Preparation of aluminium-hydroxide-modified diatomite and its fluoride adsorption mechanism

Scanning electron microscopy and energy spectrum analysis

The SEM images show the morphological structures of DA and Al-DA before and after adsorption (Fig. 1). DA and Al-DA have disk-like microstructures²⁹ with sur-faces containing both large and small pores, that is, DA and Al-DA have unique multi-level pore structures. The main component of DA and Al-DA is silica, which has a large specific surface area, good thermal stability, and is a natural green material for use as a water treatment agent with a porous structure³¹. The micrographs show that before adsorption, the DA surface is smooth with a distinct pore structure, whereas modification with aluminium hydroxide makes DA coarse and loose because of the formation of amorphous aluminium hydroxide colloids³². After adsorption, the surface pore structure is covered over for DA and completely covered over for Al-DA, which indicates that F⁻ reacts with Al³⁺ to form nanoscale precipitates²². The results of the EDS analysis (Fig. 2) show that the content of elemental Al increased from 3.96 to 12.74% after DA was modified with aluminium hydroxide, indicating that Al adhered effectively to the modified DA surface. After adsorption, the content of elemental Al decreased from 3.96 to 1.36% for DA and from 12.74 to 2.03% for Al-DA, which fully confirmed that fluorine preferentially combined with Al to form aluminium precipitates during adsorption, thereby decreasing the Al content.

XRD analysis

The surface mineral composition and crystallinity of the materials before and after adsorption were analyzed by XRD (Fig. 3). In the DA and Al-DA patterns, the wide diffraction peaks at approximately 22.0°, 26.0°, and 50.0° mainly correspond to amorphous SiO₂, and the diffraction peak at approximately 35° mainly corresponds to amorphous Al₂O₃, indicating that the material is polycrystalline²⁹. It has been re-ported that amorphous materials may be good adsorbents because of a large specific surface area and numerous active sites³³. Many Al(OH)₃ peaks and NaCl peaks appear in the XRD pattern of Al-DA, indicating the successful modification of DA by aluminium hydroxide. After adsorption, Na₃AlF₆ peaks appear in the DA pattern, and Na₃AlF₆ and AlF₃ peaks appear in the Al-DA pattern, whereas the characteristic peaks of NaCl are absent in the Al-DA pattern, which indicates the participation of NaCl in the adsorption process. It has been demonstrated that in the presence of excess sodium fluoride in the reaction solution, the generated aluminium fluoride combines with sodium fluoride to form a NaAlF₄ intermediate, which is subsequently converted to cryolite complexes by further adsorption of sodium fluoride³⁴. This result confirms the XRD mapping results.

Infrared analysis

Figure 4 shows the FTIR spectra of DA and Al-DA before and after adsorption: peaks at 3418, 1635, 1096, 791, and 538 cm⁻¹ appear in the spectrum of DA spectrum before adsorption, and peaks at 3630, 3449, 1637, 1094, 913, 793, and 538 cm⁻¹, appear in the Al-DA spectrum before adsorption. The strong and broad band centered at 3418 cm⁻¹ is due to the stretching vibration of the adsorbed water hydroxyl group (O–H) and the surface hydroxyl group, the vibrational peak at approximately 1635 cm⁻¹ is probably from bound water or the surface hydroxyl group; the peaks at 1096 cm⁻¹ and 538 cm⁻¹ correspond to siloxane groups (Si–O–Si–) and an Al–O absorption band, respectively; and the strong oscillations at 791 cm⁻¹ may be attributed to inorganic Al salts^35,36,37. The original absorption peak in the DA spectrum is shifted in the spectrum of DA modified with aluminium hydroxide, confirming the successful modification of DA. The shift of the band at 3418 cm⁻¹ in the DA spectrum to a higher frequency at 3623 cm⁻¹ in the DA spectrum after fluoride absorption is caused by fluoride bonding and has been previously reported³⁸. Another noticeable change in the spectra of DA and Al-DA before and after adsorption is the increase or decrease in the intensity of bending vibrations of specific peaks because the highly electronegative fluoride may have an inductive effect on the respective groups that leads to a blueshift, and the formation of hydrogen bonds leads to a redshift and broadening of the spectral band. The shifts and changes of these peaks indicate the interaction of fluoride with the respective groups²⁹. The new peak at approximately 1170 cm⁻¹ in the spectra of DA and Al-DA with adsorbed fluoride may be due to the formation of Al-F bonds⁶. The IR spectra show that the formation of a new bonding electronic structure by surface complexation with F⁻ is one of the main mechanisms for the adsorption of F⁻.

