INTRODUCTION
Nowadays, with the rapid growth of population and increased industrialization, water pollution poses direct negative effects to the environment as well as on human life. Various water pollutants, including dyes, heavy metal ions, organics, and so on are of interest [1]. Among these, due to the toxicity and non-degradability of almost all dyes, the study of their removal from wastewater is very important. Methylene blue (MB), as one of the commonly known cationic soluble dyes, is used in different applications to protect materials and facilitate the sale or make goods more expensive. The most common MB effects are headache, dizziness, nausea, vomiting, dyspnea, and chest discomfort [2]. Therefore, eliminating the risk of trapping it in wastewater is of great interest to protect the health and safety of the environment [3]. In recent years, various processes such as coagulation/flocculation, chemical oxidation adsorption, and ultrafiltration have been widely used to remove dyes from wastewaters. Among these processes, adsorption has been investigated as the cost-effective method due to the high removal efficiency of dyes removal in low concentration and easy process [4]. Present electrospun nanofiber membrane with promising properties, such as fine diameters, high porosity, tunable pore size, and high surface to volume ratio is one of the most important recent researches [5]. The preparation of electrospun nanofibers from various polymers and biopolymers has been considered to utilize in different applications [6-11]. Chitosan is recognized as a natural and versatile biomaterial that the existence of hydroxyl and amino groups in its structure, making it a suitable biosorbent for organic dyes and heavy metal ions by chelation or electrostatic interaction [12].  Mahmoodi et al. carried out a study on anionic dyes removal from colored wastewater using crosslinked electrospun PVA/CS blend nanofiber by glutaraldehyde vapor. Their results showed that maximum uptake capacity was 151, 95, and 114 (mg/g) for Direct Red 80, Direct Red 81, and Reactive Red 180, respectively at pH 2.1 [13]. Other researchers studied the effect of ZIF-8@chitosan/PVA electrospun nanofibers dose on the removal of Malachite green dye and they found that the removal percentage increased with increasing the adsorbent dose of ZIF-8 and from results, the optimum of its dosage was 0.03 g/L [14]. Li et al. investigated the removal of acid blue-113 from an aqueous solution using electrospun chitosan nanofibers [15]. They found that the diameter of chitosan fibers has a considerable effect on the removal efficiency and by decreasing of nanofiber diameter from micro to nanoscale, the adsorption capacity increased from 412 to 1377 mg/g. Furthermore, providing conditions for optimal use of nano-sized powder adsorbents with high porosity can increase the efficiency of removing pollutants from industrial effluents. Among these nanomaterials, CeAlO3 nanoparticle powders synthesized by combustion technique, are often spherical and have an average particle size of 14,068 nm. This polycrystalline powder is single-phase and due to its magnetic properties as well as its highly porous structure (Fig 1), it can be used in the process of removing contamination from aqueous solutions [16]. In this study, to obtain the optimum and smooth electrospun nanofibers structure, effective parameters such as voltage, distance to the collector, and feed rate of the electrospinning process were identified by the design of experiment (DOE). Then, the removal of methylene blue dye using prepared electrospun nanocomposite adsorbents that cross-linked by citric acid was investigated for the first time. The effects of pH value, contact time, and concentration on the process were studied. Also, sorption kinetics and isotherms properties of the process were obtained.  

MATERIALS AND METHODS
Materials
In this study, polyvinyl alcohol with a molecular weight of 49,000 was obtained from Fluka. CeAlO3 was prepared at Arak University (16). Triton x-114 was supplied by Applichem. Chitosan with a molecular weight of 100 KD and other high purity chemicals and solvents were purchased from Merck Company.  

Instrumentation
We used the electrospinning instrument (Nano Azma Co) for preparing nanofibers. The morphology of electrospun nanofibers was observed using a scanning electron microscope (SEM) (Beyond Technology, Korea) with a gold coating. Average diameters of nanofibers were measured by Digimizer.5.4.9 software. Fourier transform infrared spectroscopy (FTIR) analysis (Perkin Elmer) was used to investigate the structural changes using the KBr method. X-ray diffraction studies of nanocomposite samples were studied by X-ray diffractometer (Philips X-3064 Pert Pw) with Cr K-alpha-1 radiation source at ambient temperature. The pH of MB dye solution was adjusted by using HCl and NaOH and measured with a pH meter (Mi 180 Bench Meter). The residual MB concentration in the aqueous solution before and after the adsorption process was measured using a UV/ Vis spectrophotometer (Perkin-Elmar-lambada25).

