Analytical-grade chemicals [Ni(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, NaF, NH₃·H₂O, C₆H₅Na₃O₇·2H₂O, NaNO₃] were used for NiAl-LDHs preparation. Total ionic strength adjustment buffer solution (TISAB) were prepared by NaNO₃ and C₆H₅Na₃O₇·2H₂O at pH ranging from 5 to 6. Deionized water was used to prepare all solutions in this study.

NiAl-LDHs were prepared by the co-precipitation method. In a typical procedure, A total of 0.05 mol Ni (NO₃)₂·6H₂O and 0.025 mol Al (NO₃)₃·9H₂O were dissolved in 200 mL deionized water. The solution was magnetically stirred for 10 min in air at room temperature. Subsequently, the solution was titrated with an alkali solution of NH₃·H₂O (10%) under vigorous stirring at 80˚C at such a rate that the pH of the reaction mixture was maintained at 10.0 (±0.5). The resulting suspension was further stirring for 1 h, and then transferred to a water bath aged at 90˚C for 24 h. The solid was then recovered and washed with deionized water and finally dried at 80˚C for 12 h, the sample was denoted as NiAl-LDHs.

The NiAl-LDHs were treated by NTP (CTP-2000K plasma generator, Nanjing Suman electronic co., LTD, China) before calcinations. The power of NTP treatment was 60 W, and the duration time was 10 min, then the modified material was obtained, the sample was denoted as NiAl-LDHs (P).

X-ray diffraction (XRD) patterns were collected using a D/MAX-2500/PC diffractometer with CuKα radiation (λ = 1.54056 Å) from 5˚ to 80˚ (2θ) at a scanning rate of 5˚/min. For the determination of the average crystallites size the Scherrer formula [24] has been used (Equation (1)):

where D_v is the volume weighted crystallite size, λ the wavelength (CuKα), β is the full-width at the half-max- imum intensity, θ the diffraction angle corresponding to the (00l) diffraction line. FTIR were recorded on a Nicolet iS 10 spectrometer over a range from 500 to 4000 cm⁻¹. The thermal behavior of the samples was determined by thermogravimetric analysis (TGA) using a DTG 60 instrument in the temperature 30˚C - 800˚C with a heating rate of 10˚C·min⁻¹. N₂ adsorption-desorption was performed with a NOVA-4000e system utilizing Brunauer-Emmett-Teller (BET) for surface area calculation [25] . The pore size distribution and pore volume were obtained using the Barrett-Joyner-Halenda (BJH) method from the adsorption and desorption isotherms, respectively. Before analysis, the samples were pretreated at 80˚C for 12 h.

The fluoride adsorption experiments were carried out by batch experiments. All adsorption tests were carried out by a Fluoride ion selective electrode (pF-1) and reference electrode (232).

The adsorption kinetics of fluoride by the NiAl-LDHs was studied at a solid-to solution ratio of 2 g/L. The amount of fluoride ion adsorbed per unit mass of the adsorbent (q_e), (q_t) and the removal rate (R_e) were evaluated using the following equations [26] :

where C_o is the initial concentration of fluoride ions in solution (mg/L), C_e is the equilibrium concentration (mg/L), C_t is the concentration of fluoride ion at time t of adsorption (mg/L), q_e is the equilibrium adsorption capacity (mg/g), q_t is the amount adsorbed per gram of adsorbent at time t (min), V is the volume of solution (L), and m is the mass of the sorbent (g).