Zeta potential analysis

The zeta potential of the material surface plays a very key role in the adsorption process, which reflects the surface charge properties of the material under different pH conditions, and also reflects the surface properties of the material. To obtain the zero charge point of the material, we studied the potential change of the material under different pH values. The results are shown in Fig. 5. In the range of pH 3–11, the zeta potential of the two adsorbents decreased linearly with the increase in pH, and the pH_zpc of DA and Al-DA were 9.84 and 10.61, respectively. When pH < pH_zpc, the material surface is positively charged, and high electronegativity fluorine ions should be adsorbed to the material’s positively charged surface. On the other hand, when pH value > pH_zpc, the ad-sorbent surface is electronegative, resulting in electrostatic repulsion between the ad-sorbent surface and electronegative fluoride ions. Therefore, at a higher pH value, the adsorption of fluoride ions is reduced³⁹. Therefore, for the negatively charged fluoride ion, the surface of the two adsorbents is positively charged in a wide pH range, which can show a good adsorption effect, and Al-DA is better. Owing to the modification-induced formation of new aluminosilicate compounds on the diatomite surface, the pH_PZC of Al-DA is shifted to higher values. The partial substitution of Al atoms for Si atoms in silicates, owing to their different valences, generates an excess negative charge, which is compensated by the introduction of Na⁺, K⁺, Mg²⁺, or Ca²⁺ cations into the aluminosilicate crystal lattice. An increase in the Al/Si ratio leads to a decrease in the aluminosilicate ion sizes and an increase in the total negative charge per aluminum atom; as a result, the sorbent surface in solution adsorbs a larger number of protons and the pH_PZC is shifted to higher values³⁴.

Adsorption isotherms

To better understand the behavior and capacity of the adsorbents for fluoride, the Langmuir and Freundlich isotherm models were used to fit the experimental data. The Langmuir model is based on the assumption of uniform monolayer adsorption with no interactions between adsorbed species. By contrast, the Freundlich model is an empirical equation that describes a heterogeneous system with multilayer adsorption³⁸ and the formula is as follows:

where C_e is the mass concentration of F⁻ in the solution at adsorption equilibrium (mg/L); q_e is the quantity of F⁻ adsorbed at equilibrium (mg/g); Q_m is the maximum saturated adsorbed quantity (mg/g); K_L is a constant in Langmuir’s equation related to the heat of adsorption; K_F is a constant in the Freundlich equation related to the ad-sorption strength; and n > 1 for preferential adsorption, n = 1 for linear adsorption for, and n < 1 for nonpreferential adsorption.

The parameters of the isotherm model were obtained from Fig. 6 and are listed in Table 1. For the DA data, the Freundlich correlation coefficient (R² = 0.9591) was larger than the Langmuir correlation coefficient (R² = 0.8792), and the fitted curves yielded n values greater than 1, implying a strong affinity between the adsorbent and F⁻⁴⁰. Thus, the adsorption process is not simple physisorption but stable chemisorption involving the formation of chemical bonds. The fits to the data for the adsorption of F⁻ by Al-DA show that both models effectively reflect the adsorption process. However, visual inspection of the fitted correlations shows that the Langmuir isotherm model can simulate the adsorption process of F⁻ on Al-DA more effectively than the Freundlich model. The adsorption mechanism is unimolecular layer adsorption that occurs mainly via ion-exchange interactions and is dominated by chemisorption. The maximum monolayer saturation adsorption capacity Q_m is 45.8050 mg/g. Similar results were reported in a previous study⁶. In addition, a large K_F reflects a strong binding ability⁴¹: K_F for Al-DA is larger than that of DA, indicating that Al-DA has a higher binding ability for F⁻. A comparison of the n and K_F values for DA and Al-DA indicates that Al-DA has a higher adsorption capacity for F⁻ than DA.

