Synthesis of a new magnetic adsorbent using green tea leaf extract and its application in phenol removal by RSM method

Green synthesis of metal nanoparticles (NPs) has attracted a lot of attention in recent years because of lower costs and being environmentally friendly in comparison to the chemical methods. One of the most considered methods is production of metal NPs using organisms. Among these organisms, plants seem to be the best candidates and suitable for the biosynthesis of large-scale NPs. The produced NPs by plants are more stable and their synthesize are faster than microorganisms. In this research, the green synthesis of iron oxide magnetic NPs (Fe3O4-MNPs) has been performed using aqueous extract of green tea (Camelia sinensis) as a reducing and capping agent. Evaluation of properties and morphology of Fe3O4-MNPs is performed using scanning electron microscopy, X-ray diffraction, infrared spectroscopy, and vibrational magnetometer. The synthesized NPs were crystalline in nature and spin-shaped with a mean diameter of ≤ 20 nm and also has super-magnetic properties. The systematic study of the synthesis process was performed using Design Expert 10 software to determine the relationship between the four process variables, namely iron concentration, extract volume, time, and temperature. The square model was significant for the response variables. In the final, synthesized Fe3O4-MNPs were applied for magnetization of activated carbon that then used for efficient removal of phenol (up to 98%) from aqueous media. Various parameters affecting the removal process such as pH, extraction time, and adsorbent amount were optimized by using response surface methodology.


Introduction
Removal of organic compounds from industrial wastewater is difficult or impossible with biological treatment processes. Phenols, as an important class of organic compounds are structurally similar to conventional materials such as herbicides and insecticides. They are widely used for commercial production of a wide range of resins, including phenolic resins, which are used as structural materials for epoxy resins, adhesives, and polyamides for various applications [1]. Phenols can be also formed by the breakdown of bisphenol A. Among the 126 priority pollutants the designated by the US Environmental Protection Agency, phenol is ranked 11th and recommends reducing phenol in wastewater to a levels lower than 1 mg L -1 [2].
There are many methods for removing phenol from water and wastewater such as photo-Fenton process [3], ion exchange [4], electrochemical oxidation [5], reverse osmosis [6], catalytic degradation [7], and adsorption [8][9][10][11][12]. Treatment with activated carbon is an effective method for removal of phenol from wastewater because of its large surface area, micro-porosity, high adsorption capacity, high purity, and availability [13]. However, it is very difficult to separate activated carbon from solutions after its treatment. One attractive solution to this problem is magnetization of activated carbon and as a result it may be reused or recycled.
Therefore, materials contain superparamagnetic NPs, such as magnetic activated carbon (MAC) as adsorbents, have great importance for the removal of organic pollutants [14]. There are several reports on the synthesis of magnetic nanoparticles (MNPs) by using different reducing agents such as hydrazine [13][14][15][16], dimethyl formamide (DMF) [9], sodium borohydride (NaBH 4 ) [17], and carbon monoxide (CO) [18]. However, such methods are suffering from disadvantages such as using toxic chemicals as reducing agents, using toxic organic solvents, low production rate, and high cost [19]. For this reason, it is very desirable to devise alternative 'green' methods for the preparation of nanomaterial that reduce or eliminate toxic substances. Green synthesis of nanoparticles by inactivated plant tissue and plant extracts has the advantages such as using environmentally friendly, non-toxic, and safe reagents [20].
In this research, extract of green tea was used for magnetization of activated carbon. Green tea extract unquestionably is a safe, green, and environmentally friendly material, especially for herbals and medicines. The aqueous extract of green tea contains gallocatechin gallate (GCG, [ 45%), epigallocatechin gallate (EGCG, 36 to 40%), epigallocatechin (EGC 0.7-2.3%), and epicatechin (EC 0.5-2.2%) (Fig. 1S), which have antibacterial and antioxidant activity [18]. It is known that these phenolic compounds can be aid to reduce of Au, Ag, and Fe cations to NPs [21].
Response surface methodology (RSM) is used to independently minimize the absorption process. This method can also be employed to study the effect of various variables and interactions at different levels. Previously, researchers used the traditional experiments 'one variable at a time' to determine the individual effect of various factors on the adsorption process. However, factorial experimental design can be used to provide a large amount of information and reduce the number of experiments, time, and total research costs. The most important advantage of this technique is that the effect of individual parameters as well as their relative importance are obtained and also the interaction of two or more factors can be ascertained [22]. RSM has four main steps, including designing of experiments, model fitting, model verification, and determining the optimal conditions. The design of experiments, like the central composite design (CCD), is useful for RSM because it requires a small number of experiments. Optimizing of any operation means determining the optimum value of different parameter affecting that process which provide maximum desirable output. Traditionally, optimization in analytical chemistry was carried out by monitoring the influence of one factor at a time on an experimental response [22]. Its major disadvantage is that it does not include the interactive effects among the variables studied. RSM consists of several experimental techniques dedicated to the evaluation of relationship between a group of controlled experimental factors and measured responses based on one or more criteria [23]. RSM can be used to evaluate the effects of individual parameters, the interaction of variables, and the optimum conditions for responses [24].
In the current study, adsorption of phenol from aqueous solutions onto MAC was statistically analyzed using version 10 of the Design Expert software. The four-factor CCD was used to investigate the effect of the selected parameters on the phenol removal efficiency by sorbent. pH (X1), temperature (X2), initial concentration of phenol (X3), and the adsorbent dosage (X4), were selected as the independent variables, and the phenol removal efficiency was considered as the response (dependent variable). The variables were coded according to the Eq. 1: where X i , and X 0 are the coded value, and the real value at the center point of an independent variable, respectively, and DX is the step change value. RSM is a combined mathematical and statistical technique based on the fit of a polynomial equation (empirical models) to the experimental data. This method generates a polynomial function for response relating to the variables involved [25]. In doing so, it deals with the variables only at specific levels (mostly -1, 0, 1). RSM generates an experimental design for model preparation. An experimental design is a specific set of experiments defined by a matrix composed of the different levels combination of the variables studied [26]. Different methods of determining the response surface require different experimental design. The simplest model which can be used in RSM is based on a linear function (Eq. (2)): where k is the number of variables, b 0 is the constant term, bi represents the coefficients of the linear parameters, x i represents the variables, and e is the residual associated to the experiments. The next level of the polynomial model should contain additional terms, which describe the interaction between the different experimental variables [27]. In order to determine a critical point (maximum, minimum, or saddle), it is necessary for the polynomial function to contain quadratic terms according to the Eq. (3) [27].
where b ij represents the coefficients of the quadratic parameter. A second-order model can significantly improve the optimization process when a first order model suffers lack of fit due to interaction between variables. Sometimes the mathematical model obtained after fitting does not satisfactorily describe the studied experimental domain under study. So, the quality of the fitted model is evaluated by the application of analysis of variance (ANOVA). In the present work, micro-porous activated carbon was magnetized using green tea extract and used for the adsorption of phenol from aqueous solution by RSM methodology.