Batch adsorption experiments
All experiments were done at room temperature to study the effects of experimental conditions such as pH, contact time, and the initial methylene blue concentration. The process was done by using appropriately diluted 500 ppm stock solutions. At the beginning of the experiment, the pH value of each dye test solution was adjusted by the addition of HCl and NaOH solutions. In each step, 0.01 g of the nanofibers was added to the samples and shaken with the rotating speed of 150 rpm (PIT10LO). For equilibrium studies, nanocomposite was put into dye solutions with initial concentrations ranging from 400 to 600 mg/l. The equilibrium time of the process was estimated by drawing samples at desired time intervals. MB solution was separated from the adsorbent by centrifugation (model), and the residual dye concentration in the solution was determined by a UV/Vis spectrophotometer. The amount of MB adsorbed per unit mass of nanocomposite (adsorption capacity) was calculated in the following equation:

where qe is the equilibrium uptake (mg/g), C0 and Ce are the initial and the equilibrium dye concentration (mg/L), respectively, V is the volume of the solution in contact with the nanocomposite (mL), and  is the mass of the added adsorbent (g) [17].
Experimental design
In this study, controllable factors including voltage, distance from the collector, and feed rate were designed in three levels. Central Composite Design (CCD) was used to determine the optimum experimental conditions for the synthesis of smooth surface and uniform nanofiber mats. According to Table 1, PVA/CS fibers were prepared uniformly only under one condition at voltage 18 kv, distance to collector 10 cm, and feed rate of 0.3 cm3 /h with the diameter ranging about 381 nm.

Synthesis of the nanocomposite
Approximately 2 wt.% of chitosan was dissolved in concentrated acetic acid. 12 wt.% PVA solution was prepared by dissolving in deionized water. The solutions were prepared at a ratio of 1:1 at room temperature. Briefly, a clear aqueous solution of cerium nitrate, aluminum nitrate, and glycine based on the ratio stoichiometry of reactions [16] was prepared and then heated on a magnetic stirrer at 90 ° C for 1 hour. After a few seconds of the combustion reaction, all gases were released and CeAlO3 nanoparticles were prepared in the powder.  the PVA/CS blend was divided into two parts and About 1 wt.% of CeAlO3 and 1% Triton x-114 were mixed in one of the PVA/CS parts to get PVA/CS/ CeAlO3 blend while the other blend was used as a blank. Furthermore, 0.002 g of citric acid was added to both solutions to cross-link the fibers and stirred overnight. The solutions were used for electrospinning under optimum conditions. Then the prepared nanofibers dried and were cured in an oven at 140 oC for 2 h to initiate chemical crosslinking.  

RESULTS AND DISCUSSION
Characterization of the adsorbents
Scanning electron microscopy (SEM) was used to investigate the prepared optimal nanofibers before and after cross-linking, and diameter distributions of the fibers are shown in Figs (1-4). It is seen that the net and improved nanofibers before cross-linking by citric acid are very smooth and uniform. However, after cross-linking, the surface of the fibers appears slightly rough, which may be due to the evaporation of existing water in the mats. Then, 100 fibers were analyzed with Digimizer software and the size distribution histograms of the nanofibers were drawn. The diameters of the net and improved nanofibers before and after cross-linking were 381, 420, 412, and 450 nm, respectively. The viscosity of them increased by adding the citric acid and CeAlO3 nanoparticles to the polymeric solutions. Therefore, as can be seen, the diameter of the average nanofibers is larger than the net PVA/CS nanofibers. 