Adsorption kinetics

To further elucidate the adsorption mechanism, including the rate determination step, three kinetic models, the quasi-first-order kinetic, quasi-second-order kinetic, and intraparticle diffusion models, were used to investigate the adsorption of F⁻ by the adsorbents and the formula is as follows:

where q_e is the adsorbed quantity at equilibrium (mg/g), q_t is the adsorbed quantity (mg/g) at the adsorption time t (min), and K₁ is a quasi-first-order kinetic removal efficiency constant (min⁻¹), K₂ is a quasi-second-order kinetic removal efficiency constant (min⁻¹), K_pi is an internal diffusivity constant, and C is a constant related to diffusion.

The fitting results presented in Fig. 7 and Table 2 show that the quasi-secondary kinetic model effectively fits the data with quite high R² values of 0.9998 and 0.9999. In addition, the theoretical adsorbed quantities obtained using the fitted quasi-second-order kinetic equations for DA and Al-DA data were closer to the experimental values than those obtained using the other models, indicating that the adsorption of F⁻ by DA and Al-DA is well described by the quasi-second-order kinetic model. This result suggests that chemical interactions between the absorbents and F⁻ control the adsorption process, chemisorption is the underlying mechanism. This type of chemical interaction usually corresponds to strong bonding between the surface functional groups of the adsorbent and the adsorbed species³⁷. Therefore, F⁻ may form chemical bonds in the monolayer on the DA and Al-DA surfaces⁴². Xu²⁹ also confirmed that the adsorption of fluorine by diatomite is effectively modeled by quasi-second-order kinetics.

The adsorption of solid adsorbents can proceed by two types of adsorption, as well external and internal diffusion. As the type of diffusion that controls the removal efficiency differs during different phases of adsorption, the intraparticle diffusion model describes the transport of the target analyte from the aqueous phase to the surface of the adsorbent, followed by diffusion to the interior of the porous particles⁴³. Therefore, the intraparticle diffusion model was used to study the characteristics of F⁻ adsorption, and the fitted results are shown in Fig. 8 and Table 3. The linear fit to the adsorption of F⁻ by DA and Al-DA consists of three segments with high individual R² values, where K_p1 > K_p2 > K_p3: the larger K_p is, the higher the diffusion rate is, indicating a two-step process for the adsorption of F⁻ by DA and Al-DA that conforms to the internal diffusion model. External diffusion occurs at the surface of the adsorbent, and the diffusion rate gradually decreases during the internal diffusion phase while ions diffuse into the interior of the adsorbent⁴⁴. The data can be specifically interpreted as follows: the initial sharp region corresponds to external mass transport across a boundary layer; the second segment corresponds to gradual adsorption, where intraparticle diffusion is the rate-limiting step; and the third segment corresponds to equilibrium, during which intraparticle diffusion slows down^45,46. Intraparticle diffusion is the rate-limiting step if the gradual adsorption segment passes through the origin. However, the fit to the data from this study does not pass through the origin, indicating that F⁻ adsorption on DA and Al-DA is controlled by both intraparticle diffusion and surface adsorption⁴⁶.

