Optimization for removal efficiency of fluoride using La(iii)–Al(iii)-activated carbon modified by chemical route

Abstract A simple chemical route was adopted for the preparation of adsorbent to remove fluoride in the simulation solution. In this study, commercially activated carbon (CAC) was used as the raw material, HNO3 (20 wt%), La(NO3)3·2H2O (0.05 mol/L), and Al(NO3)3 (0.10 mol/L) were employed as modification reagents to successfully prepare the CAC adsorbent loaded with the elements La and Al (La(iii)–Al(iii)–CAC). The modified adsorbent was characterized by scanning electron microscopy, energy-dispersive spectroscopy, and Fourier-transform infrared spectroscopy measurements. The effects of variables such as solution pH, adsorption time, initial concentrations, and volume of solution were investigated on adsorption efficiency by the response surface method (RSM). Under optimum conditions, specified as a pH value of 6, an adsorption time of 1.2 h, an initial concentration of 30 mg/L, a solution volume of 25 mL, and the value of R F can reach at 74.11%. From the RSM analysis, the pH value was a significant factor in the adsorption process and the order for these key factors was as follows: pH > adsorption time > solution volume. Meanwhile, there were interactions among these factors. Chemical modification had an important role in pore structure generation and functional group of adsorbent to improve the adsorption efficiency. The removal performance of adsorbent on simulation solution of fluoride also showed the feasibility of adsorbent to be applied in industrial purposes.


Introduction
Fluoride is a very important trace element for our human body because it has great significance for the body's growth and development [1,2]. But it has harmful effects such as endocrine disorders [3], the formation of fluoride plaque [4], and even bone cancer [5] when the intake is excessive. Given that the source of fluorine is mainly drinking water, many people are drinking water with an excess amount of fluoride [6][7][8]. Fluoride pollution attracted great attention because it has become a worldwide problem, so it is very much necessary to carry out the research on defluoridation.
There are several methods to remove the fluoride ion (F − ), such as limestone precipitation method, ion exchange method, membrane separation method, and adsorption method. Among these strategies, adsorption method is one of the most effective and commonly used methods as it is environmentally friendly [9]. Various adsorbents have been used in the literature for defluoridation of aqueous solution such as activated carbon (AC), activated alumina, zeolites, alumina-based Zr-MOFs, iron-based adsorbents, or some biosorbents. Theoretically, they can be used as adsorbents for fluoride removal because of their unique physical or chemical composition [10][11][12]; however, there are some disadvantages such as highly pH-dependent in regeneration, material loss in adsorption process, and a gap between adsorption properties and desired effect by these strategies. AC is supposed to be a sustainable material for elimination of pollutants from drinking water.
Recently, to improve the adsorption capacity of AC, a number of metal oxides were employed to remove the anionic contaminants [13]; Fe 2 O 3 , ZrO 2 , SnO 2 , MnO 2 , and MnFe 2 O 4 can be good choice due to their biocompatibility and chemical stability [14][15][16][17]. Meanwhile, some modified methods of graphene oxide can be referred from previous studies [18][19][20]. Lanthanum is a kind of a rare-earth element, and some compounds of lanthanum were reported to be excellent absorbents of fluoride [21], such as nano-adsorbent of Al-Fe-La [22] and triplemetal composite of Fe-Mg-La [23]. Oxides and hydroxides of aluminum have also been reported to remove fluoride effectively with comparatively low cost [24][25][26]. Thus, a novel La-based and La-Al-based composite or a modified adsorbent will be a good choice for removing fluoride. However, few studies on La-Al adsorbents were reported.
In this work, the cheap AC, the low-cost reagent, and an experimental design technique were combined to explore an easier to handle, faster, more efficient, and cost-effective method to treat the simulation solution of fluoride. RSM is a statistical method used for optimization of the process and evaluation of the significance of the variables and their effects on the response [27]. The results of this study confirm the great potential and applicability of adsorbents for the removal of other ions from industrial wastewater.

Materials
Commercially activated carbon (CAC) was employed as a raw material in this experiment. The properties of CAC are listed in Table A1 in the Appendix. After being ground and screened until the particle size was less than 0.425 mm, the CAC will be washed five times and dried.

Methods and design
Preparation of La(III)-Al(III)-CAC included two steps: the acid modification process (I) and the La(III)-Al(III) loading process (II). A specific mass fraction of nitric acid (HNO 3 ) was involved in process I. A total of 100 mL of HNO 3 and CAC (20 g) were mixed in a flask, which was placed in a water bath at a constant temperature of 75°C for 2 h. Then, the CAC was washed to neutral and dried at 100°C for 3 h and the acid-modified activated carbon (MCAC) was obtained. La(NO 3 ) 3 ·2H 2 O and Al(NO 3 ) 3 were employed to load La(III) and Al(III) on MCAC. A 10 g MCAC was stirred in 100 mL of mixed solution of La (NO 3 ) 3 ·2H 2 O and Al(NO 3 ) 3 for 24 h. Then, the solid product was carbonized in a tube furnace at 300°C under a N 2 atmosphere for 2 h, and after cooling to room temperature, the preparation of La(III)-Al(III)-CAC was completed. La(III)-Al(III)-CAC will be used as an adsorbent for fluoride. The La(III)-Al(III)-CAC was put into the simulation solution of fluoride for a certain time period to perform the adsorption process. Illustration of La(III)-Al(III)-CAC preparation and adsorption processes is shown in Figure 1.
An experimental design for the preparation of La (III)-Al(III)-CAC was required to reduce the experimental cost [28]. Factors adopted here were solution concentration (mass fraction) of HNO 3 Table 1. The orthogonal array designed by the JMP method (a kind of statistical analysis software designed by SAS company) is listed in Table 2. This reduced the experimental number from 3 4 = 81 to L 15 (3 4 ) = 15.