Fourier Transform Infrared Spectroscopy
In the spectra of CS/PVA before crosslinking (Fg.6a) the adsorption peak at 3312 cm−1 in it, which was wider than that in the spectra of cross-linked CS/PVA, the reason being that CS/PVA has more O–H groups than cross-linked CS/PVA. The band in the range of 3000–3700 cm−1, represented the overlapped stretching vibration of –OH and N–H in the CS and –OH in the PVA. Furthermore, for the cross-linked CS/PVA membrane (Fig. 6b), the peak of –OH/N–H of CS shifted from 3312 cm−1 to 3351 cm−1, which was due to the formation of hydrogen bonding between the hydroxyl in PVA and amino in CS and hydroxyl groups in citric acid. Fig. 6 presents a broad and strong peak at 3351 cm-1 assigned to -OH and N-H stretching, indicating the presence of hydroxyl and amino groups, respectively. 2939 cm-1 attributed to C-H stretching of Alkanes. 1734, 1567 are attributed to the C = O aldehyde, C = C stretching bonds of alkene, respectively. C-H scissor and bending belongs to 1426-1290 cm-1. Also, 1511 and 832 cm-1 peaks belong to N-O stretching and C = C vibration bonds of alkene, respectively. 1328 and 1090 cm-1 are belonging to the C-O alcoholic and C-N stretching vibration of the chitin–chitosan and protein fractions. Fig. 6b indicates that the bands that have been observed at 3351, 2939, 1734, and 1090 cm-1 for adsorbent before process shifted to 3362, 2941, 1736, 1094 cm-1, respectively (Fig. 6c). The significant changes in the wavenumbers of these specific peaks indicate that these bonds were effective in the methylene blue adsorption process [18].   
As can be seen in the X-ray diffraction (Fig. 7), raw chitosan and polyvinyl alcohol powders revealed sharp peaks at (2θ=30.165) and (2θ =29.81) angles, respectively, indicating that the raw materials are crystalline. However, with the blend of these materials together, the intensity of the peaks and their crystallinity decreased. the reduction of the crystallinity of the nanofibers may be attributed to the hydrogen bonding interaction between molecules in the prepared mats. Furthermore, there were almost no obvious effects on the crystallinity of the nanofibers by the cross-linking process. With the addition of the CeAlo3 nanoparticles to PVA/CS, the peaks are strongly increased (2θ= 29.5). After the adsorption, as can be seen, the PVA/CS/CeAlo3 nanocomposite change in the crystallinity structure was less than the net PVA/CS nanofibers. 

Effect of pH
The effect of pH of the solution on MB sorption using the nanocomposite from an aqueous solution was investigated in an initial pH range of 2 to 10. Experiments were done with constant concentration (400 mg / l) containing 0.01 g of the adsorbent at different pHs, at room temperature, and for 30 min on a shaker with 150 rpm. according to Fig. 8. As can be seen, with increasing pH, the adsorption capacity of both adsorbents has increased and has reached its maximum value at pH 10. It can be concluded that, at low pH values and acidic conditions of the solution, the dimethylamine groups in MB ions and amine groups in the adsorbent (chitosan) are protonated. Therefore, the electrostatic attraction between the positively charged adsorbent and the positively charged dye decreases. At higher pH conditions of the solution, hydrogen bonding and electrostatic interactions between MB ions and functional groups of nanocomposite during the adsorption process increases and lead to the improved adsorption of dye from aqueous solutions. The results are consistent with the other MB removal studies [19- 23].
Effect of contact time
In order to obtain the equilibrium time for the adsorption of MB, the process was investigated at different times Fig. (9). The operating conditions including contact time (180min) and optimum pH were constant for each adsorbent. As can be seen, at the beginning of the process, due to the availability of active sites and functional groups on the external surface, the adsorption of MB increased considerably with increasing contact time. Bypassing time, uptake capacity reduced with decrease of the interior active sites and it remained constant thereafter. According to the obtained results, the equilibrium time was considered to be 45 minutes for improvement and 90 minutes for the net adsorbent.

Kinetics of sorption
Proper prediction of batch sorption kinetics is necessary to optimum operating conditions and the design of industrial sorption systems. The adsorption kinetics was studied for a better understanding of the dynamics of MB removal and obtaining predictive models that allow estimations of the amount of adsorbed dye with the contact time. This information could be used for the scale-up to a larger system. The pseudo-first-order, pseudo-second-order kinetics, and intraparticle diffusion models were thus applied to the experimental data obtained (Figs. 10-12). The pseudo-second-order kinetics model, unlike the pseudo-first-order kinetics model, predicts adsorption behavior at all times of the adsorption process [24, 25]. The linear form of different kinetics is expressed in Equations 2 to 4 as follows:

=-            (2)

=                                                              (3)

                      (4)