Materials
All chemicals used were of analytical grade and were used without further purification. Iron (II) chloride (FeCl 2 ), sodium hydroxide, and HCl were purchased from Merck (Darmstadt, Germany). A stock solution of phenol (1000.0 mg L -1 ) was prepared by dissolving 1.000 g of the analytical grade reagent of phenol in 1.0 L distilled water. Test solutions with desired concentrations were prepared by diluting this stock solution. pH of solutions was adjusted to the required value by using either 0.1 M HCl or 0.1 M NaOH solutions. Activated carbon was prepared from Glycyrrhiza glabra residue in our laboratory [12].

Instrumentation
Absorbance of phenol was measured by using a DR-5000 UV-Visible spectrophotometer (Hach Company, Canada). The JEOL (Japan) 2010F UHR was used for SEM imaging of gold-sputtered samples. A TENSOR II FTIR spectrometer from Bruker Optics (Germany) was used for FT-IR analysis in the range of 400-4000 cm -1 , employing KBr disk. The X-Ray diffractometer Model Bruker AXS D8 advance (Germany) with radiation at k Cuka1 = 1.5406 Å was employed to assesses the crystalline structure. The XMF-104 was used to determine the chemistry of a sample by measuring the fluorescent X-ray emitted from the sample. VSM Lake Shore 4700 (USA) was used to assesses magnetic properties.

Preparation of the green tea extract
The green tea extract was prepared by heating 60.0 g L -1 green tea until boiling. After cooling for 1.0 h at room temperature, it was filtered two times through a Whatman no-1 to obtain a clear extract. The green tea extract was then stored at 4°C for further use.

