution, and 0.5 M HCl or 0.5 M NaOH was applied to adjust the solution pH. The resulting solution was mixed for a predetermined time. After adsorption, MNPs were separated magnetically. The remaining diclofenac concentration in solution was determined via UV-visible spectrometer.
The diclofenac removal efficiency was calculated using Equation (1) and the adsorption capacity of adsorbent was obtained by Equation (2) where C 0 , C e and C are the liquid-phase concentration of diclofenac in solution at the initial, equilibrium and final states, respectively (mg/L). V shows the liquid phase volume (L), and M stands for the mass of the adsorbent (g).

Characterization results
The FTIR spectrum of Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic nanoparticles is shown in Figure 1. The magnetic characteristic of the Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic nanoparticles was characterized by VSM analysis and the result is shown in Figure 2. The Ni 0.5 Zn 0.5 Fe 2 O 4 MNP has a saturation magnetization of 24.90 emu/g. Within the external magnetic field, Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic nanoparticles indicated a good magnetic response and were separated in several seconds.
SEM was used to investigate the morphology of MNPs, as shown in Figure 3. Figure 3(a)-3(c) demonstrate the presence of nanosize particles, with an average diameter equal to 50 nm. As can be seen in Figure 3(a)-3(c), there are predominantly mostly regularly spherical-shaped particles. Also, small aggregates of MNPs are due to magnetic dipole-dipole interaction between nanoparticles. A similar observation was reported by Bennet and coworkers (Bennet et al. ). Also, the SEM image of Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs after the adsorption process (Figure 3(d)) shows that the adsorbent has a stable structure.
The result of elemental analysis of the Ni 0.5 Zn 0.5 Fe 2 O 4 nanoparticles was provided from the energy dispersive spectrometry, presented in Figure 4. The existence of Ni and Zn peaks is consistent with the FTIR analysis. Also, the obtained MNPs revealed a BET surface area of 168.09 m 2 /g.

Effect of contact time on diclofenac adsorption
Determining equilibrium contact time is important to study the adsorption behaviour of diclofenac on Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs. To assess the impact of contact time, adsorption tests were done up to 90 min. As presented in Figure 5, an increase in diclofenac adsorption by increasing contact time can be seen. It is seen in this figure that after 60 min no considerable changes were observed and equilibrium time was attained. This fast equilibrium time might be due to high dispersion and high specific surface area of Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs. The results revealed that kinetic of adsorption of DF could be divided into two phases: an initial rapid phase where adsorption was fast, that can be attributed to a large number of vacant sites of adsorbent and high DF concentration gradient in the first few minutes, and a slower second phase where the adsorption rate was relatively small, that can be attributed to a lower number

Effect of adsorbent dosage on diclofenac adsorption
Impact of adsorbent dosage on diclofenac removal is presented in Figure 6. The adsorbent quantity was varied within the range of 0.1 g/L to 1.8 g/L by keeping all other parameters constant. It is clear that increase in adsorbent dose results in an increase in removal efficiency from 18.83% to 57.15%, and leads to obtaining lower adsorption capacity from 37.66 mg/g to 7.01 mg/g. The increase in removal efficiency of diclofenac could be explained by the availability of more adsorbent for a certain amount of diclofenac. At higher adsorbent dose, more adsorption sites are available for DF. On the other hand, the reduction of adsorption capacity at high adsorbent dose may be due to

Effect of initial pH on diclofenac adsorption
The pH of diclofenac media may be considered as an important parameter in diclofenac removal by Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs. In order to evaluate the effect of pH on diclofenac adsorption, the initial pH of the solution was regulated to 3.2, 6, 7.7 and 10.3 by either NaOH or HCl. The outcome of the experiments is shown in Figure 7. It is clear that the lowest adsorption capacity was observed at a pH of 10.3 (4.54 mg/g) and the maximum adsorption capacity was reached at pH 3.2 (9.81 mg/g). The higher adsorption capacity of diclofenac at low pH could be due to the positively charged adsorbent surface (Zha et al. ; Demiral & Güngör ).