Where, K1 is the rate constant of the pseudo-first sorption kinetic (min)-1, K2 is the rate constant of the pseudo-second sorption kinetic ((g)/ (mg min)), kid is intraparticle diffusion constant (mg/g), and C is intraparticle diffusion constant (mg/g).  Then, using the line equations obtained from plotting and matching it with the adsorption kinetics equations, the adsorption kinetics parameters were calculated (13). According to the results, intraparticle diffusion is not the rate-controlling step. The correlation coefficient factor, R2, calculated from the pseudo-second-order model fits the experimental data very well with an amount of 0.99 for both adsorbents. The qe value derived from the second-order model (769.23) for net nanofiber is more comparable with the experimental qe value (714.61) when compared with the first-order model. The coefficient of determination (R2) of the intraparticle diffusion model reflected a poor fit to the experimental equilibrium data for both adsorbents compared to the other models. The adsorption of MB by improved and net nanofibers can thus be considered following second-order kinetic models. 

Effect of initial concentration
The effect of MB initial concentration in the range of 400 to 600 mg/L was investigated. According to the results, the uptake capacity value increased from 816.05 to 906.05 and from 714.61 to 800 by increasing initial concentration for improved and net nanofibers, respectively. This could be due to the increased driving force of mass transfer at higher concentrations of MB.
 
Isotherm of sorption
Isotherm models can describe the relationship between the amount of dye adsorbed on the external surface of the adsorbent and the dye content remaining in the solution.  The results of the experiments were analyzed using two well-known isotherm models of Langmuir and Freundlich. The Langmuir isotherm is based on monolayer adsorption on the surface of the adsorbent, while the Freundlich model investigates the multilayer adsorption on the surface of the adsorbent. Equations 5 and 6 represent the linear form of the adsorption isotherm [26].

=                                      (5)   
    

ln qe= ln kf + lnce                                                                           (6)      

Where, ce is Equilibrium concentration, (mg / L) qe adsorption capacity at equilibrium, (mg / g) Q ° equal to the maximum amount of adsorbed dye per gram of adsorbent in the Langmuir equation, kf Freundlich isotherm model capacity, and n is the amount of adsorption capacity in Freundlich equation. Using these equations and matching them with the equations of adsorption isotherm, the parameters of the adsorption isotherm were calculated (Table 2). Furthermore, if the  value (the dimensional factor) be in the range of 0 <  < 1, the adsorption process will favorable (26). 

=                                                        (7)

According to the obtained high values ​​of R2, it can be found that there is a good correlation between the obtained data and the adsorption isotherm. Therefore, the adsorption equilibrium fits the Freundlich and Langmuir isotherms on the improved nanofibers. The levels of n-1 and R in both adsorbents are in the desired range between 0 and 1, and therefore the Langmuir and Freundlich models are suitable for investigating the adsorption process. A similar result can be seen in the research of Staroń et al. [27]. Also, the value of n is higher than 1 representing favorable adsorption. Thus, the prepared nanocomposite has high efficiency in removing MB dye. 
Table 3 shows the comparison of the maximum adsorption capacity of various adsorbents to adsorbents used in this study. Both adsorbents seem to have lower uptake capacity than calcium alginate–bentonite–activated carbon and larger adsorption capacity than other adsorbents. 

CONCLUSIONS
In this study CS/CeAlO3 based nanocomposite, a new adsorbent was synthesized by electrospinning process and characterized. We used the Design expert software to find the best conditions for the synthesis of nanofibers with smooth and uniform structure. According to the obtained results, fibers were prepared uniformly under only one condition at voltage 18 kv, distance to collector 10 cm, and feed rate of 0.3 cm3 /h. Then, the removal of MB dye from solution using modified nanofibers was investigated in batch technique. The optimum process conditions for MB removal were pH 10, contact time 45 min, initial MB concentration 400 mg/L. The maximum adsorption capacity of MB was 817.81 and 714.61 mg/g onto improved and net nanofibers, respectively. The kinetics of MB adsorption on improved nanocomposite follows the pseudo-second-order model. Furthermore, the equilibrium data of it followed the Freundlich and Langmuir isotherms. The results revealed that this nanocomposite was effectively used for the removal of MB dye from the aqueous solution.

CONFLICTS OF INTEREST
The authors declare that there are no conflicts of interest regarding the publication of the manuscript.