Calculation of R F
Ion-selective electrode method was explored to calculate the concentration of F − , and a standard curve can be obtained by this method, which is detailed in the Appendix. According to the standard curve, R F can be calculated by the following equation: where R F , C 0 , and is the removal rate of F − (%), concentration of F − before and after adsorption processes, respectively (mg/L).
Optimization for removal efficiency of F −  407

SEM-EDS analysis of adsorbent
Scanning electron microscopy (SEM) was adopted to show the micromorphology of samples and the corresponding energy-dispersive spectroscopy (EDS) images to prove the successful loading of La(III) or Al(III) on La(III)-Al(III)-CAC, which were prepared under optimized conditions. As shown in Figure 2, there were more and denser pore structures on MCAC, indicating that the modification of HNO 3 was beneficial for generating the pores. The SEM images of La(III)-CAC and La(III)-Al(III)-CAC are shown in Figure 3. There were many oval pores in Figure 3a which were marked as 1. The surface of La(III)-CAC was coarse, and finer porous structures were marked as 2. However, the pore structures of La(III)-Al(III)-CAC were changed in Figure 3b. The pore structure marked as 1 became a little more regular circle, the finer pore size marked as 2 became larger, and the pores distributed more uniformly, which were suitable for better adsorption capability. The presence of elements La(III) and Al(III) were detected from the samples in Figures 3a1, a2 and b1, b2. The weight percentage of La(III) was 0.86%, which increased obviously to 2.55% after the modification of La (NO 3 ) 3 ·2H 2 O and Al(NO 3 ) 3 . More La(III) content in adsorbent will show better adsorption performance, which was probably caused by the synergistic effect of the two elements in the modification process. According to FT-IR analysis of La(III)-Al(III)-CAC, the number of oxygen-containing functional groups on adsorbent was increased, and La(III) and Al(III) were successfully loaded on CAC and existed in the form of oxide. The reaction equations that may be involved in the modification process are as follows: (4)   The other parameters of adsorption process were fixed as follows: the solution volume was 30 mL, the dosage of La(III)-Al(III)-CAC was 1 g, the adsorption time was 1.4 h, and the pH value was 6. R F was calculated after the end of the adsorption process, as shown in Figure 5. R F was improved at the first increase in the initial concentration; it can reach the value of 53.82% when the concentration was 30 mg/L, which was also the maximum value. R F decreases continuously with the increase in concentration, this is because the limited adsorption capacity of adsorbent, which can't absorb more ions [29,30], therefore, the maximum concentration that the La(III)-Al(III)-CAC can be loaded was 30 mg/L.

Effect of adsorption time
The adsorption process means the contact process, and all transfer phenomena such as adsorption are inseparable from contact; therefore, the adsorption time is a fundamental parameter. The other parameters of adsorption process were fixed as follows: the initial concentration was 30 mg/L, the solution volume was 30 mL, the dosage of La(III)-Al(III)-CAC was 1 g, and the pH value was 6. The effects of the adsorption time on R F are shown in Figure 6. Clearly, R F of solution increased quickly at the value of 71.2% in the initial stages of the contact period until 1.0 h, and then gradually increased to 71.6% with a prolonged contact time of 1.4 h until equilibrium was reached. There are usually three steps for the adsorption process as follows: external surface adsorption stage, internal surface adsorption stage, and adsorption reaction stage. The adsorption reaction stage is a chemical reaction, which is not involved here, so the decisive step is the other two physical adsorption stages. The increase in R F in initial stages before 1.0 h was contributed to the external surface adsorption process, as all of the active sites on the La(III)-Al(III)-CAC were vacant at the first stage of adsorption process and the concentration of solution was also high, so it is easy to   increase the amount of adsorption. However, after that stage, the adsorption process mainly depends on the internal surface adsorption. Few surface active sites were available, which lead to the observation of small increase in the uptake of R F . Hence, the suitable time to reach equilibrium was selected as the adsorption time for further experiments.