The XRD patterns of the NiAl-LDHs and NiAl-LDHs(P) are shown in Figure 1, one can see that the NiAl-LDHs(P) shows diffraction peaks at 2θ of 11.38˚, 23.16˚, 34.76˚, 39.30, 46.80, 60.88˚, 62.14˚, which are assigned to the (003), (006), (012), (015), (018), (110) and (113) diffraction, respectively. All of the reflections can be directly indexed to a typical rhombohedral phase hydrotalcite (JCPDS No.54-1030), and no diffraction peaks of impurities were discerned, indicating the high purity of the sample. Moreover, the features of the peaks are sharp, narrow and symmetrical with stable baseline, suggesting that the as-prepared NiAl-LDHs were highly crystallized. NiAl-LDHs (P) has higher intensity and sharper peaks which compare with the NiAl-LDHs samples, indicating that the application of NTP process promotes the dispersion of hydrotalcites particles. After modification, the shape of peak did not changed, which indicated NTP treatment process is physical effect did not changed the structure of NiAl-LDHs. The structure data determined from XRD measurements are reported in Table 1. The value of the unit cell parameter a₀ is equivalent to the mean distance between adjacent cation center in the close-packed cationic brucite-type sheets, thus its value can be correlated with the average value of radii of the metal cations in those layers [24] . The position of the (110) reflection allows the value of the lattice parameter a₀ to be determined, since a₀ = 2d (110). The basal spacing c₀ of LDHs equals the thickness of one brucite-type layer plus one interlayer. The observed reflections in Figure 1 can be indexed in a three-layer 3R polytype; in this case the basal repeat distance c₀ is c/3. The lowest reflection (003) was used to calculate c₀ (003) and c (003), respectively, as reported in Table 1. The Scherrer equation was used to estimate the crystallite size, from the width of the (110) and (003), results shown in Table 1. Crystallite size valued calculated from the (003) line show a decrease when modified by NTP in sample NiAl-LDHs (P) compared with the NiAl-LDHs. It was indicated that the NTP treatment induced the production of high crystalline LDHs with relatively small particle size; these were benefit for adsorption process. After adsorption of fluoride the basal spacing c₀ and crystallite size of NiAl-LDHs and NiAl-LDHs (P) became smaller than before. This indicated that F⁻ has been adsorbed in the gallery of NiAl-LDHs.

The FTIR spectra of the NiAl-LDHs and NiAl-LDHs (P) are given in Figure 2. The strong peak at 3440 cm⁻¹ was attributed to the stretching vibrations of -OH groups associated with the interlayer water molecules and hydrogen bonding. The H-O-H bending vibration of the interlayer molecule water was located at 1624 cm⁻¹ [27] [28] .

In addition, two intense peaks around 705 and 1387 cm⁻¹ due to the (out-of-plane deformation) and (symmetric stretching) vibrations of the interlayer carbonate anions. The bands in the range of 500 - 800 cm⁻¹ are attributed to O-M-O, and M-O-M lattice vibrations (M = Ni and Al) [16] [29] .

The TG-DTA curves of the synthesized samples were presented in Figure 3. The weight loss occurs essentially in two steps: the first one at a low temperature (30˚C - 150˚C) corresponds to the removal of weakly adsorbed water on the external surface of the crystallites as well as water intercalated in the interlayer galleries; the second one in a wide temperature range of (ca. 150˚C - 600˚C) involves dehydroxylation of the layers as well as removal of volatile species (CO₂) arising from the interlayer carbonate anion [16] . Furthermore, the similar TGA curve was observed of NiAl-LDHs(P), suggesting that NTP treatment is physical modification, the characters of the materials are not changed. According to the TG analysis, the total weight loss was 56.03%, 41.40% for NiAl-LDHs and NiAl-LDHs (P) during the thermal decomposition process, respectively.

Figure 4 shows the N₂ adsorption-desorption isotherms of NiAl-LDHs and NiAl-LDHs (P) (Figure 4(a)) and their corresponding pore size distribution (Figure 4(b)). The materials exhibited type IV isotherms according to the classification of Sing et al. [30] , which are typical for mesoporous materials. Isotherms also exhibit an H3 type hysteresis at high relative pressure, which is typical for aggregates of plate-like particles giving rise to slit-shape pores [19] . This kind of hysteresis is typical for the presence of open large pores, which allow easy diffusion of the reactants through the materials. As seen from Figure 4(b), the curves are quite broad with small mesopores at 4 nm.

The smaller mesopores reflect the presence of pores within nanosheets, while larger mesopores can be associated to the pores formed between stacked nanosheets [19] . The specific surface area of the NiAl-LDHs increased from 37.091 to 39.878 m²/g and the pore volume from 0.160 to 0.178 cm³/g by treating with NTP. It was indicated that the NTP treatment induced the production of large specific surface area.