Optimization of MNPs
The CCD, the most popular RSM, and Design Expert 10 statistical software were used to determine the optimum conditions in the co-precipitation process and obtaining the best nanoparticles. Four experimental effective parameter, i.e., iron concentration (denoted as A), extract volume (denoted as B), time (denoted as C), and temperature (denoted as D) were considered as the response. Three level of the factors were used in optimization by RSM (Table 1). Experimental factors were selected based on the most significant effect on the efficiency of the controlled precipitation method. Previous experimental were the basis of selection of factors levels.

Synthesis of magnetic activated carbon
Co-precipitation method was used for the preparation of MAC. The output of the experimental design software under optimal conditions for the synthesis of iron NPs; 1 g of activated carbon was added to 100 mL of iron (II) chloride solution in optimal molarity and stirred for 30 min. Temperature was raised to the optimum value and green tea leaf extract was added dropwise to it. The pH of the solution was adjusted to 6 and the resulting suspension was stirred for one hour. Finally, the precipitate was separated by centrifugation at 6000 rpm for 10 min. Precipitate was washed several times with distilled water and dried in an oven at 100°C and finally calcinated in an oven at 450°C for 2 h.

Adsorption procedure
In order to optimize the selected effective parameters by RSM, 22 absorption experiments were performed and the effect of the following parameters were studied. For this purpose, pH, initial concentration of phenol, time, and the adsorbent dosage on the phenol removal was studied. Each experiment was carried out in a 100 mL canonical flask installed on a shaker (160 rpm) at 25°C. An experiment with the same initial concentration of phenol and without adding MAC was always performed in parallel in order to detect any phenol loss during the experiments. The reaction initiated by adding different amounts of MAC to 30 mL solution containing 30 to 1000 mg L -1 phenol. The pH was adjusted at the range of 3.0-12.0. The effect of MAC dose on the adsorption of phenol was investigated by adding different amounts of MAC (10 to 150 mg) into different concentrations of phenol solution (based on the design of experiments) and removal efficiency was measured at contact times between 0 and 120 min. After reaching equilibrium, the mixed solution was separated from the absorbent by a magnet. By using Eq. 4, the residual phenol could be calculated.
where C o is the initial concentration of the phenol in solution (mg L -1 ) and C is the final concentration of the phenol (mg L -1 ).

Preparation and properties of MNPs
For preparation of magnetic nanoparticles (MNPs), experiments were designed based on reaction time, iron concentration, amount of green tea extract, and reaction temperature. The amounts of nanoparticles obtained in each experiment were used to evaluate the RSM effectiveness [28]. The used variables in the design of the experiments and the initial results (based on weight) for determination of influencing factors MNPs synthesis were shown in Table 2. The Statistical analysis of the data to investigate the main effects of the parameters individually and combinational on the weight of nanoparticles produced are presented in Table 3 [28]. These results show that only two factors, the concentration of iron and the amount of green tea extract, have a significant effect on the yield of nanoparticles (P \ 0.05). The correlation coefficient of these two variables shows their positive effects on the response. The time and reaction temperature alone had no effect on the production of iron nanoparticles. P value of iron concentration and amount of green tea extract indicate that they are the most important factors in the efficiency of iron nanoparticle synthesis.
The results of ANOVA test (Table 3) shows the good fit of this newly designed model due to the high value of F and the low value of P value. Also, the assesses of coefficient R-Square (= 0.9982) and the adjusted coefficient Adj R-Square (= 0.9941) show the accuracy of the model. The normal plot of residuals diagram is shown in Fig. 1 and indicate that the residues are normally distributed in a straight line.
Based on the SEM image of iron nanoparticles (Fig. 2S), the particle size of Fe 3 O 4 is approximately 23 nm.
The FT-IR spectrum of the prepared extract from green tea leaves is shown in Fig. 3S. The strong and high-intensity band at 3380 indicate the presence of  XRD is a common technique for determining the morphology and crystal structure of nanoparticles. This technique is used to determine the metallic nature of particles. The XRD pattern of the synthesized Fe 3 O 4 NPs using green tea is presented in The magnetic hysteresis curve of synthesized Fe 3 O 4 at room temperature was shown in Fig. 6S. The sample did not show any hysteresis and the magnetic curve was completely reversible so that it shows the supra-magnetic behavior of the prepared nanoparticles. The saturation magnetization of the synthesized Fe 3 O 4 was about 40 emug -1 .
The actual mechanism nanoparticle synthesis by living organisms is not yet clear; however, some studies showed that the present phenolic compounds in the plant products are responsible for the production of metallic iron NPs [29][30][31][32]. The probable mechanism for synthesis of Fe 3 O 4 NPs is showed in Fig. 3. The reduction of Fe(II) to Fe(0) and formation of the oxidized product of polyphenolic compound (Ar = O) (Fig. 3) is the first step. The standard potential values of polyphenolic compounds (Ar-OH) and iron make this reduction process feasible [33]. Since the Fe(0) is highly unstable, it is immediately oxidized to Fe(III). The next step is the precipitation of Fe(OH) 3 and Fe(OH) 2 .