Effect of solution volume
The solution volume is also a very important factor of adsorption process because of the relationship between the capacity of an adsorbent and a given initial concentration of solution. Studies were conducted to learn the effect of La(III)-Al(III)-CAC on NaF solution with the concentration of 30 mg/L, which was confirmed by experiments before. Seven different solution volumes were employed in the study (10,25,40,55,70,85, and 100 mL), the dosage of La(III)-Al(III)-CAC was 1 g, the adsorption time was 1.4 h, and the pH value was 6. The R F obtained from experiments is shown in Figure 7.
There was a sharply increasing trend as the solution volume increased from 10 to 25 mL, then continued to decline, and the maximum R F was 70.6% at the solution volume of 25 mL. The higher removal percentage in the present experiment contributed to the greater availability of exchangeable sites and larger surface area [31].
More solution volume provided more contact area of La (III)-Al(III)-CAC in order to promote the effective adsorption process. However, because of the limited adsorption capacity of La(III)-Al(III)-CAC, R F decreased when the solution volume was more than 25 mL. Therefore, 25 mL was used as the optimal solution volume of the adsorption.

Effect of pH
It is necessary to analyze the effect of pH as a single factor because the pH value of an aqueous solution is one of the most important factors during the adsorption process, especially when it can affect the existence form of fluorine. The effect of pH was investigated for pH values of 3, 4, 5, 6, 7, 9, and 11. The pH value was adjusted during the experiments by adding a few drops of HCl (1 mol/L) and NaOH (1 mol/L). All other parameters in the adsorption process were kept constant (the dosage of La(III)-Al (III)-CAC was 1 g, the initial concentration of NaF solution was 30 mg/L, and the solution volume was 25 mL). The effect of pH value on R F is shown in Figure 8. Figure 7 shows that an uptake of R F significantly increased with an increase in pH value in the solution; it can achieve the peak point at 72.1% when the pH value was 6, and then followed by a downward trend until the pH value was 11.
In the pH range of 0-7, fluorine exists in the form of F − , HF, and HF 2− , and mainly in the form of HF when pH < 3; therefore, the concentration of F − will not increase at the condition of pH > 3 theoretically. However, there is similarity in chemical bond between F − and OH − , and the OH − will affect the adsorption of F − , making the higher adsorption efficiency to be obtained at a higher pH value and the suitable pH was 6.

Design and result
The response surface method (RSM) was adopted to optimize the adsorption process and get a higher R F .  Experimental design was carried out by the custom design method in JMP software, and the response item was R F . Three parameters, including adsorption time (E), pH (F), and solution volume (G), were studied and each parameter was set according to single-factor experiment results, as shown in Table 4. In addition, R F is also shown in the last column of this table, which will be discussed later.

Analysis of model
The stepwise regression method was adopted to analyze the experimental data in Table 4. Analysis of variance and significance test for the regression model are listed in Tables 5 and 6, respectively. The smaller the P value and the larger the F value, the more significant the model. In Table 5, the value of R 2 is 0.9488, the F value is 12.3555, and the P value is 0.0031, which indicated that the stepwise regression model was suitable for the analysis of experimental data. Hence, the model can be used to analyze and optimize the adsorption process. When the P value of index was <0.0100, the factor can be interpreted as the extremely significant factor. In Table 6, the significant factor was the pH value, as well as the F (pH) squared and the E (adsorption time) squared, which means the interaction of three factors among them was not a simple linear relationship. The order for these key factors was F > E > G (solution volume).

RSM optimization and verification
The RSM method was employed to analyze the experimental data in Table 6. The quadratic regression surface model was established after eliminating the nonsignificant items. The model for removal rate (R F ) can be calculated by the obtained quadratic regression equation given by: where X 1 , X 2 , and X 3 are equal to (adsorption time: 1.2)/ 0.2, (pH value: 6)/3, and (solution volume: 25)/15. In order to get the optimization conditions for adsorption, interaction effect among three factors was explored by three-dimensional RSM of the quadratic polynomial regression equation, as shown in Figures 9-11.
The convex surfaces can be seen on the response surface in Figures 9-11, which means that the model has a stable maximum point in the test range. When the adsorption time was 1.2 h and the pH value was 6, the R F was 74.05%; when the pH value was 6 and the solution volume was 25 mL, the R F was 74.00%; when the adsorption time was 1.2 h and the solution volume was 25 mL, the R F was 73.72%. The maximum R F value was 74.05% by the prediction of Equation 2 and RSM. Therefore, the optimal adsorption process conditions are determined as the adsorption time was 1.2 h, the pH value was 6, and the solution volume was 25 mL. Three parallel adsorption experiments (No. 17, No. 18, and No.   The higher removal efficiency of this adsorbent in simulation solution of fluoride confirmed the great potential and applicability of adsorbent for the removal of other ions from industrial wastewater.   Appendix Calculation of R F Electrode is made of LaF 3 single crystal, and the potential measured by the electrode is linear with the logarithm of the F − concentration. To draw a standard curve, the standard solution and total ionic strength buffer solution (TISBS) are necessary. Standard solution was NaF (100 mg/L); TISBS was prepared from three reagents such as Na 3 C 6 H 5 O 7 ·2H 2 O (12 g), NaNO 3 (58 g), and C 2 H 4 O 2 (58 mL); and HCl (1 mol/L) and NaOH (1 mol/ L) were used to adjust the pH at 5-6. Pipette a certain volume of standard solution and TISBS, insert the electrode into the test solution, and record the potential value when the digital stable, as shown in Table A2, and the standard curve is shown in Figure A1.    Figure A1: Standard curve of F − .