The adsorption kinetics is important for adsorption studies because it can predict the rate which fluoride is removed from aqueous solutions and provide valuable data for understanding the mechanism of sorption reactions [19] . Kinetics adsorption of F⁻ on NiAl-LDHs was investigated in the presence of 2.0 g/L of adsorbent at 25˚C. Figure 7 shows that the adsorption rates were considerably fast under all the concentrations. The adsorption process almost finished within 30 min, and no significant change was observed from 30 to 90 min. This observation may be associated with the existence of a large number of vacant sites on the surface that are available for adsorption during the initial stage. However, the remaining vacant surface sites are less available for adsorption with time due to the repulsive forces that occur between the adsorbed and free molecules [28] . The phenomenon that the adsorption reached equilibrium quickly indicated that the adsorption of F⁻ on NiAl-LDHs was mainly attributed to chemical sorption or surface reaction. The binding of F⁻ anions on the surface of NiAl-LDHs may be ascribed to the strong electrostatic interaction between F⁻ and NiAl-LDHs layered structures. The other mechanism for F⁻ adsorption is due to ion exchange between the intercalated anions of NiAl-LDHs with F⁻ ion. The removal efficiencies were found to be 93.4%, 85.7%, 74.2% at the initial F⁻ concentrations of 60.0, 80.0 and 100.0 mg/L, respectively. According to Halajnia and Oustan [32] the percentage removal at equilibrium depends on the type of LDHs as well as the initial concentration. On the other hand, seen from the removal time figure, with the increase of the initial F⁻ concentrations from 60 to 100 mg/L, the total amount of F⁻ adsorbed increased from 28.02 to 37.12 mg/g. It could be attributed that more targets of F⁻ could provide higher driving force to facilitate the ions diffusion from the solution to NiAl-LDHs, and more collisions between F⁻ ions and active sites of the samples. Similar phenomena have also been observed in the literature [33] .

The experimental data were modeled by three adsorption kinetic models, pseudo-first-order model (Equation (8)), pseudo-second-order model (Equation (9)), and intra-particle diffusion model (Equation (10)) [31] [32] [34] .

Pseudo-first-order model: (8)

Pseudo-second-order model: (9)

Intra-particle diffusion model: (10)

where q_e and q_t are the amount of F⁻ adsorbed (mg/g) at equilibrium and at time t (min), respectively and k₁ is the pseudo-first-order rate constant (min⁻¹). Values of k₁ are calculated from the plots of ln (q_e − q_t) versus t for the adsorbent samples; k₂ is the pseudo-second-order rate constant (g·mg⁻¹·min⁻¹). Based on the experimental data of q_t and t, the equilibrium adsorption capacity q_e and k₂ can be determined from the slope and intercept of a plot of t/q_t versus t; k_p is intra-particular diffusion rate constant (mg·g⁻¹·min^−0.5). If intra-particular diffusion is rate-limited, then plots of adsorbate uptake q_t versus the square root of time (t^0.5) would result in a linear relationship.

From the slope and intercept of the straight line obtained, the kinetic parameters for the removal of fluoride were determined, as compiled in Table 2 and the plots of the linearized from the pseudo-second-order model are shown in Figure 8. The goodness-of-fit for the models were assessed by calculating the coefficient of determination (R²). It can be seen from Table 2, the R² values obtained from pseudo-first-order model are relatively small and the experimental q_e values do not agree with the values calculated from the linear plots, while the experimental q_e values were more close to the calculated q_e values from the pseudo-second-order model (R² ≈ 1). This indicated the better performance of the pseudo-second-order model to describe fluoride adsorption on the NiAl-LDHs. This results similar to as found in most studies [20] [35] [36] .

The rate-limiting step in the adsorption process is the “surface reaction”. The maximum percent removal could reach 93.41%, and the equilibrium concentration was 3.956 mg/L for the initial fluoride concentration of 60 mg/L. With increasing concentration from 60 to 100 mg/L, the rate constant of pseudo-second-order, k₂ decreases from 18.10 × 10⁻² to 7.656 × 10⁻² g·mg⁻¹·min⁻¹, this fact was supported by Figure 7. A similar phenomenon is observed in adsorption of methyl orange onto LDHs [30] . The adsorption kinetics of fluoride in NiAl-LDHs can be divided into two stages: a fast adsorption step and a slow adsorption step. In order to determine the rate-limiting the diffusion mechanism in the adsorption of fluoride on NiAl-LDHs can describe by the intra-particle diffusion model. It can be seen from Table 2 that the adsorption conducted in multiple steps. At first, is the external surface adsorption stage, which was driven by initial fluoride concentration different. The second is the gradual adsorption stage, where intra-particle diffusion is the rate limiting step. The third is the final equilibrium stage where intra-particle diffusion further slows down due to the fluoride concentration decreased in the solutions. These results are consistent with literature [30] [33] . It is clear from the values of R² in Table 2 that F⁻ ion adsorption kinetics on NiAl-LDHs sample was better described by pseudo-second-order model and intra-particle diffusion model.