Assesses of magnetization of activated carbon by iron nanoparticles
The crystalline structure of the MAC was shown in Fig. 7S. In X-ray diffraction pattern obtained for the  [28][29][30]. Additional, carbon's amorphous nature confirmed by one wide peak focused on around 2h = 24° [15]. Figure 8S shows ED-XRF spectra of the MAC, and as can be seen existence of C, O, N, Fe, and S elements in the MAC with weight ratios of C (82.27%), O (7.45%), N (4.41%), Fe (3.75%), and S (2.12%) was confirmed.
FT-IR analysis was carried out to determine the present functional groups onto the activated carbon   (Fig. 9S) shows vibration bands at 3500 cm -1 that attributed to stretch of O-H in polyphenols; 1623 cm -1 that assigned to C=C aromatic bonds. The presence of Fe-O bond at 696 and 641 cm -1 confirms the formation of iron oxide onto the surface of activated carbon by using green method. The absorption peaks observed around 1138 cm -1 is attributed to the C-O-C and aromatic C-OH stretching modes, respectively [34,35]. The stretching vibrations at 1623 cm -1 and 3500 cm -1 corresponding to polyphenols in the MAC. The occurrence of stretching vibrations of polyphenols in MAC confirms that the surface is magnetized by polyphenols in green tea extract.
In order to study the morphology of MAC and how iron oxide nanoparticles spread onto the surface of activated carbon, transmission electron microscopy (TEM) imaging was employed at different magnifications. These images are shown in Fig. 10S(a-c) at the specified scales of 20, 40, and 90 nm, respectively. The black and gray dots on these images are related to magnetic oxide particles alone or together [36,37]. As can be seen in Fig. 10Sa, the magnetic Fe 3 O 4 NPs are well located on the activated carbon and are mostly bonded together, but the particles that are single and non-bonded are approximately less than 20 nm in size. Also, their cohesive particles are made up of smaller nanoparticles. As can be seen in Fig. 10Sb, while most nanoparticles sticking to each other, they are also well dispersed on the surface of activated carbon and this dispersion more clearly can be seen in Fig. 10Sc. With respect to these images, it can be stated that the magnetic Fe 3 O 4 NPs are well deposited on the activated carbon sample.
The SEM of the prepared MAC and activated carbon were shown in Fig. 4. It clearly shows the formation of the magnetic Fe 3 O 4 NPs on the surface of activated carbon. The size of the magnetic Fe 3 O 4 NPs is between 12 and 20 nm.
The magnetic properties of the MAC nanocomposite were investigated with a VSM that shown in Fig. 11S and showed that the MAC nanocomposite has good magnetic properties in the magnetic field. After the preparation, a test with a magnet showed that the nanocomposite was superparamagnetic and completely attracted to the magnet. The saturation magnetization (Ms) of the nanocomposite was 21 emug -1 .

Adsorption experiments
The design of experiment for the phenol adsorption using Design Expert software is shown in Table 4 together with its effective removal percentage of phenol from the prepared stock solutions. The results obtained were utilized to further analyze the removal percentage of phenol from the prepared stock solutions by the aim of Design Expert that induces the optimum conditions for removal of phenol from the prepared stock solutions. ANOVA from the Design Expert (Tables 1 and 2) was used to study the attained optimized parameters and the successful removal percentage. Results from the ANOVA for the removal percentage of phenol from the prepared stock solutions are listed in Table 5. Based on these, the Fischer F-test value can be correlated to the mean square of the regressed model that leads to the comparison of the mean square of the residuals (errors). Efficiency of the model is validated as the F value increases. In contrast, the p value or lower  þ 0:45199D þ 1:5478E À 004AB þ 0:19643AC À 6:085E À 004AD À 0:018133BC þ 1:1785E À 005BD þ 0:1086CD À 2:7258A 2 À 2:047B 2 À 87: where ''A'' represents the coded value of pH, ''B'' represents the coded value of the concentration, ''C'' is the amount of adsorbent dosage, and ''D'' represents the coded value of the agitation time or the contact time. In addition, the coefficient of a factor represents the effect of that factor while the coefficients of double factors represent the effect and interaction between those two factors.
The obtained values from the developed model compared with the experimentally values and illustrate in Fig. 5 and indicates that the developed model for removing phenol is acceptable.
Three-dimensional diagrams (Fig. 6) were plotted to observe the relationship between optimizing conditions and the phenol removal percentage.