Effect of different influencing factors on the adsorption of F⁻

Effect of the solution pH on the adsorption

The pH of the initial solution affects the solid–liquid interface of the adsorbent and the speciation of metal ions in the solution⁴⁷. Therefore, it is crucial to investigate the effect of the initial pH on fluoride adsorption on DA and Al-DA. Therefore, adsorption experiments were carried out at different initial pH (4, 5, 6, and 7), and the results are shown in Fig. 9. The removal efficiency and adsorption capacity of F⁻ by both materials first increased with the pH and then decreased beyond a well-defined pH. For DA, the removal efficiency was 88.7% and the adsorption capacity was 22.2 mg/g at pH 6. For Al-DA, the removal efficiency was 97.0% and the adsorption capacity was 24.8 mg/g at pH 4. Subsequently, with the increase in pH, the removal efficiency and adsorption capacity of DA and Al-DA on F⁻removal efficiency and adsorption capacity gradually decreased with the increase in pH. Acidic conditions result in a high number of H⁺ on the adsorbent surface, resulting in electrostatic attraction between the positively charged adsorbent surface and F⁻. It has also been found that the removal efficiency and adsorbed quantity of F⁻ can be significantly increased by the formation of Al-F complexes^6,27. With increasing pH, F⁻ and OH⁻ compete for adsorption sites, where the higher the OH⁻ concentration is, the more significant the competitive adsorption effect is, leading to a decrease in the removal efficiency and adsorption capacity of F⁻⁴¹. The investigated materials have good fluoride removal performance over a wide pH range, but as a higher fluoride removal capacity can be obtained in acidic media than in alkaline media, the optimal pH for adsorption of F⁻ by DA and Al-DA are 6 and 4, respectively. Because Al-DA is modified with aluminum salts and some cations are introduced, resulting in an increased positive charge, the optimal pH for the adsorption of fluoride ions will be different from that of DA, and Al-DA has to show a better adsorption effect.

Effect of the dosing quantity of the adsorbent on the removal efficiency and capacity

Figure 10 shows the effect of using masses of the two adsorbents ranging from 1.6 to 12 g/L on F⁻ adsorption. The removal efficiency of F⁻ on DA and Al-DA increased with the dosage to maxima of 91.1% and 97.3% at a dosage of 4 g/L, respectively, and then started to decrease. Increasing the adsorbent dose corresponds to increasing the surface area and the number of adsorption sites available for the same quantity of fluoride ions. Therefore, the removal efficiency of F⁻ increased considerably with the dose during the initial stage^6,48. This finding is consistent with reports of the effect of the adsorbent dosage on fluoride removal^49,50. The subsequent decrease in the adsorption capacity resulted from an increase in the fixed initial fluorine concentration and the solid dose of the fixed solute load, which led to a decrease in the availability of fluoride ions per unit mass of adsorbent³⁶. Therefore, the analysis shows an optimal dose of 4 g/L for both DA and Al-DA, and higher adsorption of F⁻ by Al-DA than by DA.

Effect of coexisting anions on the adsorption of F⁻

Fluorinated natural water and wastewater typically contain a variety of anions, such as Cl⁻, NO₃⁻, SO₄²⁻, CO₃²⁻, and HCO₃⁻. During adsorption, these coexisting anions may compete with F⁻ for adsorption sites, thus reducing the removal efficiency of the adsorbent. The competitiveness for the adsorption of a coexisting anion depends on the relative concentration of the ion and its affinity for the adsorbent, which is intrinsically related to the ionic radius and charge⁴¹. Therefore, it is crucial to assess the effect of coexisting ions on the uptake of F⁻ for practical applications. Figure 11 shows the effect of the presence of Cl⁻, NO₃⁻, SO₄²⁻, CO₃²⁻, and HCO₃⁻ on F⁻ adsorption at different pHs. The highest removal efficiency was observed at pH = 4, and the interference capacity de-creased in the order SO₄²⁻  > HCO₃⁻  > CO₃²⁻  > Cl⁻  > NO₃⁻, which is consistent with the results of the previous studies^6,29. However, all the anions have a limited ability to inhibit the adsorption of F⁻. For example, at pH = 4, the adsorption of F⁻ reaches 89.8% even under the interference of SO₄²⁻. Therefore, it can be speculated that Al-DA has a high affinity for F⁻.

Fluoride removal from a real-life water sample

To examine the efficiency of the prepared adsorbent for the removal of fluoride from the real field sample, we collected water samples from a sewage treatment station of an enterprise in the fluorine chemical industrial park. The dosage was 4 g/L, the thermostatic adsorption experiment was carried out, and the effect of removing fluoride ions from actual wastewater was compared with that reported in other literature. The results are shown in Table 4. After the adsorption study, the fluoride concentrations were well below the specified limit according to WHO (< 1.5 mg/L). Compared with other adsorbents, Al-DA can reduce the concentration of fluoride ions to a relatively low safety value. Al-DA shows good advantages. Therefore, the Al-DA adsorbent developed in this study can be used to remove real fluoride-polluted water.

Source: Ecology - nature.com