Equilibrium study
The adsorption isotherm denotes equilibrium concentration of adsorbate between the adsorbent and solution. The three most common isotherm models are Langmuir, Freundlich and Temkin (   The Langmuir isotherm suggests a homogenous adsorbent surface without interactions between adsorbed molecules. The linearized form of this isotherm is represented in Equation (3): Here in Equation (3), q e (mg/g) and C e (mg/L) represent the concentration of drug on the adsorbent and in aqueous solution, respectively. q m (mg/g) represents the highest adsorption capacity on the adsorbent, and K L (L/mg) is Langmuir adsorption equilibrium constant. K L is referred to as adsorption energy. The values of q m and K L can be obtained through a plot of 1 q e versus 1 C e . The Freundlich isotherm model is generally used for heterogeneous surface and multilayer adsorption. The linearized form of this isotherm is written as Equation (4): where K F and n are Freundlich parameters that show adsorption capacity of adsorbent and degree of adsorption. A plot of ln q e versus ln C e reveals a straight line of slope 1 n and an intercept of ln K F .  The Temkin isotherm model considers the interactions of adsorbates on the adsorbent. The linearized form the Temkin model is written as Equation (5): where b T (J/mol) is the Temkin constant, which represents the adsorption heat, A T (L/g) is called the Temkin isotherm binding constant, R (8.314 J/mol K) stands for the universal gas constant and T (K) denotes temperature. A linear plot of q e against ln C e was applied to calculate b T and A T . The variation in the adsorption amount with diclofenac initial concentration by Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs was investigated. The Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs adsorption capacity with different equilibrium concentrations of diclofenac is illustrated in Figure 8. According to this figure, the adsorption capacity of adsorbent, q e (mg/g), increased from 5.25 mg/g to 44.72 mg/g when equilibrium concentration increased, which is most likely due to enhanced driving force to overcome mass transfer resistance of the diclofenac between adsorbent and solution. A similar result was also reported by Hemmati et al. (Hemmati et al. ).
The adsorption isotherm is used to analyze the equilibrium data. For this purpose, adsorption equilibrium data for diclofenac removal onto Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs have been studied with the three most common adsorption isotherms, namely the Langmuir, Freundlich and Temkin. The isotherm parameters of models were obtained and tabulated in Table 1. According to Table 1, the values of correlation coefficient (R 2 ) obtained from the Langmuir isotherm (0.9905) are higher than those from the Freundlich isotherm (0.9796) and Temkin isotherm (0.9328), indicating that the Langmuir model for experimental data could well describe the adsorption process, which reveals the monolayer coverage of diclofenac on the surface of the adsorbent. The maximum amount of adsorption capacity (q m ) was 52.91 mg/g. A similar observation was reported by Sun et al. (Sun et al. ). Furthermore, the comparisons of maximum adsorption of DF onto different adsorbents are listed in Table 2. The result shows that the q m value obtained was higher than those reported from many recent studies.

Kinetic study
Adsorption kinetic models describe the rate of the adsorption process. To provide information about the adsorption mass transfer mechanism of the adsorbent, two wellknown adsorption kinetic models were used. The experimental kinetic adsorption data were evaluated by pseudofirst-order and pseudo-second-order kinetic models (Chaudhry et al. ; Cholico-González et al. ).
The linearized mathematical form of the pseudo-firstorder kinetic model is represented by Equation (6): log (q e À q t ) ¼ log q e À K 1 2:303 t Here in Equation (6), q e (mg/g) and q t (mg/g) represent the amount of MB diclofenac adsorbed at equilibrium and time t (min), respectively. K 1 (min À1 ) is known as the pseudo-first-order adsorption rate constant. Values of q e and K 1 may be calculated from the straight-line plot log(q e À q t ) against t.
The linearized mathematical form of the pseudo-secondorder kinetic model can be expressed as Equation (7): Here in Equation (7), K 2 (g.mg À1 min À1 ) shows the rate constant of pseudo-second-order model. The q e and K 2 were calculated from the straight-line plot t q t versus t.
In this work, the pseudo-first-and second-order kinetic models have been applied to ascertain the adsorption mechanism. As outlined in Table 3, the R 2 values calculated from pseudo-first-order and pseudo-second-order kinetic models were 0.9979 and 0.9971, respectively, indicating that the pseudo-first-order kinetic model provides the better fit.

Thermodynamic study
The influence of solution temperature on diclofenac removal by Ni 0.5 Zn 0.5 Fe 2 O 4 MNPs was tested and results are shown in Figure 9. The adsorption capacity was reduced while increasing solution temperature from 298 to 328 K.
The thermodynamic parameters including Gibbs free energy variation (ΔG ), enthalpy variation (ΔH ) and entropy (ΔS ) can be obtained by the following expression (Bhatti et al. ): Enthalpy (ΔH ) and entropy (ΔS ) can be estimated from Equation (10). The obtained thermodynamic parameters are given in Table 4. The negative value of ΔH reveals that the adsorption process is an exothermic process. The negative value of ΔS can be ascribed to decreased the randomness of drug molecules while they adsorbed on the Ni 0.5 Zn 0.5 Fe 2-O 4 MNPs. Also, the negative value of ΔG indicates a spontaneous process.   Pseudo-first-order q e,calc (mg/g) 9.72 k 1 (min À1 ) 0.0853 R 2 0.9979 Pseudo-second-order q e,calc (mg/g) 11.32 k 2 (g.mg À1 min À1 ) 0.0061 R 2 0.9971