Effect of pH
Adsorption of phenol on the adsorbent depends upon the nature of the adsorbent surface and the species distribution of phenol. The surface distribution mainly depends upon the pH of the system. This is attributed to the change of charge of surface of adsorbent with the change in pH value. As can be seen in Fig. 6, removal efficiency increases when the pH increases from 3.0 to 8.0, and dropped sharply with further pH increase. The pKa value of phenol is 9.9 and the pH of the zero-point charge of the MCA is about 6.2 [38]. At lower pHs, the surface of the MAC would be protonated and as a result negatively phenolate ions could be stronger attraction via electro-static force. Non-ionized phenol molecules would also be attracted, possibly, by physical force. At high pHs, OHions would compete with the phenol molecules for sorption sites. However, sorption of excess of OH-ions could convert a positively surface into a negatively surface and as a result,

Effect of carbon dose
The removal of phenol increased with increasing adsorption capacity of MAC (Fig. 6). The increasing of phenol removal may be due to increase of adsorbent surface area and availability of more vacant sites. The removal percentages of the phenol increase when the adsorbent dosage is increased. Because, with the increment of the adsorbent, the availability of the surface area for the adsorption process increased, as a result, more phenol can be adsorbed onto the surface of the MAC [40,41]. Figure 6 shows that the power of MCA for phenol removal increases at higher contact times. It was also observed that phenol uptake is rapid in the beginning and thereafter it proceeds at a slower rate and finally becomes constant. This is probably due to larger surface area of the MAC that is available at the beginning for the adsorption of phenol. As the surface adsorption sites become occupied, 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. Figure 6 shows that the removal efficiency decreases with increasing the initial concentration of phenol. The initial concentration provides the necessary driving force to overcome the resistances to the mass transfer of phenol between the aqueous and solid phases. The increase in initial concentration also enhances the interaction between phenol and the MAC. Therefore, an increase in initial concentration of phenol enhances the adsorption of phenol. This is due to the increase in the driving force of the concentration gradient produced by the increase in the initial concentration of phenol [11,42].

Phenol removal from wastewater sample
To assess the applicability of the developed protocol for real samples, an industrial wastewater sample was collected and assessed by using ICP-OES ( Table 6). As no phenol was found in it, to evaluate the matrix effect, 50 mg L -1 of phenol was added to the sample and treated with the MAC under optimum conditions. Analysis of the supernatant showed that phenol and some other cations were successfully removed (Table 6).

Conclusion
The Fe 3 O 4 nanoparticles has been successfully synthesized by a green method using green tea extract. RSM was used to find out the best conditions in the production of Fe 3 O 4 nanoparticles. The optimum parameters for green synthesis of Fe 3 O 4 nanoparticles are iron chloride = 67 mmol, solution volume = 12.5 mL, contact time = 92 min and temperature = 53°C. This green procedure has many advantages such as high synthesis yield and ease of operation as well as being economically. No toxic reagents or surfactant is required in this protocol, as a result, the process become very environmentally friendly. The characteristic of MAC was assessed by SEM, XRD, FT-IR, and VSM. The VSM results revealed that the synthesized MAC are paramagnetic with a saturation magnetization of 21 emug -1 at room temperature and consequently could be separated from solution by a magnet. All obtained results indicated that the MAC can be used as an economical and effective adsorbent for phenol removal from wastewater samples. The RSM method was successfully used to develop a mathematical model for phenol removal. The value of R 2 [ 0.99 for the obtained quadratic model indicates the high correlation between the experimental response value and the predicted response value by the mathematical model. The obtained optimum operating conditions for phenol removal by using numerical optimization of RSM are agitation time = 65 min, phenol concentration = 500 mg L -1 , and MAC dose = 0.08 g (per 30 mL solution). Under these conditions, the phenol removal efficiency is 98%.